Azithromycin in labour: maternal benefit, no newborn survival gain

A single 2 g oral dose of azithromycin given during labour cut maternal infection by 29% across seven low- and middle-income countries (RR 0.71; 95% CI 0.64–0.79) ¹. Among 29,278 women in the A-PLUS trial, infection fell from 5.6% with placebo to 4.0% with azithromycin ¹. Later analyses examine that maternal benefit and its costs, while the larger trials have not established reductions in newborn sepsis or death ¹¹¹²¹⁵.

There is a twist that matters for policy. In 2025, the World Health Organization (WHO) reviewed the same trials and decided not to recommend routine antibiotic prophylaxis in labour for vaginal birth ¹⁶. This article explains the evidence, why WHO reached that view, and what it means for ministries that still face high rates of maternal sepsis.

The burden

About 260,000 women died during and following pregnancy and childbirth in 2023 ¹⁷. Sub-Saharan Africa accounted for about 70% of these deaths (182,000), and southern Asia for about 17% (43,000) ¹⁷. The global maternal mortality ratio was 197 per 100,000 live births. In low-income countries it was 346, against 10 in high-income countries ¹⁷.

There has been progress. The maternal mortality ratio fell by about 40% worldwide between 2000 and 2023. Southern Asia cut it by 71% ¹⁷.

Infection is one of the five complications that cause about 75% of maternal deaths, usually after childbirth ¹⁷. The last WHO systematic analysis of causes, covering 2003–2009, attributed 10.7% of maternal deaths to sepsis, about 261,000 deaths over that period ¹⁹. Only haemorrhage and hypertensive disorders caused more direct obstetric deaths ¹⁹.

Nigeria has the highest number of maternal deaths of any country, and peripartum sepsis is one of the leading causes there ⁶.

Newborns share this risk. In 2022, 2.3 million newborns died in their first 28 days, 47% of all under-5 deaths ¹⁸. Neonatal infections are one of the main causes, along with preterm birth, birth complications and congenital anomalies ¹⁸. That is why trialists hoped a single dose given to the mother in labour could protect both.

How it works

Azithromycin is an oral antibiotic. In A-PLUS, women in labour at 28 weeks or later who were planning a vaginal birth took a single 2 g dose by mouth ¹¹.

The trial counted a range of maternal infections: chorioamnionitis (infection of the membranes), endometritis (infection of the womb lining), perineal or caesarean wound infection, abdominopelvic abscess, mastitis or breast abscess, and other infections ¹.

Trialists also hoped the dose given to the mother would protect her newborn. Reduced bacterial carriage and some secondary infection outcomes did not translate into reduced newborn sepsis or mortality in the larger trials ¹²¹⁵²⁰.

For a health worker, the practical picture is simple. It is one dose, given by mouth, in labour. It does not need an injection. Its benefit for the mother did not depend on exactly when in labour it was given ².

What the evidence shows

2016: the hypotheses behind the larger trials

A Gambian trial of 829 women found that oral azithromycin during labour reduced newborn nasopharyngeal carriage of selected bacteria at day six: 28.3% versus 65.1% (prevalence ratio 0.43, 95% CI 0.36–0.52) ²⁰. That was a biological rationale for larger trials, not evidence that fewer newborns would die. The later sepsis and mortality results below test that hypothesis directly.

The same year, the US C/SOAP trial found fewer postoperative infections when intravenous azithromycin was added to standard prophylaxis for nonelective caesarean delivery: 6.1% versus 12.0% (risk ratio 0.51, 95% CI 0.38–0.68) ²¹. That different route, procedure and background treatment cannot establish the benefit of routine oral prophylaxis for vaginal birth. The neonatal composite outcome did not significantly improve ²¹.

A-PLUS main trial (7 countries, NEJM 2023)

A-PLUS was run by the NICHD Global Network with the University of Alabama at Birmingham. Sites were in the Democratic Republic of the Congo, Kenya, Zambia, Bangladesh, India, Pakistan and Guatemala ¹. It enrolled 29,278 women in labour at 28 weeks' gestation or more who planned a vaginal birth ¹¹.

The two primary outcomes were maternal sepsis or death, and stillbirth, newborn death or newborn sepsis ¹¹. The data and safety monitoring committee stopped the trial early for maternal benefit ¹¹.

Maternal sepsis or death was 1.6% with azithromycin and 2.4% with placebo (RR 0.67; 95% CI 0.56–0.79) ¹¹. The drug had little effect on the newborn outcome ¹¹. The trial is registered as NCT03871491 ¹¹. The main result, published in early 2023, is inside the expanded window, alongside its follow-up papers.

A-PLUS: any maternal infection

A 2025 post-hoc analysis in Lancet Global Health looked at any maternal infection, not only sepsis. Any maternal infection occurred in 580 of 14,558 women given azithromycin (4.0%) and 824 of 14,661 given placebo (5.6%) (RR 0.71; 0.64–0.79) ¹. Baseline characteristics were similar between groups ¹.

A-PLUS: timing in labour

A secondary analysis of 29,236 women asked whether timing mattered. The benefit for mothers was the same whether the dose was given 12 hours or less before delivery (RR 0.71; 0.64–0.79) or more than 12 hours before (RR 0.71; 0.54–0.94). The interaction p-value was 0.987 ².

In sub-Saharan Africa, there was a difference: RR 0.21 (0.08–0.54) when the dose came more than 12 hours before delivery and 0.52 (0.41–0.66) when it came within 12 hours (interaction p = 0.03) ². The authors conclude that women benefit regardless of how soon delivery is expected ².

A-PLUS: women with and without other antibiotics

Many women in A-PLUS also received antibiotics outside the study. A secondary analysis looked at whether this changed the effect ³.

Among women who received no other antibiotics, maternal infection fell by 42% (RR 0.58; 0.48–0.70). Among women who did receive other antibiotics, it fell by 20% (RR 0.80; 0.70–0.91) ³. Results for maternal sepsis were similar ³. The benefit is largest where other antibiotics are not given, which describes many rural and low-resource settings.

A-PLUS: cost-effectiveness

Researchers at UNC and RTI modelled the intervention across the A-PLUS countries with 100,000 simulations, from a health-care sector view ⁴.

Per 100,000 pregnancies, azithromycin averted 1,592 cases of maternal infection, sepsis or death (95% CI 1,139.7–2,024.1) ⁴. It averted 248.5 facility readmissions, 866.8 unplanned clinic visits and 1,816 antibiotic courses ⁴.

The model estimated net savings of US$32,661 per 100,000 pregnancies (95% CI −52,218 to 118,210) and 13.2 DALYs averted (8.3–17.9) ⁴. The authors conclude the dose is cost-saving in most scenarios ⁴. The cost of readmission, the cost of the drug and the risk of infection had the most effect on the result ⁴.

A-PLUS: resistance in mothers and babies

A Boston University substudy followed 911 women and 915 infants at 8 sites in 7 countries, with nose swabs from labour to 12 months after birth ⁵.

Azithromycin resistance in nasal Staphylococcus aureus was higher in mothers given azithromycin from day 7 to 3 months, and in infants only at 6 weeks ⁵. By 6–12 months in mothers, and after 6 weeks in infants, the difference was gone ⁵.

PregnAnZI-2 (The Gambia and Burkina Faso)

PregnAnZI-2 tested the same 2 g oral dose at 10 health facilities in The Gambia and Burkina Faso, from October 2017 to May 2021 ¹². It randomised 11,983 women in labour. Its primary outcome was newborn sepsis or death ¹².

The primary outcome was the same in both arms: 2.0% with azithromycin and 1.9% with placebo (risk difference 0.09; 95% CI −0.39 to 0.57) ¹². Newborn deaths were 0.8% in both arms, and newborn sepsis 1.3% in both ¹².

There were smaller gains. Newborns in the azithromycin group had fewer skin infections (0.8% vs 1.7%) and needed antibiotics less often (6.2% vs 7.8%) ¹². Mothers had less mastitis (0.3% vs 0.5%) and less puerperal fever (0.1% vs 0.3%) ¹². The authors concluded their results do not support routine use to prevent newborn sepsis ¹². The trial is registered as NCT03199547 ¹².

Bulabula MaPei (Fiji)

In Fiji, the Bulabula MaPei trial by the Murdoch Children's Research Institute enrolled 2,110 pregnant people at a tertiary hospital in Suva from June 2019 to January 2022 ¹⁰. Its primary outcome was infant skin and soft-tissue infection by 3 months.

That outcome was 5.8% with azithromycin and 7.8% with placebo, but the difference was not significant (RR 0.73; 0.51–1.06) ¹⁰. Serious adverse events were similar in both arms, and no infant had hypertrophic pyloric stenosis ¹⁰.

Secondary outcomes showed more. Among 2,122 infants, any infection by 3 months was 13.6% with azithromycin and 17.3% with placebo (RR 0.79; 0.63–0.99) ⁹. Infections in birthing parents were cut most at 1 week after birth (RR 0.31; 0.13–0.71), with the effect fading up to 6 months ⁹. There was no effect on antibiotic prescriptions ⁹.

The Cochrane review (2025)

A Cochrane review in August 2025 pooled four trials of a single oral dose of azithromycin in labour, with 42,846 participants ¹⁵.

Antibiotic prophylaxis probably reduces maternal sepsis, from 1.8% to 1.2% (RR 0.65; 95% CI 0.56–0.77; moderate certainty) ¹⁵. It likely makes little or no difference to maternal death (RR 1.21; 0.63–2.33; moderate certainty) ¹⁵. It makes little or no difference to newborn sepsis (RR 1.03; 0.96–1.10) or newborn death (RR 1.03; 0.87–1.21), both with high certainty ¹⁵.

Results table

LimitsWhat didn't work (or the limits)

No established reduction in newborn sepsis or death. In the A-PLUS timing analysis, infant infection did not fall at any timing: RR 1.00 (0.95–1.06) when the dose came 9 hours or less before birth, and 1.01 (0.88–1.15) when it came more than 9 hours before ². Giving the dose earlier does not fix the null newborn result.

Newborn infection was not reduced in any subgroup of the antibiotic-use analysis either ³. PregnAnZI-2 found no effect on newborn sepsis or death ¹². The Cochrane review rates the finding of little or no effect on those two outcomes as high-certainty evidence ¹⁵.

No brain-protection benefit. Some hoped azithromycin might protect newborns who suffer birth asphyxia. A UAB follow-up of 403 such children at 24 months in India, Pakistan, Zambia, the DRC and Guatemala found identical Bayley-III cognitive scores: 90.9 in both arms (mean difference 0.29; −1.77 to 2.34) ⁷. Language and motor scores also did not differ ⁷.

No effect on bleeding. A secondary analysis found no effect on postpartum haemorrhage (1.4% vs 1.6%; RR 0.88; 0.73–1.07) or blood transfusion (RR 0.90; 0.65–1.25) ⁸.

No effect on maternal deaths in pooled data. The drug prevents infections and sepsis. But the Cochrane review found little or no difference in maternal death (RR 1.21; 0.63–2.33) ¹⁵. Maternal deaths were rare in these trials, so the trials may simply be too small to show an effect on death.

Secondary infection outcomes do not establish a survival benefit. In Fiji, any infant infection by 3 months was lower with azithromycin (RR 0.79; 0.63–0.99) ⁹. But this was a secondary outcome. The primary outcome, infant skin and soft-tissue infection, was not significantly reduced ¹⁰. Fiji is also not a high-mortality setting, so this result does not transfer easily to the A-PLUS countries.

Resistance is short-lived, not zero. The nasal S. aureus rise in mothers lasted up to 3 months ⁵. In PregnAnZI-2, a substudy found that azithromycin-resistant E. coli in infants' guts was more common with azithromycin at day 6 (13.4% vs 3.6%) and day 28 (16.4% vs 9.6%) ¹³. Carriage of Klebsiella pneumoniae rose, and azithromycin-resistant K. pneumoniae was more common at day 28 (7.3% vs 2.1%) ¹³.

Changes to the infant gut. A Gambian substudy of 127 infants from PregnAnZI-2 found that, in the first month of life, the azithromycin group had a slower shift in gut bacteria, more Enterobacteriaceae and fewer Bifidobacterium ¹⁴. The authors say this may have undesirable effects on the child, though long-term harm has not been shown ¹⁴.

Cost and scale

The drug likely saves money. The UNC and RTI model found net savings of US$32,661 per 100,000 pregnancies across the A-PLUS countries, mainly from fewer readmissions, clinic visits and antibiotic courses ⁴. The confidence interval includes a net cost, so savings are not certain in every setting ⁴.

Who delivers. The dose is given in health facilities during labour. All the trials above were facility-based [1,10,12]. A-PLUS did not test delivery at home births.

Guideline status. WHO issued a new recommendation in 2025 ¹⁶. It states: "Routine antibiotic prophylaxis during labour for vaginal birth is not recommended" ¹⁶. WHO cited small reductions in maternal infections, no impact on mortality or newborn outcomes, and concern that unnecessary antibiotic use adds to antimicrobial resistance ¹⁶. Instead, it emphasises infection prevention and control, targeted antibiotic use based on risk, and prompt diagnosis and treatment of maternal infections ¹⁶.

This is the key fact for ministries. The trial evidence for a maternal benefit is solid. The WHO position, weighing that benefit against resistance and the lack of effect on deaths, is against routine use.

Implementation research continues. Nigeria is testing the dose in routine care. The AZIN-V trial will enrol 5,040 women in 48 health facilities across six states: Ebonyi, Edo, Gombe, Kano, Kwara and Lagos ⁶. Facilities are randomised to oral azithromycin in labour or usual care ⁶. A second part, in 12 facilities, compares a bottom-up Plan-Do-Study-Act approach with a one-time training workshop, to see which gets staff to follow the protocol ⁶. The study is approved by Nigeria's National Health Research Ethics Committee, the ministries of health of the six states and the National Agency for Food and Drug Administration and Control, with teaching hospitals in Lagos, Ilorin, Benin, Kano and Irrua ⁶.

What to watch

AZIN-V in Nigeria (ISRCTN16415327). This will be the first large test of the dose in routine Nigerian care, including cost-effectiveness in a setting with weak antibiotic stewardship and high resistance ⁶. Its results will show whether the A-PLUS benefit holds outside a tightly run trial.

How countries respond to the WHO recommendation. Some ministries may want to use the dose for women at higher risk of infection. The WHO guidance points to targeted, risk-based antibiotic use rather than routine use for everyone ¹⁶. How "higher risk" is defined will matter.

Resistance and the infant gut. Resistance rose for weeks to months in mothers and infants [5,13]. The Fiji authors call for more work on the infant microbiome and resistome ⁹. Long-term follow-up would help.

Target groups. The Fiji team suggests that further research should identify which populations benefit most ⁹. The A-PLUS analysis of women with no other antibiotics points one way: the benefit is biggest where other antibiotics are not used ³.

What it means

For ministries of health. Intrapartum azithromycin is a maternal intervention. It prevents about three in ten maternal infections ¹, works regardless of timing in labour ², helps most where other antibiotics are scarce ³, and likely saves health systems money ⁴. But WHO does not recommend it as routine care for vaginal birth, because it did not reduce deaths and adds antibiotic pressure [15,16].

A ministry that wants to act on maternal sepsis should start with what WHO does recommend: infection prevention and control in labour wards, prompt diagnosis, and fast treatment of infection ¹⁶. Where a ministry wants to go further, Nigeria's AZIN-V model, a trial built into routine care with cost and resistance measured, is the careful way to do it ⁶.

For clinicians. Do not give intrapartum azithromycin to prevent newborn sepsis. The evidence on that is now high-certainty and negative ¹⁵. Lower bacterial carriage and some secondary infection outcomes should not be described as established protection against newborn sepsis or death ¹²¹⁵²⁰.

For funders. The strongest case for funding is in research and implementation that can settle the open questions: whether a targeted, risk-based approach can keep the maternal benefit while limiting resistance, and whether the benefit holds in routine care [6,16].

Children still gain from healthier mothers. A mother who survives childbirth without sepsis is there to feed and care for her newborn. Keeping mothers alive and well remains one of the surest ways to protect their babies. The dose should not be presented as an established way to prevent newborn sepsis or death.