Association of Pregnancy Complications and Characteristics with Future Risk of Elevated Blood Pressure: The Västerbotten Intervention Programme

Abstract

Pregnancy characteristics are associated with risk of cardiovascular diseases, but their independent associations with hypertension or blood pressure levels remain uncertain. We linked the Swedish Medical Birth Register with Västerbotten Intervention Programme data (Northern Sweden). Using linear and logistic regression, we related pregnancy factors in any prior pregnancy with blood pressure and hypertension at age 40 years in 15,896 parous women free of prepregnancy hypertension. Pregnancy factors included parity, age at first delivery, preeclampsia, gestational diabetes, placental abruption, shortest gestational age small for gestational age baby (SGA; < 3rd percentile for birth weight) or stillbirth. We defined hypertension as systolic BP ≥ 140 and/or DBP ≥ 90 mmHg and/or antihypertensive use. Multivariable models were adjusted for all pregnancy factors and potential lifestyle and sociodemographic confounders. At age 40 years, 1,535 women (9.6%) had hypertension. In multivariable models, lower parity, younger age at first birth, preeclampsia, SGA, and placental abruption were independently associated with higher systolic and/or diastolic blood pressure levels at age 40 years. Younger age at first birth, preeclampsia, gestational age <32 versus ≥ 37 weeks and SGA, 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, 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 in mid-life may allow us to more precisely target high risk women with early lifestyle interventions to prevent or delay onset of both elevated blood pressure and subsequent CVD.

Prior data regarding associations between selected individual pregnancy characteristics with subsequent blood pressure 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 and diastolic blood pressure two decades following childbirth, independent of other pregnancy factors, including gestational diabetes and gestational age.⁴ In that report, preterm delivery (gestational age < 37 weeks) was not significantly related to future risk of developing increased systolic blood pressure.⁴ Conversely, in a separate investigation in the NHANES cohort, preterm delivery but not SGA was related to later blood pressure with stronger associations observed among African American and premenopausal women.⁶ Due to 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 Programme,⁸ we sought to determine the independent associations between several key pregnancy complications and maternal characteristics with subsequent hypertension and blood pressure at age 40 years in women.

Methods

Study Sample

The creation of the study sample is depicted in the Figure.

Figure.

Creation of the Study Sample.

Västerbotten Intervention Programme (or VIP) data linked with the Swedish Medical Birth Register

The Swedish county of Västerbotten was disproportionately affected by rising CVD rates in the early 1980’s, prompting the initiation of a public health educational campaign aimed at reducing CVD and diabetes 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 blood pressure, blood lipid collection and oral glucose tolerance testing (OGTT). 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. We included all women in VIP born between 1953 and 1971 who attended the VIP visit at age 40 years. The range of womens’ birth year (1953 to 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 age 40 years or older by the 2011 VIP examination. Of 16,009 unique women who had participated in their examination at age 40 years, 15,896 women had blood pressure data and were included in the analyses. Among these women, 2,463 (15.5%) had also participated in the examination at age 30 years.

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 ICD codes at the delivery (ICD 8 until 1986; ICD 9 from 1987-1996 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 (this information was available after 1987), placental abruption, preterm delivery (defined as very [<32 weeks] or moderately [32 to 36 weeks] preterm delivery), small for gestational age (less than the 3^rd 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^st 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 1973-1986, 1987-1996 and 1997-2011, according to the changes of the ICD-classifications in Sweden. History of hypertension and diabetes 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 International Classification of Diseases, eighth, ninth and tenth revisions [ICD-8. ICD-9, ICD-10, respectively]). Gestational diabetes mellitus was defined by ICD codes (ICD-9 648W, ICD-10 O244). Placental abruption was defined using ICD-8 6321 6514; ICD-9 641C; and ICD-10 O45. Pre-eclampsia is defined as hypertension accompanied by proteinuria and eclampsia is defined as the occurrence of preeclampsia with seizures. We included pre-eclamptic and eclamptic pregnancies in our analysis as a combined outcome (ICD-8 63703-99; ICD-9 642E-G, ICD10 O14-15. From the VIP, we obtained in later ages detailed self-reported amount and duration of smoking and snuff use, BMI, history of hypertension and diabetes 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 have not been validated against medical records in Sweden, but have demonstrated accuracy in studies conducted both the United States and Finland.¹³

Ascertainment of blood pressure and other anthropometrics in VIP

Participants visited their health center following an overnight fast. Blood pressure was measured once using a mercury column sphygmomanometer after 5 minutes rest with the subject in a seated position. Both systolic and diastolic blood pressure values were recorded. For the purpose of this study, we defined hypertension as a systolic BP ≥ 140 mmHg, or a diastolic BP ≥ 90 mmHg, or treatment with antihypertensive medications (diuretics, beta-blockers, calcium antagonists, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers and/or alpha receptor blockers). The antihypertensive medication use was confirmed using the Register of Prescribed Drugs. Body mass index 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, 1-4 times per month, or > 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 or less, 10-12 years of schooling, and 13 years or more. 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 since 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 employed linear and logistic regression in unadjusted and multivariable adjusted models. As all women were assessed at age close to age 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 body mass index at age 40 years. Linear regression analysis was employed to study associations between pregnancy characteristics and systolic and diastolic blood pressure at age 40 years, and logistic regression analysis was employed to study the association between pregnancy characteristics and incidence of hypertension at age 40 years. In order to account for the effects of antihypertensive drugs, we added 10 mm Hg to the systolic BP and 5 mmHg to the diastolic BP 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 very preterm delivery on later maternal cardiovascular disease, we explored this interaction on blood pressure and hypertension outcomes.¹⁵ Other interaction terms were selected because of hypotheses that the pregnancy characteristics may have synergistic effects on blood pressure (preeclampsia *SGA, SGA*age categories, preeclampsia *gestational age categories, and gestational diabetes*gestational age categories). In a sensitivity analysis we excluded women delivering before 1987 (n=2926) given than we did not have gestational diabetes information in these women to see if this materially changes our estimates.

Results

Among 15,896 women, 2767 (17.4%) had at least one pregnancy complication, defined as preeclampsia, gestational diabetes, placental abruption, preterm birth (gestational age < 37 weeks), small for gestational age infant or stillbirth.

There were n=1528 (9.6%) women with hypertension, of which n=472 (3.0%) reported using anti-hypertensive medications at the 40 year exam. Of these, 466 were still alive at the start of the Swedish Prescribed Drug Register in 2005, and 452 had been dispensed at least one prescription of any of the following: diuretics, beta-blockers, calcium antagonists, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers and/ or alpha receptor blocker between 2005 and 2013. The actual mean systolic and diastolic blood pressures at age 40 among study participants were 116 (SD 13) and 73 (SD 10) mmHg, respectively. The mean (SD) blood pressure remained identical on group level after adding 10 mm Hg to SBP and 5 mm Hg to DBP among women on anti-hypertensive medication.

Baseline characteristics are shown in Table 1.

Table 1.

Characteristics of Women in the Västerbotten Intervention Programme (VIP).

Variable	Mean (SD) or n%
Pregnancy characteristics
Parity, n (%)
1	2125 (13.4)
2	7837 (49.3)
≥ 3	5934 (37.3)
Age at first pregnancy, years, mean (SD)	25.6 (4.4)
Age at last pregnancy, years, mean (SD)	30.9 (4.3)
Preeclampsia, n (%)	852 (5.4)
Gestational Diabetes, n (%)	294 (1.8)
Placental Abruption, n (%)	206 (1.3)
Stillbirth, n (%)	112 (0.7)
Gestational Age, n (%)
<32 weeks	194 (1.1)
≥ 32 to 36weeks	1293(8.1)
≥ 37 weeks	14409 (90.7)
Small for Gestational Age, n (%)	733 (4.6)
Data from Statistics Sweden
Education, n (%)
≤ 9 years	737 (4.6)
9-12 years	8077 (50.8)
>12 years	7081 (44.6)
Disposable Annual Income, USD
Tertile 1	18,815 (5,235)
Tertile 2	27,016 (1,533)
Tertile 3	39,318 (2,603)
Data from VIP exam at 40 y of age
Family history of MI or stroke before age 60, n (%)	2969 (18.8)
Hypertension, n (%)	1528 (9.6)
Treatment with anti-hypertensive medications	472 (3.0)
Smoking at 40 years of age, n (%)	2180 (13.7)
Alcohol intake at 40 years of age, n (%)
Never or occasionally	4687 (29.5)
1-4 times per month	9925 (62.6)
>Once per week to daily	1254 (7.9)
Body Mass Index at 40 years of age, n (%)
<25 kg/m2	9589 (60.6)
25-to <30 kg/m2	4255 (26.9)
≥30 kg/m2	1985 (12.5)

Pregnancy characteristics and systolic and diastolic blood pressure at age 40 years

Systolic blood pressure

In the fully adjusted Model 2 (which adjusted for pregnancy characteristics, lifestyle, body mass index and socioeconomic factors), preeclampsia, preterm delivery (32 to <37 weeks), and SGA were associated with higher systolic blood pressure (Table 2). Gestational diabetes and stillbirth were related to higher systolic blood pressure in Model 1 only (which adjusted for other pregnancy characteristics). Upon adjusting for lifestyle factors, body mass index and socioeconomic factors, there was no longer an association between gestational diabetes and systolic blood pressure and the association between stillbirth and systolic blood pressure was of borderline statistical significance (p=0.05). In fully adjusted models, preeclampsia was associated with an estimated increase in systolic blood pressure of 6.95 mmHg, preterm delivery with a 0.9 mmHg increase, and SGA with a 1.02 mmHg increase in systolic blood pressure.

Table 2.

Pregnancy Factors and Blood Pressure at Age 40 in Swedish Women in the Västerbotten Intervention Programme

Pregnancy Factors	Systolic blood pressure at age 40 (mm Hg) β (SE)	P value	Diastolic blood pressure at age 40 (mm Hg) β (SE)	P value
Parity (per child)
Unadjusted	-0.05 (0.12)	0.64	-0.10 (0.09)	0.24
Model 1	-0.31 (0.14)	0.02	-0.35 (0.10)	<0.001
Model 2	-0.11 (0.13)	0.39	-0.23 (0.10)	0.02
Age at first birth (per year)
Unadjusted	-0.24 (0.02)	<0.001	-0.16 (0.02)	<0.001
Model 1	-0.18 (0.03)	<0.001	-0.16 (0.02)	<0.001
Model 2	-0.04 (0.03)	0.20	-0.07 (0.02)	0.001
Preeclampsia*
Unadjusted	8.54 (0.47)	<0.001	5.79 (0.34)	<0.001
Model 1	8.55 (0.47)	<0.001	5.76 (0.34)	<0.001
Model 2	6.95 (0.45)	<0.001	4.68 (0.33)	<0.001
Gestational diabetes*
Unadjusted	2.42 (0.79)	0.002	1.59 (0.57)	0.005
Model 1	2.32 (0.78)	0.003	1.74 (0.57)	0.002
Model 2	0.56 (0.75)	0.45	0.57 (0.55)	0.30
Stillbirth*
Unadjusted	4.26 (1.27)	<0.001	1.89 (0.92)	0.04
Model 1	4.01 (1.29)	0.002	1.77 (0.94)	0.06
Model 2	2.41 (1.24)	0.05	0.70 (0.91)	0.44
Placental abruption*
Unadjusted	2.93 (0.94)	0.002	2.71 (0.68)	<0.001
Model 1	1.54 (0.95)	0.10	1.88 (0.69)	0.006
Model 2	1.37 (0.90)	0.13	1.72 (0.66)	0.01
Gestational age, weeks†
Unadjusted
<32	3.71 (0.96)	<0.001	2.62 (0.70)	<0.001
32 to <37	1.95 (0.39)	<0.001	1.21 (0.28)	<0.001
≥ 37	Ref.	Ref.	Ref.	Ref.
Model 1
<32	1.62 (0.98)	0.10	1.19 (0.72)	0.09
32 to <37	0.82 (0.39)	0.04	0.45 (0.29)	0.12
≥ 37	Ref.	Ref.	Ref.	Ref.
Model 2
<32	1.80 (0.94)	0.05	1.33 (0.69)	0.05
32 to < 37	0.90 (0.38)	0.02	0.51 (0.28)	0.06
≥ 37	Ref.	Ref.	Ref.	Ref.
SGA *
Unadjusted	1.83 (0.50)	<0.001	1.03 (0.37)	0.005
Model 1	0.38 (0.51)	0.46	0.04 (0.37)	0.90
Model 2	1.02 (0.49)	0.04	0.51 (0.36)	0.16
Current smoking, Y/N
Unadjusted	-0.37 (0.31)	0.23	-0.81 (0.22)	<0.001
Model 1
Model 2	-1.67 (0.30)	<0.001	-1.57 (0.22)	<0.001
Alcohol consumption
Unadjusted
Never or occasionally	Ref.	Ref.	Ref.	Ref.
1-4 times per month	-1.71(0.24)	<0.001	-1.14	<0.001
>Once per week to daily	-2.36 (0.42)	<0.001	-1.84	<0.001
Model 1
Model 2
Never or occasionally	Ref.	Ref.	Ref.	Ref.
1-4 times per month	-0.32 (0.23)	0.17	-0.24 (0.17)	0.15
>Once per week to daily	0.17 (0.41)	0.68	-0.24 (0.30)	0.42
Education
Unadjusted
≤ 9 years, n (%)	3.46 (0.51)	<0.001	1.96 (0.37)	<0.001
9-12 years	2.63 (0.22)	<0.001	1.61 (0.16)	<0.001
>12 years	Ref.	Ref.	Ref.	Ref.
Model 1
Model 2
≤9 years, n (%)	1.32 (0.51)	0.01	0.62 (0.38)	0.10
9-12 years	1.11 (0.22)	<0.001	0.69 (0.16)	<0.001
>12 years	Ref.	Ref.	Ref.	Ref.
Income
Unadjusted
Lowest tertile	2.87 (0.26)	<0.001	1.69 (0.19)	<0.001
Middle tertile	1.56 (0.26)	<0.001	0.88 (0.19)	<0.001
Highest tertile	Ref.	Ref.	Ref.	Ref.
Model 1
Model 2
Lowest tertile	0.99 (0.26)	<0.001	0.49 (0.19)	0.01
Middle tertile	0.62 (0.25)	0.01	0.27 (0.18)	0.14
Highest tertile	Ref.	Ref.	Ref.	Ref.
BMI, per kg/m2
Unadjusted	0.89 (0.02)	<0.001	0.60 (0.02)	<0.001
Model 1
Model 2	0.85 (0.02)	<0.001	0.57 (0.02)	<0.001

In the n=472 women on antihypertensive medication 10 mm Hg was added to their SBP and 5mm Hg was added to their DBP.

^*If the condition was present in ≥ 1 pregnancy

^†The shortest gestational age of all a woman’s pregnancies

Model 1 adjusts for total parity, maternal age at first birth, lowest gestational age, birth of a small-for gestational age infant, preeclampsia, gestational diabetes, stillbirth, placental abruption, and year of first delivery. Model 2 adjusts for current smoking, alcohol consumption, body mass index, education, and income in addition to adjustments in Model 1.

Diastolic blood pressure

In the fully adjusted Model 2 (which adjusted for pregnancy characteristics, lifestyle, body mass index and socioeconomic factors), lower parity, younger age at first birth, preeclampsia, and placental abruption were significantly associated with higher diastolic blood pressure. Gestational diabetes was related to higher diastolic blood pressure in Model 1 only (which adjusted for other pregnancy characteristics). Upon adjusting for lifestyle factors, body mass index and socioeconomic factors, there was no longer an association between gestational diabetes and diastolic blood pressure. In fully adjusted models, preeclampsia was associated with an average increase in diastolic blood pressure of 4.68 mmHg and placental abruption with 1.72 mmHg increase whereas other pregnancy characteristics were associated with relatively smaller changes in diastolic blood pressure (0.23 mmHg increase per 1 unit decrease in parity, 0.07 mmHg increase per each one year decrease in age at first birth).

Pregnancy characteristics and at 40 years of age

In the fully adjusted Model 2 (which adjusted for pregnancy characteristics, lifestyle, body mass index 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 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. Preeclampsia was associated with a threefold (OR=3.09) odds of hypertension at age 40 years 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 associations of younger at first birth with hypertension was more modest (OR=0.98 per additional age in years of first birth).

Table 3.

Pregnancy Factors and Hypertension at Age 40 in Swedish Women in the Västerbotten Intervention Programme

Pregnancy Factors	Odds Ratio for Hypertension (95 % CI)
Parity
Unadjusted
1	1.07 (0.92-1.26)
2	1.00 (reference)
≥3	1.04 (0.93-1.17)
Adjusted for pregnancy factors
1	1.19 (1.00-1.40)
2	1.00 (reference)
≥3	0.88 (0.78-1.00)
Adjusted for pregnancy, lifestyle factors and SES
1	1.05 (0.88-1.24)
2	1.00 (reference)
≥3	0.90 (0.79-1.03)
Age at first birth (per year)
Unadjusted	0.97 (0.96-0.99)
Adjusted for pregnancy factors	0.95 (0.94-0.97)
Adjusted for pregnancy, lifestyle factors and SES	0.98 (0.96-0.99)
Preeclampsia*
Unadjusted	3.74 (3.18-4.42)
Adjusted for pregnancy factors	3.62 (3.05-4.30)
Adjusted for pregnancy, lifestyle factors and SES	3.09 (2.59-3.70)
Gestational diabetes*
Unadjusted	1.70 (1.22-2.36)
Adjusted for pregnancy factors	1.66 (1.18-2.32)
Adjusted for pregnancy, lifestyle factors and SES	1.32 (0.92-1.87)
Stillbirth*
Unadjusted	1.90 (1.14-3.16)
Adjusted for pregnancy factors	1.64 (0.96-2.82)
Adjusted for pregnancy, lifestyle factors and SES	1.35 (0.78-2.34)
Placental abruption*
Unadjusted	1.35 (0.89-2.04)
Adjusted for pregnancy factors	1.11 (0.72-1.71)
Adjusted for pregnancy, lifestyle factors and SES	1.10 (0.70-1.71)
Gestational age, weeks†
Unadjusted
<32	2.34 (1.62-3.38)
32 to <37	1.32 (1.11-1.58)
≥ 37	1.00 (referent)
Adjusted for pregnancy factors
<32	1.50 (1.00-2.23)
32 to <37	1.05 (0.87-1.27)
≥ 37	1.00 (reference)
Adjusted for pregnancy, lifestyle factors and SES
<32	1.57 (1.04-2.37)
32 to < 37	1.07 (0.88-1.30)
≥ 37	1.00 (reference)
SGA*
Unadjusted	1.61 (1.30-1.99)
Adjusted for pregnancy factors	1.21 (0.96-1.52)
Adjusted for pregnancy, lifestyle factors and SES	1.33 (1.05-1.68)
Current smoking, Y/N
Unadjusted	0.97 (0.83-1.13)
Adjusted for pregnancy, lifestyle factors and SES	0.88 (0.75-1.04)
Alcohol consumption
Unadjusted
Never or occasionally	1.00 (reference)
1-4 times per month	0.74 (0.66-0.83)
>Once per week to daily	0.68 (0.54-0.85)
Adjusted for pregnancy, lifestyle factors and SES
Never or occasionally	1.00 (reference)
1-4 times per month	0.91 (0.80-1.02)
>Once per week to daily	0.95 (0.75-1.20)
Education
Unadjusted
≤9 years, n (%)	1.63 (1.29-2.06)
9-12 years	1.45 (1.30-1.62)
>12 years	1.00 (reference)
Adjusted for pregnancy, lifestyle factors and SES
≤9 years, n (%)	1.26 (0.97-1.63)
9-12 years	1.21 (1.07-1.37)
>12 years	1.00 (reference)
Income
Unadjusted
Lowest tertile	1.48 (1.30-1.68)
Middle tertile	1.14 (1.00-1.31)
Highest tertile	1.00 (reference)
Adjusted for pregnancy, lifestyle factors and SES
Lowest tertile	1.10 (0.95-1.27)
Middle tertile	0.97 (0.84-1.13)
Highest tertile	1.00 (reference)
BMI, per kg/m2
Unadjusted	1.12 (1.11-1.14)
Adjusted for pregnancy, lifestyle factors and SES	1.11 (1.10-1.13)

^*If the condition was present in ≥ 1 pregnancy

^†The shortest gestational age of all a woman’s pregnancies

The first model adjusts for total parity, maternal age at first birth, lowest gestational age, birth of a small-for gestational age infant, preeclampsia, gestational diabetes, stillbirth, placental abruption. The second model adjusts for current smoking, alcohol consumption, body mass index, education, and income in addition to adjustments in the first model. All analyses are stratified on year of first delivery.

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*gestational age categories.

Sensitivity analysis excluding women delivering prior to 1987

When restricting analyses to women having deliveries 1987 and later, associations were slightly stronger between gestational diabetes and systolic blood pressure but were still not statistically significant (data not shown).

Discussion

Among 40 year old Swedish Women from the Vasterbotten Intervention Programme, at least one of the following pregnancy complications that we studied was present in 17.4% of women (preeclampsia, gestational diabetes, placental abruption, shortest gestational age (<32, 32-36, >=37 weeks), small for gestational age baby (SGA; < 3rd 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 systolic or diastolic blood pressure (or both) as well as the binary outcome of hypertension. We found that placental abruption was associated with higher diastolic blood pressure in women at 40 years after accounting for other pregnancy characteristics, social and life-style factors (education, income, smoking and alcohol use) but not with the binary outcome of hypertension. Preeclampsia had the strongest associations with blood pressure and hypertension at age 40 years as compared to other pregnancy characteristics.

Pregnancy characteristics and blood pressure

Consistent with our findings that parity is associated with modestly lower diastolic blood pressure at age 40, the investigators in the CARDIA study reported that increased parity is associated with lower postpartum blood pressure both proximate to and over two 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 suggests 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 diastolic blood pressure as well as hypertension at age 40 independent of socioeconomic factors and pregnancy complications. However, it also possible that residual sociodemographic factors, not fully explained 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 OR of 3.74 vs. 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 pregnancy⁴ and both prior investigations considered preterm birth dichotomously at 37 weeks. Our prior work suggests that women with very preterm deliveries (i.e. < 32 weeks of gestation) experience higher rates of CVD as compared to those with deliveries between 32 and 36 weeks.¹⁵ We observed that very preterm delivery (< 32 weeks) is related to risk of hypertension, and that moderately preterm delivery (32-36 weeks) was modestly and positively related to systolic blood pressure. These findings indicate that there may be a dose response relationship between degree of preterm delivery and subsequent blood pressure in mothers once other pregnancy characteristics are considered. Tighter blood pressure 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 or not tight blood pressure control in pregnancy led to later decreases in maternal blood pressure during the post partum 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 systolic blood pressure 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 diastolic BP 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 and maternal hypertension or blood pressure is consistent with prior findings and likely reflect the fact the gestational diabetes leads to maternal CVD primarily through its strong association of subsequent type 2 diabetes in the mother.⁴

The importance of primordial prevention of hypertension

Prior epidemiologic evidence suggests that prehypertension and increased blood pressure trajectories in young adults can lead to increased subclinical atherosclerosis in midlife.²⁵ Cumulative increases in blood pressure have been associated with carotid intimal media thickness.²⁶ Decreasing the trajectory of blood pressure by mid-life has been demonstrated to decrease one’s lifetime risk of cardiovascular disease.²⁷ Modifiable risk factors for incident hypertension include obesity, alcohol intake, systolic and diastolic blood pressure in the normal range and cigarette smoking.²⁸ A prior study suggests that having normal blood pressure following preeclampsia is associated with a cardiovascular disease risk score that is equivalent to healthy control study participants without preeclampsia.²⁹ Several lifestyle changes can be successfully undertaken among non-hypertensive individuals to decrease blood pressure 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 blood pressure and leave open the possibility for aggressive risk factor modification to prevent high blood pressure following pregnancy complications as a means to reduce later hypertension in women with the aim of decreasing her lifetime burden of cardiovascular disease.

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 blood pressure. 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 blood pressure and thus is not as accurate as multiple longitudinal blood pressure 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 blood pressure monitoring. The biologic mechanisms underlying these associations are also of scientific and public health importance and warrant further study.

Perspectives

Key pregnancy characteristics selected in our study (i.e. 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 age 40 years, 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 age 40 may be indicated. Currently, women with a complicated pregnancy tend to underutilize post-partum primary care,³² highlighting a potential healthcare utilization gap that can be targeted in the future to decrease the burden of hypertension in women.

Novelty and Significance.

What Is New?

We demonstrate that key pregnancy factors (age at first birth, preeclampsia, having a preterm and SGA-infant, lower parity, stillbirth and placental abruption), when considered together are independently related to higher blood pressure and/or hypertension at age 40 in women.

What Is Relevant?

It may be relevant to collect this pregnancy information in women of childbearing age in order 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.