Start · Catalogue · Profile · Table
Healthcare BODY HANDBOOK
Healthcare · §627
Group B Strep in Pregnancy
Around one in four pregnant women carries a bacterium called group B strep without knowing it — no symptom, no smell, no infection of her own. The risk is to the baby on the way out: a small chance of overwhelming infection in the first week of life, large enough that one swab at 36 weeks and a four-hour antibiotic drip in labor have become the standard of care across most of the world. That's the entire intervention. The choice you actually face is not should I do this — it's understanding what the antibiotic buys, what it doesn't (late infections, weeks later, aren't on the prevention list), and why the UK and the US have reached different conclusions on whether to screen everyone.
Do · Once Evidence Strong Chapter Healthcare

A swab at the 36-week visit, IV antibiotics during labor if it's positive — that's the protocol, and it cuts the chance of early newborn infection by about four-fifths. Strong evidence, near-zero burden for you (the drug is bundled into normal labor care), real but small downsides for the baby's early gut bacteria. The genuine open question is global policy, not whether to do it this pregnancy.

Group B strep — Streptococcus agalactiae, "GBS" in every chart — is a normal gut bacterium that also colonizes the vagina in about one in four women. It's not sexually transmitted in any actionable sense, not a sign of bad hygiene, and not an infection of the mother. She doesn't feel it. She doesn't have a discharge, an itch, or a fever from it. About half the time, colonization comes and goes on its own across weeks.

The vulnerability is the newborn passing through the birth canal, or being exposed to GBS-laden amniotic fluid after the water breaks. The bacteria get aspirated into the lungs and the baby's bloodstream within hours of delivery. Most colonized mothers pass GBS to the baby; only a small fraction of those exposed get sick — but when an exposed baby does get sick, it's serious: bloodstream infection, pneumonia, or meningitis presenting on day 1, 2, or 3 of life Verani et al. 2010. This is what "early-onset" group B strep disease means in the literature: the first six days. There's a second category called "late-onset," from day 7 to about three months, that the intrapartum protocol does not prevent — more on that under misconceptions.

The antibiotic works by sitting in two places at once. Given by IV to the mother during labor, penicillin crosses the placenta within minutes and reaches bactericidal levels in the baby's bloodstream before delivery. It also drives down the population of GBS in the vagina that the baby has to pass through. About four hours from first dose to delivery is the threshold where both of those effects are doing their work — the cutoff isn't arbitrary; it's where the cohort data flatten out into the prevented zone Boyer and Gotoff 1986.

Does it actually work?

Yes — at the scale of national statistics, with the temporal precision you almost never get from a public-health intervention. Early-onset group B strep disease in the United States ran at about 1.7 cases per 1,000 live births in the early 1990s before any organized prevention. By 2015, after universal screening rolled out, it was down to about 0.23 per 1,000 — a drop of roughly four-fifths, with the curve bending each time the guidelines tightened Schrag et al. 2000, Nanduri et al. 2019.

The Cochrane review is more cautious than the population data — four trials, 852 women, evidence rated "very low" because the trials had allocation and detection issues, and no mortality benefit could be shown directly Ohlsson and Shah 2014. Skeptics make a real point with that: most of the case for screening rests on ecological surveillance rather than placebo-controlled RCTs, which can no longer ethically be run now that the standard of care has converged. A 2020 meta-analysis comparing the universal-screening approach to the risk-based approach (more on that under alternatives) found universal screening associated with about half the early-onset disease rate Hasperhoven et al. 2020.

The honest summary: the strongest version of the evidence is a near-elimination of early newborn infection in countries that screen universally, anchored on one RCT, decades of surveillance, and aligned guidelines from three continents. The weakest version is that no modern placebo-controlled trial has been or will be run, the Cochrane bias rating is low, and one major country (the UK) has read the same evidence and chosen a different program.

What you're protecting against

The first week of your baby's life, if early-onset GBS happens, looks like this: feeding poorly by hour 12, breathing fast by hour 18, a fever or a strange low temperature, then the call from the nursery nurse asking whether the pediatrician can come now. Blood cultures, a lumbar puncture in a four-pound newborn, an IV antibiotic line into the foot, the NICU isolette for the next two weeks if you're lucky and the next two months if you're not. Among term babies who get it, about four to six in a hundred die; among preterm babies the share is closer to one in four to one in three Edmond et al. 2012. Among survivors of GBS meningitis specifically, roughly a third are left with permanent damage — hearing loss, developmental delay, cerebral palsy, seizure disorder.

You will not feel any of this in your body. There is no felt-experience symptom that tells you you're a carrier or that your baby is being exposed during labor. The whole protocol exists to interrupt a cascade that would otherwise play out without anyone in the room knowing it was happening until day two or three. The pre-1990s United States carried about seven thousand cases a year with around four hundred newborn deaths Nanduri et al. 2019; the same country today, after universal screening, carries about a thousand cases and fifty deaths. The difference is the swab and the four hours of antibiotics.

The trade your screen-and-treat decision is making is asymmetric in a textbook way. If you carry GBS and labor without antibiotics, the chance of your baby getting early-onset disease is about one to two in a hundred; with appropriate antibiotics, it drops to about one in a thousand. The absolute numbers are small. The consequences when they land are not small. The cost on your side is an IV that's already in your arm and a few hours of penicillin.

How the protocol actually runs

The full sequence is short and you'll mostly experience it as one extra item on the checklist at the 36-week visit and one extra IV line at admission to labor and delivery.

The antibiotic runs through the IV you'd have for fluids, blood draws, or epidural anyway. You can walk, eat ice chips, use the tub between doses, sit on a birthing ball, get the epidural. The drug stops with the placenta — there's nothing extra to take home, no postpartum course, no restriction on breastfeeding or skin-to-skin, no separation of mother and baby for monitoring beyond what the nursery does for every newborn. A scheduled cesarean before labor begins and before your water breaks doesn't need the antibiotic; if labor starts or your water breaks before the scheduled time, you go back to the standard rules.

Allergies and the one situation where you skip the antibiotic

There is no condition or medication that makes intrapartum prophylaxis itself a bad idea; what changes is which antibiotic is used.

The one situation where you don't get the antibiotic at all, even with a positive swab: a scheduled cesarean delivery that happens before labor starts and before your membranes rupture. The bacteria reach the baby through the birth canal or through ruptured amniotic membranes; if neither of those things has happened, there's nothing to interrupt. If you go into labor or your water breaks before the surgical time slot, the standard protocol resumes.

What people get wrong about a positive swab

A positive GBS swab is not a diagnosis, not an infection, not a moral verdict on your microbiome. About one in four healthy women carries it; carrying it tells you nothing about you that's wrong. It is also not an STI — sex is not how this got there and abstaining is not how it would leave. The bacterium lives in the gut and migrates to the perineum the way many normal gut bugs do.

The most expensive misreading is conflating early-onset and late-onset disease. The protocol on this page prevents the first one — infection in days 0 to 6, transmitted during labor — and has essentially no effect on the second one, infection in week 1 through about month 3, which seems to be acquired some other way (horizontal transmission, persistent maternal carriage as the baby's own antibody drops) Berardi et al. 2013, Nanduri et al. 2019. About half of all infant GBS disease in the era of universal screening is now late-onset. Adequate IAP doesn't make a baby invulnerable for three months; it closes the window from delivery through day six, which is the window with the highest attack rate.

The antibiotic is given to you, not to the baby. The baby will be observed and may have vital signs and a blood count checked depending on how the labor went, but a routinely well-appearing newborn whose mother got adequate IAP usually rooms in on the regular postpartum floor on a standard schedule Puopolo et al. 2019.

And: you can't pre-treat the colonization in pregnancy. A course of antibiotics at week 30 will not change anything by labor. The vagina recolonizes from the gut quickly. The only window where the antibiotic does anything for the baby is the labor itself, which is why the timing of the protocol is what it is.

Why the UK does this differently

The single most useful piece of context for understanding the GBS conversation is that the United Kingdom — a high-income country with excellent obstetric care and one of the world's better disease registries — has reviewed the evidence five times and declined to adopt universal screening. The UK runs a risk-based program instead: no antenatal swab, antibiotics in labor only if you have a specific risk factor (preterm birth, water broken longer than 18 hours, fever in labor, previous baby with GBS, or GBS found in your urine this pregnancy) RCOG 2017.

The case for risk-based is real. Universal screening means giving an antibiotic to about a quarter of all women in labor to prevent a relatively small absolute number of cases. The Cochrane review couldn't show a mortality benefit, the UK's absolute early-onset GBS rate is not catastrophic, and routinely exposing healthy newborns to intrapartum antibiotics has known costs on the developing gut microbiome (see failure modes) Seedat et al. 2016, Ohlsson and Shah 2014.

The case for universal screening — the US/Canadian/most-of-Europe position — is also real. Risk-based catches only about half to two-thirds of the babies who'd benefit from prophylaxis, because a substantial share of early-onset cases happen in colonized women without any other risk factor. The 2020 meta-analysis comparing the two approaches found about twice the early-onset incidence in risk-based programs Hasperhoven et al. 2020.

Neither side is reckless. If you're being cared for in a system that screens universally, the right answer is to participate in the program — the per-pregnancy math favors it. If you're in the UK or another risk-based system, you can still ask to be screened privately; some women do, and clinicians will generally honor it. There is also a maternal vaccine in late-stage trials that would, if it works, protect against both early- and late-onset disease and remove most of this conversation from the labor room Le Doare et al. 2017.

Where the protocol breaks, and the gut-bacteria question

The single most common way prophylaxis fails is timing. Labors don't run on schedule. About a third of women who are GBS-positive end up delivering less than four hours after the first antibiotic dose — the cohort cutoff where adequate prophylaxis kicks in — and that fraction is higher for second-and-later babies, whose labors are often fast. In those situations the baby may need closer observation in the first 24 hours; this isn't a treatment failure of the program, it's a known limit of how much of the labor window you can protect.

The other common gap is "unknown status": you got admitted before your swab came back, you saw a different practice, your records didn't transfer, or you're delivering preterm. The standard answer is to treat empirically based on risk factors (preterm, prolonged rupture, fever) or to run a rapid intrapartum PCR if your hospital has one.

For women with severe penicillin allergy, clindamycin works only if your particular GBS strain is susceptible to it — and roughly four in ten US strains are now resistant Cieslak et al. 2021. The fix is to ask, at the time of your 36-week swab, for clindamycin and erythromycin sensitivities to be run on a positive culture; if your strain is resistant, the labor agent shifts to vancomycin and you avoid showing up in labor with a plan that won't work.

Now the honest catch most pro-screening sources skip: the antibiotic perturbs the baby's gut bacteria. Babies exposed to intrapartum penicillin or ampicillin have lower Bifidobacterium and Bacteroides and higher Enterobacteriaceae for at least the first three to twelve months — the effect size is comparable in magnitude to delivery by cesarean Azad et al. 2016, Tapiainen et al. 2019. Whether this translates into later differences in atopy, eczema, obesity, or autoimmune disease is biologically plausible and not yet settled — the long-term cohorts haven't cleanly separated intrapartum antibiotics from other intrapartum exposures. The cost is real. It is also small relative to the avoided NICU course, and largely reversible: breastfeeding, vaginal exposure, and time recover most of the early gap. The right way to hold it is: a real cost in the margin, not a reason to skip prophylaxis for a known carrier.

What this buys you

The payoff for the mother is invisible. There's no felt change in pregnancy after a positive swab — the bacterium wasn't doing anything to you, and the protocol doesn't ask you to do anything until labor. The IV line in labor would have been there anyway. The antibiotic doesn't make labor longer or shorter, doesn't taste like anything, doesn't slow contractions, doesn't change what kind of birth you can have. It ends with the placenta.

The payoff is the call you don't get on day two. The follow-up appointment with the pediatric infectious-disease team that doesn't go on your calendar. The hearing test at six months and the developmental check-ins at eighteen months and three years that are routine, not gated by a meningitis admission. You will not get to know it didn't happen — there's no scoreboard. The cited evidence is the only reason it's visible at all: pre-program, the United States carried about seven thousand cases and four hundred newborn deaths a year from this; with screening and prophylaxis, those numbers are about one-seventh of what they were Nanduri et al. 2019.

What it doesn't buy: protection from infections after the first week. Late-onset disease — which now makes up about half of all infant GBS — is a separate problem the labor antibiotic doesn't touch Berardi et al. 2013. The warning signs in a previously well baby in the first three months — poor feeding, listless or floppy, fast breathing, fever or a strangely low temperature — deserve a same-day pediatric call no matter what your swab said.

Related

The maternal vaccine in late-stage trials — capsular-conjugate, given in pregnancy, designed to protect against both early- and late-onset disease — is the thing to watch over the next handful of years; it would change the whole picture Le Doare et al. 2017. Late-onset group B strep disease itself is a separate clinical problem with its own warning signs and its own evidence base. The broader question of how intrapartum antibiotic exposure affects long-term infant immune development sits alongside the closely-related questions of cesarean delivery and formula feeding; the signals are real and worth tracking, but disentangling them is genuinely hard. And if your reading here was triggered by an unwell newborn, not by a routine pregnancy: the warning signs in the first three months — poor feeding, unusual sleepiness, fast breathing, fever or low temperature — are not protocols, they're a same-day pediatric call.

·
627