A large Swedish pregnancy cohort is built to track microbiome shifts from early gestation to infancy
The Swedish Maternal Microbiome project, or SweMaMi, was designed to answer a question that obstetric care has struggled to capture in a systematic way: how the maternal microbiome changes across pregnancy and whether those changes track with adverse outcomes in mothers and newborns. The project combines longitudinal sampling, shotgun metagenomics, detailed questionnaires, and national registry linkage, creating a framework that can connect microbial biology with real-world obstetric and pediatric events.
Pregnancy is not a static biologic state. Hormonal changes, immune adaptation, diet, medication exposure, and shifting metabolic demands all alter the maternal environment, and the microbiome may move in parallel. The strength of SweMaMi lies in repeated sampling rather than a single snapshot. That design matters because microbial communities in the vagina, gut, and oral cavity can change quickly across gestation and again after delivery, when physiology and exposures shift abruptly.
Scale, sampling depth, and follow-up give the cohort unusual reach
Investigators enrolled pregnant women before 20 weeks of gestation in Sweden between 2017 and 2021. The final cohort included 5,439 pregnancies from 5,193 unique women. Baseline sampling was available for 3,973 pregnancies, representing 73% of the cohort, and 3,141 pregnancies, or 58%, contributed samples at all three timepoints: second trimester, third trimester, and postpartum.
The biological sampling burden was substantial. In total, the cohort generated 38,591 maternal microbiome samples from vaginal, faecal, and saliva sites, along with 3,109 infant faecal samples. Those repeated specimens are the real engine of the study. They allow analyses that can separate persistent microbial patterns from transient fluctuations and make it possible to compare maternal trajectories across pregnancy with early infant colonization.
One sentence captures the scale plainly: this is a longitudinal microbiome dataset, not a one-time specimen bank.
Questionnaires supplemented the biospecimens with information on general health, reproductive history, mental health, diet, and lifestyle. Registry linkage then extended the clinical lens beyond self-report, adding national health data and follow-up of offspring health up to age 10. For microbiome research, that linkage is unusually valuable. It gives the cohort the possibility of connecting microbial signatures to diagnoses and outcomes that matter at the bedside, rather than stopping at descriptive ecology.
Representation of the Swedish pregnant population was broad, with some expected skews
The cohort appears fairly representative of the broader Swedish pregnant population based on 2019 data. Forty-one percent of participants were first-time mothers, and the population distribution closely tracked the national obstetric setting. Some groups were slightly over-represented, including women with university-level education, women born in Sweden, those with normal body mass index, non-users of tobacco products, and women aged 30 to 34 years.
Those differences are not surprising in a study that depends on repeated sampling, questionnaire completion, and follow-up across pregnancy and postpartum. Cohort participation in reproductive health research often favors people with more stable access to care and higher engagement with longitudinal studies. For interpretation, the key point is not perfection of representativeness but breadth of capture across a large and clinically relevant population.
Microbial patterns may help explain why some pregnancies diverge from the expected course
SweMaMi was built to examine the dynamic associations between the microbiome and adverse maternal and neonatal events, including preterm birth and miscarriage. The planned analyses also include early-childhood disease. Those outcomes sit at the center of reproductive medicine because they represent points where biology, exposure, and host susceptibility intersect.
A microbiome framework is especially relevant in pregnancy because microbial communities can influence immune signaling, local inflammation, metabolic regulation, and barrier function. In the vagina, shifts in dominant bacterial species may alter the local ecological balance that protects against ascending infection. In the gut, community structure may shape systemic inflammatory tone and nutrient handling. In the oral cavity, microbial perturbations may mirror or contribute to inflammatory burden. None of these pathways stands alone, but together they offer plausible routes through which maternal microbial ecology could influence obstetric outcomes.
The study design also allows investigators to account for confounding that has complicated earlier microbiome work. Endometriosis, polycystic ovarian syndrome, diet, and drug use are all listed as factors to be examined alongside microbial measures. Those exposures can change both the microbiome and the risk of adverse outcomes, so separating correlation from causation will depend on careful modeling rather than simple comparison of groups.
Registry linkage and infant sampling extend the project beyond delivery
The inclusion of infant faecal samples pushes SweMaMi beyond maternal association studies. Early colonization patterns may reflect maternal microbial transfer, delivery mode, feeding practices, medication exposure, and the postnatal environment. By following offspring health through national registries, the cohort can examine whether early microbial patterns carry forward into later childhood outcomes.
This breadth matters because pregnancy research often stops at birth. Here, the investigators built a structure that can examine whether maternal microbiome signatures predict not only obstetric events but also infant health trajectories. That opens the door to testing whether the microbiome acts as a marker of risk, a mediator of risk, or both.
Sample and data collection were finalized in November 2021. The next phase focuses on characterizing microbial DNA and linking those results with questionnaire and registry information. The project is already one of the largest pregnancy cohorts in the world with longitudinal microbiome sampling, and the standardized collection strategy should support analyses that are hard to perform in smaller, more heterogeneous datasets.
What the cohort can answer, and what still depends on analysis
SweMaMi is not a results paper in the conventional sense; it is a cohort profile that describes a platform for future discovery. Its scientific value lies in the combination of size, repeat sampling, and linked health data. Those features create the possibility of studying how maternal microbial dynamics relate to pregnancy complications, neonatal outcomes, and early childhood disease with a level of temporal detail that is rarely available in reproductive research.
The study still leaves the central biologic questions open. Which microbial patterns precede adverse events? Which changes merely accompany them? How much of the observed variation reflects pregnancy physiology, and how much reflects modifiable exposures such as diet, tobacco use, or medication? Those questions can only be answered once the microbial sequencing data are layered onto the clinical and registry information already assembled.
For obstetric and neonatal medicine, the promise is practical. A cohort of this scale can identify candidate biomarkers, clarify timing relationships, and test whether maternal microbiome states track with complications that clinicians already watch for in routine care. The limiting factor now is not recruitment or biospecimen collection. The challenge is the analytic work of turning a vast, carefully built archive into clinically interpretable associations.