Aminofilina para el paro cardíaco bradiasistólico en adultos
Resumen
Antecedentes
En la isquemia cardíaca, la acumulación de adenosina puede provocar o exacerbar la bradiasistolia y disminuir la efectividad de las catecolaminas que se administran durante la reanimación. La aminofilina es un antagonista competitivo de la adenosina. Estudios de casos indican que la aminofilina puede ser eficaz para el paro bradiasistólico resistente a la atropina.
Objetivos
Determinar los efectos de la aminofilina en el tratamiento de los pacientes con paro cardíaco bradiasistólico, principalmente en la supervivencia al alta hospitalaria. También se consideró la supervivencia al ingreso, el retorno de la circulación espontánea, los resultados neurológicos y los eventos adversos.
Métodos de búsqueda
Para esta revisión actualizada se realizaron búsquedas en el Registro Cochrane Central de Ensayos Controlados (Cochrane Central Register of Controlled Trials), MEDLINE, EMBASE, CINAHL, LILACS, ClinicalTrials.gov y en la WHO International Clinical Trials Registry Platform en noviembre de 2014. Se verificaron las listas de referencias de los artículos recuperados, se revisaron las actas de congresos, se estableció contacto con expertos y se realizó una búsqueda adicional mediante Google.
Criterios de selección
Todos los ensayos controlados aleatorizados que compararon la administración de aminofilina intravenosa con placebo en adultos con paro cardíaco bradiasistólico, normotérmico no traumático tratados con soporte vital cardiovascular avanzado (ACLS, por sus siglas en inglés).
Obtención y análisis de los datos
Dos autores de la revisión examinaron de forma independiente los estudios y extrajeron los datos incluidos. Se estableció contacto con los autores de los estudios cuando fue necesario. Se calculó el riesgo relativo (RR) agrupado para cada resultado de los estudios. El análisis de subgrupos se predeterminó según el momento de administración de la aminofilina.
Resultados principales
En este análisis se incluyeron cinco ensayos, todos realizados en un contexto prehospitalario. El riesgo de sesgo fue bajo en cuatro de estos estudios (n = 1186). Los ensayos reclutaron 1254 participantes. Se encontró que la aminofilina no tuvo efectos sobre la supervivencia al alta hospitalaria (riesgo relativo [RR] 0,58; intervalo de confianza [IC] del 95%: 0,12 a 2,74) ni sobre el resultado secundario de supervivencia (supervivencia al ingreso hospitalario: RR 0,92; IC del 95%: 0,61 a 1,39; retorno de la circulación espontánea: RR 1,15; IC del 95%: 0,89 a 1,49). La supervivencia fue poco frecuente (6/1254), lo que hace que los datos sobre los resultados neurológicos y los eventos adversos sean bastante limitados. El análisis de subgrupos planificado de la administración temprana de aminofilina incluyó a 37 participantes. Ningún paciente del subgrupo sobrevivió al alta hospitalaria.
Conclusiones de los autores
La administración prehospitalaria de aminofilina en el paro bradiasistólico no se asocia con un mejor retorno de la circulación, supervivencia al ingreso ni supervivencia al alta hospitalaria. No se conocen los efectos beneficiosos de la aminofilina administrada en las primeras fases de la reanimación.
PICO
Resumen en términos sencillos
Aminofilina para el paro cardíaco
La aminofilina es un fármaco que podría ayudar en la reanimación de los pacientes en paro cardíaco cuando la actividad eléctrica está muy disminuida o ausente. La aminofilina puede restaurar el flujo sanguíneo al corazón, mejorar la actividad eléctrica y hacer que otros fármacos utilizados en la reanimación sean más eficaces. Se encontraron cinco estudios que incluyeron a 1254 pacientes que presentaron este tipo de paro cardíaco en un contexto prehospitalario. Cuatro de los cinco estudios (1186 pacientes) fueron bien diseñados y tuvieron bajo riesgo de sesgo. Aunque no se informaron eventos adversos, agregar la aminofilina a la práctica estándar de reanimación de los paramédicos no mostró ventajas en estos pacientes en comparación con placebo. No se sabe si la administración inmediata de aminofilina sería útil.
Authors' conclusions
Summary of findings
| Aminophylline compared to placebo for bradyasystolic cardiac arrest in adults | ||||||
| Patient or population: participants with bradyasystolic cardiac arrest in adults | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of participants | Quality of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Assumed risk | Corresponding risk | |||||
| Placebo | Aminophylline | |||||
| Survival to hospital discharge | Study population | OR 0.58 | 1254 | ⊕⊕⊕⊝ | ||
| 6 per 1000 | 4 per 1000 | |||||
| Moderate | ||||||
| 0 per 1000 | 0 per 1000 | |||||
| Return of spontaneous circulation | Study population | OR 1.15 | 1254 | ⊕⊕⊕⊕ | ||
| 234 per 1000 | 260 per 1000 | |||||
| Moderate | ||||||
| 200 per 1000 | 223 per 1000 | |||||
| Survival to admission | Study population | OR 0.92 | 1232 | ⊕⊕⊕⊝ | ||
| 87 per 1000 | 80 per 1000 | |||||
| Moderate | ||||||
| 100 per 1000 | 93 per 1000 | |||||
| Early administration of aminophylline: survival to hospital discharge | See comment | See comment | Not estimable | 37 | ⊕⊕⊝⊝ | |
| Early administration of aminophylline: survival to hospital admission | Study population | OR 0.22 | 37 | ⊕⊕⊝⊝ | ||
| 150 per 1000 | 37 per 1000 | |||||
| Moderate | ||||||
| 226 per 1000 | 60 per 1000 | |||||
| Early administration of aminophylline: return of spontaneous circulation | Study population | OR 1.61 | 37 | ⊕⊕⊝⊝ | ||
| 250 per 1000 | 349 per 1000 | |||||
| Moderate | ||||||
| 284 per 1000 | 390 per 1000 | |||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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). | ||||||
| GRADE Working Group grades of evidence: | ||||||
| aDirks 1999 has unclear risk of bias because of limited information about the methodology on which to base the assessment. However, this represents a small number of participants. | ||||||
Background
Description of the condition
The incidence of out‐of‐hospital cardiac arrest is estimated to be 58.9 to 78.8 per 100,000 person‐years (Hasegawa 2013). People suffering a cardiac arrest may have complete loss of cardiac electrical activity and heart rhythm (asystole), a slow heart rhythm but no cardiac output or a disorganised rhythm (ventricular fibrillation). Bradyasystole, defined as the absence of cardiac electrical activity confirmed in more than one lead or pulseless electrical activity at a rate of less than 60 beats per minute, is the most commonly recorded initial rhythm in cardiac arrest (Herlitz 1994; Sedgwick 1994; Steill 2004; Vaillancourt 2004). Fewer than 10% of patients suffering cardiac arrest (all rhythms) survive to discharge from hospital (Kette 1998; Hasegawa 2013). Overall survival from 'non‐shockable' rhythms is estimated to be less than 3% (Thomas 2013). The outlook for asystole, in particular, is more dismal, with fewer than 1% surviving (Steill 2004; Vaillancourt 2004). Interventions to improve outcomes in this group may have a significant impact in terms of lives saved.
Clinical trials on interventions in cardiac arrest have been hindered by variability in data definitions and outcome measures. This has been addressed in part by the Utstein style reporting template for out‐of‐hospital cardiac arrest from the International Liaison Committee on Resuscitation (ILCOR) (Cummins 1991; Cummins 1997; Jacobs 2004). These guidelines represent an international consensus on the definitions and core data needed to report and study resuscitation for both research purposes and clinical benchmarking. It is interesting to note that ILCOR could not reach a consensus on the definition of asystole or bradyasystole (Jacobs 2004).
Description of the intervention
Aminophylline is a complex of theophylline and ethylenediamine. Aminophylline's active metabolite is theophylline, and pharmacokinetic data quoted in product monographs reference the active metabolite rather than the parent compound. These methylxanthines are metabolised by the liver with about 10% excreted (unchanged) by the kidneys (Micromedex Solutions). The average plasma half‐life is 7 to 9 hours in healthy non‐smoking adults and 4 to 5 hours in adult smokers (Rudusky 2005). The median half‐life in septic patients is prolonged (18.8 hours). The specific pharmacokinetics of aminophylline administered in the setting of cardiac arrest is not known.
Methylxanthines are known to have numerous drug interactions. Prolonged clearance of theophylline has been particularly reported with cimetidine, macrolide antibiotics and ciprofloxacin. Adverse effects in the setting of cardiac arrest have not been reported.
How the intervention might work
Adenosine is an endogenous nucleotide that plays a role in the regulation of myocardial oxygen supply and demand (Mader 2000). In periods of cellular hypoxia, production of adenosine is increased, leading to accumulation in ischaemic cardiac muscle. In what is believed to be a cardioprotective mechanism, adenosine acts to increase oxygen supply through coronary vasodilation and to diminish oxygen demand by reducing intrinsic pacemaker activity, blocking conduction at the atrioventricular (AV) node and attenuating the response to catecholamines (Schrader 1977; Dobson 1983; Wesley 1989; Malcolm 1993; Belardinelli 1995; Mader 2000). These actions in the setting of cardiac ischaemia may lead to bradycardia or bradyasystole resistant to atropine, as it is independent of parasympathetic tone. Furthermore, adenosine diminishes the effectiveness of exogenous catecholamines (Belardinelli 1989; Visentin 1990).
Aminophylline is a competitive nonspecific antagonist of adenosine that enhances cardiac response to beta‐agonists and stimulates endogenous catecholamine release (Rudusky 2005). The pathophysiology of bradyasystolic arrest suggests that intravenous aminophylline may be a promising therapy for patients in whom atropine is ineffective (Mader 2000).
Why it is important to do this review
Although reported cases of successful resuscitation with aminophylline generated enthusiasm for this therapy, little has changed with respect to its use in standard care in the last 20 years (Viskin 1993; Perouansky 1998; Lee 2000; Neumar 2010). The adult advanced cardiac life support (ACLS) guidelines in 2010 did not even mention aminophylline (Neumar 2010). To date, there have been no meta‐analyses of randomised controlled trials on the use of aminophylline in bradyasystolic arrest. This systematic review of randomised controlled trials compares intravenous aminophylline with placebo in participants with bradyasystolic cardiac arrest to determine whether evidence is available to justify adopting aminophylline as a standard therapy in these patients. The low event rate (survival to hospital discharge) means that a large study or a large meta‐analysis is needed to answer this question. We will update this study on an ongoing basis to find new studies.
Objectives
To determine the effects of aminophylline in the treatment of patients in bradyasystolic cardiac arrest, primarily survival to hospital discharge. We also considered survival to admission, return of spontaneous circulation, neurological outcomes and adverse events.
Methods
Criteria for considering studies for this review
Types of studies
All randomised controlled trials comparing intravenous aminophylline with the administration of placebo.
Types of participants
Adult participants (16 years of age or older) with non‐traumatic, normothermic bradyasystolic cardiac arrest. Bradyasystole must have been present at the time of administration of aminophylline.
Types of interventions
Bradyasystole was defined as the absence of cardiac electrical activity as confirmed in more than one lead or pulseless electrical activity at a rate of fewer than 60 beats per minute. All interventions that compare intravenous aminophylline with placebo during resuscitation from bradyasystolic cardiac arrest were studied, in addition to standard ACLS. Depending on the preceding cardiac rhythms, standard care may include atropine, epinephrine, vasopressin or a combination of these. We did not analyse the combinations of drugs administered along with aminophylline. However, we planned an a priori subgroup to assess whether aminophylline was administered early or late during resuscitation.
Types of outcome measures
Primary outcome.
-
Survival to hospital discharge.
Secondary outcomes.
-
Return of spontaneous circulation.
-
Survival to hospital admission (survived event).
-
Neurological outcome.
-
Adverse events.
Survival to hospital discharge, survival to hospital admission and return of spontaneous circulation were defined according to the Utstein style guidelines and templates (Jacobs 2004).
Search methods for identification of studies
We carried out a comprehensive search to identify relevant trials, irrespective of language and publication status.
We initially searched the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (Issue 4, 2009), MEDLINE, EMBASE, CINAHL and LILACS on 26 November 2009 (Appendix 1; Hurley 2007; Hurley 2013). We performed new searches of CENTRAL, MEDLINE, EMBASE and CINAHL in November 2014 (Appendix 2). We also searched ClinicalTrials.gov (www.clinicaltrials.gov) and the World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP; http://apps.who.int/trialsearch/), using the terms 'cardiac arrest' OR 'asystole' OR 'bradyasystole' as the condition and 'aminophylline' OR 'adenosine antagonist' as the intervention.
We used similar terms to search the grey literature through Google Scholar, the Canada Institute for Scientific and Technical Information (CISTI) Catalogue and the British Library Public Catalogue. We reviewed reference lists of all available primary studies and review articles to identify potentially relevant citations. We also contacted authors of primary studies to inquire about other published or unpublished trials known to them. We contacted scientific advisors for the pharmaceutical companies that manufacture aminophylline (Omega Laboratories and Hospira Healthcare Corporation) for any unpublished results on the use of aminophylline in asystolic arrest. Finally, we reviewed reference lists of relevant trials and review articles as well as conference proceedings from the Canadian Association of Emergency Physicians, the American College of Emergency Physicians, the Society for Academic Emergency Medicine and the American Heart Association (January 1997 to December 2012) and from the European Society of Cardiology (January 2005 to December 2012).
Data collection and analysis
Selection of studies
Two review authors (KFH, RG) screened the full list of titles and abstracts retrieved from the searches, selecting all trials that appeared relevant on the basis of title, abstract and MeSH headings. These two review authors independently reviewed the full‐text articles identified as potentially relevant to select trials for inclusion. We resolved disagreements by discussion or by third party adjudication (KM).
Data extraction and management
Two review authors (KFH, KM) independently extracted data from the trials and entered them into Review Manager 5 (RevMan 2014). Data extraction included the following items.
-
Population: age, sex, aetiology of arrest, inclusion and exclusion criteria.
-
Intervention: dose and time to administration of aminophylline.
-
Control: definition of asystolic arrest, duration of resuscitation after administration of aminophylline, definition of neurological outcome.
-
Outcome: return of spontaneous circulation, survival to admission, survival to discharge, neurological outcome, adverse outcomes.
-
Design: method of randomisation.
Assessment of risk of bias in included studies
Two review authors (KFH, RG) assessed risk of bias using the criteria recommended by the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) (low risk, high risk, unclear).
-
Selection: Was the assignment truly randomised? Was the allocation sequence concealed? (selection bias)
-
Blinding: Were investigators, participants and caregivers unaware of the treatment assignment? (detection bias and performance bias)
-
Losses to follow‐up: Were all study participants accounted for and analyses carried out in an intention‐to‐treat fashion? (attrition bias)
-
Reporting: Were all outcomes reported? (reporting bias)
We were not blinded to the identity of study authors or study results.
Measures of treatment effect
We combined data in RevMan 2014. For dichotomous variables, individual and pooled statistics were calculated as risk ratios (RRs) with 95% confidence intervals (CIs). We used a fixed‐effect model for data synthesis and assessed heterogeneity using the Chi2 test and the I2 statistic (Higgins 2003). For continuous outcomes, we calculated individual and pooled statistics as mean differences (MDs) or standardised mean differences (SMDs) and 95% CIs using a random‐effects model.
Assessment of reporting biases
We planned to carry out the statistical analyses on an intention‐to‐treat basis to deal with missing data from individual trials. The data were evaluated for publication bias using graphical and statistical methods. We checked for publication bias using a funnel plot. We assessed the effects of risk of bias in the included studies by using three criteria: method of assignment, blinding and losses to follow‐up.
Subgroup analysis and investigation of heterogeneity
We planned two subgroup analyses: comparison of results based on early administration of aminophylline using five minutes as the cut‐off point between early and late administration. With respect to use of Utstein style reporting, we considered explicit mention of Utstein style reporting in the study or in correspondence with study authors.
Sensitivity analysis
Where significant heterogeneity (P < 0.1) existed, we divided the groups based on our assessment of risk of bias (high versus low).
Results
Description of studies
Results of the search
In our initial review in 2009, structured database searches yielded 285 records. Further online searches through Google and Google Scholar yielded an additional 21 records. After identifying and removing duplicates, we screened 264 records. We retrieved 22 articles for detailed review. Two authors excluded 14 articles: Nine were Chinese papers that were reported to be randomised trials yet on further inquiry were found not to be prospectively designed (Meng 2001; Ma 2003; Dong 2006; Fu 2006; Jin 2006; Ma 2006; Xiao 2006; Shi 2007; Guo 2008); five articles reported mixed comparisons of aminophylline and epinephrine (Luo 2002; Wu 2002; Luo 2003; Shi 2008; Sun 2009). Two articles were conference abstracts reporting data that were also reported in full text (Mader 2003; Abu‐Laban 2006). See details in our Characteristics of excluded studies table.
The updated search in November 2014 yielded 39 new records, all of which we screened and excluded. We did not retrieve any additional records for inclusion in this review update.
Five studies met inclusion criteria for this review (Mader 1997; Dirks 1999; Mader 1999; Mader 2003; Abu‐Laban 2006). One study, published in abstract form only, is awaiting further classification (Snell 2000). Despite repeated attempts to make contact with its authors, we were unable to garner enough additional methodological information to include any of its data in this review.
See Figure 1 for the study flow diagram.
Study flow diagram.
Included studies
We included data from five studies in this meta‐analysis, summarising each study's particular features in the Characteristics of included studies table.
Setting
The five trials included only out‐of‐hospital arrests and took place in urban, prehospital settings in Canada (n = 971, one study), the United States (n = 215, three studies) and Germany (n = 68, one study). Two studies did not report the specific study period (Mader 1997; Dirks 1999). The remaining studies were conducted from 1996 to 2003.
Participants
The studies focused on adult participants. Abu‐Laban 2006 excluded patients younger than 16 years, and each of Mader's studies excluded patients younger than 21 years (Mader 1997; Mader 1999; Mader 2003). Dirks 1999 did not specify the exclusion criteria in his published abstract, although communications with the author confirm that the trial excluded patients younger than 16 years. All trials excluded patients with suspected traumatic or hypothermic cardiac arrests. Most studies also excluded patients who were pregnant, using theophylline, known to have hypersensitivity to methylxanthines or known to have liver disease. Abu‐Laban 2006 further excluded patients who had evidence of haemorrhage or were on renal dialysis.
Mean age of included participants ranged from 65 to 76.8 years. Females represented a minority (377/1186), although this varied from 28.5% in Abu‐Laban 2006 to 45% in Mader 2003. Dirks 1999 did not report these data.
Four studies reported data about the presumed cause of the cardiac arrest (Dirks 1999; Mader 1999; Mader 2003; Abu‐Laban 2006). Three studies reported that suspected ischaemic cardiac events were the most common cause of the arrest, ranging from 64% to 82% (Dirks 1999; Mader 1999; Abu‐Laban 2006). Mader 2003 reported their suspicion that 20.7% of participants had suffered a primary non‐ischaemic cardiac event, but that most causes remained "uncertain."
Interventions
All trials compared the administration of IV aminophylline with placebo as an adjunct to standard ACLS resuscitation in bradyasystolic cardiac arrest. Four studies used 250 mg doses of aminophylline, and one study used 240 mg doses (Dirks 1999). Four studies required failure of standard ACLS interventions for one to two minutes before the study medication was administered. One study administered aminophylline along with initial doses of epinephrine and atropine (Mader 1999). Mean time to arrival of the paramedic crew ranged from 5.3 to 9.3 minutes. Because of differences in reporting, it is difficult to comment on the time from the event to administration of aminophylline, although the time from arrival of paramedics to administration of aminophylline ranged from 11 to 16.6 minutes (Mader 1997; Dirks 1999; Abu‐Laban 2006).
Risk of bias in included studies
The assessed risk of bias for each study is detailed in the Characteristics of included studies table. In general, the included studies were well‐designed randomised, placebo‐controlled trials. Random sequence generation was aided by computer (Abu‐Laban 2006) or by an investigational drug service (Mader 1997; Mader 1999; Mader 2003). Use of matched syringes, provided by a third party, assisted with both blinding and allocation concealment (Mader 1997; Mader 1999; Mader 2003; Abu‐Laban 2006). Risks of performance bias, detection bias and attrition bias were low for four studies (Mader 1997; Mader 1999; Mader 2003; Abu‐Laban 2006). The degree of risk associated with randomisation, allocation, blinding and incomplete outcomes data for Dirks 1999 remains 'unclear' because the details about study methodology that were available in the published abstract and in the poster provided by the study's lead author were minimal. Further details about methodology could not be obtained from the study's lead author despite repeated attempts to make contact by email. The risk of selective reporting bias was deemed to be low in all studies (Mader 1997; Dirks 1999; Mader 1999; Mader 2003; Abu‐Laban 2006) because each study reported the outcomes expected from a cardiac arrest study, namely, return of spontaneous circulation, survival to admission, survival to discharge and neurological outcomes. See Figure 2 and Figure 3 for a summary of risk of bias.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Effects of interventions
See: Summary of findings 1 Aminophylline compared to placebo for bradyasystolic cardiac arrest in adults
Few disagreements arose between review authors (KFH, RG) in assessing studies for inclusion, assessing risk of bias or extracting data, and they were readily resolved by discussion to reach consensus. With only five studies included in this systematic review, it is difficult to draw conclusions regarding the funnel plot. For example, only two studies included participants who experienced the primary outcome (survival to hospital discharge). The funnel plot for 'return of spontaneous circulation' was asymmetrical, possibly suggesting a publication bias (Figure 4). It is unlikely that smaller studies whose results favoured placebo would have significantly affected the pooled results of this analysis. All results discussed below are given as pooled RR and its 95% CI for aminophylline versus placebo.
Funnel plot of comparison: 1 Aminophylline versus placebo, outcome: 1.2 Return of spontaneous circulation.
Survival to discharge
All studies (including 1254 participants) reported survival to hospital discharge－the primary outcome. These findings were homogeneous (Chi2 = 0.08, degrees of freedom (df) = 1, P = 0.77). This event was rare, with only six survivors in total. We are reporting the analyses using a fixed‐effect model. As seen in Analysis 1.1, aminophylline was found to have no effect (RR 0.58, 95% CI 0.12 to 2.74). No change in the results was noted when a random‐effects model was used. Dirks 1999 reported that their lone survivor was still alive at six months postevent. Four survivors in Abu‐Laban 2006 were still alive at one year, and one refused follow‐up after discharge from the hospital.
Survival to admission
Data on the secondary outcome－survival to hospital admission－were found to be homogeneous (Chi2 = 2.73, df = 3, P = 0.43). Analysis showed that aminophylline had no effect on this secondary survival outcome (Analysis 1.3: RR 0.92, 95% CI 0.61 to 1.39).
Return of spontaneous circulation
Data on return of spontaneous circulation were also found to be homogeneous (Chi2 = 2.41, df = 4, P = 0.66). Analysis showed that aminophylline had no effect on return of spontaneous circulation (Analysis 1.2: RR 1.15, 95% CI 0.89 to 1.49).
Neurologic outcome
Data about neurological outcomes were limited. Abu‐Laban 2006 reported that all five survivors had "excellent neurological and functional outcomes", and Dirks 1999 did not comment about the neurological status of the lone survivor in their study.
Adverse events
None of the trials reported adverse events (Mader 1997; Dirks 1999; Mader 1999; Mader 2003; Abu‐Laban 2006). However, none of the data reporting schema that we reviewed specifically mentioned 'adverse events' as an outcome.
Subgroup analysis
We performed a subgroup analysis on the basis of whether aminophylline was administered early in the resuscitation, using five minutes as the cut‐off point. We used data from two studies in this analysis. Abu‐Laban 2006 provided raw study data, and we pulled relevant published data from Mader 1997 for an accumulated total of 37 participants. In this small sample of participants, no one survived to hospital discharge (Analysis 1.4). Three participants survived to admission (Analysis 1.5), and 11 had return of spontaneous circulation (Analysis 1.6). We did not find benefit derived from the use of aminophylline in this subgroup analysis. Subgroup data from the other studies were not available.
Sensitivity analysis
The results were unchanged (Analysis 1.7; Analysis 1.8; Analysis 1.9) when we only selectively pooled data taken from studies with low risk of bias (Mader 1997; Mader 1999; Mader 2003; Abu‐Laban 2006).
Discussion
Summary of main results
This systematic review did not find a survival benefit for the administration of IV aminophylline during adult, out‐of‐hospital bradyasystolic arrest. We did not find statistically significant heterogeneity between studies. Overall, the studies were well designed and the assessed risk of bias was low. Only Abu‐Laban 2006 was adequately powered to detect a significant difference in event rates for the primary outcome (return of spontaneous circulation). Because return of spontaneous circulation is a precursor to other survival outcomes, and our analysis demonstrated no significant benefit of aminophylline for this outcome, it is reasonable to conclude that a larger study is unlikely to demonstrate significant survival benefit. Nonetheless, no study was powered to detect a significant difference in our primary outcome－survival to hospital discharge. Given the relative rarity of survival to hospital admission (Analysis 1.3) and survival to hospital discharge (Analysis 1.1), drawing conclusions about neurological status and adverse events is problematic as well.
Because the studies included in this review took place in the prehospital setting, the duration of arrest before administration of aminophylline was significant, with mean times from paramedic arrival to administration of study drug as long as 16 minutes (Mader 1997; Abu‐Laban 2006). The subgroup analysis considered the effects of early administration of aminophylline but included only 37 participants with no survivors to hospital discharge (primary outcome). Mader 1999 looked specifically at early administration, with mean times (from diagnosis of asystole to administration of study drug) less than five minutes. However, the mean time from dispatch to arrival to the scene in this population, where 49% of arrests were witnessed, was 6.5 minutes, suggesting that the duration of arrest before aminophylline exceeded ten minutes for most participants. This study accumulated 82 participants, none of whom survived. Hence, the effect of early administration of aminophylline remains unknown.
The unadjusted survival for adult out‐of‐hospital cardiac arrest in a large population‐based registry was 4.9% (Hasegawa 2013). Patients presenting in 'unshockable' rhythms such as asystole and pulseless electrical activity typically have a poorer prognosis (Thomas 2013). The survival rate in this meta‐analysis, which is derived primarily from a large Canadian study (Abu‐Laban 2006), is in keeping with other published Canadian data (Steill 2004; Vaillancourt 2004). Survival from in‐hospital adult asystolic arrest appears to be in the range of 10%, with good neurological outcomes reported in 61% of survivors (Morrison 2013). The improved survival in this population may be attributable, in part, to the rapid onset of resuscitative efforts when a patient suffering a cardiac arrest is both monitored and witnessed (Brady 2011). Although 73% of in‐hospital arrests are both monitored and witnessed, out‐of‐hospital arrests are witnessed by a health care provider in fewer than 9% of cases (Brady 2011; Hasegawa 2013). Use of aminophylline has not been studied in the 'in‐hospital' cardiac arrest population.
Because of our search strategy, we recovered a number of Chinese language studies, most of which were found through extensive searching in Google and Google Scholar. Some studies have noted the poor quality of reporting of randomised controlled trials in leading medical journals and conference proceedings in China (Xu 2008; Duan 2009; Chen 2010). One study found that only 6.8% of apparent randomised controlled trials in China were authentic randomised controlled trials (Wu 2009). We asked the lead author of that study to assess whether the 'apparent' randomised controlled trials that we identified were 'authentic' and found that nine of them were not prospectively designed (Meng 2001; Ma 2003; Dong 2006; Fu 2006; Jin 2006; Ma 2006; Xiao 2006; Shi 2007; Guo 2008). None of the trials based in China met the criteria to be included in this review.
Study flow diagram.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Funnel plot of comparison: 1 Aminophylline versus placebo, outcome: 1.2 Return of spontaneous circulation.
Comparison 1: Aminophylline vs. placebo, Outcome 1: Survival to hospital discharge
Comparison 1: Aminophylline vs. placebo, Outcome 2: Return of spontaneous circulation
Comparison 1: Aminophylline vs. placebo, Outcome 3: Survival to admission
Comparison 1: Aminophylline vs. placebo, Outcome 4: Early administration of aminophylline: Survival to hospital discharge
Comparison 1: Aminophylline vs. placebo, Outcome 5: Early administration of aminophylline: Survival to hospital admission
Comparison 1: Aminophylline vs. placebo, Outcome 6: Early administration of aminophylline: Return of spontaneous circulation
Comparison 1: Aminophylline vs. placebo, Outcome 7: Sensitivity analysis: survival to discharge
Comparison 1: Aminophylline vs. placebo, Outcome 8: Sensitivity analysis: survival to admission
Comparison 1: Aminophylline vs. placebo, Outcome 9: Sensitivity analysis: return of spontaneous circulation
| Aminophylline compared to placebo for bradyasystolic cardiac arrest in adults | ||||||
| Patient or population: participants with bradyasystolic cardiac arrest in adults | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect | No of participants | Quality of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Assumed risk | Corresponding risk | |||||
| Placebo | Aminophylline | |||||
| Survival to hospital discharge | Study population | OR 0.58 | 1254 | ⊕⊕⊕⊝ | ||
| 6 per 1000 | 4 per 1000 | |||||
| Moderate | ||||||
| 0 per 1000 | 0 per 1000 | |||||
| Return of spontaneous circulation | Study population | OR 1.15 | 1254 | ⊕⊕⊕⊕ | ||
| 234 per 1000 | 260 per 1000 | |||||
| Moderate | ||||||
| 200 per 1000 | 223 per 1000 | |||||
| Survival to admission | Study population | OR 0.92 | 1232 | ⊕⊕⊕⊝ | ||
| 87 per 1000 | 80 per 1000 | |||||
| Moderate | ||||||
| 100 per 1000 | 93 per 1000 | |||||
| Early administration of aminophylline: survival to hospital discharge | See comment | See comment | Not estimable | 37 | ⊕⊕⊝⊝ | |
| Early administration of aminophylline: survival to hospital admission | Study population | OR 0.22 | 37 | ⊕⊕⊝⊝ | ||
| 150 per 1000 | 37 per 1000 | |||||
| Moderate | ||||||
| 226 per 1000 | 60 per 1000 | |||||
| Early administration of aminophylline: return of spontaneous circulation | Study population | OR 1.61 | 37 | ⊕⊕⊝⊝ | ||
| 250 per 1000 | 349 per 1000 | |||||
| Moderate | ||||||
| 284 per 1000 | 390 per 1000 | |||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (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). | ||||||
| GRADE Working Group grades of evidence: | ||||||
| aDirks 1999 has unclear risk of bias because of limited information about the methodology on which to base the assessment. However, this represents a small number of participants. | ||||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Survival to hospital discharge Show forest plot | 5 | 1254 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.58 [0.12, 2.74] |
| 1.2 Return of spontaneous circulation Show forest plot | 5 | 1254 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.15 [0.89, 1.49] |
| 1.3 Survival to admission Show forest plot | 5 | 1254 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.92 [0.61, 1.37] |
| 1.4 Early administration of aminophylline: Survival to hospital discharge Show forest plot | 2 | 37 | Odds Ratio (M‐H, Fixed, 95% CI) | Not estimable |
| 1.5 Early administration of aminophylline: Survival to hospital admission Show forest plot | 2 | 37 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.22 [0.02, 2.36] |
| 1.6 Early administration of aminophylline: Return of spontaneous circulation Show forest plot | 2 | 37 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.61 [0.38, 6.77] |
| 1.7 Sensitivity analysis: survival to discharge Show forest plot | 4 | 1186 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.66 [0.11, 3.99] |
| 1.8 Sensitivity analysis: survival to admission Show forest plot | 4 | 1186 | Odds Ratio (M‐H, Fixed, 95% CI) | 0.81 [0.52, 1.25] |
| 1.9 Sensitivity analysis: return of spontaneous circulation Show forest plot | 4 | 1186 | Odds Ratio (M‐H, Fixed, 95% CI) | 1.13 [0.86, 1.48] |
