The WHO technical consultation adopted 20 recommendations covering prioritized questions related to the prevention and treatment of maternal peripartum infections. The prevention aspect of the recommendations is focused on routine use of minor procedures (e.g. perineal/pubic shaving), antiseptic agents for vaginal and caesarean birth, and antibiotic prophylaxis for preventing bacterial infection in infection-prone conditions and obstetric procedures prelabour rupture of membranes, meconium-stained amniotic fluid, perineal tears, manual removal of the placenta, operative vaginal birth and caesarean section. When there was evidence of effectiveness regarding the use of any antibiotic, the comparative effectiveness and safety of different classes or regimens of antibiotics were considered to issue additional recommendations. The recommendations on the treatment of maternal peripartum infections are specific to antibiotic management of chorioamnionitis and postpartum endometritis.
The quality of the supporting evidence rated as “very low”, “low”, “moderate” or “high” and the strength of each recommendation assessed as “strong” or “conditional” are indicated. To ensure that each recommendation is correctly understood and appropriately implemented in practice, additional “remarks” reflecting the summary of the discussion by GDG are included under the recommendation where necessary.
Prevention of maternal peripartum infections
RECOMMENDATION 1
Routine perineal/pubic shaving prior to giving vaginal birth is not recommended.
(Conditional recommendation based on very low-quality evidence)
REMARKS
This recommendation applies to all hair shavings around the female external genital area within the context of vaginal birth. It does not apply to women being prepared for caesarean section.
The decision regarding perineal/pubic shaving should be left to the woman and not her health care giver. In situations where a woman chooses to have perineal/pubic shaving prior to birth, she could be advised to shave wherever, and by whomever she is most comfortable with (e.g. at home shortly before the time of labour and childbirth).
Review question:
Among pregnant women in labour (P), does routine perineal/pubic shaving prior to giving vaginal birth (I), compared with no perineal/pubic shaving (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on routine perineal/pubic shaving before childbirth for the prevention of maternal and neonatal infectious morbidities was extracted from a Cochrane systematic review of three randomized trials involving 1039 women (
18). The trials were conducted in hospitals in the USA (Baltimore, Dallas) and Thailand (Bangkok).
All trials included women admitted to hospital prior to giving birth. The trials compared perineal shaving versus no perineal shaving (which included clipping or cutting of perineal hair). In two trials (involving 650 women), skin preparation was performed in both intervention and control groups by scrubbing the external genitalia and inner thighs with soap and water or povidone-iodine spray; or with 4% chlorhexidine and rinsing with 1:100 savlon solution.
Perineal shaving versus no perineal shaving (EB Table 1)
Compared to no perineal shaving, perineal shaving did not reduce the risk of maternal febrile morbidity (relative risk (RR) 1.14, 95% confidence interval (CI) 0.73 to 1.76; 3 trials, 997 women). There were no differences in the number of women who were colonized by gram-positive bacteria (RR 1.16, 95% CI 0.82 to 1.64; 1 trial, 150 women), although there was a reduction in the number of women who were colonized by gram-negative bacteria (RR 0.83, 95% CI 0.70 to 0.98; 1 trial, 150 women).
There were no significant differences between comparison groups for perineal wound infection (defined as pain and erythema of the margins of perineal or episiotomy wound with/without serious or purulent discharge) (RR 1.47, 95% CI 0.80 to 2.70; 1 trial, 458 women) or perineal wound dehiscence (RR 0.33, 95% CI 0.01 to 8.00; 1 trial, 458 women).
No neonatal infections were reported in either group (1 trial, 458 women).
Women's satisfaction as measured by a five-point Likert scale did not show any difference between the comparison groups (mean difference 0.00, 95%CI -0.13 to 0.13, 1 trial, 458 women).
None of the trials reported on other critical outcomes such as cost of care, and side-effects of perineal shaving (e.g. perineal discomfort, pain during hair regrowth).
Considerations related to the strength of the recommendation
Quality of the evidence
The quality of the evidence was graded from very low and moderate. Overall the quality of evidence was graded as very low.
Balance of benefits and harms
There is no evidence of any clinical benefit of routine perineal (or pubic) shaving before childbirth, although the quality of evidence is very low. Potential complications of perineal shaving, such as irritation and redness of the perineum, multiple superficial scratches from the razor, vulval itching and burning sensation, are not clinically serious but can be discomforting to women. Non-clinical outcomes that are considered very important to women such as embarassment during the procedure and discomfort during hair regrowth were not reported by any of the studies. In the absence of any clinical benefits, it is reasonable to conclude that perineal shaving has a higher potential of leading to undesirable consequences for women.
Values and preferences
Routine shaving is a procedure that is no longer practised in some countries but is still being performed in health facilities across all settings, often as part of maintenance of obstetric tradition. The clinical implications of the substantial variations in the currently available shaving methods across contexts (e.g. shaving creams/gel versus razor) are not known. Women's preferences about perineal shaving might differ between individual women and between religious and cultural settings. Pregnant women are likely to place a high value on maintaining their autonomy and dignity and avoid possible embarrassment from being shaved by a health care provider. Therefore, most women would prefer a policy that respects their values and preferences rather than one that mandates all women to have perineal shaving prior to childbirth.
Resource implications
Implementation of this recommendation is likely to reduce costs associated with resources required for shaving women on admission into the labour room.
RECOMMENDATION 2
Digital vaginal examination at intervals of four hours is recommended for routine assessment of active first stage of labour in low-risk women.
(Strong recommendation based on very low-quality evidence)
REMARKS
There is currently no direct evidence on the most appropriate frequency of vaginal examinations to prevent infectious morbidity in the mother and baby, and this recommendation was based on consensus reached by the GDG, and in agreement with an existing recommendation in the
WHO recommendations for augmentation of labour (
19).
The recommended time intervals are consistent with timing of vaginal examination on the partograph and further reinforce the importance of using partograph as an essential tool to implement this practice. Priority must be given to restricting the frequency and total number of vaginal examinations. This is particularly crucial in situations where there are other risk factors for infection (e.g. prolonged rupture of amniotic membranes and long duration of labour).
The GDG acknowledged that the frequency of vaginal examinations is dependent on the context of care and the progress of labour. The group agreed that vaginal examinations at intervals more frequent than specified in this recommendation may be warranted by the condition of the mother or the baby.
Vaginal examinations of the same woman by multiple care givers around the same time or at different time points should be avoided. The group noted that this practice is common in teaching settings where multiple cadres of staff (or students) perform vaginal examinations for learning purposes.
Review question:
Among pregnant women undergoing labour monitoring (P), does routine vaginal examination at intervals of four hours (I), compared with shorter intervals (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on the intervals of vaginal examination during labour was extracted from a Cochrane systematic review evaluating the effectiveness of vaginal examination at term for assessing labour progress (
20). The review included two trials, each examining a different comparison. One trial conducted in Ireland in 307 women with ruptured membranes compared routine vaginal examinations (every one or two hours) with rectal examinations to assess progress in labour. A trial in the UK compared two-hourly with four-hourly vaginal examinations in nulliparous women in labour (150 women randomized, 109 included in the analysis).
Two-hourly versus four-hourly vaginal examinations in labour (no GRADE table included)
In the UK trial with 109 women comparing two-versus four-hourly vaginal examination to assess progress in labour, no maternal or neonatal critical outcomes related to infection were reported. However, the trial reported no significant differences between the two intervals for duration of labour, caesarean section, operative vaginal birth and use of labour augmentation.
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was graded as very low.
Balance of benefits and harms
There is no evidence from randomized controlled trials to evaluate the relationship between the intervals of vaginal examinations during labour and maternal and newborn infectious morbidities. Multiple vaginal examinations are recognized contributors to infectious morbidities associated with prolonged labour, so it is scientifically plausible that vaginal examinations more frequent than every four hours could potentially increase the risk of infection for both the mother and the infant. In the absence of any evidence of benefits in relation to other clinical outcomes unrelated to infections, the undesirable consequences of more frequent vaginal examinations for women are likely to overcome the potential benefits. Evidence and recommendations related to routine vaginal examination for assessing the progress of labour were included in the WHO recommendations for augmentation of labour (19), which recommends four-hour intervals for assessing the progress of labour.
Values and preferences
Pregnant women across all settings are likely to place a high value on minimal labour interventions, including less invasive procedures such as vaginal examinations. Women are unlikely to accept frequent vaginal examinations in the absence of any clinical maternal or fetal indication.
Resource implications
Implementation and adherence to this recommendation is likely to save costs related to staff time and materials that would be required to perform vaginal examinations at intervals more frequent than every four hours.
RECOMMENDATION 3
Routine vaginal cleansing with chlorhexidine during labour for preventing infectious morbidities is not recommended.
(Strong recommendation based on moderate-quality evidence)
REMARKS
This recommendation applies to women in whom vaginal birth is anticipated.
Vaginal cleansing in this context refers to vaginal douching or any mechanical irrigation or washing of the vaginal canal and cervix with chlorhexidine solution or vaginal application of chlorhexidine gel.
The infectious morbidities considered in the evidence base did not include GBS and HIV-related infections. However, this intervention is also not recommended for preventing early neonatal GBS infection in women with GBS colonization (
see
Recommendation 4).
Review question:
Among pregnant women in labour (P), does routine vaginal cleansing with an antiseptic agent (I), compared with no vaginal cleansing with an antiseptic agent (C), prevent infectious morbidities and improve outcomes following vaginal birth (O)?
Summary of evidence
Evidence on routine vaginal cleansing with an antiseptic agent during labour for the prevention of maternal and neonatal infections was extracted from a Cochrane systematic review of three trials involving 3012 women (
21). All included studies were conducted in hospitals in the USA between 1997 and 2003. The trials included women admitted to the hospital prior to birth, ≥24 weeks pregnant and excluded women with GBS infections or known allergy to chlorhexidine. The intervention was chlorhexidine vaginal douching during labour versus sterile water. Two of the studies used 200 ml of 0.2% of chlorhexidine, while the third study used 20 ml of 0.4% chlorhexidine for vaginal irrigation during labour.
Chlorhexidine vaginal douching during labour versus placebo (EB Table 2)
No differences were observed in the incidence of chorioamnionitis (RR 1.10, 95% CI 0.86 to 1.42; 3 trials, 3012 women) or postpartum endometritis (RR 0.83, 95% CI 0.61 to 1.13; 3 trials, 3012 women). No side-effects from the use of chlorhexidine were observed among women in the two groups (2 trials, 2065 women).
There was no difference in perinatal mortality between the two groups (RR 1.00, 95% CI 0.17 to 5.79; 2 trials, 2017 neonates).
For neonatal outcomes, there were no differences between groups for neonatal sepsis (RR 0.75, 95 % CI 0.17 to 3.35; 3 studies, 2987 neonates), blood cultures confirming neonatal infections, (RR 0.75, 95% CI 0.17 to 3.35; 2 trials, 2077 neonates), neonatal pneumonia (RR 0.33, 95% CI 0.01 to 8.09, 1 trial, 910 neonates) or neonatal meningitis (RR 0.34, 95% CI 0.01 to 8.29, 1 trial, 1024 neonates). No difference was observed between groups of newborns who received antibiotic treatment (RR 1.65, 95% CI 0.73 to 3.74, 1 trial, 910 neonates).
The included trials did not report on any other critical outcomes.
Considerations related to the strength of the recommendation
Quality of evidence
The quality of the evidence was graded as moderate for reported critical outcomes.
Balance of benefits and harms
There is no evidence of clinical benefits to support routine vaginal cleansing with chlorhexidine during labour in women giving vaginal birth. The possible side-effects were not reported by any of the studies. Other systematic reviews evaluating the routine use of chlorhexidine for preventing HIV and GBS infections following vaginal birth did not show any clinical benefits either. Although vaginal douching with chlorhexidine is relatively inexpensive, can be performed within minutes and is unlikely to interfere with the women's labour, the use of an additional intervention with no clinical benefit further undermines the natural process of birth. Additionally, the unnecessary use of medical disinfectants in general might contribute to antimicrobial resistance, although such a situation rarely emerges with chlorhexidine even after long-term use.
Values and preferences
Women are likely to prefer minimal interference with the process of labour, and some women may find the procedure invasive and discomforting. Health care providers and policy-makers across settings are likely to place a higher value on saving health care costs and, therefore, choose not to use the intervention.
Resource implications
Although chlorhexidine is relatively cheap, and it is technically feasible to perform vaginal cleansing during labour with minimal increase in resource use, the intervention is not cost–effective, as there is no added benefits for mother and baby.
RECOMMENDATION 4
Routine vaginal cleansing with chlorhexidine during labour in women with group B Streptococcus (GBS) colonization is not recommended for prevention of early neonatal GBS infection.
(Conditional recommendation based on very low-quality evidence)
REMARKS
This recommendation was based on the lack of clinical benefits for the neonate and not on the potential effect of the intervention on GBS-related maternal infectious morbidity.
The GDG acknowledged the considerable variations in policies regarding the screening for GBS colonization in pregnant women. Therefore, the group agreed that this recommendation should be implemented within the context of local policy and guidance on screening for GBS colonization.
Review question:
Among pregnant women with vaginal, rectal or urethral colonization with group B Streptococcus (GBS) (P), does routine vaginal cleansing with an antiseptic agent during labour (I), compared with no vaginal cleansing with an antiseptic agent (C), prevent neonatal infectious morbidities and improve neonatal outcomes (O)?
Summary of evidence
Evidence on the use of antiseptic agents for routine vaginal cleansing with an antiseptic agent in GBS-colonized women during labour to prevent neonatal infectious morbidities and improve neonatal outcomes was extracted from a Cochrane systematic review of four trials that included 1125 preterm and term infants (
22). The trials were conducted in Belgium, the Netherlands, Norway, Sweden and the UK.
Rapid screening or culture tests were used to diagnose GBS colonization. Most of the trials excluded women who received antibiotics before delivery, planned caesarean sections and fetal deaths. The interventions were different methods of application and preparations of chlorhexidine: vaginal wash, lubricated gloves with cream, or gel application around the vaginal fornices. Comparison groups included mechanical wash with placebo (sterile water) or no treatment.
The trials did not report on maternal morbidities, cost of care or maternal satisfaction.
Chlorhexidine (vaginal wash or gel/cream) versus placebo or no treatment (EB Table 3)
The two trials reporting the incidence of early onset GBS-related neonatal morbidities within the first seven days of life found no differences between groups: GBS sepsis and/or meningitis (RR 2.32, 95% CI 0.34 to 15.63; 2 trials, 987 infants) or GBS pneumonia (RR 0.35, 95% CI 0.01 to 8.60; 2 trials, 987 infants). The number of infants colonized with GBS within the first seven days of life did not differ between the chlorhexidine and placebo or no treatment groups (RR 0.65, 95% CI 0.36 to 1.18; 3 trials, 328 infants).
No neonatal deaths due to early-onset GBS infection were reported (1 trial, 190 infants).
In the three trials that reported on maternal side-effects, a significantly greater number of mothers developed minor side-effects (stinging and irritation) related to the use of chlorhexidine (RR 8.5, 95% CI 1.60 to 45.28; 3 trials, 1066 women). No adverse effects were observed in infants in either groups (3 trials; 1066 infants).
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was graded as very low for almost all critical outcomes reported. Overall, the quality of evidence was graded as very low.
Balance of benefits and harms
There is no clear evidence that routine vaginal cleansing with chlorhexidine during labour is effective in preventing early onset GBS-related disease in preterm and term neonates. Given the very low-quality evidence, there is little certainty about the estimates derived from these trials. Routine vaginal cleansing with chlorhexidine appears to increase the occurrence of stinging sensation and irritation in the vagina of treated women. Without proof of any clinical benefits, such undesirable consequences of chlorhexidine treatment are likely to be a major determinant in clinical decision-making. Although chlorhexidine has a low impact on antimicrobial resistance even with prolonged use, unnecessary use on a large scale may contribute to decreasing sensitivity of microorganisms to antimicrobial agents in the long term.
Values and preferences
The values and preferences of health care providers and women colonized with GBS may vary according to the emphasis on GBS neonatal disease in their settings. Acceptance of the minor side-effects and discomfort associated with the intervention for a relatively rare neonatal disease is also likely to vary across settings.
Resource implications
Implementation of this recommendation is likely to save costs required to perform vaginal cleansing for all women with GBS colonization during labour.
RECOMMENDATION 5
Intrapartum antibiotic administration to women with group B Streptococcus (GBS) colonization is recommended for prevention of early neonatal GBS infection.
(Conditional recommendation based on very low-quality evidence)
REMARKS
This recommendation was made based on clinical benefits for the neonates, as there was insufficient evidence on the effect of antibiotic administration on maternal infectious morbidities.
As the evidence came from studies that tested ampicillin or penicillin G, either antibiotic should first be considered for treatment except where there are contraindications (e.g. allergy history) or GBS strain has been microbiologically shown to be penicillin-resistant.
The GDG noted that although women with urethral GBS colonization were not included in the trials, the recommendation should also be applied to such women because urinary colonization is often persistent following identification and treatment during pregnancy.
The GDG acknowledged the challenges of implementing GBS screening for all pregnant women, particularly in low-resource countries and in settings where the prevalence of maternal colonization is low, coupled with the limitations in providing appropriate preventive measures and follow-up to the majority of the women screened positive. Therefore, the group agreed that this recommendation should be implemented within the context of local policy and guidance on screening for GBS colonization.
In deciding whether or not to administer antibiotics during labour to GBS-colonized women, clinicians should balance the risk and benefits of the use of antibiotics, taking into account different factors (e.g. colonization rates and factors associated with increased transmission).
Review question:
Among pregnant women with vaginal, rectal or urethral colonization with group B Streptococcus (GBS) (P), does routine administration of antibiotics during labour (I), compared with no antibiotics (C), prevent neonatal infectious morbidities and improve maternal and neonatal outcomes (O)?
Summary of evidence
Evidence on the use of antibiotics during labour or delivery for known maternal GBS colonization to prevent infectious morbidity from GBS was extracted from a Cochrane systematic review of four trials including 852 women (
23).
Trials included women with vaginal and/or rectal GBS colonization ascertained by cultures in three trials, at different postmenstrual ages, or by rapid latex agglutination test at the time the mother was giving birth in one trial. Two trials included women at 36 weeks of gestation or more. The included trials were conducted in Finland, Spain and the USA.
Two trials excluded women with rupture or prolonged rupture of membranes, and two trials excluded women undergoing planned caesarean section. Other relevant exclusion criteria varied between trials: antibiotic intake within the preceding seven days, fever prior to delivery or fetal death prior to labour.
Two trials compared ampicillin versus no treatment – using different antibiotics regimens (2 g of ampicillin IV followed by 1 g every four hours until giving birth or 500 mg of ampicillin IV every six hours until delivery). One trial compared penicillin with no treatment (5 million units of penicillin G IV every six hours during labour, and if labour lasted more than 18 hours, 1 million units of penicillin orally every eight hours until parturition). One trial compared ampicillin with penicillin.
The included trials did not report on antimicrobial resistance or maternal satisfaction.
Intrapartum antibiotics versus no treatment for GBS-positive women (EB Table 4)
Only one trial reported maternal outcomes. No significant reduction was observed between comparison groups for maternal sepsis in the peri/postpartum period (RR 0.31, 95% CI 0.01 to 7.49; 1 trial, 160 women) or puerperal infections (RR 0.16, 95% CI 0.01 to 3.03; 1 trial, 121 women).
In one small trial, intrapartum antibiotic administration did not show reductions in neonatal mortality from all causes (RR 0.19, 95% CI 0.01 to 3.82; 164 infants), neonatal mortality from early onset GBS infection (RR 0.31, 95% CI 0.01 to 7.50; 164 infants) or neonatal mortality from infections caused by bacteria other than GBS (RR 0.31, 95% CI 0.01 to 7.50; 164 infants).
There was a statistically significant reduction in the incidence of early-onset (postnatal age <7 days) GBS neonatal infection (RR 0.17, 95% CI 0.04 to 0.74; 3 trials, 488 infants, number needed to treat to benefit = 25) and probable early infection (RR 0.17, 95% CI 0.03 to 0.91; 2 trials, 324 infants). There was no difference on the late onset (≥ 7 days) and GBS neonatal infection (RR 0.36, 95% CI 0.01 to 8.69; 2 trials, 289 infants).
Analysis of the incidence of other neonatal infectious morbidities such as neonatal sepsis, meningitis, urinary tract infection or pneumonia due to bacterial organisms other than GBS showed no difference between the two comparison groups (RR 1.00, 95% CI 0.15 to 6.79; 2 trials, 289 infants).
Considerations related to the strength of the recommendation
Quality of the evidence
The quality of the evidence was graded as very low for all critical outcomes. Overall, the quality of the evidence was graded as very low.
Balance of benefits and harms
There is some evidence of reduced risk of early neonatal GBS sepsis, but with no reduction in all cause or GBS-related neonatal mortality. This lack of effect on neonatal mortality may be the result of the large size needed to demonstrate differences between the comparison groups. Available evidence was limited to three studies with relatively small sample sizes and largely at high risk of bias. In addition, most of the studies included in the review did not report on maternal or neonatal side-effects of the administration of antibiotics. Administration of antibiotics to all colonized women might expose mothers and newborns to potential side-effects of antibiotics and contribute to antimicrobial resistance. The decision to recommend the administration of antibiotics to prevent GBS also needs to take into consideration other factors related to the incidence of GBS colonization, screening and transmission of GBS from the mother to the baby. GBS vaginal colonization rates vary between populations and are not always associated with clinical symptoms in the mother or the infant. In addition, the transmission rate of GBS from the mother to her baby is known to be very low.
The available evidence is insufficient to assess the balance of benefits and harms for subgroups of women who may be at a higher risk of transmission, such as women in preterm labour, women with ruptured membranes or those with a previous baby with neonatal GBS sepsis. The current evidence was derived from high-income countries, and it is possible that the baseline incidence of GBS colonization and neonatal transmission might confer higher clinical benefits for women treated with antibiotics during labour.
Values and preferences
Pregnant women in all settings are likely to place a high value on a reduced risk of perinatal transmission of GBS, and a low value on possible side-effects of prophylactic antibiotics. However, there is likely to be significant variability in the values and preferences of health care providers and policy-makers across settings in terms of the balance between the rate of perinatal transmission of neonatal GBS infection and the potential impact of antimicrobial resistance on public health.
Resource implications
The use of antibiotics during labour for GBS-colonized women is likely to slightly increase health care costs, particularly in settings where women are routinely screened during pregnancy and labour. However, the implementation of this recommendation is likely to result in cost savings related to the management of potential adverse outcomes (e.g. intensive care admission, prolonged hospital stay) among newborns born to GBS-colonized mothers.
RECOMMENDATION 6
Routine antibiotic prophylaxis during the second or third trimester to all women with the aim of reducing infectious morbidity is not recommended.
(Strong recommendation based on very low-quality evidence)
REMARKS
This recommendation applies to an unselected population of pregnant women in the second or third trimester of pregnancy.
The GDG noted that prophylactic antibiotic use may be necessitated in a clearly defined group of women with high-risk pregnancy, but the description in the systematic review is inadequate to identify such a group.
The GDG identified the evaluation of the effects of routine antibiotics in specific groups of women with high-risk pregnancy as a research priority.
Review question:
Among women in the second or third trimester of pregnancy (P), does routine antibiotic prophylaxis (I), compared with no antibiotic prophylaxis (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on the administration of prophylactic antibiotics to pregnant women during the second or third trimester for the prevention of infectious morbidities was extracted from a Cochrane systematic review of eight trials involving approximately 4300 women (3663 included in the analysis) (
24).
One trial each was conducted in Belgium, the Netherlands and Malawi, and two trials each in India and the USA. In addition, one single-country and one multicountry trial also included data from Kenya.
Women in their second or third trimester of pregnancy (between 14 and 34 weeks of gestation) who were not in labour were eligible for inclusion in the trials. In three trials, women with high-risk pregnancies (defined variously as a history of preterm birth, low birthweight, stillbirth or early perinatal death, pre-pregnancy weight < 50 kg or previous preterm birth who had bacterial vaginosis diagnosis in current pregnancy). Women receiving antibiotics due to infection were excluded.
Trials used a range of antibiotics and administration routes: oral cefetamet-pivoxil, cephalexin, metronidazole, azythromycin or erythromycin, intramuscular (IM) ceftriaxone or clindamycin vaginal cream.
Subgroup analyses were performed on women who might be at higher risk of presenting with adverse outcomes. Authors defined high-risk women as those who had a previous spontaneous preterm delivery, history of low birthweight, a diagnosis of bacterial vaginosis in the current pregnancy (BV identified after enrolment and antibiotics used only for prophylaxis before knowing if the participant had BV or not) or a pre-pregnancy weight less than 50 kg.
There were no data reported on side-effects of antibiotics, antimicrobial resistance or cost of care.
Prophylactic antibiotics versus placebo (EB Table 5)
There was a significant reduction in postpartum endometritis (RR 0.53, 95% CI 0.35 to 0.82; 3 trials, 627 women) in the group receiving antibiotics compared to the placebo group, but no significant difference between groups regarding chorioamnionitis (RR 0.62, 95% CI 0.10 to 3.62; 1 trial, 229 women). A reduction was observed for prelabour rupture of membranes (RR 0.34 95% CI 0.15, 0.78; 1 trial, 229 women), but not for preterm prelabour rupture of membranes (RR 0.31 95% CI 0.06, 1.49; 1 trial, 229 women).
There was no significant difference between groups for perinatal mortality (RR 0.83, 95% CI 0.57 to 1.20; 4 trials, 2710 infants).
There was no significant difference between groups for preterm birth (RR 0.88 95% CI 0.72, 1.09; 6 trials, 3663 women), low birthweight (RR 0.86 95% CI 0.53, 1.39; 4 trials, 978 women) or mean birthweight (RR 41.60, 95% CI -78.20 to 161.40; 4 trials, 978 women).
The included trials did not report any serious adverse effects of antibiotic prophylaxis.
Prophylactic antibiotics versus placebo: unselected pregnant women
There were no significant reductions in the incidence of postpartum endometritis (RR 0.51 95% CI 0.24 to 1.08; 2 trials, 431 women) or chorioamnionitis (RR 0.62 95% CI 0.10 to 3.62; 1 trial, 229 women) among unselected women.
There was no significant difference between groups on perinatal mortality (RR 0.84, 95% CI 0.57 to 1.23; 2 trials, 2315 infants) among unselected women.
There were no differences in the risk of low birthweight (RR 107; 95% CI 0.71 to 1.63; three trials, 725 women), small for gestational age (RR 1.29 95% CI 0.42 to 3.96; one trial, 229 women) or congenital anomalies (RR 1.49 95% CI 0.20 to 11.14; two trials, 463 women).
Prophylactic antibiotics versus placebo: high-risk women
There was a significant reduction in postpartum endometritis in high-risk pregnant women (women with a history of preterm birth, low birthweight, stillbirth or early perinatal death) (RR 0.55; 95% CI 0.33 to 0.92; 1 trial, 196 women) and postpartum detected gonococcal infection (RR 0.35, 95% CI 0.13 to 0.94; 1 trial, 204 women) in the group receiving antibiotics compared to the placebo group.
There were no differences between subgroups of high-risk pregnant women on preterm delivery, except in the subgroup of pregnant women with a previous preterm birth who had bacterial vaginosis during the current pregnancy (RR 0.64, 95% CI 0.47 to 0.88; 1 trial, 258 women, subgroup differences P = 0.08).
There was no significant difference between groups on perinatal mortality among different high-risk groups (in women with a history of preterm birth, low birthweight, stillbirth or early perinatal death) (RR 0.53 95% CI 0.13 to 2.18; 1 trial, 253 infants) or in women with a history of preterm delivery alone (RR 3.08 95% CI 0.13 to 74.46; 1 trial, 142 women).
The risk of low birthweight was reduced in the subgroup of high-risk women who received antibiotics compared to a placebo (RR 0.57; 95% CI 0.37 to 0.88; 1 trial, 253 women),
There was no difference between control and intervention groups on neonatal sepsis (RR 11.31; 95% CI 0.64 to 200.79; 1 trial, 142 infants)
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was graded as very low for most critical outcomes. Overall, the quality of evidence was graded as very low.
Balance of benefits and harms
There is insufficient evidence to recommend the routine use of antibiotics during the second or third trimester of pregnancy to reduce infectious morbidity or adverse outcomes. Evidence suggesting some clinical benefits was observed as a result of the inclusion of women with high-risk pregnancies, based on trials with limited study design and small sample sizes. In addition, this group was heterogenous, and it was unclear from the data which of the included conditions accounted for the observed reduction in postpartum endometritis. No data were available to evaluate the potential side-effects or impact of prolonged antibiotic treatment of the mother or the newborn, including the development of antimicrobial resistance. In the light of the available evidence, potential benefits related to the use of antibiotics during pregnancy to prevent infectious morbidities do not appear to outweigh potential harms, particularly for women who are not carrying a high-risk pregnancy.
Values and preferences
Health care providers and policy-makers in all settings are likely to place a high value on the potential public health impact of administering antibiotics to an unselected population of women during their second or third trimester in the absence of evidence of any clinical benefits. Mothers will prefer to avoid the inconvenience and side-effects of antibiotic use. The panel is confident that there is no variation in this value among health care providers, policy-makers and mothers in low-, middle- and high-income settings.
Resource implications
Implementation of this recommendation is likely to significantly reduce health care costs in settings where low-risk obstetric populations are routinely provided with antibiotics during pregnancy. In the long term, adherence to this recommendation will prevent significant health care costs that might be required to combat bacterial resistance in both obstetric and the general populations.
RECOMMENDATION 7
Routine antibiotic administration is not recommended for women in preterm labour with intact amniotic membranes.
(Strong recommendation based on moderate-quality evidence)
REMARKS
This recommendation is in keeping with the WHO guideline on interventions to improve preterm birth outcomes (
25).
The GDG placed its emphasis on the potential risk of harm to the baby (i.e. cerebral palsy) and less value on the minimal benefit to mothers; therefore, it recommended against the intervention.
It is critical for women with any diagnostic or clinical signs of infection to be treated accordingly with antibiotics.
Review question:
Among pregnant women in preterm labour with intact amniotic membranes (P), does routine antibiotic prophylaxis (I), compared with no routine antibiotic prophylaxis (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on the administration of prophylactic antibiotics to women in preterm labour with intact membranes for the prevention of maternal and neonatal infection was extracted from a Cochrane systematic review of 14 trials involving 7837 women (
26).
Trials were conducted in the 1990s in low-, middle- and high-income countries: six in the USA, and one each in Canada, Chile, Denmark, Germany, Iran, South Africa, the UK and Uruguay. All included trials used similar definitions of preterm labour, which included the presence of uterine contractions and cervical dilatation, with intact membranes. Trials included women with gestational ages between 20 and 36 weeks, with a mean of 30–32 weeks. However, the majority of the women gave birth at term. All trials excluded women with symptoms or signs suggestive of overt clinical infection of the mother or fetus. Four trials included women with multiple pregnancies.
Use of tocolytics or antenatal steroids was part of the clinical protocol for the majority of the included trials (13 and 12 trials, respectively). Four of the seven trials reporting on GBS colonization reported intrapartum antibiotic administration for women with a positive GBS culture, in addition to the study drug. Only one trial reported on long-term follow-up of children.
The intervention was oral or intravenous administration of antibiotics compared to either no treatment or a placebo. Antibiotic regimes varied between trials. Seven trials used only intravenous methods, three used only oral administration, and four combined the two sequentially. Ten trials used a combination of antibiotics. The duration of treatment varied between three and 10 days, with the majority of studies using courses of five to seven days.
Trials did not report on maternal severe infectious morbidity, maternal death or antibiotic resistance.
Any antibiotics versus no antibiotics (or placebo) (EB Table 6)
Women receiving antibiotics had significantly reduced rates of maternal infection, including chorioamnionitis/endometritis (RR 0.74, 95% CI 0.63 to 0.86, number needed to treat to benefit = 34, 95% CI 24 to 63; 10 trials, 7371 women). There was no significant difference between groups for adverse drug reaction requiring cessation of treatment (RR 1.32, 95% CI 0.92 to 1.89; 5 trials, 626 women).
No statistically significant difference was found in perinatal mortality (RR 1.22, 95% CI 0.88 to 1.69; 10 trials, 7304 women) or stillbirths (RR 0.73, 95% CI 0.43 to 1.26; 8 trials, 7080 infants).
There was a trend towards an increased risk of neonatal death among those receiving prophylactic antibiotics (RR 1.57, 95% CI 1.03 to 2.40; 9 trials, 7248 infants). There was no significant difference between groups on the need for mechanical ventilation (RR 1.02, 95% CI 0.84 to 1.24; 1 trial, 6241 infants); respiratory distress syndrome (RR 0.99, 95% CI 0.84 to 1.16; 9 trials, 7200 infants); neonatal positive blood culture (RR 1.01, 95% CI 0.69 to 1.49; 3 trials, 6526 infants); or neonatal sepsis (RR 0.86, 95% CI 0.64 to 1.16; 10 trials, 7386 infants) and other preterm-related morbidities such as intraventricular haemorrhage, major cerebral abnormality, necrotizing enterocolitis or chronic lung disease. No difference was found on admission to neonatal intensive care unit (NICU) or special care (RR 0.82, 95% CI 0.62 to 1.10; 5 trials, 6875 infants).
One trial reported on outcomes at seven years of age and found no difference in infant deaths (RR 1.06, 95% CI 0.68 to 1.67; 4654 children), any functional impairment (RR 1.10, 95% CI 0.99 to 1.23; 3052 children) or moderate to severe impairment (RR 1.07, 95% CI 0.89 to 1.28; 3052 children). There was a trend towards an increased risk of cerebral palsy in the treated group (RR 1.82, 95% CI 0.99 to 3.34; 3173 children).
Considerations related to the strength of the recommendation
Quality of evidence
The evidence was graded from moderate to high. Overall, the quality of the evidence was graded as moderate.
Balance of benefits and harms
The potential harms as shown in the review, including neonatal deaths and cerebral palsy, in association with the use of routine antibiotic prophylaxis outweigh the clinical benefits of antibiotics in terms of reducing maternal infectious morbidity.
Values and preferences
Health care providers, policy-makers and pregnant women and their families in all settings are likely to place a high value on preterm survival without long-term morbidity, and less value on clinical benefits in terms of reducing maternal infection. The panel is confident that there is no variation of this value among mothers, health care providers or policy-makers in any setting.
Resource implications
Implementation of this recommendation is likely to reduce health care costs where routine antibiotic prophylaxis for women with preterm labour and intact membranes is currently the norm.
RECOMMENDATION 8
Antibiotic administration is recommended for women with preterm prelabour rupture of membranes.
(Strong recommendation based on moderate-quality evidence)
REMARKS
This recommendation is in keeping with the WHO guideline on interventions to improve preterm birth outcomes (
25).
For near-term (i.e. ≥36 weeks) PPROM where the clinical policy of immediate or early labour induction (within 12 hours of rupture) is in place, antibiotic use does not confer any benefit and should not be used (see
Recommendation 9 in this guideline).
Erythromycin is recommended as the antibiotic of choice for prophylaxis in women with preterm prelabour rupture of membranes according to the
WHO recommendations on interventions to improve preterm birth outcomes (
25).
To avoid inadvertent antibiotic administration to women with intact amniotic membranes, antibiotics should not be prescribed unless a definite diagnosis of PPROM has been made. Therefore, a policy to prescribe antibiotics for women with PPROM should be accompanied by a protocol to reliably diagnose PPROM.
Long latent phase (interval between rupture of membranes and onset of preterm labour) could predispose to intrauterine infection. Therefore, women should be closely monitored for signs of clinical chorioamnionitis.
Review question:
Among pregnant women with preterm prelabour rupture of membranes (PPROM) (P), does routine antibiotic prophylaxis (I), compared with no routine antibiotic prophylaxis (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence for routine administration of prophylactic antibiotics to women with PPROM was extracted from a Cochrane systematic review of 22 trials involving 6872 women (
27).
The majority of the trials included in the review were conducted in high-income countries: 14 in the USA, and one each in Finland, Germany and Spain. One trial was conducted in Turkey, one in Zimbabwe and one multi-country trial in Chile and the USA. Results on short- and long-term outcomes were dominated by one trial conducted in the UK. Trials were conducted in the 1990s and early 2000s.
Women were recruited between 20 and 37 weeks of gestation. Clinical definitions and methods for diagnosis of PPROM varied between trials. The majority of the women were not in active labour.
Included trials compared different antibiotic regimens with placebo, other antibiotic class or no treatment, using different routes of administration (oral alone, intravenous alone or a combination).
Any prophylactic antibiotics versus placebo (all women and babies) (EB Table 7)
Sixteen trials compared any antibiotic with placebo and randomized 6300 women. These trials tested a broad spectrum of penicillins, beta-lactam, macrolide (erythromycin) and other antibiotics (clindamycin, gentamycin) either alone or in combination.
There was a statistically significant reduction in chorioamnionitis (RR 0.66, 95% CI 0.46 to 0.96; 11 trials, 1559 women).
There was no difference in perinatal deaths for all antibiotic comparisons (RR 0.93, 95% CI 0.76 to 1.14; 12 trials, 6301 infants), but there was a significant reduction in neonatal infections, including pneumonia, for all antibiotic comparisons (for any antibiotics (RR 0.67, 95% CI 0.52 to 0.85; 12 trials, 1680 infants), all penicillins, excluding co-amoxiclav (RR 0.81, 95% CI 0.68 to 0.98; five trials, 521 infants), and other antibiotics (RR 0.71, 95% CI 0.53 to 0.95; 3 trials, 763 infants)). There was a significant reduction in the number of positive neonatal blood culture (RR 0.79, 95% CI 0.63 to 0.99; 3 trials, 4961 infants).
There was a significant reduction in the number of infants receiving surfactant (RR 0.83, 95% CI 0.72 to 0.96; 1 trial, 4809 infants), of infants requiring oxygen therapy (RR 0.88, 95% CI 0.81 to 0.96, 1 trial, 4809 infants) and of infants with abnormal cerebral ultrasound scans before discharge (RR 0.81, 95% CI 0.68 to 0.98; 12 trials, 6289 infants) in the treated group compared with placebo. The duration of NICU admission was shorter in the treated group than in the placebo group (mean difference (MD) -5.05 days, 95% CI -9.77 to -0.33; 3 trials, 255 infants).
No differences were observed for neonatal respiratory distress syndrome (RR 0.95, 95% CI 0.83 to 1.09, 12 studies, 6287 infants), the number of babies requiring ventilation (RR 0.90, 95% CI 0.80 to 1.02, 2 studies, 4924 infants), neonatal oxygenation > 28 days (RR 0.79, 95% CI 0.61 to 1.03, 3 studies, 5487 infants) or necrotizing enterocolitis (RR 1.09, 95% CI 0.65 to 1.83, 6229 infants). However, the incidence of necrotizing enterocolitis appeared to be increased only with the use of beta-lactam antibiotics (including co-amoxiclav) (RR 4.72, 95% CI 1.57 to 14.23; 2 trials, 1880 infants).
Regarding long-term outcomes, one trial showed that antibiotics seemed to have little effect on serious childhood disability at seven years (RR 1.01, 95% CI 0.91 to 1.12; 3171 children).
Maternal deaths, serious maternal morbidities, puerperal sepsis, neonatal encephalopathy, major adverse drug reactions or antibiotic resistance were not reported.
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was graded from moderate to high for all outcomes. Overall, the quality of the evidence was graded as moderate.
Balance of benefits and harms
Compared with placebo, antibiotics for PPROM reduced the risk of chorioamnionitis in the mother. Antibiotics also reduced the risk of neonatal infections, including pneumonia, and cerebral abnormality, and were associated with a shorter stay in neonatal intensive care. On the other hand, antibiotics did not appear to have an impact on other infant mortality or severe morbidity or on longer-term outcomes. Overall, there are desirable short-term benefits for the mother and preterm infants without evidence of harms on short- or long- term.
Values and preferences
Health care providers, policy-makers and pregnant women and their families in all settings are likely to place a high value on the benefits of short-term outcomes for the mother and infant (reduction in maternal and neonatal infection). The panel is confident that there is no variation of this value among mothers, health care providers and policy-makers in any setting.
Resource implications
Antibiotics are widely available in both oral and parenteral forms in all settings. It is feasible to include prophylactic antibiotic therapy into existing health structures that are designed to manage women at risk of imminent preterm birth with minimal costs.
RECOMMENDATION 9
Routine antibiotic administration is not recommended for women with prelabour rupture of membranes at (or near) term.
(Strong recommendation based on low-quality evidence)
REMARKS
“Routine” use implies administration of antibiotics in the absence of clinical signs of infection or any additional risk factors for infection.
“Near term” in this context refers to 36 weeks gestation and above.
Evidence for this recommendation was based on studies that included women with duration of ruptured membranes less than 12 hours. The GDG noted that while the available evidence clearly indicates that antibiotics do not confer any benefits under a clinical policy of immediate or early induction (within 12 hours of rupture), it is less clear for a policy of expectant or delayed induction longer than this timeframe. Nevertheless, the generally low rate of maternal infection in the control population in the included studies (< 5%) further supports the restriction of antibiotic use to women with PROM and clinical evidence of infections.
The GDG noted that evidence is lacking on the potential benefits of antibiotic prophylaxis for women with prolonged rupture of membranes (> 18 hours) and active labour where the baseline risk of infection may be higher. As the risk of infection increases with the duration of labour, it is possible that women with prolonged labour and ruptured membranes may benefit from antibiotic prophylaxis, and this underlies the common clinical practice. The group acknowledges that in the light of current obstetric practice, it is unlikely that a randomized controlled trial will address the important question on the effect of antibiotic prophylaxis in prolonged prelabour rupture of membranes at term (> 12 hours) or prolonged labour with ruptured membranes.
The GDG put its emphasis on potential side-effects of antibiotics, particularly long-term effects among exposed children, as well as bacterial resistance and, therefore, made a strong recommendation.
Review question:
Among pregnant women with prelabour rupture of membranes at or near term (P), does routine antibiotic prophylaxis (I), compared with no routine antibiotic prophylaxis(C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on the use of prophylactic antibiotics for women with prelabour rupture of membranes (PROM) at 36 weeks gestation or beyond for preventing infectious morbidities was extracted from a Cochrane systematic review of four trials involving 2639 women (
28).
The trials were conducted in Chile, Egypt, Portugal and Spain.
Gestational age inclusion criteria varied slightly between trials (≥36 weeks in two trials and ≥37 weeks in two trials). All studies excluded women with multiple pregnancy and major obstetric complications. All studies used consistent criteria for diagnosis of membrane rupture and had protocols that attempted to minimize vaginal examinations. Pregnancy management protocols varied slightly between settings, mainly for induction policies.
The trials did not report on antibiotic side-effects and antibiotic resistance.
Any antibiotic versus placebo or no antibiotic (all women) (EB Table 8a)
Two trials compared antibiotics with placebo, and two other trials compared antibiotics versus no treatment. All trials compared different antibiotics and routes of administration. Two trials tested IV ampicillin with IV or IM gentamycin, one trial tested parenteral ampicillin/sulbactam, and one trial intravenous cefuroxime and clindamycin for 48 hours then oral cefuroxime and clindamycin for a further 24 hours.
There were no significant differences between groups for maternal infectious morbidities: suspected or proven chorioamnionitis (RR 0.65, 95% CI 0.34 to 1.26; 4 trials, 2639 women), endometritis (RR 0.34, 95% CI 0.05 to 2.31; 4 trials, 2639 women) or wound infection (RR 0.79, 95% CI 0.36 to 1.72; 3 trials, 1906 women). Data on postpartum pyrexia had high levels of heterogeneity (I2 = 93%) and were presented separately for two trials (RR 0.97, 95% CI 0.58 to 1.61; RR 0.11, 95% CI 0.01 to 0.88). There was no difference in reported maternal adverse effects (RR 2.93 95% CI 0.12 to 71.63; 4 trials, 2639 women). There were no cases of serious maternal outcome, postpartum septicaemia or maternal deaths reported in any of the trials.
There was no significant difference between groups in perinatal mortality (RR 1.98, 95% CI 0.60 to 6.55; 4 trials, 2639 infants), though two studies had no cases. Furthermore, no difference in stillbirth was shown when comparing antibiotics with placebo or no antibiotics (RR 3.00, 95% CI 0.61 to 14.82; 3 trials, 1906 infants). There were no cases of neonatal mortality in the three trials reporting this outcome (1906 infants).
There was no significant difference in probable early-onset neonatal sepsis (RR 0.69, 95% CI 0.21 to 2.33; 4 trials, 2639 babies) or definite early-onset neonatal sepsis (RR 0.57, 95% CI 0.08 to 4.2; 4 trials, 2639 babies). There were no significant differences for neonatal meningitis (RR 0.33, 95% CI 0.03 to 3.11; 4 trials, 2639 infants), neonatal pneumonia (RR 0.33, 95% CI 0.01 to 7.96; 4 trials, 2639 infants), admission to NICU (RR 1.23, 95% CI 0.82 to 1.85; 3 trials, 1906 infants) or length of hospitalization in NICU (MD 0.05 days, 95% CI -0.09 to 0.19; 1 trial, 1640 infants). There were no cases of respiratory distress syndrome in two studies.
Antibiotics versus no antibiotics: by timing of induction of labour (EB Table 8b)
Three trials involving 2478 women were included in the subgroup analysis of timing of induction of labour
There was no significant difference between comparison groups for early (< 12 hr) or late (≥ 12 hr) induction subgroups with respect to chorioamnionitis and/or endometritis (early induction: RR 1.15, 95% CI 0.64 to 2.08; 1 trial, 1640 women; late induction: RR 0.34, 95% CI 0.08 to 1.47; 2 trials, 838 women).
No significant differences were found between comparison groups in the subgroups for perinatal mortality (early induction: RR 3.00, 95% CI 0.61 to 14.82; 1 trial, 1640 infants; late induction: RR 0.98, 95% CI 0.14 to 6.89; 2 trials, 838 infants). Data on stillbirths were only available for the early induction subgroup, for which no statistically significant difference was found (RR 3.00, 95% CI 0.61 to 14.82; 1 trial, 1630 women).
There were no significant differences between comparison groups in the two subgroups with respect to neonatal mortality.
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was low for the majority of critical outcomes in the main comparison. For the subgroup analysis, the quality of evidence was moderate to low for critical outcomes reported for the early induction subgroup and very low for all late induction subgroups. The overall quality of evidence was based on the main comparison (i.e. all women) and, therefore, was graded as low.
Balance of benefits and harms
There is no convincing evidence of benefits for the mothers or neonates from the routine administration of antibiotics to women with PROM at or near term. Severe maternal and neonatal morbidity or deaths were infrequent in both the intervention and the control groups. Early or late induction of labour to expedite birth does not seem to affect maternal and neonatal infection outcomes. There were no data to assess the risk of short- and long-term harms of routine antibiotic use, particularly with respect to the development of antibiotic resistance. Although the evidence is limited to women with singleton pregnancies, there is no reason to suggest that the findings would be different for women with multiple pregnancies.
Values and preferences
Health care providers and policy-makers are likely to place a high value on the potential impact of antibiotic use on antibiotic resistance, and, in the absence of evidence of effectiveness, would chose not to use the intervention. The panel is confident that there is no variation in this value across settings.
Resource implications
Implementation of this recommendation is likely to reduce health care costs in settings where routine antibiotics prophylaxis for all cases of ruptured amniotic membranes is currently the norm.
RECOMMENDATION 10
Routine antibiotic administration is not recommended for women with meconium-stained amniotic fluid.
(Conditional recommendation based on low-quality evidence)
REMARKS
In the absence of convincing evidence, the GDG puts its emphasis on the public health impact of routine administration of antibiotics (in terms of increasing antibiotic resistance) for a relatively common condition in labour and decided to recommend against the intervention.
Antibiotics should be used in a situation where the passage of meconium by the fetus may be triggered by antepartum or intrapartum infectious morbidity – e.g. chorioamnionitis — or when the characteristics of the liquor suggest intrapartum infection.
It is important that a personnel experienced in neonatal resuscitation attends the delivery of all infants in whom thick meconium liquor is noted, as the risk of meconium aspiration syndrome is higher in this situation.
Review question:
Among pregnant women with meconium-stained amniotic fluid during labour (P), does routine administration of antibiotics (I), compared with no routine antibiotics (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence for the routine use of prophylactic antibiotics among women presenting with meconium-stained amniotic fluid during labour was extracted from a Cochrane systematic review of two trials involving 362 women (
29). The two trials were led by the same investigator. One trial was reported only as a conference abstract with little methodological detail.
Both trials were conducted in the USA. The trials excluded women with evidence of active infection or allergy to penicillin and/or cephalosporin.
The two trials compared 3 g of intravenous ampicillin-sulbactam (one trial repeated every six hours until delivery) with intravenous normal saline as placebo.
Antibiotics versus placebo or no treatment (EB Table 9)
The incidence of chorioamnionitis was significantly reduced in the treated group compared with placebo (RR 0.36, 95% CI 0.21 to 0.62; 2 trials, 362 women), but no difference was observed in the incidence of postpartum endometritis (RR 0.5, 95% CI 0.18 to 1.38; 1 trial, 120 women).
No difference was found in the incidence of neonatal sepsis (RR 1.00, 95% CI 0.21 to 4.76; 1 trial, 120 infants) or NICU admission (RR 0.83, 95% CI 0.18 to 1.38; 1 trial, 120 infants).
No serious adverse effects were reported.
The trials did not report on maternal severe infectious morbidities, maternal or neonatal mortality, or antibiotic resistance.
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was graded as low for all outcomes except chorioamnionitis, which was moderate in quality. Overall, the quality of evidence was graded as low.
Balance of benefits and harms
Available evidence from trials included in the review is insufficient to support the use of prophylactic antibiotics in preventing maternal and neonatal infectious morbidities among women presenting with meconium-stained amniotic fluid during labour. Antibiotic appears to reduce the risk of chorioamnionitis but has no effect on neonatal outcomes. Data on postpartum endometritis was too small to support or refute the findings regarding chorioamnionitis. Overall, the evidence should be interpreted with caution due to its low quality and the small sample size. In the absence of clear evidence of clinical benefits, the potential risks to public health of routinely administering antibiotics for a relatively common condition during labour outweigh the potential benefits.
Values and preferences
Health care providers and policy-makers are likely to place a high value on the potential impact of antibiotic use on emerging antibiotic resistance, and, in the absence of convincing evidence of effectiveness, would chose not to use the intervention. The panel is confident that there is no variation in this value across settings.
Resource implications
Implementation of this recommendation is likely to reduce health care costs in settings where routine antibiotic prophylaxis for women with meconium-stained liquor during labour is currently the policy.
RECOMMENDATION 11
Routine antibiotic prophylaxis is recommended for women undergoing manual removal of the placenta.
(Strong recommendation based on very low-quality evidence)
REMARKS
Although there is no clear indication of benefits from the available evidence, the GDG decided to recommend prophylactic antibiotic use for this condition based on consensus after considering the potentially higher risk of infection related to the invasive nature of intrauterine manipulation required for manual placental removal. The group also considered indirect evidence of the benefit of prophylactic antibiotics from studies of caesarean section and abortion, as well as observational studies of other intrauterine manipulations.
This recommendation is based on updated evidence and is consistent with existing WHO guidance on the treatment of postpartum haemorrhage which recommends a single dose of antibiotics (ampicillin or first-generation cephalosporin) for manual placental removal (
30).
In addition to antibiotic use, health care providers should take into account other factors that could decrease the risk of infection, such as observing good hygiene and general aseptic technique during the procedure and prevention or treatment of anaemia in the woman.
This question was considered a research priority for settings in which prophylactic antibiotics are not routinely administered and those where the baseline risk of infectious morbidity is low. However, the GDG acknowledged that conducting a randomized trial may be challenging given the current clinical practice.
Review question:
Among women undergoing manual removal of retained placenta following vaginal birth (P), does antibiotic prophylaxis (I), compared with no antibiotic prophylaxis (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
An updated Cochrane review that evaluated this question did not find any eligible randomized controlled trial to include (
31). Evidence was extracted from a systematic review of non-randomized studies which included three retrospective cohort studies of 567 women (
32). This review considered only women undergoing vaginal deliveries and excluded women with prior history of fever.
The studies were conducted in Germany, Norway and Bulgaria, and all compared outcomes among women receiving antibiotic prophylaxis with no intervention.
Only two critical outcomes (endometritis and puerperal fever > 37.5 °C, > 24 hours) were reported.
Antibiotic prophylaxis versus no treatment (EB Table 10)
Compared with no antibiotic prophylaxis, antibiotic prophylaxis was not associated with significant differences in the number of women with puerperal fever [odds ratio (OR) 0.93, 95% CI 0.38 to 2.27; 1 study, 302 women] or endometritis (OR 0.84 95% CI 0.38 to 1.85; 3 studies, 567 women).
Considerations related to the strength of the recommendation
Quality of evidence
The overall quality of evidence was graded as very low.
Balance of benefits and harms revise wording
There is no evidence from randomized controlled trials to determine the effect of antibiotic use prior to manual placental removal on infectious morbidities. Low-quality evidence from observational studies in high-income countries also shows inconclusive evidence of benefits, although the odds of endometritis and puerperal fever were reduced when antibiotics were used. There was no information about potential harms of the intervention. The effect of routine use of antibiotics on antibiotic resistance is likely to be insignificant considering the very low incidence of manual removal of the placenta in women giving birth.
Values and preferences
Health care providers and policy-makers in low-income settings (where the baseline risk of infectious morbidity is high) are likely to place a high value on the potential clinical benefits of antibiotic use for an intrauterine manoeuvre and would chose to adhere to the recommendation. The panel acknowledged that there might be variation in this value in high-income settings where infection control standard is high.
Resource implications
Implementation of this recommendation is likely to require additional costs for antibiotics. However, these added costs are justified by cost savings for treatment of severe infectious morbidities that could result from an invasive intrauterine manipulation.
RECOMMENDATION 12
Routine antibiotic prophylaxis is not recommended for women undergoing operative vaginal birth.
(Conditional recommendation based on very low-quality evidence)
REMARKS
“Operative vaginal birth” is the term used to describe delivery of the fetal head assisted by either vacuum extractor or forceps.
Prophylactic antibiotics may be useful for other maternal conditions that could result from prolonged second stage of labour or the use of an instrument for vaginal birth (e.g. third- or fourth-degree perineal tear).
Review question:
Among women undergoing operative vaginal birth (P), does routine antibiotic prophylaxis (I), compared with no prophylaxis (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence regarding the routine administration of prophylactic antibiotics to women undergoing operative vaginal birth (vacuum or forceps) was extracted from a Cochrane systematic review (
33). Only one trial with a sample size of 393 women reported on critical outcomes. The trial was conducted in the USA. Women with evidence of other inflammatory infections or allergies to penicillin class of drugs were excluded.
The evidence was supplemented with another systematic review of non-randomized studies which included three retrospective cohort studies of 1293 women (
34). Two of the studies were performed in Germany, and one in France.
Antibiotic prophylaxis versus no treatment (randomized controlled trials) (EB Table 11a)
The trial investigated the use of 2 g of cefotetan intravenously after umbilical cord clamping versus no treatment in women undergoing instrumental deliveries.
The trial found no differences between the treatment and control groups regarding the incidence of endomyometritis (RR 0.07; 95% CI 0.00 to 1.21) or length of maternal hospital stay (MD 0.09 days; 95% CI -0.23 to 0.41).
The included trial did not report any other critical outcomes.
Antibiotic prophylaxis versus no treatment (non-randomized studies) (EB Table 11b)
Prophylactic antibiotic regimes varied between studies. One study used 4 g ampicillin or 4 g cephalexin or cefalotin, administered for at least five days. One study used 1 g of clamoxyl administered intravenously during an intrauterine procedure or during umbilical cord clamping and repeated two and six hours later plus 0.5 g of ornidazole. One study used mebacid sulfamerazine or 2 g chloramphenicol daily for six to 10 days.
There were no differences between the treated and untreated group in the incidence of endometritis (OR 0.67; 95% CI 0.07 to 6.04; 2 studies, 1091 women), maternal septicaemia (OR 0.32; 95% CI 0.01 to 7.90, 1 study, 336 women) or wound infection (episiotomy abscess) (OR 0.35; 95% CI 0.06 to 2.06; 2 studies, 540 women).
Considerations related to the strength of the recommendation
Quality of the evidence
Overall, the quality of evidence from both sources was graded as very low.
Balance of benefits and harms
There is low-quality evidence to suggest that antibiotic prophylaxis does not reduce the risk of maternal infections after operative vaginal birth. Neonatal outcomes were not reported in the studies included in the reviews.
Values and preferences
Health care providers and policy-makers are likely to place a high value on the potential impact of antibiotic use on antibiotic resistance, and, in the absence of evidence of effectiveness, would chose not to use the intervention. The panel is confident that there is no variation in this value across settings.
Resource implications
Implementation of this recommendation is likely to reduce additional costs for antibiotics where routine use of antibiotics for all operative births is an established practice.
RECOMMENDATION 13
Routine antibiotic prophylaxis is recommended for women with third- or fourth-degree perineal tear.
(Strong recommendation based on very low-quality evidence)
REMARKS
Despite the insufficient evidence of benefits, the GDG agreed that women with third- or fourth-degree perineal tear are at higher risk of infection in the postpartum period and took a consensus view to recommend the routine use of antibiotic prophylaxis for these conditions. The group puts its emphasis on the reduction in wound infection which might aggravate long-term consequences of third- or fourth-degree perineal tears (e.g. involuntary loss of flatus and/or faeces which affects quality of life) and, therefore, made a strong recommendation.
This recommendation is consistent with the WHO postnatal care guideline on treatment of third- or fourth-degree perineal tears (
35).
The GDG acknowledged that antibiotic administration following third- or fourth-degree tears is already a common clinical practice and, therefore, did not consider the question a research priority.
Review question:
Among women with third- or fourth-degree perineal tear after birth (P), does routine antibiotic prophylaxis (I), compared with no antibiotic prophylaxis (C), prevent maternal infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on the effectiveness and safety of routine administration of prophylactic antibiotics post-delivery to women with third- or fourth-degree perineal tear was extracted from a Cochrane systematic review that included only one trial with 147 women (
36).
The trial was conducted in the USA. Women who were less than 18 years of age, GBS- or HIV-positive, had chorioamnionitis, history of inflammatory bowel disease or were already on antibiotics at the moment of inclusion were excluded from the study. The trial planned to recruit 310 women but was terminated early because of its inability to achieve the sample size within a reasonable time period. The rate of loss to follow-up at two and six weeks following discharge among women who participated in the trial was 27.2%.
Women in the treatment arm received a single dose of second-generation cephalosporin (cefotetan, cefoxitin or penicillin, or clindamycin if allergic to penicillin), while those in the control arm received placebo. In both groups, wound disruption or purulent discharge of the perineum after repair was assessed at two and six weeks postpartum.
Antibiotic prophylaxis versus placebo (EB Table 12)
The trial showed a significant reduction in the number of women who had wound infection two weeks after delivery (RR 0.34, 95% CI 0.12 to 0.96; 107 women). However, there was no difference between the groups at six weeks after delivery (RR 0.38, 95% CI 0.13 to 1.09; 128 women).
The trial did not report on severe maternal infectious morbidity, puerperal sepsis, local discomfort, sexual dysfunction, duration of hospital stay, side-effects of antibiotics or antibiotic resistance.
Considerations related to the strength of the recommendation
Quality of evidence
Overall, the quality of evidence was graded as very low.
Balance of benefits and harms
The available evidence is insufficient to conclude on the clinical benefits of routine administration of prophylactic antibiotics in women with third- or fourth-degree perineal tear postpartum. The small sample size and high dropout rate limits the quality of the evidence. However, there is indirect evidence of benefit for prophylactic antibiotics from potentially contaminated wounds (considering the bacteria flora in the rectum) in surgical practice, and it would be reasonable to use antibiotics to reduce the risk of infection.
Values and preferences
Health care providers and policy-makers are likely to place a high value on the potential impact of antibiotic use on preventing wound complications and long-term consequences of poorly healed severe perineal tears. The panel is confident that there is no variation in this value across settings.
Resource implications
Implementation of this recommendation is likely to minimally increase health care costs (for antibiotics) where routine antibiotic use is not the current practice. However, these added costs are justified by cost savings for treatment of long-term morbidities (e.g. poor quality of life) of obstetric anal sphicteric injuries when they are complicated by perineal infections.
RECOMMENDATION 14
Routine antibiotic prophylaxis is not recommended for women with episiotomy.
(Strong recommendation based on consensus view)
REMARKS
The above recommendation was based on a consensus of the GDG in view of a high rate of episiotomy and the potential impact of antibiotics, in the absence of clinical benefits on public health. The GDG puts its emphasis on avoidance of emerging antimicrobial resistance at the global level and, therefore, made a strong recommendation.
This recommendation applies to the use of antibiotics before or immediately after episiotomy repair following vaginal birth. Antibiotics should be administered when there are clinical signs of infection of an episiotomy wound.
The GDG emphasized the need for health systems to adopt a policy of restrictive rather than routine use of episiotomy to reduce its potential complications and the use of additional resources for its treatment.
Second-degree perineal tear is anatomically similar to an episiotomy and does not warrant the use of prophylactic antibiotics.
In a situation where an episiotomy wound extends to become a third- or fourth-degree perineal tear, prophylactic antibiotics should be administered as recommended in this guideline (see
Recommendation 13).
Review question:
Among women who had an episiotomy for vaginal birth (P), does routine antibiotic prophylaxis (I), compared with no routine antibiotic prophylaxis (C), prevent maternal infectious morbidities and improve outcomes (O)?
Summary of evidence
A systematic review was conducted to evaluate the effectiveness of antibiotic prophylaxis on infectious morbidity following episiotomy in women giving vaginal birth. Based on a pre-specified protocol, a detailed search was conducted in MEDLINE, EMBASE, CENTRAL and the CINAHL databases for randomized and non-randomized studies that addressed these questions. Of the 831 citations generated by these search strategies, 38 full-text articles were retrieved for futher assessment. None of these studies met the inclusion criteria for this review (
37).
Considerations related to the strength of the recommendation
Quality of evidence
There is no direct evidence on the impact of antibiotics on infectious morbidity in women with episiotomy.
Balance of benefits and harms
There is a complete lack of evidence from randomized trials and observational studies to determine the benefit or harm of routine administration of antibiotics to women who receive an episiotomy for vaginal birth. Carefully performed episiotomies generally have a low rate of infection in settings where infection control measures are well observed. The relatively high global episiotomy rates (> 50%) means that many mothers will be exposed to antibiotics without clear evidence of benefit but a huge impact on public health in terms of emerging antimicrobial resistance.
Values and preferences
Health care providers and policy-makers in all settings are likely to place a high value on the potential public health impact of administering antibiotics to large proportion of women giving birth, in the absence of evidence of any clinical benefits. Mothers will prefer to avoid the inconvenience and side-effects of antibiotic use. The panel is confident that there is no variation in this value among health care providers, policy-makers and mothers in low-, middle- and high-income settings.
Resource implications
The implementation of this recommendation is likely to save considerable health system costs in view of the high rate of episiotomy worldwide and the widespread use of prophylactic antibiotics in women with episiotomy.
RECOMMENDATION 15
Routine antibiotic prophylaxis is not recommended for women with uncomplicated vaginal birth.
(Strong recommendation based on very low-quality evidence)
REMARKS
The GDG was concerned about the potential public health implication of the high rate of routine use of antibiotics following vaginal birth without any specific risk factors in some settings. The group puts its emphasis on the negative impact of such policy on the global efforts to contain antimicrobial resistance and, therefore, made a strong recommendation against routine antibiotic prophylaxis.
“Uncomplicated vaginal birth” in this context connotes vaginal birth in the absence of any specific risk factor for or clinical signs of maternal peripartum infection.
Careful monitoring of all women after birth is essential to promptly identify any sign of endometritis and institute appropriate antibiotic treatment (see
Recommendation 20).
Recommendations on antibiotic use for common intrapartum conditions or interventions that often raise concerns about increased risk of infection are available in this guideline.
Review question:
Among pregnant women with uncomplicated vaginal birth (P), does antibiotic prophylaxis after birth (I), compared with no prophylaxis or placebo (C) prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence on the impact of antibiotic prophylaxis in women with uncomplicated (“normal”) vaginal birth was extracted from a systematic review which identified two eligible randomized controlled trials involving 1653 women (
38). The two trials compared antibiotic prophylaxis with no prophylaxis. The trials were conducted in France and Japan.
One of the trials described women with “uncomplicated vaginal birth” as those who had vaginal delivery, no fever (> 38 °C) during labour or the hour following delivery, an interval of < 24 hours between rupture of membranes and labour onset, no evidence of extragenital infection (e.g. urinary tract infection) and no known allergy to Amox-CA or betalactam. The study excluded women with evidence of amniotic fluid infection at the time of admission. The second trial excluded women with a history of hypersensitivity to the tested antibiotics (cefteram or cephem), fourth-degree perineal lacerations, birth after PROM at term, underlying medical conditions such as gestational hypertension and diabetes mellitus and at the discretion of the physician.
One trial used a single dose of Amox-CA 1 g intravenously, while the other trial used oral 300 mg cefteram pivotal for three or five days.
In terms of outcomes, one of the trials used clinical and/or laboratory criteria for diagnosing endometritis: pyrexia > 38 °C confirmed on two separate occasions and accompanied by pain on mobilizing the uterus and/or fetid lochia, and/or leucocytosis of more than 10 000/mm3. The other trial used only clinical criteria that included the occurrence of “fever more than 37 °C for more than two days, or infected lochia, or low abdominal pain detected and diagnosed by the doctor in charge, after 24 hours from birth”.
Antibiotic prophylaxis versus no prophylaxis/placebo (EB Table 13)
Women receiving antibiotic prophylaxis after uncomplicated vaginal birth experienced significantly reduced incidence of endometritis (RR 0.26, 95% CI 0.09 to 0.73; 2 trials, 1653 women). However, no statistically significant difference was observed in the risks of puerperal fever (RR 0.26, 95% CI 0.02 to 3.97, 2 trials, 1653 women), wound infection (RR 0.80, 95% CI 0.07 to 8.68; 1 trial, 362 women), urinary tract infection (RR 0.51, 95% CI 0.18 to 1.45; 1 trial, 1291 women) and duration of hospital stay (MD -0.15 days, 95% CI -0.31 to 0.01; 1 trial, 1291 women).
All other outcomes reported in the review were not prespecified as critical outcomes for this recommendation question.
Considerations related to the strength of evidence
Quality of evidence
The quality of evidence was graded as very low for four out of the five critical outcomes reported. Overall the quality of evidence was graded as very low.
Balance of benefits and harms
There is very low-quality evidence of clinical benefit in terms of reduction in postpartum endometritis in women who received antibiotics following uncomplicated vaginal birth. No clinical benefits were observed for other critical outcomes. The studies contributing data to the endometritis outcome were at high risk of bias because of lack of blinding, given that the diagnosis of endometritis in the studies was in part subjective.
Additionally, fever, a more objective measure, which was also included as part of the diagnosis of endometritis, was not different between intervention and control arms of the trial. The incidence of postpartum endometritis in the control population of the studies was 2.2%, suggesting that only a small proportion of women were at risk of endometritis. The number needed to treat to benefit (i.e. to avoid one case of endometritis) is 58. In view of the very low rate of endometritis and the fact that endometritis in itself is more of an early sign of severe pelvic infection when left untreated, unnecessary exposure of about 98% of women who are unlikely to develop this condition will negatively impact on public health in terms of increasing antimicrobial resistance.
Values and preferences
Health care providers and policy-makers in all settings are likely to place a high value on the potential negative public health impact of administering antibiotics to a very large proportion of women giving birth who are unlikely to develop peripartum infection. Mothers will also prefer to avoid the inconvenience and side-effects of antibiotic use. The panel is confident that there is no variation in this value among health care providers, policy-makers and mothers in low-, middle- and high-income settings.
Resource implications
The implementation of this recommendation is likely to save health care costs in settings were antibiotics are routinely given to women with uncomplicated vaginal birth. Additionally, adhrence to this recommendation could potentially contribute to signficant reduction in health care costs related to combating antimicrobial resistance in the larger population.
RECOMMENDATION 16
Vaginal cleansing with povidone-iodine immediately before caesarean section is recommended.
(Conditional recommendation based on moderate-quality evidence)
REMARKS
The recommendation of the use of povidone-iodine out of the common antiseptics was because it was the only agent tested in all randomized controlled trials that evaluated the review question.
The GDG noted that the main clinical benefit (reduction in post-caesarean endometritis) demonstrated in the review was largely driven by women at higher baseline risk of infections (i.e. those who were already in labour and those with ruptured membranes). However, in consideration of the similarity in the statistical findings between subgroups and the entire study population, the group acknowledged that women at lower baseline risk of infection are also likely to benefit from the intervention.
Due to the staining of surrounding tissues, vaginal cleasing in this context may be regarded as a potentially invasive procedure, and implementation might not be easy.
The GDG considers further evaluation of the benefits in high-risk women and potential adverse effects (especially among women with ruptured membranes and those planning to breastfeed) a research priority. Additionally, the group considers it essential to identify the most appropriate timing of the intervention to achieve benefit with minimal harm and whether other antiseptic agents (e.g. chlorhexidine) have similar beneficial effects. The group noted that shorter application and contact time are likely to be associated with less maternal and fetal exposure. Therefore, the group suggested vaginal application of povidone-iodine very close to the start of caesarean section (e.g. following bladder catheterization) to mimimize the discomfort to the woman. The specified duration of vaginal cleansing with povidone-iodine in three of the seven included studies in the Cochrane review was 30 seconds.
The use of a high concentration and/or repeated applications of povidone-iodine should be avoided to minimize maternal and fetal exposure and possible interference with the results of neonatal thyroid screening.
Review question:
Among pregnant women with indications for caesarean section (P), does vaginal cleansing with an antiseptic agent prior to caesarean delivery (I), compared with no vaginal cleansing with an antiseptic agent (C), prevent post-operative maternal infectious morbidities (O)?
Summary of evidence
Evidence on vaginal preparation with antiseptic agent before caesarean section for preventing postoperative infectious morbidities was extracted from a Cochrane systematic review of seven trials involving 2816 women (2635 analysed) (
39).
Trials were conducted in low-, middle- and high-income countries: one trial each in Iran, Pakistan and Turkey, and four trials in the USA.
All seven trials compared preoperative povidone-iodine solution preparation with a control group. Where specified, the concentration of povidone-iodine applied ranged from 1% to 10%. Control group was no vaginal cleansing in six trials and saline vaginal wash in one trial.
Most trials included women undergoing either a scheduled or emergency caesarean delivery. Two trials excluded women with chorioamnionitis. Prophylactic antibiotics were used in all trials.
None of the trials reported severe maternal infectious morbidity, side-effects of antiseptic agent or the cost of care.
Vaginal preparation with antiseptic agent versus control: all women (EB Table 14a)
Women receiving vaginal cleansing with povidone-iodine experienced significantly reduced risk of post-caesarean endometritis (RR 0.45, 95% CI 0.25 to 0.81; 7 trials, 2635 women).
There was no significant difference between the comparison groups for postoperative fever (RR 0.90, 95% CI 0.74 to 1.10; 6 trials, 2475 women), wound infection (RR 0.86, 95% CI 0.54 to 1.36; 6 trials, 2205 women) or any wound complication (RR 0.63, 95% CI 0.37 to 1.07; 2 trials, 729 women).
Vaginal preparation with antiseptic agent versus control: by presence or absence of labour before caesarean section (EB Table 14b)
Four trials stratified data for women according to whether they were in labour or not before the caesarean section.
Women in labour who received vaginal preparation with povidone-iodine solution preoperatively had lower risk of endometritis (RR 0.56, 95% CI 0.34 to 0.95; 3 trials, 523 women), but there were no differences observed for postoperative fever (RR 0.68, 95% CI 0.42 to 1.08; 2 trials, 307 women) or wound infection (RR 0.72, 95% CI 0.24 to 2.21, 2 trials, 307 women).
Among women who were not in labour, no significant differences were observed between the intervention and the control groups for post-caesarean endometritis (RR 0.89, 95% CI 0.52 to 1.54; 3 trials, 871 women), postoperative fever (RR 0.96, 95% CI 0.61 to 1.49; 2 trials, 658 women) or wound infection (RR 0.64, 95% CI 0.27 to 1.56; 2 trials, 652 women).
However, the test for subgroup differences did not show evidence of any differences between the subgroups.
Vaginal preparation with antiseptic agent versus control: by status of amniotic membranes (EB Table 14c)
Four trials stratified data for women according to the status of amniotic membranes.
Women with ruptured membranes who received vaginal preparation with povidone-iodine solution had lower risk of post-caesarean endometritis (RR 0.24, 95% CI 0.10 to 0.55; 3 trials, 272 women), but there were no differences observed for postoperative fever (RR 0.62, 95% CI 0.34 to 1.12; 2 trials, 200 women) or wound infection (RR 1.22, 95% CI 0.46 to 3.20, 3 trials, 272 women).
Among women with intact membranes, no significant differences were observed between the intervention and the control groups for post-caesarean endometritis (RR 0.63, 95% CI 0.36 to 1.06; 3 trials, 857 women), postoperative fever (RR 0.93, 95% CI 0.63 to 1.36; 2 trials, 769 women) or wound infection (RR 0.72, 95% CI 0.35 to 1.52; 3 trials, 857 women).
There was evidence of subgroup differences only for post-caesarean endometritis.
Considerations related to the strength of the recommendation
Quality of the evidence
The quality of the evidence for critical outcomes was graded as moderate or high. Overall, the quality of the evidence was graded as moderate.
Balance of benefits and harms
Vaginal preparation with povidone-iodine solution immediately prior to caesarean birth has the potential to reduce postoperative endometritis, particularly in women with ruptured membranes or those who are already in labour. This benefit was demonstrated in the context of prophylactic antibiotic cover. No adverse effects were reported by any of the trials. It is likely that the clinical benefits in terms of reduced endometritis observed in the overall analysis of the review apply to both emergency and scheduled or planned caesarean sections. However, the noted reduction in post-caesarean endometritis does not appear to influence postpartum maternal febrile morbidity, which was not significantly different between women who received and those who did not receive vaginal antiseptic before caesarean birth.
The balance between harm and benefits is limited by the lack of reporting on the maternal and infant side-effects of vaginal antiseptic application, although such side-effects are known to be rare. For neonates, a few studies have reported increased iodine levels in the first few days after birth but without any important clinical consequences. Such changes in iodine levels can affect the interpretation of results of neonatal thyroid screening in settings where this is routinely performed.
The findings of this Cochrane review are limited to the use of povidone-iodine preparations, and the effect of other popular antiseptics such as chlorhexidine as a vaginal disinfectant prior to caesarean birth is unknown.
Values and preferences
Health care providers, policy-makers and pregnant women and their families are likely to place a high value on the benefits of this intervention in terms of reduction in post-caesarean endometritis and a low value on the possible inconvenience associated with vaginal cleansing. This value is unlikely to substantially vary regardless of clinical situations and baseline risk of ascending infection (e.g. intact versus ruptured membranes, elective versus emergency caesarean section).
Resource implications
Implementation of this recommendation is likely to slightly increase costs where it is not currently in practice. However, the low costs of povidone-iodine, ready availability in all settings and low resources in terms of staff time or skill needed to implement this recommendation suggest that this intervention is likely to be cost–effective.
RECOMMENDATION 17
The choice of an antiseptic agent and its method of application for skin preparation prior to caesarean section should be based primarily on the clinician's experience with that particular antiseptic agent and method of application, its cost and local availability.
(Conditional recommendation based on low-quality evidence)
REMARKS
Skin preparation is a vital part of the overall care that must be given to women undergoing surgery, to prevent surgical site infections before caesarean section. However, there is no strong evidence to recommend the use of one specific antiseptic agent over another.
Maternal allergy to the preparation must be excluded prior to surgery.
A standard preoperative skin preparation technique that is appropriate for the intended skin incision must be followed.
Review questions:
Among pregnant women undergoing caesarean delivery (P), is the use of a particular antiseptic agent for preoperative skin preparation (I), compared with other antiseptic agent(s) (C), more effective in preventing post-caesarean infectious morbidities (O)?
Among pregnant women undergoing caesarean delivery (P), is the use of a particular method of antiseptic application for preoperative skin preparation (I), compared with other methods of antiseptic application (C), more effective in preventing post-caesarean infectious morbidities (O)?
Summary of evidence
Evidence for the comparative effectiveness of different methods of application of antiseptic agents (e.g. scrub, paint, drape) was extracted from a Cochrane systematic review of six trials including 1522 women (
40). For the comparison of the use of drape versus no drape (where one trial used iodine and other used chlorhexidine), two trials conducted in Denmark and South Africa included 1294 women undergoing elective or emergency caesarean section.
Evidence on the comparative effectiveness of different antiseptic agents (e.g. alcohol, povidone-iodine) for skin preparation prior to caesarean section was extracted from the same Cochrane systematic review with data from four small trials (194 women) (
40). Three of these trials were conducted in the USA, and one in France. Trials tested different forms and concentrations of antiseptics agents. For each comparison only one trial of small sample size contributed data.
The included trials did not report on many critical outcomes: severe maternal infectious morbidity, maternal death, side-effects, maternal satisfaction, neonatal infection or severe neonatal morbidity.
Comparison of different antiseptic preparations (EB 15a–15c)
Alcohol scrub plus iodophor drape versus iodophor scrub (1 trial, 79 women)
One trial compared a one-minute scrub with 70% isoprophyl alcohol followed by application of iodophor-impregnated adhesive film in the experimental group, with a five-minute iodophor scrub followed by application of iodophor solution in the control group.
No significant difference between groups was reported in the incidence of endometritis (RR 1.62, 95% CI 0.29 to 9.16).
The trial reported no wound infection in either group.
Chlorhexidine 0.5% versus 70% alcohol plus iodophor drape (IOBAN 2) (1 trial, 22 women)
This trial reported only on neonatal outcomes and did not contribute any data to any of the comparisons included in the systematic review. Cord blood iodine concentration was significantly higher in the iodine group (18.38 ± 20.34 versus 6.44 ± 0.66 μg/100 ml, P < 0.05) than in the alcohol plus iodophor drape group. There was no significant difference in neonatal 48-hour urine iodine excretion and thyroid-stimulating hormone levels at five days.
Parachlorometaxylenol plus iodine versus iodine alone (1 trial, 50 women)
One trial compared a five-minute scrub with parachlorometaxylenol followed by povidone-iodine scrub and normal saline irrigation of the pelvis and subcutaneous tissue at uterine closure and fascial closure with povidone-iodine surgical scrub (7.5%) followed by povidone-iodine (10%) and normal saline irrigation of the pelvis and subcutaneous tissue at uterine and fascial closure.
There was no significant difference between groups in the incidence of endometritis (RR 0.88, 95% CI 0.56 to 1.38) or wound infection (RR 0.33, 95% CI 0.04 to 2.99).
Chlorhexidine gluconate versus povidone-iodine (1 trial, 60 women)
There was no significant difference between groups for wound infection at two weeks after birth (RR 2.10, 95% CI 0.20 to 21.42), although the chlorhexidine gluconate group had significantly reduced bacterial growth at 18 hours after caesarean section (RR 0.23, 95% CI 0.07 to 0.70).
Methods of application: drape versus no drape (EB Tables 15d–15e)
Two trials compared drape with no drape using different methods for preoperative skin disinfection. Incisional plastic drape was applied to the skin after preoperative skin disinfection. Preoperative skin disinsfection was performed with 0.5% chlorhexidine in 80% alcohol solution for 30 seconds in one trial, and 2.5% iodine in 70% etanol in the other trial. Rates of antibiotics prophylaxis were similar between intervention (10.7%) and control arms (8.2%) in one trial.
The comparison showed no significant differences between groups for wound infection (RR 1.29, 95% CI 0.97 to 1.71; 2 trials, 1294 women) or reduction of skin colony counts (MD 0.07 colony forming unit per plate; 95% CI -0.34 to 0.48, 1 trial, 79 women).
There was no significant difference between groups for length of hospital stay (MD 0.10 days, 95% CI -0.27 to 0.46, 1 trial, 603 women).
Considerations related to the strength of the recommendation
Quality of the evidence
The quality of evidence was graded as low for all critical outcomes for comparisons between antiseptics and for methods of application. Overall the quality of evidence was graded as low.
Balance of benefits and harms
There is insufficient evidence to demonstrate that one antiseptic agent or one method of application is better than the other for skin preparation for caesarean section. Available trials involved small number of participants and were mostly underpowered to detect statistical differences between comparison groups. Although one study showed considerable reduction in skin bacteria colony counts when chlorhexidine was used compared to povidone-iodine, the finding did not translate to reduced risk of wound infection.
Values and preferences
The choice of antiseptic preparation and the method of application of antiseptic preparation are likely to be guided by the clinician's experience, locally available options and locally common bacterial skin flora and antimicrobial sensitivity. The preferences of health care providers, policy-makers and pregnant women are likely to vary considerably across settings.
Resource implications
Antiseptic agents that are used for skin preparation are relatively cheap and available in all settings with the capacity to perform caesarean section. The implementation of this recommendation does not require a specific change in existing practice and is, therefore, unlikely to change health care costs in settings where the same preparation continues to be used. However, health care costs might be reduced in settings where the health systems choose to use a cheaper preparation that meets the criteria specified in the recommendation.
RECOMMENDATION 18.0
Routine antibiotic prophylaxis is recommended for women undergoing elective or emergency caesarean section.
(Strong recommendation based on moderate-quality evidence)
REMARKS
Antibiotic prophylaxis in this context refers to antibiotic use prior to the initiation of/or during caesarean section in the absence of clinical signs of infection. The GDG noted that it is essential for clinicians to be clear about this description to avoid using antibiotic regimens that are most applicable for treating confirmed infection - i.e. therapeutic antibiotic use.
The intravenous route should be used for antibiotic administration given that the evidence underpinning this recommendation was based on findings from trials where the majority used this route.
The GDG emphasized the importance of using the simplest and shortest antibiotic regimen for prophylaxis. As the evidence suggests that single-dose regimens are as effective as multiple-dose regimens, the GDG favoured single-dose antibiotic regimens which can easily be given prior to/during caesarean section, rather than multiple-dose regimens which sometimes extend to the postoperative period. Clinical judgement is needed to evaluate other factors that might increase the risk of developing post-caesarean infections and are, therefore, more likely to benefit from multiple antibiotic doses (e.g. prolonged duration of surgery (long “skin-to-skin” interval), difficult surgical manipulation or massive blood loss).
Review question:
Among women undergoing caesarean section (P), does routine antibiotic prophylaxis (I), compared with no antibiotic prophylaxis (C), prevent infectious morbidities and improve outcomes (O)?
Summary of evidence
Evidence for the routine use of prophylactic antibiotics for preventing infection and improving outcomes in women undergoing caesarean sections was extracted from a Cochrane systematic review of 95 trials including over 15 000 women (
41).
Trials were conducted in low-, middle- and high-income countries: 39 in the USA, six in Germany, four in Mexico, three studies each in Canada, Finland, Israel, Italy, South Africa and the UK; two studies each in Austria, France, Greece, Hong Kong, Malaysia, Nigeria, Sweden and the United Arab Emirates; and one study each in China, Denmark, Hungary, Kenya, the Netherlands, New Zealand, Saudi Arabia, Spain, Sudan, Tunisia, Turkey and Zimbabwe.
The antimicrobial agents most often used in the trials included ampicillin, first-generation cephalosporin (usually cefazolin), second-generation cephalosporin (cefamandole or cefuroxime), cefamycin (cefoxitin, cefotetan), metronidazole, penicillins with an extended spectrum of activity (e.g. ticarcillin, mezlocillin or pipericillin), beta-lactam/beta-lactamase inhibitor combination and aminoglycoside-containing combination. Antibiotics were in a majority of cases delivered intravenously. In one study, antimicrobial prophylaxis was administered by rectal suppository. In four studies, follow-up doses were administered by rectal suppository or vaginal tablet. The duration of the postoperative treatment course varied from a single intravenous dose to as long as a week.
A large proportion of the studies (n = 59, 8500 women) gave no information on the type of surgery performed. Clinical definitions for evaluated outcomes were broadly consistent across trials, except for febrile morbidity and serious infectious morbidity. No study reported on baseline risk of infection before the intervention.
Any antibiotic prophylaxis versus no antibiotic prophylaxis (EB Table 16a)
There was a reduction in cases of serious infectious complications (RR 0.31, 95% CI 0.20 to 0.49; 32 trials, 6159 women), maternal febrile morbidity (RR 0.45, 95% CI 0.40 to 0.51; 56 trials, 9046 women), endometritis (RR 0.38, 95% CI 0.34 to 0.42; 83 trials, 13 548 women), wound infections (RR 0.40, 95% CI 0.35 to 0.46; 82 trials, 14 407 women) and maternal urinary tract infections (RR 0.56, 95% CI 0.49 to 0.65; 66 trials, 10 928 women). Adverse events (rash, phlebitis at the site of the intravenous infusion) were more frequent in the treated group (RR 2.43; 95% CI 1.00 to 5.90; 13 trials, 2131 women). There were no serious drug-related adverse events reported. Maternal length of hospital stay was shorter in the treated group (MD -0.46, 95% CI -0.65 to -0.28; 19 trials, 3168 women) compared with controls.
The majority of the trials did not report on neonatal outcomes. Those trials reporting them declared few neonatal deaths but no relationship to the use of antibiotics (two trials), no complications related to drug administration (two trials) or any neonatal morbidity (five trials).
Subgroup analyses based on whether single dose only or multiple dose or either antibiotic regimens were used showed similarity in terms of effect size and direction for all maternal critical outcomes (as shown by the interaction tests) except for endometritis:
- —
Febrile morbidity: single dose (RR 0.50, 95% CI 0.42 to 0.60, 27 trials, 5410 women); multiple doses (RR 0.41, 95% CI 0.35 to 0.49, 26 trials, 3192 women); both (RR 0.33, 95% CI 0.21 to 0.53, 3 trials, 444 women); P = 0.13.
- —
Wound infection: single dose (RR 0.45, 95% CI 0.38 to 0.54, 27 trials, 7937 women); multiple doses (RR 0.35, 95% CI 0.28 to 0.43, 42 trials, 6208 women); both (RR 0.27, 95% CI 0.07 to 0.98, 2 trials, 262 women); P = 0.15, I2 = 47.6%; serious infectious complications: single dose (RR 0.50, 0.25 to 1.0, 15 trials, 3819 women); multiple doses (RR 0.24, 0.13 to 0.43, 17 trials, 2340 women); P = 0.11.
- —
Urinary tract infection: single dose (RR 0.60, 95% CI 0.49 to 0.72, 33 trials, 6941 women); multiple doses (RR 0.51, 95% CI 0.41 to 0.64, 32 trials, 3805 women); both (RR 0.70, 95% CI 0.21 to 2.39, 1 trial, 282 women); P = 0.58.
- —
Maternal hospital stay: single dose (MD -0.39 days, 95% CI -0.60 to -0.19, 12 trials, 2369 women); multiple doses (MD -0.65 days, 95% CI -1.01 to -0.30, 7 trials, 799 women); P = 0.21.
- —
Adverse effects: single dose (RR 2.12, 95% CI 0.66 to 6.75, 7 trials, 1329 women); multiple doses (RR 2.94, 95% CI 0.73 to 11.76, 6 trials, 802 women); P = 0.72.
- —
Endometritis: A significant reduction in endometritis was observed for both dosing types; however, the interaction tests showed a significant difference between subgroups: single dose (RR 0.43, 95% CI 0.38 to 0.50, 41 trials, 8487 women); multiple dose (RR 0.32, 0.27 to 0.37, 40 trials, 4799 women); both (RR 0.36, 95% CI 0.17 to 0.73, 2 trials, 262 women); P = 0.02, I2 = 75.3% – although in the same direction. Multiple-dose antibiotic prophylaxis significantly was associated with a 68% reduction in the risk of endometritis compared to a 57% reduction for single-dose antibiotics.
Antibiotic prophylaxis versus no antibiotic prophylaxis: by antibiotic class (EB Table 16b)
Approximately two thirds of studies evaluated treatment with a first- or second-generation cephalosporin, including cefamycins, or ampicillin. No study reported on monotherapy with a penicillinase-resistant penicillin, fourth-generation cephalosporin, carbapenem, tetracycline, macrolide and aminoglycosides.
There were reductions in maternal outcomes for all antibiotics subgroups, without differences between subgroups for serious infection outcomes (P = 0.93; I2 = 0%) and wound infection (P=0.17; I2 = 26.8%). Interaction tests indicated potentially significant differences among subgroups for febrile morbidity (P < 0.001; I2 = 73.8%) and endometritis (P = 0.07; I2 = 38.6%). The smallest reduction in febrile morbidity was seen for cefamycins (RR 0.73, 95% CI 0.61 to 0.88; 9 trials, 1894 women), and the largest for other regimens (RR 0.23, 95% CI 0.07 to 0.76; 1 trial, 118 women). The smallest reduction in endometritis was seen for beta-lactamase inhibitor combinations (RR 0.67, 95% CI 0.27 to 1.66; 5 trials, 788 women), though this was insignificant, and the largest reduction for natural penicillins (RR 0.19, 95% CI 0.05 to 0.65; 1 trial, 66 women).
Antibiotic prophylaxis versus no antibiotics: by type of caesarean section (EB Table 16c)
Seventeen studies (3500 women) included data on women undergoing elective caesarean sections, according to the review definition, while 22 studies (2500 women) included non-elective procedures. Two studies included both. Three subgroups were compared: elective, non-elective and both elective and non-elective or undefined caesarean section.
Interaction tests showed a significant difference (P = 0.001; I2 = 85.2%) between subgroups for wound infection (elective CS: RR 0.62, 95% CI 0.47 to 0.82; non-elective CS: RR 0.39, 95% CI 0.27 to 0.58; undefined caesarean section: RR 0.34 95% CI 0.28 to 0.40) and maternal urinary tract infection (elective CS: RR 0.92, 95% CI 0.57 to 1.50; non-elective CS: RR 0.44 95% CI 0.31 to 0.60; undefined caesarean section: RR 0.59 95% CI 0.49 to 0.70). There were no differences between subgroups for febrile morbidity (P = 0.79; I2 = 0%), endometritis (P = 0.84; I2 = 0%), febrile morbidity (P = 0.79; I2 = 0%) or serious infectious maternal outcomes (P = 0.73; I2 = 0%).
Considerations related to the strength of the recommendation
Quality of evidence
The quality of evidence was graded as moderate for most critical outcomes. Certain subgroups of antibiotics had lower-quality evidence, ranging from low to very low. Overall, the quality of the evidence for the main comparison was graded as moderate.
Balance of benefits and harms
There is moderate-quality evidence demonstrating the clinical benefits of using prophylactic antibiotics for women undergoing caesarean section, regardless of drug class and regimen or type of caesarean section (elective or non-elective). No antibiotic class or regimen seems to be more effective than the other. Subgroup analyses according to drug regimen and time of administration were observational in nature and not part of a randomized trial or meta-analysis. Serious maternal adverse effects related to the use of antibiotics were rare. However, there is little information about the impact of prophylactic antibiotic treatment on the neonates, although it is unlikely that any potential neonatal risk would outweigh maternal benefits.
Values and preferences
Health care providers, policy-makers and pregnant women and their families in all settings are likely to place a high value on the reduction in serious maternal infectious complications with minimal risk of adverse effects. The panel is confident that there is no variation in this value among mothers, health care providers and policy-makers in low-, middle-and high-income settings.
Resource implications
Antibiotics used for surgical prophylaxis in women undergoing caesarean section are inexpensive, easy to administer and readily available in all settings. It is possible that additional costs related to the use of antibiotics for caesarean section may depend on the overall caesarean section rates across settings. However, these added costs are outweighed by the cost savings for treating short- and long-term complications of post-caesarean infections (e.g. costs related to ICU admission, long hospital stay) to the woman and her family.
RECOMMENDATION 18.1
For caesarean section, prophylactic antibiotics should be given prior to skin incision, rather than intraoperatively after umbilical cord clamping.
(Strong recommendation based on moderate-quality evidence)
REMARKS
The GDG highlighted the importance of administering prophylactic antibiotics at least 15–60 minutes prior to skin incision in optimizing tissue and blood antibiotic concentrations. Based on the pharmacokinetics of common intravenous antibiotics, maximal benefit can be expected when administered between 30 and 60 minutes before skin incision.
The GDG acknowledged that evidence also supports the effectiveness of prophylactic antibiotics after umbilical cord clamping for the prevention of post-caesarean infectious morbidities. Therefore, antibiotics are still beneficial when used outside the suggested timeframe (i.e. 15–60 minutes before incision) and should be applied as circumstances demand. This is particularly important in cases of emergency caesarean section where the available time to administer a prophylactic antibiotic might be limited.
There are no data on the effects of preoperative administration on possible longer-term effects of antibiotic exposure on the baby, and women should be counselled as appropriate. The GDG considers this question a research priority and suggested that opportunities for longer-term follow-up of babies from previous trials should be explored.
Review question:
Among women receiving antibiotic prophylaxis for caesarean section (P), is preoperative administration of antibiotics (I), compared with intraoperative administration of antibiotics (after umbilical cord clamping) (C), more effective in preventing maternal and neonatal infectious morbidities (O)?
Summary of evidence
Evidence on appropriate timing (preoperative versus intraoperative) for the administration of prophylactic antibiotics for preventing infectious morbidities following caesarean section was extracted from a Cochrane systematic review of 10 trials involving 5589 women (5041 analysed) (
42).
Trials were conducted in low-, middle- and high-income countries: five in the USA, two in India and one each in Austria, Egypt and Turkey.
The target population was women undergoing caesarean section, predominantly elective and non-emergency procedures. All but two trials excluded emergency caesarean section. Most of the trials excluded women with chorioamnionitis or other signs of infection, those who had received antibiotics prior to delivery, with ruptured membranes or who delivered preterm. Two trials excluded multiple pregnancies.
Antibiotics for prophylaxis were given intravenously either before the incision or after clamping of the neonatal umbilical cord. Studies administering antibiotics before incision used different timeframes, with the majority ranging from 15 to 60 minutes.
The antibiotics used were different regimens of cephalosporins: seven trials used first-generation cephalosporin (cefazolin 1 g or 2 g), while the other three trials used third-generation cephalosporin (ceftriaxone 1 g or 2 g). Clindamycin was typically the agent of choice for women who had known allergy to cephalosporins.
Prophylactic antibiotics before skin incision versus after umbilical cord clamping (EB Table 17)
Compared with administration after umbilical cord clamping, preoperative antibiotic administration was associated with a 43% reduction in the incidence of endomyometritis (RR 0.54, 95% CI 0.36 to 0.79; 10 trials, 5041 women) and a 41% reduction in the incidence of wound infection (RR 0.59, 95% CI 0.44 to 0.81; 10 trials, 5041 women). These findings were consistent between trials testing first- and second-generation cephalosporins.
There were no significant differences between the group receiving antibiotics before incision versus after umbilical cord clamping in the incidence of urinary tract infection (RR 1.02, 95% CI 0.65 to 1.59; 8 trials, 4001 women), pelvic abscesses (RR 1.00, 95% CI 0.06 to 15.97; 1 trial, 741 women), respiratory infections (e.g. pneumonia) (RR 2.30, 95% CI 0.34 to 15.45; 4 trials, 1849 women) or febrile illness (RR 0.93, 95% CI 0.63 to 1.35; 4 trials, 2650 women).
Two trials (1274 women) collected information on septic pelvic thrombophlebitis, and one trial (874 women) on septic shock and maternal death, but reported no events.
There was evidence of a significant reduction in maternal hospital stay among women receiving antibiotics preoperatively compared with women receiving it during caesarean section (MD -0.17, 95% CI -0.30 to -0.04; 2 trials, 1342 women).
For the neonate, there was no statistically significant difference between the group receiving antibiotics before incision versus after umbilical cord clamping regarding the incidence of neonatal sepsis (RR 0.76, 95% CI 0.51 to 1.13; 5 trials, 2907 neonates), neonatal sepsis workup (RR 0.92, 95% CI 0.69 to 1.23; 4 trials, 1170 neonates), infection with a resistant organism (RR 0.70, 95% CI 0.12 to 4.14; 1 trial, 379 neonates)' febrile illness (RR 0.67, 95% CI 0.28 to 1.62, 1 trial, 953 neonates), ICU admission (RR 0.91, 95% CI 0.74 to 1.13; 6 trials, 3708 neonates) or duration of ICU stay (MD -0.07 days, 95% CI -2.60 to 2.46; 3 trials, 1731 neonates). Neonatal mortality was not reported by any of the trials.
Other critical outcomes were not reported.
Considerations related to the strength of the recommendation
Quality of evidence
The majority of critical outcomes were graded as being of moderate to high quality. Overall, quality of evidence was graded as moderate.
Balance of benefits and harms
Preoperative (pre-incision) antibiotic administration is more effective than intraoperative administration for reducing infectious morbidities in women undergoing caesarean section, without evidence of adverse neonatal outcomes. The theoretical concerns regarding antibiotic exposure of the neonate were not demonstrated by the review, although the results were limited only to short-term outcomes. Given the increasing evidence of the importance of appropriate bacterial colonization after birth for infant health and immune system development, it remains unclear whether the demostrated reduction in the risk of wound infection/endometritis outweighs the longer-term effect of prenatal antibiotic exposure on the baby.
Values and preferences
Health care providers, policy-makers and pregnant women and their families in all settings are likely to place a high value on the added clinical benefits of preoperative antibiotic admnistration in terms of further reduction in endomyometris and wound infection, without adverse effects on neonatal infections. The panel is confident that there is no variation in this value among mothers, health care providers and policy-makers in low-, middle- and high-income settings.
Resource implications
Antibiotics are widely available in parenteral forms in all settings where caesarean sections are performed. The timing of antibiotic administration for caesarean section prophylaxis is unlikely to impact the health care costs needed to implement the intervention.
RECOMMENDATION 18.2
For antibiotic prophylaxis for caesarean section, a single dose of first generation cephalosporin or penicillin should be used in preference to other classes of antibiotics.
(Conditional recommendation based on very low-quality evidence)
REMARKS
The GDG noted that the available evidence on the effectiveness of antibiotics came largely from trials that tested first-generation cephalosporin or penicillin. Based on consensus, the group favoured these classes of antibiotics over other classes of antibiotics, as they have a broad spectrum of activities and are widely available in all settings.
In acknowledgement of the lack of evidence on the comparative effectiveness of different classes of antibiotics, the GDG concluded that when the recommended antibiotic classes are not available, other classes of antibiotics may also be used. The group noted that the choice of such antibiotic class should be informed by the local bacteriologic patterns of post-caesarean infectious morbidity, the availability of such antibiotic class, the woman's allergy history, the clinician's experience with that particular class of antibiotics, and its cost.
Due to the high risk of necrotizing enterocolitis among preterm babies, the use of “co-amoxiclav” for antibiotic prophylaxis should be avoided not only for caesarean delivery of preterm infants, but it might also be safer to avoid its use for caesarean delivery of term babies.
Review question:
Among women receiving routine antibiotic prophylaxis for caesarean section (P), is the use of a particular class of antibiotics (I), compared with other classes of antibiotics (C), more effective in preventing postoperative infectious morbidities (O)?
Summary of evidence
Evidence on the comparative effectiveness and safety of different classes of antibiotics to prevent infectious morbidity in women undergoing caesarean section was extracted from a Cochrane systematic review of 35 trials (with 31 trials providing data for 7697 women) (
43).
Trials were conducted in low-, middle- and high-income countries: 12 studies in the USA, four in India, three in Italy, two in Thailand, and one each in the following countries: Argentina, Canada, Greece, Finland, the Netherlands, Malaysia, Mozambique, Rwanda, South Africa, Sudan, Switzerland, the UK, the United Arab Emirates and Zimbabwe.
The trials included women undergoing either elective or non-elective caesarean section. All but five trials administered prophylactic antibiotics after umbilical cord clamping. Antibiotics were administered preoperatively in four trials, and information about the timing of antibiotic administration was not available for one trial.
The comparisons considered in the review were those between two or more antibiotics of different classes. Comparisons of different drugs or drug regimens within the same class were excluded. The majority of the trials included compared cephalosporin with penicillin. The overall data for any cephalosporin versus any penicillin were not pooled but analysed according to the following subgroups: single cephalosporin versus single penicillin, single cephalosporin versus penicillin combination, cephalosporin combination versus single penicillin and cephalosporin combination versus penicillin combination. Three trials compared a cephalosporin or penicillin with another class of antibiotics. Few studies compared mixed antibiotic regimens (which do not include a cephalosporin or penicillin) with cephalosporin or penicillin.
Cephalosporin versus penicillin (EB Tables 18a-18j)
Single cephalosporin versus single penicillin (13 trials, 4010 women)
There were no cases of maternal sepsis in 346 women involved in two trials.
There were no significant differences between cephalosporin and penicillin regimens for endometritis (RR 1.11, 95% CI 0.81 to 1.52; 9 trials, 3130 women), maternal febrile morbidity (RR 0.89, 95% CI 0.61 to 1.30; 7 trials, 1344 women), wound infection (RR 0.83, 95% CI 0.38 to 1.81; 9 trials, 1497 women) or urinary tract infection (RR 1.48, 95% CI 0.89 to 2.48, 7 trials, 1120 women). There were no significant differences between antibiotic classes on a maternal composite of adverse effects (RR 2.02, 95% CI 0.18 to 21.96; 3 trials, 1902 women).
No cases were reported in the two trials evaluating a composite outcome of maternal serious infectious morbidity.
None of the included studies reported neonatal sepsis.
Single cephalosporin versus penicillin combination (12 trials, 2875 women)
There were no significant differences between comparison groups for maternal sepsis (RR 2.37, 95% CI 0.10 to 56.41; 1 trial, 75 women), endometritis (RR 0.90, 95% CI 0.60 to 1.35, 10 trials, 2134 women), maternal febrile morbidity (RR 0.92, 95% CI 0.56 to 1.49; 6 trials, 1824 women), wound infection (RR 0.72, 95% CI 0.40 to 1.30; 7 trials, 1608 women) or maternal urinary tract infection (RR 0.66, 95% CI 0.17 to 2.55; 6 trials, 1361 women). There were no significant differences between antibiotic classes on maternal composite adverse effects (RR 0.96, 95% CI 0.09 to 10.50; 4 trials, 1333 women).
None of the included studies reported neonatal sepsis.
Cephalosporin combination versus single penicillin (1 trial, 147 women)
From one study involving 139 women, there were no differences observed between groups for endometritis (RR 2.70, 95% CI 0.63 to 11.55), maternal febrile morbidity (RR 2.36, 95% CI 0.84 to 6.62) or wound infection (RR 2.02, 95% CI 0.42 to 9.63). The other critical outcomes were either not reported by the study or there were no events.
Cephalosporin combination versus penicillin combination (2 trials, 363 women)
There were no significant differences between groups in maternal sepsis (RR 3.21, 95% CI 0.34 to 30.45; 1 trial, 232 women), endometritis (RR 0.33, 95% CI 0.01 to 7.77; 1 trial, 83 women), maternal fever (RR 1.57, 95% CI 0.69 to 3.60; 2 trials, 315 women) or wound infection (RR 1.23, 95% CI 0.42 to 3.58; 2 trials, 315 women).
Cephalosporins versus penicillins: comparison by type of caesarean section (22 trials, 5788 women)
Most of the trials included women undergoing either elective or emergency caesarean sections.
There was no significant difference between women undergoing elective or emergency caesarean section for maternal sepsis (RR 2.91, 95% CI 0.47 to 18.10; 4 trials, 653 women).
There was a significant difference between subgroups for endometritis (I2 = 61.4%). Penicillins showed a trend towards superior effectiveness compared with cephalosporins for reducing endometritis among women undergoing non-elective caesarean section (RR 1.33, 95% CI 1.01 to 1.75, 6 trials, 2362 women). The differences were not significant for elective caesarean section (RR 2.06, 95% CI 0.66 to 6.39; 3 trials, 461 women) or when type of caesarean section was not differentiated (RR 0.85, 95% CI 0.60 to 1.19; 11 trials, 2567 women).
Cephalosporins versus penicillins: comparison by timing of administration (22 trials, 5788 women)
All but two trials administered antibiotics before umbilical cord clamping. Two trials did not report on the timing of antibiotic administration.
There were no significant differences between antibiotics for maternal sepsis (RR 2.91, 95% CI 0.47 to 18.10; 4 trials, 653 women) or endometritis (RR 1.11, 95% CI 0.90 to 1.37; 20 trials, 5390 women).
Cephalosporins versus penicillins: comparison by route of administration (22 trials, 5788 women)
Twenty trials compared antibiotics given intravenously. Two studies compared the antibiotics when administered as a lavage/irrigation during surgery.
There were no significant differences between antibiotic classes for maternal sepsis (RR 2.90, 95% CI 0.46 to 18.17, 4 trials, 653 women) or endometritis (RR 1.12, 95% CI 0.92 to 1.37, 20 trials, 5390 women) in relation to the route of administration.
First-generation cephalosporins versus extended-spectrum penicillins (2 trials, 822 women)
Extended-spectrum penicillins were more efficient in preventing endometritis than first-generation cephalosporins (RR 2.18, 95% CI 1.30 to 3.66; 2 trials, 814 women). However, no differences were reported in maternal fever (RR 2.36, 95% CI 0.84 to 6.62; 1 trial, 139 women) or wound infection (RR 2.02, 95% CI 0.42 to 9.63; 1 trial, 139 women).
First-generation cephalosporins versus aminopenicillins (8 trials, 1882 women)
There were no significant differences between groups for endometritis (RR 1.09, 95% CI 0.69 to 1.71; 7 trials, 1487 women), maternal febrile morbidity (RR 0.78, 95% CI 0.40 to 1.51; 5 trials, 883 women), wound infection (RR 0.85, 95% CI 0.36 to 2.01; 5 trials, 626 women) or urinary tract infections (average RR 1.41, 95% CI 0.54 to 3.70; 5 trials, 626 women).
A reduction in the maternal length of hospital stay was observed in the group receiving cephalosporins compared to aminopenicillins (MD -1.50, 95% CI -2.46 to -0.54; 1 trial, 132 women).
Second-generation cephalosporins versus extended spectrum penicillins (6 trials, 2077 women)
There were no significant differences between groups for endometritis (RR 1.10, 95% CI 0.78 to 1.54; 4 trials, 1334 women), maternal febrile morbidity (RR 1.08, 95% CI 0.79 to 1.47; 4 trials, 850 women), wound infection (RR 2.37, 95% CI 0.64 to 8.73; 2 trials, 438 women) or urinary tract infection (RR 1.43, 95% CI 0.67 to 3.07, 3 trials, 567 women). There were no reported events of maternal sepsis (1 trial, 287 women), post-discharge infections (3 trials, 305 women) or maternal composite adverse effects (RR 2.02, 95% CI 0.18 to 21.96; 2 trials, 1030 women).
Second-generation cephalosporins versus aminopenicillins (8 trials, 1921 women)
There were no significant differences between groups in maternal sepsis (RR 2.37, 95% CI 0.10 to 56.41; 1 trial, 75 women), endometritis (RR 1.01, 95% CI 0.75 to 1.35; 8 trials, 1890 women), maternal febrile morbidity (RR 1.17, 95% CI 0.64 to 2.15; 3 trials, 387 women), wound infection (RR 1.14, 95% CI 0.47 to 2.78; 5 trials, 638 women), maternal urinary tract infection (RR 0.63, 95% CI 0.11 to 3.66; 4 trials, 462 women) or maternal composite of adverse effects (RR 1.92, 95% CI 0.18 to 20.82; 3 trials, 1130 women).
Third-generation cephalosporins versus extended-spectrum penicillins (2 trials, 359 women)
Extended-spectrum penicillins were more efficient than third-generation cephalosporins in preventing endometritis (RR 2.14, 95% CI 1.14 to 4.00; 1 trial, 300 women). Other considered outcomes reported no events.
Third-generation cephalosporins versus aminopenicillins (7 trials, 1904 women)
There were no significant differences between groups for endometritis (RR 1.47, 95% CI 0.89 to 2.42; 5 trials, 1472 women), maternal febrile morbidity (RR 1.12, 95% CI 0.69 to 1.83; 3 trials, 1060 women), maternal urinary tract infection (RR 0.52, 95% CI 0.10 to 2.80; 2 trials, 233 women) or length of maternal hospital stay (MD -0.03, 95% CI -0.14 to 0.08; 1 trial, 746 women).
Wound infections were reduced in the group receiving third-generation cephalosporins (RR 0.49 95% CI 0.27 to 0.90; 6 trials, 1556 women).
Fluoroquinolones versus penicillin or cephalosporin (EB Tables 18k–18l)
Two very small trials tested ciprofloxacin versus ampicillin/sulbactam (72 women) or cefotetan (81 women).
No differences were found between these antibiotics regarding maternal sepsis, endometritis or wound infection.
Other antibiotic regimens versus penicillin or cephalosporins (EB Tables 18m-18n)
There were other comparisons between other antibiotic class combinations versus penicillin or cephalosporin:
- —
Lincosamide plus aminoglycoside versus penicillin (1 trial, 88 women)
- —
Beta-lactam versus cephalosporin (2 trials, 118 women)
There were no differences observed between the comparison groups for the outcomes reported: wound infection and endometritis.
Aminoglycoside plus nitroimidazole versus standard antibiotic cocktail (EB Table 18o)
One trial involving 241 women compared gentamicin (aminoglycoside) plus metronidazole (nitroimidazole) with a standard cocktail of antibiotics (containing penicillin, nitroimidazole and macrolide). There was no significant difference between the two groups with regard to endometritis (RR 0.81, 95% CI 0.29 to 2.26; 1 trial, 241 women), maternal fever (RR 1.12, 95% CI 0.69 to 1.83; 3 trials, 1060 women), wound infection (RR 3.23, 95% CI 0.34 to 30.64; 1 trial, 241 women) or maternal urinary tract infection (RR 1.08, 95% CI 0.07 to 17.03; 1 trial, 241 women).
Considerations related to the strength of the recommendation
Quality of the evidence
Several critical outcomes were graded as very low-quality evidence for the various comparisons. Overall, the quality of evidence was graded as very low.
Balance of benefits and harms
There is no evidence to demonstrate that any specific class of antibiotic is better than the other for prophylaxis in women undergoing caesarean section. However, first-generation cephalosporins and penicillin have an advantage over other classes of antibiotics in terms of cost and wide availability in all settings.
Values and preferences
Health care providers and policy-makers are likely to vary in their choice of antibiotic class for prophylaxis at caesarean section. These choices are likely to vary widely across settings but would often be dependent on local bacteriological patterns of post-caesarean infectious morbidity, patterns of bacterial resistance, availability and cost of antibiotics and common indications for caesarean section.
Resource implications
First-generation cephalosporin and penicillin are relatively cheaper than newer antibiotics, easy to administer, and readily available in all settings with the capacity to perform caesarean sections. The implementation of this recommendation is likely to save health care costs related to multiple doses of antibiotics where such practice is currently the norm.
