Pregnancy complications and later heart disease: 51 of 72 links hold

A woman leaves the postnatal ward after a pregnancy that included high blood pressure, or gestational diabetes, or a stillbirth. The discharge letter covers the baby, the stitches and the follow-up visit in six weeks. What it may not cover is the question that arrives later, sometimes years later: does what happened in that pregnancy say anything about her own heart, kidneys or blood sugar a decade on?

A paper in BMJ Medicine, published online on 6 October 2026 by Steven Wambua, Krishnarajah Nirantharakumar and colleagues, addresses that question across ten pregnancy complications, in the latest and in earlier pregnancies. Its answer is yes for many of them, with a few surprises, and with limits that the authors' own design makes plain.

The team used the Clinical Practice Research Datalink Aurum primary-care database and its pregnancy register in England, covering 1 January 2000 to 31 December 2022. They included 1,432,795 women aged 15 to 49 who had been pregnant, with or without complications, and who were registered with a general practitioner. The study is longitudinal and observational: nobody was assigned to anything, and the researchers followed what was already written in the records.

The outcomes were the first recorded diagnosis of four conditions. Cardiovascular disease meant ischaemic heart disease or cerebrovascular disease, including stroke and transient ischaemic attack. The other three were type 2 diabetes, hypertension and chronic kidney disease at stages 3 to 5. Because a woman already diagnosed with one condition cannot develop it again, each outcome had its own cohort: 1,422,287 women for cardiovascular disease, 1,389,460 for hypertension, 1,411,883 for type 2 diabetes and 1,430,965 for chronic kidney disease.

The exposures were pregnancy complications, looked at both in the latest pregnancy and in previous ones. Rates were compared using hazard ratios, which describe how much faster the first diagnosis arrived in one group than in another. The hazard ratios were adjusted for established risk factors such as age, ethnic group, deprivation, blood pressure, body mass index and smoking. With four outcomes and many complications, the authors examined 72 associations in all and corrected for multiple testing, which guards against the chance findings that appear when many comparisons are run.

Median follow-up was 4.1 to 4.3 years across the four cohorts, with an interquartile range of roughly 1.5 to 9.8 years, so many women were watched for a short time and some for much longer. Of the 72 associations, 51 were significant after correction.

Some results confirmed what clinicians already expect. Gestational hypertension and pre-eclampsia more than tripled the rate of hypertension and roughly doubled the rate of chronic kidney disease, with smaller increases for cardiovascular disease and type 2 diabetes. Gestational diabetes was associated with a sevenfold higher rate of type 2 diabetes and with higher rates of hypertension and cardiovascular disease, but not of chronic kidney disease.

Others were less established. Postnatal depression in the latest pregnancy was associated with a 35% to 56% higher rate of all four outcomes: hypertension (hazard ratio 1.36, 95% confidence interval 1.33 to 1.40), type 2 diabetes (1.41, 1.36 to 1.46), cardiovascular disease (1.56, 1.50 to 1.62) and chronic kidney disease (1.35, 1.25 to 1.45). Miscarriage, stillbirth, placental abruption and preterm birth, in the latest or an earlier pregnancy, were linked to modestly higher rates of at least one outcome. Stillbirth in the latest pregnancy, for example, carried a hazard ratio of 1.67 (1.33 to 2.08) for type 2 diabetes and 1.79 (1.19 to 2.70) for chronic kidney disease, and placental abruption carried 1.67 (1.06 to 2.62) for chronic kidney disease.

Small-for-gestational-age birth was associated with higher rates of hypertension and cardiovascular disease. Large-for-gestational-age birth was associated with higher rates of type 2 diabetes, including in women who had not had gestational diabetes.

A hazard ratio is a relative measure. A ratio of 1.56 says the rate was about half again as high in one group, but it does not say how many women that is out of a hundred. The abstract does not give absolute risks for any complication. The full paper reports that, over a median of about four years, 3.5% of women in the cohort developed hypertension, 1.5% type 2 diabetes, 1.1% cardiovascular disease and 0.4% chronic kidney disease. For conditions that are uncommon in women under 50, a large relative increase can still be a small number of actual diagnoses. The size of the relative increase and the size of the personal risk are different quantities, and this paper's summary reports only the first.

The second caveat is time. With a median follow-up near four years, the study mostly captures the early years after pregnancy. It says less about what happens at 20 or 30 years, when heart disease and kidney disease are far more common.

The third is that the data come from primary-care records linked to hospital data. A condition is counted when it is recorded, so a woman who sees her doctor often is more likely to be diagnosed, and one who rarely does may be missed. Women who have complicated pregnancies also tend to have more contact with services afterwards, which could inflate the apparent links. Postnatal depression raises a similar question: whether depression itself, the circumstances behind it, or shared underlying health explains the pattern, the design cannot separate.

The value of this work is its breadth. Looking at many complications against four outcomes in one population shows which links are large and consistent, and which, like placental abruption and kidney disease, rest on wide confidence intervals that run close to no effect at all.

The authors conclude that their findings support taking a full reproductive history and using the postpartum period as an opportunity for risk assessment and preventive cardioprotective strategies in women with a history of adverse pregnancy outcomes. That is their recommendation to health systems, and it follows from the associations; the study did not test whether such assessment improves anyone's outcomes. What the data do show is that associations with later diagnoses extend to complications in pregnancies earlier than the most recent one.

This is a paper to read for its map rather than for any single number. Its strength is scale and scope: more than 1.4 million women, many complications, four outcomes, and a correction for the fact that 72 comparisons will throw up chance results; 51 surviving that correction is meaningful. The well-known links, hypertensive and diabetic pregnancy to later disease, behave as expected, which is reassuring about the data. The less familiar ones, postnatal depression above all, deserve attention but also caution, because depression, health service contact and diagnosis can all feed each other in a primary-care record. The honest limits are short follow-up, coded diagnoses and the absence of absolute risks in the abstract. The authors' call for using the postpartum period for risk assessment is a reasonable inference from association, not a result the study tested, and a relative increase in a rare outcome in women under 50 does not by itself describe any one woman's own risk.