Probióticos para la prevención de la diabetes gestacional
Resumen
Antecedentes
La diabetes mellitus gestacional (DMG) se asocia con variados desenlaces adversos del embarazo para la madre y el lactante. La prevención de la DMG mediante intervenciones en el estilo de vida ha resultado ser difícil. La microflora intestinal (el compuesto de bacterias presentes en los intestinos) influye en las vías inflamatorias del huésped, el metabolismo de la glucosa y los lípidos y, en otros ámbitos, se ha mostrado que la alteración de la microflora intestinal repercute en estas respuestas del huésped. Los probióticos son una forma de alterar la microflora intestinal, pero se sabe poco acerca de su uso para modificar el ambiente metabólico del embarazo. Esta es una actualización de una revisión publicada por última vez en 2014.
Objetivos
Evaluar de forma sistemática los efectos de la administración de suplementos probióticos solos o en combinación con intervenciones farmacológicas y no farmacológicas en la prevención de la DMG.
Métodos de búsqueda
Se hicieron búsquedas en el Registro de ensayos del Grupo Cochrane de Embarazo y parto (Cochrane Pregnancy and Childbirth Group), ClinicalTrials.gov, la Plataforma de registros internacionales de ensayos clínicos (ICTRP) de la OMS (20 de marzo de 2020) y en las listas de referencias de los estudios identificados.
Criterios de selección
Ensayos aleatorizados y aleatorizados por conglomerados que compararon la administración de suplementos probióticos con placebo o dieta para la prevención del desarrollo de la DMG. Se consideraron elegibles para inclusión los ensayos aleatorizados por conglomerados, pero no se identificaron. Los estudios con diseños cuasialeatorizados y cruzados (crossover) no fueron elegibles para inclusión en esta revisión. Los estudios presentados solo como resúmenes sin un informe completo posterior de los resultados del estudio solo se incluyeron si los autores confirmaron que los datos del resumen procedían del análisis final. En caso contrario, el resumen quedó a la espera de ser clasificado.
Obtención y análisis de los datos
Dos autores de la revisión de forma independiente evaluaron la elegibilidad de los estudios, extrajeron los datos y evaluaron el riesgo de sesgo de los estudios incluidos. Se verificó la exactitud de los datos.
Resultados principales
En esta actualización se incluyeron siete ensayos con 1647 participantes. Dos estudios se realizaron en mujeres con sobrepeso y obesidad, dos en mujeres con obesidad y tres no excluyeron a las mujeres en función de su peso. Todos los estudios incluidos compararon probióticos con placebo. Los estudios incluidos tuvieron bajo riesgo de sesgo en general, excepto un estudio en el que el riesgo de sesgo fue poco claro. Se excluyeron dos estudios, ocho estudios estaban en curso y tres estudios están pendientes de clasificación.
Seis estudios incluidos, con 1440 participantes, evaluaron el riesgo de DMG. No está claro si los probióticos tienen algún efecto sobre el riesgo de DMG en comparación con placebo (razón de riesgos [RR] media 0,80; intervalo de confianza [IC] del 95%: 0,54 a 1,20; seis estudios, 1440 mujeres; evidencia de certeza baja). La certeza de evidencia la fue baja debido a la heterogeneidad sustancial y a los IC amplios que incluyeron efectos beneficiosos y perjudiciales apreciables.
Los probióticos aumentan el riesgo de preeclampsia en comparación con placebo (RR 1,85; IC del 95%: 1,04 a 3,29; cuatro estudios, 955 mujeres; evidencia de certeza alta) y podrían aumentar el riesgo de trastornos hipertensivos del embarazo (RR 1,39; IC del 95%: 0,96 a 2,01; cuatro estudios, 955 mujeres), aunque los IC para los trastornos hipertensivos del embarazo también indicaron que los probióticos podrían no tener efecto.
Hubo pocas diferencias entre los grupos en otros desenlaces principales. Los probióticos dan lugar a poca o ninguna diferencia en el riesgo de cesárea (RR 1,00; IC del 95%: 0,86 a 1,17; seis estudios, 1520 mujeres; evidencia de certeza alta), y probablemente dan lugar a poca o ninguna diferencia en el riesgo materno de aumento de peso durante el embarazo (DM 0,30 kg; IC del 95%: ‐0,67 a 1,26; cuatro estudios, 853 mujeres; evidencia de certeza moderada). Es probable que los probióticos den lugar a poca o ninguna diferencia en la incidencia de lactantes grandes para la edad gestacional (RR 0,99; IC del 95%: 0,72 a 1,36; cuatro estudios, 919 lactantes; evidencia de certeza moderada) y podrían dar lugar a poca o ninguna diferencia en la adiposidad neonatal (dos estudios, 320 lactantes; datos no agrupados; evidencia de certeza baja). Un estudio informó de la adiposidad como masa grasa (DM ‐0,04 kg; IC del 95%: ‐0,12 a 0,04), y un estudio informó de la adiposidad como porcentaje de grasa (DM ‐0,10%; IC del 95%: ‐1,19 a 0,99). No se sabe acerca del efecto de los probióticos sobre la mortalidad perinatal (RR 0,33; IC del 95%: 0,01 a 8,02; tres estudios, 709 lactantes; evidencia de certeza baja), una medida compuesta de morbilidad neonatal (RR 0,69; IC del 95%: 0,36 a 1,35; dos estudios, 623 lactantes; evidencia de certeza baja) o la hipoglucemia neonatal (RR media 1,15; IC del 95%: 0,69 a 1,92; dos estudios, 586 lactantes; evidencia de certeza baja). Ningún estudio incluido informó sobre el trauma perineal, la depresión posnatal, el desarrollo de diabetes en la madre y el lactante o la discapacidad neurosensorial.
Conclusiones de los autores
La evidencia de certeza baja de seis ensayos no ha identificado claramente el efecto de los probióticos sobre el riesgo de DMG. Sin embargo, evidencia de certeza alta indica que existe un aumento en el riesgo de preeclampsia con la administración de probióticos. No hubo otras diferencias claras entre los probióticos y el placebo entre los otros desenlaces principales. La certeza de la evidencia para los desenlaces principales de esta revisión varió de baja a alta; la calidad se disminuyó debido a preocupaciones por la heterogeneidad sustancial entre los estudios, los IC amplios y las bajas tasas de eventos.
Debido al riesgo de efectos perjudiciales y el escaso efecto beneficioso observado, se recomienda precaución con la administración de probióticos durante el embarazo.
El aparente efecto de los probióticos sobre la preeclampsia merece una consideración especial. Actualmente hay ocho estudios en curso y se recomienda que estos estudios tengan especial cuidado en el seguimiento y el examen del efecto sobre la preeclampsia y los trastornos hipertensivos del embarazo. Además, se debe considerar la posible fisiología subyacente de la relación entre los probióticos y el riesgo de preeclampsia.
PICO
Resumen en términos sencillos
Probióticos para prevenir la diabetes mellitus gestacional
Se analizó la evidencia de los ensayos controlados aleatorizados (estudios clínicos en los que las personas se colocan al azar en uno de dos o más grupos de tratamiento) que investigaron los suplementos probióticos solos o en combinación con intervenciones farmacológicas o no farmacológicas para prevenir la diabetes mellitus gestacional (DMG).
¿Cuál es el problema?
La DMG es una enfermedad en la que la madre desarrolla niveles elevados de azúcar en sangre, habitualmente después de las 13 semanas de embarazo. La DMG se diferencia de la diabetes tipo 2 en que los niveles de azúcar en sangre son normales antes del embarazo, y suelen volver a la normalidad después del mismo. La DMG se asocia a un mayor riesgo de desarrollar diabetes tipo 2 en etapas posteriores de la vida. Las mujeres con DMG tienen un mayor riesgo de padecer presión arterial elevada con presencia de proteínas en la orina (preeclampsia) y parto instrumental o cesárea. Sus recién nacidos tienen más probabilidades de ser grandes para su edad gestacional. Los probióticos son "bacterias buenas" que suelen tomarse en forma de cápsulas o bebidas para aumentar las bacterias intestinales. Las personas dependen de las bacterias intestinales para que les ayuden a digerir los alimentos, produzcan ciertas vitaminas, regulen el sistema inmunitario y las mantengan sanas al protegerlas de las bacterias patógenas. Los probióticos podrían cambiar el metabolismo de la persona y desempeñar un papel en la prevención de la DMG.
¿Por qué es esto importante?
Las mujeres con sobrepeso u obesidad que hayan presentado DMG en un embarazo anterior o que tengan un familiar directo con diabetes tienen un mayor riesgo de padecerla. El tratamiento actual de la DMG incluye dieta con o sin medicación, pero no siempre evita los problemas asociados con la DMG. Los probióticos podrían ser un método sencillo para prevenir la DMG. Esta revisión analizó si existe evidencia que demuestre que lo anterior es cierto.
¿Qué evidencia se encontró?
En marzo de 2020 se buscó evidencia de ensayos controlados aleatorizados y se identificaron siete estudios con 1647 mujeres embarazadas que compararon probióticos con placebo inactivo (tratamiento simulado). Dos estudios se realizaron en mujeres con sobrepeso y obesidad, dos en mujeres con obesidad y tres no excluyeron a las mujeres en función de su peso. El riesgo general de sesgo fue bajo, excepto en un estudio en el que el riesgo de sesgo fue incierto.
No está claro cómo los probióticos afectan al riesgo de desarrollar DMG debido a la gran variación en los resultados de seis estudios (1440 mujeres, evidencia de calidad baja). Los probióticos aumentan el riesgo de desarrollar preeclampsia (cuatro estudios, 955 mujeres; evidencia de calidad alta). Los probióticos dan lugar a poca o ninguna diferencia en el riesgo de necesitar una cesárea (seis estudios, 1520 mujeres; evidencia de calidad alta), y probablemente dan lugar a poca o ninguna diferencia en el aumento de peso durante el embarazo (cuatro estudios, 853 mujeres; evidencia de calidad moderada) o en el riesgo de dar a luz a un recién nacido grande (cuatro estudios, 919 mujeres; evidencia de calidad moderada). Ninguno de los estudios aportó información sobre el riesgo de traumatismo perineal (desgarros durante el parto vaginal o una incisión quirúrgica [episiotomía]), depresión posnatal o desarrollo de diabetes posterior.
No se sabe si los probióticos afectan que el lactante tenga problemas médicos después del parto debido a la variación de los resultados entre los estudios (dos estudios, 623 lactantes; evidencia de calidad baja). Tampoco se sabe con certeza cómo los probióticos afectan la mortalidad infantil (antes del parto o cuando son recién nacidos) (tres estudios, 709 recién nacidos; evidencia de certeza baja), el bajo nivel de azúcar en sangre (dos estudios, 586 recién nacidos; evidencia de certeza baja) o la grasa corporal (dos estudios, 320 niños; evidencia de certeza baja). Ninguno de los estudios aportó información sobre el riesgo de que los recién nacidos desarrollen diabetes o afecciones a largo plazo que afecten al desarrollo del cerebro.
¿Qué significa esto?
La evidencia de calidad baja de seis ensayos no ha identificado claramente el efecto de los probióticos sobre el riesgo de DMG. Sin embargo, evidencia de alta calidad indica que los probióticos probablemente aumentan el riesgo de preeclampsia. Por lo tanto, actualmente hay evidencia de posibles efectos perjudiciales con pocos efectos beneficiosos observados del uso generalizado de los probióticos en el embarazo.
Actualmente hay ocho estudios en curso que podrían contribuir a aclarar los efectos de los probióticos. También es importante explorar más la relación entre los probióticos y la preeclampsia.
Authors' conclusions
Summary of findings
| Probiotics compared to placebo for preventing gestational diabetes (maternal outcomes) | ||||||
| Patient or population: preventing gestational diabetes Setting: Australia, Finland, Iran, Ireland, and New Zealand Intervention: probiotics Comparison: placebo | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with Probiotics | |||||
| Diagnosis of gestational diabetes mellitus | Study population | Mean RR 0.80 | 1440 | ⊕⊕⊝⊝ | — | |
| 191 per 1000 | 153 per 1000 | |||||
| Hypertensive disorders of pregnancy (pre‐eclampsia) | Study population | RR 1.85 | 955 | ⊕⊕⊕⊕ | — | |
| 35 per 1000 | 65 per 1000 | |||||
| Caesarean section | Study population | RR 1.00 | 1520 | ⊕⊕⊕⊕ | — | |
| 285 per 1000 | 285 per 1000 | |||||
| Perineal trauma | — | — | — | — | — | Not reported |
| Weight gain during pregnancy | The mean weight gain during pregnancy was 9.4–14.8 kg | MD 0.30 kg higher | — | 853 | ⊕⊕⊕⊝ | — |
| Postnatal depression | — | — | — | — | — | Not reported |
| Development of subsequent diabetes | — | — | — | — | — | Not reported |
| *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; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio. | ||||||
| GRADE Working Group grades of evidence | ||||||
| aDowngraded one level for serious concerns about inconsistency due to substantial unexplained heterogeneity between studies. | ||||||
| Probiotics compared to placebo for preventing gestational diabetes (infant outcomes) | ||||||
| Patient or population: preventing gestational diabetes Setting: Australia, Finland, Iran, Ireland, and New Zealand Intervention: probiotics Comparison: placebo | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with probiotics | |||||
| Large‐for‐gestational age | Study population | RR 0.99 | 919 | ⊕⊕⊕⊝ | — | |
| 142 per 1000 | 141 per 1000 | |||||
| Perinatal mortality (stillbirth and neonatal mortality) | Study population | RR 0.33 | 709 | ⊕⊕⊝⊝ | — | |
| 3 per 1000 | 1 per 1000 | |||||
| Mortality or morbidity composite | Study population | RR 0.69 | 623 | ⊕⊕⊝⊝ | — | |
| 61 per 1000 | 42 per 1000 | |||||
| Hypoglycaemia as defined by trialists | Study population | Mean RR 1.15 | 586 | ⊕⊕⊝⊝ | — | |
| 135 per 1000 | 155 per 1000 | |||||
| Adiposity | The included studies showed no appreciable difference in fat mass or % fat between groups. 1 study reported adiposity as fat mass (MD –0.04 kg, 95% CI –0.12 to 0.04) 1 study reported adiposity as % fat (MD –0.10%, 95% CI –1.19 to 0.99) | — | 320 | ⊕⊕⊝⊝ | — | |
| Diabetes | — | — | — | — | — | Not reported |
| Neurodisability | — | — | — | — | — | Not reported |
| *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; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio. | ||||||
| GRADE Working Group grades of evidence | ||||||
| aDowngraded one level for serious concerns about imprecision due to a wide confidence intervals that included both appreciable benefit and harm. | ||||||
Background
Description of the condition
According to the American Diabetes Association (ADA), gestational diabetes mellitus (GDM) is diabetes in pregnancy that is diagnosed during the second or third trimester and was not clearly present prior to pregnancy (ADA 2019). There are multiple sets of diagnostic criteria that are used worldwide, which has caused estimates of prevalence to vary greatly (Buchanan 2012). According to the International Association of Diabetes in Pregnancy Study Group (IADPSG) criteria, approximately 14% of pregnancies worldwide were affected by GDM in 2017 (Cho 2018), and multiple studies have observed that the incidence is rising (Noh 2021; López‐de‐Andrés 2020; Abouzeid 2014; Dabelea 2005).
GDM is associated with a number of maternal and fetal adverse outcomes, and the risk of these outcomes increases with higher fasting plasma glucose levels (HAPO 2008). Women with GDM have higher rates of pre‐eclampsia and need for a caesarean section, and their infants have higher rates of macrosomia, shoulder dystocia, neonatal hypoglycaemia and respiratory distress syndrome (Carr 2011; Dodd 2007; Esakoff 2009; HAPO 2008). In addition, there is an increased risk for metabolic dysfunction for both mother and infant in the long term including diabetes, obesity and metabolic syndrome (Malcolm 2012). Large randomised controlled trials have demonstrated the benefits of treating GDM for preventing many of the associated adverse outcomes (Crowther 2005; Landon 2009), but it is not known if treatment prevents the long‐term adverse effects. In addition, there are substantial costs associated with the treatment of GDM, and cost‐effectiveness has not been clearly demonstrated (Fitria 2019). Therefore, prevention of GDM is favourable.
Prevention efforts have primarily focused on lifestyle interventions such as diet and exercise. The Cochrane Review evaluating the combination of diet and exercise interventions for the prevention of GDM concluded that there is moderate‐quality evidence that this combination of lifestyle interventions can reduce the risk of GDM (Shepherd 2017). However, the Cochrane Reviews evaluating diet and exercise for GDM prevention independently were inconclusive and suggested the need for higher‐quality evidence (Han 2012; Tieu 2017). All these reviews reported that studies involving these interventions were difficult to interpret given heterogeneity and small sample sizes. In addition, there was concern about adherence to these interventions on a population level (Sui 2013).
Due to these concerns, dietary supplements such as probiotics and myo‐inositol are being studied. The Cochrane Review evaluating myo‐inositol use for the prevention of GDM concluded there may be a reduction in GDM risk with its use, but the review authors ultimately suggested the need for further research due to low‐quality evidence (Crawford 2015). One overview of Cochrane Reviews for the prevention of gestational diabetes looked at all these interventions, and they found that no studied intervention resulted in clear benefit or harm, and many of these interventions did not have enough high‐quality evidence to determine an effect (Griffith 2020).
Description of the intervention
According to the World Health Organization, probiotics are defined as "live microorganisms which when administered in adequate amounts confer a health benefit on the host" (FAO/WHO 2006). The health effects provided by probiotics vary depending on the specific species and strain of probiotic used, and, therefore, have been investigated in a wide variety of health conditions. Among the most common probiotics used are members of the genera Lactobacillus, Bifidobacterium and Enterococcus, but products differ greatly in the strains and concentrations used (Syngai 2016; FAO/WHO 2006). Probiotics are available in a variety of food products, such as yoghurt or fermented milks, or as dietary supplements that can be purchased as capsules without a prescription.
How the intervention might work
Studies of the human microbiome have revealed a complex relationship between the microbiome and an individual's overall health and wellbeing. The microbiome is altered by a variety of factors including diet and various health conditions (David 2014), and in turn, the microbiome may influence host metabolism and contribute to the development of obesity and diabetes (Musso 2011). Many studies of the gut microbiome in obese people have revealed an increase in the proportion of bacteria in the Firmicutes phylum and a decrease in bacteria belonging to the Bacteroidetes phylum (John 2016). Similarly, studies in people with type 2 diabetes and glucose intolerance have revealed a reduction in Akkermansia muciniphila and butyrate‐producing bacteria such as Lactobacillus and Bifidobacterium in their microbiome (Brunkwall 2017). These changes in the gut microbiome may be linked to obesity and diabetes through the role bacteria play in host glucose and lipid metabolism (Musso 2011).
Increases in body fat and decreases in insulin sensitivity are normal changes in pregnancy, and these changes appear to be linked to changes in the microbiome as well. Koren 2012 found that the gut microbiome became less diverse as pregnancy progressed, and the microbiome in the third trimester was associated with increased adiposity and insulin insensitivity when transplanted into mice (Koren 2012). In women with GDM, insulin sensitivity is impaired beyond normal levels, leading to hyperglycaemia. Crusell 2018 have looked at the microbiome in women with GDM, and found that the microbiome in GDM differed slightly from that in normal pregnancy and may have resembled that of non‐pregnant women with type 2 diabetes (Crusell 2018).
Given the role the microbiome plays in host glucose and lipid metabolism (Musso 2011), probiotics have been suggested as an intervention for improving glycaemic control in diabetes by helping to restore balance among species of bacteria in the microbiome (Tiderencel 2020). Many randomised controlled trials have examined the use of probiotics in people with type 2 diabetes, and one meta‐analysis of these trials revealed that probiotics were helpful in improving glycaemic control and may have improved glucose metabolism (Tiderencel 2020). GDM is like type 2 diabetes in that there are similar changes in insulin resistance and possibly in the microbiome (Crusell 2018), and, therefore, probiotics may have similar effects for prevention or treatment of GDM. A Cochrane Review evaluating the use of probiotics to treat GDM was published in 2020 (Okesene‐Gafa 2020).
Why it is important to do this review
The incidence of GDM is increasing (Noh 2021; López‐de‐Andrés 2020; Abouzeid 2014; Dabelea 2005), and GDM is associated with significant health implications for both mother and infant (HAPO 2008). Therefore, prevention of GDM is ideal. One study evaluating the use of probiotics in pregnancy suggested that probiotics may reduce the incidence of GDM (Laitinen 2009). Since this initial study, other studies have tried to address this question with mixed results (Asgharian 2020; Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). Therefore, a systematic review is necessary to synthesise the available evidence for or against the use of probiotics for preventing GDM in pregnancy.
Objectives
To systematically assess the effects of probiotic supplements used either alone or in combination with pharmacological and non‐pharmacological interventions on the prevention of GDM.
Methods
Criteria for considering studies for this review
Types of studies
We included randomised and cluster‐randomised trials; however, we found no cluster‐randomised trials. Quasi‐randomised and cross‐over design studies were not eligible for inclusion in this review. Studies presented only as abstracts with no subsequent full report of study results were only included if we received confirmation from the study authors that the data in the abstract were final. Otherwise, they were listed as awaiting classification.
Types of participants
Studies that included pregnant women not previously diagnosed with diabetes mellitus. Studies of women with GDM in a previous pregnancy but no evidence of diabetes mellitus or GDM in the current pregnancy before entering the trial were eligible for inclusion.
Types of interventions
Probiotic supplementation for prevention of GDM, either alone or in combination with pharmacological (e.g. metformin) or non‐pharmacological (e.g. diet/lifestyle) interventions.
Probiotic supplementation (administered by any method) should have been commenced prior to the diagnosis of GDM and continued for any duration.
Comparison interventions of placebo or diet were eligible.
Trials may have used other interventions in a comparison arm or in combination with the probiotic. These other interventions may have included pharmaceutical probiotic supplements as well as food items supplemented with probiotics.
Types of outcome measures
The outcomes for this review are from the Cochrane Core Outcome Set for GDM prevention.
Primary outcomes
Maternal
-
Diagnosis of GDM
-
Hypertensive disorders of pregnancy (including pre‐eclampsia, pregnancy‐induced hypertension and eclampsia)
-
Caesarean section
Infant
-
Large‐for‐gestational age
-
Perinatal mortality (including stillbirth and neonatal mortality)
-
Mortality or morbidity composite
Secondary outcomes
Maternal
-
Induction of labour
-
Perineal trauma
-
Placental abruption
-
Postpartum haemorrhage
-
Postpartum infection
-
Weight gain during pregnancy
-
Adherence to the intervention
-
Behaviour changes associated with the intervention
-
Relevant biomarker changes associated with the intervention (including adiponectin, free fatty acids, triglycerides, high‐density lipoproteins, low‐density lipoproteins, insulin, etc.)
-
Sense of wellbeing and quality of life
-
Views of the intervention
-
Breastfeeding
Long‐term maternal
-
Postnatal depression
-
Postnatal weight retention or return to prepregnancy weight
-
Body mass index (BMI)
-
GDM in a subsequent pregnancy
-
Type 1 diabetes
-
Type 2 diabetes
-
Impaired glucose tolerance
-
Cardiovascular health as defined by trialists (including blood pressure, hypertension, cardiovascular disease and metabolic syndrome)
Infant
-
Stillbirth
-
Neonatal mortality
-
Gestational age at birth
-
Preterm birth (less than 37 weeks' gestation and less than 32 weeks' gestation)
-
Apgar score (less than seven at five minutes)
-
Macrosomia
-
Small‐for‐gestational age (SGA)
-
Birthweight and z‐score
-
Head circumference and z‐score
-
Length and z‐score
-
Ponderal index
-
Adiposity
-
Shoulder dystocia
-
Bone fracture
-
Nerve palsy
-
Respiratory distress syndrome
-
Hypoglycaemia as defined by trialists
-
Hyperbilirubinaemia
Later infant and childhood
-
Weight and z‐scores
-
Height and z‐scores
-
Head circumference and z‐scores
-
Adiposity (including BMI and skinfold thickness)
-
Blood pressure
-
Type 1 diabetes
-
Type 2 diabetes
-
Impaired glucose tolerance
-
Dyslipidaemia or metabolic syndrome
-
Neurodisability
-
Educational achievement
Child as an adult
-
Weight
-
Height
-
Adiposity (including BMI and skinfold thickness)
-
Cardiovascular health as defined by trialists (including blood pressure, hypertension, cardiovascular disease and metabolic syndrome)
-
Type 1 diabetes
-
Type 2 diabetes
-
Impaired glucose tolerance
-
Dyslipidaemia or metabolic syndrome
-
Employment, education and social status/achievement
Health service use
-
Number of hospital or health professional visits (including midwife, obstetrician, physician, dietician and diabetic nurse)
-
Number of antenatal visits or admissions
-
Length of antenatal stay
-
Neonatal intensive care unit admission
-
Length of postnatal stay (mother)
-
Length of postnatal stay (baby)
-
Costs to families associated with the management provided
-
Costs associated with the intervention
-
Cost of maternal care
-
Cost of offspring care
Search methods for identification of studies
The following methods section of this review is based on a standard template used by Cochrane Pregnancy and Childbirth.
Electronic searches
For this update, we searched Cochrane Pregnancy and Childbirth's Trials Register by contacting their Information Specialist (20 March 2020).
The Register is a database containing over 25,000 reports of controlled trials in the field of pregnancy and childbirth. It represents over 30 years of searching. For full current search methods used to populate Pregnancy and Childbirth's Trials Register including the detailed search strategies for CENTRAL, MEDLINE, Embase and CINAHL; the list of handsearched journals and conference proceedings; and the list of journals reviewed via the current awareness service, see the Cochrane Pregnancy and Childbirth's Trials Register (pregnancy.cochrane.org/pregnancy-and-childbirth-groups-trials-register).
Briefly, Cochrane Pregnancy and Childbirth's Trials Register is maintained by their Information Specialist and contains trials identified from:
-
monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
-
weekly searches of MEDLINE (Ovid);
-
weekly searches of Embase (Ovid);
-
monthly searches of CINAHL (EBSCO);
-
handsearches of 30 journals and the proceedings of major conferences;
-
weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.
Two people screen search results and review the full text of all relevant trial reports identified through the searching activities. Based on the intervention described, each trial report is assigned a number that corresponds to a specific Pregnancy and Childbirth review topic (or topics), and is then added to the Register. The Information Specialist searches the Register for each review using this topic number rather than keywords. This results in a more specific search set that has been fully accounted for in the relevant review sections (Included studies; Excluded studies; Studies awaiting classification; Ongoing studies).
In addition, we searched ClinicalTrials.gov and the World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch/) for unpublished, planned and ongoing trial reports (20 March 2020) using the search methods detailed in Appendix 1.
Searching other resources
We searched the reference lists of all retrieved studies.
We applied no language or date restrictions.
Data collection and analysis
For methods used in the previous version of this review, see Barrett 2014.
For this update, we used the following methods for assessing the 43 reports that were identified as a result of the updated search.
The following methods section is based on a standard template used by Cochrane Pregnancy and Childbirth.
Selection of studies
Two review authors (SJD and MDN) independently assessed for inclusion all the potential studies we identified as a result of the search strategy. We resolved any disagreement through discussion or, if required, we consulted another review author (LC). Since three review authors were also authors on one of the identified studies (Callaway 2019), the two review authors not involved with this study assessed it for inclusion (SJD and SAP).
We created a study flow diagram to map out the number of records identified, included, excluded or awaiting classification (Figure 1).
Study flow diagram.
Screening eligible studies for scientific integrity/trustworthiness
Two review authors evaluated all studies meeting our inclusion criteria against predefined criteria to select studies that, based on available information, were deemed to be sufficiently trustworthy to be included in the analysis. The criteria are as follows.
Research governance
-
No prospective trial registration for studies published after 2010 without plausible explanation.
-
When requested, trial authors refuse to provide/share the protocol or ethics approval letter (or both).
-
Trial authors refuse to engage in communication with the Cochrane Review authors.
-
Trial authors refuse to provide individual participant data (IPD) data upon request with no justifiable reason.
Baseline characteristics
-
Characteristics of the study participants being too similar (distribution of mean (standard deviation (SD)) excessively narrow or excessively wide, as noted by Carlisle 2017.
Feasibility
-
Implausible numbers (e.g. 500 women with severe cholestasis of pregnancy recruited in 12 months).
-
(Close to) zero losses to follow‐up without plausible explanation.
Results
-
Implausible results (e.g. massive risk reduction for main outcomes with small sample size).
-
Unexpectedly even numbers of women 'randomised' including a mismatch between the numbers and the methods (e.g. if they say no blocking was used but still end up with equal numbers, or they say they used blocks of four but the final numbers differ by six).
Studies assessed as being potentially 'high risk' were not included in the review. Where a study was classified as 'high risk' for one or more of the above criteria, we attempted to contact the study authors to address any possible lack of information/concerns. If adequate information remained unavailable, the study remained in 'awaiting classification' and the reasons and communications with the author (or lack of) described in detail.
The process is described fully in Figure 3.
Abstracts
Data from abstracts were only included if, in addition to the trustworthiness assessment, the study authors confirmed in writing that the data to be included in the review had come from the final analysis and will not change. If such information was not available/provided, the study remained 'awaiting classification' (as above).
Data extraction and management
We designed a form to extract data. For eligible studies, at least two review authors (SJD and MDN for most studies, SJD and SAP for Callaway 2019) extracted data using the agreed form. We resolved discrepancies through discussion, or, if required, through consultation with a third review author. We entered data into Review Manager 5 and checked for accuracy (Review Manager 2014). When information regarding any of the above was unclear, we attempted to contact authors of the original reports to request further details.
Assessment of risk of bias in included studies
Two review authors (SJD and MDN) independently assessed risk of bias for the included studies using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved any disagreement by discussion or by involving a third review author. Different review authors (SJD and SAP) independently assessed risk of bias for Callaway 2019 to limit the effect of conflict of interest.
1. Random sequence generation (checking for possible selection bias)
We described the method used to generate the allocation sequence in sufficient detail to assess whether it produced comparable groups.
We assessed the method as:
-
low risk of bias (any truly random process, e.g. random number table; computer random number generator);
-
high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);
-
unclear risk of bias.
2. Allocation concealment (checking for possible selection bias)
We described the methods used to conceal allocation to interventions prior to assignment and assessed whether the intervention allocation could have been foreseen in advance of, or during, recruitment or changed after assignment.
We assessed the methods as:
-
low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
-
high risk of bias (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);
-
unclear risk of bias.
3.1. Blinding of participants and personnel (checking for possible performance bias)
We described the methods used to conceal the allocation sequence and determine whether intervention allocation could have been foreseen in advance of, or during, recruitment or changed after assignment.
We assessed the methods as:
-
low, high or unclear risk of bias for participants;
-
low, high or unclear risk of bias for personnel.
3.2. Blinding of outcome assessment (checking for possible detection bias)
We described the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We considered that studies would be at low risk of bias if they were blinded, or if we judged that the lack of blinding would be unlikely to affect results. We assessed blinding separately for different outcomes or classes of outcomes.
We assessed the methods as:
-
low, high or unclear risk of bias.
4. Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)
We described the completeness of data including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported, the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes. Where sufficient information was reported, or could be supplied by the trial authors, we re‐included missing data in the analyses that we undertook.
We assessed methods as:
-
low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);
-
high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; 'as treated' analysis done with substantial departure of intervention received from that assigned at randomisation);
-
unclear risk of bias.
5. Selective reporting (checking for reporting bias)
We described how we investigated the possibility of selective outcome reporting bias and what we found.
We assessed the methods as:
-
low risk of bias (where it was clear that all the study's prespecified outcomes and all expected outcomes of interest to the review were reported);
-
high risk of bias (where not all the study's prespecified outcomes were reported; one or more reported primary outcomes were not prespecified; outcomes of interest were reported incompletely and so could not be used; study failed to include results of a key outcome that would have been expected to have been reported);
-
unclear risk of bias.
6. Other bias (checking for bias due to problems not covered by 1. to 5. above)
We described any important concerns we had about other possible sources of bias.
We assessed whether the included study was free of other problems that could have put it at risk of bias:
-
low risk of other bias;
-
high risk of other bias;
-
unclear whether there is risk of other bias.
7. Overall risk of bias
We made explicit judgements about whether the included study was at high risk of bias, according to the criteria given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). With reference to 1. to 6. above, we assessed the likely magnitude and direction of the bias and whether we considered it was likely to impact on the findings. We planned to explore the impact of the level of bias through undertaking sensitivity analyses (see Sensitivity analysis).
Measures of treatment effect
Dichotomous data
For dichotomous data, we presented results as summary risk ratio (RR) with 95% confidence intervals (CI).
Continuous data
For continuous data, we used the mean difference (MD) with 95% CIs if outcomes were measured in the same way between trials. We planned to use the standardised mean difference with 95% CIs to combine trials that measured the same outcome, but used different scales.
Unit of analysis issues
Cluster‐randomised trials
We identified no cluster‐randomised trials for inclusion. However, if we identify cluster‐randomised trials in updates of this review, we will include them in the analyses along with individually randomised trials. We will adjust their effect measure using the methods described in the Cochrane Handbook for Systematic Reviews of Interventions using an estimate of the intracluster correlation coefficient (ICC) derived from the trial (if possible), from a similar trial or from a study of a similar population. Where the cluster‐randomised trial properly accounts for the cluster design, we will extract an estimate of the effect measure directly. Where the cluster‐randomised trial does not properly account for the clustering, we will calculate the effective sample size of the intervention and placebo groups by dividing the sample size by the design effect. The design effect is 1 + (m – 1) × ICC where m is the mean cluster size. We will assess the cluster‐randomised trials and the calculation of the effective sample size will be performed with the assistance of a statistician. If we use ICCs from other sources, we will report this and conduct sensitivity analyses to investigate the effect of variation in the ICC. If we identify both cluster‐randomised trials and individually randomised trials, we plan to synthesise the relevant information. We will consider it reasonable to combine the results from both if there is little heterogeneity between the study designs and the interaction between the effect of intervention and the choice of randomisation unit is unlikely.
We will also acknowledge heterogeneity in the randomisation unit and perform a subgroup analysis to investigate the effects of the randomisation unit.
Studies with more than two intervention groups
In studies with more than two groups, only the two groups that best fit the comparisons used in this review were chosen. When there were more than two groups due to a secondary intervention, the groups without the secondary intervention were chosen if possible to minimise the effect of the secondary intervention on the comparison. If this was not possible, the groups were chosen so that the secondary intervention was the same in both groups. For 2×2 factorial trials, groups were combined where appropriate given the participants were independently randomised to the intervention of interest.
Dealing with missing data
For included studies, we noted levels of attrition. We explored the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analysis.
For all outcomes, we carried out analyses, as far as possible, on an intention‐to‐treat basis (i.e. we attempted to include all participants randomised to each group in the analyses, and analysed all participants in the group to which they were allocated, regardless of whether or not they received the allocated intervention). The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes were known to be missing.
Assessment of heterogeneity
We assessed statistical heterogeneity in each meta‐analysis using the I² and Chi² statistics. For random‐effects meta‐analyses we also considered the Tau² statistic. We regarded heterogeneity as substantial if the I² statistic was greater than 30% and either the Tau² was greater than zero, or there was a low P value (less than 0.10) in the Chi² test for heterogeneity.
Assessment of reporting biases
Given we included only seven studies, reporting bias analysis was not undertaken. In updates of this review, if there are 10 or more studies in the meta‐analysis, we will investigate reporting biases (such as publication bias) using funnel plots. We will assess funnel plot asymmetry visually. If asymmetry is suggested by a visual assessment, we will perform exploratory analyses to investigate it.
Data synthesis
We carried out statistical analysis using the Review Manager 5 (Review Manager 2014). We used fixed‐effect meta‐analysis for combining data where it was reasonable to assume that studies were estimating the same underlying treatment effect (i.e. where trials examined the same intervention, and the trials' populations and methods were judged sufficiently similar). If there was clinical heterogeneity sufficient to expect that the underlying treatment effects differed between trials, or if there was substantial statistical heterogeneity, we used random‐effects meta‐analysis to produce an overall summary, if a mean treatment effect across trials was considered clinically meaningful. The random‐effects summary was treated as the mean range of possible treatment effects and we discussed the clinical implications of treatment effects differing between trials. If the mean treatment effect was not clinically meaningful, we did not combine trials.
If we used random‐effects analyses, we presented the results as the mean treatment effect with 95% CIs, and the estimates of the Tau² and I² statistics.
Subgroup analysis and investigation of heterogeneity
Where we identified substantial heterogeneity, we investigated it using subgroup analyses and sensitivity analyses. We considered whether an overall summary was meaningful, and if it was, used random‐effects analysis to produce it.
We planned to carry out the following subgroup analyses.
-
History of GDM or family history of type 2 diabetes (yes versus no).
-
Probiotic dose (more than five billion colony‐forming units (CFU) versus less than five billion CFU).
-
Probiotic bacterial species (one species versus another species).
-
Probiotic treatment starting in early pregnancy versus starting at more than 20 weeks' gestation.
-
Probiotic mode of delivery (capsule versus other).
-
Probiotic frequency of administration (daily versus other).
In this update of the review, subgroup analysis by history of GDM or family history of type 2 diabetes was not conducted as outcome data were not available in these subgroups. The subgroup analysis by probiotic mode of delivery and frequency of administration were also not conducted since all included studies administered the intervention daily as a capsule. These subgroups will be included in future updates of the review if possible.
Subgroup analysis was restricted to the review's primary outcomes. We assessed subgroup differences by interaction tests available within Review Manager 5 (Review Manager 2014). We reported the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² statistic.
Sensitivity analysis
Sensitivity analysis was carried out, where necessary, to explore the influence of diagnostic criteria for GDM. Sensitivity analysis was restricted to the review's primary outcomes.
We planned to conduct a sensitivity analysis to explore the influence of high dropout rates (more than 20%); however, we identified no such studies. This may be possible in updates of this review.
Summary of findings and assessment of the certainty of the evidence
For this update, we used the GRADE approach to assess the certainty of the evidence as outlined in the GRADE Handbook to assess the quality of the body of evidence relating to the following outcomes for the main comparison probiotics versus placebo. If studies comparing probiotics and diet are identified in future updates, we will evaluate this comparison.
Maternal
-
Diagnosis of GDM
-
Hypertensive disorders of pregnancy (pre‐eclampsia)
-
Caesarean section
-
Perineal trauma
-
Weight gain during pregnancy
-
Postnatal depression
-
Development of subsequent diabetes
Infant
-
Large‐for‐gestational age
-
Perinatal mortality
-
Mortality or morbidity composite
-
Hypoglycaemia as defined by trialists
-
Adiposity
-
Diabetes
-
Neurodisability
We used GRADEpro GDT to import data from Review Manager 5 (Review Manager 2014) in order to create 'Summary of findings' tables. We produced a summary of the intervention effect and a measure of certainty for each of the above outcomes was produced using the GRADE approach. The GRADE approach uses five considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the certainty of the body of evidence for each outcome. The evidence can be downgraded from 'high certainty' by one level for serious (or by two levels for very serious) limitations, depending on assessments for risk of bias, indirectness of evidence, serious inconsistency, imprecision of effect estimates or potential publication bias.
Results
Description of studies
See Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification; and Characteristics of ongoing studies tables.
Results of the search
See: Figure 1
We assessed 43 new trial reports from an updated search in March 2020. We also reassessed the four studies (five reports) that were ongoing in the previous version of the review. We included six new trials (25 reports), added four new reports to the previously included study (Laitinen 2009), and excluded two trials (six reports). Three studies (four reports) are awaiting further classification and we added eight studies (nine reports) to the Ongoing studies section.
Screening eligible studies for scientific integrity/trustworthiness
See: Figure 2
Applying the trustworthiness screening tool
One study is awaiting classification since it was only available in abstract form and confirmation that the presented data came from the final analysis was not received (Charles 2018). Another study is awaiting classification because it was unclear whether the study met our inclusion criteria; we sought clarification from the authors but received no response (Si 2019).
Two studies were at high risk according to the prespecified trustworthiness criteria. One study had almost no losses to follow‐up (Asgharian 2020), and there was insufficient information provided by the study authors for us to make a definitive classification. Therefore, this study remains in studies awaiting classification. The other study had no losses to follow‐up and had limited information regarding their randomisation methods (Jamilian 2016). However, the study authors provided more detail regarding these concerns, and the study was included since it was determined to be at low risk. See Characteristics of included studies; Characteristics of studies awaiting classification for further information.
Included studies
Design
All included studies were parallel randomised controlled trials. Four studies had one intervention arm and one control arm (Callaway 2019; Jamilian 2016; Lindsay 2014; Wickens 2017). Laitinen 2009 had three arms for two interventions. Pellonpera 2019 had four arms for two interventions, and Okesene‐Gafa 2019 was designed as a 2×2 factorial with two interventions.
Sample sizes
The number of women recruited in the included studies ranged from 60 (Jamilian 2016) to 438 women (Pellonpera 2019). The other studies recruited 175 (Lindsay 2014), 230 (Okesene‐Gafa 2019), 256 (Laitinen 2009), 423 (Wickens 2017), and 433 women (Callaway 2019).
Setting
The included studies in this review were conducted in Iran (Jamilian 2016), Australia (Callaway 2019), Finland (Laitinen 2009; Pellonpera 2019), Ireland (Lindsay 2014), and New Zealand (Okesene‐Gafa 2019; Wickens 2017).
Participants
All included studies were conducted in pregnant women with singleton pregnancies without pre‐existing diabetes or other significant health conditions, although two studies included women with a history of atopic disease (Laitinen 2009; Wickens 2017). Two studies were conducted in overweight and obese pregnant women (Callaway 2019; Pellonpera 2019), two in obese pregnant women only (Lindsay 2014; Okesene‐Gafa 2019), and three did not exclude women based on their body mass index (Jamilian 2016; Laitinen 2009; Wickens 2017).
Interventions and comparisons
All seven trials compared probiotics versus placebo. In four trials, women were only randomised to either probiotics or placebo (Callaway 2019; Jamilian 2016; Lindsay 2014; Wickens 2017). Three studies included a second intervention, two of which included a dietary intervention (Laitinen 2009; Okesene‐Gafa 2019), and one included a fish oil capsule (Pellonpera 2019). Laitinen 2009 randomised women to probiotics plus dietary intervention, placebo plus dietary intervention, or placebo plus routine dietary advice. Okesene‐Gafa 2019 first randomised all women to the dietary intervention or routine dietary advice, then randomised all women again to either probiotics or placebo. Pellonpera 2019 randomised women to one of four study arms (probiotics plus fish oil, probiotics plus placebo, placebo plus fish oil or placebo plus placebo). Although two trials included diet as a secondary intervention, the trials did not directly compare probiotics versus diet (Laitinen 2009; Okesene‐Gafa 2019). Therefore, no conclusions could be drawn about this comparison.
Six trials started the intervention prior to 20 weeks' gestation (Callaway 2019; Jamilian 2016; Laitinen 2009; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017), and one trial started the intervention at 20 weeks' gestation or later (Lindsay 2014). One study initially started the intervention before 16 weeks' gestation, but it was later changed to before 20 weeks' gestation due to a change in hospital policy (Callaway 2019). Women received the intervention daily in all studies.
All included studies delivered the intervention as a capsule. The dose of probiotic used in the studies varied, with three studies reporting a dose of less than five billion CFUs per species (Callaway 2019; Jamilian 2016; Lindsay 2014), and four studies reporting a dose of greater than five billion CFUs per species (Laitinen 2009; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). However, it is important to note that given decay in probiotics over time from the date of manufacturing, the dose can change. Therefore, studies reported either the minimum or mean dose, so individual participants in each study may have received doses that varied from the reported study dose.
Studies used a variety of different bacterial species and strains, and most used a combination of species for their probiotics. The species were Lactobacillus rhamnosus GG (Callaway 2019; Laitinen 2009; Okesene‐Gafa 2019), Lactobacillus rhamnosus HN001 (Pellonpera 2019; Wickens 2017), Lactobacillus acidophilus LA5 (Jamilian 2016), Lactobacillus casei (Jamilian 2016), Lactobacillus salivarius UCC118 (Lindsay 2014), Bifidobacterium animalis subspecies lactis BB12 (Callaway 2019; Laitinen 2009; Okesene‐Gafa 2019), Bifidobacterium animalis subspecies lactis 420 (Pellonpera 2019), and Bifidobacterium bifidum (Jamilian 2016).
Outcomes
Studies were required to have either the diagnosis of GDM or a marker of glucose metabolism in the third trimester of pregnancy as a reported outcome to be eligible for inclusion. Six studies reported the incidence of GDM (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017), while one reported laboratory measures of glucose metabolism such as fasting plasma glucose and insulin levels (Jamilian 2016). The studies that reported the incidence of GDM used several different diagnostic criteria, and some studies reported results according to more than one set of diagnostic criteria. Four studies used the IADPSG criteria (Callaway 2019; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017; criteria: IADPSG 2010); one study used the Carpenter and Coustan criteria (Lindsay 2014; criteria: Carpenter 1982), one used the modified Fourth International Workshop‐Conference on GDM criteria (Laitinen 2009; criteria: Metzger 1998), one used the local New Zealand criteria (Australasian Diabetes in Pregnancy Society) (Wickens 2017; criteria: Ministry of Health 2014), and one used the local Finnish criteria (Pellonpera 2019; criteria: The Finnish Medical Society Duodecim 2013). The details for each set of diagnostic criteria can be found in Table 1.
| Parameter | IADPSG (IADPSG 2010 ) | Carpenter and Coustan (Carpenter 1982 ) | Modified Fourth International Workshop‐Conference (Metzger 1998 ) | New Zealand Guidelines (Ministry of Health 2014 ) | Finnish Current Care Guidelines (The Finnish Medical Society Duodecim 2013 ) |
|---|---|---|---|---|---|
| OGTT (g) | 75 | 100 | 75 | 75 | 75 |
| Fasting (mmol/L) | 5.1 | 5.3 | 4.8 | 5.5 | 5.3 |
| 1 hour (mmol/L) | 10.0 | 10.0 | 10.0 | — | 10.0 |
| 2 hours (mmol/L) | 8.5 | 8.6 | 8.7 | 9.0 | 8.6 |
| 3 hours (mmol/L) | — | 7.8 | — | — | — |
| Elevated values required | 1 | 2 | 1 | 1 | 1 |
IADPSG: International Association of Diabetes and Pregnancy Study Groups; OGTT: oral glucose tolerance test.
At least one study reported the other primary outcomes in this review. Four studies reported hypertensive disorders of pregnancy (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019), six studies reported caesarean sections (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017), four studies reported large‐for‐gestational‐age infants (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019), three studies reported perinatal mortality including stillbirth and neonatal mortality (Callaway 2019; Laitinen 2009; Lindsay 2014), and two studies reported a neonatal mortality or morbidity composite measure (Callaway 2019; Okesene‐Gafa 2019).
Further details on the primary and secondary outcomes of each study can be found in the Characteristics of included studies table.
Dates of study
The studies were all conduced between 2002 and 2017, with the following trial dates: April 2002 to November 2005 (Laitinen 2009), March 2012 to March 2013 (Lindsay 2014), commencement in November 2012 with no end date provided (Callaway 2019), December 2012 to November 2014 (Wickens 2017), October 2013 to July 2017 (Pellonpera 2019), March 2015 to July 2015 (Jamilian 2016), and April 2015 to June 2017 (Okesene‐Gafa 2019).
Funding sources
Study authors reported the following sources of funding: National Health and Medical Research Council (Callaway 2019), the Royal Brisbane and Women's Hospital Foundation (Callaway 2019), Vice‐Chancellor for Research, AUMS, Iran (Jamilian 2016), Academy of Finland (Laitinen 2009; Pellonpera 2019), Sigrid‐Juselius Foundation (Laitinen 2009), Juho Vainio Foundation (Laitinen 2009; Pellonpera 2019), Social Insurance Institution of Finland (Laitinen 2009), Raisio (Laitinen 2009), Chr. Hansen A/S (Laitinen 2009; Okesene‐Gafa 2019), Valio Ltd (Laitinen 2009), National Maternity Hospital Medical Fund Ivo Drury Award (Lindsay 2014), Alimentary Health Ltd (Lindsay 2014), Counties Manukau Health (Okesene‐Gafa 2019), Cure Kids Grant (Okesene‐Gafa 2019), Lottery Health Research (Okesene‐Gafa 2019), RANZCOG Two Mercia Barnes Trust (Okesene‐Gafa 2019), Gravida National Centre for Growth and Development (Okesene‐Gafa 2019), University of Auckland Faculty Development Research Fund and Reinvestment Fund (Okesene‐Gafa 2019), Nurture Foundation (Okesene‐Gafa 2019), Heart Foundation of New Zealand (Okesene‐Gafa 2019), Roche Diagnostics International Ltd (Okesene‐Gafa 2019), Turku University Hospital Expert Responsibility Area (Pellonpera 2019), Diabetes Research Foundation (Pellonpera 2019), Business Finland (Pellonpera 2019), the Finnish Medical Foundation (Pellonpera 2019), the University of Turku (Pellonpera 2019), DuPont (Pellonpera 2019), Croda Europe Ltd (Pellonpera 2019), the Health Research Council of New Zealand (Wickens 2017), and Fonterra (Wickens 2017).
Declarations of interest
All included studies stated they had no declarations of interest.
Further details on each study can be found in the Characteristics of included studies table.
Excluded studies
We excluded two studies from this review since the intervention was not started until the third trimester of pregnancy, after GDM would have been diagnosed (Asemi 2013; Taghizadeh 2014). See Characteristics of excluded studies table.
Risk of bias in included studies
Our risk of bias assessment is summarised in Figure 3. The included studies were at low risk of bias in all domains except for Jamilian 2016, which was at unclear risk of selection bias. Three review authors were authors on one of the included studies (HB, MDN and LC) (Callaway 2019). Therefore, the other two review authors (SJD and SAP) assessed risk of bias in Callaway 2019 to minimise any effects from conflicts of interest.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Allocation
Random sequence generation
All included studies used computer‐generated randomisation procedures. Some studies used block randomisation with blocks of four (Pellonpera 2019), six (Laitinen 2009), or random block sizes (Okesene‐Gafa 2019; Wickens 2017). Other studies used simple 1:1 randomisation (Lindsay 2014), or did not state whether blocking methods were used (Callaway 2019; Jamilian 2016). All studies were classified at low risk of bias for random sequence generation.
Allocation concealment
All studies stated that allocation concealment was used. Three studies used sealed, opaque envelopes to conceal the allocation sequence prior to recruitment (Callaway 2019; Laitinen 2009; Lindsay 2014), and three studies assigned participants sequentially using a randomisation sequence generated by a third party and unknown to the study staff responsible for enrolment (Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). One study claimed to use allocation concealment but provided no further information, so this study was classified at unclear risk of bias (Jamilian 2016). All other studies were at low risk of bias (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017).
Blinding
All studies blinded both participants and personnel to probiotic/placebo allocation. While Laitinen 2009 had one group where the probiotic/placebo intervention was not blinded to study staff (placebo plus routine dietary advice), this group was not used for comparison in this review. In addition, in the two studies that used a secondary dietary intervention, participants and personnel were not blinded to the dietary intervention (Laitinen 2009; Okesene‐Gafa 2019). However, the dietary intervention was the same in both the probiotic and placebo groups, and, therefore, the lack of blinding for this intervention did not affect the probiotics versus placebo comparison. Therefore, all included studies were at low risk for bias for both performance and detection bias.
Incomplete outcome data
There was minimal loss to follow‐up at the time of testing for GDM or third trimester measurements of glucose metabolism in all studies. Loss to follow‐up rates ranged from 0% to 13.9% and were similar between groups. All studies were classified as low risk for attrition bias. Attrition rates for each study are shown in the Characteristics of included studies table.
Selective reporting
All studies reported their prespecified or mentioned outcomes. All studies were at low risk for reporting bias.
Other potential sources of bias
The studies had no other sources of bias.
Effects of interventions
See: Summary of findings 1 Probiotics compared to placebo for preventing gestational diabetes (maternal outcomes); Summary of findings 2 Probiotics compared to placebo for preventing gestational diabetes (infant outcomes)
All seven included studies compared probiotics versus placebo. While two studies included a secondary dietary intervention (Laitinen 2009; Okesene‐Gafa 2019), the studies were not conducted in a way that facilitated our second comparison of probiotics versus diet. This comparison will be included in review updates.
Probiotics versus placebo
Four studies had one intervention with the comparison probiotics versus placebo (Callaway 2019; Jamilian 2016; Lindsay 2014; Wickens 2017). In the three studies that had two simultaneous interventions, the study groups used in this review were chosen to balance the effect of the secondary intervention between the probiotic and placebo groups (Laitinen 2009; Okesene‐Gafa 2019; Pellonpera 2019). Laitinen 2009 had three arms, and for this comparison we used two of these arms to isolate the effect of the probiotic intervention (probiotics plus dietary intervention for probiotics, placebo plus dietary intervention for placebo). Okesene‐Gafa 2019 conducted a 2×2 factorial study where all participants were separately randomised to the dietary intervention and probiotics, so we used all groups for this comparison (probiotics with or without dietary intervention for probiotics, placebo with or without dietary intervention for placebo). Pellonpera 2019 was a four‐arm study of two interventions, so we only used the groups without the fish oil intervention for this comparison to isolate the effect of the probiotics (probiotics plus placebo for probiotics, placebo plus placebo for placebo).
Primary outcomes
Maternal
Diagnosis of gestational diabetes mellitus
Six studies reported GDM (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). We used a random‐effects model given the substantial heterogeneity present (I² = 64%). It is uncertain if probiotics have any effect on the risk for GDM compared to placebo (average RR 0.80, 95% CI 0.54 to 1.20; 1440 women; I2 = 64%; Tau² = 0.15; Analysis 1.1). Given the substantial heterogeneity and the wide CI including both appreciable benefit and appreciable harm, this evidence was low certainty (summary of findings Table 1).
The studies that reported GDM used different criteria for the diagnosis. Four studies used IADPSG criteria (Callaway 2019; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017), one study used local criteria based on the Fourth International Workshop‐Conference on Gestational Diabetes criteria (Laitinen 2009), and one study used Carpenter and Coustan criteria (Lindsay 2014). Sensitivity analysis was performed based on these criteria, and the results were largely unchanged. There was a reduced risk of GDM when using the Fourth International Workshop‐Conference on Gestational Diabetes criteria, but this is only based on one study and, therefore, should be interpreted with caution.
Subgroup analyses based on whether the reported dose of probiotics was less than five billion CFU (2 studies, 547 participants) or greater than five billion CFU (4 studies, 911 participants) found a difference between the subgroups (Chi² = 6.92, P = 0.009, I² = 85.5%; Analysis 1.2). However, there was still substantial heterogeneity in the subgroup with a dose greater than five billion CFU (Tau² = 0.07, I² = 51%), and the subgroup with a dose less than five billion CFU had only two studies. In addition, given the decay in probiotics over time, this subgroup analysis was conducted based on reported minimum or mean dose and there was no guarantee all participants received the reported dose. Therefore, this subgroup analysis should be interpreted with caution. The subgroup analysis based on bacterial species and duration of treatment revealed no clear differences, although both had subgroups with only one trial (Analysis 1.3; Analysis 1.4).
Hypertensive disorders of pregnancy (including pre‐eclampsia, pregnancy‐induced hypertension and eclampsia)
Four studies reported hypertensive disorders of pregnancy (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019). Probiotics may increase the risk of hypertensive disorders of pregnancy compared to placebo (RR 1.39, 95% CI 0.96 to 2.01; 955 women; I2 = 0%; Analysis 1.5). Subgroup analyses found no differences in the results, although most subgroups only included one or two studies (Analysis 1.6; Analysis 1.7; Analysis 1.8).
Four studies reported pre‐eclampsia (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019). Probiotics increase the risk of pre‐eclampsia compared to placebo (RR 1.85, 95% CI 1.04 to 3.29; 955 women; I2 = 0%; Analysis 1.9; high‐certainty evidence; summary of findings Table 1).
Caesarean section
Six studies reported caesarean section (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). Probiotics make little to no difference in the rate of caesarean sections compared to placebo (RR 1.00, 95% CI 0.86 to 1.17; 1520 women; I2 = 0%; Analysis 1.10; high‐certainty evidence; summary of findings Table 1). Subgroup analyses revealed no differences in the results, although most subgroups only included one or two studies (Analysis 1.11; Analysis 1.12; Analysis 1.13).
Infant
Large‐for‐gestational age
Four studies reported large‐for‐gestational age (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019). One study defined large‐for‐gestational age as greater than 90th percentile on customised percentile charts (Okesene‐Gafa 2019), while the other three studies also defined large‐for‐gestational age as greater than 90th percentile but did not specify what charts were used (Callaway 2019; Lindsay 2014; Pellonpera 2019). Probiotics probably make little to no difference in the risk of being large‐for‐gestational age compared to placebo (RR 0.99, 95% CI 0.72 to 1.36; 919 infants; I2 = 0%; Analysis 1.14; moderate‐certainty evidence; summary of findings Table 2). Subgroup analyses revealed no differences in the results, although most subgroups only included one or two studies (Analysis 1.15; Analysis 1.16; Analysis 1.17).
Perinatal mortality (including stillbirth and neonatal death)
Three studies reported perinatal mortality (Callaway 2019; Laitinen 2009; Lindsay 2014). However, two of these studies had no stillbirths or neonatal deaths in either group (Laitinen 2009; Lindsay 2014), and the other study had only one perinatal death across groups (Callaway 2019). We do not know if probiotics have an effect on perinatal mortality compared to placebo because the wide CI crossed the line of no effect (RR 0.33, 95% CI 0.01 to 8.02; 3 studies, 709 infants; I2 = 0%; Analysis 1.18; low‐certainty evidence; summary of findings Table 2). This evidence was of low certainty due to the small number of events and very wide CIs. Given the lack of data on this outcome, subgroup analyses would not be meaningful and were not performed.
Mortality or morbidity composite
Two studies reported a composite measure of neonatal morbidity (Callaway 2019; Okesene‐Gafa 2019). Callaway 2019 used a composite measure of birth injury including nerve injury, bone fracture and intracranial haemorrhage. Okesene‐Gafa 2019 used a composite measure of morbidity including birth trauma, hypoxic‐ischaemic encephalopathy, sepsis, respiratory distress requiring continuous positive airway pressure and hypoglycaemia requiring intravenous therapy. It is uncertain if probiotics have any effect on neonatal morbidity compared to placebo because the CIs were consistent with appreciable harm and appreciable benefit (RR 0.69, 95% CI 0.36 to 1.35; 623 infants; I2 = 0%; Analysis 1.19; low‐certainty evidence; summary of findings Table 2). Subgroup analyses were not performed because only two studies reported this outcome.
Secondary outcomes
Maternal
Induction of labour
Two studies reported induction of labour (Callaway 2019; Lindsay 2014). Probiotics may make little to no difference in induction of labour rates compared to placebo (RR 1.08, 95% CI 0.85 to 1.39; 544 women; I2 = 23%; Analysis 1.20).
Perineal trauma
No studies reported perineal trauma.
Placental abruption
No studies reported placental abruption.
Postpartum haemorrhage
Two studies reported postpartum haemorrhage (Lindsay 2014; Pellonpera 2019). One study defined postpartum haemorrhage as greater than 1000 mL (Pellonpera 2019), while the other study provided no definition (Lindsay 2014). It is uncertain if probiotics have any effect on the risk of postpartum haemorrhage compared to placebo (RR 1.05, 95% CI 0.60 to 1.85; 324 women; I2 = 0%; Analysis 1.21).
Postpartum infection
No studies reported postpartum infection.
Weight gain during pregnancy
Four studies reported weight gain during pregnancy (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019). Two studies specified the reported weight gain was from baseline to 36 weeks' gestation (Callaway 2019; Okesene‐Gafa 2019), while the other two studies stated "total" weight gain over the course of the pregnancy (Laitinen 2009; Lindsay 2014). There was substantial heterogeneity using a random‐effects model (I² = 40%). Probiotics probably make little to no difference in weight gain during pregnancy compared to placebo (MD 0.30 kg, 95% CI –0.67 to 1.26; 853 women; Analysis 1.22).
Adherence to the intervention
Six studies reported intervention adherence (Jamilian 2016; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). However, it was reported differently by each study, and, therefore, was not amenable to meta‐analysis. Jamilian 2016 reported that all participants received all capsules throughout the intervention (probiotics 30/30 women, placebo 30/30 women). Laitinen 2009 reported that participants reported 99.5% of capsules taken at visit two, 99% at visit three and 95% at visit four; these data were not separated out by group. Lindsay 2014 reported that the number of missed capsules was similar between groups, with 9/63 participants missing three or more capsules in the probiotics group and 12/75 participants missing three or more capsules in the placebo group. Okesene‐Gafa 2019 reported that over 75% of capsules were taken by 87/115 participants, but these data were not separated by group. Pellonpera 2019 reported good compliance by 88.4% of the entire study cohort, and stated that adherence was similar between groups. Wickens 2017 reported median adherence rates with interquartile ranges (IQR), and found no clear difference between groups (probiotics: median 94.9%, IQR 85.7 to 98.8; placebo: median 94.0%, IQR 85.9 to 98.8). Overall, there may be little to no difference in adherence rates between probiotics and placebo.
Behaviour changes associated with the intervention
No studies reported behaviour changes associated with the intervention.
Relevant biomarker changes associated with the intervention (e.g. adiponectin, free fatty acids, triglycerides, high‐density lipoproteins, low‐density lipoproteins, insulin)
All included studies reported at least one relevant biomarker.
Seven studies reported fasting plasma glucose in the third trimester. Given the substantial heterogeneity, we used a random‐effects model (I² = 69%). Probiotics may make little to no difference in fasting plasma glucose levels in the third trimester compared to placebo (MD –0.04 mmol/L, 95% CI –0.12 to 0.05; 1519 women; I2 = 69%; Analysis 1.23).
Four studies reported plasma glucose at one hour of a 75 g oral glucose tolerance test (OGTT) (Callaway 2019; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). Probiotics may make little to no difference in one‐hour OGTT results compared to placebo (MD –0.07 mmol/L, 95% CI –0.27 to 0.13; 1110 women; I2 = 0%; Analysis 1.24).
Four studies reported plasma glucose at two hours of a 75 g OGTT (Callaway 2019; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). Probiotics may make little to no difference in two‐hour OGTT results compared to placebo (0.02 mmol/L, 95% –0.13 to 0.18; 1186 women; I2 = 0%; Analysis 1.25).
Two studies reported triglycerides at the end of the intervention period (Jamilian 2016; Lindsay 2014). Probiotics may slightly reduce triglyceride levels compared to placebo (MD –0.21 mmol/L, 95% CI –0.40 to –0.02; 198 women; I2 = 0%; Analysis 1.26). However, given this is based on two small studies with a wide CI, the certainty of this evidence was low.
Two studies reported high‐density lipoproteins at the end of the intervention period (Jamilian 2016; Lindsay 2014). Probiotics may make little to no difference in high‐density lipoprotein levels compared to placebo (MD 0.02 mmol/L, 95% CI –0.08 to 0.11; 198 women; I2 = 0%; Analysis 1.27).
Two studies reported low‐density lipoproteins at the end of the intervention period (Jamilian 2016; Lindsay 2014). Probiotics may slightly reduce low‐density lipoprotein levels compared to placebo, but the CIs indicates probiotics may make little or no difference (MD –0.22 mmol/L, 95% CI –0.48 to 0.04; 198 women; I2 = 0%; Analysis 1.28).
Two studies reported total cholesterol at the end of the intervention period (Jamilian 2016; Lindsay 2014). Probiotics may slightly reduce total cholesterol levels compared to placebo, but the 95% CI included zero indicating the possibility that probiotics may make little or no difference (MD –0.31 mmol/L, 95% CI –0.62 to –0.00; 198 women; I2 = 0%; Analysis 1.29).
Four studies reported insulin levels in the third trimester (Jamilian 2016; Laitinen 2009; Lindsay 2014; Pellonpera 2019). Probiotics may reduce insulin levels slightly compared to placebo (MD –1.95 mU/L, 95% CI –3.01 to –0.88; 538 women; I2 = 0%; Analysis 1.30).
Sense of wellbeing and quality of life
One study reported sense of wellbeing and quality of life (Okesene‐Gafa 2019). This study reported several different measures of wellbeing and quality of life at 36 weeks' gestation. It is uncertain if probiotics have any effect compared to placebo in Edinburgh Postnatal Depression scores at 36 weeks (MD 0.42, 95% CI –0.89 to 1.73; 164 women; Analysis 1.31.1), Spielberger State‐Trait Anxiety Inventory Short Form scores at 36 weeks (MD –0.94, 95% CI –4.09 to 2.21; 164 women; Analysis 1.31.2), Short‐Form Health Survey scores Mental Component Score (MD 0.31, 95% CI –2.54 to 3.16; 164 women; Analysis 1.31.3) or Physical Component Score (MD 0.87, 95% CI –1.94 to 3.68; 164 women; Analysis 1.31.4).
Views of the intervention
One study reported views of the intervention (Lindsay 2014). Specifically, Lindsay 2014 reported if participants thought the "intervention was an inconvenience" (probiotics: 11/56 women, placebo: 12/63 women), "capsules were difficult to swallow" (probiotics: 1/56 women, placebo: 5/64 women), and if they "would consider taking a probiotic in a future pregnancy" (probiotics: 55/55 women, placebo: 60/64 women). Overall, Lindsay 2014 reported no differences in views of the intervention between groups.
Breastfeeding (e.g. at discharge, six weeks postpartum)
Two studies reported women who were breastfeeding any amount at six months (Laitinen 2009; Wickens 2017). It is uncertain if probiotics have any effect on numbers of women breastfeeding at six months compared to placebo (non‐event RR 1.08, 95% CI 0.77 to 1.50; 552 women; I2 = 0%; Analysis 1.32).
Long‐term maternal
Postnatal depression
No studies reported postnatal depression.
Postnatal weight retention or return to prepregnancy weight
One study reported weight at four to seven days postpartum (Wickens 2017). Probiotics may make little to no difference in postpartum weight compared with placebo (MD –0.10 kg, 95% CI –0.91 to 0.71; 391 women; Analysis 1.33). Laitinen 2009 also reported that weight decreased similarly between groups but provided no data.
Body mass index
One study reported body mass index at four to seven days postpartum (Wickens 2017), one at one‐year postpartum (Laitinen 2009), and one at four years postpartum (Laitinen 2009). At all time points, there was no clear difference between groups (4 to 7 days: MD –0.10 kg/m², 95% CI –0.38 to 0.18; 391 women; 12 months: MD –0.10 kg/m², 95% CI –0.65 to 0.45; 128 women; 4 years: MD 0.70 kg/m², 95% CI –0.18 to 1.58; 80 women; Analysis 1.34).
Gestational diabetes mellitus in a subsequent pregnancy
No studies reported GDM in a subsequent pregnancy.
Type 1 diabetes
No studies reported type 1 diabetes.
Type 2 diabetes
No studies reported type 2 diabetes.
Impaired glucose tolerance
No studies reported glucose tolerance.
Cardiovascular health as defined by trialists (including blood pressure, hypertension, cardiovascular disease and metabolic syndrome)
No studies reported cardiovascular health.
Infant
Stillbirth
Five studies reported stillbirth (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019). Of the 1128 participants, there were six stillbirths. Therefore, we do not know if probiotics reduce stillbirth compared to placebo because there were so few events and the CIs were very wide (RR 0.59, 95% CI 0.14 to 2.46; 1128 women; I2 = 0%; Analysis 1.35).
Neonatal mortality
Three studies reported neonatal mortality (Callaway 2019; Laitinen 2009; Lindsay 2014). However, there were no neonatal mortalities in any study, so the effect of probiotics could not be estimated (Analysis 1.36). Therefore, we do not know if probiotics reduce neonatal mortality compared to placebo.
Gestational age at birth
Six studies reported gestational age at birth (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). One study reported gestational age at birth as median and IQR, so this study was not included in the meta‐analysis (Wickens 2017). However, Wickens 2017 found no difference between groups (probiotics: 39.7 weeks, IQR 38.7 to 40.7; placebo: 39.6 weeks, IQR 38.7 to 40.4). We included the other five studies in the meta‐analysis (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019). Probiotics may make little to no difference in gestational age at birth compared to placebo (MD 0.01 weeks, 95% CI –0.19 to 0.21; 5 studies, 1073 infants; I2 = 26%; Analysis 1.37).
Preterm birth (less than 37 weeks' gestation and less than 32 weeks' gestation)
Six studies reported preterm birth (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). All studies defined preterm birth as prior to 37 weeks' gestation. It is uncertain if probiotics have an effect on the risk of preterm birth compared to placebo (RR 1.32, 95% CI 0.86 to 2.01; 1484 participants; I2 = 0%; Analysis 1.38).
Apgar score (less than seven at five minutes)
Three studies reported Apgar score (Laitinen 2009; Pellonpera 2019; Wickens 2017). However, all studies reported this outcome differently, so the data were not conducive to a meta‐analysis. Laitinen 2009 reported the median and range of five‐minute Apgar scores, which showed no difference between groups (probiotics: median 9, range 6 to 10; placebo: median 9, range 3 to 10). Pellonpera 2019 reported the mean and SD of five‐minute Apgar scores, which showed no difference between groups (probiotics: mean 9.0, SD 0.7; placebo: 9.0, SD 0.8). Finally, Wickens 2017 reported the proportion of infants who had an Apgar score of seven or greater at five minutes and found no difference between groups (probiotics: 200/203; placebo: 198/202). Overall, probiotics may make little to no difference in five‐minute Apgar scores compared to placebo.
Macrosomia
Three studies reported macrosomia defined as a birthweight greater than 4000 g (Callaway 2019; Lindsay 2014; Wickens 2017). Probiotics may make little to no difference in the risk of macrosomia compared to placebo (RR 1.13, 95% CI 0.86 to 1.48; 952 infants; I2 = 20%; Analysis 1.39).
Small‐for‐gestational age
Three studies reported SGA (Callaway 2019; Okesene‐Gafa 2019; Pellonpera 2019). One study defined SGA as less than 10th percentile using customised percentile charts (Okesene‐Gafa 2019), while the other two studies also used less than 10th percentile but did not state which charts they used (Callaway 2019; Pellonpera 2019). Probiotics may reduce the incidence of SGA infants compared to placebo (RR 0.51, 95% CI 0.30 to 0.85; 814 infants; Analysis 1.40).
Birthweight and z‐score
Six studies reported birthweight and z‐score (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). We used a random‐effects model due to substantial heterogeneity (I² = 42%). It is uncertain if probiotics have an effect on birthweight compared to placebo (MD 26.87 g, 95% CI –49.52 to 103.26; 1524 infants; Analysis 1.41). Two studies also reported birthweight z‐scores (Okesene‐Gafa 2019; Pellonpera 2019), and one study reported birthweight percentiles (Lindsay 2014). Two of these studies reported no difference between groups (Lindsay 2014; Pellonpera 2019), while one study reported a slight increase in birthweight z‐score in the probiotics group compared to placebo (Okesene‐Gafa 2019). We do not know if probiotics affect birthweight z‐score compared to placebo.
Head circumference and z‐score
Three studies reported head circumference and z‐score (Laitinen 2009; Okesene‐Gafa 2019; Wickens 2017). Probiotics may make little to no difference in head circumference compared to placebo (MD –0.04 cm, 95% CI –0.27 to 0.18; 789 infants; I2 = 21%; Analysis 1.42). One study reported head circumference z‐scores and also found no difference between groups (Okesene‐Gafa 2019).
Length and z‐score
Three studies reported length and z‐score (Laitinen 2009; Okesene‐Gafa 2019; Wickens 2017). We used a random‐effects model due to substantial heterogeneity (I² = 59%). Probiotics may make little to no difference in length compared to placebo (MD 0.02 cm, 95% CI –0.54 to 0.59; 786 infants; Analysis 1.43). One study reported length z‐scores and also found no difference between groups (Okesene‐Gafa 2019).
Ponderal index
Two studies reported ponderal index (Lindsay 2014; Wickens 2017). Probiotics may make little to no difference in ponderal index compared to placebo (MD 0.25 kg/m3, 95% CI –0.21 to 0.70; 539 infants; I2 = 0%; Analysis 1.44).
Adiposity
Two studies reported adiposity (Callaway 2019; Okesene‐Gafa 2019). Okesene‐Gafa 2019 reported adiposity as fat mass (MD –0.04 kg, 95% CI –0.12 to 0.04; 110 infants; Analysis 1.45), and Callaway 2019 reported adiposity as percentage fat (MD –0.10%, 95% CI –1.19 to 0.99; 210 infants; Analysis 1.46), which both showed no difference between probiotics and placebo.
Shoulder dystocia
No studies reported shoulder dystocia.
Bone fracture
No studies reported bone fracture.
Nerve palsy
No studies reported nerve palsy.
Respiratory distress syndrome
No studies reported respiratory distress syndrome.
Hypoglycaemia as defined by trialists
Two studies reported hypoglycaemia (Callaway 2019; Pellonpera 2019). One study defined hypoglycaemia as a fasting blood sugar level of less than 2.2 mmol/L (Callaway 2019), and one study defined hypoglycaemia as a fasting blood sugar level of less than 2.4 mmol/L (Pellonpera 2019). We used a random‐effects model due to substantial heterogeneity (I² = 37%). Given the heterogeneity and the wide CIs, we do not know if probiotics reduce the risk of hypoglycaemia compared to placebo (mean RR 1.15, 95% CI 0.69 to 1.92; 586 infants; Analysis 1.47).
Hyperbilirubinaemia
Two studies reported hyperbilirubinaemia, both as hyperbilirubinaemia requiring phototherapy (Callaway 2019; Pellonpera 2019). It is uncertain if probiotics have any effect on the risk of hyperbilirubinaemia compared to placebo (RR 0.95, 95% CI 0.66 to 1.38; 593 infants; I2 = 10%; Analysis 1.48).
Later infant and childhood
Weight and z‐scores
One study reported weight gain in grams per month from 0 to 6, 6 to 12 and 12 to 24 months (Laitinen 2009). Another study reported that weights and z‐scores were similar between groups at five months (Okesene‐Gafa 2019). Probiotics may make little to no difference in weight compared to placebo (0 to 6 months: MD –3 g/month, 95% CI –53.07 to 47.07; 6 to 12 months: MD 27 g/month, 95% CI –0.76 to 54.76; 12 to 24 months: MD –19 g/month, 95% CI –42.62 to 4.62; Analysis 1.49).
Height and z‐scores
One study reported growth in centimetres per month from 0 to 6, 6 to 12 and 12 to 24 months (Laitinen 2009). Another study reported that heights and z‐scores were similar between groups at five months (Okesene‐Gafa 2019). Probiotics may make little to no difference in height compared to placebo (0 to 6 months: MD –0.05 cm/month, 95% CI –0.15 to 0.05; 6 to 12 months: MD 0.02 cm/month, 95% CI –0.04 to 0.08; 12 to 24 months: MD 0.01 cm/month, 95% CI –0.04 to 0.06; Analysis 1.50).
Head circumference and z‐scores
One study reported head circumference with 119 participants at six months of age (Laitinen 2009). Another study reported head circumferences and z‐scores were similar between groups at five months (Okesene‐Gafa 2019). Probiotics may make little to no difference in head circumference compared to placebo (MD 0.30 cm, 95% CI –0.26 to 0.86; Analysis 1.51).
Adiposity (including body mass index and skinfold thickness)
One study reported that infant body mass indexes and skinfold thicknesses were similar between groups at five months (Okesene‐Gafa 2019). There were no data for meta‐analysis. Probiotics may have no effect on long‐term adiposity in children compared with placebo.
Blood pressure
One study reported mean blood pressures in infants aged six months (Laitinen 2009). Probiotics may make little to no difference in mean blood pressure compared to placebo (MD –1.00 mmHg, 95% CI –4.19 to 2.19; 114 participants; Analysis 1.52).
Type 1 diabetes
No studies reported type 1 diabetes.
Type 2 diabetes
No studies reported type 2 diabetes.
Impaired glucose tolerance
One study reported impaired glucose tolerance (Laitinen 2009). This study used 32–33 split proinsulin as a measure of glucose tolerance at six months, and levels were defined as abnormal if they were above the 85th percentile. We do not know if probiotics affect impair glucose tolerance in children (RR 0.95, 95% CI 0.34 to 2.69; Analysis 1.53).
Dyslipidaemia or metabolic syndrome
One study reported dyslipidaemia or metabolic syndrome (Laitinen 2009). Laitinen 2009 stated there were no differences in lipid levels with probiotics in children at one, two and four years of age compared to placebo. There were no data for meta‐analysis.
Neurodisability
No studies reported neurodisability.
Educational achievement
No studies reported educational achievement.
Child as an adult
Weight
No studies reported weight.
Height
No studies reported height.
Adiposity (including body mass index and skinfold thickness)
No studies reported adiposity.
Cardiovascular health as defined by trialists (including blood pressure, hypertension, cardiovascular disease and metabolic syndrome)
No studies reported cardiovascular health.
Type 1 diabetes
No studies reported type 1 diabetes.
Type 2 diabetes
No studies reported type 2 diabetes.
Impaired glucose tolerance
No studies reported impaired glucose tolerance.
Dyslipidaemia or metabolic syndrome
No studies reported dyslipidaemia or metabolic syndrome.
Employment, education and social status/achievement
No studies reported employment, education and social status/achievement.
Health service use
Number of hospital or health professional visits (including midwife, obstetrician, physician, dietician and diabetic nurse)
No studies reported health service use outcomes.
Number of antenatal visits or admissions
No studies reported number of antenatal visits or admissions.
Length of antenatal stay
No studies reported length of antenatal stay.
Neonatal intensive care unit admission
Five studies reported neonatal intensive care unit admission (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017). All studies reported neonatal intensive care unit admission alone except for Callaway 2019, which reported special care unit admission. Probiotics may make little to no difference in neonatal intensive care unit admissions compared to placebo (RR 0.97, 95% CI 0.75 to 1.26; 1354 infants; I2 = 0%; Analysis 1.54).
Length of postnatal stay (mother)
No studies reported length of postnatal stay (mother).
Length of postnatal stay (baby)
No studies reported length of postnatal stay (baby).
Costs to families associated with the management provided
No studies reported costs to families associated with the management provided.
Costs associated with the intervention
No studies reported costs associated with the intervention.
Cost of maternal care
No studies reported cost of maternal care.
Cost of offspring care
No studies reported cost of offspring care.
Probiotics versus diet
We found no trials comparing probiotics versus diet.
Discussion
Summary of main results
The aim of this review was to determine the effect of probiotic supplementation during pregnancy on the risk of developing gestational diabetes. Seven trials met our inclusion criteria, and all included studies compared probiotics with placebo.
Six included studies in this review reported the incidence of GDM in 1440 participants. It is uncertain if probiotics have any effect on the risk of GDM compared to placebo because there was substantial heterogeneity between studies and wide CIs that included both appreciable benefit and harm (low‐certainty evidence). Two of these studies reported a reduction in the risk of GDM with probiotics, while the other four studies reported no difference. This heterogeneity was explored through subgroup analysis, and identified no clear causes for the heterogeneity.
Among the other primary outcomes for this review, we found probiotics increase the risk of pre‐eclampsia (high‐certainty evidence) and may increase the risk of hypertensive disorders of pregnancy, although the CIs for hypertensive disorders of pregnancy included the possibility of no effect.
There were few differences between groups for this review's other main outcomes. Probiotics make little to no difference in the risk of caesarean section (high‐certainty evidence), and probably make little to no difference in maternal weight gain during pregnancy (moderate‐certainty evidence). Probiotics probably make little to no difference in the incidence of large‐for‐gestational age infants (moderate‐certainty evidence) and may make little to no difference in neonatal adiposity (low‐certainty evidence, data from two studies not pooled). We do not know the effect of probiotics on perinatal mortality (low‐certainty evidence), a composite measure of neonatal morbidity (low‐certainty evidence) or neonatal hypoglycaemia (low‐certainty evidence) because of serious imprecision in the effect estimates and CIs, which are consistent with possible benefit and possible harm. No included studies reported on perineal trauma, postnatal depression, maternal and infant development of diabetes, and neurosensory disability.
There were few differences between groups for most of the secondary outcomes reported in this review. Among the markers for glucose tolerance, there was no difference for fasting plasma glucose and OGTT results. While there was a reduction in insulin levels with probiotics, this finding was of minimal importance given there were no differences in glucose levels or the diagnosis of GDM. The only other notable finding was that probiotics may reduce the risk of SGA infants compared to placebo. However, SGA was not one of our selected outcomes for inclusion in the 'Summary of findings' tables, therefore, we have not assessed the certainty of the evidence contributing to this outcome. The SGA data were based on relatively few events and it is not certain if the effect estimate is due to chance or a real difference between the intervention and control groups.
Overall completeness and applicability of evidence
This review included seven studies with 1647 participants. Given the incidence of GDM in the study population, interpretation of the results was limited by wide CIs and substantial heterogeneity between studies. We identified eight ongoing or unpublished studies as part of our search, which will add to the body of evidence when published and will hopefully help to overcome some of the present limitations.
The studies that reported the incidence of GDM were conducted in women at high risk of developing GDM due to being overweight or obese, or women with a history of atopic disease of any body mass index. Although this was not formally explored through subgroup analysis, the difference in study populations may explain some of the heterogeneity observed between studies. In particular, the authors would like to note that the two studies conducted in women of any weight with a history of atopic disease were the only two studies to detect a reduction in the risk of GDM with probiotics. Therefore, women with atopic disease may be a population where probiotics are effective. Another possibility is that probiotics are more effective in normal‐weight women compared to the higher‐risk overweight and obese women. Genetic differences or differences in diet between populations may also be responsible, although both Finland and New Zealand have each had one trial that showed benefit and one trial that showed no effect. More data are required before any conclusions can be made, but we plan to explore the effect of probiotics in these populations through subgroup analyses in updates of this review.
While multiple studies reported all our primary outcomes, many of our secondary outcomes were not reported or were reported by only one study. In particular, data were lacking in most of the long‐term outcomes since only one included study had significant long‐term follow‐up. There were also almost no data on health service use apart from neonatal intensive care unit admission. More studies will need to include long‐term follow‐up and data on health service use before any conclusions can be drawn about these outcomes.
Quality of the evidence
Overall, risk of bias among the included studies was low across all domains, apart from one study (Jamilian 2016), which had an unclear risk of bias in relation to allocation concealment. While some studies lacked blinding for secondary interventions, all studies were double blind for the probiotics versus placebo comparison.
We assessed certainty of evidence using GRADE methodology, and this assessment is given in summary of findings Table 1 and summary of findings Table 2. Overall, the certainty of evidence ranged from low to high, with downgrading due to concerns around inconsistency and imprecision. Certainty varied widely across the assessed outcomes, with evidence of high certainty for caesarean section, pre‐eclampsia; moderate certainty for weight gain during pregnancy, large‐for‐gestational age and composite neonatal morbidity; and low certainty for GDM, perinatal mortality, neonatal hypoglycaemia and adiposity.
Potential biases in the review process
There are possible sources of bias with any review, and we minimised these sources of bias. We performed the search for studies in this area using the Cochrane Pregnancy and Childbirth Group's Trials Register, which is updated weekly to monthly with information from CENTRAL, MEDLINE, Embase, handsearches from 30 journals and conference proceeding of major conferences, and alerts for a further 44 journals to minimise the risk of missing eligible studies. Two review authors independently assessed studies for inclusion and risk of bias, and resolved any discrepancies through discussion or with the entire author team if necessary.
This review was also uniquely susceptible to bias given three of the review authors (MDN, LC and HB) were also authors on a study that was identified for possible inclusion. For this Cochrane Review, the two review authors who were not involved with the study (SJD and SAP) independently assessed the relevant trial reports for inclusion, determined risk of bias and collected all data to minimise any conflict of interest.
Agreements and disagreements with other studies or reviews
Many reviews have evaluated the impact of probiotics on GDM, either as part of a specific review or a general review on prevention strategies for GDM or the use of probiotics during pregnancy. Since many of the studies included in this review were only published in the past few years, many older reviews on the topic include only one or two studies (Agha‐Jaffar 2016; Barrett 2012; Facchinetti 2014; Gomez Arango 2015; Griffin 2015; Jarde 2018; Lindsay 2013; Plows 2019; Rogozinska 2015; Simmons 2015; van de Vusse 2013). Depending on inclusion criteria and publication date, some combinations of Laitinen 2009, Lindsay 2014, and Wickens 2017, and one study we excluded (Asemi 2013) were included in each review. Overall, these reviews concluded that probiotics may reduce the risk of GDM, but agreed that more data are needed to make firm conclusions.
Two recent systematic reviews have been published that include the newer trials. Masulli 2020 published a systematic review and meta‐analysis including 17 trials evaluating the use of probiotics during pregnancy on various metabolic outcomes in both women with and without GDM. For the diagnosis of GDM outcome, this review included seven studies: the six included in this review (Callaway 2019; Laitinen 2009; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019; Wickens 2017) and one awaiting classification (Asgharian 2020). In agreement with our review, Masulli 2020 reported no benefit with probiotics on the incidence of GDM, although they also demonstrated substantial heterogeneity between studies (RR 0.77, 95% CI 0.51 to 1.16; I2 = 62%).
The other recently published systematic review on the topic was a network meta‐analysis evaluating a variety of interventions to prevent GDM specifically in overweight and obese pregnant women (Chatzakis 2019). In total, the review included 23 studies, five of which evaluated the use of probiotics. Four of these studies were included in our review (Callaway 2019; Lindsay 2014; Okesene‐Gafa 2019; Pellonpera 2019), while the other study is awaiting classification (Asgharian 2020). Overall, Chatzakis 2019 found no interventions to be superior to placebo. Their direct comparison of probiotics and placebo revealed no clear differences between groups in agreement with our review (RR 1.02, 95% CI 0.78 to 1.32). However, there was no substantial heterogeneity between the studies as seen in our review, which is likely reflective of the fact that Chatzakis 2019 limited their review to overweight and obese women. The effect of body mass index was not evaluated through subgroup analysis in our review, but further supports our observation that probiotics may not be effective for preventing GDM in overweight and obese women.
Study flow diagram.
Applying the trustworthiness screening tool
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Comparison 1: Probiotics versus placebo, Outcome 1: Gestational diabetes mellitus
Comparison 1: Probiotics versus placebo, Outcome 2: Gestational diabetes mellitus (by dose)
Comparison 1: Probiotics versus placebo, Outcome 3: Gestational diabetes mellitus (by bacterial species)
Comparison 1: Probiotics versus placebo, Outcome 4: Gestational diabetes mellitus (by duration of treatment)
Comparison 1: Probiotics versus placebo, Outcome 5: Hypertensive disorders of pregnancy
Comparison 1: Probiotics versus placebo, Outcome 6: Hypertensive disorders of pregnancy (by dose)
Comparison 1: Probiotics versus placebo, Outcome 7: Hypertensive disorders of pregnancy (by bacterial species)
Comparison 1: Probiotics versus placebo, Outcome 8: Hypertensive disorders of pregnancy (by duration of treatment)
Comparison 1: Probiotics versus placebo, Outcome 9: Pre‐eclampsia
Comparison 1: Probiotics versus placebo, Outcome 10: Caesarean section
Comparison 1: Probiotics versus placebo, Outcome 11: Caesarean section (by dose)
Comparison 1: Probiotics versus placebo, Outcome 12: Caesarean section (by bacterial species)
Comparison 1: Probiotics versus placebo, Outcome 13: Caesarean section (by duration of treatment)
Comparison 1: Probiotics versus placebo, Outcome 14: Large‐for‐gestational age
Comparison 1: Probiotics versus placebo, Outcome 15: Large‐for‐gestational age (by dose)
Comparison 1: Probiotics versus placebo, Outcome 16: Large‐for‐gestational age (by bacterial species)
Comparison 1: Probiotics versus placebo, Outcome 17: Large‐for‐gestational age (by duration of treatment)
Comparison 1: Probiotics versus placebo, Outcome 18: Perinatal mortality (stillbirth and neonatal mortality)
Comparison 1: Probiotics versus placebo, Outcome 19: Mortality or morbidity composite
Comparison 1: Probiotics versus placebo, Outcome 20: Induction of labour
Comparison 1: Probiotics versus placebo, Outcome 21: Postpartum haemorrhage
Comparison 1: Probiotics versus placebo, Outcome 22: Weight gain during pregnancy (kg)
Comparison 1: Probiotics versus placebo, Outcome 23: Fasting plasma glucose (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 24: 1‐hour oral glucose tolerance test (OGTT) plasma glucose (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 25: 2‐hour OGTT plasma glucose (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 26: Triglycerides (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 27: High‐density lipoprotein (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 28: Low‐density lipoprotein (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 29: Total cholesterol (mmol/L)
Comparison 1: Probiotics versus placebo, Outcome 30: Insulin (mU/L)
Comparison 1: Probiotics versus placebo, Outcome 31: Sense of wellbeing and quality of life
Comparison 1: Probiotics versus placebo, Outcome 32: Breastfeeding at 6 months
Comparison 1: Probiotics versus placebo, Outcome 33: Postnatal weight retention (kg)
Comparison 1: Probiotics versus placebo, Outcome 34: Body mass index (kg/m2)
Comparison 1: Probiotics versus placebo, Outcome 35: Stillbirth
Comparison 1: Probiotics versus placebo, Outcome 36: Neonatal mortality
Comparison 1: Probiotics versus placebo, Outcome 37: Gestational age at birth (weeks)
Comparison 1: Probiotics versus placebo, Outcome 38: Preterm birth
Comparison 1: Probiotics versus placebo, Outcome 39: Macrosomia
Comparison 1: Probiotics versus placebo, Outcome 40: Small‐for‐gestational age
Comparison 1: Probiotics versus placebo, Outcome 41: Birthweight (g)
Comparison 1: Probiotics versus placebo, Outcome 42: Head circumference (cm)
Comparison 1: Probiotics versus placebo, Outcome 43: Length (cm)
Comparison 1: Probiotics versus placebo, Outcome 44: Ponderal index (kg/m3)
Comparison 1: Probiotics versus placebo, Outcome 45: Adiposity – fat mass (kg)
Comparison 1: Probiotics versus placebo, Outcome 46: Adiposity – % fat
Comparison 1: Probiotics versus placebo, Outcome 47: Hypoglycaemia
Comparison 1: Probiotics versus placebo, Outcome 48: Hyperbilirubinaemia
Comparison 1: Probiotics versus placebo, Outcome 49: Infant weight gain (g/month)
Comparison 1: Probiotics versus placebo, Outcome 50: Infant height (cm/month)
Comparison 1: Probiotics versus placebo, Outcome 51: Infant head circumference – 6 months (cm)
Comparison 1: Probiotics versus placebo, Outcome 52: Infant mean blood pressure – 6 months (mmHg)
Comparison 1: Probiotics versus placebo, Outcome 53: 32–33 split proinsulin > 85th percentile – 6 months
Comparison 1: Probiotics versus placebo, Outcome 54: Neonatal intensive care unit admission
| Probiotics compared to placebo for preventing gestational diabetes (maternal outcomes) | ||||||
| Patient or population: preventing gestational diabetes Setting: Australia, Finland, Iran, Ireland, and New Zealand Intervention: probiotics Comparison: placebo | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with Probiotics | |||||
| Diagnosis of gestational diabetes mellitus | Study population | Mean RR 0.80 | 1440 | ⊕⊕⊝⊝ | — | |
| 191 per 1000 | 153 per 1000 | |||||
| Hypertensive disorders of pregnancy (pre‐eclampsia) | Study population | RR 1.85 | 955 | ⊕⊕⊕⊕ | — | |
| 35 per 1000 | 65 per 1000 | |||||
| Caesarean section | Study population | RR 1.00 | 1520 | ⊕⊕⊕⊕ | — | |
| 285 per 1000 | 285 per 1000 | |||||
| Perineal trauma | — | — | — | — | — | Not reported |
| Weight gain during pregnancy | The mean weight gain during pregnancy was 9.4–14.8 kg | MD 0.30 kg higher | — | 853 | ⊕⊕⊕⊝ | — |
| Postnatal depression | — | — | — | — | — | Not reported |
| Development of subsequent diabetes | — | — | — | — | — | Not reported |
| *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; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio. | ||||||
| GRADE Working Group grades of evidence | ||||||
| aDowngraded one level for serious concerns about inconsistency due to substantial unexplained heterogeneity between studies. | ||||||
| Probiotics compared to placebo for preventing gestational diabetes (infant outcomes) | ||||||
| Patient or population: preventing gestational diabetes Setting: Australia, Finland, Iran, Ireland, and New Zealand Intervention: probiotics Comparison: placebo | ||||||
| Outcomes | Anticipated absolute effects* (95% CI) | Relative effect | № of participants | Certainty of the evidence | Comments | |
|---|---|---|---|---|---|---|
| Risk with placebo | Risk with probiotics | |||||
| Large‐for‐gestational age | Study population | RR 0.99 | 919 | ⊕⊕⊕⊝ | — | |
| 142 per 1000 | 141 per 1000 | |||||
| Perinatal mortality (stillbirth and neonatal mortality) | Study population | RR 0.33 | 709 | ⊕⊕⊝⊝ | — | |
| 3 per 1000 | 1 per 1000 | |||||
| Mortality or morbidity composite | Study population | RR 0.69 | 623 | ⊕⊕⊝⊝ | — | |
| 61 per 1000 | 42 per 1000 | |||||
| Hypoglycaemia as defined by trialists | Study population | Mean RR 1.15 | 586 | ⊕⊕⊝⊝ | — | |
| 135 per 1000 | 155 per 1000 | |||||
| Adiposity | The included studies showed no appreciable difference in fat mass or % fat between groups. 1 study reported adiposity as fat mass (MD –0.04 kg, 95% CI –0.12 to 0.04) 1 study reported adiposity as % fat (MD –0.10%, 95% CI –1.19 to 0.99) | — | 320 | ⊕⊕⊝⊝ | — | |
| Diabetes | — | — | — | — | — | Not reported |
| Neurodisability | — | — | — | — | — | Not reported |
| *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; MD: mean difference; RCT: randomised controlled trial; RR: risk ratio. | ||||||
| GRADE Working Group grades of evidence | ||||||
| aDowngraded one level for serious concerns about imprecision due to a wide confidence intervals that included both appreciable benefit and harm. | ||||||
| Parameter | IADPSG (IADPSG 2010 ) | Carpenter and Coustan (Carpenter 1982 ) | Modified Fourth International Workshop‐Conference (Metzger 1998 ) | New Zealand Guidelines (Ministry of Health 2014 ) | Finnish Current Care Guidelines (The Finnish Medical Society Duodecim 2013 ) |
|---|---|---|---|---|---|
| OGTT (g) | 75 | 100 | 75 | 75 | 75 |
| Fasting (mmol/L) | 5.1 | 5.3 | 4.8 | 5.5 | 5.3 |
| 1 hour (mmol/L) | 10.0 | 10.0 | 10.0 | — | 10.0 |
| 2 hours (mmol/L) | 8.5 | 8.6 | 8.7 | 9.0 | 8.6 |
| 3 hours (mmol/L) | — | 7.8 | — | — | — |
| Elevated values required | 1 | 2 | 1 | 1 | 1 |
| IADPSG: International Association of Diabetes and Pregnancy Study Groups; OGTT: oral glucose tolerance test. | |||||
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
| 1.1 Gestational diabetes mellitus Show forest plot | 6 | 1440 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.54, 1.20] |
| 1.2 Gestational diabetes mellitus (by dose) Show forest plot | 6 | 1440 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.54, 1.20] |
| 1.2.1 < 5 billion CFU | 2 | 547 | Risk Ratio (M‐H, Random, 95% CI) | 1.47 [0.94, 2.30] |
| 1.2.2 > 5 billion CFU | 4 | 893 | Risk Ratio (M‐H, Random, 95% CI) | 0.67 [0.46, 0.98] |
| 1.3 Gestational diabetes mellitus (by bacterial species) Show forest plot | 6 | 1440 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.54, 1.20] |
| 1.3.1 Lactobacillusrhamnosus + Bifidobacteriumanimalis | 4 | 931 | Risk Ratio (M‐H, Random, 95% CI) | 0.83 [0.50, 1.37] |
| 1.3.2 Lactobacillus rhamnosus | 1 | 373 | Risk Ratio (M‐H, Random, 95% CI) | 0.59 [0.32, 1.08] |
| 1.3.3 Lactobacillus salivarius | 1 | 136 | Risk Ratio (M‐H, Random, 95% CI) | 1.19 [0.25, 5.70] |
| 1.4 Gestational diabetes mellitus (by duration of treatment) Show forest plot | 6 | 1440 | Risk Ratio (M‐H, Random, 95% CI) | 0.80 [0.54, 1.20] |
| 1.4.1 Started early pregnancy | 5 | 1304 | Risk Ratio (M‐H, Random, 95% CI) | 0.78 [0.51, 1.20] |
| 1.4.2 Started ≥ 20 weeks' gestation | 1 | 136 | Risk Ratio (M‐H, Random, 95% CI) | 1.19 [0.25, 5.70] |
| 1.5 Hypertensive disorders of pregnancy Show forest plot | 4 | 955 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.39 [0.96, 2.01] |
| 1.6 Hypertensive disorders of pregnancy (by dose) Show forest plot | 4 | 955 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.39 [0.96, 2.01] |
| 1.6.1 < 5 billion CFU | 2 | 545 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.35 [0.87, 2.12] |
| 1.6.2 > 5 billion CFU | 2 | 410 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.47 [0.77, 2.81] |
| 1.7 Hypertensive disorders of pregnancy (by bacterial species) Show forest plot | 4 | 955 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.39 [0.96, 2.01] |
| 1.7.1 Lactobacillusrhamnosus + Bifidobacteriumanimalis | 3 | 819 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.35 [0.92, 1.98] |
| 1.7.2 Lactobacillus salivarius | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.99 [0.49, 7.99] |
| 1.8 Hypertensive disorders of pregnancy (by duration of treatment) Show forest plot | 4 | 955 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.39 [0.96, 2.01] |
| 1.8.1 Started early pregnancy | 3 | 819 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.35 [0.92, 1.98] |
| 1.8.2 Started ≥ 20 weeks' gestation | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.99 [0.49, 7.99] |
| 1.9 Pre‐eclampsia Show forest plot | 4 | 955 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.85 [1.04, 3.29] |
| 1.10 Caesarean section Show forest plot | 6 | 1520 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.00 [0.86, 1.17] |
| 1.11 Caesarean section (by dose) Show forest plot | 6 | 1520 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.00 [0.86, 1.17] |
| 1.11.1 < 5 billion CFU | 2 | 547 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.91 [0.73, 1.14] |
| 1.11.2 > 5 billion CFU | 4 | 973 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.09 [0.87, 1.36] |
| 1.12 Caesarean section (by bacterial species) Show forest plot | 6 | 1520 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.00 [0.86, 1.17] |
| 1.12.1 Lactobacillusrhamnosus + Bifidobacteriumanimalis | 4 | 977 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.98 [0.81, 1.19] |
| 1.12.2 Lactobacillus rhamnosus | 1 | 407 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.09 [0.79, 1.51] |
| 1.12.3 Lactobacillus salivarius | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.95 [0.59, 1.55] |
| 1.13 Caesarean section (by duration of treatment) Show forest plot | 6 | 1520 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.00 [0.86, 1.17] |
| 1.13.1 Started early pregnancy | 5 | 1384 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.01 [0.86, 1.19] |
| 1.13.2 Started ≥ 20 weeks' gestation | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.95 [0.59, 1.55] |
| 1.14 Large‐for‐gestational age Show forest plot | 4 | 919 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.72, 1.36] |
| 1.15 Large‐for‐gestational age (by dose) Show forest plot | 4 | 919 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.72, 1.36] |
| 1.15.1 < 5 billion CFU | 2 | 509 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.08 [0.72, 1.62] |
| 1.15.2 > 5 billion CFU | 2 | 410 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.88 [0.53, 1.46] |
| 1.16 Large‐for‐gestational age (by bacterial species) Show forest plot | 4 | 919 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.72, 1.36] |
| 1.16.1 Lactobacillusrhamnosus + Bifidobacteriumanimalis | 3 | 783 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.71, 1.38] |
| 1.16.2 Lactobacillus salivarius | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.02 [0.36, 2.89] |
| 1.17 Large‐for‐gestational age (by duration of treatment) Show forest plot | 4 | 919 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.72, 1.36] |
| 1.17.1 Started early pregnancy | 3 | 783 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.99 [0.71, 1.38] |
| 1.17.2 Started ≥ 20 weeks' gestation | 1 | 136 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.02 [0.36, 2.89] |
| 1.18 Perinatal mortality (stillbirth and neonatal mortality) Show forest plot | 3 | 709 | Risk Ratio (IV, Fixed, 95% CI) | 0.33 [0.01, 8.02] |
| 1.19 Mortality or morbidity composite Show forest plot | 2 | 623 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.69 [0.36, 1.35] |
| 1.20 Induction of labour Show forest plot | 2 | 544 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.08 [0.85, 1.39] |
| 1.21 Postpartum haemorrhage Show forest plot | 2 | 324 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.05 [0.60, 1.85] |
| 1.22 Weight gain during pregnancy (kg) Show forest plot | 4 | 853 | Mean Difference (IV, Random, 95% CI) | 0.30 [‐0.67, 1.26] |
| 1.23 Fasting plasma glucose (mmol/L) Show forest plot | 7 | 1519 | Mean Difference (IV, Random, 95% CI) | ‐0.04 [‐0.12, 0.05] |
| 1.24 1‐hour oral glucose tolerance test (OGTT) plasma glucose (mmol/L) Show forest plot | 4 | 1110 | Mean Difference (IV, Fixed, 95% CI) | ‐0.07 [‐0.27, 0.13] |
| 1.25 2‐hour OGTT plasma glucose (mmol/L) Show forest plot | 4 | 1186 | Mean Difference (IV, Fixed, 95% CI) | 0.02 [‐0.13, 0.18] |
| 1.26 Triglycerides (mmol/L) Show forest plot | 2 | 198 | Mean Difference (IV, Fixed, 95% CI) | ‐0.21 [‐0.40, ‐0.02] |
| 1.27 High‐density lipoprotein (mmol/L) Show forest plot | 2 | 198 | Mean Difference (IV, Fixed, 95% CI) | 0.02 [‐0.08, 0.11] |
| 1.28 Low‐density lipoprotein (mmol/L) Show forest plot | 2 | 198 | Mean Difference (IV, Fixed, 95% CI) | ‐0.22 [‐0.48, 0.04] |
| 1.29 Total cholesterol (mmol/L) Show forest plot | 2 | 198 | Mean Difference (IV, Fixed, 95% CI) | ‐0.31 [‐0.62, ‐0.00] |
| 1.30 Insulin (mU/L) Show forest plot | 4 | 538 | Mean Difference (IV, Fixed, 95% CI) | ‐1.95 [‐3.01, ‐0.88] |
| 1.31 Sense of wellbeing and quality of life Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 1.31.1 Edinburgh Postnatal Depression Score – 36 weeks | 1 | 164 | Mean Difference (IV, Fixed, 95% CI) | 0.42 [‐0.89, 1.73] |
| 1.31.2 Spielberger State‐Trait Anxiety Inventory Short Form Score – 36 weeks | 1 | 164 | Mean Difference (IV, Fixed, 95% CI) | ‐0.94 [‐4.09, 2.21] |
| 1.31.3 12‐Item Short‐Form Health Survey – Mental Component Score, 36 weeks | 1 | 164 | Mean Difference (IV, Fixed, 95% CI) | 0.31 [‐2.54, 3.16] |
| 1.31.4 12‐Item Short‐Form Health Survey – Physical Component Score, 36 weeks | 1 | 164 | Mean Difference (IV, Fixed, 95% CI) | 0.87 [‐1.94, 3.68] |
| 1.32 Breastfeeding at 6 months Show forest plot | 2 | 552 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.08 [0.77, 1.50] |
| 1.33 Postnatal weight retention (kg) Show forest plot | 1 | 391 | Mean Difference (IV, Fixed, 95% CI) | ‐0.10 [‐0.91, 0.71] |
| 1.34 Body mass index (kg/m2) Show forest plot | 2 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 1.34.1 4–7 days postpartum | 1 | 391 | Mean Difference (IV, Fixed, 95% CI) | ‐0.10 [‐0.38, 0.18] |
| 1.34.2 12 months postpartum | 1 | 128 | Mean Difference (IV, Fixed, 95% CI) | ‐0.10 [‐0.65, 0.45] |
| 1.34.3 4 years postpartum | 1 | 80 | Mean Difference (IV, Fixed, 95% CI) | 0.70 [‐0.18, 1.58] |
| 1.35 Stillbirth Show forest plot | 5 | 1128 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.59 [0.14, 2.46] |
| 1.36 Neonatal mortality Show forest plot | 3 | 709 | Risk Ratio (M‐H, Fixed, 95% CI) | Not estimable |
| 1.37 Gestational age at birth (weeks) Show forest plot | 5 | 1073 | Mean Difference (IV, Fixed, 95% CI) | 0.01 [‐0.19, 0.21] |
| 1.38 Preterm birth Show forest plot | 6 | 1484 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.32 [0.86, 2.01] |
| 1.39 Macrosomia Show forest plot | 3 | 952 | Risk Ratio (M‐H, Fixed, 95% CI) | 1.13 [0.86, 1.48] |
| 1.40 Small‐for‐gestational age Show forest plot | 3 | 814 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.51 [0.30, 0.85] |
| 1.41 Birthweight (g) Show forest plot | 6 | 1524 | Mean Difference (IV, Random, 95% CI) | 26.87 [‐49.52, 103.26] |
| 1.42 Head circumference (cm) Show forest plot | 3 | 789 | Mean Difference (IV, Fixed, 95% CI) | ‐0.04 [‐0.27, 0.18] |
| 1.43 Length (cm) Show forest plot | 3 | 786 | Mean Difference (IV, Random, 95% CI) | 0.02 [‐0.54, 0.59] |
| 1.44 Ponderal index (kg/m3) Show forest plot | 2 | 539 | Mean Difference (IV, Fixed, 95% CI) | 0.25 [‐0.21, 0.70] |
| 1.45 Adiposity – fat mass (kg) Show forest plot | 1 | 110 | Mean Difference (IV, Fixed, 95% CI) | ‐0.04 [‐0.12, 0.04] |
| 1.46 Adiposity – % fat Show forest plot | 1 | 210 | Mean Difference (IV, Fixed, 95% CI) | ‐0.10 [‐1.19, 0.99] |
| 1.47 Hypoglycaemia Show forest plot | 2 | 586 | Risk Ratio (M‐H, Random, 95% CI) | 1.15 [0.69, 1.92] |
| 1.48 Hyperbilirubinaemia Show forest plot | 2 | 593 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.95 [0.66, 1.38] |
| 1.49 Infant weight gain (g/month) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 1.49.1 at 0–6 months | 1 | 162 | Mean Difference (IV, Fixed, 95% CI) | ‐3.00 [‐53.07, 47.07] |
| 1.49.2 at 6–12 months | 1 | 162 | Mean Difference (IV, Fixed, 95% CI) | 27.00 [‐0.76, 54.76] |
| 1.49.3 at 12–24 months | 1 | 130 | Mean Difference (IV, Fixed, 95% CI) | ‐19.00 [‐42.62, 4.62] |
| 1.50 Infant height (cm/month) Show forest plot | 1 | Mean Difference (IV, Fixed, 95% CI) | Subtotals only | |
| 1.50.1 at 0–6 months | 1 | 156 | Mean Difference (IV, Fixed, 95% CI) | ‐0.05 [‐0.15, 0.05] |
| 1.50.2 at 6–12 months | 1 | 156 | Mean Difference (IV, Fixed, 95% CI) | 0.02 [‐0.04, 0.08] |
| 1.50.3 at 12–24 months | 1 | 130 | Mean Difference (IV, Fixed, 95% CI) | 0.01 [‐0.04, 0.06] |
| 1.51 Infant head circumference – 6 months (cm) Show forest plot | 1 | 119 | Mean Difference (IV, Fixed, 95% CI) | 0.30 [‐0.26, 0.86] |
| 1.52 Infant mean blood pressure – 6 months (mmHg) Show forest plot | 1 | 114 | Mean Difference (IV, Fixed, 95% CI) | ‐1.00 [‐4.19, 2.19] |
| 1.53 32–33 split proinsulin > 85th percentile – 6 months Show forest plot | 1 | 131 | Risk Ratio (M‐H, Fixed, 95% CI) | 0.95 [0.34, 2.69] |
| 1.54 Neonatal intensive care unit admission Show forest plot | 5 | 1354 | Risk Ratio (IV, Fixed, 95% CI) | 0.97 [0.75, 1.26] |
