PFAS and Miscarriage: What the Norwegian Study Found

A large Norwegian study found no measurable increase in miscarriage odds with higher PFAS exposure. That is the best direct evidence available for sporadic miscarriage at the exposure levels represented in this cohort. It also does not settle whether PFAS affect reproductive health, because the same chemicals have measurable associations with sex-steroid hormones, and the study measured exposure after the miscarriage it sought to explain.

PFAS are a large family of persistent industrial chemicals used in products designed to repel water, grease, or stains. Their environmental persistence has earned them the nickname "forever chemicals," which is more memorable than reassuring. For people planning pregnancy, the question is narrower: do these exposures raise the risk of miscarriage?

The honest answer is that human evidence remains mixed. The strongest recent prospective study found a null result. Studies of recurrent pregnancy loss in different populations have reported positive associations. Hormone studies give the concern a plausible biological basis, but they cannot prove that a hormone change causes a miscarriage.

The Norwegian result was null, with useful precision

Grindstad et al.'s 2026 Norwegian Mother, Father and Child Cohort Study analysis used a nested case-control design in 2,525 nulliparous women enrolled between 1999 and 2008. It included 253 women with a recent miscarriage and 2,272 controls. Researchers measured seven PFAS compounds: PFHxS, PFHpS, PFOS, PFOA, PFNA, PFDA, and PFUnDA.

For the joint PFAS mixture, each quartile increase corresponded to an odds ratio of 0.95 for miscarriage, with a 95% confidence interval of 0.83 to 1.09. The interval includes 1.0, so the result was not statistically significant. It does not show protection, either. It indicates that this analysis found no detectable increase in miscarriage odds and makes a large increase per quartile unlikely in this population.

Adjusted single-compound models also found no clear associations for the seven chemicals. Additional adjustment for body mass index, menstrual-cycle characteristics, and oral-contraceptive use did not change the conclusion. The publication's available results do not provide compound-specific odds ratios and confidence intervals, so readers cannot inspect whether any individual compound showed a small imprecise signal beneath the overall null result.

This is meaningful evidence, not a result to wave away because PFAS are unpopular chemicals. Prospective blood collection, a relatively large sample for this literature, mixture analysis, and covariate adjustment give the study real weight.

Another large cohort does not supply a clear contrary answer

The Norwegian finding is not alone. A 2025 analysis from the Japan Environment and Children's Study examined exposure during early pregnancy and late miscarriage risk and reported inconclusive results. Its endpoint was late miscarriage, not the first-trimester losses often meant by "miscarriage," so it cannot be treated as an exact replication of MoBa. Still, both large prospective cohorts fail to show a consistent positive association in their respective settings.

That pattern fits earlier reviews. A 2021 systematic review and meta-analysis described PFAS associations with adverse pregnancy and birth outcomes as inconsistent. A review focused on early pregnancy loss called the evidence controversial. Inconsistency does not prove safety; it argues against a strong, uniform effect that appears across populations, exposure profiles, and outcome definitions.

Why the null result is not a clean exoneration

The key design problem is timing. MoBa measured PFAS during the pregnancy in which women enrolled, while the "recent miscarriage" had occurred before that pregnancy, potentially months or more than two years earlier. PFAS concentrations reflect long-term body burden, but they can still change as environmental and dietary exposures change. A current measurement is therefore an imperfect stand-in for exposure around conception and implantation in the earlier pregnancy.

Pregnancy also expands maternal plasma volume, which can dilute circulating analytes. A sample collected during an ongoing pregnancy may underestimate the concentration relevant to the earlier periconceptional period. Both the timing gap and dilution can blur exposure differences between women, which generally makes an association harder to detect.

Selection may matter as well. To become a MoBa case, a woman had to experience a miscarriage, conceive again, and enroll during a viable subsequent pregnancy. If a high PFAS burden affected both pregnancy maintenance and the ability to conceive again, the study would miss women who did not reach that later pregnancy. The enrolled cases could then be a more resilient subset than all women who had miscarried.

The analysis also focused on nulliparous women, which reduces confounding related to prior births but limits its relevance to recurrent pregnancy loss. Recurrent loss has different immunological, coagulation, and clinical determinants than a single sporadic loss. A null result for one phenotype should not be stretched into a verdict on the other.

The positive studies address a different clinical and exposure context

Nian et al. in 2022 reported positive associations between legacy PFAS and unexplained recurrent spontaneous abortion in a Chinese cohort. That finding does not directly contradict MoBa. It concerns recurrent, unexplained loss rather than sporadic recent miscarriage, and the population likely had a different PFAS mixture and concentration range.

A 2025 environmental-wide association study in Northwest China also examined PFAS mixtures and pregnancy loss. Its approach is useful for detecting possible signals but is hypothesis-generating; it cannot establish that findings apply to Norwegian or Japanese populations without comparable exposure data. These studies raise a reasonable hypothesis that risk, if present, could depend on dose, chemical profile, or susceptibility. They do not establish a population-wide causal effect.

A Swedish cohort study examined PFAS exposure in early pregnancy and sporadic first-trimester miscarriage and discussed placental dysfunction and disrupted endometrium-blastocyst signaling as plausible pathways. The available quantitative information does not permit a reliable characterization of its effect estimates, so the Swedish study should not be counted as evidence for either direction here.

The endocrine evidence is real, but it is not a miscarriage trial

PFAS can affect reproductive biology in humans. A 2024 pregnancy study by Pacyga et al. found associations between PFAS concentrations and maternal early-second-trimester sex-steroid hormones, including progesterone. Progesterone supports the endometrium and early pregnancy, making this association biologically relevant. But a hormone measurement in the second trimester cannot show that PFAS caused an earlier miscarriage.

A 2025 analysis of U.S. women likewise associated PFAS exposure with hormones involved in hypothalamic-pituitary-gonadal axis function, including FSH, estradiol, progesterone, and AMH. These are human biomarker associations, not proof of a clinical outcome.

Other pregnancy endpoints add context without answering the miscarriage question. In the MADRES cohort, detected prenatal PFOA exposure was associated with smaller fetal head biometric parameters among participants reporting higher perceived stress. That study involved 335 prenatal specimens and fetal-growth measurements, not pregnancy loss. Reviews and cohort work have also linked PFAS exposure with hypertensive disorders of pregnancy, including an association between prenatal PFOA and gestational hypertension in an Atlanta African American cohort. Those findings support concern about placental and vascular biology, but they do not convert an association with hypertension into evidence that PFAS cause miscarriage.

What is reasonable to do before pregnancy

Long-chain PFAS levels in maternal plasma have declined after U.S. regulatory phase-outs, according to a 2025 study of maternal plasma trends. Replacement and short-chain PFAS remain a concern because their reproductive toxicity profile is less certain. The available evidence does not identify a safe reproductive threshold, and it does not support detox regimens, supplements, or chelation as ways to improve pregnancy outcomes.

Reducing future intake is more defensible than trying to rapidly clear an established body burden after conception. Practical steps include using a water filter designed to remove PFAS, such as reverse osmosis or an NSF/ANSI 58-certified activated-carbon-block system; reducing contact with PFAS-treated food packaging; and replacing damaged or PFAS-coated cookware when practical. Research in pregnant U.S. women links reported food consumption patterns with serum PFAS concentrations, supporting food as one exposure route.

The current evidence does not demonstrate that PFAS raise sporadic miscarriage risk at the exposure levels measured in large Norwegian and Japanese prospective cohorts. It does show enough endocrine and pregnancy-health concern to make low-cost exposure reduction sensible before conception, without pretending that filters or "detox" products can erase past exposure.