The Prevention of Early-onset Neonatal GBS Disease in UK ... · The Prevention of Early-onset Neonatal Group B Streptococcal Disease in UK Obstetric Units An audit of reported practice

The Prevention of Early-onsetNeonatal Group B StreptococcalDisease in UK Obstetric Units

An audit of reported practice in England, Scotland, Wales and Northern Ireland

January 2007

Royal College of Obstetricians and Gynaecologists

and London School of Hygiene and Tropical Medicine

Commissioned by the National Screening Committee

The Prevention of Early-onsetNeonatal Group B Streptococcal

Disease in UK Obstetric Units

Royal College of Obstetricians and Gynaecologists

and London School of Hygiene and Tropical Medicine

The Prevention of Early-onset Neonatal Group BStreptococcal Disease in UK

Obstetric Units

An audit of reported practice in England,Scotland, Wales and Northern Ireland

Commissioned by the National Screening Committee

Dr David Cromwell

Dr Tracey Joffe

Dr Jan van der Meulen

Mrs Charnjit Dhillon

Dr Rhona Hughes

Professor Deirdre Murphy

January 2007

© 2007 Royal College of Obstetricians and Gynaecologists

First published 2007

All rights reserved. No part of this publication may be reproduced, stored ortransmitted in any form or by any means, without the prior written permission ofthe publisher or, in the case of reprographic reproduction, in accordance with theterms of licences issued by the Copyright Licensing Agency in the UK[www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated hereshould be sent to the publisher at the UK address printed on this page.

Published by RCOG Press at the Royal College of Obstetricians and Gynaecologists27 Sussex Place, Regent’s ParkLondon NW1 4RGRegistered Charity No. 213280

ISBN 978-1-904752-37-0

RCOG Press Editor: Jane MoodyDesign & typesetting: FiSH Books, LondonPrinted by Print Direct (Maidenhead) Ltd, Unit 1, Priors Way, Maidenhead, UK

Contents

Acknowledgements vi

Abbreviations vii

Executive summary ix

1. Introduction 1

2. An international comparison of guidelines on the prevention of early-onset GBS disease 4

3. Audit design and data collection 13

4. Content of maternity unit protocols 16

5. Audit results 23

6. Patient vignettes 36

7. Conclusion 40

References 44

Guidelines reviewed

Appendix 1. RCOG guideline classifications for levels of evidence and strength of recommendation 46

Appendix 2. Search strategy for the review of international GBS guidelines 47

Appendix 3. The midwife questionnaire 48

Appendix 4. Agreement among midwives and obstetricians on the management of women with GBS risk factors 59

Appendix 5. Clinician responses to patient vignettes 62

Acknowledgements

The audit was a collaborative project of the Royal College of Obstetricians andGynaecologists (RCOG) and the London School of Hygiene and Tropical Medicine(LSHTM). The National Screening Committee provided funding.

Project Team

Dr Tracey Joffe, Research Fellow, LSHTMDr David Cromwell, Lecturer, LSHTMProfessor Deirdre Murphy, Chair of the Guidelines and Audit Committee, RCOGDr Rhona Hughes, Lead author on RCOG GBS guideline, RCOGDr Jan van der Meulen, Reader, LSHTMMrs Charnjit Dhillon, Director of Standards, RCOG

Contributions

CD, DM and RH initiated the Audit; JVM and DC developed the protocol for theaudit; TJ carried out the systematic search and comparison of internationalguidelines, with support from DC to perform the AGREE analysis; TJ designed theaudit questionnaire with contributions from all team members; TJ and DC undertookthe collection, management and analysis of the audit data and hospital protocols; TJand DC wrote the final report with contributions from all team members.

We would also like to thank Elaine Garrett, Assistant Librarian, RCOG, whoconducted the initial literature search and Nicola Betton and Emily Symington(RCOG) who provided additional administration support.

Finally, we would like to thank Dr Ben Stenson, Consultant Neonatologist, RoyalInfirmary of Edinburgh at Little France, for his clinical advice and comments on thedraft report.

vi

Abbreviations

ACNM American College of Nurse-Midwives

ACOG American College of Obstetricians and Gynecologists

AGREE Appraisal of Guidelines for Research & Evaluation

ANAES Agence Nationale d’Accréditation et d’Evaluátion en Santé

BCRCP British Columbia Reproductive Care Program

CDC Centers for Disease Control and Prevention

CSH Conseil Supérieur d’Hygiène

CTF Canadian Task Force

GBS group B streptococcus

GBSS Group B Strep Support

HSCIC Health and Social Care Information Centre

IAP intrapartum antibiotic prophylaxis

IV intravenous

LSHTM London School of Hygiene and Tropical Medicine

NHS National Health Service

NNT number needed to treat

NZCOM New Zealand College of Midwives

NZCWG New Zealand GBS Consensus Working Group

PHLS Public Health Laboratory Service

RANZCOG Royal Australian and New Zealand College of Obstetricians andGynaecologists

RCM Royal College of Midwives

RCOG Royal College of Obstetricians and Gynaecologists

RCPCH Royal College of Paediatrics and Child Health

ROCR Review of Central Returns

ROM rupture of membranes

SOGC Society of Obstetricians and Gynaecologists of Canada

SROM spontaneous rupture of membranes

vii

Executive summary

An audit was established to evaluate practice in UK obstetric units on the preventionof early-onset neonatal group B streptococcus (GBS) disease against the recommend-ations of the RCOG Green-top guideline.1 The RCOG guideline was published inNovember 2003 and it was thought that practice would have changed since thesurveys conducted in 1999 and 2001.2 The audit examined the organisation ofscreening, the use of intrapartum antibiotic prophylaxis (IAP) and the managementof neonates born with increased risk of GBS disease.

A systematic review of national guidelines was also undertaken to determine thevariation in recommended practice between countries. The recommendations in theRCOG Green-top guideline were similar to guidelines by other UK organisations.However, there were differences in the recommendations of guidelines from differentcountries, particularly in relation to the strategies used to determine which womenshould be offered IAP. This variation appears to reflect the limited evidence base onthe prevention of early-onset GBS disease as well as concerns about the applicabilityof research in different contexts and cultural perspectives on childbirth.

The clinical directors of UK obstetric units were contacted in December 2005 andwere asked to supply the details of a senior labour ward midwife, lead obstetricianand lead neonatologist who could complete a questionnaire on their current practicefor preventing early-onset GBS disease. They were also asked to provide a copy oftheir local protocol on the prevention of GBS disease. The nominated midwives,obstetricians and neonatologists were sent the questionnaire in February 2006.Responses were received from 202 of the 227 units and response rates over 70% wereachieved for each profession.

Content of maternity unit protocols

The audit reviewed the protocols from 171 units. The RCOG guideline was cited by80% of the 120 protocols created in 2004 or after, and was the most citedpublication. The clinical recommendations in the protocols were broadly consistentwith the RCOG guideline. No protocol appeared to recommend universal bacterio-logical screening and most units (78%) had protocols recommending a risk-basedintrapartum antibiotic prophylaxis strategy. The other units (22%) recommended acombination of risk-based IAP and risk-based bacteriological testing, in whichwomen with certain risk factors had a bacteriological test and received IAP if foundto be GBS-positive.

The risk factors for which IAP was recommended in the protocols generally reflectedthe guideline, with 93% of units recommending IAP for women with a previous babyaffected by early-onset GBS disease, and around 85% recommending IAP for GBSbacteriuria or an incidental finding of GBS in the current pregnancy. A lowerproportion of protocols recommended IAP for fever during labour, preterm labourand prolonged rupture of membranes (61%, 49% and 52%, respectively). Thesethree risks were often included in combinations of risks for which IAP was advised.However, while the RCOG guideline states that the argument for prophylaxis

ix

becomes stronger with two or more risk factors, many combinations contained in theprotocols did not conform to the options presented in the guideline. Another notabledifference between the protocols and guideline was the inclusion in 29 unit protocolsof ‘maternal GBS colonisation in a previous pregnancy’ as an indication for IAP,although the RCOG guideline states that there is no good evidence to support thispractice.

The primary antibiotic regimens specified in the protocols generally conformed to theRCOG guideline, with 74% of units recommending penicillin G (now termedbenzylpenicillin) with a 3 g loading dose, followed by 1.5 g every 4 hours. The mainvariants, typically different doses of benzylpenicillin rather than an alternativeantibiotic, did not appear evidence-based. For women allergic to penicillin, 83% ofunit protocols specified the recommended clindamycin regimen.

Reported practice

A similar pattern of practice was observed in the responses from midwives andobstetricians, being broadly consistent with the RCOG guideline. Only 2% ofmidwives and obstetricians reported using a universal bacteriological screening, while18% of midwives and 24% of obstetricians said that they routinely took swabs totest for GBS in selected groups of women (risk-based bacteriological testing).

Interestingly, a higher proportion of clinicians reported taking swabs for GBS whenasked about their practice in specific clinical situations. This result was unexpectedfor women with a previous baby affected by GBS disease or who had GBS bacteriuriain the current pregnancy. IAP is recommended by almost all international guidelinesfor a woman with these risk factors, regardless of whether or not she is colonisedwith GBS. Consequently, a significant proportion of clinicians may be incorrectlyinterpreting a negative swab as indicating that IAP is unnecessary.

In terms of which women are offered IAP, reported practice generally reflected theRCOG guideline. Over 90% of clinicians would offer IAP to women with one of therisk factors with the strongest recommendations (previous baby with early-onset GBSdisease, GBS bacteriuria or incidental GBS finding in current pregnancy). There wasgreater variation between units in the proportions of clinicians who would offer IAPfor intrapartum fever, preterm labour or prolonged rupture of membranes, a patternconsistent with that observed among the protocols for these risk factors. In particular,less than 50% of midwives and obstetricians would offer IAP for preterm labour.Another area of practice that differed from the RCOG guideline concerned womenwho were colonised with GBS in a previous pregnancy. IAP is not recommended forthese women. However, among those clinicians who would not take a swab if thisrisk factor was present, 70% of midwives and 75% of obstetricians reported thatthey would offer IAP. If a woman in this situation had a swab that proved GBSpositive, proportion of clinicians who would offer IAP exceeded 90%.

There was also some variation in reported practice regarding the antibiotic prophyl-axis regimen. While over 60% of respondents followed the recommended benzyl-penicillin regimen, a sizeable majority reported using penicillin with an alternativedose. There was also some variation in the regimens used for women with a penicillinallergy, with the recommended clindamycin regimen reportedly used by 73% ofmidwives and 64% of obstetricians.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

x

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Finally, the audit found variation in the management of neonates at risk of GBSdisease. Clinicians either preferred to assess and observe ‘at risk’ babies or to take aculture and start antibiotic treatment while awaiting the result. This variation reflectsthe lack of available evidence to guide practice in this area and current practicefollowed the established patterns of care described in the RCOG guideline.

Conclusion

The practice reported by midwives and obstetricians is in broad agreement with therisk-based IAP strategy described in the guideline. There would appear to be a slightimprovement in the proportion of units offering IAP to appropriate women since theprevious surveys in 1999 and 2001. Nonetheless, variation in reported practice wasevident in each aspect of the care process. This variation would appear to arise froma number of sources. First, the variation in which women are offered IAP seems toincrease as the benefits of IAP becomes smaller or more uncertain. This uncertaintyis associated with a lack of research evidence on the risk posed by individual factorsor combinations thereof. Second, some variation in practice is probably due to thelimitations of the RCOG guideline. In particular, some recommendations werephrased in terms of the quality of the evidence rather than specific behaviour andcould therefore lead to inconsistencies in their interpretation. A third source ofvariation is the variety of recommendations contained in the protocols. While someof these differences may reflect local conditions, others appear to be due to flaws inthe development process. Finally, variation may be due to difficulties inherent inimplementing a risk-based IAP strategy.

Recommendations

1. Obstetric units should continue to offer intrapartum antibiotic prophylaxis towomen with risk factors in accordance with the RCOG guideline.

2. A local GBS protocol should be readily available to staff and units shouldensure that it is interpreted and implemented consistently.

3. Units should ensure that their protocols are up to date and consistent withnational guidelines albeit adapted for the local context. Protocols should alsocite the most important sources of evidence and give the date of developmentand review.

4. When revising the RCOG Green-top guideline, care should be taken to ensurethat recommendations are unambiguous and comprehensive.

5. The revised RCOG Green-top guideline should include clearly defined auditcriteria.

6. Research should be commissioned to fill the current gaps in the evidence baseon which strategies to identify women for IAP are most effective in preventingearly-onset GBS disease in the UK.

Executive Summ

ary

xi

Executive Summ

ary

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

1. Introduction

1.1 Background

Group B streptococcus (GBS) or Streptococcus agalactiae is a leading causeof early-onset neonatal infection, resulting in sepsis, pneumonia and menin-gitis.3 Early-onset GBS disease is generally defined as an infection appearingwithin 7 days of a baby being born, although 90% of cases occur within 24hours.4 Infection is predominantly caused by exposure to maternal GBSduring childbirth5 but the risk is increased if the mother has one or more ofthe following factors:

� previous infant with early-onset GBS disease

� GBS bacteriuria in current pregnancy or GBS colonisation at term

� prolonged rupture of membranes (an interval of 18 hours or morebetween rupture and delivery)

� preterm labour at less than 37 weeks of gestation

� maternal temperature higher than 38˚C.

The transmission of GBS from mother to baby can be reduced with the use ofintrapartum antibiotic prophylaxis (IAP).6,7 There is, however, uncertaintyabout the most effective way of identifying women who should receive IAP.One approach is to offer IAP to women with one or more risk factors.Another approach is to screen all pregnant women using a bacteriological testat 35–37 weeks and offer IAP to women who are found to be colonised withGBS. It is also possible to combine these two methods to form otherstrategies.

The prevalence of early-onset GBS diease in the UK is not precisely known,nor is the number of neonatal deaths that it causes. During a surveillancestudy from February 2000 to February 2001, the reported prevalence ratewas 0.48 cases/1000 births, with a 10% mortality rate.3 However, the studywas recognised as suffering from possible under-reporting and the totalburden of disease in the UK may be two to three times higher. Rates are alsolikely to vary across the country because of regional differences in the localpopulation and patterns of treatment.8

Surveys into the practice of all UK obstetric maternity units with respect tothe prevention and intrapartum management of GBS disease were conductedin 1999 and 2001.2 These showed that the overwhelming majority of unitsoffered IAP based on whether or not women had particular risk factors forearly-onset GBS disease. Only a few units (2–3%) offered IAP based on theresults of universal bacteriological screening. However, in 2001, only 20%of units indicated that they used one of the antibiotic regimens recom-mended by the American College of Obstetricians and Gynecologists.9

Among the other units, approximately 55% used the appropriate antibiotic

1

but with an incorrect dosage, while approximately 23% used an alternativeantibiotic incorrectly.

The variation in practice identified by these surveys highlighted theuncertainty among obstetricians about the most effective way to preventearly-onset GBS disease. This may have been partly due to the lack of a UKguideline, which was not published until June 2001.10 Before then, the onlyclinical guidelines available had been developed overseas, the mostprominent by the US Centers for Disease Control and Prevention (CDC). Inits 1996 guideline,11 the CDC had recommended that clinicians selectwomen as candidates for intrapartum antibiotic prophylaxis using eitheruniversal bacteriological screening or risk-based strategy. However, therewas uncertainty about whether such prevention strategies should be adoptedin UK maternity units10 and the results of the surveys by Kenyon et al.8

suggested that this guideline had not been widely adopted.

The UK guideline published in 2001 was produced by the Public HealthLaboratory Service (PHLS) Group B Streptococcus Working Group.10 Itrecommended IAP for women with specific risk factors but did not supportuniversal bacteriological screening. However, it was regarded as an ‘interim’guideline because of the paucity of UK data on maternal GBS carriage, neo-natal infection rates and attributable mortality and morbidity. Consequently,a national surveillance study and a case–control study of risk factors wereinitiated to overcome these deficiencies. Guidance was finally provided in2003, when the Royal College of Obstetricians and Gynaecologists pub-lished its Green-top guideline on the prevention of early-onset neonatal GBSdisease.1 However, it is not clear to what extent practice has shifted inresponse to its publication.

1.2 Aim and objectives of the audit

An audit was established to evaluate practice in UK obstetric units on theprevention of early-onset neonatal GBS diease against the recommendationsof the RCOG guideline. The audit aimed to survey all NHS and independentsector obstetric units in the UK and to examine the organisation ofscreening, the use of intrapartum prophylaxis and the management ofneonates born with increased risk of GBS disease.

Specific objectives were:

� to carry out an international comparison and evaluation of existingnational guidelines on the prevention of early-onset GBS disease

� to assess the consistency of the local clinical protocols provided byNHS and independent sector obstetric units in the UK

� to evaluate practice on preventing neonatal GBS disease against theRCOG guideline and to assess whether it has changed since thesurveys carried out in 1999 and 2001.

In this report, we describe the results of this audit.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

2

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

1.3 Terminology

Various terms are used in the literature to describe the different strategiesthat are used to prevent the transmission of GBS from a mother to her baby.In this report, the following terms are used:

Universal bacteriological screening Used to refer to the practice of takingswabs from all pregnant women,which are then cultured to seewhether or not GBS is present.Alternative phrases for this activityinclude routine antenatal screening,routine microbiological screening,universal screening, universalprenatal screening, prenatal culturescreening, culture-based screeningand a (universal) screening-basedapproach to prevention.

Risk-based bacteriological testing Used to refer to the practice of takingswabs in selected groups of women,the groups being defined by thepresence of a defined set of riskfactors.

Risk-based IAP Used to refer to the practice ofoffering intrapartum antibioticprophylaxis to women who have oneor more GBS risk factors. Alternativephrases for this activity include risk-based screening, risk-basedassessment, and a risk-basedapproach to prevention. The term‘risk factor-based’ can be used insteadof ‘risk-based’.

It is also worth noting that ‘universal bacteriological screening’ will be usedto refer to such strategies that include risk-based IAP when a culture resultis unknown. Similarly, ‘risk-based bacteriological testing’ will be used torefer to strategies that include risk-based IAP to cover situations in whichtaking a swab is not indicated.

Introduction

3

Introduction

2. An international comparison ofguidelines on the prevention ofearly-onset GBS disease

2.1 Summary of the RCOG Green-top guideline

The RCOG guideline aimed to provide guidance for obstetricians, midwivesand neonatologists.1 Recommendations were formulated on screening, IAPand the management of newborn infants (Box 1). Few recommendationsreceived a strong grading because of the limited number of randomisedcontrolled trials.

Box 1. Recommendations from RCOG Green-topGuideline no. 36 (summary version)

ANTENATAL

1. Routine screening (either bacteriological or risk based) for ante-natal GBS carriage is not recommended.

2. Antenatal treatment with penicillin is not recommended.

INTRAPARTUM

1. Clinicians should discuss the use of intrapartum antibiotic prophyl-axis in the presence of known risk factors including incidentalcarriage. Risk factors include:�� prematurity less than 37 weeks

�� prolonged rupture of membranes 18 hours or more

�� fever in labour higher than 38˚C.

The argument for prophylaxis becomes stronger in the presence oftwo or more risk factors.

2. Intrapartum antibiotic prophylaxis should be considered if GBS isdetected incidentally in the vagina or the urine in the current preg-nancy.

3. Intrapartum antibiotic prophylaxis should be offered to womenwith a previous baby with neonatal GBS disease.

4. There is no good evidence to support the administration of intra-partum antibiotic prophylaxis to women in whom GBS carriagewas detected in a previous pregnancy or to women undergoingplanned caesarean section.

5. Antibiotic prophylaxis for GBS is unnecessary for women withpreterm rupture of membranes unless they are in established labour.

4

CB

C

C

C

C

C

PROPHYLAXIS REGIMENS

1. If antibiotic prophylaxis is used, benzylpenicillin (penicillin G)should be administered as soon as possible after the onset of labourand at least 2 hours before delivery. The recommended dose is 3 gintravenously followed by 1.5 g every 4 hours during labour. Intra-venous clindamycin 900 mg 8-hourly should be administered tothose women allergic to penicillin.

2. If chorioamnionitis is suspected, broad-spectrum antibiotic therapy,including an agent active against GBS, should replace GBS-specificantibiotic prophylaxis.

NEONATAL

1. Newborn infants with clinical signs compatible with infectionshould be treated promptly with broad-spectrum antibiotics whichprovide cover against early-onset GBS disease as well as othercommon pathogens.

2. For an infant whose mother had a previous infant with GBS diseaseor has other risk factors for early-onset GBS (especially more thantwo), either clinical evaluation after birth and observation for atleast 12 hours are necessary, or blood cultures should be obtainedand the infant treated with penicillin until the culture results areavailable.

3. Postnatal antibiotic prophylaxis is not recommended for low-riskterm infants.

4. It is not necessary to perform routine surface cultures or bloodcultures on well infants.

5. There is no evidence to suggest that mode of infant feeding affectsthe risk of neonatal GBS disease and women concerned about late-onset disease should be given the usual advice about breastfeeding.

NB: Grade of recommendation indicated in brackets, see Appendix 1 forexplanation.

Many recommendations were similar to those in the PHLS interim guide-line.10 Universal bacteriological screening was not recommended. IAP wasrecommended for women with the major risk factors: GBS bacteriuria in thecurrent pregnancy, suspected chorioamnionitis and a previous baby with neo-natal GBS disease. In addition, to guide clinical decisions, the RCOG guide-line included estimates on the numbers of women needed to receive IAP inorder to prevent one case of GBS for other risk factors (reproduced in Table1); the presence of two or more risk factors was recognised as strengtheningthe argument for prophylaxis.

Both guidelines recommended the same antibiotic regimens for GBS, althoughthe RCOG guideline changed the minimum time IAP should be administeredfrom 4 hours to 2 hours. The RCOG guideline also only recommended broad-

An international comparison of guidelines on the prevention of early-onset G

BSdisease

5


BSdisease

B

A

C

C

B

C

✔

spectrum antibiotics for women with suspected chorioamnionitis (but, unlikethe interim guideline, did not recommend this for women with prelabourprolonged rupture of membranes).

2.2 Comparison of international guidelines for early-onsetGBS disease

A systematic search was performed for English-language GBS guidelines (seeAppendix 2 for search strategy). The search identified 13 publications bynational bodies from the USA, Canada, Australia, New Zealand, France andBelgium (Table 2). One guideline originated from a UK patient support group.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

6

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Table 1. Estimate of numbers needed to treat for identified risk factors, based on 680 000 births/year,incidence of GBS in UK of 0.5/1000 life births, 80% effectiveness of intrapartum antibiotic prophylaxis(IAP) in preventing group B streptococcus (GBS)

Risk factor Prevalence of GBS cases NNT with GBS risk factor (%) treated/ IAP to cases/year

10 000 prevent in UKIn In babies untreated 1 case

pregnancy with GBS women GBS

Intrapartum fever > 38°C 1.6 19.0 60 208 52

Preterm labour < 35 weeks 3.2 22.0 35 357 61

Preterm labour < 37 weeks 7.4 37.0 25 500 101

Prolonged rupture of membranes at term (≥ 18 hours) 8.0 34.0 21 595 91

NNT = number needed to treat

Table 2. English-language guidelines on the prevention of early-onset neonatal GBS disease

Country Author of guideline Abbrev. Date of guideline

Current Original

Canada Canadian Task Force on Preventive Health Care CTF 200120

France Agence Nationale d’Accréditation et d’Evaluátion en Santé ANAES 200119

USA Centers for Disease Control and Prevention CDC 200212 199611

USA American College of Obstetricians and Gynecologists ACOG 20029 1997

Belgium Conseil Supérieur d’Hygiène (Belgian Health Council) CSH 200318

Canada British Columbia Reproductive Care Program BCRCP 200316

UK Royal College of Obstetricians and Gynaecologists RCOG 20031

USA American College of Nurse-Midwives ACNM 200315 1997

Canada Society of Obstetricians and Gynaecologists of Canada SOGC 200417 1997

New Zealand New Zealand College of Midwives NZCOM 200413

New Zealand New Zealand GBS Consensus Working Group NZCWG 200422

Australia and The Royal Australian and New Zealand New Zealand College of Obstetricians and Gynaecologists RANZOG 200521 2003

UK Group B Strep Support Organisation GBSS Not stated14

The guideline published by the CDC seemed the most influential.12 Thecurrent or superseded CDC guideline was cited by all but two guidelines, theexceptions being the NZCOM13 and GBSS14 publications. The guidelinesfrom ACOG,9 ACNM,15 BCRCP16 and SOGC17 essentially followed therecommendations in the 2002 CDC guideline. It was also endorsed by theAmerican Academy of Pediatrics.

Recommendations on strategies for identifying women to be offeredIAP

The guidelines recommended various strategies for selecting which womenshould be offered IAP (Table 3):

� Seven guidelines (CDC,12 ACOG,9 ACNM,15 SOGC,17 BCRCP,16 CSH18

and ANAES19) recommended universal bacteriological screening, inwhich the offer of IAP was predominantly based on the results acultured swab.

� The CTF guideline20 recommended universal bacteriological screeningbut women would only be offered IAP if they had a positive GBSculture and another risk factor.

� The RANZOG guideline21 gave no strong preference but discussedhow to implement a strategy based on universal bacteriologicalscreening in more detail.

� Four guidelines (RCOG,1 GBSS,14 NZCOM,13 NZCWG22)recommended using a risk-based IAP strategy. However, the NZCWGguideline22 suggested following the CDC strategy if a woman opts fora bacteriological screen.

Among guidelines recommending universal bacteriological screening, therewas general agreement in sample collection and microbiology recommend-ations. All but one guideline recommended swabs be taken between 35–37weeks from the lower vagina and rectum and inoculated into a selective brothmedium. The exception (ANAES)19 recommended that swabs be takenbetween 34–38 weeks from the whole vaginal cavity without use of a selectivebroth medium.

The different strategies result in different proportions of pregnant womenbeing offered IAP.1,22 The proportion of women in labour estimated to receiveIAP under the universal bacteriological screening recommended by the CDCis 30–35%. The proportion is estimated to be 15–20% under the risk-basedIAP strategy.

Despite this, the strategies for selecting women for IAP shared somesimilarities:

� All but two guidelines (CTF,20 NZCOM13) recommended IAP forwomen who had a previous baby with neonatal GBS disease or forwomen who had bacteriuria in the current pregnancy, regardless ofthe result of a cultured swab.


BSdisease

7


BSdisease

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

8

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Tabl

e 3.

Sum

mar

y of

gui

delin

e re

com

men

datio

ns fo

r sel

ectin

g w

omen

for i

ntra

part

um a

ntib

iotic

pro

phyl

axis

(IAP

)

Sour

ce

UBS

Pr

evio

us b

aby

Bact

eriu

ria

Posi

tive

Inci

dent

alIn

trap

artu

mPr

eter

mPr

olon

ged

GBS

inPl

anne

d CS

PPRO

Mw

ith

GBS

G

BS

findi

ng o

ffe

ver

labo

urRO

Mpr

evio

usdi

seas

e cu

ltur

eG

BS(<

37 w

eeks

)(≥

18 h

ours

)pr

egna

ncy

colo

nisa

tion

RCO

G1

Not

reco

mm

ende

d Y

Y N

A Y

DD

D N

o N

o N

o

GBS

S14N

ot s

tate

dY

YN

AY

YY

YN

o N

o Y

NZC

OM

13N

ot re

com

men

ded

DD

NA

?D

DD

??

?

NZC

WG

22N

ot re

com

men

ded

YY

NA

YY

YY

?N

oN

o

RAN

ZOG

21Po

ssib

leY*

Y*Y

NA

CUCU

CU?

No

?

CDC12

Reco

mm

ende

dY*

Y*Y

NA

CUCU

CUN

oN

oN

o

ACO

G9

Reco

mm

ende

dY*

Y*Y

NA

CUCU

CUN

oN

oN

o

ACN

M15

Reco

mm

ende

dY*

Y*Y

NA

CUY*

CUN

oN

oN

o

CTF20

Reco

mm

ende

dY*

C+N

oN

AC+

C+C+

??

?

SOG

C17Re

com

men

ded

Y*Y*

YN

AY*

CUCU

??

CU

BCRC

P16Re

com

men

ded

Y*Y*

YN

ACU

CUCU

No

No

?

CSH18

Reco

mm

ende

dY*

Y*Y

NA

CUCU

CUN

oN

oN

o

ANAE

S19Re

com

men

ded

Y*Y*

YN

ACU

CUCU

a?

?N

o

CS =

cae

sare

an s

ectio

n,G

BS =

gro

up B

str

epto

cocc

us,P

PRO

M =

pre

term

pre

labo

ur ru

ptur

e of

mem

bran

es,R

OM

= ru

ptur

e of

mem

bran

es,U

BS =

uni

vers

al b

acte

riolo

gica

l scr

eeni

ng;a

Prol

onge

d RO

M d

efin

ed a

s >

12 h

ours

KEY:

Y =

Offe

r or s

houl

d co

nsid

er IA

P;Y*

= O

ffer I

AP re

gard

less

of c

ultu

re re

sult;

No

= D

o no

t offe

r IAP

;C+

= O

ffer I

AP o

nly

if cu

lture

resu

lt po

sitiv

e;CU

= O

ffer I

AP if

cul

ture

resu

lt un

know

n or

pos

itive

;D

= D

iscu

ss IA

P w

ith w

oman

;? =

Not

sta

ted;

N/A

= N

ot a

pplic

able

� In the seven guidelines recommending universal bacteriologicalscreening, risk-based IAP was recommended when the results of aculture were unknown.

� There was general agreement that IAP for GBS is not required for:�� women undergoing a planned caesarean section (in the absence

of labour and ruptured membranes), regardless of GBS carriage�� women with a history of GBS colonisation in a previous

pregnancy.

If a woman suffered preterm prelabour rupture of membranes, seven guide-lines recommended that antibiotic treatment is not required before labour hasbegun. However, when birth was imminent, guidance was not prescriptive.The CDC and related guidelines suggested antibiotic treatment unless theresults of a cultured swab showed that a woman did not have GBS colon-isation. The RCOG advised that antibiotic treatment be considered especiallyif ROM occurred before 37 weeks.

Recommended antibiotic regimens

The primary antibiotic regimen recommended by all but two guidelines wasbenzylpenicillin (penicillin G), with a loading dose of 3 g intravenously,followed by 1.5 g every 4 hours until delivery. The exceptions were theNZCWG22 and NZCOM13 guidelines; both recommended benzylpenicillin,with an initial dose of 1.2 g intravenously, followed by 0.6 g every 4 hours.Although ampicillin was regarded as an acceptable alternative in manyguidelines, benzylpenicillin was preferred because of its narrow spectrum.The RCOG guideline recommended that the use of broad-spectrum anti-biotics such as ampicillin should be avoided if possible. However, the needfor a broad-spectrum antibiotic was generally recognised for women withsuspected chorioamnionitis. Nine guidelines included this recommendation(RCOG,1 CDC,12 ACOG,9 ACNM,15 CTF,20 SOGC,17 BCRCP,16 CSH,18

NZCWG22).

Eleven guidelines also considered the requirements of women with apenicillin allergy:

� nine recommended a regimen of clindamycin 900 mg intravenously 8-hourly, while one other suggested clindamycin 600 mg intravenously8-hourly

� erythromycin or vancomycin were recommended as alternatives toclindamycin, especially for women with a history of anaphylaxis orimmediate hypersensitivity

� cefazolin was recommended for women with a penicillin allergy but alow risk (no history) of anaphylaxis.

Neonatal management

Recommendations for neonatal management were provided for eight of the 13guidelines (RCOG,1 GBSS,14 NZCOM,13 NZCWG,22 CDC,12 SOGC,17 BCRCP,16

9


BSdisease


BSdisease

CSH18). For neonates with clinical signs of sepsis, most recommended animmediate evaluation and treatment while awaiting culture results. Guidelinesgenerally suggested that a single management strategy was not advisable withtreatment decisions dictated by clinical circumstances. Recommended anti-biotic treatments were similarly varied or imprecise, and only two guidelinesgave recommendations on treatment duration (GBSS,14 CSH18).

The guidelines also made recommendations for managing neonates withoutclinical signs of sepsis but considered to be at risk of early-onset GBSdisease. Being ‘at risk’ was not consistently defined, with guidelines usingdifferent algorithms that consisted of different criteria. The criteria included:

� mother with suspected chorioamnionitis (NZCOM,13 NZCWG,22

CDC,12 SOGC,17 BCRCP,16 CSH18)

� maternal GBS risk factors (RCOG,1 NZCWG22)

� maternal antibiotic prophylaxis administered for less than theminimum appropriate time (GBSS,14 NZCOM,13 NZCWG,22 CDC,12

SOGC,17 BCRCP,16 CSH18)

� preterm prelabour rupture of membranes (CSH18)

� neonate born at less than 35 weeks of gestation (GBSS,14 NZCWG,22

CDC,12 BCRCP,16 CSH18).

Guidance on neonatal treatment also varied between guidelines, with strat-egies including observation, limited evaluation, full diagnostic evaluation andtreatment while awaiting culture results.

Exploratory analysis to identify reasons for differences inrecommendations

An analysis of the guideline development process and the recommendationswas undertaken with the aim of identifying reasons for differences in theguidelines and the recommendations on which women should receive IAP.The quality of the guideline development process was analysed using theAppraisal of Guidelines for Research & Evaluation (AGREE) tool.23 Thistool assesses guidelines on 23 criteria, grouped into six dimensions (clarityof purpose, stakeholder involvement, rigour of development, clarity ofpresentation, applicability and editorial independence).

In terms of rigour of development, there was no apparent associationbetween the described development process and the formulated recom-mendations. Nonetheless, there was considerable variation in the quality ofthe process described. Only three guidelines achieved high scores: the CTF20

(score 25/28), NZCWG22 (score 21/28) and RCOG1 (score 20/28). All otherguidelines scored between 7/28 and 12/28. On specific criteria in this domain:

� only three guidelines (CDC,12 NZCOM13 and RCOG1) were externallyreviewed by experts prior to publication

� eight guidelines gave no indication of systematically searching forevidencePr

even

tion

of E

arly

-ons

et N

eona

tal G

roup

B S

trep

toco

ccal

Dis

ease

in U

K O

bste

tric

Uni

ts

10

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

� Only one guideline (CTF)20 clearly described their selection criteriaused in the literature review

� Only two guidelines clearly described the methods for formulatingrecommendations (CTF and NZCWG)20,22

� Only four guidelines (RCOG,1 CDC,12 CTF20 and NZCWG22)adequately formed an explicit link between their recommendationsand the supporting evidence.

It was clear that the interpretation of the evidence was influential informulating some recommendations. This was particularly evident in relationto the recommendations about how to select women for IAP. The CDCrevised its guideline in 2002 to only recommend universal bacteriologicalscreening, largely due to the results of a large population-based study thatfound universal screening prevented more cases of GBS disease than a risk-based approach.24 The CTF20 and ANAES19 guidelines were the only onesproduced before this study was published.

The CDC guideline recommendation for universal bacteriological screeningwas graded A ‘strongly recommended’ and the evidence was rated II: ‘atleast one well designed clinical trial without randomisation; cohort or casecontrolled analytic studies (preferably from more than one centre); multipletime-series studies; dramatic results from uncontrolled studies; or someevidence from laboratory experiments’.12 In contrast, the NZCWG22 andRCOG1 recommendations for risk-based IAP were less strong, although theyreferenced the key publication used by the CDC to reach its recommend-ation. Both gave reasons that would limit the applicability of the study toNew Zealand and the UK, respectively, including:

� the prevalence of the disease

� the accuracy of the bacteriological test

� the effectiveness of the antibiotics and the risk of antibiotic resistance

� the financial costs of universal bacteriological screening

� the cultural perspective on the degree to which childbirth should be‘medicalised’.

Consequently, the RCOG rated the evidence to support its recommendationagainst universal bacteriological screening as IV: ‘Evidence obtained fromexpert committee reports or opinions and/or clinical experience of respectedauthorities’. The NZCWG guideline still rated its evidence as B: ‘at least onewell-designed, non-randomised clinical trial; cohort or case–control studies,or multiple time-series experiments’.22

Conclusion

The RCOG guideline is one of a number of international guidelines that makerecommendations on the prevention of early-onset GBS disease. It is generallyconsistent with the guideline produced by the other UK organisation (GroupB Strep Support), with both suggesting IAP for similar groups of women with

11


BSdisease


BSdisease

risk factors and both recommending the same antibiotic regimens. TheRCOG recommendation for a risk-based IAP strategy is also shared withguidelines published in New Zealand.21,22

There is, however, no international consensus regarding which of the variousprevention strategies health professionals should adopt. In particular, incontrast to the UK and New Zealand guidelines, the guidelines from mostother countries recommended identifying women for IAP using universalbacteriological screening. Depending upon which guidelines are selected byUK maternity units, differences could arise in terms of:

� the criteria for selecting women to receive IAP

� the dose of the primary antibiotic

� the minimum duration of prophylaxis

� the groups of women who should receive broad-spectrum antibiotics.

Consequently, the practice at a unit could differ substantially from thatrecommended by the RCOG. In particular, the proportion of women whoreceive IAP would be expected to be higher at any unit adopting the influentialCDC guideline in comparison with those adopting the RCOG guideline.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

12

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

3. Audit design and data collection

3.1 Audit organisation

The audit was designed to evaluate practice in UK obstetric units on theprevention of early-onset neonatal GBS against the recommendations of theRCOG guideline. It was established as a collaborative project between theRCOG and the Health Services Research Unit of the London School ofHygiene and Tropical Medicine. Funding was provided by the NationalScreening Committee.

Various stakeholders were consulted on the design of the audit. These included:

� the Health Protection Agency

� the National Screening Committee

� the Association of Medical Microbiologists

� the Royal College of Midwives (RCM)

� the Royal College of Paediatrics and Child Health (RCPCH)

� the RCOG Guidelines and Audit Committee

� the RCOG Consumers’ Forum

� other academic institutions with expertise in this area, notably theNational Perinatal Epidemiology Unit, Oxford.

Stakeholders were also given the opportunity to comment on the draftreport.

Ethics approval was not required for this project as it was a national audit.The audit was developed in consultation with the Health and Social CareInformation Centre Review of Central Returns Committee, who consideredthe data collection to be useful and reasonable.

3.2 Audit participants

All NHS and independent obstetric maternity units in the UK were eligibleto participate in the audit. A contact database was compiled using theRCOG database of maternity unit clinical directors. Its contents wereverified and updated using information published by BirthChoiceUK, DrFoster and the Hospital Episode Statistics database. The information onindependent healthcare providers was checked against the list published bythe Healthcare Commission.

All clinical directors were contacted in December 2005. A letter explainedthe purpose of the audit and clinical directors were asked to supply thenames and contact details of a senior midwife, lead obstetrician and leadneonatologist to complete a questionnaire on their current practice for

13

preventing early-onset GBS disease. Nominated individuals had the optionof receiving the questionnaire via the postal service or in an electronicformat via email. The clinical director was also asked to provide a copy ofthe local labour ward protocol on the prevention of early-onset GBS disease.The nominated midwives, obstetricians and neonatologists were sent thequestionnaire in February 2006. Both the clinical directors and thenominated clinicians received two reminders. These were sent at approxi-mately 4-week intervals. If no response was received, a follow-up telephonecall was made.

Information about the audit was posted on the websites of the RCOG andthe RCM and was included in an email newsletter to members of theRCPCH. It was also advertised at the annual meeting of maternity unitclinical directors held at the RCOG in November 2005.

3.3 The questionnaire

A questionnaire was developed to assess the agreement between reportedpractice and the recommendations of the RCOG guideline. It was based onthe previous surveys used by Kenyon et al.,2 so that the results from thecurrent audit could be compared to previous patterns of practice. Additionalquestions were added to collect contextual information and to improve itscorrespondence with the various topics covered in the RCOG guideline.Respondents were asked about:

� the selection strategies used to identify women who should be offeredIAP

� the IAP regimens administered

� the preferences of pregnant women for prophylaxis

� the management of neonates at risk of early-onset GBS disease

� the characteristics of the maternity unit

� the local incidence of early-onset GBS disease.

Finally, five case vignettes were included in the questionnaire describingclinical scenarios for which the screening and treatment policies ranged fromuncontroversial to controversial. The purpose of this exercise was first tovalidate the responses to the survey questionnaire and second to investigatein more detail the professionals’ interpretation of the recommendations inthe context of individual patient care.

3.4 Pilot of the questionnaire

The questionnaire was piloted among midwives, obstetricians andneonatologists at four hospitals. A total of 18 clinicians participated in thepilot and the questionnaire was revised in response to the comments receivedon its clarity and speed/ease of completion.

The midwife questionnaire used in the audit is reproduced in Appendix 3.The obstetrician questionnaire did not contain the section on management

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

14

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

of neonates as feedback from obstetricians in the pilot suggested that thisaspect of care was not within their usual role. Likewise, the neonatologistquestionnaire did not contain the questions relating to the antenatal andintrapartum management of pregnant women. The midwife questionnaireincluded all sections because a midwife could be involved in all aspects ofcare.

3.5 Collection and analysis of the protocols and audit data

A Microsoft Access® database was developed to hold the data collected onthe protocols and from the audit questionnaires. Data from the electronicquestionnaires were imported directly into the database. The responses frompaper questionnaires were entered centrally using a structured Access form.

The content of each protocol was assessed using a structured template.Information was extracted on:

� the protocols’ date of development and the quality of the protocols’evidence base

� the recommended strategy for selecting women for IAP

� the timing and swab sites advised for bacteriological screening, if used

� the risk factors for which women were recommended to receive IAP,either individually or in combination

� the antibiotic regimen for particular groups of women

� whether the protocol mentioned women’s preferences or requests for abacteriological screen.

The analysis of the questionnaire and vignette data derived the proportionof participants (or maternity units) undertaking an action. Denominatorsreflect the number of participants answering the question. These could varybetween questions owing to non-responses, erroneous data or because thequestion only applied to a subset of units.

Data were analysed in Microsoft Excel® and Stata® v8.

15

Audit design and data collectionAudit design and data collection

4. Content of maternity unitprotocols

We asked the clinical directors of 227 maternity units with obstetric facilitiesfor a copy of their protocol on the prevention of early-onset GBS disease. Aresponse was received from 177 units, with 161 units sending the audit theirprotocol. Only four stated that their unit did not have a protocol, while 12units stated that they had a protocol but did not provide the audit with acopy.

Among the 161 returned protocols, 11 protocols were received twice fromdifferent units in the same NHS trust. With one exception, the protocolsreceived from multi-unit trusts were written as trust-level documents.Consequently, if a protocol had only been returned by only one unit withina multi-unit trust, this protocol was assumed to also apply to the other units.This enabled the policies of 171 maternity units to be analysed.

4.1 Characteristics of the unit protocols

Among the units that returned a protocol, almost all developed their owndocument from available guidelines and published research. Two unitsindicated they used the RCOG guideline,1 while the PHLS10 and GBSS14

guidelines were each supplied by a unit once.

The protocols from 156 of the 171 units contained a development date. Justover 75% had been developed in 2004 or later (120 of 156). As the RCOGguideline was published in November 2003, it is reasonable to assume thatthe guideline would have been available to those developing or revising theseprotocols. It was cited in 96 (80%) of the 120 protocols. Of the 29 protocolsdeveloped in 2003, ten cited the final or draft RCOG guideline. Seven unitsused protocols developed before 2003.

Evidence to support the recommendations was cited by 150 protocols(88%). The number citing key publications is summarised in Table 4. ThePHLS guideline,10 which the RCOG guideline superseded, was cited in one-third of protocols. Surprisingly, 47 of these protocols were published in2004 or later, after the PHLS guideline had become obsolete, although theRCOG guideline was also referenced in 39 of these 47 protocols. Similarly,the CDC 1996 guideline11 was cited by 11 protocols published after 2003,by which time it had been superseded by the 2002 revision.12

The date for review was included in 117 (68%) unit protocols, although intwo cases this was not accompanied by a date of publication. Most units hadprotocols with a shelf-life of 2 or 3 years, respectively, 28 and 77 protocols.Eight units had protocols that were scheduled for review prior to 2006 butwere yet to be revised.

16

4.2 Strategies for identifying women for intrapartumantibiotics

The protocols from 78% of the 171 units advocated a risk-based IAP strategyfor identifying women who should be offered IAP. The protocols from 37units (22%) recommended using a risk-based bacteriological testing strategyand no unit had a protocol that recommended universal bacteriologicalscreening. Of the 134 units that adopted a risk-based IAP strategy, 58explicitly advised against offering universal bacteriological screening.

There were no common sets of risk factors used to indicate which women toswab among the 37 units with protocols recommending risk-based bacterio-logical testing. The most frequent strategy involved only a single risk factor,as follows:

1. Swab a woman who had been GBS colonised in a previous pregnancyand offer IAP if the cultured swab was GBS positive.

2. Offer IAP to a woman who had one or more risk factor.

This approach was recommended at seven units. Table 5 describes thefrequency with which risk factors were defined as indicating when a swabshould be taken.

In all but three maternity units, a woman would be advised to have IAP ifthe cultured swab was GBS positive or if she had one or more risk factors.The exceptions were as follows:

� At two units, IAP should be offered if (1) the cultured swab waspositive; (2) the culture result was unknown and the woman hadeither a fever, preterm labour or prolonged ROM; or (3) the womanhad bacteriuria or suspected chorioamnionitis regardless of the cultureresult. This was similar to the CDC strategy but the protocol did notadvise all women to be offered bacteriological screening.

� At one other unit, IAP should be offered if the cultured swab wasGBS positive or when the culture result was unknown, if the womenhad a risk factor.

17

Table 4. Frequency with which selected publications were cited in unit protocols

Publication (year of publication) Units citing publications

n %

RCOG (2003)1 111 65

PHLS (2001)10 64 37

GBSS (n/s)14 49 29

CDC (2002)12 27 16

CDC (1996)11 15 9

Kenyon et al. (2001)27 23 13

Kenyon et al. (2001)28 17 10



The circumstances in which IAP should be offered to women recommendedby the protocols were typically a selection of the accepted risk factors (Table6). In addition, while no protocol recommended IAP for a planned caesareansection, 88 protocols (51%) included a statement against offering IAP in thissituation.

Whether or not a specific risk factor was included in a protocol as anindication for IAP did not appear to depend upon whether the protocolrecommended a risk-based IAP strategy or risk-based bacteriological testing.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

18

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Table 5. Number of units in which the protocol recommended that women with the listed risk factorshould have a bacteriological swab for group B streptococcus (GBS)

Risk factor (current pregnancy unless otherwise stated) Units (n)

GBS colonisation in previous pregnancy 13

Preterm labour1 14

Preterm rupture of membranes a 6

Prelabour rupture of membranes at term 9

Prolonged rupture of membranes b 8

Intrapartum fever 7

Suspected chorioamnionitis 7

Bacteriuria 6

Vaginal discharge 6

Situations listed by two units:previous stillbirth, antenatal admission, maternal GBS infection prior to pregnancy

Situations listed by one unit:home birth, multiparous woman, woman with diabetes, vaginal bleeding, suspected infection, sexuallytransmitted disease, history of preterm labour, history of preterm prelabour rupture of membranes,previous adverse event to mother or babya Specified as either less than 37 weeks or less than 36 weeksb If specified, durations were over 18 hours, over 24 hours, over 48 hours, over 72 hours

Table 6. Number of units whose protocol recommended intrapartum antibiotic prophylaxis for women withspecific risk factors

Risk factor RCOG recommendation Unitsn %

Previous baby affected by early-onset disease Offer 159 93

Bacteriuria in current pregnancy Offer 146 85

Incidental GBS finding in current pregnancy Offer 144 84

Fever during labour > 38˚C Discuss 104 61

Preterm labour less than 37 weeks Discuss 84 49

Prolonged ROM 18 hours or over a Discuss 89 52

Suspected chorioamnionitis Offer 66 39

Maternal GBS colonisation in previous pregnancy Not supported 29 17

Planned caesarean section (in absence of labour or ROM) Not supported 0 0a Includes those protocols that define prolonged ROM as 24 hours or more; GBS = group B streptococcus; ROM = rupture ofmembranes

The only exception was an incidental GBS finding. Intrapartum prophylaxiswas recommended for this factor by 90% of the ‘risk-based IAP’ protocolsand 70% of the ‘risk-based bacteriological testing’ protocols.

The consistency of the protocol recommendations with the RCOG guidelinevaried according to the strength of the evidence. On their own, fever duringlabour, preterm labour and prolonged ROM were regarded as an indicationfor IAP by less than two-thirds of the units. However, it was these factorsthat typically featured when a combination of factors were given as anindication for IAP (Table 7). If the frequencies given in Table 6 are alteredto include instances of when these factors were also used in combination,then the proportions of units with a protocol recommending IAP for thesefactors were 78%, 64%, and 70%, respectively.

Nonetheless, many protocols did not conform to the RCOG recommendations:

� Only 66 units had protocols that mentioned suspectedchorioamnionitis as an individual risk factor, although it wasmentioned in combination with other risks in protocols at nineadditional units.

� Maternal GBS colonisation in a previous pregnancy was mentioned asan individual risk factor by protocols at 29 units, although theevidence does not support IAP in this situation. It was also used inrisk-factor combinations at a further seven units.

� Prolonged ROM was specified as exceeding 24 hours by protocolsfrom 34 units instead of the more usual limit of 18 hours.

� While the RCOG guideline states the argument for prophylaxisbecomes stronger with two or more risk factors, the combinationsused at many units do not conform to the options presented in theguideline.

19

Table 7. Risk factor combinations for which intrapartum antibiotic prophylaxis was recommended thatwere found in the protocols used by UK maternity units

Combinations Unitsn %

Two or more risks including fever, preterm labour and prolonged rupture of membranes (ROM) a 15 9

Preterm labour and another risk factor, such as fever,incidental group B streptococcus (GBS) finding 11 6

Prolonged ROM and factors indicative of suspected chorioamnionitis b 9 5

Incidental GBS finding and prelabour ROM 6 4

Incidental finding and a risk including fever, preterm labour and prolonged ROM a 6 4

Previous maternal GBS and another risk factor such as fever, preterm labour and prolonged ROM 6 4a While these three factors always appeared, other factors were sometimes mentioned, including bacteriuria, preterm ROM not inlabour, leucocytosis, preterm prelabour ROM, stillbirth; b the factors indicative of suspected chorioamnionitis were fever, uterinetenderness, maternal tachycardia, fetal tachycardia



Other situations in which IAP was recommended included:

� preterm labour (variously defined as less than 34, 35 or 36 weeks)

� preterm ROM, preterm prelabour ROM, preterm prolonged ROM

� prolonged or prelabour ROM (either undefined or with a limit of 36,48 or 72 hours).

Protocols were inconsistent in describing when IAP should be offered forpreterm ROM and it was not always clear whether the recommendationapplied to women in labour or to women not yet in labour.

Finally, few protocols contained advice on considering maternal preferencesfor antibiotic prophylaxis and maternal requests for a swab. Of thoseprotocols recommending risk-based IAP, 50 protocols (37%) included astatement about the need to consider a woman’s preference for antibiotics,while only one mentioned offering IAP on a maternal request. Among the 35protocols recommending risk-based bacteriological testing, 11 mentionedmaternal preferences for antibiotics and 6 considered maternal requests.

4.3 Recommended practice for bacteriological testing

Among the protocols that recommended risk-based bacteriological testing,the majority described when swabs should be taken and from what sites. Ofthe 29 units whose protocol contained guidance on the timing of testing, thiswas after 35 weeks of gestation in 19 cases. However, the suggested rangesvaried and only six units used the recommended 35–37 week period.

Recommended swab sites were contained in the protocols of 31 units.Again, there was no consistency among protocols:

� six units recommended a low vaginal swab

� six units recommended a high vaginal swab

� four units recommended a whole vaginal swab (low and high)

� four units recommended swabbing the low vagina and rectum, whilefour recommended swabbing the whole vagina and rectum.

Other recommendations included either a low or high vaginal swab, or aperineal swab. No protocol made recommendations on the use of an enrichedculture medium.

4.4 Protocol antibiotic regimens

The primary antibiotic regimens specified in the maternity unit protocolstended to conform to the recommended approaches in the RCOG guideline(Table 8). The main variants reflected different doses of penicillin G (benzyl-penicillin) rather than an alternative antibiotic. An even higher proportion ofunits used the recommended regimen for women allergic to penicillin.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

20

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Only 75 units had protocols that included recommendations for when to usebroad-spectrum antibiotics (Table 9). The one clinical situation specified inthe RCOG guideline, suspected chorioamnionitis, was mentioned by proto-cols from 72 units (this includes protocols that mention suspectedchorioamnionitis and/or fever).

4.5 Conclusion

While most protocols contained individual recommendations that wereconsistent with those in the RCOG guideline, only a minority of units hadprotocols that were entirely consistent with the RCOG guideline. Among theresponding units, 78% had adopted risk-based IAP but there was variationin the sets of risk factors used for identifying which women should beoffered IAP. Some protocols omitted a number of the established riskfactors, although these tended to be those associated with a lower risk. Someprotocols also recommended IAP for women with additional risk factors orcombinations of factors, recommendations for which there is little

21

Table 8. Antibiotic regimens specified in the maternity unit protocols

Unitsn %

Primary regimens specified:Penicillin G (benzylpenicillin), 3 g loading dose followed by 1.5 g every 4 hours (recommended by RCOG) 126 74

Alternative specific benzylpenicillin regimens reported:3 g loading dose followed by 1.2 g every 4 hours 10 63 g loading dose followed by 1.8 g every 4 hours 6 42.4 g loading dose followed by 1.2 g every 4 hours 6 4

Penicillin G, unusual, imprecise or incompletely specified 22 13

Other drug regimen 1 1

Protocol regimens for women allergic to penicillin:Clindamycin, 900mg 8-hourly (recommended by RCOG) 142 83

Alternative regimens specified for women allergic to penicillin:clindamycin, 600 mg 6-hourly 6 4other antibiotic (e.g. erythromycin, vancomycin) 11 6

None specified 12 7

Table 9. Situations where broad-spectrum antibiotics were recommended by protocols

Clinical situation Unitsn %

Suspected chorioamnionitis and/or fever 72 42

Preterm prolonged rupture of membranes 17 10

Preterm prelabour rupture of membranes 6 4

Preterm labour 3 2

Prolonged rupture of membranes at term 4 2

No recommendation made 96 56



supporting evidence. In addition, 22% of units recommended risk-basedbacteriological testing. At these units, there was variation in thecombinations of risk factors used to identify those women who should havea bacteriological screen. This also led to variation in the sets of risk factorsused for the risk-based IAP component of these strategies.

There was less variation in the antibiotic regimens proposed by the protocols.In all but one protocol, the primary antibiotic regimen was penicillin G(benzylpenicillin), with clindamycin being recommended by most units forwomen with a penicillin allergy. The alternatives for clindamycin included inthe protocols tended to conform to regimens recommended by otherguidelines (for example CDC).12 Of more concern is the variation in thedosages proposed for penicillin G because these did not conform to anypublished guideline.

The variation in the protocol recommendations would not be of concern ifit had been clearly linked to local circumstances or published evidence. Thisdid not appear to be the case. There are reasons to believe that some of thevariation is due to weaknesses in the procedures used at some NHS trusts todevelop their protocols. First, 4% of unit protocols had not been updatedsince the publication of the RCOG guideline. Second, 12% of unit protocolscontained no reference to any clinical evidence. Finally, in those protocolsthat included references, 26% of unit protocols did not cite the RCOGguideline and 51% cited guidelines that had been superseded. These weak-nesses need to be addressed.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

22

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

5. Audit results

5.1 Audit participants and response rates

The audit contacted clinical directors representing 227 maternity units withobstetric facilities. A response was received from clinical directors of 186(82%) maternity units; all agreed to participate.

Of the 41 units from which no response was received, 27 were contacteddirectly through the lead obstetrician. The 19 remaining units were notcontacted either because the lead obstetrician was also the clinical directorwho had previously not responded or because no contact details for the leadobstetrician were available.

Questionnaires were sent to the clinicians at 213 of the 227 units (94%).The response rates achieved for each discipline were:

� midwives: 71% (161/227)

� obstetricians: 78% (177/227)

� neonatologists: 71% (161/227).

A response from at least one clinician was received from 202 units (89%).A response from either a midwife or obstetrician, thereby giving informationon maternal care, was received from 198 units (87%). A response fromeither a midwife or neonatologist, thereby giving information on neonatalcare, was received from 186 units (82%). Responses from one midwife, oneobstetrician and two neonatologists could not be attributed to a unit andwere therefore excluded from the analysis.

Participation by NHS hospitals by region is summarised in Table 10. Aresponse from one obstetrician and one neonatologist in an independenthospital was received.

23

Table 10. NHS hospitals participating in the audit by region

Specialty England Scotland Wales N. Ireland

n % n % n % n %

Midwives 128 70 11 58 10 71 11 100

Obstetricians 144 80 16 84 7 50 8 73

Neonatologists 132 73 11 58 7 50 8 73

Total 181 19 14 11

5.2 Intrapartum management

Screening policy

RCOG recommendation

Routine screening (either bacteriological or risk based) for antenatalcarriage is not recommended.

Information on the screening strategy practiced by was provided by 159midwives and 172 obstetricians from 197 units. Few clinicians reportedfollowing a universal bacteriological screening strategy but about 20% ofrespondents reported that they routinely swabbed for GBS in selectedwomen (Table 11). All other respondents said that they did not routinelytake bacteriological swabs.

At least one respondent from 63 units (32%) reported that they routinelyswabbed for GBS in selected women. However, there was disagreementbetween midwives and obstetricians at the 50 units from which bothresponses were received (Table 11). A universal bacteriological screeningapproach was only reported by at least one respondent at five units (3%)but, among the four units for which both responses were available, therewas again disagreement between midwives and obstetricians. Where adisagreement occurred, the other respondent always stated that they did notroutinely swab for GBS. All respondents from the other 129 units (65%)

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

24

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Table 11. Summary of bacteriological screening practice reported

A

Midwife Obstetrician(n = 159) (n = 172)

n % n %

Respondent routinely took swabs for GBS in all women (universal bacteriological screening) 3 2 3 2

Respondent routinely took swabs for GBS in selectedwomen (risk-based bacteriological testing) 29 18 42 24

Respondent did not routinely take swabs for GBS 127 80 127 74

B

Routinely took swabs for GBS Units

In all women (universal bacteriological screening):by midwife but not obstetrician 1by obstetrician but not midwife 2by both midwife and obstetrician 1

In selected women (risk-based bacteriological testing):by midwife but not obstetrician 17by obstetrician but not midwife 25by both midwife and obstetrician 8

GBS = group B streptococcus

reported that they did not practice either universal screening or risk-basedbacteriological testing.

As noted earlier, none of the protocols were judged to recommend universalbacteriological screening. However, for the two units with protocols thatappeared to advocate risk-based bacteriological testing (using a CDC-stylestrategy, see section 2.2), three of the four respondents reported screening allwomen. Protocols were provided by two of the other three units at which atleast one respondent reported universal bacteriological screening. Bothprotocols recommended a risk-based IAP strategy.

Among the 37 protocols judged to recommend risk-based bacteriologicaltesting, the majority of clinicians reported that they did not routinely swabfor GBS in any women. Only 14 midwives and 12 obstetricians reportedroutinely taking swabs in selected women. The reasons for this are not clear.However, it demonstrates that there are different interpretations of whatconstitutes universal screening and risk-based bacteriological testing amonghealth professionals.

Reported practice for bacteriological screening

For units that took bacteriological swabs from either all or selected women,the audit questionnaire included questions about how this was implemented.The RCOG guideline reported the risk of early-onset GBS disease if practicefollowed the process recommended in the CDC guideline (namely, thatswabs be taken from the rectum and lower vagina at 35–37 weeks andinoculated into an enriched medium). However, the RCOG guideline notedthat the risk of disease is probably higher if a positive swab was obtainedfrom the upper vagina or was cultured using a non-enriched medium. If apositive swab is obtained at an early gestational age, the risk of a babygetting early-onset GBS disease is probably lower.

Among the 68 units reporting performing universal screening or risk-basedbacteriological testing:

� One midwife and four obstetricians reported taking swabs between 35and 37 gestational weeks. The variation in the periods reported byobstetricians and midwives is shown in Figure 1.

� There was no consistency in the site of swabbing. The most frequentlyreported practice in both professions was a high-vaginal swab (11 of32 midwives and 14 of 45 obstetricians, respectively). Just threemidwives and six obstetricians reported taking swabs from the lowervagina and rectum.

� Most staff did not know if the recommended selective-enrichedmedium was used in the laboratory to process the GBS cultures.

� Results were typically available in 48 hours.

Among the five hospitals reporting universal bacteriological screening, nomidwife or obstetrician reported taking swabs between 35 and 37 gestationalweeks, as recommended by the CDC guideline.12

25

Audit resultsAudit results

Intrapartum management of women with specific risk factors

RCOG recommendation

Clinical situation Recommendation regardingintrapartum antibioticprophylaxis

Previous baby affected by GBS disease Should be offered

GBS bacteriuria in current pregnancy Should be offered

Incidental finding in current pregnancy Should be considered

Fever over 38˚C Discuss

Preterm labour less than 37 weeks Discuss

Prolonged ROM (18 hours or more) Discuss

Maternal GBS carriage in previous No good evidence to pregnancy support

Planned caesarean section at term Not required(in absence of labour and with intact membranes)

Prelabour preterm ROM (less than Unnecessary unless in 37 weeks) established labour

Suspected chorioamnionitis Offer IAP using a broad-spectrum antibiotic.

The questionnaire asked midwives and obstetricians about the clinicalindications for intrapartum antibiotic prophylaxis among pregnant womenwhose GBS status was not known. For women with different risk factors,clinicians were asked whether or not they took a swab for GBS. Those that

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

26

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

40

35

30

25

20

15

10

5

0

45

ob obob ob ob ob obob ob ob ob obob ob ob obob ob ob mw mw mw mw mw mw mw mw mw mw mwRespondents (ob = obstetrician, mw = midwife)

Wee

ks o

fg

esta

tio

n

Figure 1. Range over which respondents reported taking swab-based cultures

would were then asked if IAP would be offered if the culture from the swabwas GBS positive or negative. Those that would not take a swab were askedif they would offer IAP on the basis of the risk factor alone. The practicesreported by midwives and obstetricians are described in Tables 12 and 13,respectively. The right-hand columns give the number of responses to thesequestions.

The proportion of clinicians who would take a swab was generally higherthan expected, given that clinicians at only 32% of units reported routinelyswabbing for GBS from either all or selected women to determine whichwomen are offered IAP. The reason for this is not clear. Clinicians may haveinterpreted the word ‘routinely’ differently. It is also possible that the swabwould be used for other diagnostic purposes, either to inform the manage-ment of the neonate or to check for other infections. The high proportion ofrespondents who would offer IAP even when a swab for GBS was negativewould seem to support this.

The proportion of women offered IAP in the clinical situations are generallyin line with the RCOG recommendations, with the highest proportionsbeing observed for the strongest recommendations. The proportions ofwomen offered IAP are also generally higher when GBS colonisation hasbeen confirmed than when the offer is based solely on the presence of a riskfactor. However, practice differs from the recommendation for women whowere colonised with GBS in a previous pregnancy. For women with this riskfactor, 70% of midwives and 75% of obstetricians reported that they wouldoffer IAP on that basis alone. The proportions exceeded 90% for womenwith a GBS positive culture, and over 30% of women would be offered IAPeven if the culture was negative.

27

Table 12. Management of pregnant women whose group B streptococcus (GBS) status was unknown bymidwives in specific clinical situations

Risk factor Percentage Percentage who Valid responses who swab would offer IAP (n)a

+ swab – swab No Swab IAP IAPswabb taken? swab risk

Previous baby affected by GBS disease 58 99 52 100 137 79 53

GBS bacteriuria in current pregnancy 73 95 13 80 137 99 30

Incidental finding in current pregnancy c – – – 94 – – 144

Maternal GBS carriage in previous pregnancy 48 95 39 70 131 59 54

Planned caesarean section at term 9 75 8 10 130 12 62

Preterm prelabour ROM (< 37 weeks) 94 98 15 – 144 135 6

Vaginal discharge 88 95 0 – 144 124 9

Suspected chorioamnionitis 92 97 22 – 146 133 6a Not all clinicians gave responses to both parts of the question; b Proportions not derived for samples of less than ten; c Questionon swab not applicable; IAP = intrapartum antibiotic prophylaxis; ROM = rupture of membranes


At units for which the audit had both responses, the agreement amongmidwives and obstetricians on how to manage women varied according tothe various risk factors for which women would be swabbed (see Appendix4). However, there was good agreement among clinicians in terms of whichwomen would be offered IAP, which was expected given the high percent-ages in Tables 12 and 13.

Midwives and obstetricians were also asked about their management ofwomen whose GBS status was unknown in intrapartum situations.Compared to the responses about the risk factors in Tables 12 and 13, theproportion of clinicians who would offer IAP to women with an intrapartumrisk factor was much lower (Table 14). This is consistent with the patternobserved among the protocols for these risk factors. Agreement amongmidwives and obstetricians at the same unit on how to manage thesesituations was also only moderate (see Appendix 4).

Antibiotic prophylaxis regimen

RCOG recommendations

Penicillin G (benzylpenicillin) should be administered as soon aspossible after the onset of labour. The recommended dose is 3 g intra-venously followed by 1.5 g every 4 hours during labour.

Intravenous clindamycin 900 mg 8-hourly should be administered tothose women allergic to penicillin.Pr

even

tion

of E

arly

-ons

et N

eona

tal G

roup

B S

trep

toco

ccal

Dis

ease

in U

K O

bste

tric

Uni

ts

28

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Table 13. Management of pregnant women whose group B streptococcus (GBS) status was unknown byobstetricians in specific clinical situations

Risk factor Percentage Percentage who Valid responses who swab would offer IAP (n)a

+ swab – swab No Swab IAP IAPswabb taken? swab risk

Previous baby affected by GBS disease 38 98 81 100 162 59 98

GBS bacteriuria in current pregnancy 44 97 39 96 163 72 85

Incidental finding in current pregnancy c 99 155

Maternal GBS carriage in previous pregnancy 39 95 32 75 162 63 92

Planned caesarean section at term 8 38 8 3 163 13 105

Preterm prelabour ROM (< 37 weeks) 98 98 32 – 170 164 4

Vaginal discharge 93 96 2 – 166 152 5

Suspected chorioamnionitis 98 94 44 – 168 161 3a Not all clinicians gave responses to both parts of the question; b Proportions not derived for samples of less than ten; c Questionon swab not applicable; IAP = intrapartum antibiotic prophylaxis; ROM = rupture of membranes

Penicillin G was used by the bulk of midwives and obstetricians as theprimary antibiotic prophylaxis (Table 15). Indeed, at the majority of the 198responding units, the obstetrician or midwife (or both) reported usingpenicillin G in all but five institutions. Moreover, at 139 units (70%), at leastone respondent reported offering intrapartum prophylaxis in accordancewith the RCOG recommendation.

Several alternative penicillin regimens were reported, however. At 13 units,both the midwife and obstetrician reported using the same alternative. Thistended to match the regimen recommended in the protocol. Among the unitsfor which both clinician and protocol data were available, 50% of midwivesand 62% of obstetricians reported following the alternative regimen recom-mended in the protocols (respectively, 8 of 16 midwives and 13 of 21obstetricians).

The antibiotic regimens reportedly used by midwives and obstetricians forwomen with a penicillin allergy were again generally consistent with theRCOG recommendation (Table 16). Clindamycin 900 mg, delivered every 8hours, was reported as the substitute for penicillin by at least one respondentat 157 units (79%) among the 195 for which data were available. Obstet-rician and midwife responses were received from 126 of these units and thestandard regimen was given by both clinicians at 62 units (49%).

29

Table 14. Proportion of clinicians who would offer antibiotic prophylaxis to women in specificintrapartum situations when their group B streptococcus status is unknown

Midwife Obstetricians

Resp. (n) IAP (%) Resp. (n) IAP (%)

Preterm labour 35–37 weeks of gestation 155 31 171 43

Preterm labour < 35 weeks of gestation 154 35 172 50

Intrapartum fever at term > 38˚C 158 82 173 87

Prolonged ROM ≥ 18 hours 158 53 171 61

Preterm prelabour ROM, followed by labour at term 152 49 169 50

IAP = intrapartum antibiotic prophylaxis; Resp. = respondents; ROM = rupture of membranes

Table 15. Primary antibiotic prophylaxis regimens used by midwives and obstetricians

Antibiotic regimens reported Midwives Obstetricians

n % n %

Penicillin G (benzylpenicillin), 3 g IV followed by 1.5 g every 4 hours 107 68 105 61

Penicillin G, 3 g IV followed by 1.2 g every 4 hours 11 7 9 5

Penicillin G, 3 g IV followed by 1.8 g every 4 hours 5 3 6 3

Penicillin G, 2.4 g IV followed by 1.2 g every 4 hours 3 2 6 3

Other penicillin G regimen 27 17 45 26

Other drug regimen 5 3 1 1

IV = intravenous


Several alternative doses were reported for clindamycin, most commonly 600mg at 6-hourly intervals. Reported alternatives to clindamycin includederythromycin and vancomycin. At seven units both the midwife and obstet-rician reported using the same alternative. As before, this tended to match theregimen recommended in the protocol. Among the units for which data wereavailable, 100% of midwives and 57% of obstetricians reported followingthe alternative regimen recommended in the protocols (respectively, eight ofeight midwives and four of seven obstetricians).

Preferences of pregnant women

Midwives and obstetricians were asked to report the percentage of womenwho declined an offer of IAP in the past year as well as the percentage ofwomen who requested a GBS swab. Units did not appear to collect this dataroutinely and where they were provided responses were generally estimates.

Both professions thought that few women declined an offer of IAP (Table17), with most estimates falling between 0% and 1%. Slightly higher esti-mates were given for the percentage of women who request a GBS swab butless than 25% of both midwives and obstetricians estimated the proportionwas above 5% of women.

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

30

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Table 16. Antibiotic regimens used when a pregnant woman is allergic to penicillin

Antibiotic regimens reported Midwives Obstetricians

n % n %

Clindamycin 900 mg, every 8 hours 111 73 108 64

Alternative clindamycin regimen specified 12 8 12 7

Unspecified clindamycin regimen 15 10 26 15

Other drug regimen 14 9 23 14

Table 17. Estimates of women’s preferences for intrapartum antibiotic prophylaxis (IAP) andbacteriological testing

Proportion of women Midwives Obstetricians

n % n %

Declining an offer of IAP:0–1% 104 73 91 55> 1–5% 29 20 53 32> 5% 10 7 22 13

Total 143 166

Requesting a swab for group B streptococcus:0–1% 63 49 53 32> 1–5% 36 28 79 47> 5% 32 24 36 21

Total 131 168

These figures seem fairly speculative. Using the three categories shown inTable 17, a kappa statistic was derived to test the level of agreement betweenthe figures given by midwives and obstetricians at the same unit. For bothquestions, the level of agreement was poor. For the question about theproportion of women who decline IAP, the kappa statistic was 0.07(P = 0.15). For the question about the proportion of women who request aswab, the statistic was 0.22 (P < 0.01).

5.3 Comparison of maternity practice with previous studies

Kenyon et al. surveyed the clinical directors of maternity units with obstetriccare about their antenatal screening practice for GBS and intrapartumantibiotic regimens in 1999 and 2001.2 In comparison with these earliersnapshots of practice, the results of the audit show that in areas wherepractice varied between units, the amount of variation has reduced since thepublication of the RCOG guideline, with a higher proportion of unitsreporting practice that is consistent with the RCOG recommendations.

There has been minimal change in the number of units offering universalbacteriological screening. This was offered by six units (3%) in 1999 andfour units (4%) in 2001, which is comparable to the situation found in thisaudit. There would appear to be a slight improvement in the proportion ofunits offering IAP to women with specific risk factors, although thecomparison is limited because the earlier survey only asked all units whetherintrapartum prophylaxis was offered in two circumstances (Table 18). In2006, 99% of obstetricians reported offering IAP in these situations.

Finally, reported practice has noticeably improved in relation to theproportion of units using recommended IAP regimens. In 1999, only 9% ofunits indicated that they were using one of the recommended IAP regimens.In 2001, it was still only 20% of units. In 2006, over 60% of obstetriciansreported using the recommended penicillin regimen.

31

Table 18. Response of units when asked about the situations in which they offer intrapartum antibioticprophylaxis (IAP)

Risk factor 1999 2001 2006

Units Offer Units Offer Units Offer(n) IAP (%) (n) IAP (%) (n) IAP (%)

Previous baby affected by group B streptococcal disease 207 85 203 95 162 99

Incidental finding in current pregnancy 207 86 203 95 155 99


5.4 Neonatal management

RCOG recommendations

Newborn infants with clinical signs compatible with infection shouldbe treated promptly with broad-spectrum antibiotics which providecover against early-onset GBS disease as well as other commonpathogens.

For an infant whose mother had a previous infant with GBS disease orhas other risk factors for early-onset GBS (especially two or more),either clinical evaluation after birth and observation for at least 12hours are necessary or blood cultures should be obtained and theinfant treated with penicillin until the culture results are available.

Postnatal antibiotic prophylaxis is not recommended for low-risk terminfants.

There is little research evidence to support decisions about the treatment ofnewborn babies at risk of GBS disease and the RCOG guideline acknowledgedthat not enough is currently known to make strong recommendations. Inparticular, for babies who are apparently well but who have a risk factor, twotreatment strategies were recognised as being commonly used:

� clinical evaluation after birth and observation for at least 12 hours, or

� obtaining blood cultures and treating the infant with penicillin untilthe culture results are available.

The questionnaire to neonatologists and midwives contained a series ofquestions about the management of babies who may be at risk of early-onsetGBS disease. The questions asked about their management practices for anapparently well baby born (a) after 37 weeks of gestation, (b) between 35 and37 weeks of gestation and (c) before 35 weeks of gestation, in the followingsituations:

� the mother received IAP for a positive GBS swab during the currentpregnancy

� the mother received IAP because a previous baby had early-onset GBSdisease

� the mother was GBS positive but failed to receive IAP

� the mother was GBS positive and received suboptimal IAP because shegave birth before receiving 2 hours of antibiotic treatment.

The responses by neonatologists and midwives to these four clinical situa-tions are shown in Figures 2–5. Each graph shows the response for babies ofdifferent gestational age from both professions. The number of responses toeach question is shown in brackets along the horizontal axis.Pr

even

tion

of E

arly

-ons

et N

eona

tal G

roup

B S

trep

toco

ccal

Dis

ease

in U

K O

bste

tric

Uni

ts

32

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

The results show that the two strategies included in the guideline are themost widely practised when active management is judged necessary. Theresults also show that perceptions of risk were associated with the likelyeffectiveness of prophylaxis given to a baby’s mother, the gestational age ofthe baby and the types of maternal risk factor to which the baby is exposed.

33

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

>37

wee

ks(n

=15

3)

35–3

7 w

eeks

(n=

150)

<35

wee

ks(n

=15

1)

>37

wee

ks(n

=14

5)

35–3

7 w

eeks

(n=

143)

<35

wee

ks(n

=14

4)

Gestation in weeks at birth

Take no action

Culture and wait for results before treatmentDon’t know

Assess clinically and observe for 12 hours

Culture and treat until results availableOther

Neonatologists Midwives

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

>37

wee

ks(n

=15

6)

35–3

7 w

eeks

(n=

155)

<35

wee

ks(n

=15

2)

>37

wee

ks(n

=14

4)

35–3

7 w

eeks

(n=

143)

<35

wee

ks(n

=14

2)


Take no action





Figure 2. Reported practice for a baby whose mother received intrapartum antibioticprophylaxis for a positive group B streptococcus swab during the current pregnancy

Figure 3. Reported practice for a baby whose mother received intrapartum antibioticprophylaxis because a previous baby had early-onset group B streptococcal disease


In particular, increased prematurity was associated with a shift fromobservation towards antibiotic treatment until the possibility of infectioncan be ruled out. In addition, the perception of risk posed by no or subopt-imal maternal IAP was judged to be the same. Moreover, more clinicians

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

34

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

>37

wee

ks(n

=15

6)

35–3

7 w

eeks

(n=

155)

<35

wee

ks(n

=15

2)

>37

wee

ks(n

=14

4)

35–3

7 w

eeks

(n=

143)

<35

wee

ks(n

=14

2)


Take no action





100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%

>37

wee

ks(n

=15

6)

35–3

7 w

eeks

(n=

155)

<35

wee

ks(n

=15

2)

>37

wee

ks(n

=14

4)

35–3

7 w

eeks

(n=

143)

<35

wee

ks(n

=14

2)


Take no action





Figure 4. Reported practice for a baby whose mother tested positive group B streptococcusbut did not receive intrapartum antibiotic prophylaxis

Figure 5. Reported practice for a baby whose mother tested positive group B streptococcusand received suboptimal intrapartum antibiotic prophylaxis because she received less than 2

hours of antibiotic treatment

judged this risk to be greater than that posed to a baby whose mother hadappropriate IAP and who had a previous baby with GBS disease.

The reported practice in managing the risk to newborn babies did not differgreatly between neonatologists and midwives overall. There were slightlymore neonatologists who would take no action than midwives in the ‘lowrisk’ clinical situation. However, as the risk increased, slightly more neo-natologists would culture and treat until results became available. A higherproportion of midwives would culture and wait for the results beforetreatment was begun.

In conclusion, current practice for managing neonates at risk of GBS diseasecontinues to follow the established patterns of care at the time the RCOGguideline was published. There is little agreement between clinicians in termsof whether they should assess and observe ‘at risk’ babies or take a moreinterventionist approach; that is, take a culture and start antibiotic treat-ment while awaiting the result. This reflects the lack of available evidence toguide practice in this area.

5.5 Conclusion

The responses to the audit questionnaire on the management of pregnantwomen show that the practice of midwives and obstetricians is generallyconsistent with the RCOG guideline recommendations:

� 2% of midwives and 2% of obstetricians performed universalbacteriological screening

� over 90% of midwives and obstetricians offer IAP to women that hadany of the ‘high risk’ factors:�� a previous baby affected by GBS disease�� GBS bacteriuria in the current pregnancy�� an incidental finding of GBS colonisation in the current

pregnancy.

� over 96% of midwives and obstetricians used benzylpenicillin(penicillin G) as the primary IAP regimen, although only 68% ofmidwives and 61% of obstetricians used the dose recommended in theRCOG guideline.

The limited variation in practice in these areas compared to the RCOGrecommendations might have been expected given the content of thereviewed protocols. The higher level of variation in practice reported inother areas is also consistent with the findings of the protocol review. Asbefore, the proportions of clinicians who report offering IAP to women forpreterm labour (less than 37 weeks) and prolonged ROM (18 hours orlonger) are noticeably lower than the proportions offering IAP for the ‘high-risk’ factors. In addition, around 70% of midwives and obstetricians wouldoffer IAP to women who were colonised in a previous pregnancy, althoughthe RCOG guideline did not recommend this practice.

35


6. Patient vignettes

Five case vignettes were developed to investigate what treatment options cliniciansconsidered appropriate for women with a combination of maternal risk factors. Thevignettes covered decisions related to the care of the mother and newborn andclinicians were asked whether they considered particular actions as appropriatemanagement by selecting one of the following categories: ‘strongly disagree’,‘disagree’, ‘agree’, ‘strongly agree’ or ‘don’t know’. Questions related to the care ofmother focused on the administration of antibiotics and were linked to the RCOGrecommendations about IAP in labour. Others related to the recommendation relatedto antenatal treatment.

The options about the management of the newborn were the same in each vignette,with respondents asked to rate the appropriateness of (1) clinically observing thenewborn infant for signs of early-onset GBS disease and (2) giving the newborn infantintravenous antibiotics pending the results of blood cultures. These options wereselected because the RCOG guideline recognised that the risks and benefits oftreatment are uncertain for well infants with a risk factor and clinicians can prefereither approach.

The vignettes were included in the questionnaire sent to each clinical profession. Theresponse rate was high overall and the proportion of ‘don’t knows’ or missing valueswas generally below 5%. Detailed results are provided in Appendix 5.

VIGNETTE 1

A 24-year-old primigravida presents at 37 weeks of gestation with a history ofvaginal discharge. A high vaginal swab confirms the presence of GBS.

In this situation, the RCOG guideline recommends that IAP should be considered ifGBS is detected incidentally. Over 90% of respondents in each profession agreed orstrongly agreed that the prescription of IAP in labour was appropriate management.This is consistent with the 94% of midwives and 99% of obstetricians who reportedoffering IAP to women with GBS colonisation diagnosed incidentally. The proportionof midwives and obstetricians who regarded prescribing oral antibiotics immediatelyas appropriate was 12% and 29%, respectively.

In each profession, over 87% of respondents considered clinically observing thenewborn to be appropriate management. Among neonatologists, 24% of respondentsalso considered that giving the newborn antibiotics until the results of blood culturesbecame available was appropriate management. This is significantly higher than the1% of neonatologists who reported that they would take this course of action in theearlier part of the questionnaire (see Figure 2). For midwives and obstetricians, theproportions were 12% and 18%, respectively.

36

VIGNETTE 2

A 36-year-old parous woman attends for booking at 11 weeks of gestation. Herprevious baby was diagnosed with early-onset GBS disease.

The RCOG guideline recommends that IAP should be offered to women if a previousbaby had neonatal GBS disease and 96% of both midwives and obstetricians and88% of neonatologists agreed that prescribing IAP in labour was appropriatemanagement. Less than 6% of respondents in each profession agreed with prescribingoral antibiotics at the booking visit. There was greater variation among units inwhether or not a vaginal swab was considered appropriate management. A swabwould be taken by 32% of obstetricians compared with 47% of midwives. This isconsistent with the responses to the questionnaire, where 38% of obstetricians and58% of midwives reported offering a GBS swab to pregnant women.

As in vignette 1, most respondents considered clinically observing the newborn asappropriate management, the proportions being 92%, 91% and 76% for midwives,obstetricians and neonatologists, respectively. Among neonatologists, 44% ofrespondents would also consider giving the newborn antibiotics until the results ofblood cultures became available. Again, this is higher than the 22% of neonatologistswho recommended this action earlier (see Figure 3). For midwives and obstetricians,the proportions were 28% and 41%, respectively.

VIGNETTE 3

A 27-year-old primigravida (with no known allergies) is admitted in spontaneouslabour at 38 weeks of gestation. She makes slow progress in labour, despiteartificial rupture of the membranes, and has a recorded temperature of 38.5°C.

The raised temperature is consistent with suspected chorioamnionitis, for which theRCOG guideline recommends a broad-spectrum antibiotic therapy including GBScover. Practice among obstetricians was generally consistent with this recommend-ation with 84% considering this option appropriate management. However,approximately 25% of obstetricians also considered offering intravenous penicillin.Among midwives, 74% would prescribe broad-spectrum antibiotics, while 30%would also offer intravenous penicillin. On its own, using penicillin would beinsufficient but would be appropriate if offered in addition to a broad-spectrumantibiotic.

As before, the favoured approach to managing newborns was to clinically observethem for signs of GBS disease, the proportions being 91%, 93% and 82% formidwives, obstetricians and neonatologists, respectively. Among neonatologists, 47%of respondents agreed that giving antibiotics to a newborn until receiving the resultsof blood cultures was also appropriate management. For midwives and obstetricians,the proportions were 38% and 27%, respectively. No question in the questionnairedirectly corresponded to this vignette for comparison.

37

Patient vignettesPatient vignettes

VIGNETTE 4

A 19-year-old primigravida (with no known allergies) presents with prelabourrupture of membranes at 31 weeks of gestation. She receives oral erythromycin,in keeping with the hospital protocol, and a high vaginal swab confirms thepresence of GBS.

The RCOG guideline does not recommend antibiotics for women with pretermrupture of membranes prior to labour. Over 90% of respondents in each professionagreed or strongly agreed that the prescription of IAP in labour is appropriate man-agement. This is consistent with the 98% of midwives and obstetricians whorecommended IAP in their responses to the questionnaire. Around 25% of midwivesand obstetricians regarded prescribing oral antibiotics immediately as appropriate.

Among obstetricians, 91% considered clinically observing the newborn as appro-priate management. Sixty-five percent of neonatologists agreed with this approachbut there was a tendency to prefer giving the newborn antibiotics until the results ofblood cultures became available. Eighty-one percent of neonatologists agreed thatthis was appropriate management, compared with just 51% of obstetricians. Theproportion of midwives agreeing with each approach was similar to the obstetricians.

VIGNETTE 5

A 25-year-old parous patient at 41 weeks of gestation experiences a 30-hourspontaneous rupture of membranes (SROM).

In this situation, the RCOG guideline recommends that decisions about the use ofIAP incorporate information on the numbers needed to treat included in theguideline. Between 36% and 40% of respondents in each profession agreed that theprescription of IAP at the onset of labour was appropriate management. This differsfrom responses to the earlier part of the questionnaire (see Table 14) when 53%midwives and 61% obstetricians reported that they would offer IAP under thesecircumstances. A minority of midwives (14%) and obstetricians (21%) regardedprescribing intravenous penicillin 18 hours after SROM as appropriate.

As before, the favoured approach to managing newborns was to clinically observethem for signs of GBS disease, the proportions being 84%, 86% and 91% formidwives, obstetricians and neonatologists, respectively. Less than 17% ofrespondents in each profession agreed that giving antibiotics to a newborn untilreceiving the results of blood cultures was appropriate management.

Conclusion

There was broad support for the RCOG recommendation on prescribing IAP duringlabour among midwives, obstetricians and neonatologists. The only exception arosewith respect to prolonged rupture of membranes, where opinions were divided onwhether or not IAP was appropriate. Variation in the management of this conditionwas also evident in the responses to the questionnaire (see Section 5) suggesting thatclinicians are uncertain about how best to manage the condition. This may be due to

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

38

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

the fact that prolonged rupture of membranes at term is a common condition(prevalence 8%) and the number needed to treat to prevent one early-onset GBSdeath in this group is very high (10 416 women).1

Responses to the case vignettes were generally consistent with the informationprovided in the questionnaire. However, a lower proportion of obstetricians and mid-wives reported that they would offer IAP for prolonged ROM (vignette 5) in thevignette than in the questionnaire. Conversely, the neonatologists become more inter-ventionist in their responses to the vignettes than when answering the questionnaire.This pattern was seen throughout, with neonatologists being generally more active intheir management of the preterm neonate than obstetricians and midwives. Thisprobably reflects a lack of knowledge in this area by professionals not directlyinvolved in the care of the preterm neonate, as well as different perspectives on riskby these groups of health professionals. Interestingly, while the majority of cliniciansgave a preference for one action or another in the vignettes, some regarded bothcourses of action appropriate. This may be because they recognise that both coursesof action are justifiable but they have a preference for one over the other. Their choicemight also be based upon factors not included in the vignettes, for example,prematurity or low birth weight.

39

Patient vignettesPatient vignettes

7. Conclusion

7.1 Prevention of early-onset GBS disease

This audit of UK obstetric units has found that practice aimed at preventingearly-onset GBS disease in newborn babies is broadly consistent with therecommendations in the RCOG Green-top guideline. The majority of unitshave updated their GBS protocols since the RCOG guideline was publishedin November 2003 and the guideline was the most widely cited source ofevidence. Moreover, the protocols generally recommend a risk-based IAPstrategy as advocated in the guideline.

The practice reported by midwives and obstetricians in their responses toboth the questionnaire and vignettes was similarly in broad agreement withthe risk-based IAP strategy described in the guideline. Only 2% of midwivesand obstetricians reported that they followed a universal bacteriologicalscreening strategy. In addition, over 90% of clinicians said that they offeredIAP to women that had any of the ‘high risk’ indications. There wouldappear to be a slight improvement in the proportion of units offering IAP towomen with these risk factors since the previous surveys although the resultsof this audit can only be compared to a limited extent.

Nonetheless, variation was found between units in the practice reported byclinicians, between the three professions and between the practice ofclinicians within the same unit. Moreover, variation was evident in eachaspect of the care process. It was observed in:

� the sets of risk-based indications used to identify which groups ofwomen would be offered a bacteriological swab for GBS

� the timing of the swab and the site from which it is taken

� the sets of risk-based indications for IAP

� the antibiotic regimens used

� the management of neonates at risk of early-onset GBS disease.

This variation may arise from various sources. First, the variation in therecommendations on giving IAP increased as the benefit becomes small (thenumber of women needed to receive IAP to prevent one case of GBS diseaseincreases). The variation also increases as the benefits of IAP becomeuncertain because of the lack of research evidence on the precise risk posedby individual factors or combinations thereof. Indeed, the lack of evidenceseems to be the fundamental problem. It appears to have limited the strengthof the recommendations in the guideline. It may have also resulted in theprotocols containing different sets of recommendations, especially on whento offer IAP.

Second, variation in practice may be attributable to the limitations of theRCOG guideline. While the AGREE assessment revealed its documented

40

development process to be among the best of the reviewed guidelines, it stillsuffered from various weaknesses. In particular, its recommendations werejudged against the AGREE criteria to be less specific and unambiguous thanother guidelines.

Another source of variation in practice is the variety of different recom-mendations in the protocols. It is likely that these differences are partly dueto the weak evidence-base and the limitations of the RCOG guideline. Theymay also reflect the differences in local circumstances. However, thesereasons cannot account for all the variation in the protocol recommend-ations. Some may be due to flaws in the development process. For example,12% of unit protocols contained no reference to any clinical evidence and,in those protocols that did, 26% of unit protocols did not cite the RCOGguideline and 51% cited guidelines that had been superseded. It is thereforenot surprising that some protocols recommended IAP or bacteriologicaltesting in situations for which there is no evidence on effectiveness (such aswomen who had GBS in a previous pregnancy), while others specified IAPregimens that were not in any reviewed guideline.

Finally, variation in practice may be due to the difficulties inherent inimplementing a risk-based IAP strategy. The approach requires midwives andobstetricians to remember a series of actions for various circumstances and ismore complex for clinicians to implement than universal bacteriologicalscreening.25,26 This might be one explanation for the within-unit variation inpractice, although this might also arise from differences in individualperceptions of risk as well as professional attitudes.

7.2 Implications of the audit findings

The audit results have implications for various aspects of clinical practiceand governance. First, units should seek to maintain and improve theiroverall performance in the prevention of early-onset GBS disease bycontinuing to offer IAP to women with risk factors in accordance with theRCOG guideline. To reduce variation among clinicians within local units,clinical leads should ensure that all staff are familiar with their localprotocol and interpret it consistently. Consideration should also be given tohow this information might be made readily available for consultationbecause the variety of situations covered by the guideline.

RECOMMENDATIONS

� Obstetric units should continue to offer IAP to women with riskfactors in accordance with the RCOG guideline.

� A local GBS protocol should be readily available to staff andunits should ensure that it is interpreted and implementedconsistently.

Second, units need to examine their process of developing protocols. Aminority of protocols were old and may be perpetuating outdated practices.

41

ConclusionConclusion

Others contained recommendations about which women should be offeredIAP for which there is no evidence. The inclusion of such recommendationsmay reflect the wish to provide guidance in the absence of evidence but thiscan lead to suboptimal practice and reduce a clinician’s ability to cater forpatient preferences. It can also lead to a postcode lottery as inconsistenciesbecome institutionalised.

RECOMMENDATION

� Units should ensure that their protocols are up to date andconsistent with national guidelines, albeit adapted for the localcontext. Protocols should also cite the most important sourcesof evidence and give the date of development and review.

Third, consideration should be given to the following issues when theRCOG guideline is revised. The language used in some recommendationswas vague, which may give rise to inconsistencies in their interpretation. Forexample, some recommendations were phrased in terms of the quality of theevidence rather than specific behaviour (for example, ‘should offer’). Thelack of evidence for specific situations undoubtedly makes it more difficultto formulate recommendations and highlights the problems associated withthe extent to which a consensus statement should be included in the absenceof evidence. However, while it is recognised that firm recommendationscannot be made because of the lack of evidence, using behaviourally specificlanguage where possible is recognised as improving the likelihood of imple-mentation. One approach to reducing the risk of inconsistent interpretationwould be to pilot the revision before publication. Another approach wouldbe to include guidance on local audit by including suitable criteria.

Consideration should also be given to providing more advice on particularissues where variation was observed. This could include risk-based bacterio-logical testing, the procedures for taking a bacteriological swab, (inappro-priate) combinations of risk factors, practice for preterm labour, preterm ROMand preterm prelabour ROM and alternative antibiotic regimens.

RECOMMENDATIONS

� When revising the RCOG Green-top guideline, care should betaken to ensure that recommendations are unambiguous andcomprehensive.

� The revised RCOG Green-top guideline should include clearlydefined audit criteria.

Finally, the audit highlights the need for better evidence on the effectivenessof different prevention strategies. The current lack of evidence is the rootcause of much of the observed variation in practice. Consequently, fundingshould be made available for more clinical and observational research in thisarea.Pr

even

tion

of E

arly

-ons

et N

eona

tal G

roup

B S

trep

toco

ccal

Dis

ease

in U

K O

bste

tric

Uni

ts

42

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

RECOMMENDATION

� Research should be commissioned to fill the current gaps in theevidence base on which strategies to identify women for IAP aremost effective in preventing early-onset GBS disease in the UK.

7.3 Limitations of the audit method

The audit relied on two data sources to capture practice on preventionstrategies and clinical practice: the hospital protocols and the auditquestionnaire. As such, the audit relied on statements about what practiceshould occur within units and what staff reported their practice to be. Theprincipal weakness of the audit, therefore, was the lack of actual data onhow pregnant women are treated. This was unavoidable given the durationand resources available. The audit tried to collect information on the annualuse of antibiotics within the maternity units (in particular, penicillin G) toobtain an estimate of how many women might be receiving prophylaxis butthis information was not available. Consequently, the audit possibly under-estimates the level of variation in actual practice between units andindividual clinicians because respondents may have completed the question-naire by giving answers that were consistent with the guideline or protocolrather than how they actually deliver care.

The audit had several other methodological limitations. First, the analysis ofthe international guidelines and the protocols required a subjective assess-ment of their content. As format, layout and terminology were not entirelyconsistent between the documents, particularly among the protocols,judgement was necessary to extract the data. To improve the reliability of theanalysis, the guidelines and protocols were examined by two researchers.Agreement between the researchers was reasonable in both situations.

Second, it is possible that the non-respondents were systematically differentfrom those that responded. In particular, one might expect units withoutprotocols or clinicians who know that their practice differs from the UKguideline to be less likely to participate. There did not appear to be anyobvious differences in the characteristics of the units that responded andthose that did not; however, the statistical power of the comparison wasweak owing to the small numbers involved. Moreover, the response rates tothe clinical director survey and clinician questionnaires were high, eachexceeding 70%, and so the size of any bias is likely to be small.

43

ConclusionConclusion

References

1. Royal College of Obstetricians and Gynaecologists. Prevention of Early-onsetNeonatal Group B Streptococcal Disease. Green-top Guideline No. 36. London:RCOG ; 2003.

2. Kenyon S, Brocklehurst P, Blackburn A, Taylor DJ. Antenatal screening andintrapartum management of Group B Streptococcus in the UK. BJOG2004;111:226–30.

3. Heath PT, Balfour G, Weisner AM, Efstratiou A, Lamagni TL, Tighe H, et al. Group Bstreptococcal disease in UK and Irish infants younger than 90 days. Lancet2004;363:292–4.

4. Law MR, Palomaki G, Alfirevic Z, Gilbert R, Heath P, McCartney C, et al. Theprevention of neonatal group B streptococcal disease: a report by a working group ofthe Medical Screening Society. J Med Screen 2005;12:60–8.

5. Schrag SJ, Zywicki S, Farley MM, Reingold AL, Harrison LH, Lefkowitz LB, et al.Group B streptococcal disease in the era of intrapartum antibiotic prophylaxis. N EnglJ Med 2000;342:15–20.

6. Boyer KM, Gotoff SP. Prevention of early-onset group B streptococcal disease withselective intrapartum chemoprophylaxis. N Engl J Med 1986;314:1665–9.

7. Smaill F. Intrapartum antibiotics for group B streptococcal colonisation. CochraneDatabase Syst Rev 2002;CD000115.

8. Luck S, Torny M, d’Agapeyeff K, Pitt A, Heath P, Breathnach A, et al. Estimated early-onset group B streptococcal neonatal disease. Lancet 2003;361:1953–4.

9. American College of Obstetricians and Gynecologists. Committee Opinion. Preventionof early-onset group B streptococcal disease in newborns. Obstet Gynecol2002;100:1405–12.

10. Public Health Laboratory Services Group B Streptococcus Working Group. Interim‘Good Practice’ Recommendations for the Prevention of Early-onset Neonatal GroupB Streptococcal (GBS) Infection in UK. London: Public Health Service Laboratory;2001 [www.hpa.org.uk].

11. Centers for Disease Control and Prevention. Prevention of perinatal Group Bstreptococcal disease: a public health perspective. Morb Mortal Wkly Rep1996;45(RR-7):1–24.

12. Centers for Disease Control and Prevention (CDC). Prevention of perinatal group Bstreptococcal disease: revised guidelines from CDC. Morbid Mortal Wkly Rep2002;51(RR-11):1–24.

13. New Zealand College of Midwives (NZCOM). Consensus Statement. Group BStreptococcus (GBS). Christchurch: NZCOM; September 2004[www.midwife.org.nz/content/documents/127/GBS-_NZCOM.pdf].

14. Group B Strep Support [www.gbss.org.uk].

15. American College of Nurse-Midwives. Clinical Bulletin. Early-onset group B strepinfection in newborns: prevention and prophylaxis. J Midwifery Womens Health2003;2:375–81.

16. British Columbia Reproductive Care Program. Obstetric Outline 12. Group BStreptococcus in the Perinatal Period. July 2003 [www.rcp.gov.bc.ca/guidelines.htm#2].

17. Society of Obstetricians and Gynaecologists. The prevention of early-onset neonatalgroup B streptococcal disease. SOGC Clinical Practice Guidelines No. 149, September2004. J Obstet Gynaecol Can 2004;26:826–32.

18. Conseil Supérieur d’Hygiène [Belgian Health Council]. Prévention des InfectionsPérinatale à Streptocoques du Groupe B [Prevention of Perinatal Group BStreptococcal Infections]. CSH 7721. Brussels: CSH; 2003

44

[www.health.fgov.be/pls/portal/docs/PAGE/INTERNET_PG/HOMEPAGE_MENU/ABOUTUS1_MENU/INSTITUTIONSAPPARENTEES1_MENU/HOGEGEZONDHEIDSRAAD1_MENU/ADVIEZENENAANBEVELINGEN1_MENU/ADVIEZENENAANBEVELINGEN1_DOCS/7721_GBS_2003_FR.PDF].

19. Agence Nationale d’Accréditation et d’Evaluátion en Santé (ANAES). AntenatalPrevention of the Risk of early neonatal Bacterial Infection. Clinical PracticeGuidelines. Paris: ANAES; September 2001.

20. Shah, V, Ohlsson, A with the Canadian Task Force on Preventive Health Care.Prevention of Early-onset Group B Streptococcal (GBS) Infection in the Newborn:Systematic Review and Recommendations. CTFPHC Technical Report #01-6. London,ON: Canadian Task Force; 2001 [www.ctfphc.org/].

21. Royal Australian and New Zealand College of Obstetricians and Gynaecologists(RANZCOG). Swabbing for Group B Streptococcus. Statement No. C-Obs 19.Melbourne: RANZCOG; July 2005[www.ranzcog.edu.au/publications/statements/C-obs19.pdf].

22. Campbell N, Eddy A, Darlow B, Stone P, Grimwood K, New Zealand GBS ConsensusWorking Party. The prevention of early-onset neonatal group B streptococcus infection:technical report from the New Zealand GBS Consensus Working Party. N Z Med J2004;117(1200):U1023.

23. AGREE Collaboration. Development and validation of an international appraisalinstrument for assessing the quality of clinical practice guidelines: the AGREE project.Qual Saf Health Care 2003;12:18–23.

24. Schrag SJ, Zell ER, Lynfield R, Roome A, Arnold KE, Craig AS, et al. A population-based comparison of strategies to prevent early-onset group B streptococcal disease inneonates. N Engl J Med 2002;347:233–9.

25. Lieu TA, Mohle-Boetani JC, Ray GT, Ackerson LM, Walton DL. Neonatal group Bstreptococcal infection in a managed care population. Perinatal Group B StreptococcalInfection Study Group. Obstet Gynecol. 1998;92:21–7.

26. Schuchat A, Roome A, Zell ER, Linardos H, Zywicki S, O’Brien KL. Integratedmonitoring of a new group B streptococcal disease prevention program and otherperinatal infections. Matern Child Health J 2002;6:107–14.

27. Kenyon SL, Taylor DJ, Tarnow-Mordi W, ORACLE Collaborative Group. Broad-spectrum antibiotics for preterm, prelabour rupture of fetal membranes: the ORACLEI randomised trial. ORACLE Collaborative Group. Lancet 2001;357:979–88. Erratum:Lancet 2001;358:156.

28. Kenyon SL, Taylor DJ, Tarnow-Mordi W, ORACLE Collaborative Group. Broad-spectrum antibiotics for spontaneous preterm labour: the ORACLE II randomisedtrial. ORACLE Collaborative Group. Lancet 2001;357:989–94.

45

ReferencesReferences

Appendix 1. RCOG guideline classifications forlevels of evidence and strength ofrecommendation

Strength of recommendations

Requires at least one randomised controlled trial as part of a body of literatureof overall good quality and consistency addressing the specific recommend-ation. (Evidence levels Ia, Ib)

Requires the availability of well-controlled clinical studies but no randomisedclinical trials on the topic of recommendations. (Evidence levels IIa, IIb, III)

Requires evidence obtained from expert committee reports or opinions and/orclinical experiences of respected authorities. Indicates an absence of directlyapplicable clinical studies of good quality. (Evidence level IV)

Good practice point. Recommended best practice based on the clinicalexperience of the guideline development group.

Levels of evidence

Ia Evidence obtained from meta-analysis of randomised controlled trials.

Ib Evidence obtained from at least one randomised controlled trial.

IIa Evidence obtained from at least one well-designed controlled study withoutrandomisation.

IIb Evidence obtained from at least one other type of well-designed quasi-experi-mental study.

III Evidence obtained from well-designed non-experimental descriptive studies,such as comparative studies, correlation studies and case studies.

IV Evidence obtained from expert committee reports or opinions and/or clinicalexperience of respected authorities.

46

A

B

C

✔

Appendix 2. Search strategy for the review ofinternational GBS guidelines

A systematic search for international guidelines on the prevention of early-onsetneonatal group B streptococcal disease was performed. The abstracting databasesEmbase and Medline (1966 to August 2005) were searched using the phrases ‘guide-line$ or consensus’ and ‘streptococc$ and agalactiae’ or ‘streptococc$ and group B orBeta Strep$’ or ‘Beta Haemolytic Strep$’. The Cochrane Library was searched usingMeSH headings ‘Streptococcal infections’, ‘Streptococcus agalactiae’ and free text‘group b streptococc*’.

Guideline websites were searched, using the same set of phrases, following allrelevant nested links (see below for list of sites visited). In addition, an internet searchusing Google™ was conducted for specialist or professional agencies from English-speaking countries or countries likely to have an English version website (such asmedical colleges, associations or societies). Guidelines or consensus statements fromthese websites were assessed for relevance.

The search was limited to English language guidelines.

Organisation Internet address Date searched

Canadian Medical Association http://mdm.ca/cpgsnew/cpgs/ 11 August 2005Infobase

Guidelines International Network www.g-i-n.net/ 11 August 2005

Health Services Technology www.ncbi.nlm.nih.gov/books/ 25 August 2005Assessment Texts bv.fcgi?rid=hstat

National Electronic Library http://libraries.nelh.nhs.uk/ 18 August 2005for Health Guideline Finder guidelinesFinder/

National Institute for Health www.nice.org.uk/ 18 August 2005and Clinical Excellence

National Institutes of Health http://consensus.nih.gov/ 25 August 2005Consensus Development Program

New Zealand Guidelines Group www.nzgg.org.nz/ 18 August 2005

Scottish Intercollegiate Guidelines www.sign.ac.uk/ 25 August 2005Network

US National Guideline www.guidelines.gov/ 11 August 2005Clearinghouse

47

Appendix 3. The midwife questionnaire

48

49

Appendix 3Appendix 3

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

50

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

51


Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

52

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

53


Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

54

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

55


Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

56

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

57


Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

58

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Appendix 4. Agreement among midwives andobstetricians on the management ofwomen with GBS risk factors

Previous baby affected by GBS disease

Obstetricians Obstetricians

Midwives No swab Swab Midwives No IAP IAP

No swab 34 14 No IAP 0 1Swab 36 26 IAP 1 101

GBS bacteriuria in current pregnancy




Incidental finding in current pregnancy



No swab . . No IAP 0 5Swab . . IAP 2 103

Maternal GBS carriage in previous pregnancy




Planned caesarean section at term




59

Prelabour preterm rupture or membranes (< 37 weeks)




Vaginal discharge



No Swab 3 13 No IAP 2 8Swab 9 93 IAP 5 85

Suspected chorioamnionitis



No Swab 0 9 No IAP 1 3Swab 4 105 IAP 7 98

IAP = IAP offered if swabbed and result positive or IAP on risk factor only

Intrapartum clinical situations

Preterm labour (35–37 weeks of gestation)

Obstetricians

Midwives No IAP IAP

No IAP 57 32IAP 14 26

Preterm labour (< 35 weeks of gestation)

Obstetricians

Midwives No IAP IAP

No IAP 51 34IAP 10 34

Intrapartum fever at term (> 38˚C)

Obstetricians

Midwives No IAP IAP

No IAP 8 18IAP 8 100Pr

even

tion

of E

arly

-ons

et N

eona

tal G

roup

B S

trep

toco

ccal

Dis

ease

in U

K O

bste

tric

Uni

ts

60

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Prolonged ROM (18 hours or more)

Obstetricians

Midwives No IAP IAP

No IAP 39 24IAP 11 59

Preterm prelabour ROM, with term labour

Obstetricians

Midwives No IAP IAP

No IAP 47 19IAP 17 43

61


Appendix 5. Clinician responses to patientvignettes

Vignette 1A 24-year-old primigravida presents at 37 weeks of gestation with a history of vaginaldischarge. A high vaginal swab confirms the presence of GBS.

Midwives Obstetricians Neonatologists(%) (%) (%)

Prescribe oral Agree 12 29 27antibiotics now Disagree 83 69 49

Don’t know/missing 5 2 24

Prescribe intravenous Agree 94 95 91in antibiotic Disagree 6 5 4prophylaxis labour Don’t know/missing 1 1 4

Clinically observe the Agree 91 93 88newborn infant for Disagree 6 5 9signs of early-onset Don’t know/missing 4 2 3GBS disease

Give the newborn infant Agree 12 18 24intravenous antibiotics Disagree 80 71 63pending the results Don’t know/missing 8 11 12of blood cultures

Vignette 2A 36-year-old parous woman attends for booking at 11 weeks of gestation whose previousbaby was diagnosed with early-onset GBS disease.


Take vaginal swab Agree 47 32 43Disagree 47 67 37Don’t know/missing 7 1 19



Prescribe intravenous Agree 96 96 88antibiotic prophylaxis Disagree 1 2 6in labour Don’t know/missing 2 2 6

Clinically observe the Agree 92 91 76newborn infant for signs Disagree 2 8 16of early-onset GBS disease Don’t know/missing 6 1 9

Give the newborn infant Agree 28 41 44intravenous antibiotics Disagree 58 42 39pending the results of Don’t know/missing 14 18 17blood cultures62

Vignette 3A 27-year-old primigravida (with no known allergies) admitted in spontaneous labour at 38weeks of gestation. She makes slow progress in labour despite artificial rupture of themembranes and has a recorded temperature of 38.5°C.


Prescribe intravenous Agree 30 29 37penicillin (narrow) Disagree 53 66 35


Prescribe broad-spectrum Agree 74 84 50antibiotics Disagree 12 16 27



Give the newborn infant Agree 27 38 47intravenous antibiotics Disagree 61 46 43pending the results of Don’t know/missing 11 16 9blood cultures

Vignette 4A 19-year-old primigravida (with no known allergies) presents with prelabour rupture ofmembranes at 31 weeks of gestation. She receives oral erythromycin in keeping with thehospital protocol and a high vaginal swab confirms the presence of GBS.




Prescribe intravenous Agree 94 97 91antibiotic prophylaxis Disagree 3 2 4in labour Don’t know/missing 3 1 5


Give the newborn infant Agree 54 51 81intravenous antibiotics Disagree 34 32 11pending the results of Don’t know/missing 12 17 8blood cultures

63


Vignette 5A 25-year-old parous woman at 41 weeks of gestation experiences a 30-hour spontaneousrupture of membranes (SROM).


Prescribe IV penicillin Agree 14 21 1818 hours after SROM Disagree 75 73 49


Prescribe intravenous Agree 36 40 40penicillin at the onset Disagree 55 56 34of labour Don’t know/missing 9 4 27


Give the newborn infant Agree 9 15 16intravenous antibiotics Disagree 79 69 75pending the results Don’t know/missing 12 16 10of blood cultures

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

64

Prev

entio

n of

Ear

ly-o

nset

Neo

nata

l Gro

up B

Str

epto

cocc

al D

isea

se in

UK

Obs

tetr

ic U

nits

Documents

The Prevention of Early-onset Neonatal GBS Disease in UK ... · The Prevention of Early-onset Neonatal Group B Streptococcal Disease in UK Obstetric Units An audit of reported practice