Pregnancy Loss and Later Risk of Atherosclerotic DiseaseClinical Perspective
Background—Pregnancy losses and atherosclerotic disease may be etiologically linked through underlying pathology. We examined whether miscarriage and stillbirth increase later risk of myocardial infarction, cerebral infarction, and renovascular hypertension.
Methods and Results—Among women pregnant at least once between 1977 and 2008, we identified a cohort of women with miscarriages, stillbirths, or live singleton births. These women were followed from the end of pregnancy for incident myocardial infarction, cerebral infarction, and renovascular hypertension. Using Poisson regression, we estimated incidence rate ratios for each of the outcomes by history of miscarriage and stillbirth. Among 1 031 279 women followed for >15 928 900 person-years, we identified 27 98 myocardial infarctions, 40 53 cerebral infarctions, and 1269 instances of renovascular hypertension. Women with stillbirths had 2.69 (95% confidence interval, 2.06–3.50), 1.74 (1.32–2.28), and 2.42 (1.59–3.69) times the rates of myocardial infarction, cerebral infarction, and renovascular hypertension, respectively, as women with no stillbirths. Compared with women with no miscarriages, women with miscarriages had 1.13 (1.03–1.24), 1.16 (1.07–1.25), and 1.20 (1.05–1.38) times the rates of these same outcomes, respectively; these associations were dose dependent, with each additional miscarriage increasing the rates of myocardial and cerebral infarction and renovascular hypertension by 9% (3% to 16%), 13% (7% to 19%), and 19% (9% to 30%), respectively. Associations were strongest in younger women (<35 years).
Conclusions—Pregnancy losses were associated with subsequent risks of myocardial infarction, cerebral infarction, and renovascular hypertension, consistent with either shared etiology or the initiation of pathological processes by a pregnancy loss leading to atherosclerosis.
- abortion, spontaneous
- cerebral infarction
- hypertension, renovascular
- myocardial infarction
Female sex is generally considered to be protective against atherosclerosis. However, underlying systemic inflammation and vascular pathology might link certain adverse pregnancy events with conditions that have atherosclerotic changes in the vascular bed in common (eg, myocardial and cerebral infarction and renovascular hypertension).1–3 Although the link between preeclampsia and ischemic heart disease is fairly well established,2 the evidence linking pregnancy loss and atherosclerotic disease, which could plausibly share pathological mechanisms independent of preeclampsia, is limited and more equivocal. An association between early pregnancy loss before first live birth and increased risk of ischemic heart disease has been reported,4 and a study linking endothelial dysfunction with adverse pregnancy events supplied a possible underlying mechanism.5 Recently, links have been reported between a parental history of ischemic heart disease and recurrent miscarriage before first live birth6 and between self-reported number of pregnancy losses and later risk of myocardial infarction.7 However, studies that also examined stroke risk were unable to demonstrate an association between pregnancy loss and stroke.4,7 Clear evidence of an association between pregnancy loss and major atherosclerotic manifestations would have important implications for later clinical follow-up of women who suffer pregnancy losses, as well as for understanding the pathogenesis of these conditions. To this end, we conducted a population-based cohort study with long-term follow-up of >1 million women to compare the rates of atherosclerotic outcomes in the heart (ischemic heart disease and myocardial infarction, in particular), the brain (cerebral infarction), and the kidney (renovascular hypertension) in women who had and had not experienced miscarriages and stillbirths.
Clinical Perspective on p 1782
Denmark has universal free public health care, and all hospital contacts are registered in national health registers. Since April 1, 1968, the Danish Civil Registration System8 has assigned a unique personal identification number to each Danish resident and regularly updates vital status and kinship information. The personal identification number is used in all of Denmark’s nationwide population-based registers and permits accurate linkage of individual-level information across registers.
The National Patient Register contains information from all Danish hospitals on inpatient diagnoses assigned since January 1, 1978, and outpatient diagnoses assigned from 1995 onward. Diagnoses are registered with the use of International Classification of Diseases codes, with International Classification of Diseases, Eighth Revision (ICD-8) codes used from 1978 to 1993 and International Classification of Diseases, Tenth Revision (ICD-10) codes used from 1994 onward.9 Registration of cerebral and myocardial infarction diagnoses has been validated previously. The National Patient Register has a sensitivity of 97% for myocardial infarction, and the positive predictive value of a registered myocardial infarction diagnosis is 93%, whereas the positive predictive value of a cerebral infarction diagnosis is 97% (sensitivity not reported).10,11
The Medical Birth Register contains detailed information on pregnancy outcomes, including molar pregnancies, missed abortions, spontaneous abortions, induced abortions, extrauterine pregnancies, stillbirths, and live births, from 1973 to the present. Smoking status at the beginning of pregnancy has been registered since 1991.12 From 1977 onward, parts of the register have been based on the National Patient Register, in which the positive predictive value of a spontaneous abortion diagnosis is 97%.13
The National Diabetes Register is based on hospital diabetes mellitus diagnoses, filled prescriptions for diabetes mellitus treatment, primary care measurements of blood glucose, and diabetic chiropody. The register contains nationwide information on diabetes mellitus diagnoses from 1991 onward. Validation of diabetes mellitus diagnoses showed a sensitivity of 96% and a positive predictive value of 89%.14
The Register of Medicinal Product Statistics allows tracking of individual prescription medication histories from 1994 onward. Dispensed medications are identified by Anatomic Therapeutic Chemical (ATC) classification codes.15
All females aged ≥12 years with at least 1 pregnancy ending in a live singleton birth, miscarriage, or stillbirth registered in the Medical Birth Register between 1977 and 2008 were identified. Women free of any cardiovascular disease (ICD-8 codes 390.00–429.99 or ICD-10 codes I00–I52 registered in the National Patient Register) at the time of first registered pregnancy outcome were included in the study cohort. For our purposes, miscarriage included isolated missed abortions and spontaneous abortions; miscarriages registered within 8 weeks of molar pregnancies, induced abortions (surgical or medical), and extrauterine pregnancies were ignored in the analysis. Exposure to pregnancy loss was considered a time-dependent variable. A woman whose first registered pregnancy outcome was a live birth was considered unexposed until such time as she experienced a pregnancy loss (if ever), at which point her exposure status changed to exposed. A woman whose first registered pregnancy outcome was a miscarriage/stillbirth was considered exposed from the start of follow-up. Once exposed, a woman could not revert to unexposed. Numbers of miscarriages, live births, and stillbirths were likewise tracked as time-dependent variables.
Myocardial infarctions (ICD-8 code 410 and ICD-10 codes I21–I22), cerebral infarctions (ICD-8 code 436 and ICD-10 codes I63–I64), and renovascular hypertension (ICD-8 code 400.39 and ICD-10 codes I12–I13 and I15.0) were identified with the use of the National Patient Register.
We considered number of live births, attained age, and calendar period as potential confounders in all analyses. In addition, we identified women with atrial fibrillation (ICD-8 codes 427.93 and 427.94 and ICD-10 code I48) using the National Patient Register and women with diabetes mellitus using the National Diabetes Register. Medical treatments were used as proxies for cardiovascular risk; as a proxy for hyperlipidemia, we used the Register of Medicinal Product Statistics to identify women with elevated blood lipid levels, defined as having filled ≥2 prescriptions for any lipid-lowering medication (ATC code C01). As proxies for arterial hypertension and other risk factors for cardiovascular disease, we identified women who had filled ≥2 prescriptions for the following medications: β-blockers (ATC code C07), calcium antagonists (ATC code C08), diuretics (ATC code C03), angiotensin-converting enzyme inhibitors, and angiotensin receptor blockers (ATC code C09). Smoking status was obtained from the Medical Birth Register.
The study period was from 1977 to 2008, and females aged ≥12 years were followed from the first registered live singleton birth, miscarriage, or stillbirth until the first of the following events: (1) atherosclerotic outcome; (2) death; (3) emigration; (4) designated missing in the Danish Civil Registration System; or (5) December 31, 2008 (end of follow-up). Using log-linear Poisson regression, we estimated incidence rate ratios (IRRs) separately for each of the 3 outcomes by a woman’s history of miscarriage and stillbirth. Exposure and confounding variables were treated as time-dependent variables. All analyses were adjusted for number of live births (0, 1, 2, and ≥3), age (12–29 years, 3-year age groups from 30–59 years, ≥60 years), and calendar period (1977–1993, 1994–1999, 2000–2004, 2005–2008), with additional adjustment for the potential confounders in subanalyses. In analyses including certain potential confounders, follow-up started in 1991 (smoking and diabetes mellitus) and 1994 (medication use), depending on when registration of the confounder of interest began. In some women, confounder information was missing (most often smoking from the Medical Birth Register); these women were excluded, as appropriate, from the subanalyses starting in 1991 and 1994. In the analysis of cerebral infarction risk, we conducted a subanalysis censoring women on diagnosis of atrial fibrillation (if any) because the risk of cerebral infarction is known to increase with atrial fibrillation. Significance testing was performed with the use of likelihood ratio tests. IRRs for each additional miscarriage (live birth) were estimated by assigning the categories 0, 1, 2, 3, and ≥4 miscarriages (live birth) the scores 0, 1, 2, 3, or 4 and estimating a linear effect (ie, linear trend). Tests for departure from linearity were performed before estimating a linear trend. IRRs for atherosclerotic disease for each additional miscarriage by age (ie, interaction) were modeled with the use of restricted cubic splines with knots at ages 24, 32, 37, 42, and 52 years; the knots were chosen such that person-time was evenly distributed across the age intervals. All statistical analyses were performed with the use of SAS version 9.2 (SAS Institute Inc, Cary, NC) and STATA version 12.1 (StataCorp, College Station, TX).
Our cohort included 1 031 279 women; depending on the outcome, the cohort was followed for at least 15 928 934 person-years. By the end of follow up, 842 616 women had had only live births (no pregnancy losses registered), 8191 women had had ≥1 stillbirths, and 188 663 women had had ≥1 miscarriages; the latter 2 groups could overlap (ie, a woman could appear in both groups). One miscarriage was registered for 151 880 women, and 28 398 and 5979 women had 2 and 3 miscarriages, respectively; ≥4 miscarriages were registered for 2406 women. One stillbirth was registered in 8078 women, and 113 women experienced ≥2 stillbirths. We identified 2798 myocardial infarctions, 4053 cerebral infarctions, and 1269 diagnoses of renovascular hypertension during follow-up. Table 1 shows event numbers and years of follow-up by history of miscarriage and stillbirth, years since latest miscarriage, current age, calendar period, and potential confounders.
Associations With Stillbirth and Miscarriage
When we compared women with ≥1 stillbirths with women who never had a stillbirth, the IRRs were 2.69 (95% confidence interval [CI], 2.06–3.50) for myocardial infarction, 1.74 (95% CI, 1.32–2.28) for cerebral infarction, and 2.42 (95% CI, 1.59–3.69) for renovascular hypertension. The corresponding IRRs for ≥1 miscarriages, compared with no miscarriages, were 1.13 (95% CI, 1.03–1.24), 1.16 (95% CI, 1.07–1.25), and 1.20 (95% CI, 1.05–1.38). The rates for all 3 outcomes increased with increasing number of miscarriages (Table 2); each additional miscarriage increased the rate of the outcomes of interest by 9% to 20%, depending on the outcome (Table 3). The estimates for cerebral infarction were not altered when women with an atrial fibrillation diagnosis were censored (data not shown). There were too few stillbirths to perform the corresponding burden-of-loss analyses for stillbirths.
Age and Number of Live Births
The rates of myocardial infarction, cerebral infarction, and renovascular hypertension in women aged <35 years increased by 35% to 55% per additional miscarriage, whereas in those aged ≥35 years, the rates increased by only 6% to 7% (Table 3). To further illustrate the effect of age on the associations between miscarriage and these 3 outcomes, we present IRRs showing the effect of each additional miscarriage on rates of myocardial infarction, cerebral infarction, and renovascular hypertension, by age, in the Figure. For the sake of comparison, we also examined associations between live births and these 3 outcomes (Figure I in the online-only Data Supplement). For each additional live birth, there was a corresponding increase in the rates of myocardial and cerebral infarction; however, these associations were observed only in the youngest women and were of much lesser magnitude than the associations with miscarriages.
Years Since Latest Miscarriage
We did not observe excess rates of any of the outcomes in the year after miscarriage. Using the rate of myocardial infarction in the first year after a miscarriage as the reference, we observed that the rate increased with time since miscarriage, with the strongest effect ≥5 years after the most recent miscarriage. For both cerebral infarction and renovascular hypertension, the rate did not differ by time since most recent miscarriage (Table 2).
IRRs per additional miscarriage in women who had only had miscarriages (consecutive miscarriages) and in women whose miscarriages were interspersed with live births are presented in Table 3. The IRRs for cerebral infarction and renovascular hypertension were greater for women with consecutive miscarriages.
Available potential confounders included diabetes mellitus, cardiovascular medication, and ever smoking. Table I in the online-only Data Supplement shows distributions of events and years of follow-up by history of pregnancy losses for these potential confounders. Adjustment for these variables did not affect our results (Tables II and III in the online-only Data Supplement).
In our population-based cohort of >1 million women, stillbirths and miscarriages were associated with increased rates of myocardial infarction, cerebral infarction, and renovascular hypertension. Overall, stillbirths were associated with greater risks than miscarriages, although in women with a history of ≥4 miscarriages, the rate increases were similar. The increase in rate associated with a history of miscarriage was of similar magnitude for all 3 atherosclerotic conditions, with increasing numbers of miscarriages producing corresponding increases in rate; unfortunately, there were too few stillbirths to allow for corresponding analyses for stillbirths. Consecutive miscarriages yielded greater IRRs than nonconsecutive miscarriages for cerebral infarction and renovascular hypertension, although the differences were only borderline significant. The associations were robust to adjustment for potential confounders.
Recurrent pregnancy losses before first live birth have been associated previously with an increased risk of ischemic heart disease or myocardial infarction in the woman4 and in her parents.6 The large (>10 000 women) European Prospective Investigation Into Cancer and Nutrition–Heidelberg cohort also reported significant associations between pregnancy loss and myocardial infarction but no significant associations with stroke, although this study had only a limited number of events and conclusions were limited by a self-reported history of pregnancy loss and a participation rate of 40%.7
Our findings suggest an etiologic link between pregnancy loss and atherosclerotic outcomes, with these disparate events connected either by pathological processes initiated by a miscarriage or stillbirth that then lead to atherosclerosis or by common underlying mechanisms, possibly with genetic underpinnings. The finding that pregnancy losses are associated with later atherosclerotic events in 3 different organs indicates that any underlying common pathology is unlikely to be organ specific. The immune system and inflammatory mechanisms, including Th1-type inflammatory responses, appear to play a role in miscarriage, suggesting that pregnancy loss could resemble an autoimmune condition16,17; several autoimmune conditions (inflammatory bowel disease, psoriasis, and rheumatoid arthritis) have recently been associated with an increased risk of atherosclerosis.18–20 Atherosclerotic plaques on the endothelium contain active immune cells, including activated T cells and macrophages, along with signaling molecules promoting a Th1 response21; endothelial dysfunction has been associated with previous miscarriages.5 Consequently, inflammatory processes are good candidates for underlying, genetically influenced mechanisms common to both pregnancy loss and atherosclerotic conditions.
Whether the mechanisms underlying the associations observed for miscarriage and stillbirth overlap or whether separate processes are at play is unclear. Although there are distinct causes of early and late pregnancy loss, there are also factors known to increase the risk of fetal loss throughout gestation (eg, fetal genetic abnormalities and birth defects, maternal infection, maternal diabetes, and other chronic diseases).22 Of particular relevance, given the possible mechanism outlined above, endothelial dysfunction produced by systemic inflammatory processes is a plausible underlying mechanism common to miscarriage (late and recurrent miscarriage in particular), stillbirth, and the 3 study outcomes. The resulting vascular pathology would be expected to contribute both to poor placentation during pregnancy, resulting in losses, and to an increased likelihood of myocardial infarction, cerebral infarction, and renovascular hypertension. Under this scenario, the stronger overall associations observed for stillbirth (74% to 169% increases in risk associated with stillbirth versus the 13% to 20% increases associated with miscarriage) could be explained by the fact that placental problems comprise a bigger proportion of stillbirths than miscarriages. Because the biggest contributors to miscarriage overall are fetal genetic abnormalities, infection, and maternal uterine and reproductive tract abnormalities,22 none of which is likely to be associated with later maternal atherosclerotic outcomes, the overall miscarriage estimates were subject to more noise from miscarriages whose causes were unrelated to the outcomes. However, miscarriages due to fetal genetic abnormalities and infection are likely to be sporadic events, whereas those due to vascular pathology are likely to repeat. Consistent with this assumption, we observed stronger associations for recurrent miscarriage, suggesting that women with late or recurrent miscarriages are likely to have driven the overall associations.
Although vascular pathology provides a plausible link between pregnancy losses and the atherosclerotic outcomes we considered, it is difficult to determine whether preeclampsia, a known risk factor for later ischemic heart disease and cerebral infarction,2 is causally involved or whether miscarriage or stillbirth is independently linked to the later events. Preeclampsia is known to recur,23 such that very early preeclampsia or its antecedents cannot be ruled out as a cause of, for example, myocardial infarction associated with recurrent miscarriage. However, by definition miscarriage occurs before preeclampsia can be diagnosed, and therefore this is difficult to disentangle, particularly in a register-based study. Likewise, preeclampsia almost certainly contributed to some of the stillbirths in the cohort, but there were too few stillbirths to allow for categorization by presumed etiology.
Separate processes might also contribute to the associations with miscarriage and stillbirth. Hormonal factors and abnormal maternal immune responses could plausibly link miscarriage but not stillbirth and the outcomes, whereas high maternal blood pressure might provide a link for stillbirth but not miscarriage. However, such discussions are purely speculative and would require clinical data not available in our registers to confirm or refute.
Inherited forms of myocardial infarction are well known to present at younger ages than nonfamilial forms. Consequently, the recent finding of Smith and colleagues6 of an association between a parental history of ischemic heart disease and recurrent miscarriage in daughters (an association that was robust to adjustment for socioeconomic deprivation and smoking among other potential confounders) is of particular interest, suggesting as it does a possible genetic link between the 2 conditions. Previous studies of the associations between pregnancy losses and risk of ischemic heart disease and stroke have had limited possibilities for examining these associations specifically among younger women4,7,24; examining 3 different atherosclerotic end points, we found the strongest associations with miscarriages in the youngest women. Taken together with the findings of Smith and colleagues, our findings support a genetic component to the link between miscarriage and atherosclerotic disease and suggest that previous pregnancy losses might be considered when one evaluates risk of atherosclerotic disease, at least in younger women.
We compared consecutive and nonconsecutive miscarriages because recurrent consecutive miscarriages can differ etiologically from miscarriages interspersed between successful pregnancies. For cerebral infarction and renovascular hypertension, we observed a tendency toward stronger associations for consecutive miscarriages than for nonconsecutive miscarriages. In contrast, associations between miscarriage and myocardial infarction were similar for consecutive and nonconsecutive miscarriages. Consequently, our results do not completely support the idea that only consecutive miscarriages are associated with increased rates of later atherosclerotic outcomes.
Strengths and Limitations
The major strengths of our study were the unparalleled size of the cohort and duration of follow-up, with an average of 15 years of follow-up per woman in a cohort of >1 million women. The completeness of follow-up was ensured by use of regularly updated Danish national registers.8,9 The completeness of registration and validity of spontaneous abortion, myocardial infarction, and cerebral infarction diagnoses are excellent, minimizing the likelihood of misclassification.10,11,13 Furthermore, we used a conservative algorithm to identify miscarriages and excluded all miscarriages with coregistration of molar pregnancy, induced abortion, and extrauterine pregnancy. Unfortunately, information on timing of miscarriage (in gestational weeks) was not available, making it impossible to examine early and late miscarriages separately.
Potential confounders of the observed associations include number of live births, alcohol consumption, hypertension, smoking, diabetes mellitus, socioeconomic status, and body mass index/overweight.17,25 Information on many of these variables was unavailable at a population level; however, we were able to adjust our analyses for several of these factors and observed no changes in either the pattern or the strength of the observed associations. Residual confounding may nevertheless influence our reported associations, but it is reassuring that previous studies that had the ability to adjust for factors such as education, weight, social class, and birth characteristics found that adjustment had no major impact on their results.4,7 In addition, the consistent pattern across different atherosclerotic conditions and the dose-response increase in risk with increasing numbers of miscarriages make it unlikely that unmeasured confounding could account for the associations we report.
We considered several other atherosclerotic end points for this study, including cerebral hemorrhage, unstable angina, and percutaneous coronary interventions. However, although ≈10% of cerebrovascular events are hemorrhagic rather than thrombotic, we opted not to include cerebral hemorrhage as an end point because there was insufficient information available in the National Patient Register to differentiate relevant hemorrhagic events (those caused by hypertension and thus related to atherosclerosis) from those caused by trauma and vascular malformations (and thus not atherosclerotic in origin). We chose not to use unstable angina as an end point because it is likely that registration of this condition is at best incomplete. However, analyses using the broader end point ischemic heart disease, which included unstable angina, yielded results similar to those obtained from the analysis of myocardial infarction only, despite concerns about the validity of a proportion of registered ischemic heart disease diagnoses. Percutaneous coronary interventions would also have been obvious outcomes for this study, but information on percutaneous procedures was only available from 2000 onward.
Although the results of the analyses treating number of miscarriages as a categorical variable strongly support a dose-response relationship between number of miscarriages and the rates of all 3 outcomes, single estimates depart from the pattern of increasing rates with increasing numbers of miscarriages. However, tests for departure from linearity indicated that it was appropriate to fit a linear model, and analyses of the effect of each additional miscarriage showed a highly significant trend. It is likely that occasional departures from the overall linear trend were simply due to the fact that the numbers of women in certain miscarriage groups were small. However, we cannot rule out that repeated miscarriages (≥3) are even more strongly associated with the outcomes than our estimates suggest (ie, that the dose-response pattern is not linear but curves more sharply upward).
It could be argued that the observed associations were due simply to number of pregnancies rather than to pregnancy losses in particular. However, the observed associations between number of live births and myocardial infarction, cerebral infarction, and renovascular hypertension were much smaller in magnitude than those for stillbirths and miscarriages, indicating that our findings cannot simply be ascribed to the effect of parity.
Miscarriage and stillbirth are associated with an increased risk of myocardial infarction, cerebral infarction, and renovascular hypertension, conditions that have atherosclerotic changes in the vascular bed as a common feature, making pregnancy loss a possible candidate risk factor for atherosclerotic disease in women. Our findings support either a shared etiology (likely with a genetic component) for pregnancy loss and atherosclerotic conditions or the initiation of pathological processes by a miscarriage or stillbirth that then lead to these conditions. In either case, inflammatory processes might be the common denominator.
Sources of Funding
This work was funded by grants from the Danish Heart Foundation (grant number 11-04-R84-A3387-22642) and from the A.P. Møller Foundation for the Advancement of Medical Sciences. The sponsors were not involved in the design and conduct of the study; collection, management, analysis, and interpretation of the data; or preparation, review, and approval of the manuscript.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.112.000285/-/DC1.
- Received December 25, 2012.
- Accepted March 20, 2013.
- © 2013 American Heart Association, Inc.
- Fraser A,
- Nelson SM,
- Macdonald-Wallis C,
- Cherry L,
- Butler E,
- Sattar N,
- Lawlor DA
- Calleja-Agius J,
- Jauniaux E,
- Pizzey AR,
- Muttukrishna S
- Smith GC,
- Pell JP,
- Walsh D
- Kharazmi E,
- Dossus L,
- Rohrmann S,
- Kaaks R
- Carstensen B,
- Kristensen JK,
- Marcussen MM,
- Borch-Johnsen K
- Wallach Kildemoes H,
- Toft Sørensen H,
- Hallas J
- Rungoe C,
- Basit S,
- Ranthe MF,
- Wohlfahrt J,
- Langholz E,
- Jess T
- Lindhardsen J,
- Ahlehoff O,
- Gislason GH,
- Madsen OR,
- Olesen JB,
- Torp-Pedersen C,
- Hansen PR
Pregnancy losses and atherosclerotic disease may be etiologically linked through underlying vascular pathology. In a population-based cohort of >1 million women, we found increased rates of myocardial infarction, cerebral infarction, and renovascular hypertension in women with previous stillbirths or miscarriages. Given a history of only 1 miscarriage, the risk was increased 10%, and we observed increasing risks with more miscarriages; risks for all outcomes increased 10% to 20% with each additional miscarriage. Given a previous stillbirth or ≥4 miscarriages, relative risks were increased ≥2-fold for all 3 outcomes. Associations were even stronger in the youngest women. Our findings support either a shared etiology (likely with a genetic component) for pregnancy loss and later atherosclerotic conditions or the initiation of pathological processes by a miscarriage or stillbirth that then lead to these conditions. In either case, inflammatory processes might be the common denominator. Risks conferred by 1 or 2 miscarriages were statistically significant but too small to be of clinical relevance. Women with 2 or 3 miscarriages had markedly increased risk in the range of 1.1- to 1.8-fold. Given ≥4 miscarriages or a previous stillbirth, risks were 2- to 4-fold for all outcomes, suggesting that these kinds of events should be considered when one evaluates a woman’s future risk of atherosclerotic events.