Intrapartum antibiotics for known maternal Group B streptococcal colonization

  • Comment
  • Review
  • Intervention

Authors

  • Arne Ohlsson,

    Corresponding author
    1. University of Toronto, Departments of Paediatrics, Obstetrics and Gynaecology and Institute of Health Policy, Management and Evaluation, Toronto, ON, Canada
    • Arne Ohlsson, Departments of Paediatrics, Obstetrics and Gynaecology and Institute of Health Policy, Management and Evaluation, University of Toronto, 600 University Avenue, Toronto, ON, M5G 1X5, Canada. aohlsson@mtsinai.on.ca.

    Search for more papers by this author
  • Vibhuti S Shah

    1. University of Toronto, Department of Paediatrics and Institute of Health Policy, Management and Evaluation, Toronto, ON, Canada
    Search for more papers by this author

Abstract

Background

Maternal colonization with group B streptococcus (GBS) during pregnancy increases the risk of neonatal infection by vertical transmission. Administration of intrapartum antibiotic prophylaxis (IAP) during labor has been associated with a reduction in early onset GBS disease (EOGBSD). However, treating all colonized women during labor exposes a large number of women and infants to possible adverse effects without benefit.

Objectives

To assess the effect of intrapartum antibiotics for maternal Group B haemolytic streptococci (GBS) colonization on mortality from any cause, from GBS infection and from organisms other than GBS.

Search methods

We updated the search of the Cochrane Pregnancy and Childbirth Group's Trials Register on 11 March 2014.

Selection criteria

Randomized trials assessing the impact of maternal IAP on neonatal GBS infections were included.

Data collection and analysis

We independently assessed eligibility and quality of the studies.

Main results

We did not identify any new trials from the updated search so the results remain unchanged as follows.

We included four trials involving 852 women.

Three trials (involving 500 women) evaluating the effects of IAP versus no treatment were included. The use of IAP did not significantly reduce the incidence of all cause mortality, mortality from GBS infection or from infections caused by bacteria other than GBS. The incidence of early GBS infection was reduced with IAP compared to no treatment (risk ratio (RR) 0.17, 95% confidence interval (CI) 0.04 to 0.74, three trials, 488 infants; risk difference -0.04, 95% CI -0.07 to -0.01; number needed to treat to benefit 25, 95% CI 14 to 100, I² 0%). The incidence of LOD or sepsis from organisms other than GBS and puerperal infection was not significantly different between groups.

One trial (involving 352 women) compared intrapartum ampicillin versus penicillin and reported no significant difference in neonatal or maternal outcomes.

We found a high risk of bias for one or more key domains in the study methodology and execution.

Authors' conclusions

Intrapartum antibiotic prophylaxis appeared to reduce EOGBSD, but this result may well be due to bias as we found a high risk of bias for one or more key domains in the study methodology and execution. There is lack of evidence from well designed and conducted trials to recommend IAP to reduce neonatal EOGBSD.

Ideally the effectiveness of IAP to reduce neonatal GBS infections should be studied in adequately sized double-blind controlled trials. The opportunity to conduct such trials has likely been lost, as practice guidelines (albeit without good evidence) have been introduced in many jurisdictions.

Résumé scientifique

Antibiotiques intrapartum en cas de colonisation maternelle connue par le streptocoque du groupe B

Contexte

La colonisation maternelle par le streptocoque du groupe B (SGB) pendant la grossesse augmente le risque d'infection néonatale par transmission verticale. L'administration d'une prophylaxie antibiotique intrapartum (PAI) pendant le travail a été associée à une diminution de la maladie à SGB d'apparition précoce. Toutefois, traiter toutes les femmes présentant une infection pendant le travail expose un grand nombre de femmes et de nouveau-nés à de possibles effets indésirables sans bénéfice.

Objectifs

Évaluer l'effet des antibiotiques intrapartum pour la colonisation maternelle à streptocoque hémolytique du groupe B (SGB) sur la mortalité toutes causes confondues, la mortalité due à une infection au SGB ou à d'autres bactéries que le SGB.

Stratégie de recherche documentaire

Nous avons mis à jour la recherche dans le registre des essais cliniques du groupe Cochrane sur la grossesse et l'accouchement le 11 mars 2014.

Critères de sélection

Les essais randomisés évaluant l'impact de la PAI chez la mère sur les infections néonatales au SGB ont été inclus.

Recueil et analyse des données

Nous avons évalué de façon indépendante l'éligibilité et la qualité des études.

Résultats principaux

Nous n'avons pas identifé de nouveaux essais depuis la mise à jour de la recherche. Les résultats restent donc inchangés.

Nous avons inclus quatre essais portant sur 852 femmes.

Trois essais (impliquant 500 femmes) évaluant les effets de la PAI comparée à aucun traitement ont été inclus dans la revue. L'utilisation de la PAI n'a pas réduit de façon significative l'incidence de la mortalité toutes causes confondues, de la mortalité due à une infection au SGB ou à d'autres infections par des bactéries autres que le SGB. L'incidence des infections précoces au SGB a été réduite avec la PAI par rapport à l'absence de traitement (risque relatif (RR) 0,17, intervalle de confiance (IC) à 95 % 0,04 à 0,74 ; trois essais, 488 nouveau-nés ; différence de risque -0,04, IC à 95 % -0,07 à -0,01 ; nombre de mères à traiter pour obtenir un bénéfice 25, IC à 95 % 14 à 100, I2 0 %). L'incidence de maladie d'apparition tardive ou de sepsis dû à des organismes autres que le SGB et d'infection puerpérale n'était pas significativement différente entre les groupes.

Un essai (impliquant 352 femmes) a comparé l'ampicilline intrapartum à la pénicilline intrapartum et aucune différence significative des résultats cliniques de la mère ou de l'enfant n'a été observée.

Nous avons évalué le risque de biais comme étant élevé pour un ou plusieurs domaines clés dans la méthodologie et la réalisation des études.

Conclusions des auteurs

La prophylaxie antibiotique intrapartum (PAI) a semblé réduire la maladie à SGB d'apparition précoce, mais ce résultat pourrait bien être dû à un biais, car le risque de biais était élevé pour un ou plusieurs domaines clés dans la méthodologie et la conduite de l'étude. Les preuves issues d'essais conçus et menés de façon adéquate sont insuffisantes pour permettre de recommander la PAI pour réduire la maladie à SGB d'apparition précoce chez les nouveau-nés.

Dans l'idéal, l'efficacité de la PAI pour réduire les infections néonatales au SGB devrait être étudiée dans le cadre d'essais contrôlés, en double aveugle et sur un échantillon de taille adéquate. Il n'est probablement plus possible de mener de tels essais car des recommandations pour la pratique clinique (malgré l'absence de données probantes) ont été introduites dans de nombreuses juridictions.

Notes de traduction

Translated by: French Cochrane Centre

Translation supported by: Ministère du Travail, de l'Emploi et de la Santé Français

Resumo

Antibioticoterapia intraparto para gestantes colonizadas por estreptococos do grupo B.

Introdução

A colonização materna por Streptococcus do grupo B (EGB) durante a gravidez aumenta o risco de infecções neonatais através da transmissão vertical. A administração de antibiótico profilático intraparto (AIP) tem sido associada com redução da infecção neonatal precoce por EGB. No entanto, ao tratarmos todas as mulheres colonizadas durante o trabalho de parto, expomos um grande número de mulheres e crianças a possíveis efeitos adversos sem qualquer benefício.

Objetivos

Avaliar os efeitos do uso intraparto de antibióticos em gestantes colonizadas com Streptococcus hemolítico do grupo B (EGB) sobre a mortalidade neonatal por qualquer causa, mortalidade neonatal por infecção por EGB e mortalidade neonatal por infecções causadas por outros microrganismos que não EGB.

Métodos de busca

Em março de 2014, atualizamos a busca na base de dados Cochrane Pregnancy and Childbirth Group's Trials Register.

Critério de seleção

Incluímos na revisão todos os ensaios clínicos randomizados que avaliaram o impacto do uso de antibióticos intraparto sobre as infecções neonatais por EGB.

Coleta dos dados e análises

Avaliamos de forma independente a elegibilidade e a qualidade dos estudos incluídos.

Principais resultados

A busca atualizada não identificou nenhum estudo novo. Portanto os resultados da revisão sistemática permanecem inalterados e são apresentados a seguir.

Incluímos quatro estudos envolvendo 852 mulheres.

Três estudos (envolvendo 500 mulheres) avaliaram os efeitos dos AIP versus nenhum tratamento. O uso de AIP não reduziu significativamente a incidência de mortalidade neonatal por todas causas, mortalidade neonatal por infecção pelo EGB ou mortalidade neonatal por outras infecções bacterianas. O uso de AIP, comparado com nenhum tratamento, reduziu a incidência de infecção neonatal precoce por EGB: razão de risco (RR) 0,17, intervalo de confiança (IC) 95% 0,04 a 0,74, 3 estudos, 488 crianças; diferença de risco -0,04, IC 95% -0,07 a -0,01;número necessário para tratar para ter benefício 25, IC 95% 14 a 100, I² 0%. Não houve diferença significativa entre os grupos na incidência de doença neonatal de início tardio ou de sepse neonatal por outros organismos, nem na taxa de infecção puerperal.

Um estudo (envolvendo 352 mulheres) comparou o uso intraparto de ampicilina versus penicilina e não encontrou diferenças significantes nos desfechos maternos ou neonatais.

Esse ensaio clínico teve um alto risco de viés para um ou mais domínios chaves na metodologia e na execução do estudo.

Conclusão dos autores

O uso de antibiótico profilático intraparto parece reduzir a doença estreptocócica neonatal precoce, porém este resultado pode ser devido a vieses, pois encontramos um alto risco de viés em um ou mais domínios chaves relacionados à metodologia e execução dos estudos. Falta evidência proveniente de estudos bem desenhados e conduzidos para recomendar o uso de antibiótico intraparto para reduzir a doença estreptocócica neonatal precoce.

Seria ideal avaliar a efetividade dos antibióticos intraparto para reduzir as infecções EGB através de ensaios clínicos controlados e duplos-cegos com tamanho amostral adequado. A oportunidade para conduzir esses estudos provavelmente já passou, uma vez que em muitos locais já existem diretrizes recomendando essa prática (apesar da falta de evidências científicas de qualidade sobre isso).

Notas de tradução

Tradução do Centro Cochrane do Brasil (Fernando Takashi Kojima Marques). Contato: tradutores@centrocochranedobrasil.org.br

Plain language summary

Intrapartum antibiotics for known maternal Group B streptococcal colonization

Women, men and children of all ages can be colonized with Group B streptococcus (GBS) bacteria without having any symptoms. Group B streptococcus are particularly found in the gastrointestinal tract, vagina and urethra. This is the situation in both developed and developing countries. About one in 2000 newborn babies have GBS bacterial infections, usually evident as respiratory disease, general sepsis, or meningitis within the first week. The baby contracts the infection from the mother during labor. Giving the mother an antibiotic directly into a vein during labor causes bacterial counts to fall rapidly, which suggests possible benefits but pregnant women need to be screened. Many countries have guidelines on screening for GBS in pregnancy and treatment with antibiotics. Some risk factors for an affected baby are preterm and low birthweight; prolonged labor; prolonged rupture of the membranes (more than 12 hours); severe changes in fetal heart rate during the first stage of labor; and gestational diabetes. Very few of the women in labor who are GBS positive give birth to babies who are infected with GBS and antibiotics can have harmful effects such as severe maternal allergic reactions, increase in drug-resistant organisms and exposure of newborn infants to resistant bacteria, and postnatal maternal and neonatal yeast infections.

This review finds that giving antibiotics is not supported by conclusive evidence. The review identified four trials involving 852 GBS positive women. Three trials, which were more than 20 years old, compared ampicillin or penicillin to no treatment and found no clear differences in newborn deaths although the occurrence of early GBS infection in the newborn was reduced with antibiotics. The antibiotics ampicillin and penicillin were no different from each other in one trial with 352 GBS positive women. All cases of perinatal GBS infections are unlikely to be prevented even if an effective vaccine is developed.

Résumé simplifié

Antibiotiques intrapartum en cas de colonisation maternelle connue par le streptocoque du groupe B

Les femmes, les hommes et les enfants de tous âges peuvent être colonisés par les bactéries streptocoques du groupe B (SGB) sans développer de symptômes. Les streptocoques du groupe B se retrouvent particulièrement dans le tractus gastro-intestinal, le vagin et l'urètre. Ces bactéries sont présentes dans les pays développés aussi bien que dans les pays en développement. Environ 1 nouveau-né sur 2 000 présente une infection bactérienne au SGB, qui se manifeste généralement par une maladie respiratoire, un sepsis généralisé ou une méningite au cours de la première semaine de vie. Le bébé contracte l'infection par la mère au cours de l'accouchement. L'administration d'un antibiotique à la mère directement par voie intraveineuse au cours du travail permet de faire diminuer rapidement le nombre de bactéries, ce qui suggère des bénéfices possibles, mais les femmes enceintes doivent être dépistées. Dans de nombreux pays, il existe des recommandations pour le dépistage des SGB au cours de la grossesse et pour le traitement par antibiotiques. Les facteurs de risque pour la contamination du bébé sont notamment : naissance prématurée et faible poids à la naissance ; travail prolongé ; rupture prolongée des membranes (plus de 12 heures) ; modifications sévères du rythme cardiaque du fœtus pendant le premier stade du travail ; et diabète gestationnel. Parmi les femmes qui accouchent en présence d'une infection au SGB, très peu donnent naissance à un bébé également infecté au SGB, et les antibiotiques peuvent avoir des effets néfastes tels que de graves réactions allergiques de la mère, une augmentation des organismes pharmacorésistants et l'exposition du nouveau-né à des bactéries résistantes, ainsi que des infections post-natales à levures chez la mère et l'enfant.

Cette revue a déterminé que l'administration d'antibiotiques n'est pas corroborée par des données probantes. La revue a identifié quatre essais impliquant 852 femmes infectées au SGB. Trois essais, qui dataient de plus de 20 ans, ont comparé l'ampicilline ou la pénicilline à l'absence de traitement et aucune différence significative n'a été observée dans le nombre de décès des nouveau-nés, bien que le taux d'infection précoce au SGB chez le nouveau-né ait été réduit avec les antibiotiques. Les antibiotiques, ampicilline et pénicilline, étaient identiques l'un à l'autre dans un essai portant sur 352 femmes infectées au SGB. Il est très improbable de parvenir à une prévention de tous les cas d'infection périnatale au SGB, même si un vaccin efficace est développé.

Notes de traduction

Translated by: French Cochrane Centre

Translation supported by: Ministère du Travail, de l'Emploi et de la Santé Français

Резюме на простом языке

Антибиотики во время родов при известной колонизации матери стрептококками группы В

Женщины, мужчины и дети всех возрастов могут быть колонизированы бактериями - стрептококками группы B (СГВ), не имея никаких симптомов. Стрептококки группы В наиболее часто обнаруживают в желудочно-кишечном тракте, влагалище и уретре. Такова ситуация и в развитых, и в развивающихся странах. Примерно у одного из 2000 новорожденных развиваются бактериальные инфекции, вызванные СГВ, обычно проявляющиеся как респираторные заболевания, сепсис или менингит в течение первой недели. Ребенок получает инфекцию от матери во время родов. Введение антибиотика непосредственно в вену матери во время родов вызывает быстрое уменьшение числа бактерий, что свидетельствует о возможной пользе, но [до введения антибиотика] беременные женщины должны быть обследованы (подвергнуты скринингу). Во многих странах существуют клинические рекомендации по скринингу на предмет наличия СГВ во время беременности и лечению антибиотиками. Факторами риска развития у ребенка инфекции, вызванной СГВ, являются преждевременные роды и низкий вес при рождении; затяжные роды; длительный безводный промежуток (интервал времени после разрыва плодных оболочек более 12 часов); значительные изменения частоты сердечных сокращений плода во время первого периода родов; гестационный диабет. Только у небольшого числа женщин, являющихся носителями СГВ, рождаются дети, инфицированные СГВ. В тоже время антибиотики могут иметь вредные эффекты, такие как тяжёлые аллергические реакции у матери, могут увеличить число микроорганизмов, устойчивых к антибиотикам, оказать воздействие на новорожденных и привести к устойчивости бактерий, а также привести к развитию послеродовых грибковых инфекций у матери и новорожденного.

В этом обзоре обнаружено, что введение антибиотиков не поддерживается убедительными доказательствами. При подготовке этого обзора было выявлено четыре испытания с участием 852 женщин, являющихся носителями СГВ (СГВ-положительных). В трех испытаниях, проведенных более 20 лет назад, сравнивали ампициллин или пенициллин с отсутствием лечения и не обнаружили явных различий в смертности новорожденных, хотя возникновение у новорожденных ранней инфекции, вызванной СГВ, сократилось при использовании антибиотиков. Антибиотики - ампициллин и пенициллин не показали различий в одном испытании с участием 352 СГВ-положительных женщин. Маловероятно, что можно предотвратить все случаи перинатальной инфекции, вызванной СГВ, даже если будет разработана эффективная вакцина.

Заметки по переводу

Перевод: Чупрова Наталья Сергеевна. Редактирование: Кораблева Анна Александровна, Юдина Екатерина Викторовна. Координация проекта по переводу на русский язык: Cochrane Russia - Кокрейн Россия (филиал Северного Кокрейновского Центра на базе Казанского федерального университета). По вопросам, связанным с этим переводом, пожалуйста, обращайтесь к нам по адресу: cochrane.russia.kpfu@gmail.com; cochranerussia@kpfu.ru

Laički sažetak

Davanje antibiotika za vrijeme poroda u slučaju kolonizacije majke streptokokima grupe B

Žene, muškarci i djeca svih dobi mogu biti kolonizirani streptokoknim bakterijama grupe B (GBS) bez ikakvih simptoma. Najviše streptokoka grupe B može se naći u probavnom sustavu, rodnici i mokraćovodu. Ovo se odnosi na razvijene zemlje i zemlje u razvoju Otprilike 1 od 2000 novorođenih djece ima GBS bakterijsku infekciju, koja se obično uočava unutar prvog tjedna kao respiratorna bolest, sustavna sepsa ili meningitis. Dijete skupi infekciju od majke tijekom poroda. Davanjem antibiotika majci direktno u venu tijekom porođaja izaziva nagli pad broja bakterija, što sugerira moguću prednost, ali kod trudnica se isto treba prethodno provesti probir. Mnoge zemlje imaju smjernice za probir u slučaju GBS-a u trudnoći i terapiju antibioticima. Neki rizični čimbenici za zahvaćeno dijete su prerani porod i niska porođajna težina; produžena trudnoća; „otvorenost“ dulja od 12 sati; ozbiljne promjene pulsa djeteta u prvim satima porođaja; i gestacijski dijabetes. Vrlo malo GBS-pozitivnih trudnica rađaju djecu inficiranu GBS bakterijama i antibiotici mogu imati štetne učinke kao što su alergijske reakcije majke, povećana bakterijska rezistencija na lijekove i izloženost novorođenčeta rezistentnim bakterijama te postnatalna infekcija majke i novorođenčeta gljivicama.

Zaključak Cochrane sustavnog pregleda je da davanje antibiotika nije utemeljeno na čvrstim dokazima. Sustavni pregled uključuje četiri studije u kojima su sudjelovale 852 GBS-pozitivne žene. Tri studije, više od 20 godina stare, uspoređivale su primjenu ampicilina ili penicilina u odnosu na nikakvo liječenje i nisu pronađene jasne razlike u smrtnosti novorođenčadi, iako je pojavnost ranih GBS infekcija u novorođenčadi smanjena upotrebom antibiotika. Antibiotici ampicilin i penicilin nisu pokazali međusobnu razliku u djelovanju u jednom istraživanju s 352 GBS-pozitivne žene. Nije vjerojatno da se svi slučajevi perinatalne GBS infekcije mogu spriječiti čak ni u slučaju razvijanja učinkovitog cjepiva.

Bilješke prijevoda

Cochrane Hrvatska
Prevela: Ivana Miošić
Ovaj sažetak preveden je u okviru volonterskog projekta prevođenja Cochrane sažetaka. Uključite se u projekt i pomozite nam u prevođenju brojnih preostalih Cochrane sažetaka koji su još uvijek dostupni samo na engleskom jeziku. Kontakt: cochrane_croatia@mefst.hr

Resumo para leigos

Antibióticos intraparto para mães colonizadas por Estreptococos B.

Homens, mulheres e crianças de todas as idades podem ser colonizados por bactérias do tipo Streptococcus do grupo B (EGB), sem que isso cause qualquer sintoma. Os EGB geralmente vivem no trato gastrintestinal, vagina e uretra. Isso ocorre tanto nas pessoas que moram nos países em desenvolvimento quanto nos países desenvolvidos. Cerca de 1 em cada 2.000 recém-nascidos terá uma infecção por EGB, que pode se manifestar durante a primeira semana de vida como um problema respiratório, sepse (infecção generalizada) ou meningite. Os bebês contraem essa infecção de suas mães, durante o trabalho de parto. Se a mãe com EGB receber injeções intravenosas com antibióticos durante o trabalho de parto, haverá uma queda rápida do número de bactérias. Isso seria potencialmente benéfico, mas as gestantes precisam fazer exames para saber se têm ou não o EGB (“rastreamento”). Muitos países têm diretrizes sobre o rastreamento do EGB na gravidez e sobre o tratamento dessas mulheres com antibióticos. Alguns fatores de risco para o bebê ter problemas devido ao EGB são: nascer prematuro, ser de baixo peso (< 2,5 kg ao nascer), trabalho de parto prolongado, tempo prolongado de bolsa rota (mais de 12 horas); grandes alterações nos batimentos cardíacos fetais durante o primeiro estágio do parto e diabetes gestacional. Porém, apenas uma minoria das mulheres em trabalho de parto e com exame positivo para EGB terá um bebê infectado com EGB. Por outro lado, o uso de antibióticos pode causar efeitos prejudiciais, como reações alérgicas maternas graves, surgimento de organismos resistentes aos antibióticos, exposição dos recém-nascidos a bactérias resistentes e surgimento de infecções maternas e neonatais por fungos.

Esta revisão não encontrou evidências conclusivas que apoiam o uso de antibióticos nesses casos. A revisão identificou 4 ensaios clínicos randomizados envolvendo 852 mulheres em trabalho de parto com exames positivos para EGB. Três estudos, feitos há mais de 20 anos atrás, compararam ampicilina ou penicilina versus não dar nenhum tratamento. Esses estudos não encontraram diferenças claras no risco de os bebês morrerem, embora a ocorrência de infecção neonatal precoce por EGB foi menor no grupo dos antibióticos. Um estudo envolvendo 352 mulheres com exame positivo para EGB não encontrou diferenças entre usar ampicilina versus penicilina. Mesmo que uma vacina efetiva seja desenvolvida algum dia, é pouco provável que todos os casos de infecção perinatal por EGB sejam prevenidos.

Notas de tradução

Tradução do Centro Cochrane do Brasil (Fernando Takashi Kojima Marques). Contato: tradutores@centrocochranedobrasil.org.br

Ringkasan bahasa mudah

Antibiotik intrapartum untuk ibu yang dikenalpasti dengan pengkolonian streptococcal Kumpulan B

Wanita, lelaki dan kanak-kanak daripada semua peringkat umur boleh dikolonisasi oleh streptococcus Kumpulan B (GBS) bakteria tanpa mempunyai apa-apa gejala. Streptococcus Kumpulan B terutamanya didapati di dalam saluran gastrousus, vagina dan uretra. Ini adalah situasi di kedua-dua negara maju dan membangun. Kira-kira satu dalam 2000 bayi yang baru lahir mempunyai jangkitan bakteria GBS, biasanya terbukti sebagai penyakit pernafasan, sepsis umum, atau meningitis dalam tempoh minggu pertama. Bayi itu mendapat jangkitan daripada ibu semasa proses bersalin. Memberi ibu antibiotik terus ke dalam vena semasa proses bersalin menyebabkan bilangan bakteria kurang dengan cepat, yang menunjukkan faedah tetapi wanita hamil perlu diperiksa. Banyak negara mempunyai garis panduan mengenai pengujian GBS masa mengandung dan rawatan dengan antibiotik. Beberapa faktor risiko untuk bayi yang terlibat ialah pramatang dan berat kelahiran yang rendah; proses bersalin yang berpanjangan; pecah membran yang berpanjangan (lebih daripada 12 jam); perubahan yang teruk dalam kadar jantung janin semasa peringkat pertama proses bersalin; dan kencing manis semasa mengandung. Hanya sedikit wanita dalam proses bersalin mempunyai jangkitan GBS melahirkan bayi yang dijangkiti GBS dan antibiotik boleh memberi kesan yang berbahaya seperti reaksi alahan yang teruk pada ibu , peningkatan dalam organisma menentang ubat dan pendedahan bayi yang baru lahir kepada bakteria penentang, dan jangkitan yis selepas bersalin pada ibu dan neonatal.

Ulasan ini mendapati bahawa pemberian antibiotik tidak disokong oleh bukti kukuh. Ulasan ini mengenalpasti empat kajian yang melibatkan 852 wanita yang GBS positif. Tiga kajian, yang berusia lebih daripada 20 tahun, berbanding ampicillin atau penisilin untuk tiada rawatan dan mendapati tiada perbezaan yang jelas dalam kematian bayi baru lahir walaupun berlakunya jangkitan awal GBS pada bayi yang baru lahir telah dikurangkan dengan antibiotik. Antibiotik ampicillin dan penisilin tidak berbeza antara satu sama lain dalam satu kajian dengan 352 wanita yang GBS positif. Semua kes-kes jangkitan GBS perinatal tidak mungkin dicegahkan walaupun dibuat vaksin yang berkesan.

Catatan terjemahan

Diterjemahkan oleh Nila Pillai (Penang Medical College). Disunting oleh Sharifah Halimah Jaafar (Regency Specialist Hosp Sdn Bhd, JB) Untuk sebarang pertanyaan berkaitan terjemahan ini sila hubungi nila@pmc.edu.my

Background

Description of the condition

Historical background

The etiology of neonatal sepsis varies with geographical location and changes over time (Nyhan 1958; Ohlsson 1986). Although, asymptomatic vaginal carriage of Group B haemolytic streptococci (GBS) was described in 1935 (Lancefield 1935), the first report of GBS sepsis in a neonate did not appear until 1964 (Eickhoff 1964). Since the 1970s, GBS is one of the most common causes of neonatal infectious morbidity and mortality in the US (McCracken 1973), Canada (Allardice 1982), UK (Lloyd 1976), other European countries (Bergqvist 1974; Cayeux 1972; Schröder 1979; Vesikari 1989) and Australia (Fliegner 1990).

GBS causes significant maternal and perinatal morbidity (Institute of Medicine 1985), asymptomatic bacteriuria in pregnancy (Hastings 1986), as well as urinary tract and other infections in the adult non-pregnant population (Ho 2006). 

In 1990, the Group B Strep Association, an advocacy group, was formed by parents (Group B Strep Association 2008). Broad media coverage followed and in 1992 the first guidelines for GBS prevention were published in the US (AAP 1992; ACOG 1992).

Maternal colonization with Group B streptococci and transmission

The gastrointestinal tract, vagina and urethra serve as reservoirs for GBS. An overview in 1992, reported maternal colonization rates from 19 studies (1980 to 1991) ranging from 1.6% in Israel, to 28% in England. The transmission rate for GBS colonization from mother to infant varied from 35% in England, to 69% in Brazil, with the incidence of early onset GBS disease varying from 0.2 per 1000 live births in Israel, to 5.0 per 1000 live births in the US (Ohlsson 1992). Stoll (Stoll 1998) reviewed maternal GBS colonization rates in developing countries where the overall colonization rate was 12.7% (34 studies with 7730 women). Among only those studies in which the methods to ascertain colonization was adequate, the colonization rate was 17.8% (675 of 3801 women). The colonization rates were; in the Middle East/North Africa, 22%; Asia/Pacific, 19%; Sub-Saharan Africa, 19%; India/Pakistan, 12%; and the Americas, 14%, respectively. The authors concluded that the range of colonization reported from developing countries is similar to that identified in population studies in the United States (Stoll 1998). There is likely an increasing rate of GBS neonatal infections in developing countries (Osrin 2004).

A systematic review on the prevalence of maternal GBS colonization in European countries, identified 21 studies published between 1996 and 2006, that reported on 24,093 women (Barcaite 2008). Among the studies, GBS vaginal colonization rates ranged from 6.5% to 36%, with one-third of the studies reporting rates of 20% or greater. The carriage rates varied with Eastern Europe 19.7% to 29.3%, Western Europe 11% to 21%, Scandinavia 24.3% to 36%, and Southern Europe 6.5% to 32% (Barcaite 2008).

Early onset GBS neonatal disease (EOD)

Early onset disease (EOD) occurs, by definition, during the first seven days of life, with the vast majority of cases (approximately 90%) presenting during the first 24 hours of life (Garland 1991; Yagupsky 1991). Neonates with EOD present with respiratory disease (54%), sepsis without focus (27%) and meningitis (15%) (Yagupsky 1991). Risk factors for EOD include: GBS bacteriuria during pregnancy; gestational age less than 37 weeks (Håkansson 2006); previous infant with invasive GBS disease (Schrag 2002); preterm labor/delivery (Dillon 1987; Garland 1991; Yagupsky 1991); birthweight less than 2500 g (Baker 1973; Dillon 1987; Schuchat 1990; Yagupsky 1991); prolonged labor (Dillon 1987); prelabor rupture of the membranes (Dillon 1987; Garland 1991); prolonged membrane rupture (more than 12 hours) (Baker 1973; Garland 1991); black race (Schuchat 1990); teenage mother (Schuchat 1990); previous miscarriage (Schuchat 1990); maternal infection including chorioamnionitis (Dillon 1987); bacteraemia (Dillon 1987); sepsis (Garland 1991); urinary tract infection (Dillon 1987; Wood 1981); maternal fever in labor (Dillon 1987); gestational age more than 42 weeks (Christensen 1983); severe changes in fetal heart rate during the first stage of labor (Christensen 1983) and gestational diabetes (Håkansson 2006).

A number of infants born to GBS negative mothers, but infected with GBS at birth, have been reported (Hamada 2008; Mereghetti 2007). This may be due to false negative tests in the mother or to a change in GBS colonization status between the time the test was performed and the time when the mother gave birth. False negative culture results may be due to inappropriate sampling methods, the choice of media the sample was plated on, or the method of transporting the sample to the laboratory. In a study from two tertiary perinatal centres in Canada, antepartum or intrapartum predisposing factors for neonatal GBS infection were recognized in 62% of cases (Hamada 2008). In the same study, all infants born at term survived, but the mortality rate for preterm neonates with early symptomatic disease (and who presented with shock and thrombocytopenia) was 6% (Hamada 2008).

Late onset GBS disease (LOD)

Late onset disease (LOD) occurs beyond seven days of life and can develop up to three months of age (Yagupsky 1991). Risk factors for late onset disease include non-white race and preterm birth (Yagupsky 1991).  Neonates with LOD present with sepsis (46%), meningitis (37%), urinary tract infection (7%), osteoarthritis (6%), respiratory disease (4%) and cellulitis (4%) (Yagupsky 1991).

The organism and its detection

Streptococci are Gram-positive cocci that occur in pairs or chains (Lancefield 1933).They are divided into three groups by the type of haemolysis on blood agar plates: β-haemolytic, α-haemolytic or ϒ-haemolytic streptococci. GBS is a β-haemolytic streptococcus. Serologic grouping is based on the polysaccharide capsule in GBS. GBS strains isolated in the 1970s were serotypes I, II, and III; but new serotypes (IV, V, VI, VIII) have since emerged. A proposed IX serotype was reported in 2007 (Slotved 2007).

Polymerase chain reaction (PCR) and optical immunoassay are candidates for rapid near patient intrapartum GBS testing to determine whether women in labor are colonized with GBS (Abdelazim 2013; Gavino 2007; Honest 2006; Håkansson 2014; Poncelet-Jasserand 2013).

Burden of illness

In the US in the early 1980s, the total number of maternal GBS infections was estimated to be 47,885 and the number of neonatal cases 11,074 (7198 with EOD). The total cost of disease burden was estimated at $726.8 million (US) per year. The incidence of EOD in the UK, in the absence of systematic screening or widespread use of intrapartum antibiotics, was 0.5 per 1000 births with vaginal carriage rates comparable to those in the US (RCOG 2003). In a population-based cohort in Sweden (1997 to 2001), the incidence of EOD was 0.4 per 1000 live births with the total burden of illness of early onset GBS morbidity approximately three times higher (Håkansson 2006). A considerable number of infants are diagnosed with probable early-onset GBS neonatal sepsis possibly as a result of maternal treatment intrapartum that inhibits growth in blood and cerebrospinal fluid but does not alleviate clinical symptoms, signs, nor death (Carbonell-Estrany 2008).

Preventive measures

Several GBS vaccine candidates have been developed against the nine currently identified GBS serotypes (Johri 2006) and a type III conjugate vaccine has been found to be safe and immunogenic in pregnant women (Johri 2006). Further advances in GBS vaccine development are likely through combining genomics with newer proteomic technologies (Johri 2006). The projected health benefits of maternal GBS vaccination in the era of chemoprophylaxis could be considerable (Sinha 2005). Immunization of pregnant women against influenza and pertussis is becoming acceptable in the US (Baker 2013). In 2013 Dr Carol Baker wrote: "This changing landscape provides the perfect opportunity to introduce a glycoconjugate GBS vaccine for use in pregnant women. Like influenza, a GBS maternal immunization program will reduce some adverse pregnancy outcomes and intrapartum infections in the mother in addition to early- and late-onset infections in infants" (Baker 2013).

Chlorhexidine vaginal treatment, with or without neonatal wash, reduced GBS bacterial load but showed no impact on EOD (Stade 2004). A systematic review including non-randomized studies suggests that important reductions in maternal and neonatal sepsis (not restricted to GBS as a cause) in developing countries may be achieved using this method (Goldenberg 2006).

Induction of labor with intravenous (IV) oxytocin may be preferable for GBS positive women with prelabor rupture of membranes at term as infections are reduced (Hannah 1997).

To date, the most commonly used prevention intervention is intrapartum chemoprophylaxis with antibiotics to mothers with known GBS colonization. This review will focus on this aspect of prevention of GBS neonatal disease. To date, four approaches have been recommended for the prevention of neonatal GBS infections: a risk-based strategy; a screening- (vaginal/rectal GBS cultures) based strategy; a combined risk/screening-based strategy and a combined risk/screening-based strategy using the PCR test (Akker van Marle 2005). This topic has been previously reviewed (Shah 2001), but deserves to be updated in a separate review using Cochrane guidelines.

Guidelines for prevention of GBS neonatal infections

In 1992, the first guidelines for GBS prevention were published in the US (AAP 1992; ACOG 1992). Since then numerous guidelines with different recommendations have been published by various organizations (AAP 1997; ACOG 1996; CDC 1996; CDC 2002; CDC 2010 RCOG 2003; Shah 2001; SOGC 1994; SOGC 1997; SOGC 2004; Money 2013 for SOGC).

Although these current guidelines are based on studies of poor quality (Ohlsson 1994), there seems to be a temporary association between the introduction of guidelines and a decline in the GBS EOD rate (CDC 2005; CDC 2007; Schrag 2002). The incidence of invasive early-onset GBS disease decreased from 1.8 cases/1000 live births in the early 1990s to 0.26 cases/1000 live births in 2010 (Schrag 2013). However, there has been no reduction in LOD GBS disease in infants (CDC 2007; Schrag 2013). Mortality has decreased. The same literature has been interpreted differently by different professional organizations. All cases of EOD cannot be prevented.

Description of the intervention

In 1976, chemoprophylaxis was first proposed for reducing maternal GBS colonization in labor to reduce neonatal disease (Ablow 1976). Non-randomized studies showed that intravenous ampicillin given during labor to GBS positive women could significantly reduce neonatal GBS colonization, and a non-significant reduction in GBS neonatal invasive disease was reported (Allardice 1982; Yow 1979). Currently in the US, penicillin is the drug of choice for intrapartum prophylaxis given every four hours intravenously until the baby is born (CDC 2002).

Adverse effects

Severe allergic reaction to antibiotics has been reported among mothers giving birth (Berthier 2007; Jao 2006). The incidence of postnatal maternal and neonatal yeast infections may increase with the use of intrapartum antibiotics (Dinsmoor 2005). There is a growing concern about antibiotic resistance to erythromycin (3.8% to 21.2%) and clindamycin (2.7% to 20%) (Barcaite 2008). Intrapartum antibiotic prophylaxis (IAP) may increase exposure of neonates to ampicillin resistant Enterobacteriaceae (Edwards 2002b).

How the intervention might work

Vaginal GBS colony counts fall rapidly after intrapartum penicillin-G administration which may, to some degree, explain the possible effectiveness of chemoprophylaxis (McNanley 2007).

Why it is important to do this review

It is important to know if intrapartum antibiotics do more good than harm in trying to reduce mortality and morbidity from neonatal GBS infections. Most women colonized with GBS are asymptomatic, so screening is necessary if these women are to be identified. However, of the women in labor who are GBS positive, very few will give birth to babies who are infected with GBS. Hence, giving IV antibiotics to all women in labor who are GBS positive will put a large number of women and babies at risk of adverse effects unnecessarily. These adverse effects include potentially fatal anaphylaxis, increase in drug-resistant organisms and the medicalization of labor and the neonatal period (RCOG 2003).

A critical review of randomized controlled trials of intrapartum chemoprophylaxis of perinatal GBS infections identified numerous methodological flaws (Ohlsson 1994). Whether we are using the optimal strategy for GBS management in pregnancy has been questioned (Yudin 2006). A Cochrane review adopting high-quality methodology is, therefore, justified.

Objectives

Primary objective

  • To assess the effect of intrapartum antibiotics for maternal Group B haemolytic streptococci (GBS) colonization on mortality from any cause, from GBS infection and from organisms other than GBS.

Secondary objectives

  • To assess the effect of intrapartum antibiotics for maternal GBS colonization on neonatal morbidity from early onset neonatal GBS infection (as defined under outcomes below).

  • To assess the effect of intrapartum antibiotics for maternal GBS colonization on probable early (postnatal age less than seven days) neonatal GBS infection.

  • To assess the effect of intrapartum antibiotics for maternal GBS colonization on late onset GBS sepsis (sepsis due to GBS in an infant at least seven days old).

  • To assess the effect of intrapartum antibiotics for maternal GBS colonization on long-term child development (motor and cognitive).

  • To assess the effect of intrapartum antibiotics for maternal GBS colonization on maternal outcomes including; chorioamnionitis, sepsis, urinary tract infection, hospital stay and allergic reactions to antibiotics.

Methods

Criteria for considering studies for this review

Types of studies

Randomized or quasi-randomized trials assessing the impact of intrapartum antibiotics on neonatal GBS colonization and infections.

Types of participants

Mothers known to be colonized with GBS in the vaginal/intestinal tract and/or the urinary tract at any time during the pregnancy and at < 35 weeks and > 35 weeks' gestation. Mothers giving birth vaginally or by caesarean section were included.

Types of interventions

Intervention

Intrapartum administration of antibiotics to mothers known to be GBS positive (by culture or rapid detection test) from a vaginal or rectal swab (or both), or from urine.

Comparison

Placebo or no treatment to mothers known to be GBS positive (by culture or rapid detection test) from a vaginal or rectal swab (or both), or from urine. In a deviation from our protocol we decided to include studies that compared the effectiveness of one antibiotic versus another.

Types of outcome measures

Primary outcomes
Neonatal

All cause mortality

Mortality from early (postnatal age less than seven days) onset culture positive neonatal GBS infection including one or more of the following conditions.

a) Sepsis - defined as symptoms and signs of sepsis and a bacterial culture positive for GBS (obtained in a sterile manner from normally sterile body fluids such as blood, cerebrospinal fluid or urine, or culture from internal organs at autopsy).

b) Pneumonia in the neonate (postnatal age less than seven days) - defined as symptoms and signs and radiographic findings consistent with pneumonia and positive culture for GBS (obtained from tracheal aspirate or by culture of lung tissue at autopsy).

Mortality from infections (as per a and b above) caused by bacteria other than GBS.

Secondary outcomes
Neonatal

Early (postnatal age less than seven days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and positive GBS bacterial cultures (from normally sterile body fluids obtained from the neonate).

Probable early (postnatal age less than seven days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and bacterial cultures from normally sterile body fluids obtained from the neonate that were negative for GBS.

Late onset GBS sepsis - sepsis due to GBS in an infant at least seven days old.

Neonatal sepsis, meningitis, urinary tract infection or pneumonia due to bacterial organisms other than GBS, to drug-resistant bacteria and to fungi.

Initial hospital stay.

Long-term follow-up assessments of infants, at an age of 12 months or later, using a validated assessment tool for motor or cognitive functions (or both).

Maternal
  1. Chorioamnionitis defined as a temperature of more than 38oC on two occasions in labor with uterine tenderness or chorioamnionitis diagnosed on placental histopathology.

  2. Sepsis in the peri/postpartum period.

  3. Urinary tract infection with any bacteria in the peri/postpartum period.

  4. Hospital stay.

  5. Allergic reactions to antibiotics.

  6. Puerperal infection defined according to clinical criteria - uterine tenderness, uterine subinvolution and fever in the absence of any other known cause of infection in the postpartum period. We included this outcome in deviation from our protocol as it was reported in one study.

The methods section of this review is based on a standard template used by the Cochrane Pregnancy and Childbirth Group.

Search methods for identification of studies

Electronic searches

We contacted the Trials Search Co-ordinator to search the Cochrane Pregnancy and Childbirth Group’s Trials Register (11 March 2014). 

The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co-ordinator and contains trials identified from: 

  1. monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);

  2. weekly searches of MEDLINE;

  3. weekly searches of Embase;

  4. handsearches of 30 journals and the proceedings of major conferences;

  5. weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Details of the search strategies for CENTRAL, MEDLINE and Embase, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group

Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co-ordinator searches the register for each review using the topic list rather than keywords. 

We did not apply any language restrictions. 

No new studies were included for this update (2014). The following methods were used in the previous update.

Data collection and analysis

Both review authors assessed the validity of each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). The review authors were not blinded to authors, institution or journal of publication for articles considered for inclusion.

All abstracts and published full reports identified as potentially relevant by the literature search were assessed by both review authors for inclusion in the review. Each review author independently extracted data using a pre-designed data extraction form, and then compared results and resolved differences. Arne Ohlsson (AO) entered data into RevMan (RevMan 2008) and Vibhuti Shah (VS) cross checked the printouts against her own data extraction forms. Differences were resolved by consensus.

In future updates of this review, where studies are identified as abstracts, the primary authors will be contacted to ascertain whether a full publication is available, if the full paper was not identified in an electronic database. Information from the primary author will be sought if the published article does not provide adequate information for the review.

Selection of studies

All potential studies identified from the search strategy were assessed for eligibility for inclusion independently by the two review authors (AO and VS). We resolved any disagreement through discussion or consulted a third person as an arbitrator.

Data extraction and management

We designed a form to extract data. For eligible studies, we extracted the data using the agreed form. We resolved discrepancies through discussion or, if required, we would have consulted a third person. Data were entered into Review Manager software (RevMan 2008) and checked for accuracy.

When information regarding any of the above was unclear, we attempted to contact authors of the original reports to provide further details.

Assessment of risk of bias in included studies

We independently assessed the risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Any disagreement was resolved by discussion or by involving a third assessor as an arbitrator.

(1) Random sequence generation (checking for possible selection bias)

We describe for each included study the methods used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.

We assessed the methods 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 describe for each included study the method used to conceal the allocation sequence in sufficient detail 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 risk of bias (e.g. telephone or central randomization; 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 describe for each included study all the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We provide information on whether the intended blinding was effective. Where blinding was not possible, we assessed whether the lack of blinding was likely to have introduced bias. Blinding was assessed separately for different outcomes or classes of outcomes.

We assessed the methods as:

  • low risk of bias, high risk of bias, unclear risk of bias for participants;

  • low risk of bias, high risk of bias or unclear risk of bias for personnel.

(3.2) Blinding of outcome assessment (checking for possible detection bias)

We describe for each included study the methods used, if any, to blind outcome assessors from knowledge of which intervention a participant received. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed the methods used to blind outcome assessment as:

  • low, high or unclear risk of bias.

(4) Incomplete outcome data (checking for possible attrition bias through withdrawals, dropouts, protocol deviations)

We describe for each included study and for each outcome or class of outcomes the completeness of data including attrition and exclusions from the analysis. We state whether attrition and exclusions were reported, the numbers included in the analysis at each stage (compared with the total randomized 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 was supplied by the trial authors, we re-included missing data in the analyses which we undertook.

(5) Selective reporting bias

We describe for each included study how the possibility of selective outcome reporting bias was examined by us and what we found.

We assessed the methods as:

  • low risk of bias (where it is clear that all of the study’s prespecified outcomes and all expected outcomes of interest to the review have been reported);

  • high risk of bias (where not all the study’s prespecified outcomes have been reported; one or more reported primary outcomes were not prespecified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);

  • unclear risk of bias.

(6) Other sources of bias

We describe for each included study any important concerns we have about other possible sources of bias.

We assessed whether each study was free of other problems that could put it at risk of bias:

  • low risk of other bias;

  • high risk of other bias;

  • unclear risk of other bias.

(7) Overall risk of bias

We make explicit judgements about whether studies were at high risk of bias, according to the criteria given in the Handbook (Higgins 2011). With reference to (1) to (6) above, we assessed the likely magnitude and direction of the bias and whether we considered it is 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 present results as summary risk ratio (RR) with 95% confidence intervals (CI). When appropriate we present risk difference (RD) with 95% CI. If the RD was found to be statistically significant, we calculated the number needed to treat to benefit (NNTB) and in the case of a harmful effects (if they had been identified) we would have calculated the number needed to treat to harm (NNTH).  

Continuous data

For continuous data, we used the mean difference (MD) if outcomes were measured in the same way between trials. Standardized mean difference (SMD) will be used in future updates of this review to combine trials that measured the same outcome, but used different methods.  

Unit of analysis issues

Cluster-randomized trials

No cluster-randomized trials were identified.

Cross-over trials

Cross-over trials were not considered appropriate for this review topic.

Dealing with missing data

For included studies, we noted levels of attrition. The impact of including studies with high levels of missing data in the overall assessment of treatment effect was explored by using sensitivity analyses.

For all outcomes analyses were carried out, as far as possible, on an intention-to-treat basis, i.e. we attempted to include all participants randomized to each group in the analyses. The denominator for each outcome in each trial is the number randomized minus any participants whose outcomes were known to be missing.

Assessment of heterogeneity

We used the I² statistic to measure heterogeneity among the trials in each analysis (Higgins 2003). If we identified substantial (moderate or high) heterogeneity, we explored it by prespecified sensitivity analyses. We used the adjectives of low (25%), moderate (50%) and high (75%) assigned to values for I² by Higgins (Higgins 2003).

Assessment of reporting biases

Where we suspected reporting bias (see 'Selective reporting bias' above), we attempted to contact study authors asking them to provide missing outcome data. Where this was not possible, and the missing data were thought to introduce serious bias, the impact of including such studies in the overall assessment of results was explored by a sensitivity analysis

Data synthesis

We carried out statistical analyses using the Review Manager software (RevMan 2008). We used the fixed-effect inverse variance meta-analysis for combining data where trials were examining the same intervention, and the trials’ populations and methods were judged sufficiently similar. In future updates if we identify clinical or methodological heterogeneity between studies sufficient to suggest that treatment effects may differ between trials, we will use the random-effects meta-analysis.

In future updates, if substantial heterogeneity is identified in a fixed-effect meta-analysis this will be noted and the analysis repeated using a random-effects method, as a sensitivity analysis.

Subgroup analysis and investigation of heterogeneity

We planned to perform subgroup analyses when possible based on:

  1. timing of bacterial cultures in the mother at < 35 weeks' post menstrual age verses bacterial cultures in the mother at 35 weeks' gestation or more;

  2. GBS colonization ascertained by bacterial culture (from the vagina or the rectum, or both) versus GBS colonization ascertained by a rapid screening test.  

We planned to perform subgroup analyses only for the primary outcomes (infant mortality from any cause, infant mortality from GBS infection and infant mortality from infections other than GBS). As only one study reported on these outcomes (Boyer 1986), subgroup analyses were not possible based on the two criteria listed above.

We did not conduct subgroup analyses based on heterogeneity, as for many outcomes there was only one study included and for those outcomes with more than one study there was no important heterogeneity (i.e. I² values were less than 25 %).

In future updates, we will assess subgroup differences by interaction tests available within RevMan (RevMan 2012). We will report the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² value.

Sensitivity analysis

As we have noted, discrepancies between numbers enrolled in trials as reported in abstracts and full text reports of the same trial (Ohlsson 1999), we planned to perform sensitivity analyses excluding abstracts. As no abstracts were included we did not conduct these planned sensitivity analyses.

No additional sensitivity analyses were planned a priori.

Results

Description of studies

The search conducted in December 2008 resulted in 24 references to 13 studies (several studies were reported on at different stages during the study period and/or different aspects of the study were reported on in different publications). An updated search on 10 November 2012 did not identify any additional studies. The updated search on March 11 2014 did not identify any additional trials. Of the 13 studies identified from the 2008 search, nine were excluded as two studied the effect of chlorhexidine wash (Dykes 1987; Facchinetti 2002); two were not true randomized controlled trials (Morales 1986; Sáez-Llorens 1995); three (Belady 1996; Easmon 1983; Gibbs 1996) did not report on outcomes of interest for this review; and in two studies (Merenstein 1980; Pinette 2005) the treatment with antibiotics started outside of the intrapartum period. One study identified from our personal files (Håkansson 2014) was excluded for this update (2014) because it did not compare intrapartum antibiotic prophylaxis (IAP) with placebo. For further details see Characteristics of excluded studies.

Only four studies involving 852 women) qualified for inclusion in this review. Two studies (Boyer 1986; Matorras 1990) compared ampicillin with no treatment for GBS intrapartum prophylaxis and one study compared penicillin with no treatment (Tuppurainen 1989). None of these three studies used a placebo treatment in the control group. One study compared the effects of ampicillin versus penicillin G (Edwards 2002a). For further details please see Characteristics of included studies.

Risk of bias in included studies

See Figure 1 and Figure 2 for a summary of 'Risk of bias' assessments.

Figure 1.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 2.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Overall, the quality of these four studies was poor and the risk of bias high as defined by Higgins 2011; there was plausible bias that seriously weakens confidence in the results; there was high risk of bias for one or more key domains; and the proportion of information from all studies was at high risk of bias sufficient to affect the interpretation of the results.

No study reported on a pre-set sample size. The total number of women enrolled in these three studies comparing intrapartum prophylaxis with no treatment was 500 (239 in the treatment group and 261 in the control group) (Boyer 1986; Matorras 1990; Tuppurainen 1989). These three studies were published between 24 and 28 years ago. In the study by Tuppurainen (Tuppurainen 1989) there was imbalance in allocation of participants to the two groups; 44% of the participants were allocated to the intervention group and 56% to the control group. This represents a large deviation from the expected ratio of 50:50 possibly due to mothers dropping out of the intervention group but not the control group.

We used judgement in assessing the neonatal and maternal outcomes as definitions of outcomes were often not clear. In the only included study (involving 352 women) that compared ampicillin with penicillin for GBS prophylaxis (Edwards 2002a), the authors did not state in which group the only neonatal infection occurred and they do not provide a definition for their outcomes of suspected infection and chorioamnionitis. The risk of bias was lower in this study (Edwards 2002a) compared to the three studies that assessed the effectiveness of intrapartum antibiotics versus no treatment (Boyer 1986; Matorras 1990; Tuppurainen 1989).

Allocation

In two studies a random sequence was appropriately generated (Boyer 1986; Edwards 2002a); in the other two included studies there was not enough information provided to allow for a judgment (Matorras 1990; Tuppurainen 1989). There was high risk for bias in two of the four included studies for allocation (selection) bias (Boyer 1986; Tuppurainen 1989). The risk was unclear for Matorras 1990) and was low for the study by Edwards 2002a.

Blinding

No placebo was used in the three studies comparing one antibiotic versus no treatment (Boyer 1986; Matorras 1990; Tuppurainen 1989). Consequently, patients, care providers and researchers in these three studies were not blinded to group assignment. In the study by Edwards 2002a the risk for performance and detection bias was unclear as the authors did not state whether the women and the care-givers were aware of what antibiotic was administered.

Incomplete outcome data

In the Boyer 1986 study, women who developed intrapartum fever were excluded as were their offspring from the analyses, which is remarkable in a study that attempted to prevent infections. In 11% of the women randomized, the maternal and neonatal outcomes were not reported. In the studies by Edwards 2002a, Matorras 1990 and Tuppurainen 1989 all randomized women were accounted for.

Selective reporting

Two studies reported on results after different numbers of women had been enrolled (Boyer 1986; Tuppurainen 1989). In one study (Boyer 1986) the authors clearly waited for an additional neonatal outcome in the control group (Gotoff 1984) and when this outcome occurred they published their final report (Boyer 1986). In addition, they changed their level of significance from a two-tailed to an one-tailed statistical test and thus reached statistical significance from a previous report of the same ongoing study. There was no apparent reporting bias in the study by Edwards 2002a and our judgment was unclear risk for the study by Matorras 1990.

Other potential sources of bias

As indicated in the 'Risk of bias' tables, we raised serious concerns for other areas of bias in the trials by Boyer 1986 and Tuppurainen 1989 (see that table for details).The study by Edwards 2002a appeared free of other forms of bias. The risk was unclear for the study by Matorras 1990

Effects of interventions

Four trials involving 852 women were included. Three trials (n = 500) compared ampicillin or penicillin versus no treatment (Boyer 1986; Matorras 1990; Tuppurainen 1989) and one trial enrolling 352 women compared ampicillin with penicillin for GBS positive women. Group B streptococcus carriage was ascertained by vaginal/rectal cultures in three studies (Boyer 1986; Edwards 2002a; Matorras 1990). The cultures were performed at variable post-menstrual ages. In one study a rapid latex agglutination test was performed at the time of the mother giving birth (Tuppurainen 1989).

(1) Intrapartum antibiotics versus placebo or no treatment for GBS positive women

Primary outcomes for the infant

Only one study reported on the three primary outcomes (Boyer 1986).

  • There was no statistically significant effect of intrapartum antibiotics compared to no treatment on neonatal mortality from all causes (risk ratio (RR) 0.19, 95% confidence interval (CI) 0.01 to 3.82, one trial, 164 infants, test for heterogeneity not applicable) (seeAnalysis 1.1).

  • There was no statistically significant effect of intrapartum antibiotics compared to no treatment on neonatal mortality from GBS infection (RR 0.31, 95% CI 0.01 to 7.50, one trial, 164 infants, test for heterogeneity not applicable) (seeAnalysis 1.2).

  • There was no statistically significant effect of intrapartum antibiotics compared to no treatment on neonatal mortality from infections caused by bacteria other than GBS (RR 0.31, 95% CI 0.01 to 7.50, one trial, 164 infants, test for heterogeneity not applicable) (seeAnalysis 1.3).

Secondary outcomes for the infant

One or more studies reported on most of the pre-determined secondary outcomes. The number of infants included varied from 289 to 488.

  • There was a statistically significant reduction in the incidence of early (postnatal age less than seven days) GBS infection in neonates following intrapartum antibiotics compared to no treatment (RR 0.17, 95% CI 0.04 to 0.74, three trials, 488 infants, I² 0%; risk difference (RD) -0.04, 95% CI -0.07 to -0.01, I² 0%; number needed to treat to benefit (NNTB) 25, 95% CI 14 to 100, I² 0%) (seeAnalysis 1.4).

  • There was a statistically significant reduction in the incidence of probable early (postnatal age less than seven days) GBS infection in a neonate following intrapartum antibiotics compared to no treatment (RR 0.17, 95% CI 0.03 to 0.91, two trials, 324 infants, I² 0%; RD -0.05, 95% CI -0.09 to -0.01; NNTB 20, 95% CI 11 to 100, I² 0%) (seeAnalysis 1.5).

  • There was no statistically significant reduction in the incidence of late onset (seven days old or more) GBS infection in a neonate following intrapartum antibiotics compared to no treatment (RR 0.36, 95% CI 0.01 to 8.69, two trials, 289 infants, test for heterogeneity not applicable) (seeAnalysis 1.6).

  • There was no statistically significant difference in the incidence of neonatal sepsis, meningitis, urinary tract infection or pneumonia due to bacterial organisms other than GBS following intrapartum antibiotics compared to no treatment (RR 1.00, 95% CI 0.15 to 6.79, two trials, 289 infants, I² 4%) (seeAnalysis 1.7).

  • No other predetermined neonatal outcomes of interest were reported.

Secondary outcomes for the mother
  • There was no statistically significant effect on maternal sepsis in the peri/postpartum period (RR 0.31, 95% CI 0.01 to 7.49, one study, 160 women, test for heterogeneity not applicable) (seeAnalysis 1.8).

  • There was no statistically significant effect on puerperal infection (RR 0.16, 95% CI 0.01 to 3.03, one study, 121 women, test for heterogeneity not applicable) (seeAnalysis 1.9).

  • No other predetermined maternal outcomes of interest were reported.

As there was only one study that reported on our primary outcome, we did not perform subgroup analyses based on heterogeneity. As no abstracts were included we did not perform any analyses excluding abstracts. The authors did not provide sufficient information for us to do separate analyses based on timing of antenatal test to detect GBS or timing of the administration of antibiotics to the mother.

(2) Intrapartum ampicillin versus penicillin for GBS positive women

One study qualified for inclusion (Edwards 2002a) enrolling 352 participants. Most of the outcomes reported lacked a definition. We contacted the first author but were unable to obtain information from him.

Primary outcomes for the infant
  • There was no statistically significant effect on neonatal mortality from all causes comparing intrapartum administration of ampicillin with penicillin (RR 3.03, 95% CI 0.12 to 73.98, one study 352 infants, test for heterogeneity not applicable). The one death was caused by a lethal congential heart condition and not due to an infection (seeAnalysis 2.1).

Secondary outcomes for the infant
  • There was no statistically significant effect on suspected neonatal infection (definition not provided by the authors) comparing intrapartum administration of ampicillin with penicillin (RR 0.85, 95% CI 0.49 to 1.46, one study 352 infants) (seeAnalysis 2.2).

  • There was no statistically significant difference in the length of initial hospital stay for neonates comparing intrapartum administration of ampicillin with penicillin (mean difference (MD) 0.20 days , 95% CI -0.28 to 0.68) (seeAnalysis 2.3).

Secondary outcomes for the mother
  • There were no allergic reactions reported in either the ampicillin nor the penicillin group (seeAnalysis 2.4).

  • There was no statistically significant effect on chorioamnionitis (definition not provided) comparing intrapartum administration of ampicillin with penicillin (RR 0.91, 95% CI 0.38 to 2.19, one study, 352 women, test for heterogeneity not applicable) (seeAnalysis 2.5).

  • There was no statistically significant effect on endometritis (definition not provided) comparing intrapartum administration of ampicillin with penicillin (RR 3.03, 95% CI 0.32 to 28.89, one study, 352 women, test for heterogeneity not applicable) (seeAnalysis 2.6).

Discussion

Summary of main results

Acknowledging our serious concerns about bias in three of the included trials, we did combine these studies and found a statistically significant reduction in early group B streptococcus (GBS) neonatal infection. A similar statistically significant point estimate was obtained for RR for probable early GBS infection. Intrapartum antibiotic prophylaxis (IAP) appeared to reduce early onset GBS disease (EOGBSD), but this result may well be a result of bias as we found a high risk of bias for one or more key domains in the study methodology and execution. It should be noted that the attack rate in the control groups were 4.7% (47/1000 infants) and 5.7 % (57/1000 infants), respectively for these two outcomes, which seems exceedingly high.

Overall completeness and applicability of evidence

It is remarkable that in North America the commonly implemented practice of IAP to GBS colonized women has been so poorly studied. Only three randomized controlled trials conducted more than 20 years ago in three different countries and enrolling a total of 500 women have been published. We identified serious concerns of bias in these trials affecting our ability to draw conclusions from this systematic review. Concerns include no preset sample sizes, the lack of a placebo in the control groups, women and care-providers not blinded to group assignment, reporting on outcomes while the trials were ongoing, and exclusion of women who developed signs of infections in labor. As these trials were conducted and published prior to the CONSORT guidelines, the description of important aspects of study design, execution and the reporting of data are missing (Begg 1996).

Quality of the evidence

The attack rate for neonatal GBS sepsis has been reported as 0.5 per 1000 live births in the UK in the absence of systematic screening or widespread use of antibiotics (RCOG 2003). It is therefore not possible to draw meaningful conclusions from studies that today have only included a total of 500 women, even when appreciating the fact that some of these women were at a higher risk for neonatal GBS infections than the pregnant population at large.

Potential biases in the review process

We are not aware of any potential biases in our review process.

Agreements and disagreements with other studies or reviews

In one of our previous reviews on the topic published 20 years ago we decided not to combine the results of the same studies that are included in this current review, as we raised similar serious concerns about bias (Ohlsson 1994). The conclusion remains the same "Intrapartum chemoprophylaxis to reduce perinatal GBS infections are not supported by conclusive evidence from well designed and conducted randomised controlled trials" (Ohlsson 1994). All cases of perinatal GBS infections cannot be prevented. The UK National Screening Committee examined the issue of strategies for the prevention of EOGBS disease in November 2008 and recommended that routine screening using bacteriological culture or near-patient testing techniques should not be introduced into UK practice (UK National Screening Committee 2012). The latest update of the guidelines from the Royal College of Obstetricians and Gynaecologist regarding GBS do not recommend routine bacteriological screening of all pregnant women for antenatal GBS carriage (RCOG 2012).

Authors' conclusions

Implications for practice

In the three studies investigating the effects of intrapartum antibiotics versus no treatment for Group B haemolytic streptococci (GBS) colonized women, we identified high risks of bias for one or more key domains in the study methodology and execution. The information from all studies was at high risk of bias sufficient to affect the interpretation of the results. Based on this review, we conclude that there is no valid information from these three small, old and biased trials to inform clinical practice.

Information on whether intrapartum ampicillin is preferable to penicillin for GBS colonized women is lacking.

Implications for research

Ideally, the effectiveness of intrapartum antibiotics to GBS colonized women to reduce neonatal GBS infections should be studied in adequately sized double blind controlled trials. The opportunities to conduct such trials have likely been lost as practice guidelines have been introduced in many jurisdictions. It should be noted that the guidelines have changed many times, indicating that they are not based on clear evidence informing best clinical practice. Even if an effective vaccine to prevent GBS infections is developed in the future, a need for intrapartum prophylaxis (if proven effective) is still likely to be present as all women will not be immunized and the vaccine may not be effective in women giving birth preterm.

Acknowledgements

We acknowledge with appreciation the support from Mrs G Gyte in completing previous versions of this review.

Data and analyses

Download statistical data

Comparison 1. Intrapartum antibiotics versus placebo or no treatment for GBS positive women
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Neonatal mortality from all causes1164Risk Ratio (M-H, Fixed, 95% CI)0.19 [0.01, 3.82]
2 Neonatal mortality from early onset GBS infection1164Risk Ratio (M-H, Fixed, 95% CI)0.31 [0.01, 7.50]
3 Neonatal mortality from infections caused by bacteria other than GBS1164Risk Ratio (M-H, Fixed, 95% CI)0.31 [0.01, 7.50]
4 Early (postnatal age less than 7 days) GBS infection in a neonate3488Risk Ratio (M-H, Fixed, 95% CI)0.17 [0.04, 0.74]
5 Probable early (postnatal age less than 7 days) GBS infection in a neonate2324Risk Ratio (M-H, Fixed, 95% CI)0.17 [0.03, 0.91]
6 Late onset (7 days old or more) GBS infection in a neonate2289Risk Ratio (M-H, Fixed, 95% CI)0.36 [0.01, 8.69]
7 Neonatal sepsis due to bacterial organisms other than GBS2289Risk Ratio (M-H, Fixed, 95% CI)1.00 [0.15, 6.79]
8 Maternal sepsis in the peri/postpartum period1160Risk Ratio (M-H, Fixed, 95% CI)0.31 [0.01, 7.49]
9 Puerperal infection (not prespecified outcome, definition not provided)1121Risk Ratio (M-H, Fixed, 95% CI)0.16 [0.01, 3.03]
Analysis 1.1.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 1 Neonatal mortality from all causes.

Analysis 1.2.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 2 Neonatal mortality from early onset GBS infection.

Analysis 1.3.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 3 Neonatal mortality from infections caused by bacteria other than GBS.

Analysis 1.4.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 4 Early (postnatal age less than 7 days) GBS infection in a neonate.

Analysis 1.5.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 5 Probable early (postnatal age less than 7 days) GBS infection in a neonate.

Analysis 1.6.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 6 Late onset (7 days old or more) GBS infection in a neonate.

Analysis 1.7.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 7 Neonatal sepsis due to bacterial organisms other than GBS.

Analysis 1.8.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 8 Maternal sepsis in the peri/postpartum period.

Analysis 1.9.

Comparison 1 Intrapartum antibiotics versus placebo or no treatment for GBS positive women, Outcome 9 Puerperal infection (not prespecified outcome, definition not provided).

Comparison 2. Intrapartum ampicillin versus penicillin for GBS positive women
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Neonatal mortality from all causes1352Risk Ratio (M-H, Fixed, 95% CI)3.03 [0.12, 73.98]
2 Suspected neonatal infection (definition not provided)1352Risk Ratio (M-H, Fixed, 95% CI)0.85 [0.49, 1.46]
3 Initial hospital stay (days) for neonates1352Mean Difference (IV, Fixed, 95% CI)0.20 [-0.28, 0.68]
4 Maternal allergic reactions to antibiotics1352Risk Ratio (M-H, Fixed, 95% CI)0.0 [0.0, 0.0]
5 Chorioamnionitis (not prespecified outcome, definition not provided)1352Risk Ratio (M-H, Fixed, 95% CI)0.91 [0.38, 2.19]
6 Endometritis (not prespecified outcome, definition not provided)1352Risk Ratio (M-H, Fixed, 95% CI)3.03 [0.32, 28.89]
Analysis 2.1.

Comparison 2 Intrapartum ampicillin versus penicillin for GBS positive women, Outcome 1 Neonatal mortality from all causes.

Analysis 2.2.

Comparison 2 Intrapartum ampicillin versus penicillin for GBS positive women, Outcome 2 Suspected neonatal infection (definition not provided).

Analysis 2.3.

Comparison 2 Intrapartum ampicillin versus penicillin for GBS positive women, Outcome 3 Initial hospital stay (days) for neonates.

Analysis 2.4.

Comparison 2 Intrapartum ampicillin versus penicillin for GBS positive women, Outcome 4 Maternal allergic reactions to antibiotics.

Analysis 2.5.

Comparison 2 Intrapartum ampicillin versus penicillin for GBS positive women, Outcome 5 Chorioamnionitis (not prespecified outcome, definition not provided).

Analysis 2.6.

Comparison 2 Intrapartum ampicillin versus penicillin for GBS positive women, Outcome 6 Endometritis (not prespecified outcome, definition not provided).

Feedback

Tony Ades, April 2016

Summary

I have a number of comments and questions concerning this review.

Outcome measures . Obviously neonatal early onset GBS infection is an important outcome, but why is infant colonisation not mentioned in this update, although it was prominent in the previous version originally published in 1996.1 Although colonisation of the newborn is not the critical outcome in itself, it may be a better outcome to look at. First because the numerators are larger, so that treatment effects can be estimated more accurately, and second, if treatment can prevent colonisation of the newborn, then logically it must also prevent neonatal infection, as well as the sequelae of neonatal infection, and mortality associated with neonatal infection.

Bias. The review comments on the potential biases in the original trials. Accepting that many of these trials may have suspect features, one still has to ask: can the degree of bias really be so great as to explain away the extremely low Odds ratios / Relative Risks reported for both these outcomes in every study.

Estimates of odds ratios and risk ratios. Both the current review and the earlier one appear to have underestimated the extent of risk reduction, by almost a factor of two. The Peto method used for calculating the summary statistic in the previous version of the review1 is known to be biased towards unity, especially when the true risk is far from 1. In the update the reviewers have used the Mantel Haenzel method, but have added 0.5 to zero cells which is not necessary or advisable. This has biased the reported risk ratio towards 1, quite markedly in view of the high proportion of trials with numerators in the intervention arm.

Conclusions. The updated review states "conclusions remain the same" as the previous review, namely that "Intrapartum chemoprophylaxis to reduce perinatal GBS infections are not supported by conclusive evidence from well-designed and conducted RCTs". The UK National Screening Committee (2008, 2012) and RCOG (2012) are cited as not recommending “routine bacteriological screening”. But the question of whether all women, or some women, should be screened is very different from the question whether intrapartum antibacterial prophylaxis is effective, and UK guidelines continue to recommend this for women at higher risk. In fact, most countries routinely use intrapartum antibacterial prophylaxis, either for higher risk groups or as part of a screening strategy. The recommendations in this update, and the previous version of the review, therefore appear to diverge markedly from accepted practice and opinion.

References

1. Smaill FM. Intrapartum antibiotics for Group B streptococcal colonisation. Cochrane Database Syst Rev. 2010 Jan 20;(1):CD000115. doi: 10.1002/14651858.CD000115.pub2.

Summary of comment from Tony Ades, April 2016

Reply

Responses by Arne Ohlsson and Vibhuti Shah, July 2016

Outcome measures: We were not involved in the previous Cochrane review by Smail.l1 Our Cochrane review, first published in 2009 and updated in 2013 and 20142--4, does not include group B streptococcal (GBS) neonatal colonization as an outcome because it was not considered a clinically important outcome. Neonatal colonization with GBS is common, whereas clinical disease in colonized infants is very low. Hence we do not agree that neonatal GBS colonization should be used as an outcome “….because the numerators are larger, so that treatment effects can be estimated more accurately…”. Early-onset neonatal GBS infection is a clinically relevant and important outcome for clinicians, and neonatal mortality due to early-onset neonatal GBS infection is one of our primary outcomes.

Bias: The risk of bias in the included studies is high and could explain away the extremely low Odds ratios / Relative Risks reported for both these outcomes.

In the review intrapartum antibiotic prophylaxis did not significantly reduce the incidence of all-cause mortality, mortality from GBS infection or from infections caused by bacteria other than GBS compared with no prophylaxis. The incidence of early GBS infection was reduced with intrapartum antibiotic prophylaxis. The quality of the included studies was poor and the risk of bias was high, sufficient to affect the interpretation of the results.

Study quality is summarized in the ‘characteristics of included studies’ table in the review. Concerns regarding risk of bias include: no reported sample size estimate; the lack of a placebo in the control groups; women, care-providers and outcome assessors not blinded to group assignment; researchers reporting on outcomes while the trials were ongoing; and exclusion of women who developed signs of infections in labour. In the one study comparing ampicillin with penicillin, treatment allocation for the only neonatal infection is not reported and there is no definition of suspected infection and chorioamnionitis. These trials were conducted and published prior to the CONSORT guidelines, and the description of important aspects of study design, execution and the reporting of data are missing.

Acknowledging the high risk of bias in the three trials comparing intrapartum antibiotic prophylaxis with no prophylaxis, we did combine these studies and found a statistically significant reduction in the risk of early group B streptococcus (GBS) neonatal infection and in probable early GBS infection. However, the attack rate in the control group was 4.7% (47/1000 infants) and 5.7 % (57/1000 infants) respectively for these two outcomes, which seems exceedingly high.

Intrapartum antibiotic prophylaxis appears to reduce early onset GBS disease, but this result may well be due to bias.

Estimates of odds ratios and risk ratios: We are not statisticians, but assume that the statistical methods used in RevMan 5.2 are correct.5 In the landmark book ‘Effective Care in Pregnancy and Child Birth’ the Mantel-Haentzel-Peto method was used to report on dichotomous outcomes as odds ratios.6 In ‘Effective Care of the Newborn’ Bracken advocates the use of risk ratio, risk difference and number needed to treat to benefit or harm.7 Most Cochrane review groups now use these statistics. In our review we used the Pregnancy and Childbirth Review Group methodological guidelines, including the statistics for risk ratio and risk difference. In the case of no events, the Cochrane Handbook8 states: “Neither the risk ratio nor the odds ratio can be calculated for a study if there are no events in the control group. This is because, ….. we would be trying to divide by zero. The odds ratio also cannot be calculated if everybody in the intervention group experiences an event. In these situations, and others where standard errors cannot be computed, it is customary to add ½ to each cell of the 2×2 table (RevMan automatically makes this correction when necessary)”. The addition of 0.5 events does not seem to show up in the RevMan analyses.

Conclusions: In our latest update we conclude that “Information on whether intrapartum ampicillin is preferable to penicillin for GBS colonized women is lacking”. Although we state: “There is lack of evidence from well designed and conducted trials to recommend intrapartum antibiotic prophylaxis to reduce neonatal early onset GBS disease”, we do not make any recommendations for clinical practice. We simply present the evidence and it is up to Guideline developing authorities/caregivers and/or parents to be make decisions on the evidence we have provided. We did not evaluate different screening programs for GBS maternal colonization.

References

  1. Smaill F. Intrapartum antibiotics for group B streptococcal colonisation. Cochrane Database Systematic Reviews 1996;(1):CD000115.

  2. Ohlsson A, Shah VS. Intrapartum antibiotics for known maternal Group B streptococcal colonization. Cochrane Database of Systematic Reviews 2009, Issue 3. DOI: 10.1002/14651858.CD007467.pub2.

  3. Ohlsson A, Shah VS. Intrapartum antibiotics for known maternal Group B streptococcal colonization. Cochrane Database of Systematic Reviews 2013, Issue 1. DOI: 10.1002/14651858.CD007467.pub3.

  4. Ohlsson A, Shah VS. Intrapartum antibiotics for known maternal Group B streptococcal colonization. Cochrane Database of Systematic Reviews 2014, Issue 6. Art. No.: CD007467. DOI: 10.1002/14651858.CD007467.pub4.

  5. Review Manager (RevMan) [Computer program]. Version 5.2. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012.

  6. Effective Care in Pregnancy and Childbirth (1989). Volume 1 & 2. Chalmers I, Enkin M, Keirse MJNC (eds.). Oxford University Press, Oxford, New York, Toronto.

  7. Bracken MB. Statistical methods for analysis of effects of treatment in overviews of randomized trials. Pp 13-8. In: Effective Care of the Newborn Infant (1992). Sinclair JC, Bracken MB (eds.) Oxford University Press, Oxford, New York, Tokyo.

  8. Higgins JPT, Green S (editors). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 [updated March 2011]. The Cochrane Collaboration, 2011. Available from www.cochrane-handbook.org.

Contributors

Responses by Arne Ohlsson and Vibhuti Shah, July 2016

What's new

DateEventDescription
8 August 2016Feedback has been incorporatedResponse from Arne Ohlsson and Vibhuti Shah added.

History

DateEventDescription
18 July 2016Feedback has been incorporated Feedback 1 has been received from Tony Ades.
11 March 2014New citation required but conclusions have not changedNo new trial reports identified by the updated search.
11 March 2014New search has been performedSearch updated. 'Risk of bias' summary figures added and text within 'Risk of bias' results section.
10 November 2012New search has been performedSearch updated.
10 November 2012New citation required but conclusions have not changedNo new trial reports identified by the updated search.

Contributions of authors

Both review authors contributed to the protocol and the review. Arne Ohlsson wrote the text and Vibhuti Shah made important contributions to the protocol and edited the text. Both review authors contributed to all steps of the full review. Both authors contributed to the current and previous updates of the review.

Declarations of interest

None known.

Sources of support

Internal sources

  • Department of Paediatrics, Mount Sinai Hospital, Canada.

External sources

  • No sources of support supplied

Differences between protocol and review

In a deviation from our protocol we included studies that compared one antibiotic versus another for GBS prophylaxis. We considered this inclusion to be important to ascertain whether one antibiotic would be preferable to another and to be able to estimate the expected attack rate of GBS infection in infants exposed to any antibiotic. We included the outcome of puerperal infection as it was reported in one study. We included the outcome of chorioamnionitis from one study although the authors did not provide a definition. From the same study we included endometritis (no definition) as an outcome in spite of the fact that this was not one of our pre-determined outcomes.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Boyer 1986

MethodsRandomized controlled trial.
I. Blinding of randomization - no.
II. Blinding of intervention - no.
III. Complete follow-up - no.
IV. Blinding of outcome measurement - no.
Participants

180 women with vaginal, rectal or both specimens positive for group B streptococcus (in the majority of women the interval between the time of cultures and parturition was > 10 weeks) and at the time of giving birth had the following risk factors (selective intrapartum prophylaxis): preterm labor (< 37 weeks of gestation) or prolonged rupture of membranes (> 12 hours).

If women in the control group developed intrapartum fever (temperature > 37.5ºC) they were excluded from the study group and were treated with ampicillin.

Exclusion criteria: penicillin allergy or need for other antimicrobial agents.

Study period: prior to May 1982 (first report in abstract form by Boyer 1982).

Study period: May 1979 to September 1981 (second report Boyer 1983).

Study period: May 1979 to September 1984 (final report Boyer 1986).

Multicenter study, USA (Private obstetrician’s clinics, a health maintenance organization and the obstetric clinics of Micheal Reese Hospital and Medical Center).

Interventions

Women in the treatment group (n = 94): received 2 g of ampicillin intravenously followed by 1 g every 4 hours until giving birth.

Women in the control group (n = 86) received no ampicillin.

If the mother had received ampicillin, the infant was treated with 4 doses of intramuscular ampicillin (50 mg/kg) every 12 hours.

Infants born to untreated women received antibiotics only if symptoms of sepsis were observed.

In all symptomatic infants (presence of respiratory distress, asphyxia, or signs of infection at birth regardless of maternal treatment) cerebrospinal fluid examination was performed and treatment with ampicillin and kanamycin commenced until the results of blood and surface cultures were available.

Outcomes

Primary outcomes

Neonatal

All cause mortality

Mortality from early (postnatal age less than 7 days) onset culture positive neonatal GBS infection including 1 or more of the following conditions.

a) Sepsis - defined as symptoms and signs of sepsis and a bacterial culture positive for GBS (obtained in a sterile manner from normally sterile body fluids such as blood, cerebrospinal fluid or urine, or culture from internal organs at autopsy).

b) Pneumonia in the neonate (postnatal age less than 7 days) - defined as symptoms and signs and radiographic findings consistent with pneumonia and positive culture for GBS (obtained from tracheal aspirate or by culture of lung tissue at autopsy).

Mortality from infections (as per a and b above) caused by bacteria other than GBS.

Secondary outcomes

Neonatal

Early (postnatal age less than 7 days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and positive GBS bacterial cultures (from normally sterile body fluids obtained from the neonate).

Late onset GBS sepsis - sepsis due to GBS in an infant at least 7 days old.

Neonatal sepsis due to bacterial organisms other than GBS.

Maternal

Sepsis in the peri/postpartum period.

NotesDetermined the incidence of group B streptococcus bacteraemia in infants born to 1648 women with prenatal colonization who did not participate in the randomized study. Antibiotics were administered to 232 of these women and blood culture obtained from mother or their infant if sepsis was suspected.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskParticipants were assigned to ampicillin or control groups (with an allocation ratio of 1:1) on the basis of sequential selection of sealed opaque envelopes containing assignments generated from a table of random numbers.
Allocation concealment (selection bias)High riskSequential selection of opaque envelopes. Although there was adequate sequence generation it would very soon become obvious to which group the next mother would be assigned as the allocation ratio was 1:1. Although the allocation ratio was 1:1 there is an imbalance in the numbers of women allocated to the 2 groups: 94 in the ampicillin group and 86 in the control group. After excluding 20 women (13 women who developed intrapartum fever and 7 for whom there were randomization errors or incomplete data, 83 women (85 infants) remained in the ampicillin group and 77 (79 infants) in the control group.
Blinding of participants and personnel (performance bias)
All outcomes
High risk"Neither the patient nor the obstetricians were blinded to the assignment to study groups." We have interpreted the information as that the sequence generation was adequate but from then on the study was open to patients and care-givers.
Blinding of outcome assessment (detection bias)
All outcomes
High riskOutcomes were not assessed by staff blinded to the intervention.
Incomplete outcome data (attrition bias)
All outcomes
High risk13 women were excluded as they developed intrapartum fever (7 in the ampicillin group and 6 in the control group) and 7 in whom there were randomization errors or incomplete data (4 in the ampicillin group and 3 in the control group).
Selective reporting (reporting bias)High riskConcerns addressed in the cells above and below.
Other biasHigh risk

The results of this ongoing study has been reported on 3 occasions. In the first report (Boyer 1982) there were 71 infants in the ampicillin group and there were 128 infants in the control group (the number of mothers randomized was not reported). Blood cultures were positive in 4 heavily colonized infants whose mothers were not treated with ampicillin; no blood cultures were positive among infants whose mothers were treated (P = .17).

In the second report (Boyer 1983), 80 women were randomized; 43 received ampicillin chemoprophylaxis and 37 did not. 1 infant in the control group had GBS bacteraemia.

Between the second (Boyer 1983) and third (final, Boyer 1986) publication Gotoff in a letter to the editor (Gotoff 1984) wrote "In order to show efficacy in preventing GBS disease, we need an additional case in our control group". It seems clear that the researchers were aware of study results throughout the study and stopped recruitment when statistical significance (1-tailed) had been achieved. Between the second (Boyer 1983) and the third publication (Boyer 1986) the authors changed their test of significance for comparisons of colonization, bacteraemia, and the rate of postpartum febrile morbidity from a 2-tailed to a 1-tailed test.

It is remarkable that in a study of perinatal infections 13 women were excluded as they developed intrapartum fever (7 in the ampicillin group and 6 in the control group).

Edwards 2002a

MethodsRandomized controlled trial.
I. Blinding of randomization - yes.
II. Blinding of intervention - cannot tell.
III. Complete follow-up - yes.
IV. Blinding of outcome measurement - cannot tell.
Participants

Women who were at a gestational age of 36 weeks or more, were in spontaneous or induced labor and were culture-proven carriers of group B streptococci. Cultures for GBS were obtained at the time of admission for spontaneous or induced labor. Exclusion criteria included planned caesarean section, antibiotics taken within the preceding 7 days, a history of allergy to penicillins, multifetal gestation, or antepartum fetal death.

Study period 26 February 2000 to 22 May 2001.

Interventions175 women received ampicillin (2 g of ampicillin IV followed by 1 g every 4 hours until giving birth) and 177 received penicillin (5 million units of penicillin G IV, followed by 2.5 million units every 4 hours until giving birth).
OutcomesAll cause mortality.
Suspected infection (the authors do not provide a definition).
Initial hospital stay (neonatal).
Chorioamnionitis (definition not provided).
Endometritis (definition not provided).
Allergic reactions to antibiotics (maternal).
NotesWe contacted (5 January 2009) the primary author to provide us with information in which treatment group the early-onset neonatal infection occurred and what their definition of suspected infection was. As of 7 May 2009, we have not received an answer.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskA random number-generating software program was used to assign participants to groups.
Allocation concealment (selection bias)Low riskParticipants were randomized, by selection of the next opaque envelope containing an order sheet, to receive intrapartum antibiotic prophylaxis with ampicillin or penicillin.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskIt is not stated whether the women and the care-givers were aware of what antibiotic was administered.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskIt is not stated whether the women and the care-givers were aware of what antibiotic was administered to the participants.
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcomes reported for all women randomized.
Selective reporting (reporting bias)Low riskThe authors report on an intent-to-treat analysis and a per-protocol analysis (Changed to low risk).
Other biasLow riskThe study appears to be free of other sources of bias.

Matorras 1990

MethodsRandomized controlled trial.
I. Blinding of randomization - no.
II. Blinding of intervention - no.
III. Complete follow up - yes.
IV. Blinding of outcome measurement - no.
Participants

121 women with group B streptococcal colonization (vaginal and/or rectal swabs). The gestational age at which cultures were obtained ranged from 17 and 42 weeks (mean 32.98 weeks; standard deviation +/- 5.03 weeks).
Study period: not reported.  

Single centre study, Spain.

Interventions

Women in the treatment group (n = 57) received 500 mg of ampicillin IV every 6 hours until delivery. If induction of labor, antibiotics were administered at the beginning of induction, and if caesarean section without labor, 2 hours prior to surgery.

Patients allergic to penicillin received erythromycin.

In the control group (n = 64) no ampicillin prophylaxis or placebo was administered.

Outcomes

Information on the different parameters analyzed was obtained retrospectively from clinical histories.

Early (postnatal age less than 7 days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and positive GBS bacterial cultures (from normally sterile body fluids obtained from the neonate) (Matorras 1991).

Probable early (postnatal age less than 7 days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and bacterial cultures from normally sterile body fluids obtained from the neonate that were negative for GBS (Matorras 1991).

Late onset GBS sepsis - sepsis due to GBS in an infant at least 7 days old.

Neonatal sepsis due to bacterial organisms other than GBS (Matorras 1991).

Late onset sepsis (Matorras 1991).

Puerperal infection: defined according to clinical criteria- uterine tenderness, uterine subinvolution and fever in the absence of any other known cause of infection.

NotesIn order to assess the impact of GBS maternal colonization for infective puerperal morbidity, the non-carrier patients were compared with the GBS carrier patients who did not receive prophylaxis.
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo information was provided on how the random sequence was generated.
Allocation concealment (selection bias)Unclear riskIt is stated that "....women were randomly divided" into 2 groups. Insufficient information to permit judgement of "yes" or "no".
Blinding of participants and personnel (performance bias)
All outcomes
High riskThe control group received no intervention (no placebo).
Blinding of outcome assessment (detection bias)
All outcomes
High riskThe placebo group received no intervention and therefore outcomes were assessed by staff aware of group assignment.
Incomplete outcome data (attrition bias)
All outcomes
Low riskOutcomes reported for all women randomized.
Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement of "yes" or "no".
Other biasUnclear riskThe study appears to be free of other sources of bias. The results of the randomized part of this study, which formed the part of a larger cohort study has been reported on at least 3 occasions. The number of randomized women was the same in all 3 reports (n = 121). The rate of puerperal infection was reported in Matorras 1990. In an abstract (Omenaca 1987) the rate of neonatal sepsis caused by GBS was reported as 3 % (2/57) in babies whose mothers received prophylaxis and 13.8% (9/64) in infants of untreated mothers. In the final report (Matorras 1991), the authors report that there was no case of GBS sepsis in the prophylaxis group compared to 3 cases in the control group. Clinically infected newborns represented 3.3% in the prophylaxis group vs 13.8% in the control group.

Tuppurainen 1989

  1. a

    GBS: Group B haemolytic streptococci
    IV: intravenous
    vs: versus

MethodsRandomized controlled trial.
I. Blinding of randomization - no.
II. Blinding of intervention - no.
III. Complete follow up - yes.
IV. Blinding of outcome measurement - no.
Participants

Women with a positive group B streptococcus latex test before giving birth (in labor), no history of penicillin allergy and no elective term caesarean section without labor or rupture of fetal membranes were eligible.

In woman with a positive streptolatex test admitted for induction of labor but who did not give birth and returned home, the test was repeated if 3 or more days had passed since the first test.

Study period: December 1983 and January 1986 except for a 6-month period in 1984 to 1985.

Single centre, Finland.

Interventions

Women in the treatment group (n = 88) received 5 million units penicillin G IV every 6 hours during labor. If labor lasted more than 18 hours, 1 million units penicillin V administered orally every 8 hours until parturition (penicillin was chosen as no resistant strains of group B streptococcus had been detected in the institution where the study was conducted).

In the control group (n = 111) women received no prophylaxis or placebo.

Newborns born to group B streptococcus positive women were evaluated as follows.

  1. Blood culture was obtained within 2 hours of birth.

  2. Pharyngeal aspirate and other superficial samples for culture from the external ear canal, eye, and umbilicus were taken within 30 minutes of birth.

  3. Urine sample for streptolatex text was collected.

Cerebrospinal fluid examined was performed if the infant had symptoms and/or signs of sepsis or of meningeal involvement.

Outcomes

Early (postnatal age less than 7 days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and positive GBS bacterial cultures (from normally sterile body fluids obtained from the neonate).

Probable early (postnatal age less than 7 days) GBS infection in a neonate - defined as symptoms and signs of sepsis or pneumonia in a neonate born to a GBS positive mother, and bacterial cultures from normally sterile body fluids obtained from the neonate that were negative for GBS.

Notes

No mention of severity of illness in the newborn or details provided regarding duration of hospitalization and antibiotic administration.

No mention of benefits/adverse reactions to the mother.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskInsufficient information about sequence generation process to permit judgement of "yes" or "no".
Allocation concealment (selection bias)High riskParticipants were assigned to the penicillin or control group based on sequential selection of sealed envelopes containing the treatment instructions. However, the authors state "There was no blinding in the assignment to study groups". There was imbalance in allocation of participants to the two groups; 44% of the participants were allocated to the intervention group and 56% to the control group.
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo blinding as the control group received no intervention (no placebo was used).
Blinding of outcome assessment (detection bias)
All outcomes
High riskThe control group received no intervention and no placebo and staff were not blinded to group assignment. Thus the outcome assessors were not blinded to group assignment.
Incomplete outcome data (attrition bias)
All outcomes
Low riskNo missing outcome data.
Selective reporting (reporting bias)Low riskThe authors reported outcomes for women with positive streptolatex test who were randomized (199) and those who were not randomized (157 women who gave birth before the test results were available and 21 women with a history of penicillin allergy. Outcomes for 8565 women who were streptolatex negative were reported. These women gave birth to six neonates with early-onset GBS disease.
Other biasHigh riskResults of the study were first reported in abstract form in 1986 (Tuppurainen 1986) when 94 patients had been randomized, 36 received intrapartum penicillin and 58 did not. 7 of the 58 (12%) neonates whose mother did not receive penicillin developed early onset GBS disease. 1 neonate whose mother received penicillin had intrauterine pneumonia probably due to GBS. It appears that results of the study were known on an ongoing basis.

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    GBS: Group B haemolytic streptococci
    IAP: intrapartum antibiotic prophylaxis
    PCR: polymerase chain reaction

Belady 1996Randomized controlled trial comparing ampicillin versus penicillin for group B streptococcus prophylaxis. No placebo or untreated control group was included. Reports only on colonization rates. In a separate abstract from the same study (Davies 1998) the authors do not report on maternal infections as per randomized groups.
Dykes 1987This study reported on the effects of chlorhexidine for prevention of neonatal colonization with group B streptococci. This intervention is the topic of another Cochrane review.
Easmon 1983This randomized controlled study reports only on maternal and neonatal colonization, which were not considered important outcomes in our review.
Facchinetti 2002The objective of this randomized controlled trial was to study the efficacy of intrapartum vaginal flushing with chlorhexidine compared with ampicillin in preventing group B streptococcus transmission to neonates. This is the topic of a separate Cochrane review.
Gibbs 1996This randomized controlled study reported on the effect of 2% clindamycin cream administered intravaginally during labor to group B streptococcal-colonized pregnant women. Outcomes included only maternal and neonatal colonization, which were not considered important outcomes for this review.
Håkansson 2014This study consisted of 2 phases (Phase 1 was a randomized controlled trial and Phase 2 was non-randomized, assessing an improved version of the assay). Phase 1 involved 6 Swedish delivery units. The women included in phase 1 were randomized into 2 groups (1A or 1B). The randomization was performed in advance and each hospital had received sealed and consecutively numbered envelopes containing protocols for either group 1A or 1B. Randomization was performed in blocks of 4. In the PCR group (group 1A) swabs in duplicate (Copan Italia S.P.A., Brescia, Italy) were obtained from vagina and rectum (V/R) in this order. 1 swab was used for PCR and the other for conventional, enhanced culture for detection of GBS [16]. IAP was administered if the PCR assay was positive for GBS or indeterminate. In the control group" (group 1B), swabs were taken from V/R for conventional culture with the intention to treat all with IAP according to recommended guidelines. The study did not compare an intervention to placebo. The study (Phase 1) was excluded as both groups received IAP. Phase 2 was not a randomized controlled trial.
Merenstein 1980In this study treatment with antibiotics started at 38 weeks postmenstrual age not intrapartum.
Morales 1986In this study the control group included randomly selected patients and those with a history of ampicillin allergy.
Pinette 2005In this study pregnant women positive for GBS at 35 to 37 weeks' postmenstrual age were randomized to receive intramuscular benzathine penicillin G suspension versus no treatment. Intrapartum all the women received prophylaxis according to CDC guidelines.
Sáez-Llorens 1995This study was an open, non-randomized trial.

Ancillary