Jump to
Abstract
Pregnancy characteristics are associated with risk of cardiovascular diseases, but their independent associations with hypertension or blood pressure (BP) levels remain uncertain. We linked the Swedish Medical Birth Register with Västerbotten Intervention Program data (Northern Sweden). Using linear and logistic regression, we related pregnancy factors in any prior pregnancy with BP and hypertension at 40 years of age in 15 896 parous women free of prepregnancy hypertension. Pregnancy factors included parity, age at first delivery, preeclampsia, gestational diabetes mellitus, placental abruption, shortest gestational age small for gestational age baby (<third percentile for birth weight) or stillbirth. We defined hypertension as systolic BP ≥140 mm Hg and diastolic BP ≥90 mm Hg or antihypertensive use. Multivariable models were adjusted for all pregnancy factors and potential lifestyle and sociodemographic confounders. At 40 years of age, 1535 women (9.6%) had hypertension. In multivariable models, lower parity, younger age at first birth, preeclampsia, small for gestational age, and placental abruption were independently associated with higher systolic and diastolic BP levels at 40 years of age. Younger age at first birth, preeclampsia, gestational age <32 versus ≥37 weeks, and small for gestational age were independently associated with hypertension. Our findings raise the possibility that earlier and more frequent BP screening may be desirable in women with these pregnancy characteristics.
Introduction
Hypertension is a leading risk factor for myocardial infarction, stroke, and congestive heart failure among women.1 Maternal characteristics and complications during pregnancy can provide an early window into future risk of cardiovascular disease (CVD) in women.2,3 Prior data suggest that hypertension subsequent to pregnancy may be an important factor mediating the association between pregnancy factors and CVD in later life.4–6
Maternal characteristics that have been linked to incident CVD include parity and a younger age at first delivery.7 Individual pregnancy and birth complications that have been linked to CVD in later life in women include preeclampsia, gestational diabetes mellitus, placental abruption, preterm delivery, and having a small-for-gestational age (SGA) baby.2 A better understanding of which specific pregnancy factors that are independently associated with high blood pressure (BP) in midlife may allow us to more precisely target high-risk women with early lifestyle interventions to prevent or delay onset of both elevated BP and subsequent CVD.
Prior data regarding associations between selected individual pregnancy characteristics with subsequent BP levels have yielded conflicting results. For instance, findings from the Avon Longitudinal Study of Parents and Children demonstrate that pregnancy-induced hypertension and giving birth to an SGA infant predicted future increases in both systolic BP (SBP) and diastolic BP (DBP) 2 decades after childbirth, independent of other pregnancy factors, including gestational diabetes mellitus and gestational age.4 In that report, preterm delivery (gestational age <37 weeks) was not significantly related to future risk of developing increased SBP.4 Conversely, in a separate investigation in the NHANES cohort (National Health and Nutrition Examination Survey), preterm delivery but not SGA was related to later BP, with stronger associations observed among black and premenopausal women.6 Because of lack of prior studies, it is uncertain whether hypertension mediates the reported association between placental abruption or age at first delivery and later CVD, especially when related pregnancy complications, such as preeclampsia, gestational age, and SGA, are accounted for.
Using the Swedish Medical Birth Register and other nation-wide registers combined with a cohort in Northern Sweden called the Västerbotten Intervention Program (VIP),8 we sought to determine the independent associations between several key pregnancy complications and maternal characteristics with subsequent hypertension and BP at 40 years of age in women.
Methods
Study Sample
The creation of the study sample is depicted in Figure.
Creation of the study sample.
VIP Data Linked With the Swedish Medical Birth Register
The Swedish county of Västerbotten was disproportionately affected by rising CVD rates in the early 1980s, prompting the initiation of a public health educational campaign aimed at reducing CVD and diabetes mellitus rates. The VIP is a CVD screening primary prevention project that started in 1985 with a pilot in a small municipality and was then disseminated to the rest of the county from 1990.8 The study examinations consisted of information on sociodemographic, past medical history, anthropometric measurements, and CVD risk factor measurements pertinent to our study, including BP, blood lipid collection, and oral glucose tolerance testing. Invitations to participate in VIP were sent out to men and women in the Västerbotten county the year they turned 30 (in the whole county only until 1995), 40, 50, or 60 years of age. We included all women in VIP born between 1953 and 1971 who attended the VIP visit at 40 years of age. The range of women’s birth year (1953–1971) was chosen to ascertain their full reproductive history in the Medical Birth Register, starting in 1973, and to include all women who would be ≥40 years of age by the 2011 VIP examination. Of 16 009 unique women who had participated in their examination at 40 years of age, 15 896 women had BP data and were included in the analyses. Among these women, 2463 (15.5%) had also participated in the examination at 30 years of age.
Swedish Medical Birth Register
The Swedish Medical Birth Register was started in 1973 as a means to collect detailed maternal, obstetric, and neonatal data during pregnancy. More than 98% of all births in Sweden are registered in the Medical Birth Register.9 The maternal and infant diagnoses are coded using International Classification of Diseases (ICD) codes at the delivery (ICD-8 until 1986; ICD-9 from 1987 to 1996, and ICD-10 thereafter). The validity and the coverage of the maternal diagnoses has been shown to be high (96%–98%) when validated against medical records.10
Pregnancy Exposures
The pregnancy characteristics evaluated as potential risk factors for hypertension include the following: preeclampsia or eclampsia (hypertension in pregnancy with proteinuria—with seizures in the case of eclampsia), gestational diabetes mellitus (this information was available after 1987), placental abruption, preterm delivery (defined as very [<32 weeks] or moderately [32–36 weeks] preterm delivery), SGA (less than the third percentiles of mean birth weight for gestational age according to the Swedish reference curve for normal fetal growth),11 and stillbirth (at 28 weeks of gestation or later until July 1, 2008, and thereafter at 22 weeks of gestation or later). Maternal age at first delivery was obtained from the Medical Birth Register and used as a continuous variable in the analyses. Year of first delivery was categorized as 1973 to 1986, 1987 to 1996, and 1997 to 2011, according to the changes of the ICD in Sweden. History of hypertension and diabetes mellitus prior to pregnancy was based on self-report and discharge diagnosis codes, obtained from the Swedish Medical Birth Register (using the Swedish versions of the ICD-8, ICD-9, and ICD-10). Gestational diabetes mellitus was defined by ICD codes (ICD-9 648 W and ICD-10 O244). Placental abruption was defined using ICD-8 6321 6514, ICD-9 641C, and ICD-10 O45. Preeclampsia is defined as hypertension accompanied by proteinuria, and eclampsia is defined as the occurrence of preeclampsia with seizures. We included preeclamptic and eclamptic pregnancies in our analysis as a combined outcome (ICD-8 63703-99, ICD-9 642E-G, and ICD-10 O14-15). From the VIP, we obtained in later ages detailed self-reported amount and duration of smoking and snuff use, body mass index (BMI), history of hypertension, and diabetes mellitus from the time of pregnancy. The positive predictive value of hypertensive disorders in pregnancy (including preeclampsia) has been demonstrated to be as high as 98%.12 The accuracy of placental abruption ICD codes has not been validated against medical records in Sweden, but have demonstrated accuracy in studies conducted both in the United States and in Finland.13
Ascertainment of BP and Other Anthropometrics in VIP
Participants visited their health center after an overnight fast. BP was measured once using a mercury column sphygmomanometer after 5 minutes rest, with the subject in a seated position. Both SBP and DBP values were recorded. For the purpose of this study, we defined hypertension as an SBP ≥140 mm Hg or a DBP ≥90 mm Hg or treatment with antihypertensive medications (diuretics, β-blockers, calcium antagonists, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, or α-receptor blockers). The antihypertensive medication use was confirmed using the Register of Prescribed Drugs. BMI was calculated by height and weight in light clothing in kilogram per square meter. Detailed participant smoking and alcohol intake history were assessed on participant questionnaire. Smoking in both the Medical Birth register and in VIP was defined as any amount of current smoking. Alcohol use from the VIP database was defined in the following categories: never or occasionally, 1 to 4 times per month, or more than once per week to daily.
Socioeconomic Factors
Maternal level of education was obtained from the Education Register held at Statistics Sweden (www.scb.se). It was categorized as ≤9 years of schooling, 10 to 12 years of schooling, and ≥13 years of schooling. The household’s income closest to the 40-year visit was retrieved from the Income Register, also held at Statistics Sweden, and categorized into tertiles of the distribution that year. Absolute numbers could not be used because the variable was skewed and redefined in 2004, hampering the possibility to compare incomes before and after 2004.
Ethical approval was obtained in September 2012 by the Regional Ethical Review Board in Stockholm, Sweden (IRB/KI 2012/1325–31/3).
Statistical Methods
Descriptive statistics were used to describe the study population, including numbers, percentages, means, and standard deviations. We used linear and logistic regression in unadjusted and multivariable adjusted models. Because all women were assessed at age close to 40 years (mean =40.1; 95% CI, 39.5–40.6), age adjustment was not done in our primary models. Model 1 was adjusted for all other pregnancy characteristics and year of first delivery category, and model 2 was additionally adjusted for household disposable income, education, smoking, alcohol intake, and BMI at 40 years of age. Linear regression analysis was used to study associations between pregnancy characteristics and SBP and DBP at 40 years of age, and logistic regression analysis was used to study the association between pregnancy characteristics and incidence of hypertension at 40 years of age. To account for the effects of antihypertensive drugs, we added 10 mm Hg to the SBP and 5 mm Hg to the DBP for women taking these medications. This method has been used previously to conservatively estimate the effect of antihypertensive medication on BP.14 A Hosmer–Lemeshow goodness-of-fit test was calculated to assess whether there was evidence for lack of fit of the logistic models. In secondary analysis, we adjusted for age to ensure that even our small age range did not affect the estimates. Given our prior study that demonstrated a synergistic effect of a short gestational age and preterm delivery on later maternal CVD, we explored this interaction on BP and hypertension outcomes.15 Other interaction terms were selected because of the hypotheses that the pregnancy characteristics may have synergistic effects on BP (preeclampsia×SGA, SGA×age categories, preeclampsia×gestational age categories, and gestational diabetes mellitus×gestational age categories). In a sensitivity analysis, we excluded women delivering before 1987 (n=2926), given that we did not have gestational diabetes mellitus information in these women to see if this materially changes our estimates.
Results
Among 15 896 women, 2767 (17.4%) had at least 1 pregnancy complication, defined as preeclampsia, gestational diabetes mellitus, placental abruption, preterm birth (gestational age <37 weeks), SGA infant, or stillbirth.
There were n=1528 (9.6%) women with hypertension, of which n=472 (3.0%) reported using antihypertensive medications at the 40-year examination. Of these, 466 were still alive at the start of the Swedish Prescribed Drug Register in 2005, and 452 had been dispensed at least 1 prescription of any of the following: diuretics, β-blockers, calcium antagonists, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, or α-receptor blocker between 2005 and 2013. The actual mean SBP and DBPs at 40 years of age among study participants were 116 (SD=13) and 73 (SD=10) mm Hg, respectively. The mean (SD) BP remained identical on group level after adding 10 mm Hg to SBP and 5 mm Hg to DBP among women on antihypertensive medication. Baseline characteristics are shown in Table 1.
Characteristics of Women in the Västerbotten Intervention Program (VIP)
Pregnancy Characteristics and SBP and DBP at 40 Years of Age
Systolic BP
In the fully adjusted model 2 (which adjusted for pregnancy characteristics, lifestyle, BMI, and socioeconomic factors), preeclampsia, preterm delivery (32 to <37 weeks), and SGA were associated with higher SBP (Table 2). Gestational diabetes mellitus and stillbirth were related to higher SBP in model 1 only (which adjusted for other pregnancy characteristics). On adjusting for lifestyle factors, BMI, and socioeconomic factors, there was no longer an association between gestational diabetes mellitus and SBP, and the association between stillbirth and SBP was of borderline statistical significance (P=0.05). In fully adjusted models, preeclampsia was associated with an estimated increase in SBP of 6.95 mm Hg, preterm delivery with a 0.9 mm Hg increase, and SGA with a 1.02 mm Hg increase in SBP.
Pregnancy Factors and Blood Pressure at Age 40 Years in Swedish Women in the Västerbotten Intervention Program
Diastolic BP
In the fully adjusted model 2 (which adjusted for pregnancy characteristics, lifestyle, BMI, and socioeconomic factors), lower parity, younger age at first birth, preeclampsia, and placental abruption were significantly associated with higher DBP. Gestational diabetes mellitus was related to higher DBP in model 1 only (which adjusted for other pregnancy characteristics). On adjusting for lifestyle factors, BMI, and socioeconomic factors, there was no longer an association between gestational diabetes mellitus and DBP. In fully adjusted models, preeclampsia was associated with an average increase in DBP of 4.68 mm Hg and placental abruption with 1.72 mm Hg increase, whereas other pregnancy characteristics were associated with relatively smaller changes in DBP (0.23 mm Hg increase per 1 U decrease in parity, 0.07 mm Hg increase per each 1 year decrease in age at first birth).
Pregnancy Characteristics and Hypertension at 40 Years of Age
In the fully adjusted model 2 (which adjusted for pregnancy characteristics, lifestyle, BMI, and socioeconomic factors), younger age at first birth, preeclampsia, gestational age <32 weeks, and SGA were associated with increased risks of hypertension (Table 3). Gestational diabetes mellitus and stillbirth were related to hypertension in unadjusted models but not in multivariable models. Placental abruption was not related to risk of hypertension at 40 years of age. Preeclampsia was associated with a 3-fold (odds ratio=3.09) odds of hypertension at 40 years of age in fully adjusted model 2, gestational age <32 weeks with a 1.57-fold increase, and SGA with a 1.33-fold increase odds of hypertension (Table 3), and the association of age at first birth with hypertension was more modest (odds ratio=0.98 per additional age in years of first birth).
Pregnancy Factors and Hypertension at Age 40 Years in Swedish Women in the Västerbotten Intervention Program
In secondary analysis, we additionally adjusted for participant age, and it did not materially change our estimates or levels of statistical significance (data not shown).
Effect Modification
There was no evidence for effect modification for the following: preeclampsia×SGA, SGA×age categories, preeclampsia×gestational age categories, or gestational diabetes mellitus×gestational age categories.
Sensitivity Analysis Excluding Women Delivering Prior to 1987
When restricting analyses to women having deliveries in 1987 and later, associations were slightly stronger between gestational diabetes mellitus and SBP but were still not statistically significant (data not shown).
Discussion
Among 40-year-old Swedish women from the VIP, at least one of the following pregnancy complications that we studied was present in 17.4% of women (preeclampsia, gestational diabetes mellitus, placental abruption, shortest gestational age [<32, 32–36, ≥37 weeks], SGA baby [<third percentile for birthweight], or stillbirth). Age at first birth, preeclampsia, having a preterm infant, SGA infant, and a lower parity were all associated with both continuous increases in either SBP or DBP (or both), as well as the binary outcome of hypertension. We found that placental abruption was associated with higher DBP in women aged 40 years after accounting for other pregnancy characteristics, social and lifestyle factors (education, income, smoking, and alcohol use) but not with the binary outcome of hypertension. Preeclampsia had the strongest associations with BP and hypertension at 40 years of age as compared with other pregnancy characteristics.
Pregnancy Characteristics and BP
Consistent with our findings that parity is associated with modestly lower DBP at 40 years of age, the investigators in the CARDIA study (the Coronary Artery Risk Development in Young Study) reported that increased parity is associated with lower postpartum BP both proximate to and over 2 decades following childbirth.16 A separate prior investigation demonstrated that this inverse association may only be present in younger versus older cohorts of women.17 We extended prior findings by accounting for several pregnancy complications, including preeclampsia.
Having a child at a young age can lead to adverse cardiometabolic effects in mothers.18,19 Recent data suggest that a younger age at first birth is associated with coronary artery disease in postmenopausal women.7 Underlying this association is the observation that young mothers, who have not yet achieved their full adolescent growth, compete with their fetuses for both nutrition and energy.20 Our data suggest that a young age at first birth is related to both higher DBP, as well as hypertension at 40 years of age, independent of socioeconomic factors and pregnancy complications. However, it is also possible that residual sociodemographic factors, not fully explained in our data, underlie the association between younger age at first pregnancy and CVD.
Preeclampsia has a well-established association with incident hypertension and later life CVD.2,21 Our findings suggest that these associations are only slightly attenuated after adjusting for other pregnancy complications, socioeconomic, and lifestyle factors (unadjusted versus adjusted odds ratio of 3.74 versus 3.09 for hypertension). Preterm delivery (<37 weeks) has been linked with modest increases in later BP levels in prior investigations.4,5 One of these studies did not account for hypertension in pregnancy,4 and both prior investigations considered preterm birth dichotomously at 37 weeks. Our prior work suggests that women with preterm deliveries (ie, <32 weeks of gestation) experience higher rates of CVD as compared with those with deliveries between 32 and 36 weeks.15 We observed that preterm delivery (<32 weeks) is related to risk of hypertension and that moderately preterm delivery (32–36 weeks) was modestly and positively related to SBP. These findings indicate that there may be a dose–response relationship between degree of preterm delivery and subsequent BP in mothers once other pregnancy characteristics are considered. Tighter BP control during pregnancies that are complicated by hypertension did not affect adverse serious maternal or fetal outcomes but did protect against severe maternal hypertension during pregnancy.22 Whether tight BP control in pregnancy led to later decreases in maternal BP during the postpartum period is an important area of future research.
Our prior work demonstrates that SGA is a risk marker for incident CVD that is independent of preeclampsia and gestational age.15 Our current findings suggest that increases in SBP and hypertension may underlie this association. Two recent studies have demonstrated an association between placental abruption and later CVD mortality in women.23,24 Our study did not demonstrate an independent association between placental abruption and later maternal hypertension. However, although placental abruption is a rare condition, it was associated with a rise in DBP even after accounting for lifestyle factors, socioeconomic factors, and pregnancy characteristics. Placental abruption is a disorder of placentation and, thus, may reflect vascular dysfunction in mothers in a similar way as preeclampsia, another placental disorder. The lack of significant associations between gestational diabetes mellitus and maternal hypertension or BP is consistent with prior findings and likely reflect the fact the gestational diabetes mellitus leads to maternal CVD primarily through its strong association of subsequent type 2 diabetes mellitus in the mother.4
Importance of Primordial Prevention of Hypertension
Prior epidemiological evidence suggests that prehypertension and increased BP trajectories in young adults can lead to increased subclinical atherosclerosis in midlife.25 Cumulative increases in BP have been associated with carotid intimal media thickness.26 Decreasing the trajectory of BP by midlife has been demonstrated to decrease one’s lifetime risk of CVD.27 Modifiable risk factors for incident hypertension include obesity, alcohol intake, SBP, and DBP in the normal range and cigarette smoking.28 A prior study suggests that having normal BP after preeclampsia is associated with a CVD risk score that is equivalent to that of healthy control study participants without preeclampsia.29 Several lifestyle changes can be successfully undertaken among nonhypertensive individuals to decrease BP and delay the onset of hypertension.30 These studies taken together with our present findings provide evidence for the early identification of risk factors for high BP and leave open the possibility for aggressive risk factor modification to prevent high BP after pregnancy complications as a means to reduce later hypertension in women with the aim of decreasing her lifetime burden of CVD.
Strengths and Limitations
A key strength of our study is the relatively large, unselected sample of women with well-characterized pregnancy data and longitudinal follow-up, including measured BP. Moreover, we were able to account for important socioeconomic cardiovascular risk factors, such as income, education, smoking, and alcohol use. Limitations include the fact that not all women invited to participate in the VIP accepted. Participation rates have varied over the years, ranging between 59.8% and 64.6% for women at 40 years of age from 1990 to 2006.31 Moreover, this population is ethnically and geographically homogeneous. This may increase internal validity, but may also limit the generalizability of our findings to other populations. Finally, our definition of hypertension was partially based on a one-time study visit BP and, thus, is not as accurate as multiple longitudinal BP measurements.
Directions for Future Research
Future studies in ethnically diverse populations should be undertaken to replicate these results. Studies assessing the utility of using these pregnancy risk factors for hypertension screening and for risk stratification would be important in terms of better capturing the population of women who may potentially benefit from more frequent BP monitoring. The biological mechanisms underlying these associations are also of scientific and public health importance and warrant further study.
Perspectives
Key pregnancy characteristics selected in our study (ie, preeclampsia, baby’s size for gestational age, preterm delivery, etc) are easily recorded during a reproductive history and represent a noninvasive way to ascertain a woman’s later hypertension risk. Therefore, our findings that key pregnancy complications are independently related to hypertension at 40 years of age suggest that knowledge of pregnancy history may provide an early window into a woman’s risk for development of hypertension. More frequent monitoring of women with pregnancy complications that are related to increased BP and hypertension at 40 years of age may be indicated. Currently, women with a complicated pregnancy tend to underutilize postpartum primary care,32 highlighting a potential healthcare utilization gap that can be targeted in the future to decrease the burden of hypertension in women.
Acknowledgments
We thank the participants of the Västerbotten Intervention Program.
Sources of Funding
This work was supported by National Institutes of Health R21 7R21HL115398 (N.I. Parikh) and American Heart Association grant 13CRP17350002 (N.I. Parikh), and Swedish Research Council for Health, Working Life, and Welfare 2010-0643 (A.-K. Edstedt Bonamy).
Disclosures
None.
- Received July 21, 2016.
- Revision received August 10, 2016.
- Accepted January 3, 2017.
- © 2017 American Heart Association, Inc.
References
Novelty and Significance
What Is New?
We demonstrate that key pregnancy factors (age at first birth, preeclampsia, having a preterm and small for gestational age infant, lower parity, and placental abruption) when considered together are independently related to higher blood pressure or hypertension at 40 years of age in women.
What Is Relevant?
It may be relevant to collect this pregnancy information in women of childbearing age to more closely monitor them for the onset of high blood pressure and to modify their risk of developing hypertension.
Summary
In summary, pregnancy is an early window into later hypertension risk in women.
This Issue
Jump to
Article Tools
Share this Article
- Association of Pregnancy Complications and Characteristics With Future Risk of Elevated Blood PressureNovelty and Significance
