First aid interventions by laypeople for acute oral poisoning

Summary of findings for the main comparison. SDAC versus no intervention for first aid in patients with acute oral poisoning

SDAC versus no intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (paracetamol or not specified) Setting: hospital setting Intervention: single‐dose activated charcoal (SDAC) Comparison: no intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with no intervention	Risk with SDAC
Incidence of mortality	No studies collected or reported this outcome
Incidence of adverse events	Control group: 0/236; intervention group: 4/240 (Peto OR 4.17, 95% CI 0.30 to 57.26)		—	476 (2 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain of the effect of SDAC on the incidence of adverse events.
Incidence and severity of symptoms of poisoning: incidence of clinical deterioration during stay in the hospital	—		—	451 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,d	The relative effect was not estimable due to the absence of events in the intervention (0/220) and the control group (0/231). We are uncertain of the effect of SDAC on incidence and severity on poisoning.
Duration of toxic symptoms	No studies collected or reported this outcome
Drug absorption	No studies collected or reported this outcome
Incidence of hospitalization	No studies collected or reported this outcome
Incidence of ICU admission	Control group: 0/231; intervention group: 1/220 (Peto OR 7.77, 95% CI 0.15 to 391.93)		—	451 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain of the effect of SDAC on the incidence of ICU admission.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; MD: mean difference; RCT: randomized controlled trial; RR: risk ratio; SDAC: single‐dose activated charcoal; OR: odds ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious limitations in study design: high risk of selection bias. ^bDowngraded one level for serious indirectness: study conducted in a hospital setting. ^cDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^dDowngraded one level for serious imprecision: low number of events.

Summary of findings 2. SDAC + hospital intervention versus hospital intervention alone for first aid in patients with acute oral poisoning

SDAC + hospital intervention versus hospital intervention alone for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (not specified, tricyclic antidepressants, combinations of different drugs or yellow oleander) Setting: hospital setting Intervention: single‐dose activated charcoal (SDAC) + hospital intervention Comparison: hospital intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with hospital intervention	Risk with SDAC + hospital intervention
Incidence of mortality	Study population		Peto OR 1.04 (0.79 to 1.37)	3425 (2 RCTs)	⊕⊕⊝⊝ Low^a,b	SDAC in addition to hospital treatments may make little or no difference on incidence of mortality.
	62 per 1000	64 per 1000 (49 to 85)
Incidence of adverse events	Incidence of vomiting: intervention group: 118/570 and control group: 163/1236 (RR 1.44, 95% CI 0.88 to 2.37; 1806 participants; 2 studies). Incidence of absent bowel sounds: intervention group: 7/1544 and control group: 17/1554 (RR 0.41, 95% CI 0.17 to 1.00, 1 study, 3098 participants).		—	4904 (3 RCTs)	⊕⊝⊝⊝ Very low^a,c,d	Statistically significant heterogeneity was found, which may be explained partially by subgroup analyses per type of adverse event. We are uncertain about the effect of SDAC in addition to hospital treatments on incidence of adverse events.
Incidence and severity of symptoms of poisoning: incidence of need for intubation	Patients that received gastric lavage prior to SDAC: intervention group: 80/1578 and control group: 87/1597 (RR 0.95, 95% CI 0.70 to 1.27, 2 studies, 3175 participants). Patients that did not receive gastric lavage prior to SDAC: intervention group: 24/194 and control group: 10/193 (RR 2.61, 95% CI 1.38 to 4.93, 1 study, 387 participants).		—	3562 (4 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	Statistically significant heterogeneity was found, which may be explained by subgroup analyses in patients receiving or not receiving gastric lavage. We are uncertain about the effect of SDAC on incidence of need for intubation.
Duration of toxic symptoms: duration of intubation (h)	Eddleston 2008: intervention group median (IQR): 112.0 (36.6 to 234.9) h and control group median (IQR): 88.5 (38.5 to 203.1) h (median difference: 23.5 h, P > 0.05). Merigian 2002: intervention group mean: 54.6 h and control group mean: 39.9 h (MD: 14.7 h, P = 0.70).		—	(2 RCTs)	⊕⊕⊝⊝ Low^a,e	Data were reported as median with IQR in one or means without measure of spread in another study, without information on participant numbers. SDAC in addition to hospital treatments may make little or no difference on the duration of intubation.
Drug absorption: cardenolide: AUC (µg/L) × h Follow‐up: 1 days	The median (IQR) in intervention group was 17.7 (11.1 to 21.8) (µg/L) × h and in the control group 19.0 (13.7 to 24.3) (µg/L) × h (median difference: −1.3 h, P > 0.05)		—	68 (1 RCT)	⊕⊝⊝⊝ Very low^a,f,g	We are uncertain about the effect of SDAC in addition to hospital treatments on cardenolide absorption.
Incidence of hospitalization	125 per 1000	196 per 1000 (152 to 252)	RR 1.57 (1.22 to 2.02)	1479 (1 RCT)	⊕⊝⊝⊝ Very low^a,g,h	We are uncertain about the effect of SDAC in addition to hospital treatments on incidence of hospitalization.
Incidence of ICU admission	30 per 1000	69 per 1000 (42 to 114)	RR 2.33 (1.42 to 3.82)	1479 (1 RCT)	⊕⊝⊝⊝ Very low^a,g,h	We are uncertain about the effect of SDAC in addition to hospital treatments on incidence of ICU admission.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; RCT: randomized controlled trial; RR: risk ratio; SDAC: single‐dose activated charcoal; OR: odds ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious indirectness: study conducted in a hospital setting. ^bDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^cDowngraded one level for serious inconsistency: large and statistically significant heterogeneity present (I² > 60%, P < 0.10). ^dDowngraded one level for serious imprecision: wide confidence intervals. ^eDowngraded one level for serious imprecision: lack of data on the number of patients analysed. ^fDowngraded one level due to serious limitations in study design: high risk of other bias: it is not entirely clear what is measured with the assay used. The fact that both active cardenolides and (inactive) metabolites might be detected by the assay compromise the results of these analyses, as they might explain the wide variability observed. ^gDowngraded one level for serious imprecision: low number of events. ^hDowngraded one level for serious limitations in study design: high risk of selection bias.

Summary of findings 3. MDAC + hospital intervention versus SDAC + hospital intervention for first aid in patients with acute oral poisoning

MDAC + hospital intervention versus SDAC + hospital intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (carbamazepine, yellow oleander, or combinations of different drugs) Setting: hospital setting Intervention: multiple dose of activated charcoal (MDAC) + hospital intervention Comparison: single‐dose activated charcoal (SDAC) + hospital intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with SDAC + hospital intervention	Risk with MDAC + hospital intervention
Incidence of mortality	Study population		RR 0.59 (0.21 to 1.63)	3476 (2 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	Combining the studies resulted in statistically significant heterogeneity, for which explanations remain speculative. We are uncertain about the effects of MDAC in addition to hospital treatment, compared to SDAC, in addition to hospital treatment.
	72 per 1000	42 per 1000 (15 to 117)
Incidence of adverse events	Study population		Peto OR 3.55 (1.85 to 6.79)	3476 (2 RCTs)	⊕⊕⊝⊝ Low^b,c	There was statistically significant heterogeneity, which may be attributable to different adverse events measured in individual studies. MDAC in addition to hospital treatment may increase abdominal discomfort/diarrhoea and absent bowel sounds, compared to SDAC in addition to hospital treatment.
	4 per 1000	14 per 1000 (7 to 27)
Incidence and severity of symptoms of poisoning: incidence of need for intubation	Study population		RR 1.01 (0.75 to 1.38)	3097 (2 RCTs)	⊕⊕⊝⊝ Low^b,c	MDAC in addition to hospital treatment may make little or no difference in the incidence of need for intubation, compared to SDAC in addition to hospital treatment.
	49 per 1000	49 per 1000 (37 to 67)
Duration of toxic symptoms: duration of intubation (h)	Brahmi 2006: intervention group: 24.1 (SD 4.2 h and control group 36.4 (SD 3.6 h (MD: 12.30 h lower, 95% CI −18.56 to −6.04, 6 participants). Eddleston 2008: intervention group median (IQR): 83.8 (35.0 to 173.0) h and control group median (IQR): 112.0 (36.6 to 234.9) h (median difference: 28.2 h), unclear number of participants		—	(2 RCTs)	⊕⊝⊝⊝ Very low^b,d,e	Data were reported as means with SD in one study or medians with IQR in another study, without information on participant numbers or statement of significance. We are uncertain about the effects of MDAC in addition to hospital treatment on duration of intubation, compared to SDAC in addition to hospital treatment.
Drug absorption: cardenolide: AUC (µg × L/h) Follow‐up: 1 days	The median (IQR) in intervention group was 17.3 (12.8 to 21.7) (µg/L) × h and in the control group 17.7 (11.1 to 21.8) (µg/L) × h (median difference −0.4, P > 0.05).		—	64 (1 RCT)	⊕⊝⊝⊝ Very low^b,e,f	We are uncertain about the effects of MDAC in addition to hospital treatment on cardenolide drug absorption, compared to SDAC in addition to hospital treatment.
Incidence of hospitalization	No studies collected or reported this outcome
Incidence of ICU admission	Study population		RR 0.31 (0.12 to 0.83)	401 (1 RCT)	⊕⊕⊝⊝ Low^b,g	MDAC in addition to hospital treatment may result in a decreased incidence of ICU admission, compared to SDAC in addition to hospital treatment.
	80 per 1000	25 per 1000 (10 to 66)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). AUC: area under the receiver operating curve; CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; MDAC: multi‐dose activated charcoal; RCT: randomized controlled trial; RR: risk ratio; SD: standard deviation; SDAC: single‐dose activated charcoal; OR: odds ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious inconsistency: combining results resulted in a considerable and statistically significant degree of heterogeneity (I² > 60%, P < 0.10). ^bDowngraded one level for serious indirectness: study conducted in a hospital setting. ^cDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^dDowngraded one level for other limitations: inconsistent conclusions made by the studies. ^eDowngraded one level for serious imprecision: low sample size and lack of data. ^fDowngraded one level for serious study limitations: high risk of other bias: it is not entirely clear what is measured with the assay used. The fact that both active cardenolides and (inactive) metabolites might be detected by the assay compromise the results of these analyses, as they might explain the wide variability observed. ^gDowngraded one level for serious imprecision: low number of events.

Summary of findings 4. SDAC versus syrup of ipecac for first aid in patients with acute oral poisoning

SDAC versus syrup of ipecac for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (anti‐inflammatory drugs, analgesics or psychotropic drugs) Setting: hospital setting Intervention: single‐dose activated charcoal (SDAC) Comparison: syrup of ipecac
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with syrup of ipecac	Risk with SDAC
Incidence of mortality	No studies collected this outcome
Incidence of adverse events	Study population		RR 1.24 (0.26 to 5.83)	34 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain about the effect of SDAC, compared to syrup of ipecac on incidence of adverse events.
	154 per 1000	191 per 1000 (40 to 897)
Incidence and severity of symptoms of poisoning: level of coma assessed with Glasgow Coma Scale Scale from: 3 to 15 Follow‐up: 1 h	The mean incidence and severity of symptoms of poisoning: level of coma was 14.91	MD 0.15 lower (0.43 lower to 0.13 higher)	—	34 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,d	We are uncertain about the effect of SDAC, compared to syrup of ipecac on the level of coma.
Duration of toxic symptoms	No studies collected this outcome
Drug absorption	No studies collected this outcome
Incidence of hospitalization	No studies collected this outcome
Incidence of ICU admission	No studies collected this outcome
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; MD: mean difference; RCT: randomized controlled trial; RR: risk ratio; SDAC: single‐dose activated charcoal.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious study limitations: high risk of selection bias. ^bDowngraded one level for serious indirectness: study conducted in a hospital setting. ^cDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^dDowngraded one level for serious imprecision: low sample size.

Summary of findings 5. MDAC + hospital intervention versus hospital intervention for first aid in patients with acute oral poisoning

MDAC + hospital intervention versus hospital intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (benzodiazepines, yellow oleander or combinations of different drugs) Setting: hospital setting Intervention: multi‐dose activated charcoal (MDAC) + hospital intervention Comparison: hospital intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with hospital intervention	Risk with MDAC + hospital intervention
Incidence of mortality	Study population		RR 0.94 (0.72 to 1.22)	3085 (1 RCT)	⊕⊕⊝⊝ Low^a,b	MDAC in addition to hospital treatment may make little or no difference in incidence of mortality.
	68 per 1000	64 per 1000 (49 to 82)
Incidence of adverse events	Study population		RR 1.02 (0.52 to 1.98)	3085 (1 RCT)	⊕⊕⊝⊝ Low^a,b	MDAC in addition to hospital treatment may make little or no difference in incidence of adverse events.
	11 per 1000	11 per 1000 (6 to 22)
Incidence and severity of symptoms of poisoning: incidence of need for intubation	Study population		RR 0.97 (0.71 to 1.33)	3085 (1 RCT)	⊕⊕⊝⊝ Low^a,b	MDAC in addition to hospital treatment may make little or no difference in incidence of need for intubation.
	49 per 1000	47 per 1000 (35 to 65)
Duration of toxic symptoms: length of intubation (h)	The median (IQR) length of intubation in the intervention group was 83.8 (35.0 to 173.0) h and 88.5 (38.5 to 203.1) h in the control group and was reported not to differ significantly (P > 0.05); unclear number of participants		—	(1 RCT)	⊕⊕⊝⊝ Low^a,c	The number of participants analysed was not reported. MDAC in addition to hospital treatment may make little or no difference in length of intubation
Drug absorption: cardenolide: AUC (µg/L × h) Follow‐up: 1 day	The median (IQR) cardenolide AUC in the intervention group was 17.3 (12.8 to 21.7) (µg/L) × h and 19.0 (13.7 to 24.3) (µg/L) × h in the control group.		—	76 (1 RCT)	⊕⊝⊝⊝ Very low^a,c,d	We are uncertain about the effects of MDAC in addition to hospital treatment on cardenolide drug absorption.
Incidence of hospitalization	No studies collected or reported this outcome
Incidence of ICU admission	No studies collected or reported this outcome
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). AUC: area under the receiver operating curve; CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; MDAC: multi‐dose activated charcoal; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious indirectness: study conducted in a hospital setting. ^bDowngraded one level for serious imprecision: low number of events and wide confidence interval. ^cDowngraded one level for serious imprecision: low sample size and lack of data. ^dDowngraded one level for serious study limitations: high risk of other bias: it is not entirely clear what is measured with the assay used. The fact that both active cardenolides and (inactive) metabolites might be detected by the assay compromise the results of these analyses, as they might explain the wide variability observed.

Summary of findings 6. Syrup of ipecac versus no intervention for first aid in patients with acute oral poisoning

Syrup of ipecac versus no intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (toxic berries) Setting: pre‐hospital setting Intervention: syrup of ipecac Comparison: no intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with no intervention	Risk with Syrup of ipecac
Incidence of mortality	No studies collected this outcome
Incidence of adverse events: diarrhoea Follow‐up: 1 day	Study population		RR 4.08 (1.66 to 10.04)	103 (1 RCT)	⊕⊕⊝⊝ Low^a,b	Syrup of ipecac may result in an increased incidence of diarrhoea.
	96 per 1000	392 per 1000 (160 to 965)
Incidence and severity of symptoms of poisoning	No studies collected this outcome
Duration of toxic symptoms	No studies collected this outcome
Drug absorption	No studies collected this outcome
Hospitalization: incidence of hospitalization Follow‐up: 1 days	—		—	103 (1 RCT)	⊕⊕⊝⊝ Low^a,b	The effect was not estimable due to the absence of events in the intervention (0/52) and the control group (0/51).
ICU admission	No studies collected this outcome
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious study limitations: high risk of selection bias, high risk of detection bias, and high risk of other bias (no confirmation of actual ingestion or uptake, reporting of dichotomous outcomes while measuring with an ordinal scale). ^bDowngraded one level for serious imprecision: low number of events.

Background

Description of the condition

Poisoning can be defined as exposure of the body to exogenous substances, in sufficiently large amounts to cause harm to the individual. This can happen through chronic exposure to low doses of a substance, or more acutely through sudden exposure to a harmful dose. Acute poisoning can happen either accidentally or voluntarily, as a way to end one's own or another's life or as a 'cry for help'.

Poisoning inflicts a major burden of morbidity and mortality worldwide. The World Health Organization (WHO) estimates that 108,000 deaths a year are caused by unintentional poisoning (WHO 2016), accompanied by the loss of a staggering 6,558,000 disability adjusted life years (DALYs) (WHO 2016). In addition to this, auto‐intoxication is one of the most common methods to attempt suicide. Yearly, around 800,000 people worldwide commit suicide, and around 30% of their attempts occur through the intake of pesticides, a phenomenon typically occurring in rural areas in lower‐ and middle‐income countries (WHO 2018). As most attempted suicides are unsuccessful, the actual burden will be much higher (Albert 2015). Poisoning can happen via different routes of exposure, such as through inhalation, injection or dermal absorption, but by far the most common is through deliberate or accidental ingestion of a toxic substance (Mowry 2016), which is the focus of this review. An important patient group to suffer from unintentional poisoning are young children. Roughly 20% of all accidental poisonings are thought to occur in children aged under 5 years (WHO 2016). In high‐income countries, this proportion is even larger: up to 47% of the incoming calls to the American Poison Control Centers concern exposures in this age group (Mowry 2016). This is most likely because young children are curious to explore their environment and do not realize the dangers of putting unknown and potentially harmful things in their mouth. Indeed, large numbers of exposures are to cosmetics and household products (25% of all reported exposures in children aged 5 years or younger; Mowry 2016).

Hospital treatment of acute oral poisoning focuses initially on supportive therapy: hypertonic glucose infusion, maintaining the victim's vital parameters and keeping poison‐induced symptoms under control (Isbister 2016; Nelson 2011). If practitioners can identify a toxin syndrome, they can administer a poison‐specific antidote, for example N‐acetylcysteine for a paracetamol overdose or naloxone for an opioid overdose (Chiew 2018; Wilkerson 2016). Third‐line treatment options include gastrointestinal decontamination procedures: activated charcoal can adsorb the poisonous substance (Corcoran 2016), while gastric lavage or whole bowel irrigation are procedures that attempt to eliminate the poison out of the gastrointestinal tract before absorption into the blood (Donkor 2016; Thanacoody 2015). A final treatment strategy is to eliminate toxins that have already been absorbed through multiple doses of activated charcoal, haemodialysis or blood/urinary alkalinization (Decker 2015; Gaudreault 2005; Proudfoot 2003; Roberts 2005).

In cases of acute oral poisoning, a swift reaction is crucial. For activated charcoal (AC), experimental studies have shown that its efficacy in limiting drug absorption decreases dramatically over time (Chyka 2005). Therefore, treatment guidelines recommend using AC within an hour after ingestion of the poison, although AC may still produce effects after that time, especially in drugs administered in a delayed release formula (Chyka 2005; Juurlink 2015). However, it is difficult to adhere to these guidelines in emergency services, mainly due to the delay between ingestion and presentation at the emergency department (Karim 2001; LoVecchio 2007; Tuuri 2009). Thus, any effective first aid measure that would neutralize, limit or delay uptake, or promote evacuation from the gastrointestinal tract in case of acute oral poisoning, could save precious time for professionals, potentially making the difference between life and death, or serious morbidity, for the poisoned patient.

First aid, as defined by the International Liaison Committee On Resuscitation (ILCOR), is the immediate help provided to a sick and injured person until professional help arrives. First aid interventions seek to preserve life, alleviate suffering, prevent further illness or injury and promote recovery (Zideman 2015). This definition implies that a first aid intervention must be both available to and feasible for a layperson in a pre‐hospital setting. Of the previously mentioned hospital interventions, only activated charcoal, which is relatively easy to administer orally and available without prescription, is feasible. In addition to these, other suggested first aid techniques include administering emetics, such as syrup of ipecac (or ipecacuanha) (Quang 2000); using cathartics, such as sorbitol (Keller 1990), drinking water, milk, vinegar or citrus juice to dilute and/or neutralize the poison (Rumack 1977); or adjusting the poisoned victim's body position to slow down the uptake of the poison (Vance 1992).

In case of ingestion of toxic alcohols (e.g. methanol, ethylene glycol), ethanol could be considered a potential home remedy due to its wide availability. However, the use of large volumes of ethanol is dangerous and needs to be monitored carefully (Rietjens 2014). Therefore, it is not recommended for use in a lay setting without professional guidance. Current recommendations for laypeople are limited to placing the victim in the lateral decubitus position and seeking professional help (e.g. contacting poison control centres if available) and following their advice (IFRC 2016).

Description of the intervention

The focus of this Cochrane Review is any intervention that is readily available to and administrable by laypeople before professional help arrives, targeted at neutralizing, limiting or delaying the absorption, or promoting the evacuation of a poison.

Limiting the absorption of a poison can be achieved by administering an adsorbent, such as activated charcoal. This black powder is produced through pyrolysis of carbon‐rich materials and activation by steam to remove already adsorbed substances (Olson 2010). This process results in a material with a very large surface area and hence adsorbing capacity. It needs to be mixed with water to form a slurry that can be ingested.

Placing the poisoning victim on their left side might be another method to decrease absorption of the poison (Vance 1992).

Substances that can promote the evacuation of a poison from the gastrointestinal tract firstly include emetics. The best known and most recommended is syrup of ipecacuanha, or ipecac. This syrup is derived from the roots and rhizome of Cephaelis ipecacuanha (Lee 2008). Other suggested emetics are apomorphine and copper sulphate. Apomorphine is believed to induce vomiting faster than ipecac (MacLean 1973), but it is not feasible to administer in a home setting and can cause central nervous depression, so it is not recommended, especially in children (MacLean 1973). Copper sulphate has also been used to induce vomiting in people with oral poisoning due to its action as a local irritant in the stomach (Karlsson 1965). However, it is a common source of intoxication itself, hence its use is also discouraged (Nastoulis 2017).

A second class of substances that can theoretically speed up the evacuation of an ingested poison from the gastrointestinal tract are cathartics. Patients who have ingested slow‐absorbing materials might benefit most from these treatments, although current guidelines suggest not using cathartics without activated charcoal (American Academy of Clinical Toxicology 2004). Examples of suggested cathartics include sugars, such as mannitol, lactulose and sorbitol, or salts, including magnesium sulphate, magnesium citrate and sodium sulphate.

In addition, diluting and neutralizing poisons, especially caustic substances such as lye, could occur through the intake of water, milk, vinegar or citrus juice (Rumack 1977). Milk might also have some adsorbing capacity (Chin 1969).

How the intervention might work

First aid interventions to treat poisoning can be categorized in four groups:

those that either limit or delay absorption of the poison in the body, such as activated charcoal or certain body positions;
interventions that evacuate the poison from the gastrointestinal tract, either by vomiting or by defecation;
combinations of first aid interventions that limit uptake and promote evacuation of the poison, e.g. sorbitol and activated charcoal;
first aid interventions that neutralize or dilute the poison, such as drinking water, milk, vinegar or citrus juice.

Furthermore, other combinations of first aid interventions may also be used.

A. First aid interventions that limit or delay the absorption of the poison in the body

One way to limit the absorption of a poison is to administer a substance that binds to the poison, thus preventing it from being absorbed by the body. Activated charcoal (AC) is one such adsorbent. Its enormous surface area can adsorb large quantities of drugs through the generation of Van der Waals forces between the charcoal and the adsorbed molecule (Olson 2010). Not all substances are equally effectively bound by AC. For example, lithium, iron, cyanide or alcohols bind to AC only to a minor extent, which means its appropriateness needs to be carefully considered in these cases (Bateman 1999; Juurlink 2015; Olson 2010). The optimal dose regimen for activated charcoal administration is not entirely clear, but 25 g to 100 g is considered to be a standard dose for adults (Chyka 2005). In practice, ingesting more than 50 g seems to be difficult to achieve for patients.

A certain body position might also slow down the uptake of the poison. The primary site of absorption for most pharmacologic substances is the small intestine, because of its large surface area and thin epithelium. Therefore, any factor that would delay gastric emptying into the small intestine should decrease the rate of absorption and limit the potential toxic effects of the ingested drug. Studies indicate that laying on the right side accelerates gastric emptying (Loots 2013; Valeur 2015; Van Wijk 2007). In contrast, placing the patient in the left lateral decubitus position might slow the rate of absorption of the ingested poison, because the anatomy of the stomach, combined with gravity, would allow the gastric content to stay in the greater curvature of the stomach (Vance 1992).

B. First aid interventions that evacuate the poison from the gastrointestinal tract

Evacuation of the poison from the gastrointestinal tract as quickly as possible can be achieved by inducing vomiting or accelerating defecation. Two types of drugs can be considered: emetics induce vomiting, while cathartics accelerate defecation.

As mentioned before, syrup of ipecac is the best known type of emetic. The main active substances of the ipecacuanha plant are emetine and cephaeline, which induce emesis and diarrhoea by acting both as a local irritant in the upper gastrointestinal tract and by targeting the chemoreceptor trigger zone in the medulla oblongata of the brain, the body's vomiting centre (Lee 2008). A potential risk associated with the use of emetics is lung injury through vomit aspiration (Höjer 2013).

Cathartics draw water into the large intestine, thereby stimulating bowel movements and thus accelerating defecation (American Academy of Clinical Toxicology 2004).

C. First aid interventions that limit uptake and promote evacuation of the poison from the gastrointestinal tract

Cathartics can be used with activated charcoal. This combination is thought to reduce drug uptake by accelerating evacuation out of the small bowel (Moon 2015). Furthermore, cathartics counteract the constipating effects of AC (James 1995). On the other hand, in vitro studies have suggested that cathartics might influence the adsorbing capacity of AC (Orisakwe 2001).

D. First aid interventions that neutralize or dilute the poison

A commonly used home remedy for poisoning by caustic substances is drinking large amounts of fluids, such as water, milk, vinegar or citrus juice (Rumack 1977). The rationale behind this is not only to dilute the poison, but also to change the pH in the stomach, thereby neutralizing the caustic effects of the ingested poison. Considerations that need to be made when using this approach are the chemical properties of the ingested substance (acidic or basic), the heat production that might occur during neutralization and sufficient availability of the neutralizing substance. In addition to its potential neutralizing effect, in vitro data suggest that milk has some adsorbing capacity (Chin 1969). However, increasing the volume of fluids in the stomach might also increase the rate of emptying into the small bowel, where the absorption of the poison takes place (Blain 2011). Furthermore, drinking large amounts of water might cause water intoxication (Lai 2016). A final consideration is that drinking large volumes of fluids might increase the risk of vomiting, which could be problematic in cases of caustic poisoning, as the caustic substance would contact the oesophagus for a second time.

Why it is important to do this review

There are several Cochrane Reviews concerning the treatment and prevention of poisoning. Kendrick 2012 provided evidence on interventions to prevent injuries at home, including cases of oral poisoning, while Hawton 2015 investigated potential interventions to decrease self‐harm in children, adolescents and adults. A review by Nussbaumer‐Streit 2016 documented potential household interventions to prevent domestic lead exposure in children. In addition, numerous Cochrane Reviews have investigated the use of hospital interventions to treat a range of specific intoxications, such as for example paracetamol or lithium poisoning (Chiew 2018; Lavonas 2015).

This Cochrane Review fills the gap between prevention and hospital treatment of poisoning, by investigating which pre‐hospital interventions, available and feasible for a lay person in a non‐healthcare setting, are effective in cases of acute oral poisoning. Identified interventions can be used in first aid guidelines targeted at lay people in settings such as nightclubs, childcare centres or the workplace, to be applied before arrival of professional help. As time is a crucial factor in acute oral poisoning, effective interventions conducted by laypeople would save valuable time and could therefore be crucial to survival (Chyka 2005; Juurlink 2015).

Objectives

To assess the effects of pre‐hospital interventions (alone or in combination) for treating acute oral poisoning, available to and feasible for laypeople before the arrival of professional help.

Methods

Criteria for considering studies for this review

Types of studies

We considered randomized controlled trials (RCTs) in actual poisoning patients. We excluded studies involving healthy volunteers and preclinical studies (animal studies, in vitro research).

In order to be eligible for inclusion in the review, all RCTs taking place after 2010 must have been prospectively registered (Roberts 2015). All RCTs conducted prior to 2010 were eligible for inclusion.

Types of participants

We included participants poisoned via oral ingestion, both deliberately and accidentally. In addition to studies in a community setting, we considered studies conducted in a healthcare setting, including a hospital setting or ambulatory care, as most studies identified would likely have been performed in a controlled setting. Although this may be a source of indirectness, we feel that excluding these studies would result in selection bias.

Types of interventions

All identified first aid interventions, alone or in combination and feasible for a layperson in a pre‐hospital setting, were eligible. These included, among others, activated charcoal and other adsorbents (single‐ or multi‐dose); syrup of ipecac and other emetics (single or multi‐dose); cathartics (single or multi‐dose); body positioning; and water, milk, vinegar or citrus juice.

We compared the interventions to each other or to no intervention. We did not compare them to typical hospital interventions such as gastric lavage, whole bowel irrigation or the use of antidotes. However, if pre‐hospital treatments were used in adjuvant to an established hospital treatment, we included these studies. The reason for not considering established hospital treatments as comparisons is that we are interested in the most efficacious treatments in a non‐healthcare setting. It is likely that these would be less efficient than a hospital treatment, but they might still be useful as a first aid measure, which typically takes place before presentation to a healthcare facility.

Co‐interventions were allowed if all groups received them in equal doses. We separately explored interventions aiming to limit or delay absorption of poison, evacuate poison, limit uptake and evacuate poison, and neutralize or dilute poison.

Types of outcome measures

Timings of outcomes are defined as early (within 24 h after poisoning), intermediate (24 h to one week after poisoning) and late (more than one week and less than one year after poisoning).

Primary outcomes

Incidence of mortality
Incidence of adverse events due to the intervention
Incidence and severity of symptoms of poisoning, reported for example with the Poisoning Severity Score (PSS) (Persson 1998)

Secondary outcomes

Duration of toxic symptoms
Drug absorption: measured as maximal concentration of drug in the blood (C_max), time to C_max (T_max) or area under the curve (AUC) of drug concentration versus time
Incidence of hospitalization
Incidence of intensive care unit (ICU) admission

Search methods for identification of studies

Electronic searches

We searched the following databases on 4 December 2018, without any language restrictions or date limits.

The Cochrane Library (2018, Issue 11, searched 4 December 2018; www.cochranelibrary.com), including the following databases.
- The Cochrane Database of Systematic Reviews.
- The Cochrane Central Register of Controlled Trials (CENTRAL), for reports of RCTs from MEDLINE, Embase and records submitted from Cochrane Specialized Registers, including the Cochrane Injuries Group.
- DARE (Database of Abstracts of Reviews of Effect).
MEDLINE, using the PubMed interface (1966 to 4 December 2018).
Embase, using the Embase.com interface (1947 to 4 December 2018).
CINAHL, using the EBSCO host interface (1982 to 4 December 2018).
ISI Web of Science: Science Citation Index Expanded (SCI‐EXPANDED) and Conference Proceedings Citation Index‐Science (CPCI‐S) (1900 to 4 December 2018).
International Pharmaceutical Abstracts, using the Ovid interface (1970 to 4 December 2018).
Clinicaltrials.gov (clinicaltrials.gov).
EU Clinical Trials Register (www.clinicaltrialsregister.eu).
WHO International Clinical Trials Registry Platform (ICTRP) (www.who.int/ictrp/en/).

For each of the articles included, we did a search in MEDLINE (via the PubMed interface) and screened the first 20 similar articles for additional relevant publications. Search strategies can be found in Appendix 1. Furthermore, we searched previously published systematic reviews and evidence‐based guidelines that were identified during the database searches (Table 1).

Table 1. Sources of individual studies

Author and year of publication	Title
Abrass 2012	The evidence for activated charcoal in resource poor settings: a systematic review
American Academy of Clinical Toxicology 1999	Position statement and practice guidelines on the use of multi‐dose activated charcoal in the treatment of acute poisoning
American Academy of Clinical Toxicology 2004	Position paper: cathartics
Chiew 2018	Interventions for paracetamol (acetaminophen) overdoses
Chyka 2005	Position paper: single‐dose activated charcoal
Eddleston 2003	Does gastric lavage really push poisons beyond the pylorus? A systematic review of the evidence
Blain 2011	Organophosphorus poisoning (acute)
Höjer 2013	Position paper update: ipecac syrup for gastrointestinal decontamination.
Jones 2002	Towards evidence based emergency medicine: best BETs from the Manchester Royal Infirmary. Activated charcoal and gastric absorption of iron compounds
Manoguerra 2005	Guideline on the use of ipecac syrup in the out‐of‐hospital management of ingested poisons
Qureshi 2011	Adverse effects of activated charcoal used for the treatment of poisoning
Roberts 2011	Enhanced elimination in acute barbiturate poisoning ‐ a systematic review

We included relevant conference abstracts retrieved from searches in the above‐mentioned databases in the review.

Searching other resources

We searched the reference list of included articles, retrieved with the above searches, to identify other studies.

Data collection and analysis

Selection of studies

Two authors (BA and VB or AV) independently screened the titles and abstracts of all references yielded by the search. Subsequently, we retrieved full texts of selected articles, using a study selection form to assess eligibility. We resolved any discrepancies between authors through discussion. In cases where no consensus could be reached, we consulted a third author (EDB or AV). We documented the included studies in the appropriate sections within the review and summarized studies that were excluded after full‐text assessment in the Characteristics of excluded studies table, together with the reason for exclusion. We describe identified studies that were selected based on study design, study population and intervention of interest, but which reported no outcome of interest or outcome data, in the Results section of the review. We tried to contact the authors to ascertain whether the data for our outcomes of interest were unavailable due to lack of measurement or lack of reporting.

Data extraction and management

Two authors (BA and VB or AV) independently extracted data from all studies using a standardized and piloted data extraction form.

They extracted the following information from each study.

General information: author, year of publication, year of study, country of study lead author.
Study characteristics.
- Study design.
- Information on study population: number of participants, age, sex, country of study and poisoning characteristics (type and dose of intoxication, deliberate or accidental intoxication, time elapsed between intoxication and intervention, experimental or community setting).
- Details of the intervention and the comparison: type of intervention, dose, route of administration, duration of the treatment.
- Outcome(s) measured.
Study findings.
- Effects of the intervention on the outcome: effect measure, confidence interval, P value.
- Number of events and participants in intervention and comparison groups.

Assessment of risk of bias in included studies

Two authors (BA and VB or AV) independently assessed risk of bias in the included studies using Cochrane's 'Risk of bias' tool (Higgins 2011). They assessed the domains of sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data addressed, selective reporting and possible other bias, rating each domain as being at low, high or unclear risk of bias.

Measures of treatment effect

We used Review Manager 5 (RevMan 5) to manage data and conduct analyses (RevMan 2014). We reported continuous outcomes as mean differences (MD) with 95% confidence intervals (CIs) and dichotomous outcomes as risk ratios (RR) with 95% CIs, or Peto odds ratios (OR) when events were rare and criteria were satisfied.

Unit of analysis issues

We identified studies that had a multi‐arm design. We were cautious during the analysis of these data, ensuring that the same group of participants was not included twice in the meta‐analysis. We achieved this by ensuring that separate interventions were not included in a single meta‐analysis. Secondly, if multiple doses or administration times of an intervention were compared to a control group, we combined groups to create a single pair‐wise comparison in the case of dichotomous outcomes, according to the recommendations of the Cochrane Handbook for Systematic Reviews of Interventions on the analysis of multi‐arm trials (Higgins 2011). We did not identify multi‐arm trials reporting continuous outcomes.

Dealing with missing data

In case of missing data, we attempted to contact the authors to obtain these data at least twice, if contact details were available.

Where possible, we calculated missing values (such as SDs) from the available data (P values, t values, CIs or standard errors) (Higgins 2011).

If insufficient data were available to calculate missing values, we only analysed the available data. We narratively described results from studies with missing data. We addressed the issue of the missing data and their potential impact on the findings of the study in the Discussion.

Assessment of heterogeneity

The target population of this review, patients with oral poisoning, is inherently heterogeneous with respect to the type, dose and timing of poison intake. However, the target audience for delivering this intervention, laypeople, are likely not capable of differentiating between these differences. Therefore, a certain degree of heterogeneity in the results is unavoidable.

Our analyses are stratified based on type of intervention. We assessed heterogeneity by inspection of the forest plot and by using the Chi²‐test and the I² statistic. We considered the Chi² statistic to be significant at P < 0.10. For interpretation of the I², we followed the guidance of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).

When I² was at least 80% and the P value of the Chi² test was less than 0.1, we considered heterogeneity to be substantial, whereas for I² values below 40%, we considered heterogeneity to be unimportant. When heterogeneity was substantial, we examined the direction of the effects before making a decision whether to report the pooled result or describe the effects narratively.

Assessment of reporting biases

We planned methods for assessing reporting biases, but we could not perform them (New Reference). See Differences between protocol and review section.

Data synthesis

Where possible, we performed meta‐analyses. We pooled data if there were two or more studies on the same intervention that assessed the same outcome and provided sufficient data. We did not combine outcomes with different timings into a single meta‐analysis. We analysed different comparisons as separate analyses. We performed meta‐analyses using a random‐effects model, given the anticipated variation between studies. For dichotomous outcomes, we used the Mantel‐Haenszel method, while for continuous outcomes, we used the inverse variance method. In case of dichotomous outcomes with no or few events in one of the test groups, we used the Peto OR method, if criteria were met according to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Given the large number of interventions in this review, we considered the possibility of a network meta‐analysis (NMA). However, due to the paucity of data and the heterogeneity in reported outcomes, this was not an option. In future updates of this review, we will consider this possibility again if there are sufficient data.

Subgroup analysis and investigation of heterogeneity

To investigate potential heterogeneity, we could have theoretically performed four possible subgroup analyses.

Different drugs taken. We hypothesized that for the intervention activated charcoal, drugs with a higher or a lower affinity for activated charcoal would be taken up to a lesser or higher extent in the body, while for cathartics, drugs that are absorbed faster would be less effectively flushed out of the body than drugs with a slower absorption rate.
Time point of the intervention. We hypothesized that the later an intervention is performed, the less efficacious it is in lowering the uptake of the drug.
Co‐interventions administered. We hypothesized that differing co‐interventions, such as the administration of a hospital treatment (e.g. gastric lavage), could influence the efficacy of the intervention investigated.
Type of adverse event experienced. We hypothesized that for the combined outcome 'occurrence of adverse events', different types of adverse events might be experienced to a different degree for a certain intervention.

Of these potential analyses, we could perform only the latter two because of the paucity of data.

Sensitivity analysis

We had planned to perform a sensitivity analysis by excluding studies at high or unclear risk of bias for sequence generation, allocation concealment, incomplete outcome reporting or other sources of bias, and then comparing the results with the initial analysis. However, we were not able to combine sufficient studies into a meta‐analysis for this analysis based on risk of bias of the individual studies.

We had also planned to perform sensitivity analyses in case we were required to impute data for some studies to enable meta‐analysis. We would have excluded the studies with imputed data and compared the results to the initial analysis. However, we were not able to impute data.

See Differences between protocol and review section.

'Summary of findings' table

We assessed the certainty of the body of evidence from the included studies according to the methodology described by the GRADE working group (Atkins 2004). The GRADE approach assesses the certainty of evidence for separate outcomes across the different studies in five domains: limitations in study design, consistency, imprecision, indirectness and publication bias. RCTs start with a level of high‐certainty evidence, which can be downgraded by one (serious limitations) or two (very serious limitations) levels for each of these domains. The certainty of evidence can therefore be high, moderate, low or very low. For the assessment of the GRADE domain 'limitations in study design', we decided to downgrade the certainty of evidence for an outcome if we judged one of the studies contributing to this outcome to be at high risk of bias in one of following domains: selection bias, detection bias, attrition bias or other bias. We decided not to take into account domains with unclear risk of bias to make this judgment. For the assessment of the GRADE domain 'imprecision' according to the guidance of the GRADE working group (Guyatt 2011), we decided to downgrade the certainty of evidence for an outcome:

if the optimal information size criterion was not met and total sample size of studies contributing to the outcome was low (fewer than 400 participants) for continuous outcomes or there was a low number of events (fewer than 300 events) for dichotomous outcomes;
if the CIs were wide (including both the line of no effect and an appreciable benefit or harm, i.e. a 25% increase or decrease in risk for dichotomous outcomes or a 50% increase or decrease in mean difference for continuous outcomes); or
if there was a lack of data to judge the prior two criteria.

We created a 'Summary of findings' table, using the online GRADEpro Guideline Development Tool (GRADEpro GDT 2015), for the most relevant comparison of interventions in a first aid setting: single‐dose activated charcoal (SDAC) versus no intervention. We created additional 'Summary of findings' tables for the other most clinically relevant comparisons involving single‐ and multi‐dose activated charcoal: SDAC plus hospital intervention versus hospital intervention alone, MDAC plus hospital intervention versus SDAC plus hospital intervention, MDAC plus hospital intervention versus hospital intervention alone, and syrup of ipecac versus no intervention. We also created 'Summary of findings' tables for the other identified comparisons, but we placed these in the Appendices.

We have included all primary and secondary outcomes of this review in our 'Summary of findings' tables. For outcomes such as severity of symptoms, studies reported multiple outcomes (e.g. incidence of clinical improvement, incidence of intubation requirement, incidence of convulsions etc.). As 'Summary of findings' tables should include no more than seven outcomes, we chose the clinically most relevant outcomes together with a clinical expert (PD).

Results

Description of studies

Results of the search

Our search strategies identified a total of 11,582 references. After removing 1859 duplicates and screening titles and abstracts, we assessed 78 full‐text records for eligibility. At this stage we included 20 studies, reported in 27 records, and we then included an additional four studies after screening reference lists of included studies and systematic reviews retrieved with the database searches and similar articles in PubMed. Figure 1 shows the flowchart of the study selection.

Figure 1

Study selection flow diagram.

Included studies

We included 24 studies reported in 31 publications and involving a total of 7099 participants randomized to different treatment groups. Only one study took place in a pre‐hospital setting (Wax 1999), whereas the rest were in hospitals.

Nineteen studies assessed the effects of single‐dose activated charcoal (SDAC), either administered alone (Amigó Tadín 2002; Merigian 1990; Underhill 1990), in adjuvant to hospital treatment (Behnoush 2009; Brahmi 2006; Comstock 1982; Cooper 2005; Crome 1983; De Silva 2003; Eddleston 2008; Hultén 1988; Merigian 2002; Roberts 2006), combined with a cathartic (James 1995; Passeron 1989), or combined with syrup of ipecac (Albertson 1989; Kornberg 1991; Kulig 1985; Pond 1995).

Seven studies looked at the effect of multi‐dose activated charcoal (MDAC) either in adjuvant to hospital treatment (Behnoush 2009; Bouget 1989; Brahmi 2006; De Silva 2003; Eddleston 2008; Roberts 2006), or combined with cathartics and in adjuvant to hospital treatment (Montoya‐Cabrera 1999).

Six studies investigated syrup of ipecac alone (Amigó Tadín 2002; Wax 1999), or followed by SDAC and a cathartic (Albertson 1989; Kornberg 1991; Kulig 1985; Pond 1995).

Table 2 contains an overview of the comparisons made in the different studies.

See: Summary of findings for the main comparison SDAC versus no intervention for first aid in patients with acute oral poisoning; Summary of findings 2 SDAC + hospital intervention versus hospital intervention alone for first aid in patients with acute oral poisoning; Summary of findings 3 MDAC + hospital intervention versus SDAC + hospital intervention for first aid in patients with acute oral poisoning; Summary of findings 4 SDAC versus syrup of ipecac for first aid in patients with acute oral poisoning; Summary of findings 5 MDAC + hospital intervention versus hospital intervention for first aid in patients with acute oral poisoning; Summary of findings 6 Syrup of ipecac versus no intervention for first aid in patients with acute oral poisoning

Table 2. Overview of comparisons

Comparison	Type of poisoning	Study
A. First aid interventions that limit or delay the absorption of the poison in the body
SDAC vs no intervention	Not specified	Merigian 1990
SDAC vs no intervention	Paracetamol	Underhill 1990
SDAC + hospital intervention vs hospital intervention	Not specified	Comstock 1982
	Benzodiazepines + paracetamol or other drug combinations	Cooper 2005
	Tricyclic antidepressants	Crome 1983
	Yellow oleander, organophosphorus/carbamate pesticide, organochlorine, other/unknown pesticide or paraquat, medicine or unknown	Eddleston 2008
	Amitriptyline, clomipramine, mianserin, imipramine, dothiepin, doxepin, nortriptyline, mixed overdoses with most commonly benzodiazepines or alcohol	Hultén 1988
	Not specified	Merigian 2002
	Yellow oleander	Roberts 2006
MDAC + hospital intervention vs SDAC + hospital intervention	Carbamazepine	Behnoush 2009
	Carbamazepine	Brahmi 2006
	Yellow oleander	De Silva 2003
	Yellow oleander, organophosphorus/carbamate pesticide, organochlorine, other/unknown pesticide or paraquat, medicine or unknown	Eddleston 2008
	Yellow oleander	Roberts 2006
SDAC vs syrup of ipecac	Anti‐inflammatory drugs, analgesics or psychotropic drugs	Amigó Tadín 2002
MDAC + hospital intervention vs hospital intervention	Benzodiazepine	Bouget 1989
	Yellow oleander, organophosphorus/carbamate pesticide, organochlorine, other/unknown pesticide or paraquat, medicine or unknown	Eddleston 2008
	Yellow oleander	Roberts 2006
B. First aid interventions that evacuate the poison from the gastrointestinal tract
Emetics
Syrup of ipecac vs no intervention	Toxic berries	Wax 1999
Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic	Not specified	Albertson 1989
	Wide variety, most commonly paracetamol	Kornberg 1991
	Not specified	Kulig 1985
	Paracetamol, salicylate, phenothiazines or ethanol, or other drugs	Pond 1995
Syrup of ipecac 15 mL vs syrup of ipecac 30 mL (dose)	Benzodiazepine tranquillizers or hypnotics, other tranquillizers, other hypnotics, antidepressants, analgesics, antihistamines, miscellaneous drugs and chemicals	Ilett 1977
Cathartics
SDAC + cathartic vs SDAC	Not specified	Sue 1994
SDAC + cathartic vs SDAC	Analgesics, anticonvulsants, antihistamines and decongestants, asthma therapies, automotive products, cardiovascular drugs, gastrointestinal preparations, insecticides, mushrooms, psychotropic drugs, rodenticides, topicals, miscellaneous drugs	James 1995
SDAC + cathartic vs SDAC + cathartic (dose)	Not specified	Sue 1994
SDAC + cathartic vs SDAC + cathartic (type)	Analgesics, anticonvulsants, antihistamines and decongestants, asthma therapies, automotive products, cardiovascular drugs, gastrointestinal preparations, insecticides, mushrooms, psychotropic drugs, rodenticides, topicals, miscellaneous drugs	James 1995
C. Combined first aid interventions that limit uptake and promote evacuation of the poison from the gastrointestinal tract
SDAC + cathartic + hospital intervention vs hospital intervention	Benzodiazepines, barbiturates or imipramine	Passeron 1989
MDAC + cathartic + hospital intervention vs hospital intervention	Paracetamol	Montoya‐Cabrera 1999
D. First aid interventions that neutralize or dilute the poison
No studies were identified

APF: Australian Pharmaceutical Formulary; MDAC: multi‐dose activated charcoal; SDAC: single‐dose activated charcoal; USP: United States Pharmacopeia.

Excluded studies

We excluded 47 studies after full‐text evaluation (Characteristics of excluded studies). We excluded 11 studies because of an ineligible study population (not oral poisoning patients or patients with chronic poisoning), 30 studies because of an intervention that did not meet our selection criteria and 5 because of an inappropriate comparison. Furthermore, we excluded one recent study, published as an abstract only (Escalante 2016), because of a lack of prospective trial registration, in accordance to the editorial policies of the Cochrane Injuries review group.

Risk of bias in included studies

We did not judge any study to be at low risk of bias on all domains investigated. We scored two studies as having a low risk of bias for all but one domain: in one study there was a risk of selective reporting (De Silva 2003), and in the other there was a risk of performance bias (Eddleston 2008). All other studies were at high or unclear risk of bias for two or more domains. Six studies were at high risk of bias in at least four domains (Albertson 1989; Crome 1983; Kornberg 1991; Merigian 2002; Pond 1995; Wax 1999), whereas 12 studies were at unclear risk of bias in three or more domains (Amigó Tadín 2002; Behnoush 2009; Bouget 1989; Brahmi 2006; Comstock 1982; Crome 1983; Hultén 1988; Ilett 1977; Montoya‐Cabrera 1999; Passeron 1989; Sue 1994; Underhill 1990). Figure 2 and Figure 3 provide an overview of the risk of bias across domains and studies, and detailed judgments by domain can be found for each included study in the Characteristics of included studies table.

Figure 2

Figure 3

Allocation

In general, randomization and allocation concealment was inadequately performed or poorly reported. The population was sufficiently randomized and adequately reported in six studies only (Bouget 1989; De Silva 2003; Eddleston 2008; Hultén 1988; Ilett 1977; Roberts 2006). In four studies the allocation concealment was adequate (Cooper 2005; De Silva 2003; Eddleston 2008; Roberts 2006).

Blinding

Most studies either did not blind or did not report on blinding of the participants and personnel. This is likely due to the nature of the interventions, which makes it difficult to perform adequate blinding. However, this might lead to performance bias, for example, because of differential administration of co‐interventions. One study that combined activated charcoal with different cathartics reported blinding both participants and personnel (James 1995), while another study testing multiple versus single doses of activated charcoal blinded the treating physicians by making sure research assistants cleaned the participants and their bedclothes after each activated charcoal treatment (De Silva 2003). Blinding of outcome assessors was not common, but seven studies did take this step (Comstock 1982; De Silva 2003; Eddleston 2008; James 1995; Montoya‐Cabrera 1999; Roberts 2006; Underhill 1990).

Incomplete outcome data

Only three studies were at high risk of attrition bias (Comstock 1982; Crome 1983; Passeron 1989), and three were at unclear risk (Amigó Tadín 2002; Behnoush 2009; Bouget 1989). All other studies showed no evidence of incomplete outcome data.

Selective reporting

Overall there was a high risk of reporting bias. Only a third of the studies were at low risk (Amigó Tadín 2002; Cooper 2005; Eddleston 2008; Hultén 1988; Kornberg 1991; Merigian 1990; Roberts 2006; Wax 1999).

Other potential sources of bias

Fifteen studies were at low risk of other potential sources of bias, and we assessed nine studies as being at high risk of bias for reasons other than those mentioned above.

In Albertson 1989, actual poisoning was not verified for 25% of the participants by means other than history. Furthermore, in Wax 1999, there was no confirmation of actual ingestion or uptake of the drug.

In another study, investigators suspected a clinical difference between the groups receiving the MDAC intervention versus the SDAC control, based on divergent carbamazepine kinetics during the initial six hours of the treatment period, when both groups had received only one dose of activated charcoal (Brahmi 2006). Also in Montoya‐Cabrera 1999, the hepatic toxicity marker values suggest there might be a clinically meaningful difference between the two treatment groups. This could create a bias in effectiveness of the treatment, because of differences in degree and type of poisoning.

In Comstock 1982 there was a potential bias in the selection of the study population since participants were selected at the discretion of the attending physician.

Crome 1983 did not find significant amounts of any drugs in 11 of the 48 participants, and 7 of them had not taken any tricyclic antidepressant (although this was a criterion for inclusion). Furthermore, the role of the study funder was not clear.

Two studies included only asymptomatic participants, who are less likely to experience a benefit from any treatment (Merigian 1990; Wax 1999).

Merigian 2002 performed only post hoc analyses according to clinical severity, and there was no follow‐up after discharge from the hospital.

In Roberts 2006 it is not entirely clear, even to the authors, what exactly is measured with the digoxin assay used in the study. The fact that both active cardenolides and metabolites might be detected by the assay compromise the results of these analyses, potentially explaining the wide variability observed. Furthermore, only participants with mild intoxication were included in this analysis, as the severe cases were treated with Fab antitoxin or transferred to a tertiary hospital, but these patients might have shown the biggest effect (Roberts 2006).

In Wax 1999, the authors reported dichotomous outcomes but performed measurement using an ordinal scale.

Effects of interventions

See summary of findings Table for the main comparison; summary of findings Table 2; summary of findings Table 3; summary of findings Table 4; summary of findings Table 5; summary of findings Table 6; Appendix 2; Appendix 3; Appendix 4; Appendix 5; Appendix 6; Appendix 7.

A. First aid interventions that limit or delay the absorption of the poison in the body

1. Single‐dose activated charcoal versus no intervention

Two studies compared a single dose of activated charcoal versus no intervention (Merigian 1990; Underhill 1990). Underhill 1990 included 25 participants presenting in the emergency department with acute paracetamol overdose. Recruitment in the control group was stopped early for ethical reasons, as blood levels of paracetamol kept rising over time. Merigian 1990 included 820 participants presenting at the emergency department with self‐reported oral overdose in general. This study subdivided participants into a symptomatic and an asymptomatic group, for which treatments differed. Only the 451 asymptomatic participants, who received either a single dose of activated charcoal or were kept for observation, were within scope of this review. For a detailed summary of the outcomes, we refer to the Data and analyses section. Below we provide a narrative overview. See summary of findings Table for the main comparison.

Primary outcomes

Incidence of mortality

The identified studies either did not collect or did not report outcomes related to mortality.

Adverse events

The only adverse event Underhill 1990 reported in response to SDAC was vomiting, which occurred in 4/20 participants compared to 0/5 in the control group (Peto OR 4.17, 95% CI 0.30 to 57.26; Analysis 1.1). Merigian 1990 reported no adverse events in any treatment group (451 participants). We assessed this evidence as being of very low certainty because of limitations in study design, imprecision due to a low event number and indirectness.

Incidence and severity of poisoning symptoms

Only Merigian 1990 reported an outcome related to symptom severity. In 451 asymptomatic participants presenting to the emergency department, no participants experienced events of clinical deterioration in either group (Table 3). We assessed this evidence as being of very low certainty because of limitations in study design, imprecision due to a low event number and indirectness.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Table 3. Additional pre‐defined outcomes reported in the included studies

A. First aid interventions that limit or delay the absorption of the poison in the body
SDAC vs no intervention
Incidence of clinical deterioration
	SDAC		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Merigian 1990	0	220	0	231	Not estimable
Incidence of ICU admission
	SDAC		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Merigian 1990	0	220	0	231	7.77 (0.15 to 391.93)
SDAC + hospitalintervention vs hospitalintervention
Incidence of clinical deterioration
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	Peto OR (95% CI)
Merigian 2002	0	455	0	1075	Not estimable
Grade of coma (4 h after admission)
	SDAC + hospital treatment (N = 9)		Hospital treatment (N = 7)
Study	Median	IQR	Median	IQR	Median difference (P value)
Crome 1983	2	(1 to 3)	2	(1 to 2.5)	0 (P = 0.55)
Grade of coma (8 h after admission)
	SDAC + hospital treatment (N = 9)		Hospital treatment (N = 7)
Study	Median	IQR	Median	IQR	Median difference (P value)
Crome 1983	2	(1 to 3)	1	(0.5 to 2)	1 (P = 0.38)
Grade of coma (24 h after admission)
	SDAC + hospital treatment (N = 9)		Hospital treatment (N = 7)
Study	Median	IQR	Median	IQR	Median difference (P value)
Crome 1983	1	(0 to 2)	0	(0 to 0.5)	1 (P = 0.27)
Incidence of coma grade III (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	7	34	6	43	1.48 (0.55 to 3.98)
Incidence of coma grade IV (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	7	43	0.18 (0.02 to 1.40)
Incidence of coma grade III (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	2	34	2	43	1.26 (0.19 to 8.52)
Incidence of coma grade IV (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	0	34	2	43	0.16 (0.01 to 2.70)
Incidence of coma grade III (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	Peto OR (95% CI)
Hultén 1988	0	34	0	43	Not estimable
Incidence of coma grade IV (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	Peto OR (95% CI)
Hultén 1988	0	34	0	43	Not estimable
Incidence of need for cardiac pacing/antitoxin
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Eddleston 2008	101	549	101	555	1.01 (0.79 to 1.30)
Incidence of need for respirator
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	2	34	9	43	0.28 (0.06 to 1.22)
Incidence of systolic blood pressure < 100 mmHg (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	4	34	8	43	0.63 (0.21 to 1.92)
Incidence of systolic blood pressure < 100 mmHg (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	5	43	0.25 (0.03 to 2.06)
Incidence of systolic blood pressure < 100 mmHg (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	0	34	2	43	0.16 (0.01 to 2.70)
Incidence of heart rate > 100 bpm (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	10	34	15	43	0.84 (0.44 to 1.63)
Incidence of heart rate > 100 bpm (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	8	34	10	43	1.01 (0.45 to 2.28)
Incidence of heart rate > 100 bpm (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	10	34	10	43	1.26 (0.60 to 2.68)
Incidence of cardiac arrhythmias (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	4	43	0.32 (0.04 to 2.70)
Incidence of cardiac arrhythmias (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	3	43	0.42 (0.05 to 3.87)
Incidence of cardiac arrhythmias (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	2	43	0.63 (0.06 to 6.68)
Incidence of intubation > 8 h
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	4	34	9	43	0.56 (0.19 to 1.67)
B. First aid interventions that evacuate the poison from the gastrointestinal tract
Syrup of ipecac vs no intervention
Incidence of referrals to the emergency department
	Syrup of ipecac		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Wax 1999	1	51	0	52	7.54 (0.15 to 378.83)
Incidence of hospitalizations
	Syrup of ipecac		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Wax 1999	0	51	0	52	Not estimable
SDAC + cathartic vs SDAC (higher dose)
Incidence of adverse events (6 mL vs 4 mL)
	SDAC + 6 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	0	16	0	16	Not estimable
Incidence of adverse events (8 mL vs 4 mL)
	SDAC + 8 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	0	18	0	16	Not estimable
Incidence of adverse events (8 mL vs 6 mL)
	SDAC + 8 mL cathartic		SDAC + 6 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	0	18	0	16	Not estimable
Incidence of hospitalization (6 mL vs 4 mL)
	SDAC + 6 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	1	16	0	16	7.39 (0.15 to 372.38)
Incidence of hospitalization (8 mL vs 4 mL)
	SDAC + 8 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	2	18	0	16	7.01 (0.42 to 117.63)
C. Combined first aid interventions that limit uptake and promote removal of the poison
MDAC + cathartic + hospitalintervention vs hospitalintervention
Incidence of adverse events
	MDAC + cathartic + hospital intervention		Hospital intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Montoya‐Cabrera 1999	0	7	0	7	Not estimable
T_1/2 (h)
	MDAC + cathartic + hospital intervention (N = 7)		Hospital intervention (N = 7)
Study	Mean	SD	Mean	SD	MD (95% CI)
Montoya‐Cabrera 1999	10	N/A	17	N/A	−7 (not estimable)

bpm: beats per minute; CI: confidence interval; IQR: interquartile range; OR: odds ratio; RR: risk ratio; MDAC; multi‐dose activated charcoal; SDAC: single‐dose activated charcoal.

Secondary outcomes

Duration of symptoms

The identified studies either did not collect or did not report outcomes related to symptom duration.

Drug absorption

Underhill 1990 measured drug levels of paracetamol before treatment and at several time points after treatment. However, the study reported none of the pre‐defined outcomes of interest in our protocol, nor could we derive them from the data provided; therefore we could make no reliable estimation of drug absorption from the available data.

Incidence of hospitalization

The identified studies either did not collect or did not report outcomes related to incidence of hospitalization.

Incidence of ICU admission

Merigian 1990 reported the number of participants admitted to the ICU department, which was 1/220 in the intervention group and 0/231 participants in the control group (Peto OR 7.77; 95% CI 0.15 to 391.93; Table 3). We assessed this evidence as being of very low certainty, downgraded for limitations in study design, imprecision due to a low number of events and indirectness.

Summary for this comparison

We were uncertain about the effect of SDAC compared to no intervention on the incidence of ICU admission or the incidence of clinical deterioration (very low certainty due to limitations in study design, imprecision and indirectness). One study described a single type of adverse event in response to the treatment, vomiting, but we are uncertain about the effect (very low‐certainty evidence due to limitations in study design, imprecision and indirectness).

2. Single‐dose activated charcoal plus hospital intervention versus hospital intervention alone

Seven trials used single‐dose activated charcoal (SDAC) in adjuvant to established hospital treatments (Comstock 1982; Cooper 2005; Crome 1983; Eddleston 2008; Hultén 1988; Merigian 2002; Roberts 2006). These hospital treatments consisted of supportive treatments to maintain vital parameters plus poison‐specific treatments, but in most studies this also included gastric lavage (Comstock 1982; Crome 1983; Eddleston 2008; Hultén 1988; Roberts 2006). Crome 1983 and Hultén 1988 specifically included participants with tricyclic antidepressant overdose, while the other studies did not define a specific toxic syndrome. Roberts 2006 investigated drug uptake in a subpopulation of participants entering the Eddleston 2008 study, with yellow oleander seed poisoning. For a detailed summary of the outcomes, we refer to the Data and analyses section. Below we provide a narrative overview. See summary of findings Table 2.

Primary outcomes

Incidence of mortality

Two studies reported the impact of SDAC on mortality (Cooper 2005; Eddleston 2008). Both studies included participants with a variety of toxic syndromes, so we considered it appropriate to pool these results. Moreover, from a layperson's perspective, it is usually impossible to distinguish different toxic syndromes, let alone to decide on the appropriateness of administering SDAC in case of a specific syndrome. The statistical results were mainly determined by the large study, Eddleston 2008 (Peto OR 1.04, 95% CI 0.79 to 1.37; 3425 participants; 2 studies; Analysis 2.1). We did not identify important heterogeneity (P = 0.30, I² = 7%). Evidence was of low certainty, downgraded for imprecision due to wide confidence intervals and indirectness.

Adverse events

Three studies reported on adverse events in relation to administering SDAC (Cooper 2005; Eddleston 2008; Merigian 2002). Two studies reported the occurrence of vomiting (Cooper 2005; Merigian 2002), while Eddleston 2008 reported the absence of bowel sounds as a proxy for constipation. We considered a combined estimate of adverse events to be appropriate, given the wide variety in toxic syndromes included in the different studies and the inability of laypeople to distinguish between different toxic syndromes. However, this resulted in considerable heterogeneity (P = 0.002, I² = 83%) with different directions of effect. We performed a subgroup analysis by reported symptom, showing between‐group differences (P = 0.02, I² = 83%), which decreased but did not fully eliminate the heterogeneity. The risk ratio for the outcome occurrence of vomiting was 1.44 (95% CI 0.88 to 2.37; 1806 participants; 2 studies) and still showed substantial unexplained heterogeneity (P = 0.08, I² = 68%). The RR for the sub‐outcome, absence of bowel sounds, was 0.41 (95% CI 0.17 to 1.00; 3098 participants; 1 study; Analysis 2.2). The evidence on adverse events was of very low certainty because of inconsistency between studies, imprecision (low number of events) and indirectness.

Incidence and severity of symptoms of poisoning

One study evaluated the incidence of clinical deterioration, which was absent in both an intervention group of 455 participants and a control group of 1075 participants (Merigian 2002; Table 3). This evidence was of very low certainty, downgraded due to limitations in study design, imprecision (low number of events) and indirectness.

Crome 1983 and Hultén 1988 expressed the grade of coma using the Matthew‐Lawson coma scale at 4 h, 8 h and 24 h after hospital admission. The median coma score scales of intervention and control groups were similar in the study of 16 participants (Crome 1983; Table 3). The proportion of participants with a coma scale of III or IV were also similar in Hultén 1988, with 77 participants (Table 3). Evidence was of very low certainty, downgraded for limitations in study design, imprecision due to a low number of events and a low sample size and indirectness.

Eddleston 2008 reported the incidence of participants with yellow oleander seed poisoning, who needed specialized treatment, namely cardiac pacing or Fab antitoxin treatment (Eddleston 2008). The risk ratio was 1.01 (95% CI 0.79 to 1.30, P = 0.93, 1104 participants; Table 3). This evidence was of low certainty, downgraded for imprecision due to a low number of events and indirectness.

Four studies reported on the need for intubation and/or ventilation (Cooper 2005; Eddleston 2008; Hultén 1988; Merigian 2002). In Merigian 2002 these data were only available for participants admitted to the ICU. We considered it appropriate to combine the results, given the multiple or unspecified toxic syndromes included in three out of four individual studies, and the inability of laypeople to distinguish between ingested toxins.The combined result showed substantial heterogeneity and different directions of effect, so we do not present it (P = 0.04; I² = 63%; Analysis 2.3). A possible reason for the observed heterogeneity may have been the co‐interventions, i.e. whether participants received gastric lavage as part of the hospital treatments. A subgroup analysis suggested there may be between‐group differences between participants who received gastric lavage prior to receiving SDAC and those who did not (P = 0.005, I² = 87.6%). The summary estimate in the subgroup with gastric lavage was RR 0.95 (95% CI 0.70 to 1.27, 3175 participants, 2 studies, P = 0.71). In the subgroup without gastric lavage the RR was 2.61 (95% CI 1.38 to 4.93, 387 participants, 2 studies, P = 0.003), in favour of not receiving SDAC. In addition, Hultén 1988 made a subcomparison of the need for ventilation with a respirator (RR 0.28, 95% CI 0.06 to 1.22, 77 participants, P = 0.09; Table 3). Evidence on ventilation was of low certainty, downgraded for imprecision (low number of events and wide confidence intervals) and indirectness.

Two studies, Eddleston 2008 and Hultén 1988, studied the incidence of convulsions. We considered combining these outcomes appropriate, given the inability of laypeople to make a distinction between different toxic syndromes. A combined estimate of these studies, however, had substantial heterogeneity (P = 0.03; I² = 79%) and a different direction of effect, so we only report the individual study results. The individual estimates were RR 1.87 (95% CI 0.75 to 4.67, 3098 participants; Eddleston 2008) and RR 0.28 (95% CI 0.06 to 1.22, 77 participants; Hultén 1988; Analysis 2.4). Exploring heterogeneity in a meta‐analysis with only two studies is difficult, so explanations for the observed differences between studies remain speculative. Possible reasons might be the small study sample in Hultén 1988. Alternatively, true differences in patient population might explain the differences, as Hultén 1988 specifically recruited participants with tricyclic antidepressant poisoning, while most participants in Eddleston 2008 took an overdose of pesticides or yellow oleander seeds. This evidence was of very low certainty, downgraded for inconsistency between studies, imprecision (wide confidence intervals and low number of events) and indirectness.

Hultén 1988 recorded some additional clinical parameters in their 77 participants, as measures of poisoning severity: systolic blood pressure, heart rate and incidence of cardiac arrhythmias, at 4 h, 8 h and 24 h after treatment (Table 3). These numbers were not shown to differ between treatments. The certainty of evidence for these parameters was rated as very low, due to limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Secondary outcomes

Duration of symptoms

Three studies measured length of intubation (Eddleston 2008; Hultén 1988; Merigian 2002). Combination of the studies' results in a meta‐analysis was not feasible, due to differences in reporting. Eddleston 2008 reported medians with interquartile ranges, and there was no demonstrable difference. Authors did not report the number of participants in this analysis. Merigian 2002 reported length of intubation as means without a measure of spread or the number of participants in the analysis, and likewise, authors could not show a difference between intervention and control (Analysis 2.5). Hultén 1988 reported the proportion of participants that were intubated for longer than 8 h, which was similar between intervention and control (RR 0.56, 95% CI 0.19 to 1.67, 77 participants, P = 0.30; Hultén 1988; Table 3). This evidence was rated to be of low certainty, downgraded for imprecision due to a low number of events or lack of data, plus indirectness.

Drug absorption

Three included studies measured drug absorption (Comstock 1982; Hultén 1988; Roberts 2006). Comstock 1982, with 339 participants, only reported increases in blood drug concentrations over time, without reporting any of our pre‐specified outcomes of interest. Hultén 1988 presented the course of tricyclic antidepressant levels in the blood of 77 participants graphically, and reported narratively that there was no demonstrable difference in AUC, C_max or T_1/2 between treatments (Hultén 1988). Therefore, the only numeric data available were on the AUC, C_max and T_max of cardenolides from yellow oleander seeds, measured by Roberts 2006 in a subset of 68 participants from the Eddleston study. Authors reported results as median with interquartile ranges (IQR) and could not show a difference between treatments (Analysis 2.6; Analysis 2.7; Analysis 2.8). The evidence on the pharmacokinetic parameters was of very low certainty, due to limitations in study design, imprecision (low sample size) and indirectness.

Incidence of hospitalization

Merigian 2002 reported the incidence of hospitalization. The results favoured no treatment over SDAC (RR 1.57, 95% CI 1.22 to 2.02, 1479 participants, P < 0.001; Analysis 2.9). We assessed this evidence as being of very low certainty, downgraded for limitations in study design, imprecision (low number of events) and indirectness.

Incidence of ICU admission

Merigian 2002 reported the incidence of ICU admission in favour of no treatment with SDAC (RR 2.33, 95% CI 1.42 to 3.82, 1479 participants, P < 0.001; Analysis 2.10). This evidence was of very low certainty, downgraded for limitations in study design, imprecision (low number of events) and indirectness.

Summary for this comparison

SDAC as adjuvant to supportive hospital treatments may have little or no influence on one of our primary outcomes, incidence of mortality, while we are uncertain about its effect on another primary outcome, adverse events due to the intervention. In addition, SDAC plus hospital treatments may have little or no influence on the primary outcomes of need for intubation, need for cardiac pacing or antitoxin treatment in cases of yellow oleander poisoning, or the secondary outcome, length of intubation. We are uncertain about the effect of SDAC in addition to hospital treatments on the incidence of clinical deterioration, Matthew‐Lawson coma scale scores, incidence of convulsions, blood pressure, heart rate, cardiac arrhythmias, and the secondary outcomes of drug absorption and incidence of hospital or ICU admission. The evidence collected is of low to very low certainty, due to limitations in study design, indirectness and/or imprecision.

3. Multi‐dose activated charcoal plus hospital intervention versus single‐dose activated charcoal plus hospital intervention

Five trials compared single‐dose (SDAC) versus multi‐dose activated charcoal (MDAC), in adjuvant to hospital treatments (Behnoush 2009; Brahmi 2006; De Silva 2003; Eddleston 2008; Roberts 2006). The identified trials studied the following toxic overdoses: carbamazepine (Behnoush 2009; Brahmi 2006), yellow oleander (De Silva 2003; Roberts 2006), and a combination of toxic syndromes (Eddleston 2008). In all studies except for Brahmi 2006, supportive treatments included gastric lavage. Behnoush 2009 did not report any outcomes of interest for this review. See summary of findings Table 3.

Primary outcomes

Incidence of mortality

Incidence of mortality was an outcome of interest in De Silva 2003 and Eddleston 2008. We considered it appropriate to combine the findings given the similar populations studied. Nevertheless, the meta‐analysis resulted in an estimate with substantial heterogeneity (P = 0.04, I² = 76%), albeit the same direction of effect. The pooled risk ratio was 0.59 (95% CI 0.21 to 1.63; 3476 participants; 2 studies; Analysis 4.1). Reasons for the observed heterogeneity were not immediately clear, as both the provided interventions and the population studied are remarkably similar. Eddleston 2008 included a broader range of toxic syndromes; however, including only the subpopulation of participants with yellow oleander poisoning (SDAC: 26/549 and MDAC: 23/541) would not change the conclusions made (P = 0.06, I² = 71%). There are some factors that might explain the differences observed, such as a longer treatment in De Silva 2003 (activated charcoal up to 72 h) compared to Eddleston 2008 (activated charcoal up to 24 h) or differences in the compliance rate with the treatment (reported to drop to 66% by the final dose by Eddleston 2008 but claimed to be ensured in all cases by De Silva 2003). However, these explanations remain speculative, and if there are more studies in a future update, we may be able to show more robust evidence and clarify the heterogeneity issue. We considered the evidence here to be of very low certainty, due to inconsistency between studies, imprecision (wide confidence intervals) and indirectness.

Adverse events

Two studies reported the incidence of adverse events in response to the intervention (De Silva 2003; Eddleston 2008). We considered combining the results of these studies appropriate, given the similar study population. The combined result (Peto OR 3.55, 95% CI 1.85 to 6.79; 3476 participants; 2 studies) contained a substantial degree of heterogeneity (P = 0.08, I² = 66.8%), but with the same direction of effect. Reasons for this heterogeneity remain speculative, but it could be due to differing definitions and diagnostic methods for adverse events, for example absent bowel sounds in Eddleston 2008 (Peto OR 2.34, 95% CI 1.05 to 5.21; 3075 participants, P = 0.04) versus abdominal discomfort/diarrhoea in De Silva 2003 (Peto OR 7.82, 95% CI 2.59 to 23.58; 401 participants, P < 0.001; Analysis 4.2). Both studies suggest that the number of adverse events may increase in case of MDAC, compared to SDAC. This evidence was of low certainty, downgraded for imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence and severity of symptoms of poisoning

Both De Silva 2003 and Eddleston 2008 reported on the need for cardiac pacing or administration of a Fab antitoxin, treatments used in severe cases of yellow oleander poisoning. We considered it appropriate to combine results, given the similar patient populations studied. The combined effect estimate (RR 0.26, 95% CI 0.02 to 4.18; 1490 participants; 2 studies) resulted in considerable heterogeneity (P = 0.005, I² = 87%; Analysis 4.3). As for mortality, the reasons for the observed heterogeneity are not immediately clear and remain speculative. We considered evidence to be of very low certainty, due to imprecision (low numbers of events and wide confidence intervals), inconsistency between studies and indirectness.

De Silva 2003 also recorded the incidence of life‐threatening arrhythmias after 24 h, which may be lower for the group receiving SDAC (RR 0.21, 95% CI 0.06 to 0.71, 385 participants, P = 0.01; Analysis 4.4). Evidence was of low certainty, downgraded for imprecision due to a low number of events and indirectness.

Another outcome De Silva 2003 reported in their study with 401 participants was the need for atropine, expressed as both the amount of atropine administered and the number of boluses administered. Both the amount (mg) of atropine administered (MD −1.60, 95% CI −2.25 to −0.95, P < 0.001; Analysis 4.5) and the median number of boluses (Analysis 4.6) were higher in the group receiving SDAC. We assessed the evidence as being of moderate certainty, downgraded for indirectness.

Two studies reported the need for intubation (Brahmi 2006; Eddleston 2008). We decided to combine these results, as from the point of view of laypeople, the focus of this review, it is usually impossible to distinguish between toxic syndromes or adapt the provided intervention accordingly. There may be little or no effect on the need for intubation (RR 1.01, 95% CI 0.75 to 1.38, 3097 participants, P = 0.93; Analysis 4.7). There was no heterogeneity between studies (P = 0.98, I² = 0%). The evidence was of low certainty, downgraded due to imprecision (low number of events and wide confidence intervals) and indirectness.

Finally, Eddleston 2008 reported the incidence of convulsions (RR 1.09, 95% CI 0.52 to 2.32, 3085 participants, P = 0.82; Analysis 4.8). Evidence was of low certainty, downgraded due to imprecision (low number of events and wide confidence intervals) and indirectness.

Secondary outcomes

Duration of symptoms

One study in six participants reported the duration of coma as a measure of symptom duration (Brahmi 2006). The mean difference was −9.00 h (95% CI −14.79 to −3.21, P = 0.002), in favour of MDAC (Analysis 4.9). We considered this to be evidence of very low certainty, downgraded for limitations in study design, imprecision (small sample size) and indirectness.

Brahmi 2006 and Eddleston 2008 also reported duration of intubation. Due to differences in reporting, it was not possible to combine the estimates. The mean difference reported by Brahmi 2006 is −12.30 h (95% CI −18.56 to −6.04, 6 participants, P < 0.001). Eddleston 2008 reported medians with IQR; however, authors made no statement of effectiveness. The number of participants in the analysis was not clear either (Analysis 4.10). This evidence was of very low certainty, downgraded for inconsistency between studies, imprecision (low sample size and lack of data) and indirectness.

Drug absorption

Two studies, Brahmi 2006 and Roberts 2006, reported on pharmacokinetic parameters as measures of drug absorption. Differences in reporting precluded meta‐analysis.

Both studies reported the C_max. The mean difference reported by Brahmi 2006 was 0.40 mg/L (95% CI −4.89 to 5.69, 6 participants, P = 0.88). Roberts 2006 reported C_max, as medians with IQR, demonstrating no difference between intervention and control (participants = 64, P > 0.05; Analysis 4.11). AUC and T_max were similar between treatments (64 participants, P > 0.05; Analysis 4.12; Analysis 4.14). In addition, Brahmi 2006 measured the T_1/2 which was in favour of MDAC (MD −15.32 h (95% CI −21.84 to −8.80, 6 participants, P < 0.001; Analysis 4.13). The evidence was of very low certainty, downgraded for limitations in study design, imprecision (low sample size) and indirectness.

Incidence of hospitalization

The identified studies either did not collect or did not report outcomes related to incidence of hospitalization.

Incidence of ICU admission

De Silva 2003 reported the incidence of ICU admissions. The risk ratio suggests MDAC may have a strong effect on reducing ICU admissions (RR 0.31, 95% CI 0.12 to 0.83, 401 participants, P = 0.02; Analysis 4.15). The presented evidence was of low certainty, downgraded for imprecision (low number of events) and indirectness.

Summary for this comparison

The evidence that we have collected concerning the use of single‐ versus multi‐dose activated charcoal in adjuvant to hospital treatments is of moderate to very low certainty. There may be little or no difference in the incidence of convulsions and the need for intubation between MDAC plus hospital treatments and SDAC plus hospital treatments. On the other hand, there may be a favourable effect for MDAC on the incidence of life‐threatening cardiac arrhythmias and ICU admissions, while it probably decreases the number of atropine boluses and total amount of atropine administered. Low‐certainty evidence suggests, however, that MDAC may come with an increased risk of adverse events. We are uncertain about the effects of MDAC plus hospital treatments on mortality, the need for cardiac pacing or antitoxin treatment, symptom duration and drug absorption.

4. Single‐dose activated charcoal versus syrup of ipecac

One study with 34 participants compared SDAC versus syrup of ipecac in oral poisoning participants with mild levels of intoxication (defined as a Glasgow Coma Scale score of more than 12; Amigó Tadín 2002). The study specifically included participants presenting with oral overdoses of anti‐inflammatory, psychotropic or analgesic drugs. Participants received no additional treatments. See summary of findings Table 4.

Primary outcomes

Incidence of mortality

The identified study did not collect outcomes related to mortality.

Adverse events

Amigó Tadín 2002 reported the incidence of adverse events encountered (RR 1.24, 95% CI 0.26 to 5.83, 34 participants, P = 0.79; Analysis 3.1). Evidence was of very low certainty, downgraded for limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence and severity of symptoms of poisoning

The identified study measured participants' poisoning symptoms 1 h after treatment (Amigó Tadín 2002). The mean difference in Glasgow Coma Scale scores between treatments was −0.15 (95% CI −0.43 to 0.13, 34 participants, P = 0.29; Analysis 3.2). Furthermore, the study reported mean arterial blood pressure (MD 7.00 mmHg, 95% CI −3.56 to 17.56, 34 participants, P = 0.19; Analysis 3.3), heart rate (MD −2.39 bpm, 95% CI −15.58 to 10.80, 34 participants, P = 0.72; Analysis 3.4) and respiratory rate (MD 1.12 breaths/min, 95% CI −1.69 to 3.93, 34 participants, P = 0.44; Analysis 3.5) as measures of intoxication. The collected evidence was of very low certainty, downgraded for limitations in study design, imprecision (low sample size) and indirectness.

Secondary outcomes

Duration of symptoms

The identified study did not collect outcomes related to symptom duration.

Drug absorption

The identified study did not collect outcomes related to drug absorption.

Incidence of hospitalization

The identified study did not collect outcomes related to incidence of hospitalization.

Indicence of ICU admission

The identified study did not collect outcomes related to incidence of ICU admission.

Summary for this comparison

We identified evidence of very low certainty, originating from one study (Amigó Tadín 2002). We are uncertain about any difference between SDAC and syrup of ipecac concerning poisoning symptoms or incidence of adverse events.

5. Multi‐dose activated charcoal plus hospital treatment versus hospital treatment alone

We found three studies comparing the administration of MDAC in adjuvant to hospital treatments versus hospital treatments alone (Bouget 1989; Eddleston 2008; Roberts 2006). For most participants in Eddleston 2008, who had a variety of intoxications, hospital treatments included gastric lavage. Roberts 2006 studied a subgroup of participants from the Eddleston study, those with yellow oleander seed poisoning. Bouget 1989 included 36 participants with benzodiazepine poisoning, but no numeric outcomes were reported in this study. See summary of findings Table 5.

Primary outcomes

Incidence of mortality

Eddleston 2008 reported the incidence of mortality, which may not differ between intervention and control (RR 0.94, 95% CI 0.72 to 1.22, 3085 participants, P = 0.64; Analysis 5.1). The evidence was of low certainty, downgraded for imprecision (low number of events and wide confidence intervals) and indirectness.

Adverse events

Eddleston 2008 reported one adverse event, incidence of absent bowel sounds (RR 1.02, 95% CI 0.52 to 1.98, 3085 participants, P = 0.97; Analysis 5.2). Evidence was of low certainty, downgraded for imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence and severity of symptoms of poisoning

Eddleston 2008 reported several markers of intoxication, including the need for intubation (RR 0.97, 95% CI 0.71 to 1.33, 3085 participants, P = 0.87; Analysis 5.3), seizures (RR 2.03, 95% CI 0.82 to 5.02, 3085 participants, P = 0.12; Analysis 5.4) and need for cardiac pacing/Fab antitoxin treatment in participants with yellow oleander poisoning (RR 0.86, 95% CI 0.66 to 1.13, 1095 participants, P = 0.28; Analysis 5.5). Evidence was of low certainty, downgraded for imprecision (low number of events and wide confidence intervals) and indirectness.

Secondary outcomes

Duration of symptoms

Eddleston 2008 reported length of intubation as a measure of symptom duration. It was expressed as a median plus IQR, and it may not be different between groups (Analysis 5.6). The evidence was of low certainty, downgraded for imprecision (lack of data) and indirectness.

Drug absorption

Roberts 2006 analysed cardenolide pharmacokinetic parameters in a subgroup of 76 participants from the Eddleston study, with yellow oleander poisoning. AUC, C_max and T_max were reported as medians with IQR, and there was no demonstrable difference between treatment groups (Analysis 5.7; Analysis 5.8; Analysis 5.9). Evidence was of very low certainty due to limitations in study design, imprecision (low sample size) and indirectness.

Incidence of hospitalization

The identified studies either did not collect or did not report outcomes related to incidence of hospitalization.

Incidence of ICU admission

The identified studies either did not collect or did not report outcomes related to incidence of ICU admission.

Summary for this comparison

The identified evidence on the use of MDAC in addition to hospital treatment is of low to very low certainty and originates from two studies, one of which is a subgroup analysis of the larger study (Eddleston 2008; Roberts 2006). MDAC plus hospital treatments may not be better than hospital treatments alone for mortality, symptoms of intoxication (need for intubation, seizures, cardiac pacing or antitoxin treatment), duration of intubation or absence of bowel sounds as adverse events. We are uncertain about the effects of MDAC on pharmacokinetic parameters of cardenolides.

B. First aid interventions that evacuate the poison from the gastrointestinal tract

B1. Emetics

6. Syrup of ipecac versus no intervention

One study compared ipecac versus observation in 103 cases of paediatric, asymptomatic toxic berry ingestion (Wax 1999). This study took place in a pre‐hospital setting and did not include any other treatments. See summary of findings Table 6.

Primary outcomes

Incidence of mortality

The identified study did not collect outcomes related to incidence of mortality.

Adverse events

The identified study reported on the incidence of several adverse events separately. As it was likely that one patient could encounter multiple adverse events, it was not possible to combine these into a composite outcome. Therefore, we present the risk ratios for the individually described adverse events. Participants receiving ipecac may show an increased risk of diarrhoea (RR 4.08, 95% CI 1.66 to 10.04, 103 participants, P = 0.002; Analysis 6.1) and sedation (RR 5.10, 95% CI 1.17 to 22.13, 103 participants, P = 0.03; Analysis 6.3), while there may be little or no difference for abdominal pain (RR 1.02, 95% CI 0.07 to 15.87, 103 participants, P = 0.99; Analysis 6.2) or agitation (RR 1.53, 95% CI 0.27 to 8.77, 103 participants, P = 0.63; Analysis 6.4). The evidence was of low certainty, downgraded for limitations in study design and imprecision (low number of events and wide confidence intervals).

Incidence and severity of symptoms of poisoning

The identified study did not collect outcomes related to incidence and severity of poisoning symptoms.

Secondary outcomes

Duration of symptoms

The identified study did not collect outcomes related to duration of poisoning symptoms.

Drug absorption

The identified study did not collect outcomes related to drug absorption.

Incidence of hospitalization

One patient in the intervention group was referred to the emergency department, compared to none in the control group (Wax 1999). The Peto OR was 7.54 (95% CI 0.15 to 379.83, 103 participants, P = 0.31; Table 3). None of the 103 participants were hospitalized (Table 3). The evidence was of low certainty, downgraded for limitations in study design and imprecision due to a small sample size and wide confidence intervals.

Incidence of ICU admission

As none of the participants were hospitalized, none could have been admitted to the ICU.

Summary for this comparison

We identified one study that compared syrup of ipecac versus home observation (Wax 1999). Authors did not report any clinical outcomes, but there may be little or no difference in emergency department referrals. In contrast, the incidence of adverse events (diarrhoea and sedation) may be larger. Evidence is of low certainty (downgraded for limitations in study design and imprecision).

7. Syrup of ipecac plus single‐dose activated charcoal plus cathartics versus single‐dose activated charcoal plus cathartics

We identified four studies that compared SDAC plus a cathartic (sorbitol or magnesium sulphate), preceded or not preceded by syrup of ipecac in participants presenting to an emergency department (Albertson 1989; Kornberg 1991; Kulig 1985; Pond 1995). None of the studies selected participants on a specific toxic syndrome. Kornberg 1991 focused on children under 6 years old, while Albertson 1989 and Pond 1995 included adults (specified as more than 18 years old or more than 13 years old, respectively). Kulig 1985 did not specify a certain age range. See Appendix 2.

Primary outcomes

Incidence of mortality

Two of the included studies reported on incidence of mortality, but neither study noted any events in the 573 participants across treatment groups (Kornberg 1991; Kulig 1985; Analysis 7.1). We considered this evidence to be of very low certainty, downgraded for limitations in study design, imprecision (low number of events) and indirectness.

Adverse events

Two studies reported the incidence of adverse events (Albertson 1989; Pond 1995), while one study reported the number of activated charcoal that was vomited as an adverse event (Kornberg 1991). We combined these outcomes, as all studies included a wide variety of toxic syndromes. The meta‐analysis favoured not using ipecac (RR 2.59, 95% CI 1.37 to 4.91, 764 participants, 3 studies, P = 0.003). We found no important heterogeneity (P = 0.29, I² = 19%; Analysis 7.2). This was evidence of very low certainty, downgraded for limitations in study design, imprecision (low number of events) and indirectness.

Incidence and severity of symptoms of poisoning

Three studies reported on the incidence of participants with clinical improvement during their stay in the emergency department (Kornberg 1991; Kulig 1985; Pond 1995). We considered a meta‐analysis appropriate, given the wide variety of toxic syndromes included in the individual studies. The combined risk ratio was 1.00 (95% CI 0.83 to 1.21, 989 participants, 3 studies, P = 0.98), without evidence of important heterogeneity (P = 0.21, I² = 36%; Analysis 7.3). The evidence was of low certainty, downgraded for limitations in study design and indirectness.

Two trials studied the incidence of clinical deterioration during the emergency department stay (Kulig 1985; Pond 1995); we combined these in a meta‐analysis, given the wide variety of toxic syndromes included in the individual studies. The pooled RR was 0.88 (95% CI 0.46 to 1.69, 970 participants, 2 studies, P = 0.70), with no apparent heterogeneity (P = 0.38, I² = 0%; Analysis 7.4). We considered this to be evidence of very low certainty, downgraded for limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Secondary outcomes

Duration of symptoms

The identified studies either did not collect or did not not report outcomes related to duration of poisoning symptoms.

Drug absorption

The identified studies either did not collect or did not report outcomes related to drug absorption.

Incidence of hospitalization

The incidence of hospitalization was an outcome of interest in three studies (Albertson 1989; Kornberg 1991; Kulig 1985). Given the wide variety of toxic syndromes included in the individual studies, and the inability of laypeople to distinguish between ingested toxins, we considered a meta‐analysis appropriate. The Peto OR was 1.17 (95% CI 0.69 to 1.98, 746 participants, 3 studies, P = 0.56), without any important heterogeneity (P = 0.15, I² = 47%; Analysis 7.5). We considered this evidence to be of very low certainty, due to limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence of ICU admission

One trial, Albertson 1989, reported the incidence of ICU admission (RR 1.38, 95% CI 0.44 to 4.38, 200 participants, 1 study, P = 0.58; Analysis 7.6). We considered this evidence to be of very low certainty, downgraded for limitations in study design, imprecision (low number of events or low sample size, and wide confidence intervals) and indirectness.

Summary for this comparison

Evidence from four studies suggests that adding syrup of ipecac to SDAC plus cathartics may make little difference for clinical improvement. On the other hand, we are uncertain about its impact on mortality, adverse events, clinical deterioration, hospitalization or ICU admission. Evidence was of low to very low certainty, downgraded for limitations in study design, imprecision and/or indirectness.

8. Syrup of ipecac versus syrup of ipecac (different doses)

We found one study using different types of syrup of ipecac, manufactured according to the American Pharmacopeia (USP) or the Australian Pharmaceutical Formulary (AFP) (Ilett 1977). More interestingly, this study also compared two different doses of the AFP syrup of ipecac: 15 mL or 30 mL, followed by 200 mL water. This study involved 120 participants presenting to the emergency department of a hospital with various intoxications.

Primary outcomes

Incidence of mortality

The identified study either did not collect or did not report on outcomes related to incidence of mortality.

Adverse events

The identified study either did not collect or did not report on outcomes related to adverse events due to the intervention.

Incidence and severity of symptoms of poisoning

The identified study either did not collect or did not report on outcomes related to incidence and severity of poisoning symptoms.

Secondary outcomes

Duration of symptoms

The identified study either did not collect or did not report on outcomes related to duration of poisoning symptoms.

Drug absorption

The identified study either did not collect or did not report on outcomes related to drug absorption.

Incidence of hospitalization

The identified study either did not collect or did not report on outcomes related to incidence of hospitalization.

Incidence of ICU admission

The identified study either did not collect or did not report on outcomes related to incidence of ICU admission.

Summary for this comparison

We identified one study that compared the use of different doses of syrup of ipecac (Ilett 1977), but it did not report any outcome of interest to our review.

B2. Cathartics

9. Single‐dose activated charcoal plus cathartics versus single‐dose activated charcoal alone

Two trials compared SDAC plus cathartics versus SDAC alone (James 1995; Sue 1994). Both trials studied children presenting to the emergency department following various or unspecified toxic ingestions, requiring SDAC. See Appendix 3.

Primary outcomes

Incidence of mortality

The identified studies either did not collect or did not report outcomes related to incidence of mortality.

Adverse events

In Sue 1994 the only measured adverse event due to the intervention was lethargy during follow‐up. Authors reported no cases of lethargy in the SDAC plus magnesium citrate group (50 participants) or in the SDAC group (14 participants). James 1995 recorded the incidence of participants vomiting upon receiving activated charcoal with or without sorbitol, magnesium citrate or magnesium sulphate. The pooled RR was 1.46 (95% CI 0.61 to 3.49, 116 participants, P = 0.39; Analysis 8.1). We considered the evidence to be of very low certainty, due to limitations in study design, imprecision (low number of events) and indirectness.

Incidence and severity of symptoms of poisoning

The identified studies either did not collect or did not report outcomes related to occurrence and severity of symptoms of poisoning.

Secondary outcomes

Duration of symptoms

The identified studies either did not collect or did not report outcomes related to symptom duration.

Drug absorption

The identified studies either did not collect or did not report outcomes related to drug absorption.

Incidence of hospitalization

Sue 1994 studied the incidence of hospitalization. Three participants required hospitalization in the group receiving SDAC plus either 4 mL/kg, 6 mL/kg or 8 mL/kg of magnesium citrate, versus one patient in the activated charcoal group (RR 0.84, 95% CI 0.09 to 7.46, 64 participants, P = 0.88; Analysis 8.2). The evidence was of very low certainty, downgraded for limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence of ICU admission

The identified studies did not report outcomes related to incidence of ICU admission.

Summary for this comparison

We are uncertain if adding SDAC to a cathartic has an effect on adverse events or hospitalization. The evidence was of very low certainty, downgraded for limitations in study design, imprecision and indirectness.

10. Single‐dose activated charcoal plus different doses of cathartics

Sue 1994 assessed SDAC with different doses of cathartics. This trial studied 64 children presenting to the emergency department following an unspecified toxic ingestion requiring SDAC. See Appendix 4.

Primary outcomes

Incidence of mortality

The identified study either did not collect or did not report outcomes related to incidence of mortality.

Adverse events

The only measured adverse event due to the intervention was lethargy during follow‐up. Authors reported no cases of lethargy in any of the treatment groups (Table 3). We considered this evidence to be of very low certainty, downgraded for limitations in study design, imprecision (low number of events) and indirectness.

Incidence and severity of symptoms of poisoning

The identified study either did not collect or did not report outcomes related to incidence and severity of symptoms of poisoning.

Secondary outcomes

Duration of symptoms

The identified study either did not collect or did not report outcomes related to symptom duration.

Drug absorption

The identified study either did not collect or did not report outcomes related to drug absorption.

Incidence of hospitalization

No patient required hospitalization in the group receiving 4 mL/kg of magnesium citrate plus SDAC compared to one participant in the 6 mL/kg magnesium citrate plus SDAC group and two participants in the 8 mL/kg magnesium citrate plus SDAC group. When comparing 6 mL/kg or 8 mL/kg magnesium citrate versus 4 mL/kg, the ORs were 7.39 (95% CI 0.15 to 372.38, 32 participants, P = 0.32; Table 3) and 7.01 (95% CI 0.42 to 117.63, 34 participants, P = 0.18; Table 3), respectively. The RR when comparing 6 mL/kg to 8 mL/kg magnesium citrate was 1.78 (95% CI 0.18 to 17.80, 34 participants, P = 0.62; Analysis 9.1). We considered this to be evidence of very low certainty, due to limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence of ICU admission

The identified study either did not collect or did not report outcomes related to incidence of ICU admission.

Summary for this comparison

We are uncertain about the effects of higher doses of magnesium citrate combined with SDAC, compared to lower doses combined with SDAC, with respect to the incidence of adverse events or hospitalization. These results were of very low certainty, downgraded for limitations in study design, imprecision and indirectness.

11. Single‐dose activated charcoal plus different types of cathartics

One trial compared SDAC plus different types of cathartics, namely sorbitol, magnesium citrate and magnesium sulphate (James 1995). This trial studied 119 children who ingested a variety of toxins (analgesics, anticonvulsants, antihistamines and decongestants, asthma therapies, automotive products, cardiovascular drugs, gastrointestinal preparations, insecticides, mushrooms, psychotropic drugs, rodenticides, topicals, or miscellaneous drugs). See Appendix 5.

Primary outcomes

Incidence of mortality

The identified study either did not collect or did not report outcomes related to incidence of mortality.

Adverse events

Emesis occurred in 13 of 32 children receiving SDAC plus sorbitol; 6 of 33 children receiving SDAC plus magnesium citrate, and 4 of 23 children receiving SDAC plus magnesium sulphate.

When comparing sorbitol with magnesium sulphate, the RR was 2.34 (95% CI 0.87 to 6.25, 55 participants, P = 0.09; Analysis 10.1). Sorbitol versus magnesium citrate resulted in an RR of 2.23 (95% CI 0.97 to 5.16, 55 participants, P = 0.06; Analysis 10.2). For magnesium sulphate versus magnesium citrate, the RR is 0.96 (95% CI 0.30 to 3.01, 55 participants, P = 0.94; Analysis 10.3). We considered this evidence to be of very low certainty, due to limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence and severity of symptoms of poisoning

The identified study either did not collect or did not report outcomes related to incidence and severity of symptoms of poisoning.

Secondary outcomes

Duration of symptoms

The identified study either did not collect or did not report outcomes related to symptom duration.

Drug absorption

The identified study either did not collect or did not report outcomes related to drug absorption.

Incidence of hospitalization

The identified study either did not collect or did not report outcomes related to incidence of hospitalization.

Incidence of ICU admission

The identified study either did not collect or did not report outcomes related to incidence of ICU admission.

Summary for this comparison

Only one study assessed the effect of different types of cathartics in combination with SDAC in the treatment of poisoning. The only relevant outcome measured is incidence of emesis as an adverse event, for which any effect of using different types of cathartics is uncertain. Evidence is of very low certainty, downgraded for limitations in study design, imprecision and indirectness.

C. Combined first aid interventions that limit uptake and promote evacuation of the poison from the gastrointestinal tract

12. Single‐dose activated charcoal plus cathartic plus hospital intervention versus hospital intervention alone

One study compared SDAC plus a cathartic plus hospital interventions versus hospital interventions alone (Passeron 1989). Passeron 1989 included participants presenting with a confirmed overdose of benzodiazepines, barbiturates or imipramine. All participants in this study received gastric lavage prior to the SDAC plus sorbitol or no additional intervention. See Appendix 6.

Primary outcomes

Incidence of mortality

The identified study either did not collect or did not report outcomes related to incidence of mortality.

Adverse events

The included study reported the incidence of vomiting (Passeron 1989). The Peto OR was 9.94 (95% CI 1.52 to 65.02, 32 participants; Analysis 11.1). The level of evidence was very low, downgraded due to limitations in study design, imprecision (low number of events and wide confidence intervals) and indirectness.

Incidence and severity of symptoms of poisoning

Passeron 1989 monitored participants' symptoms using the Glasgow Coma Scale. They did not report any numeric data but reported no difference for the course of the Glasgow Coma Scale scores between treatments (P = 0.49). The evidence was of very low certainty, downgraded for limitations in study design, imprecision due to a lack of data and indirectness.

Secondary outcomes

Duration of symptoms

The identified study either did not collect or did not report outcomes related to symptom duration.

Drug absorption

Passeron 1989 measured drug levels in their participants, but did not report any of our pre‐defined outcomes of interest (AUC, C_max, T_max), so we could not make a reliable estimate of effect on drug absorption.

Incidence of hospitalization

The identified study either did not collect or did not report outcomes related to incidence of hospitalization.

Incidence of ICU admission

The identified study either did not collect or did not report outcomes related to incidence of ICU admission.

Summary for this comparison

We found only evidence of very low certainty (downgraded for limitations in study design, imprecision and indirectness) concerning the use of SDAC plus a cathartic, in adjuvant to established hospital treatments. Thus we are uncertain about the effect on Glasgow Coma Scale scores or incidence of vomiting.

13. Multi‐dose activated charcoal plus cathartic plus hospital intervention versus hospital intervention alone

One study in 14 participants compared MDAC plus magnesium sulphate as an adjuvant to oral N‐acetylcysteine versus N‐acetylcysteine alone for paediatric paracetamol overdose (Montoya‐Cabrera 1999). See Appendix 7.

Primary outcomes

Incidence of mortality

The identified study either did not collect or did not report outcomes related to incidence of mortality.

Adverse events

Montoya‐Cabrera 1999 reported that no adverse events occurred in any treatment group (Table 3). This is evidence of very low certainty, downgraded for limitations in study design, imprecision (low number of events) and indirectness.

Incidence and severity of symptoms of poisoning

The identified study either did not collect or did not report outcomes related to occurrence and severity of poisoning symptoms.

Secondary outcomes

Duration of symptoms

The identified study either did not collect or did not report outcomes related to symptom duration.

Drug absorption

Montoya‐Cabrera 1999 measured T_1/2 of paracetamol in the plasma, reporting a decrease (P < 0.05; Table 3). However, they failed to report any measure of spread, so we cannot report any summary estimate with 95% CI. We consider the evidence to be of very low certainty due to limitations in study design, imprecision (low sample size and lack of data) and indirectness.

Incidence of hospitalization

The identified study either did not collect or did not report outcomes related to incidence of hospitalization.

Incidence of ICU admission

The identified study either did not collect or did not report outcomes related to incidence of ICU admission.

Summary for this comparison

We identified one study, which provided evidence of very low certainty (downgraded for limitations in study design, imprecision and indirectness) (Montoya‐Cabrera 1999). Any effect on the incidence of adverse events or plasma half‐life of paracetamol after receiving MDAC in adjuvant to oral administration of N‐acetylcysteine, compared to oral administration of N‐acetylcysteine alone is uncertain.

D. First aid interventions that neutralize or dilute the poison

We identified no studies comparing interventions aiming at neutralizing or diluting orally ingested poisons in a patient setting.

Discussion

Summary of main results

Out of a total of 11,582 potentially relevant references, we identified 24 studies reported in 31 publications. All but one study took place in a hospital setting (Wax 1999 was in a pre‐hospital setting). A total of 7099 participants were randomized to different treatment groups.

A. First aid interventions that limit or delay the absorption of the poison in the body

Activated charcoal

A commonly used intervention in poisoning is activated charcoal. Due to its enormous surface area it can adsorb large quantities of drugs, thus preventing the absorption of the poison by the body. Included studies made different comparisons, either using single‐dose activated charcoal (SDAC) alone, combined with other pre‐hospital treatments, or in adjuvant to hospital interventions. Furthermore, multi‐dose activated charcoal (MDAC) was used in adjuvant to hospital interventions.

We found very low‐certainty evidence from two studies, involving 476 participants and comparing single‐dose activated charcoal to no intervention, which is our main comparison (Merigian 1990; Underhill 1990). These studies included participants with unspecified exposures in Merigian 1990 or paracetamol overdoses in Underhill 1990. Any effect on the incidence of adverse events, clinical deterioration or ICU admission is uncertain. See summary of findings Table for the main comparison.

Seven trials, providing evidence of low to very low certainty and including 5383 participants, investigated the effect of SDAC in adjuvant to established hospital interventions such as supportive treatment and in most cases also gastric lavage (Comstock 1982; Cooper 2005; Crome 1983; Eddleston 2008; Hultén 1988; Merigian 2002; Roberts 2006). All but three studies either did not specify a toxic syndrome or recruited participants with different toxic syndromes (Crome 1983; Hultén 1988;Roberts 2006). There may be little or no difference in the incidence of mortality or the need for and length of intubation. We are uncertain about the effects of SDAC in addition to hospital treatments with regard to adverse events, drug absorption and incidence of hospitalization or ICU admission. See summary of findings Table 2.

Five trials including 3568 participants compared MDAC versus SDAC, all in adjuvant to hospital interventions (Behnoush 2009; Brahmi 2006; De Silva 2003; Eddleston 2008; Roberts 2006). Two trials included participants with carbamazepine overdose (Behnoush 2009; Brahmi 2006), and two trials studied participants with yellow oleander poisoning (De Silva 2003; Roberts 2006). Eddleston 2008 included participants with different types of overdose, including yellow oleander and pesticide poisoning. There were some discrepancies between studies, preventing us from drawing any conclusions regarding incidence of mortality. This is evidence of very low certainty. In addition, low‐certainty evidence suggests that MDAC may result in decreased incidence of ICU admissions and an increase in abdominal discomfort or diarrhoea, but it may have no influence on the need for intubation. Furthermore, we are uncertain about the effects of MDAC on drug absorption or length of intubation, evidence of very low certainty. See summary of findings Table 3.

We identified one study, involving 34 participants, with very low‐certainty evidence that compared SDAC with syrup of ipecac in participants with anti‐inflammatory, analgesic or psychotropic drug overdose (Amigó Tadín 2002). We are uncertain about the effect of SDAC versus ipecac on the incidence of adverse events or Glasgow Coma Scale scores. See summary of findings Table 4.

We identified three studies comparing MDAC in adjuvant to hospital intervention versus hospital intervention alone in 3121 participants (Bouget 1989; Eddleston 2008; Roberts 2006). This low‐ to very low‐certainty evidence suggests there may be no difference in the incidence of mortality, symptoms of intoxication or length of intubation, while we are uncertain about its effects on pharmacokinetic parameters. See summary of findings Table 5.

None of the above‐mentioned evidence could show added value for the use of activated charcoal, either administered as a single dose or as multiple doses, or in adjuvant or not to hospital interventions.

Body position

Another possible intervention to slow down the uptake of the poison is a certain body position. The theory is that placing a patient on the left lateral decubitus position would allow the gastric content to stay in the greater curvature of the stomach, due to the combination of gravity and the anatomy of the stomach, which might slow down the rate of absorption of the poison. However, we did not identify any studies performed in poisoning participants that compared different kinds of body position.

B. First aid interventions that evacuate the poison from the gastrointestinal tract

Vomiting or accelerated defecation might induce the quick evacuation of the poison from the gastrointestinal tract. Interventions that might obtain this effect are emetics, which induce vomiting, or cathartics, for the acceleration of defecation.

Emetics

Five included studies looked at the effectiveness of syrup of ipecac as a first aid measure for poisoning.

One study including 103 participants provided evidence of low certainty on the use of ipecac versus no intervention in asymptomatic participants with toxic berry ingestion (Wax 1999). This study took place in a pre‐hospital setting and reported no clinical outcomes. While there may be little or no difference in emergency department referral, there may be an increase in adverse events. See summary of findings Table 6.

Four studies, involving 1240 participants, assessed the addition of syrup of ipecac to SDAC plus a cathartic (Albertson 1989; Kornberg 1991; Kulig 1985; Pond 1995). All studies either did not specify or included multiple types of overdose. Low‐certainty evidence suggests there may be little or no difference in the incidence of clinical improvement. On the other hand, we are uncertain about any effect on the incidence of mortality, adverse events, clinical deterioration, hospitalization or ICU admission. See Appendix 2.

We identified one study with 120 participants comparing different doses of ipecac (Ilett 1977), but it did not report any outcomes of interest.

One study compared syrup of ipecac versus SDAC. We describe this study above (interventions that limit or delay the absorption of the poison in the body). See summary of findings Table 4.

None of the evidence on the use of syrup of ipecac as a first aid intervention shows any benefit, and it may even cause harm.

Cathartics

Cathartics are often used in combination with activated charcoal, where activated charcoal is used to adsorb the poison and the cathartic to accelerate the evacuation from the gastrointestinal tract. We identified two studies in 183 participants that looked at the combination of different types or doses of cathartic with SDAC in children with unspecified or various intoxications (James 1995; Sue 1994).

These trials provided evidence of very low certainty comparing the use of SDAC plus a cathartic versus SDAC alone in either unspecified or multiple toxic syndromes. We are uncertain whether adding a cathartic to the treatment influences the incidence of adverse events or the incidence of hospitalization. See Appendix 3.

Sue 1994 assessed SDAC plus different doses of magnesium citrate in 64 participants, but we are uncertain whether this would result in a difference regarding the incidence of hospitalization. See Appendix 4.

James 1995 studied the effects of different types of cathartics (sorbitol, magnesium citrate or magnesium sulphate) in combination with SDAC, in 119 participants. We are uncertain whether a different type of cathartic in adjuvant to SDAC influences the incidence of vomiting as an adverse event. See Appendix 5.

In summary, we did not identify any trials that looked at the use of cathartics alone compared with no intervention. Cathartics were always used in combination to SDAC. From the limited evidence available, we are not able to draw conclusions regarding the use of cathartics in addition to SDAC.

C. Combinations of first aid interventions

One study including 32 participants looked at the effects of combining SDAC with cathartics as an adjuvant to hospital intervention, compared to hospital intervention alone in participants with overdoses of benzodiazepines, barbiturates or imipramine (Passeron 1989). The study provided evidence of very low certainty, so we are uncertain about the impact of adding SDAC plus cathartics in adjuvant to a hospital intervention on Glasgow Coma Scale scores or the incidence of vomiting. See Appendix 6.

One study in 14 participants used MDAC in combination with magnesium sulphate as a cathartic, in adjuvant to hospital treatment (Montoya‐Cabrera 1999). The evidence of very low certainty precluded us from drawing conclusions about the effects of MDAC and magnesium sulphate in adjuvant to hospital treatments on the plasma half‐life of paracetamol. See Appendix 7.

D. First aid interventions that neutralize or dilute the poison

We did not identify any studies in poisoning patients that looked at the effects of commonly used home remedies such as drinking milk, water, vinegar or citrus juice to neutralize or dilute the poison.

Overall completeness and applicability of evidence

The objective of this review was to assess the effects of pre‐hospital interventions, alone or in combination, that laypeople could feasibly provide victims of acute oral poisoning before professional help arrives. We identified only one study from a pre‐hospital setting (Wax 1999). As we anticipated that this would be the case in advance, we also included studies performed in a hospital setting as indirect evidence. Furthermore, half of the studies compared the intervention of interest in adjuvant to a hospital intervention versus hospital interventions alone. These considerations limit the applicability of our findings.

With regard to the interventions of interest, about two‐thirds of the identified studies looked at SDAC and MDAC alone or in combination with cathartics, and sometimes in adjuvant to hospital treatment. We identified six studies assessing syrup of ipecac. We found little evidence on the use of cathartics, and in the two studies we did identify cathartics were used in adjuvant to another treatment (i.e. SDAC), making it difficult to make a judgment on the use of cathartics by themselves for oral poisoning. Finally, we found no studies of the effect of body position or interventions that might dilute or neutralize the poison, such as drinking water or milk.

The 24 identified studies described a wide range of outcomes. However, the primary outcomes of interest, mortality, severity of symptoms due to poisoning and adverse events, were very variably and often incompletely reported. Useable data on these outcomes were thus limited. This also precluded combining the different interventions in this review in a network meta‐analysis (NMA), which would have allowed us to compare the relative efficacy of different interventions and potentially rank the interventions for efficacy. We will perform an NMA for future updates of this review, if more useable data becomes available.

Furthermore, most studies were over 10 years old, with the oldest study being performed in 1977 (Ilett 1977). Only two studies took place in the past decade (Behnoush 2009; Eddleston 2008).

A major limitation in most of the identified studies is the substantial heterogeneity of the included participants. This might obscure potential benefits in subgroups of participants, for example participants with severe poisoning, specific toxic syndromes or those presenting early (Juurlink 2015). On the other hand, in a first aid setting it might often be unclear what type of patient a caregiver is dealing with, with regard to the type, dose and timing of intoxication, further complicating conclusions with respect to the lay setting.

Overall, the identified evidence is scarce and of low to very low certainty, which precludes any firm conclusions about the added value of any of the first aid interventions discussed in this review. However, almost all of these studies were performed in a hospital setting, which means there is a delay in presentation and thus treatments are started at a later time than when given in a home setting. It could therefore be possible that treatments were not effective because of their delayed administration.

In addition, evidence is too scarce to be able to draw conclusions about the safety of most of the first aid interventions. The exception is syrup of ipecac, for which low‐certainly evidence suggests that the number of complications increases when using it compared to no intervention.

Quality of the evidence

A key methodological limitation in the included studies is that most studies used inappropriate or unclear methods of randomization. Furthermore, most studies reported outcomes poorly, and the reporting format was highly heterogeneous. This makes the studies difficult to compare with one another. The variation between and within studies with respect to the population further complicates the comparison of different studies.

First aid interventions that limit or delay the absorption of the poison in the body

For most of the comparisons including single‐ or multi‐dose activated charcoal, the evidence is of low to very low certainty. In most cases, we downgraded the evidence for indirectness (since most studies were performed in a hospital setting), imprecision (limited sample size, low number of events and/or wide confidence intervals) and limitations in study design.

In the comparison of SDAC versus MDAC (both in adjuvant to hospital interventions), we identified two studies at low risk of bias; however, there was inconsistency between the studies' findings, which makes it difficult to draw any conclusions. There is no clear cause of this inconsistency, but possible explanations might be that De Silva 2003 included participants up to 72 h after poison ingestion, whereas Eddleston 2008 included participants only up to 24 h after ingestion. Furthermore, Eddleston 2008 included less severely poisoned participants than De Silva 2003 (Glasgow Coma Scale of less than 13).

First aid interventions that evacuate the poison from the gastrointestinal tract

For interventions that promote the evacuation of the poison from the gastrointestinal tract, we found the most evidence on the use of syrup of ipecac. However, all studies were at high risk of selection bias, and most were at high or unclear risk of detection bias. We further downgraded studies due to imprecision and indirectness, leading to low or very low certainty evidence.

The evidence on the use of cathartics was limited and of very low certainty due to a high risk of reporting bias, indirectness and imprecision. The identified studies always used cathartics in combination with other interventions, making it difficult to draw conclusions about cathartics alone as a treatment for oral poisoning.

First aid interventions that neutralize or dilute the poison

There are no available data that we identified on interventions that neutralize or dilute the poison.

Potential biases in the review process

This review intended to assess interventions that are feasible for laypeople to use in situations of oral poisoning. This means that the interventions should be feasible to use in a pre‐hospital setting by people without any medical knowledge. According to these criteria, we excluded interventions such as gastric lavage or intravenous drug administrations.

Most identified studies took place in a hospital setting, which means that we had to downgrade the level of evidence due to indirectness. Only one included study was in a pre‐hospital community setting, but it included only asymptomatic poisoning patients.

Although the interventions were mostly in a hospital setting, we included only studies that used interventions feasible by laypeople. We only allowed comparisons with hospital interventions if the treatment group received the same hospital interventions in adjuvant to the possible first aid treatment under investigation.

We did not include the many available volunteer studies. These are studies in which healthy volunteers receive a drug in a therapeutic or supratherapeutic dose in an attempt to simulate oral poisoning, in a controlled setting such as a laboratory environment, mostly without co‐ingestion of other drugs or alcohol, on an empty stomach. We believed this was even more indirect than including studies performed in actual oral poisoning patients, although they were performed in a hospital setting.

As mentioned earlier, most studies were over 10 years old, with the oldest study from 1977. Only two studies took place within 10 years of our literature search. Many of the studies were poorly reported: data were missing, and our attempts to contact the authors were often unsuccessful because no contact details were available, authors did not respond, or data were no longer available. This could introduce a bias on the completeness of the data and the risk of bias assessment, leading to perhaps a more strict judgment of bias for some studies.

Agreements and disagreements with other studies or reviews

An existing Cochrane Review on interventions for paracetamol (acetaminophen) overdose included not only possible first aid interventions such as activated charcoal or syrup of ipecac, but also hospital interventions such as gastric lavage, charcoal haemoperfusion, antidotes such as N‐acetylcysteine or cimetidine, and liver transplantations (Chiew 2018). The review included randomized controlled trials as well as observational studies, and studies performed in healthy volunteers as well as in patients. In our review, the focus is on first aid interventions feasible for laypeople. This excludes all types of hospital interventions. Furthermore, we did not focus on a specific toxin, and since sufficient studies in patients were available, we decided to exclude studies performed in healthy volunteers. We agree with the conclusions of Chiew 2018 that the use of activated charcoal seems a safer option than syrup of ipecac to reduce uptake, although research still needs to demonstrate a clear clinical benefit.

There might be some overlap with the different position papers published on the use of SDAC (American Academy of Clinical Toxicology 2005), MDAC (American Academy of Clinical Toxicology 1999), cathartics (American Academy of Clinical Toxicology 2004), and syrup of ipecac (Höjer 2013). These papers give a comprehensive overview of the interventions and discuss the published literature, from preclinical to clinical research. However, most are out of date, and it is not clear if the literature search was systematic. Our systematic review does highlight that in the decade preceding publication, there has been very little research on this important topic.

Furthermore, the position papers only give a description of the different identified studies, whereas in our review we combined studies in a meta‐analysis where possible, to give an overall effect size. In any case, our conclusions are largely similar: there is insufficient evidence for a clinical benefit and thus for the routine use of any of the investigated treatments. Few studies have been published since the publication of these position papers, and while our review includes them, all fail to show a clear clinical benefit for the use of these first aid treatments in a hospital setting. However, one important difference is that the recommendations made in these position papers were designed for a professional care setting. The relevance for a pre‐hospital setting remains unclear.

A systematic literature search and meta‐analysis on the effect of activated charcoal in healthy volunteers showed that administration of activated charcoal was most effective when administered immediately after drug intake, but it was still effective up until four hours after drug ingestion (Jürgens 2009). There was no information on adverse events due to the intervention. This meta‐analysis demonstrates the theoretical capacity of activated charcoal to reduce uptake of a variety of toxins. However, the actual clinical benefit for oral poisoning patients remains speculative, as demonstrated by the studies included in our review, which fail to show a clear clinical benefit. Reasons for the discrepancy between data collected from healthy volunteers and actual patients might include the time passed between ingestion of the drugs and the start of the treatment. Also, the meta‐analysis looked at activated charcoal as the only treatment. It did not assess any combination treatments, such as activated charcoal plus a cathartic. Furthermore, the controlled setting where studies in healthy volunteers take place excludes certain confounding variables, for example ingesting the drugs with alcohol, intake of a cocktail of different kinds of drugs or not knowing which drugs were taken. These considerations imply that the use of activated charcoal is still a therapeutic option in emergency departments, but clinicians should carefully consider its use for individual patients (Juurlink 2015). From the available evidence, it is unclear whether a layperson would be capable of making these considerations in a pre‐hospital setting.

Figure 1

Study selection flow diagram.

Figure 2

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 3

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.1

Comparison 1 SDAC vs no intervention, Outcome 1 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.1

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 1 Incidence of mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.2

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 2 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.3

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 3 Incidence of need for intubation.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.4

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 4 Incidence of convulsions.

Ir a la figura de la revisiónAbrir en una pestaña nueva


Study	SDAC	No SDAC	Summary estimate (P value)	# participants
Eddleston 2008	median (IQR): 112.0 (36.6–234.9)	median (IQR): 88.5 (38.5–203.1)	median difference: 23.5 (P > 0.05)	No information
Merigian 2002	mean: 54.6	mean: 39.9	mean difference: 14.7 (P = 0.70)	No information

Analysis 2.5

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 5 Duration of intubation (h).


Study	SDAC (median (IQR))	no SDAC (median (IQR))	Summary estimate (P value)	# participants
Roberts 2006	17.7 (11.1;21.8)	19.0 (13.7;24.3)	‐1.3 (P > 0.05)	28 vs 40

Analysis 2.6

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 6 AUC ((µg/L) × h).

max
Study	SDAC (median (IQR))	no SDAC (median (IQR))	Summary estimate (P value)	# participants
Roberts 2006	0.98 (0.72;1.50)	1.05 (0.75;1.40)	‐0.07 (P > 0.05)	28 vs 40

Analysis 2.7

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 7 C_max (µg/L).

max
Study	SDAC (median (IQR))	no SDAC (median (IQR))	summary estimate (P value)	# participants
Roberts 2006	7.2 (5.7;13.8)	12.1 (5.4;17.4)	‐4.9 (P > 0.05)	28 vs 40

Analysis 2.8

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 8 T_max (h).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.9

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 9 Incidence of hospitalization.

Analysis 2.10

Comparison 2 SDAC + hospital intervention vs hospital intervention, Outcome 10 Incidence of ICU admission.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.1

Comparison 3 SDAC vs syrup of ipecac, Outcome 1 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.2

Comparison 3 SDAC vs syrup of ipecac, Outcome 2 Glasgow Coma Scale score.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.3

Comparison 3 SDAC vs syrup of ipecac, Outcome 3 Mean arterial blood pressure (mmHg).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.4

Comparison 3 SDAC vs syrup of ipecac, Outcome 4 Heart rate (bpm).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 3.5

Comparison 3 SDAC vs syrup of ipecac, Outcome 5 Respiratory rate (breaths/min).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.1

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 1 Incidence of mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.2

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 2 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.3

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 3 Incidence of need for cardiac pacing/antitoxin treatment.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.4

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 4 Incidence of life‐threatening arrhythmias.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.5

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 5 Amount of atropine administered (mg).

Ir a la figura de la revisiónAbrir en una pestaña nueva


Study	MDAC (median (range))	SDAC (median (range))	Median difference [95% CI]	P value	# participants
De Silva 2003	1 (1‐6)	2 (1‐12)	0.0 (0.0‐1.0)	P < 0.0001	201 vs 200

Analysis 4.6

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 6 Number of atropine boluses administered.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.7

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 7 Incidence of need for intubation.

Analysis 4.8

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 8 Incidence of convulsions.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.9

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 9 Duration of coma (h).

Ir a la figura de la revisiónAbrir en una pestaña nueva


Study	MDAC	SDAC	Summary estimate	P value	# participants
Brahmi 2006	mean±SD: 24.1±4.2	mean±SD: 36.4±3.6	MD: ‐12.30, 95%CI [‐18.56;‐6.04]	P = 0.0001	3 vs 3
Eddleston 2008	median [IQR]: 83.8 (35.0–173.0)	median [IQR]: 112.0 (36.6–234.9)	median difference: ‐28.2	No information	No information

Analysis 4.10

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 10 Duration of intubation (h).

max
Study	MDAC	SDAC	Summary estimate	P value	# participants
Brahmi 2006	mean±SD: 33±3.46	mean±SD: 32.6±5.63	MD: 0.40, 95%CI [‐4.89;5.69]	P = 0.88	6 vs 6
Roberts 2006	median (IQR): 1.13 (0.86;1.47)	median (IQR): 0.98 (0.72;1.50)	median difference: 0.15	P > 0.05	36 vs 28

Analysis 4.11

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 11 C_max (µg/L).

max
Study	MDAC	SDAC	Summary estimate	P value	# participants
Roberts 2006	median (IQR): 8.3 (4.8;15.0)	median (IQR): 7.2 (5.7;13.8)	median difference: 1.1	P > 0.05	36 vs 28

Analysis 4.12

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 12 T_max (h).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.13

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 13 T_1/2 (h).


Study	MDAC	SDAC	Summary estimate	P value	# participants
Roberts 2006	median (IQR): 17.3 (12.8;21.7)	median (IQR): 17.7 (11.1;21.8)	median difference: ‐0.4	P > 0.05	36 vs 28

Analysis 4.14

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 14 AUC ((µg/L) × h).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 4.15

Comparison 4 MDAC + hospital intervention vs SDAC + hospital intervention, Outcome 15 Incidence of ICU admission.

Analysis 5.1

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 1 Incidence of mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.2

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 2 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.3

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 3 Incidence of need for intubation.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.4

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 4 Incidence of seizures.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 5.5

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 5 Incidence of need for cardiac pacing/antitoxin treatment.

Ir a la figura de la revisiónAbrir en una pestaña nueva


Study	MDAC (median IQR)	no intervention (median IQR)	Summary estimate	P value	# participants
Eddleston 2008	83.8 (35.0–173.0)	88.5 (38.5–203.1)	median difference: ‐4.7	P > 0.05	No information

Analysis 5.6

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 6 Length of intubation (h).


Study	MDAC (median IQR)	no intervention (median IQR)	Summary estimate	P value	# participants
Roberts 2006	17.3 (12.8;21.7)	19.0 (13.7;24.3)	median difference: ‐1.7	P > 0.05	36 vs 40

Analysis 5.7

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 7 AUC.

max
Study	MDAC (median IQR)	no intervention (median IQR)	Summary estimate	P value	# participants
Roberts 2006	1.13 (0.86;1.47)	1.05 (0.75;1.40)	median difference: 0.08	P > 0.05	36 vs 40

Analysis 5.8

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 8 C_max.

max
Study	MDAC (median IQR)	no intervention (median IQR)	Summary estimate	P value	# participants
Roberts 2006	8.3 (4.8;15.0)	12.1 (5.4;17.4)	median difference: ‐3.8	P > 0.05	36 vs 40

Analysis 5.9

Comparison 5 MDAC + hospital intervention vs hospital intervention, Outcome 9 T_max.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 6.1

Comparison 6 Syrup of ipecac vs no intervention, Outcome 1 Incidence of diarrhoea.

Analysis 6.2

Comparison 6 Syrup of ipecac vs no intervention, Outcome 2 Incidence of abdominal pain.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 6.3

Comparison 6 Syrup of ipecac vs no intervention, Outcome 3 Incidence of sedation.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 6.4

Comparison 6 Syrup of ipecac vs no intervention, Outcome 4 Incidence of agitation.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.1

Comparison 7 Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic, Outcome 1 Incidence of mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.2

Comparison 7 Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic, Outcome 2 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.3

Comparison 7 Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic, Outcome 3 Incidence of clinical improvement.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.4

Comparison 7 Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic, Outcome 4 Incidence of clinical deterioration.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.5

Comparison 7 Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic, Outcome 5 Incidence of hospitalization.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 7.6

Comparison 7 Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic, Outcome 6 Incidence of ICU admission.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 8.1

Comparison 8 SDAC + cathartic vs SDAC, Outcome 1 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 8.2

Comparison 8 SDAC + cathartic vs SDAC, Outcome 2 Incidence of hospitalization.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 9.1

Comparison 9 SDAC + cathartic vs SDAC + cathartic (higher dose), Outcome 1 Incidence of hospitalization (8 mL vs 6 mL).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 10.1

Comparison 10 SDAC + cathartic vs SDAC + cathartic (different type), Outcome 1 Incidence of vomiting (sorbitol vs magnesium sulphate.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 10.2

Comparison 10 SDAC + cathartic vs SDAC + cathartic (different type), Outcome 2 Incidence of vomiting (sorbitol vs magnesium citrate).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 10.3

Comparison 10 SDAC + cathartic vs SDAC + cathartic (different type), Outcome 3 Incidence of vomiting (magnesium sulphate vs magnesium citrate).

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 11.1

Comparison 11 SDAC + cathartic + hospital intervention vs hospital intervention, Outcome 1 Incidence of adverse events.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Summary of findings for the main comparison. SDAC versus no intervention for first aid in patients with acute oral poisoning

SDAC versus no intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (paracetamol or not specified) Setting: hospital setting Intervention: single‐dose activated charcoal (SDAC) Comparison: no intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with no intervention	Risk with SDAC
Incidence of mortality	No studies collected or reported this outcome
Incidence of adverse events	Control group: 0/236; intervention group: 4/240 (Peto OR 4.17, 95% CI 0.30 to 57.26)		—	476 (2 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain of the effect of SDAC on the incidence of adverse events.
Incidence and severity of symptoms of poisoning: incidence of clinical deterioration during stay in the hospital	—		—	451 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,d	The relative effect was not estimable due to the absence of events in the intervention (0/220) and the control group (0/231). We are uncertain of the effect of SDAC on incidence and severity on poisoning.
Duration of toxic symptoms	No studies collected or reported this outcome
Drug absorption	No studies collected or reported this outcome
Incidence of hospitalization	No studies collected or reported this outcome
Incidence of ICU admission	Control group: 0/231; intervention group: 1/220 (Peto OR 7.77, 95% CI 0.15 to 391.93)		—	451 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain of the effect of SDAC on the incidence of ICU admission.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; MD: mean difference; RCT: randomized controlled trial; RR: risk ratio; SDAC: single‐dose activated charcoal; OR: odds ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious limitations in study design: high risk of selection bias. ^bDowngraded one level for serious indirectness: study conducted in a hospital setting. ^cDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^dDowngraded one level for serious imprecision: low number of events.

Summary of findings for the main comparison. SDAC versus no intervention for first aid in patients with acute oral poisoning

Summary of findings 2. SDAC + hospital intervention versus hospital intervention alone for first aid in patients with acute oral poisoning

SDAC + hospital intervention versus hospital intervention alone for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (not specified, tricyclic antidepressants, combinations of different drugs or yellow oleander) Setting: hospital setting Intervention: single‐dose activated charcoal (SDAC) + hospital intervention Comparison: hospital intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with hospital intervention	Risk with SDAC + hospital intervention
Incidence of mortality	Study population		Peto OR 1.04 (0.79 to 1.37)	3425 (2 RCTs)	⊕⊕⊝⊝ Low^a,b	SDAC in addition to hospital treatments may make little or no difference on incidence of mortality.
	62 per 1000	64 per 1000 (49 to 85)
Incidence of adverse events	Incidence of vomiting: intervention group: 118/570 and control group: 163/1236 (RR 1.44, 95% CI 0.88 to 2.37; 1806 participants; 2 studies). Incidence of absent bowel sounds: intervention group: 7/1544 and control group: 17/1554 (RR 0.41, 95% CI 0.17 to 1.00, 1 study, 3098 participants).		—	4904 (3 RCTs)	⊕⊝⊝⊝ Very low^a,c,d	Statistically significant heterogeneity was found, which may be explained partially by subgroup analyses per type of adverse event. We are uncertain about the effect of SDAC in addition to hospital treatments on incidence of adverse events.
Incidence and severity of symptoms of poisoning: incidence of need for intubation	Patients that received gastric lavage prior to SDAC: intervention group: 80/1578 and control group: 87/1597 (RR 0.95, 95% CI 0.70 to 1.27, 2 studies, 3175 participants). Patients that did not receive gastric lavage prior to SDAC: intervention group: 24/194 and control group: 10/193 (RR 2.61, 95% CI 1.38 to 4.93, 1 study, 387 participants).		—	3562 (4 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	Statistically significant heterogeneity was found, which may be explained by subgroup analyses in patients receiving or not receiving gastric lavage. We are uncertain about the effect of SDAC on incidence of need for intubation.
Duration of toxic symptoms: duration of intubation (h)	Eddleston 2008: intervention group median (IQR): 112.0 (36.6 to 234.9) h and control group median (IQR): 88.5 (38.5 to 203.1) h (median difference: 23.5 h, P > 0.05). Merigian 2002: intervention group mean: 54.6 h and control group mean: 39.9 h (MD: 14.7 h, P = 0.70).		—	(2 RCTs)	⊕⊕⊝⊝ Low^a,e	Data were reported as median with IQR in one or means without measure of spread in another study, without information on participant numbers. SDAC in addition to hospital treatments may make little or no difference on the duration of intubation.
Drug absorption: cardenolide: AUC (µg/L) × h Follow‐up: 1 days	The median (IQR) in intervention group was 17.7 (11.1 to 21.8) (µg/L) × h and in the control group 19.0 (13.7 to 24.3) (µg/L) × h (median difference: −1.3 h, P > 0.05)		—	68 (1 RCT)	⊕⊝⊝⊝ Very low^a,f,g	We are uncertain about the effect of SDAC in addition to hospital treatments on cardenolide absorption.
Incidence of hospitalization	125 per 1000	196 per 1000 (152 to 252)	RR 1.57 (1.22 to 2.02)	1479 (1 RCT)	⊕⊝⊝⊝ Very low^a,g,h	We are uncertain about the effect of SDAC in addition to hospital treatments on incidence of hospitalization.
Incidence of ICU admission	30 per 1000	69 per 1000 (42 to 114)	RR 2.33 (1.42 to 3.82)	1479 (1 RCT)	⊕⊝⊝⊝ Very low^a,g,h	We are uncertain about the effect of SDAC in addition to hospital treatments on incidence of ICU admission.
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; RCT: randomized controlled trial; RR: risk ratio; SDAC: single‐dose activated charcoal; OR: odds ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious indirectness: study conducted in a hospital setting. ^bDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^cDowngraded one level for serious inconsistency: large and statistically significant heterogeneity present (I² > 60%, P < 0.10). ^dDowngraded one level for serious imprecision: wide confidence intervals. ^eDowngraded one level for serious imprecision: lack of data on the number of patients analysed. ^fDowngraded one level due to serious limitations in study design: high risk of other bias: it is not entirely clear what is measured with the assay used. The fact that both active cardenolides and (inactive) metabolites might be detected by the assay compromise the results of these analyses, as they might explain the wide variability observed. ^gDowngraded one level for serious imprecision: low number of events. ^hDowngraded one level for serious limitations in study design: high risk of selection bias.

Summary of findings 2. SDAC + hospital intervention versus hospital intervention alone for first aid in patients with acute oral poisoning

Summary of findings 3. MDAC + hospital intervention versus SDAC + hospital intervention for first aid in patients with acute oral poisoning

MDAC + hospital intervention versus SDAC + hospital intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (carbamazepine, yellow oleander, or combinations of different drugs) Setting: hospital setting Intervention: multiple dose of activated charcoal (MDAC) + hospital intervention Comparison: single‐dose activated charcoal (SDAC) + hospital intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with SDAC + hospital intervention	Risk with MDAC + hospital intervention
Incidence of mortality	Study population		RR 0.59 (0.21 to 1.63)	3476 (2 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	Combining the studies resulted in statistically significant heterogeneity, for which explanations remain speculative. We are uncertain about the effects of MDAC in addition to hospital treatment, compared to SDAC, in addition to hospital treatment.
	72 per 1000	42 per 1000 (15 to 117)
Incidence of adverse events	Study population		Peto OR 3.55 (1.85 to 6.79)	3476 (2 RCTs)	⊕⊕⊝⊝ Low^b,c	There was statistically significant heterogeneity, which may be attributable to different adverse events measured in individual studies. MDAC in addition to hospital treatment may increase abdominal discomfort/diarrhoea and absent bowel sounds, compared to SDAC in addition to hospital treatment.
	4 per 1000	14 per 1000 (7 to 27)
Incidence and severity of symptoms of poisoning: incidence of need for intubation	Study population		RR 1.01 (0.75 to 1.38)	3097 (2 RCTs)	⊕⊕⊝⊝ Low^b,c	MDAC in addition to hospital treatment may make little or no difference in the incidence of need for intubation, compared to SDAC in addition to hospital treatment.
	49 per 1000	49 per 1000 (37 to 67)
Duration of toxic symptoms: duration of intubation (h)	Brahmi 2006: intervention group: 24.1 (SD 4.2 h and control group 36.4 (SD 3.6 h (MD: 12.30 h lower, 95% CI −18.56 to −6.04, 6 participants). Eddleston 2008: intervention group median (IQR): 83.8 (35.0 to 173.0) h and control group median (IQR): 112.0 (36.6 to 234.9) h (median difference: 28.2 h), unclear number of participants		—	(2 RCTs)	⊕⊝⊝⊝ Very low^b,d,e	Data were reported as means with SD in one study or medians with IQR in another study, without information on participant numbers or statement of significance. We are uncertain about the effects of MDAC in addition to hospital treatment on duration of intubation, compared to SDAC in addition to hospital treatment.
Drug absorption: cardenolide: AUC (µg × L/h) Follow‐up: 1 days	The median (IQR) in intervention group was 17.3 (12.8 to 21.7) (µg/L) × h and in the control group 17.7 (11.1 to 21.8) (µg/L) × h (median difference −0.4, P > 0.05).		—	64 (1 RCT)	⊕⊝⊝⊝ Very low^b,e,f	We are uncertain about the effects of MDAC in addition to hospital treatment on cardenolide drug absorption, compared to SDAC in addition to hospital treatment.
Incidence of hospitalization	No studies collected or reported this outcome
Incidence of ICU admission	Study population		RR 0.31 (0.12 to 0.83)	401 (1 RCT)	⊕⊕⊝⊝ Low^b,g	MDAC in addition to hospital treatment may result in a decreased incidence of ICU admission, compared to SDAC in addition to hospital treatment.
	80 per 1000	25 per 1000 (10 to 66)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). AUC: area under the receiver operating curve; CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; MDAC: multi‐dose activated charcoal; RCT: randomized controlled trial; RR: risk ratio; SD: standard deviation; SDAC: single‐dose activated charcoal; OR: odds ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious inconsistency: combining results resulted in a considerable and statistically significant degree of heterogeneity (I² > 60%, P < 0.10). ^bDowngraded one level for serious indirectness: study conducted in a hospital setting. ^cDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^dDowngraded one level for other limitations: inconsistent conclusions made by the studies. ^eDowngraded one level for serious imprecision: low sample size and lack of data. ^fDowngraded one level for serious study limitations: high risk of other bias: it is not entirely clear what is measured with the assay used. The fact that both active cardenolides and (inactive) metabolites might be detected by the assay compromise the results of these analyses, as they might explain the wide variability observed. ^gDowngraded one level for serious imprecision: low number of events.

Summary of findings 3. MDAC + hospital intervention versus SDAC + hospital intervention for first aid in patients with acute oral poisoning

Summary of findings 4. SDAC versus syrup of ipecac for first aid in patients with acute oral poisoning

SDAC versus syrup of ipecac for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (anti‐inflammatory drugs, analgesics or psychotropic drugs) Setting: hospital setting Intervention: single‐dose activated charcoal (SDAC) Comparison: syrup of ipecac
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with syrup of ipecac	Risk with SDAC
Incidence of mortality	No studies collected this outcome
Incidence of adverse events	Study population		RR 1.24 (0.26 to 5.83)	34 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain about the effect of SDAC, compared to syrup of ipecac on incidence of adverse events.
	154 per 1000	191 per 1000 (40 to 897)
Incidence and severity of symptoms of poisoning: level of coma assessed with Glasgow Coma Scale Scale from: 3 to 15 Follow‐up: 1 h	The mean incidence and severity of symptoms of poisoning: level of coma was 14.91	MD 0.15 lower (0.43 lower to 0.13 higher)	—	34 (1 RCT)	⊕⊝⊝⊝ Very low^a,b,d	We are uncertain about the effect of SDAC, compared to syrup of ipecac on the level of coma.
Duration of toxic symptoms	No studies collected this outcome
Drug absorption	No studies collected this outcome
Incidence of hospitalization	No studies collected this outcome
Incidence of ICU admission	No studies collected this outcome
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; MD: mean difference; RCT: randomized controlled trial; RR: risk ratio; SDAC: single‐dose activated charcoal.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious study limitations: high risk of selection bias. ^bDowngraded one level for serious indirectness: study conducted in a hospital setting. ^cDowngraded one level for serious imprecision: low number of events and wide confidence intervals. ^dDowngraded one level for serious imprecision: low sample size.

Summary of findings 4. SDAC versus syrup of ipecac for first aid in patients with acute oral poisoning

Summary of findings 5. MDAC + hospital intervention versus hospital intervention for first aid in patients with acute oral poisoning

MDAC + hospital intervention versus hospital intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (benzodiazepines, yellow oleander or combinations of different drugs) Setting: hospital setting Intervention: multi‐dose activated charcoal (MDAC) + hospital intervention Comparison: hospital intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with hospital intervention	Risk with MDAC + hospital intervention
Incidence of mortality	Study population		RR 0.94 (0.72 to 1.22)	3085 (1 RCT)	⊕⊕⊝⊝ Low^a,b	MDAC in addition to hospital treatment may make little or no difference in incidence of mortality.
	68 per 1000	64 per 1000 (49 to 82)
Incidence of adverse events	Study population		RR 1.02 (0.52 to 1.98)	3085 (1 RCT)	⊕⊕⊝⊝ Low^a,b	MDAC in addition to hospital treatment may make little or no difference in incidence of adverse events.
	11 per 1000	11 per 1000 (6 to 22)
Incidence and severity of symptoms of poisoning: incidence of need for intubation	Study population		RR 0.97 (0.71 to 1.33)	3085 (1 RCT)	⊕⊕⊝⊝ Low^a,b	MDAC in addition to hospital treatment may make little or no difference in incidence of need for intubation.
	49 per 1000	47 per 1000 (35 to 65)
Duration of toxic symptoms: length of intubation (h)	The median (IQR) length of intubation in the intervention group was 83.8 (35.0 to 173.0) h and 88.5 (38.5 to 203.1) h in the control group and was reported not to differ significantly (P > 0.05); unclear number of participants		—	(1 RCT)	⊕⊕⊝⊝ Low^a,c	The number of participants analysed was not reported. MDAC in addition to hospital treatment may make little or no difference in length of intubation
Drug absorption: cardenolide: AUC (µg/L × h) Follow‐up: 1 day	The median (IQR) cardenolide AUC in the intervention group was 17.3 (12.8 to 21.7) (µg/L) × h and 19.0 (13.7 to 24.3) (µg/L) × h in the control group.		—	76 (1 RCT)	⊕⊝⊝⊝ Very low^a,c,d	We are uncertain about the effects of MDAC in addition to hospital treatment on cardenolide drug absorption.
Incidence of hospitalization	No studies collected or reported this outcome
Incidence of ICU admission	No studies collected or reported this outcome
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). AUC: area under the receiver operating curve; CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; MDAC: multi‐dose activated charcoal; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious indirectness: study conducted in a hospital setting. ^bDowngraded one level for serious imprecision: low number of events and wide confidence interval. ^cDowngraded one level for serious imprecision: low sample size and lack of data. ^dDowngraded one level for serious study limitations: high risk of other bias: it is not entirely clear what is measured with the assay used. The fact that both active cardenolides and (inactive) metabolites might be detected by the assay compromise the results of these analyses, as they might explain the wide variability observed.

Summary of findings 5. MDAC + hospital intervention versus hospital intervention for first aid in patients with acute oral poisoning

Summary of findings 6. Syrup of ipecac versus no intervention for first aid in patients with acute oral poisoning

Syrup of ipecac versus no intervention for first aid in patients with acute oral poisoning
Patient or population: first aid in patients with acute oral poisoning (toxic berries) Setting: pre‐hospital setting Intervention: syrup of ipecac Comparison: no intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	№ of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with no intervention	Risk with Syrup of ipecac
Incidence of mortality	No studies collected this outcome
Incidence of adverse events: diarrhoea Follow‐up: 1 day	Study population		RR 4.08 (1.66 to 10.04)	103 (1 RCT)	⊕⊕⊝⊝ Low^a,b	Syrup of ipecac may result in an increased incidence of diarrhoea.
	96 per 1000	392 per 1000 (160 to 965)
Incidence and severity of symptoms of poisoning	No studies collected this outcome
Duration of toxic symptoms	No studies collected this outcome
Drug absorption	No studies collected this outcome
Hospitalization: incidence of hospitalization Follow‐up: 1 days	—		—	103 (1 RCT)	⊕⊕⊝⊝ Low^a,b	The effect was not estimable due to the absence of events in the intervention (0/52) and the control group (0/51).
ICU admission	No studies collected this outcome
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; ICU: intensive care unit; RCT: randomized controlled trial; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.
^aDowngraded one level for serious study limitations: high risk of selection bias, high risk of detection bias, and high risk of other bias (no confirmation of actual ingestion or uptake, reporting of dichotomous outcomes while measuring with an ordinal scale). ^bDowngraded one level for serious imprecision: low number of events.

Summary of findings 6. Syrup of ipecac versus no intervention for first aid in patients with acute oral poisoning

Table 1. Sources of individual studies

Author and year of publication	Title
Abrass 2012	The evidence for activated charcoal in resource poor settings: a systematic review
American Academy of Clinical Toxicology 1999	Position statement and practice guidelines on the use of multi‐dose activated charcoal in the treatment of acute poisoning
American Academy of Clinical Toxicology 2004	Position paper: cathartics
Chiew 2018	Interventions for paracetamol (acetaminophen) overdoses
Chyka 2005	Position paper: single‐dose activated charcoal
Eddleston 2003	Does gastric lavage really push poisons beyond the pylorus? A systematic review of the evidence
Blain 2011	Organophosphorus poisoning (acute)
Höjer 2013	Position paper update: ipecac syrup for gastrointestinal decontamination.
Jones 2002	Towards evidence based emergency medicine: best BETs from the Manchester Royal Infirmary. Activated charcoal and gastric absorption of iron compounds
Manoguerra 2005	Guideline on the use of ipecac syrup in the out‐of‐hospital management of ingested poisons
Qureshi 2011	Adverse effects of activated charcoal used for the treatment of poisoning
Roberts 2011	Enhanced elimination in acute barbiturate poisoning ‐ a systematic review

Table 1. Sources of individual studies

Table 2. Overview of comparisons

Comparison	Type of poisoning	Study
A. First aid interventions that limit or delay the absorption of the poison in the body
SDAC vs no intervention	Not specified	Merigian 1990
SDAC vs no intervention	Paracetamol	Underhill 1990
SDAC + hospital intervention vs hospital intervention	Not specified	Comstock 1982
	Benzodiazepines + paracetamol or other drug combinations	Cooper 2005
	Tricyclic antidepressants	Crome 1983
	Yellow oleander, organophosphorus/carbamate pesticide, organochlorine, other/unknown pesticide or paraquat, medicine or unknown	Eddleston 2008
	Amitriptyline, clomipramine, mianserin, imipramine, dothiepin, doxepin, nortriptyline, mixed overdoses with most commonly benzodiazepines or alcohol	Hultén 1988
	Not specified	Merigian 2002
	Yellow oleander	Roberts 2006
MDAC + hospital intervention vs SDAC + hospital intervention	Carbamazepine	Behnoush 2009
	Carbamazepine	Brahmi 2006
	Yellow oleander	De Silva 2003
	Yellow oleander, organophosphorus/carbamate pesticide, organochlorine, other/unknown pesticide or paraquat, medicine or unknown	Eddleston 2008
	Yellow oleander	Roberts 2006
SDAC vs syrup of ipecac	Anti‐inflammatory drugs, analgesics or psychotropic drugs	Amigó Tadín 2002
MDAC + hospital intervention vs hospital intervention	Benzodiazepine	Bouget 1989
	Yellow oleander, organophosphorus/carbamate pesticide, organochlorine, other/unknown pesticide or paraquat, medicine or unknown	Eddleston 2008
	Yellow oleander	Roberts 2006
B. First aid interventions that evacuate the poison from the gastrointestinal tract
Emetics
Syrup of ipecac vs no intervention	Toxic berries	Wax 1999
Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic	Not specified	Albertson 1989
	Wide variety, most commonly paracetamol	Kornberg 1991
	Not specified	Kulig 1985
	Paracetamol, salicylate, phenothiazines or ethanol, or other drugs	Pond 1995
Syrup of ipecac 15 mL vs syrup of ipecac 30 mL (dose)	Benzodiazepine tranquillizers or hypnotics, other tranquillizers, other hypnotics, antidepressants, analgesics, antihistamines, miscellaneous drugs and chemicals	Ilett 1977
Cathartics
SDAC + cathartic vs SDAC	Not specified	Sue 1994
SDAC + cathartic vs SDAC	Analgesics, anticonvulsants, antihistamines and decongestants, asthma therapies, automotive products, cardiovascular drugs, gastrointestinal preparations, insecticides, mushrooms, psychotropic drugs, rodenticides, topicals, miscellaneous drugs	James 1995
SDAC + cathartic vs SDAC + cathartic (dose)	Not specified	Sue 1994
SDAC + cathartic vs SDAC + cathartic (type)	Analgesics, anticonvulsants, antihistamines and decongestants, asthma therapies, automotive products, cardiovascular drugs, gastrointestinal preparations, insecticides, mushrooms, psychotropic drugs, rodenticides, topicals, miscellaneous drugs	James 1995
C. Combined first aid interventions that limit uptake and promote evacuation of the poison from the gastrointestinal tract
SDAC + cathartic + hospital intervention vs hospital intervention	Benzodiazepines, barbiturates or imipramine	Passeron 1989
MDAC + cathartic + hospital intervention vs hospital intervention	Paracetamol	Montoya‐Cabrera 1999
D. First aid interventions that neutralize or dilute the poison
No studies were identified
APF: Australian Pharmaceutical Formulary; MDAC: multi‐dose activated charcoal; SDAC: single‐dose activated charcoal; USP: United States Pharmacopeia.

Table 2. Overview of comparisons

Table 3. Additional pre‐defined outcomes reported in the included studies

A. First aid interventions that limit or delay the absorption of the poison in the body
SDAC vs no intervention
Incidence of clinical deterioration
	SDAC		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Merigian 1990	0	220	0	231	Not estimable
Incidence of ICU admission
	SDAC		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Merigian 1990	0	220	0	231	7.77 (0.15 to 391.93)
SDAC + hospitalintervention vs hospitalintervention
Incidence of clinical deterioration
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	Peto OR (95% CI)
Merigian 2002	0	455	0	1075	Not estimable
Grade of coma (4 h after admission)
	SDAC + hospital treatment (N = 9)		Hospital treatment (N = 7)
Study	Median	IQR	Median	IQR	Median difference (P value)
Crome 1983	2	(1 to 3)	2	(1 to 2.5)	0 (P = 0.55)
Grade of coma (8 h after admission)
	SDAC + hospital treatment (N = 9)		Hospital treatment (N = 7)
Study	Median	IQR	Median	IQR	Median difference (P value)
Crome 1983	2	(1 to 3)	1	(0.5 to 2)	1 (P = 0.38)
Grade of coma (24 h after admission)
	SDAC + hospital treatment (N = 9)		Hospital treatment (N = 7)
Study	Median	IQR	Median	IQR	Median difference (P value)
Crome 1983	1	(0 to 2)	0	(0 to 0.5)	1 (P = 0.27)
Incidence of coma grade III (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	7	34	6	43	1.48 (0.55 to 3.98)
Incidence of coma grade IV (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	7	43	0.18 (0.02 to 1.40)
Incidence of coma grade III (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	2	34	2	43	1.26 (0.19 to 8.52)
Incidence of coma grade IV (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	0	34	2	43	0.16 (0.01 to 2.70)
Incidence of coma grade III (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	Peto OR (95% CI)
Hultén 1988	0	34	0	43	Not estimable
Incidence of coma grade IV (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	Peto OR (95% CI)
Hultén 1988	0	34	0	43	Not estimable
Incidence of need for cardiac pacing/antitoxin
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Eddleston 2008	101	549	101	555	1.01 (0.79 to 1.30)
Incidence of need for respirator
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	2	34	9	43	0.28 (0.06 to 1.22)
Incidence of systolic blood pressure < 100 mmHg (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	4	34	8	43	0.63 (0.21 to 1.92)
Incidence of systolic blood pressure < 100 mmHg (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	5	43	0.25 (0.03 to 2.06)
Incidence of systolic blood pressure < 100 mmHg (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	0	34	2	43	0.16 (0.01 to 2.70)
Incidence of heart rate > 100 bpm (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	10	34	15	43	0.84 (0.44 to 1.63)
Incidence of heart rate > 100 bpm (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	8	34	10	43	1.01 (0.45 to 2.28)
Incidence of heart rate > 100 bpm (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	10	34	10	43	1.26 (0.60 to 2.68)
Incidence of cardiac arrhythmias (4 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	4	43	0.32 (0.04 to 2.70)
Incidence of cardiac arrhythmias (8 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	3	43	0.42 (0.05 to 3.87)
Incidence of cardiac arrhythmias (24 h after admission)
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	1	34	2	43	0.63 (0.06 to 6.68)
Incidence of intubation > 8 h
	SDAC + hospital treatment		Hospital treatment
Study	Events	Total	Events	Total	RR (95% CI)
Hultén 1988	4	34	9	43	0.56 (0.19 to 1.67)
B. First aid interventions that evacuate the poison from the gastrointestinal tract
Syrup of ipecac vs no intervention
Incidence of referrals to the emergency department
	Syrup of ipecac		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Wax 1999	1	51	0	52	7.54 (0.15 to 378.83)
Incidence of hospitalizations
	Syrup of ipecac		No intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Wax 1999	0	51	0	52	Not estimable
SDAC + cathartic vs SDAC (higher dose)
Incidence of adverse events (6 mL vs 4 mL)
	SDAC + 6 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	0	16	0	16	Not estimable
Incidence of adverse events (8 mL vs 4 mL)
	SDAC + 8 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	0	18	0	16	Not estimable
Incidence of adverse events (8 mL vs 6 mL)
	SDAC + 8 mL cathartic		SDAC + 6 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	0	18	0	16	Not estimable
Incidence of hospitalization (6 mL vs 4 mL)
	SDAC + 6 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	1	16	0	16	7.39 (0.15 to 372.38)
Incidence of hospitalization (8 mL vs 4 mL)
	SDAC + 8 mL cathartic		SDAC + 4 mL cathartic
Study	Events	Total	Events	Total	Peto OR (95% CI)
Sue 1994	2	18	0	16	7.01 (0.42 to 117.63)
C. Combined first aid interventions that limit uptake and promote removal of the poison
MDAC + cathartic + hospitalintervention vs hospitalintervention
Incidence of adverse events
	MDAC + cathartic + hospital intervention		Hospital intervention
Study	Events	Total	Events	Total	Peto OR (95% CI)
Montoya‐Cabrera 1999	0	7	0	7	Not estimable
T_1/2 (h)
	MDAC + cathartic + hospital intervention (N = 7)		Hospital intervention (N = 7)
Study	Mean	SD	Mean	SD	MD (95% CI)
Montoya‐Cabrera 1999	10	N/A	17	N/A	−7 (not estimable)
bpm: beats per minute; CI: confidence interval; IQR: interquartile range; OR: odds ratio; RR: risk ratio; MDAC; multi‐dose activated charcoal; SDAC: single‐dose activated charcoal.

Table 3. Additional pre‐defined outcomes reported in the included studies

Table 4. Pre‐defined outcomes that were extracted but not included in the review

A. First aid interventions that limit or delay the absorption of the poison in the body
SDAC vs no intervention
Length of stay in the emergency department (min)
	SDAC			No intervention
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Merigian 1990	252	279	220	230	166	231	22.00 (−20.63 to 64.63)
SDAC + hospitalintervention vs hospitalintervention
Length of ICU stay (h)
	SDAC + hospital intervention			Hospital intervention
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Merigian 2002	54.4	93.15	28	45.5	36.3	32	8.90 (−27.82 to 45.62)
Incidence of ICU stay > 3 days
	SDAC + hospital intervention			Hospital intervention
Study	Events	Total		Events	Total		RR (95% CI)
Hultén 1988	0	34		5	43		0.11 (0.01 to 2.00)
Length of hospital stay (h)
	SDAC + hospital intervention			Hospital intervention
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Merigian 2002	63.8	79.8	51	91.7	103.97	102	−27.90 (−57.68 to 1.88)
Study	Median	IQR	N	Median	IQR	N	Median difference (P value)
Cooper 2005	6.8	(4.0 to 14.0)	166	5.5	(3.0 to 12.0)	161	1.3 (P = 0.11)
Incidence of hospital stay > 3 days
	SDAC + hospital intervention			Hospital intervention
Study	Events	Total		Events	Total		RR (95% CI)
Hultén 1988	1	34		4	43		0.32 (0.04 to 2.70)
Length of stay in the emergency department (h)
	SDAC + hospital intervention			Hospital intervention
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Merigian 2002	6.2	3.9	325	5.3	3.9	941	0.90 (0.41 to 1.39)
MDAC + hospitalintervention vs hospitalintervention
Length of hospital stay (h)
	MDAC + hospital intervention			Hospital intervention
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Behnoush 2009	31.0	8.9	30	55.0	8.5	30	−24.00 (−28.40 to −19.60)
Brahmi 2006	30.3	3.4	6	39.7	7.3	6	−9.40 (−15.84 to −2.96)
	Median	IQR	N	Median	IQR	N	Median difference (P value)
De Silva 2003	3	(0.25 to 24)	201	3	(0.5 to 10)	200	0 (P = 0.90)
SDAC vs syrup of ipecac
Length of stay in the emergency department (min)
	Syrup of ipecac			SDAC
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Amigó Tadín 2002	113.21	66.0	21	81.46	27.92	13	31.75 (−0.30 to 63.80)
B. First aid interventions that evacuate the poison from the gastrointestinal tract
Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic
Length of hospital stay (days)
	Syrup of ipecac + SDAC + cathartic			SDAC + cathartic
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Albertson 1989	2.4	5.8	13	1.7	5.2	12	0.70 (−3.61 to 5.01)
Length of stay in the emergency department (h)
	Syrup of ipecac + SDAC + cathartic			SDAC + cathartic
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Albertson 1989	6.8	2.9	93	6.0	2.1	107	0.80 (0.09 to 1.51)
Kornberg 1991	4.1	1.1	29	3.4	1.2	38	0.70 (0.15 to 1.25)
Length of ICU stay (h)
	Syrup of ipecac + SDAC + cathartic			SDAC + cathartic
Study	Mean	SD	N	Mean	SD	N	MD (95% CI)
Albertson 1989	1.8	3.9	6	1.0	0.0	5	Not estimable
CI: confidence interval; ICU: intensive care unit; IQR: interquartile range; MD: mean difference; SD: standard deviation; SDAC: single‐dose activated charcoal.

Table 4. Pre‐defined outcomes that were extracted but not included in the review

Comparison 1. SDAC vs no intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of adverse events Show forest plot	2	476	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.17 [0.30, 57.26]

1.1 Incidence of adverse events	1	451	Peto Odds Ratio (Peto, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Incidence of vomiting	1	25	Peto Odds Ratio (Peto, Fixed, 95% CI)	4.17 [0.30, 57.26]

Comparison 1. SDAC vs no intervention

Comparison 2. SDAC + hospital intervention vs hospital intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of mortality Show forest plot	2	3425	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.04 [0.79, 1.37]

2 Incidence of adverse events Show forest plot	3		Risk Ratio (M‐H, Random, 95% CI)	Subtotals only

2.1 Incidence of vomiting	2	1806	Risk Ratio (M‐H, Random, 95% CI)	1.44 [0.88, 2.37]
2.2 Incidence of absent bowel sounds	1	3098	Risk Ratio (M‐H, Random, 95% CI)	0.41 [0.17, 1.00]
3 Incidence of need for intubation Show forest plot	4	3562	Risk Ratio (M‐H, Random, 95% CI)	1.37 [0.76, 2.47]

3.1 Gastric lavage prior to SDAC	2	3175	Risk Ratio (M‐H, Random, 95% CI)	0.95 [0.70, 1.27]
3.2 No gastric lavage prior to SDAC	2	387	Risk Ratio (M‐H, Random, 95% CI)	2.61 [1.38, 4.93]
4 Incidence of convulsions Show forest plot	2		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

5 Duration of intubation (h) Show forest plot			Other data	No numeric data

6 AUC ((µg/L) × h) Show forest plot			Other data	No numeric data

7 C_max (µg/L) Show forest plot			Other data	No numeric data

8 T_max (h) Show forest plot			Other data	No numeric data

9 Incidence of hospitalization Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

10 Incidence of ICU admission Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 2. SDAC + hospital intervention vs hospital intervention

Comparison 3. SDAC vs syrup of ipecac

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of adverse events Show forest plot	1		Risk Difference (M‐H, Random, 95% CI)	Totals not selected

2 Glasgow Coma Scale score Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

3 Mean arterial blood pressure (mmHg) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

4 Heart rate (bpm) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

5 Respiratory rate (breaths/min) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

Comparison 3. SDAC vs syrup of ipecac

Comparison 4. MDAC + hospital intervention vs SDAC + hospital intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of mortality Show forest plot	2	3476	Risk Ratio (M‐H, Random, 95% CI)	0.59 [0.21, 1.63]

2 Incidence of adverse events Show forest plot	2	3476	Peto Odds Ratio (Peto, Fixed, 95% CI)	3.55 [1.85, 6.79]

3 Incidence of need for cardiac pacing/antitoxin treatment Show forest plot	2	1490	Risk Ratio (M‐H, Random, 95% CI)	0.26 [0.02, 4.18]

4 Incidence of life‐threatening arrhythmias Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

5 Amount of atropine administered (mg) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

6 Number of atropine boluses administered Show forest plot			Other data	No numeric data

7 Incidence of need for intubation Show forest plot	2	3097	Risk Ratio (M‐H, Random, 95% CI)	1.01 [0.75, 1.38]

8 Incidence of convulsions Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

9 Duration of coma (h) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

10 Duration of intubation (h) Show forest plot			Other data	No numeric data

11 C_max (µg/L) Show forest plot			Other data	No numeric data

12 T_max (h) Show forest plot			Other data	No numeric data

13 T_1/2 (h) Show forest plot	1		Mean Difference (IV, Random, 95% CI)	Totals not selected

14 AUC ((µg/L) × h) Show forest plot			Other data	No numeric data

15 Incidence of ICU admission Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 4. MDAC + hospital intervention vs SDAC + hospital intervention

Comparison 5. MDAC + hospital intervention vs hospital intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of mortality Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

2 Incidence of adverse events Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

3 Incidence of need for intubation Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

4 Incidence of seizures Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

5 Incidence of need for cardiac pacing/antitoxin treatment Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

6 Length of intubation (h) Show forest plot			Other data	No numeric data

7 AUC Show forest plot			Other data	No numeric data

8 C_max Show forest plot			Other data	No numeric data

9 T_max Show forest plot			Other data	No numeric data

Comparison 5. MDAC + hospital intervention vs hospital intervention

Comparison 6. Syrup of ipecac vs no intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of diarrhoea Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

2 Incidence of abdominal pain Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

3 Incidence of sedation Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

4 Incidence of agitation Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 6. Syrup of ipecac vs no intervention

Comparison 7. Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of mortality Show forest plot	2		Peto Odds Ratio (Peto, Fixed, 95% CI)	Totals not selected

2 Incidence of adverse events Show forest plot	3	764	Risk Ratio (M‐H, Random, 95% CI)	2.59 [1.37, 4.91]

3 Incidence of clinical improvement Show forest plot	3	989	Risk Ratio (M‐H, Random, 95% CI)	1.00 [0.83, 1.21]

4 Incidence of clinical deterioration Show forest plot	2	970	Risk Ratio (M‐H, Random, 95% CI)	0.88 [0.46, 1.69]

5 Incidence of hospitalization Show forest plot	3	746	Peto Odds Ratio (Peto, Fixed, 95% CI)	1.17 [0.69, 1.98]

6 Incidence of ICU admission Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 7. Syrup of ipecac + SDAC + cathartic vs SDAC + cathartic

Comparison 8. SDAC + cathartic vs SDAC

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of adverse events Show forest plot	2		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

1.1 Vomiting	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
1.2 Lethargy during follow‐up	1		Risk Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2 Incidence of hospitalization Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 8. SDAC + cathartic vs SDAC

Comparison 9. SDAC + cathartic vs SDAC + cathartic (higher dose)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of hospitalization (8 mL vs 6 mL) Show forest plot	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Comparison 9. SDAC + cathartic vs SDAC + cathartic (higher dose)

Comparison 10. SDAC + cathartic vs SDAC + cathartic (different type)

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of vomiting (sorbitol vs magnesium sulphate Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

2 Incidence of vomiting (sorbitol vs magnesium citrate) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

3 Incidence of vomiting (magnesium sulphate vs magnesium citrate) Show forest plot	1		Risk Ratio (M‐H, Random, 95% CI)	Totals not selected

Comparison 10. SDAC + cathartic vs SDAC + cathartic (different type)

Comparison 11. SDAC + cathartic + hospital intervention vs hospital intervention

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of adverse events Show forest plot	1		Peto Odds Ratio (Peto, Fixed, 95% CI)	Totals not selected

Comparison 11. SDAC + cathartic + hospital intervention vs hospital intervention