| | Parity and risk of later-life maternal cardiovascular diseaseReceived 18 September 2009; accepted 18 November 2009. BackgroundPrior studies relating parity with maternal cardiovascular disease (CVD) have been performed in relatively small study samples without accounting for pregnancy-related complications associated with CVD. MethodsWe examined the associations between parity and maternal risk of later-life CVD in a population-based cohort study using data from the Swedish population registers. Women born from 1932 to 1955 were followed until the occurrence of CVD, death, emigration, or end of follow-up (December 31, 2005). Cox proportional hazards models were used to estimate associations between parity and risk of CVD accounting for birth year, yearly income, education level, country of birth, hypertension (pregestational hypertension or gestational hypertension, with or without proteinuria), diabetes (type 1, type 2, or gestational diabetes), preterm birth, small for gestational age, and stillbirth. ResultsDuring a median follow-up time of 9.5 years (range 0-23.5), there were 65,204 CVD events in the full sample of women. Among 1,332,062 women, parity was associated with CVD in a J-shaped fashion, with 2 births representing the nadir of risk (global P value < .0001). Upon accounting for pregnancy-related complications in a subset of women with at least 1 childbirth after 1973 (n = 590,725), the association of parity with CVD was similar. Compared with women with 2 childbirths, the multivariable-adjusted hazard ratios (95% CIs) for women with 1 and ≥5 births were 1.09 (1.03-1.15) and 1.47 (1.37-1.57), respectively. ConclusionsIn conclusion, parity was associated with incident maternal CVD in a J-shaped fashion, even after accounting for socioeconomic factors and pregnancy-related complications. Pregnancy is associated with physiologic changes affecting multiple cardiovascular disease (CVD) pathways, including inflammation, endothelial function, and hemostasis. Prior studies relating parity (number of births) with later-life maternal CVD have yielded conflicting results.1, 2, 3, 4, 5, 6 Whereas some studies report no association between parity and CVD,1, 6 others have demonstrated that the number of childbirths increases CVD risk.2, 3, 4, 5 Although selected prior studies have considered socioeconomic status as confounder of the association between parity and CVD,4, 5 they have not accounted for history of pregnancy-related conditions such as pregnancy-induced hypertension, preeclampsia, gestational diabetes, and intrauterine growth restriction. Such pregnancy-related conditions have been linked to CVD risk factors and CVD7, 8 and may also be associated with lower9, 10 or higher number of pregnancies.11 In addition, many prior studies have been based on relatively small study samples and have not been able to study different CVD outcomes. The Swedish population registries provide unique opportunities to overcome limitations of previous studies relating parity and maternal CVD. We hypothesized that parity would be associated with later-life maternal CVD, but that accounting for pregnancy-related complications such as hypertension, preeclampsia, eclampsia, diabetes, and intrauterine growth restriction would attenuate these associations. Methods  Study population In all, there were 1,429,532 women born from 1932 to 1955 who were registered in Sweden with a personal identity number (PIN) between January 1, 1973, and December 31, 2005. As the objective of the present investigation was to study parity in relation to later-life maternal CVD, the follow-up started at age 50 years. Thus, individuals who emigrated, experienced a CVD event, or died before age 50 years were not considered at risk (n = 48,900). In addition, to make the study population more homogenous and to avoid potential misclassification of exposure, we excluded all women immigrating to Sweden (uncertain parity estimates; n = 24,363) and women with any multiple pregnancy (n = 24,207). After these exclusions, 1,332,062 women were eligible and constituted the full study population. Furthermore, we a priori decided to also perform our analyses in a subsample of 590,725 women giving at least 1 birth between 1973 and 2005, where information on pregnancy and birth characteristics was available from the Swedish Medical Birth Register. The study was approved by the Ethics Committee of Uppsala University, Uppsala, Sweden. Data sources A detailed description of data sources is given in the online Supplement. Briefly, participant data were combined using their unique Swedish PIN number. We combined data from the Multi-Generation Register (for parity and information of changed paternity), the Medical Birth Register (for information about the pregnancy and delivery), and the Hospital Discharge Register and Cause of Death Register (for CVD outcomes). Additional data were obtained from the Swedish Censuses, the Educational Register, the Income Register, the Register of Emigrations and Immigrations, the Register of Stillborn, and the Register of the Total Population. Ascertainment of exposure and covariates A detailed description of exposure and covariates is given in the online Supplement. Briefly, parity (number of births including stillbirths) was derived from the Multi-Generation Register at Statistics Sweden. Highest total income was considered before age 50 years. Highest registered educational level before age 50 years was categorized into 4 groups: primary and secondary school, 2 years of high school (manual, clerical), 3 years of high school (theoretical), and college/university studies. Country of birth was categorized as follows: Sweden, other Nordic countries, other European countries, and non-European countries. From the Medical Birth Register, information about maternal diseases was retrieved from self-reports to the midwife at registration to antenatal care and from International Classification of Diseases (ICD) codes provided by the physician when the women were discharged from hospital after birth. Hypertension (pregestational hypertension or gestational hypertension, with or without proteinuria) was defined by self-reported hypertension at first antenatal visit or by ICD codes. Diabetes (type 1, type 2, or gestational diabetes) was defined by self-reported diabetes at first antenatal visit or by ICD codes. Preterm birth was defined as occurring earlier than 37 completed weeks of pregnancy. Small for gestational age (SGA) was defined as birth weight at least 2 SDs less than the mean birth weight for the gestational age according to the Swedish reference curve for fetal growth.12 Information on stillbirths was collected through the Register of Stillborn and the Register of the Total Population. Follow-up and outcomes A detailed description of follow-up and outcomes is given in the online Supplement. Briefly, follow-up started July 1 in the year the women turned 50 years old. The end of follow-up was December 31, 2005, or date of first occurrence of the following: any CVD event (as defined below), emigration from Sweden, or death. Incidence of CVD was defined as first hospitalization (assessed from the Hospital Discharge Register) or death (assessed from the Cause of Death Register) caused by coronary heart disease, stroke, or heart failure. All CVD outcomes were classified by the International Classification of Diseases, Ninth Revision (ICD-9) and International Classification of Diseases, 10th Revision (ICD-10) codes. Coronary heart disease was defined as unstable angina or acute myocardial infarction. Stroke was defined as cerebral infarction, cerebral hemorrhage, subarachnoidal hemorrhage, transient ischemic attack, or other acute stroke. We only considered hospitalizations or deaths with the above diagnoses as primary diagnosis of hospitalization or primary cause of death to maximize diagnoses validity.13, 14 Statistical methods Parity was analyzed as a categorical variable with 6 levels: 0, 1, 2, 3, 4, and ≥5 births. Birth year was stratified into 4 categories (1932-1937, 1938-1943, 1944-1949, and 1950-1955). We examined associations between parity and 4 outcomes (total CVD, coronary heart disease, stroke, and heart failure) by calculating age-adjusted incidence rates, and by age-adjusted and multivariable-adjusted Cox proportional hazards regressions (with maternal age as underlying time variable). The proportionality of hazards was assessed by plotting smoothed scaled Schoenfeld residuals against time. The associations between parity and outcomes were investigated in 3 models: A—adjusted only for maternal age; B—adjusted for maternal age, birth year, highest income before age 50 years, education level, and country of birth; and C—adjusted for hypertension, diabetes, preterm birth, SGA, and stillbirth (in any pregnancy) in addition to covariates from model B. Models A and B were analyzed in the full study population (N = 1,332,062), whereas all 3 models could be analyzed in the subsample of women giving at least 1 birth in 1973 or later (n = 590,725). We performed a number of prespecified secondary and sensitivity analyses. To decrease the influence of residual social confounding, we performed additional analyses after exclusion of all women having children with different partners (n = 1,208,805 in the full sample and n = 502,088 in the sample of women giving birth after 1973). Second, in a smaller subset (n = 197,630) of women included in the Birth Register, information about smoking habits at registration to antenatal care was available. We added smoking (no, 1-9 cigarettes per day, and >10 cigarettes per day) as a covariate to the fully adjusted model. We examined effect modification by testing the significance of 2-way interaction terms incorporating parity and: hypertension, diabetes, preterm birth, SGA, or stillbirth. Based on the results of the interaction analyses, we performed separate analyses in women without (n = 585,530) and with (n = 5,195) diabetes. Two-tailed 95% CIs and P values were given. To account for multiple statistical testing, we used a Bonferroni-corrected significance level of .0025 (.05 divided by 20; the number of statistical tests for all models in the main sample and subsample). The statistical software package STATA 10.0 MP (Stata Corporation, College Station, TX) was used for all analyses. This work was supported through the Foundation of the National Board of Health and Welfare. E. I. is supported by grants from the Swedish Research Council, the Swedish Heart-Lung Foundation, the Swedish Society of Medicine, the Swedish Foundation for Strategic Research, and the Royal Swedish Academy of Science. J. F. L. was supported by a grant from Örebro University Hospital while contributing to this article. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the article, and its final contents. Results The baseline characteristics of the study samples are shown in Table I. | | |  | | Full sample (N = 1 332 062) | Subsample of women giving birth 1973-2005 (n = 590 725) | |
|---|
| Parity (no. of deliveries), n (%) | | | | |  0 | 186 240 (14) | 0 (0) | | | 1 | 219 160 (16) | 83 628 (14) | | | 2 | 544 062 (41) | 276 610 (47) | | | 3 | 271 264 (20) | 161 558 (27) | | | 4 | 78 922 (6) | 48 695 (8) | | | ≥5 | 32 414 (2) | 20 234 (3) | | | Birth year, n (%) | | | | | 1932-1937 | 255 536 (19) | 12 392 (2) | | | 1938-1943 | 309 561 (23) | 74 399 (13) | | | 1944-1949 | 400 707 (30) | 228 430 (39) | | | 1950-1955 | 366 258 (28) | 275 504 (47) | | | Highest income before age 50 y (USD/y)†, median (25th; 75th percentile) | 21 881 (12 554; 31 240) | 28 125 (20 340; 36 300) | | | Educational level, n (%) | | | | | 0-9 y of primary and secondary school | 472 562 (37) | 129 323 (22) | | | 2 y of high school | 406 329 (32) | 205 603 (35) | | | 3 y of high school | 97 838 (8) | 55 284 (10) | | | College or university studies | 313 554 (24) | 191 022 (33) | | | Country of birth, n (%) | | | | | Sweden | 1 116 558 (84) | 517 323 (88) | | | Other Nordic countries‡ | 110 800 (8) | 40 224 (7) | | | Other European countries | 71 325 (5) | 22 942 (4) | | | Countries from rest of the world | 33 228 (3) | 10 213 (2) | | | Hypertension, n (%) | Not applicable§ | 16 637 (3) | | | Diabetes, n (%) | 5195 (1) | | | Preterm birth, n (%) | 44 569 (8) | | | SGA, n (%) | 35 657 (6) | | | Stillbirth, n (%) | 7100 (1) | | | | |
| ⁎ Data are numbers (percentages) for categorical variables or medians (25th, 75th percentile) for continuous variables. †Originally recorded in Swedish crowns (1 SEK = 0.1245 USD, as of December 19, 2008). ‡Finland, Norway, Denmark, and Iceland. §Data on these parameters were not available until births after 1973. |
Full sample Parity was associated with total CVD, coronary heart disease, stroke, and heart failure in age-adjusted and multivariable-adjusted models (Table II; models A and B, respectively). The associations between parity and risk of CVD were J-shaped, with a nadir in women giving 2 births for all 4 outcomes (Figure 1). Nulliparous, primiparous, and women with 3 pregnancies had about 10% increased risk of total CVD when compared with women giving 2 births. The CVD risk was steeply increased in women giving 4 or more births. The patterns were similar for coronary heart disease and stroke, whereas low and high parities were associated with even more pronounced risks for heart failure. Sample of women giving birth after 1973 In the smaller sample of women giving birth after 1973, the associations between parity and CVD events were similar to those for the full sample (Table III). Additional adjustment for pregnancy and delivery complications (hypertension, diabetes, preterm birth, SGA, and stillbirth in any pregnancy) did not have a major impact on the parity-related risks of CVD (Table III; model C). Additional analyses To reduce the risk of residual social confounding, we investigated the effect of parity in a subsample of women that only had children with the same partner. The associations between parity and CVD risks were similar, but generally slightly attenuated in these subsamples (Supplementary Tables I and II). In analyses adding smoking as a covariate to the fully adjusted models in a considerably smaller subsample where data on smoking were available (n = 197,630), the association of parity with total CVD still was J-shaped, although weaker and of borderline significance (multivariable-adjusted hazard ratio 1.03 and 1.28 for 1 and ≥5 births; women giving 2 births as reference; global P = .032). The interaction term between parity and diabetes was highly significant (P < .0001), whereas the other 4 interaction terms were nonsignificant (P > .25). As expected, the results in women without history of diabetes were very similar to the main results (Supplementary Table III, left panel). In contrast, in women with a history of diabetes, there was a consistent decrease in CVD risk with increasing parity for all 4 outcomes, although these associations were not significant, presumably because of low statistical power (Supplementary Table III, right panel). Finally, we also analyzed associations of parity with different subtypes of stroke (ischemic, hemorrhagic, unspecified, and transient ischemic attack). The results were consistent with the composite stroke end point for the 3 stroke categories, whereas parity was associated with risk of future transient ischemic attacks in a more linear fashion with lowest risk in nulliparous women and a gradually increasing risk with higher parity (Supplementary Table IV). Discussion Principal findings Among 1,332,062 women in the Swedish population registers, we report that parity was related to later-life maternal CVD in a J-shaped fashion, with the nadir of risk among women giving 2 births and with the highest risk among women giving ≥5 births. Results were independent of adjustment for potential confounding by pregnancy-related complications or socioeconomic factors; and the associations were similar in specific analyses of coronary heart disease, stroke, and heart failure. Comparison with the prior literature Our findings complement prior longitudinal data that have demonstrated a positive association between parity and later-life maternal CVD.2, 3, 4, 5 An association between nulliparity and CVD has also been previously noted, although it was not statistically significant.1 A British study demonstrated a similar J-shaped association between parity and coronary heart disease among 4,286 women and 4,252 men studied.4 One prior study, combining data from 4,890 women in the Framingham Heart Study and the National Health and Nutrition Examination Survey, demonstrated a slight CVD risk increase in women having 6 or more pregnancies, after accounting for age, CVD risk factors, and educational level.5 Prior reports from relatively socioeconomically homogenous populations did not find such an association after accounting for CVD risk factors.1, 6 Reasons for the lack of association in these studies could include inability to account for history of pregnancy-related cardiometabolic conditions and residual effects of socioeconomic confounding. Although we accounted for educational level, income, and changed paternity, it is possible that there was some residual social confounding in our analyses. However, one prior study demonstrating an association between parity and later-life maternal CVD did not show a similar association between number of children and later-life paternal CVD.15 Although one prior study examined effects of adverse pregnancy outcomes on the association between parity and CVD, this study was limited by a small sample size.16 Our results demonstrate that pregnancy-induced hypertension, preeclampsia, gestational diabetes, and intrauterine growth restriction do not explain the J-shaped association between parity and CVD. In women with diabetes, there was a nonsignificant but consistent trend of decreasing CVD risk with increasing parity. This should not be interpreted to mean that having many children will decrease the CVD risk among women with diabetes. Rather, the plausible explanation of this finding is that women with more severe diabetes (and higher CVD risk) are discouraged, unwilling, or unable to have children, in contrast to women with milder forms of diabetes. Possible mechanisms Epidemiologic studies have demonstrated that increasing number of pregnancies is associated with maternal dyslipidemia,17 adiposity,18 and type 2 diabetes.19 Two prior studies that accounted for these CVD risk factors still demonstrated a significant positive association between parity and CVD.4, 5 Normal pregnancy is associated with physiologic changes that alter mediators in various cardiovascular pathways. For example, during pregnancy, there is up-regulation of the renin-angiotensin-aldosterone system20; increased peripheral insulin resistance21; lipid changes including increased free fatty acids, triglycerides, and cholesterol22; and alterations in immune modulation, hemostasis, and endothelial function.23 It is possible that these alterations may result in cumulative derangements over successive pregnancies that eventually lead to increased CVD risk. Nulliparity conferred a modest risk of maternal CVD in our analyses. It is unclear what proportion of nulliparous women had a history of involuntary infertility and what proportion had made an active decision not to have children. Assuming that there was a relatively high incidence of infertility among nulliparous women in our population, it is possible that conditions such as polycystic ovarian syndrome, which itself is related to cardiometabolic risk factors,24 may have contributed to our findings. Thyroid hormone disorders, which have increased prevalence among women experiencing infertility,25 have also been modestly linked to incident CVD26 and could partly mediate the excess risk associated with nulliparity. The excess risk for women having 1 compared with 2 pregnancies may also reflect involuntary infertility among women having only 1 term pregnancy. As this is a population-based study, it is not possible to tease out details about which mechanisms connect parity and specific CVD outcomes. Rather, it should be noted that total CVD, as well as the more specific stroke and heart failure end points, comprises a whole range of different pathophysiologies. It is likely that no matter what the mechanisms by which parity is associated with different types of CVD, those mechanisms will be operative in only some of the pathophysiologic processes that can cause disease. Strengths and limitations The strengths of our study include the very large sample size that allowed us to study the association between parity and CVD among several parity categories. The study sample included almost all Swedish women born during the study period and thus was highly representative of a community-based sample in Northern Europe. Our detailed data on socioeconomic factors and our ability to account for pregnancy-related complications were other unique strengths of this study. However, several limitations should also be acknowledged. We were unable to assess the association of parity and CVD in different ethnic groups because of the constraints of studying the relatively homogenous population of women living in Sweden. Although the Swedish Medical Birth, the Hospital Discharge, and the Cause of Death Registers are very nearly complete and the validity of the studied CVD end points is known to be high, the registers only capture hospitalized cases of CVD. Thus, we may have missed milder cases of CVD that only were treated as outpatient patients. On the other hand, this would be expected to increase the validity of the end points, limiting the inclusion of false-positive cases. However, we acknowledge the possibility that milder types of CVD might not have as strong association with parity as the more severe forms requiring hospitalization. Our study can only establish an association of parity and these more severe appearances of CVD requiring hospitalization. Finally, we were unable to account for maternal obesity and dyslipidemia, which are important CVD risk factors. However, if at all associated with parity, these factors might rather be on the causal pathway between parity and maternal CVD and could therefore be considered potential mediators of the associations observed. Implications Parity is associated with incident maternal CVD in a J-shaped fashion. Gaining a better understanding of the mechanisms that lead to CVD among nulliparous women as well as women with a large number of childbirths may lead to the uncovering of novel CVD pathways. Furthermore, CVD risk stratification with the “classic” CVD risk markers of lipids, diabetes, hypertension, and smoking does not fully characterize CVD risk in women.27 Whether or not reproductive risk factors such as parity may be useful additions to the existing tools for CVD risk stratification is not known, but may be an important area of future research given the relative ease of collecting such information during a routine cardiovascular risk factor screening examination. Disclosures None of the authors have any conflicts of interest to declare. Appendix A | | | | Parity | Without history of diabetes (n = 585 530)† | With history of diabetes (n = 5195)† | |
|---|
| No. of CVD events (Rate/1000 PYAR) | HR (95% CI) | P value | No. of CVD events (Rate/1000 PYAR) | HR (95% CI) | P value | |
|---|
| All CVDs (no. of events 13 224 [without history of diabetes]/339 [with history of diabetes]) | | | 1 | 1641 (3.13) | 1.08 (1.03-1.15) | <.0001 | 58 (13.37) | 1.04 (0.76-1.43) | .083 | | | 2 | 4964 (2.81) | 1 | 122 (12.40) | 1 | | | 3 | 3873 (3.40) | 1.11 (1.06-1.16) | 90 (12.17) | 0.92 (0.70-1.22) | | | 4 | 1665 (4.39) | 1.26 (1.19-1.33) | 43 (10.98) | 0.74 (0.52-1.06) | | | ≥5 | 1081 (6.20) | 1.51 (1.41-1.62) | 26 (9.87) | 0.60 (0.39-0.93) | | | Coronary heart disease (no. of events 5310/181) | | | 1 | 612 (1.17) | 1.05 (0.96-1.15) | <.0001 | 32 (7.38) | 1.03 (0.67-1.57) | .18 | | | 2 | 1900 (1.07) | 1 | 69 (7.01) | 1 | | | 3 | 1600 (1.40) | 1.18 (1.10-1.26) | 41 (5.55) | 0.74 (0.50-1.10) | | | 4 | 716 (1.89) | 1.36 (1.25-1.49) | 25 (6.39) | 0.75 (0.47-1.20) | | | ≥5 | 482 (2.77) | 1.36 (1.25-1.49) | 14 (5.31) | 0.56 (0.31-1.01) | | | Stroke (no. of events 6967/128) | | | 1 | 888 (1.70) | 1.07 (0.99-1.15) | <.0001 | 22 (5.07) | 1.09 (0.65-1.83) | .38 | | | 2 | 2744 (1.55) | 1 | 44 (4.47) | 1 | | | 3 | 2014 (1.77) | 1.06 (1.00-1.13) | 39 (5.28) | 1.16 (0.75-1.79) | | | 4 | 830 (2.19) | 1.18 (1.09-1.28) | 14 (3.58) | 0.72 (0.39-1.33) | | | ≥5 | 491 (2.82) | 1.33 (1.20-1.47) | 9 (3.42) | 0.65 (0.31-1.37) | | | Heart failure (no. of events 949/30) | | | 1 | 141 (0.27) | 1.42 (1.16-1.73) | <.0001 | 4 (0.92) | 0.94 (0.29-3.11) | .90 | | | 2 | 320 (0.18) | 1 | 9 (0.91) | 1 | | | 3 | 260 (0.23) | 1.10 (0.93-1.30) | 10 (1.35) | 1.22 (0.48-3.06) | | | 4 | 120 (0.32) | 1.27 (1.02-1.58) | 4 (1.02) | 0.77 (0.23-2.55) | | | ≥5 | 108 (0.62) | 1.92 (1.52-2.43) | 3 (1.14) | 0.71 (0.18-2.74) | | | | |
| ⁎ The median follow-up time was 6.5 years (range 0-23.5, 25th-75th percentile 3.5-9.5) contributing to 4.0 × 106 person-years at risk in women without diabetes and 4.5 years (range 0-23.5, 25th-75th percentile 2.2-7.5) contributing to 28 118 person-years at risk in women with diabetes. †Values are unadjusted incidence rates and Cox proportional hazards ratios (95% CIs) for parity (number of children); women with parity = 2 were used as a reference level. P values are from likelihood ratio tests for differences across levels of parity. The hazard ratios are adjusted for age, birth year, highest income before age 50 years, education level, country of birth, hypertension, diabetes, preterm birth, SGA, and stillbirth (in any pregnancy; model C). |
Online supplementary methods Study population In all, there were 1,429,532 women born from 1932 to 1955 who were registered in Sweden with a PIN between January 1, 1973, and December 31, 2005. As the objective of the present investigation was to study parity in relation to later-life maternal CVD, the follow-up started at age 50 years. Thus, individuals who emigrated, experienced a CVD event, or died before age 50 years were not considered at risk (n = 48,900). In addition, to make the study population more homogenous and to avoid potential misclassification of exposure, we excluded all women immigrating to Sweden (uncertain parity estimates, n = 24,363) and women with any multiple pregnancy (n = 24,207). After these exclusions, 1,332,062 women were eligible and constituted the full study population. Furthermore, we a priori decided to also perform our analyses in a subsample of 590,725 women giving at least 1 birth between 1973 and 2005, where information on pregnancy and birth characteristics was available from the Swedish Medical Birth Register. The study was approved by the Ethics Committee of Uppsala University, Uppsala, Sweden. Data sources The Multi-Generation Register The Multi-Generation Register (maintained at Statistics Sweden; www.scb.se) is made up of all individuals with a Swedish PIN born in 1932 or later who have been registered in Sweden since 1961. The registry connects these individuals with their biological or adoptive parents. The study population was defined using this registry and then linked to information in the other registries by use of unique PINs. The Medical Birth Register, Hospital Discharge Register, and Cause of Death Register The Swedish Medical Birth Register, Hospital Discharge Register, and Cause of Death Register are maintained at the National Board of Health and Welfare (http://www.socialstyrelsen.se/en/). These registers use the International Classification of Diseases, Eighth Revision (ICD-8) until 1986, ICD-9 from 1987 to 1996, and ICD-10 from 1997 and onward. The Medical Birth Register was initiated in 1973 as a means to compile information on maternal, obstetric, and infant factors. Between 1973 and 1998, >98% of all births in Sweden were registered in the Medical Birth Register.28 The Swedish Hospital Discharge Register started to collect data on inpatients treated at public hospitals beginning in the 1960s. From 1978, it covered a geographical area corresponding to 85% of the Swedish population; and from 1987, the coverage is nationwide. The registry contains admission and discharge dates, and discharge diagnosis codes, the first representing the principal cause of hospitalization. The Swedish Cause of Death Register includes information about death date, and primary and contributing causes of death. Other registries In addition, from Statistics Sweden, we received information from the Swedish Censuses of 1970 and 1990, the Educational Register (years 1995, 2000, and 2005), the Income Register (years 1973, 1980, 1985, 1990, 1995, 2000, and 2005), the Register of Emigrations and Immigrations (1961-), the Register of Stillborn (1961-1997), and the Register of the Total Population (1998-2005). Ascertainment of exposure and covariates Parity (number of births including stillbirths) was derived from the Multi-Generation Register at Statistics Sweden. Highest total income before age 50 years was collected from the Income Register. Educational level according to the Swedish Educational Nomenclature was collected from the Swedish Censuses of 1970 and 1990 and the Educational Register in the years 1995, 2000, and 2005. The highest registered educational level before age 50 years was categorized into 4 groups: primary and secondary school, 2 years of high school (manual, clerical), 3 years of high school (theoretical), and college/university studies. Country of birth was categorized into 4 groups: Sweden, other Nordic countries, other European countries, and non-European countries. From the Medical Birth Register, information about maternal diseases was retrieved from self-reports to the midwife at registration to antenatal care and from ICD codes provided by the physician when the women were discharged from hospital after birth. Hypertension (pregestational hypertension or gestational hypertension, with or without proteinuria) was defined by self-reported hypertension at first antenatal visit or by ICD codes (ICD-8 codes 400-404 and 637; ICD-9 codes 401-405 and 642; ICD-10 codes I10-15, O10-O11, and O13-O16). Diabetes (type 1, type 2, or gestational diabetes) was defined by self-reported diabetes at first antenatal visit or by ICD codes (ICD-8 code 250; ICD-9 codes 250, 648A, and 648W; and ICD-10 codes E10-E11 and O24). Preterm birth was defined as occurring earlier than 37 completed weeks of pregnancy. Small for gestational age was defined as birth weight at least 2 SDs less than the mean birth weight for the gestational age according to the Swedish reference curve for fetal growth.12 Information on stillbirths was collected through the Register of Stillborn and the Register of the Total Population. Follow-up and outcomes Follow-up started July 1 in the year the women turned 50 years old, for the following reasons: (1) The aims of the study were to examine parity in relation to later-life maternal CVD, not pregnancy complications with acute or subacute CVD. (2) We could ascertain correct parity estimates by starting follow-up after childbearing years. (3) The Swedish Hospital Discharge Register did not have full coverage until the 1980s; so by choosing the current design, we minimized left truncation of data. End of follow-up was December 31, 2005, or date of first occurrence of the following: any CVD event (as defined below), emigration from Sweden, or death. Incidence of CVD was defined as first hospitalization (assessed from the Hospital Discharge Register) or death (assessed from the Cause of Death Register) caused by coronary heart disease, stroke, or heart failure. Coronary heart disease was defined as unstable angina (ICD-8 code 411, ICD-9 code 411B, ICD-10 code I20.0) or acute myocardial infarction (ICD-8 and ICD-9 code 410, ICD-10 codes I21-I22). Stroke was defined as cerebral infarction (ICD-8 codes 432-434, ICD-9 codes 433-434, ICD-10 code I63), cerebral hemorrhage (ICD-8 code 431, ICD-9 codes 431-432, ICD-10 codes I61-I62), subarachnoidal hemorrhage (ICD-8 and ICD-9 code 430, ICD-10 code I60), transient ischemic attack (ICD-8 and ICD-9 code 435, ICD-10 code G45), or other acute stroke (ICD-8 and ICD-9 code 436, ICD-10 code I64). Heart failure was defined by ICD-8 codes 427.00 and 427.10, ICD-9 code 428, or ICD-10 code I50. We only considered hospitalizations or deaths with the above diagnoses as primary diagnosis of hospitalization or primary cause of death. The positive predictive values (ie, validity) of the myocardial infarction,13 stroke,13 and heart failure14 diagnoses in the Swedish hospital discharge register have been demonstrated to be around 95% when only primary diagnoses are considered. References 1. 1Colditz GA, Willett WC, Stampfer MJ, et al. A prospective study of age at menarche, parity, age at first birth, and coronary heart disease in women. Am J Epidemiol. 1987;126:861–870. MEDLINE 2. 2Dekker JM, Schouten EG. Number of pregnancies and risk of cardiovascular disease. N Engl J Med. 1993;329:1893–1894. MEDLINE |
CrossRef
3. 3Green A, Beral V, Moser K. Mortality in women in relation to their childbearing history. BMJ. 1988;297:391–395. 4. 4Lawlor DA, Emberson JR, Ebrahim S, et al. Is the association between parity and coronary heart disease due to biological effects of pregnancy or adverse lifestyle risk factors associated with child-rearing? Findings from the British Women's Heart and Health Study and the British Regional Heart Study. Circulation. 2003;107:1260–1264.
CrossRef
5. 5Ness RB, Harris T, Cobb J, et al. Number of pregnancies and the subsequent risk of cardiovascular disease. N Engl J Med. 1993;328:1528–1533. MEDLINE |
CrossRef
6. 6Steenland K, Lally C, Thun M. Parity and coronary heart disease among women in the American Cancer Society CPS II population. Epidemiology. 1996;7:641–643. MEDLINE 7. 7Carr DB, Utzschneider KM, Hull RL, et al. Gestational diabetes mellitus increases the risk of cardiovascular disease in women with a family history of type 2 diabetes. Diabetes Care. 2006;29:2078–2083. MEDLINE |
CrossRef
8. 8Irgens HU, Reisaeter L, Irgens LM, et al. Long term mortality of mothers and fathers after pre-eclampsia: population based cohort study. BMJ. 2001;323:1213–1217. 9. 9Duckitt K, Harrington D. Risk factors for pre-eclampsia at antenatal booking: systematic review of controlled studies. BMJ. 2005;330:565. 10. 10Ego A, Subtil D, Grange G, et al. Should parity be included in customised fetal weight standards for identifying small-for-gestational-age babies? Results from a French multicentre study. BJOG. 2008;115:1256–1264.
CrossRef
11. 11Ben-Haroush A, Yogev Y, Hod M. Epidemiology of gestational diabetes mellitus and its association with type 2 diabetes. Diabet Med. 2004;21:103–113. MEDLINE |
CrossRef
12. 12Marsal K, Persson PH, Larsen T, et al. Intrauterine growth curves based on ultrasonically estimated foetal weights. Acta Paediatr. 1996;85:843–848. MEDLINE |
CrossRef
13. 13Lindblad U, Rastam L, Ranstam J, et al. Validity of register data on acute myocardial infarction and acute stroke: the Skaraborg Hypertension Project. Scand J Soc Med. 1993;21:3–9. MEDLINE 14. 14Ingelsson E, Ärnlöv J, Sundström J, et al. The validity of a diagnosis of heart failure in a hospital discharge register. Eur J Heart Fail. 2005;7:787–791. MEDLINE |
CrossRef
15. 15Ness RB, Cobb J, Harris T, et al. Does number of children increase the rate of coronary heart disease in men?. Epidemiology. 1995;6:442–445. MEDLINE 16. 16Catov JM, Newman AB, Sutton-Tyrrell K, et al. Parity and cardiovascular disease risk among older women: how do pregnancy complications mediate the association?. Ann Epidemiol. 2008;18:873–879. Abstract | Full Text |
Full-Text PDF (130 KB)
|
CrossRef
17. 17Lewis CE, Funkhouser E, Raczynski JM, et al. Adverse effect of pregnancy on high density lipoprotein (HDL) cholesterol in young adult women. The CARDIA Study. Coronary Artery Risk Development in Young Adults. Am J Epidemiol. 1996;144:247–254. MEDLINE 18. 18Smith DE, Lewis CE, Caveny JL, et al. Longitudinal changes in adiposity associated with pregnancy. The CARDIA Study. Coronary Artery Risk Development in Young Adults Study. JAMA. 1994;271:1747–1751. MEDLINE 19. 19Nicholson WK, Asao K, Brancati F, et al. Parity and risk of type 2 diabetes: the Atherosclerosis Risk in Communities Study. Diabetes Care. 2006;29:2349–2354. MEDLINE |
CrossRef
20. 20Gallery EDM, Lindheimer MD. Alterations in volume homeostasis. In: Lindheimer ML, Roberts JM, Cunningham FG editor. Chelsey's hypertensive disease in pregnancy. 2 ed.. Stamford (CT): Appelton and Lange; 1999;. 21. 21Catalano PM, Huston L, Amini SB, et al. Longitudinal changes in glucose metabolism during pregnancy in obese women with normal glucose tolerance and gestational diabetes mellitus. Am J Obstet Gynecol. 1999;180:903–916. Abstract | Full Text |
Full-Text PDF (200 KB)
|
CrossRef
22. 22Brizzi P, Tonolo G, Esposito F, et al. Lipoprotein metabolism during normal pregnancy. Am J Obstet Gynecol. 1999;181:430–434. Abstract | Full Text |
Full-Text PDF (40 KB)
|
CrossRef
23. 23Stewart FM, Freeman DJ, Ramsay JE, et al. Longitudinal assessment of maternal endothelial function and markers of inflammation and placental function throughout pregnancy in lean and obese mothers. J Clin Endocrinol Metab. 2007;92:969–975. 24. 24Wild RA. Polycystic ovary syndrome: a risk for coronary artery disease?. Am J Obstet Gynecol. 2002;186:35–43. Abstract | Full Text |
Full-Text PDF (94 KB)
|
CrossRef
25. 25Arojoki M, Jokimaa V, Juuti A, et al. Hypothyroidism among infertile women in Finland. Gynecol Endocrinol. 2000;14:127–131. MEDLINE |
CrossRef
26. 26Ochs N, Auer R, Bauer DC, et al. Meta-analysis: subclinical thyroid dysfunction and the risk for coronary heart disease and mortality. Ann Intern Med. 2008;148:832–845. 27. 27Lakoski SG, Greenland P, Wong ND, et al. Coronary artery calcium scores and risk for cardiovascular events in women classified as “low risk” based on Framingham risk score: the multi-ethnic study of atherosclerosis (MESA). Arch Intern Med. 2007;167:2437–2442.
CrossRef
28. 28Centre for Epidemiology The National Board of Health and Welfare . The Swedish Medical Birth Register—a summary of content and quality. 2003;. a Cardiovascular Division and Cardiovascular Epidemiology and Research Unit, Department of Medicine, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA b Clinical Epidemiology Unit, Department of Medicine, Karolinska University Hospital, Karolinska Institutet, Sweden c Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden d Department of Pediatrics, Örebro University Hospital, Örebro, Sweden  Reprint requests: Erik Ingelsson, MD, PhD, Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Box 281, SE-171 77 Stockholm, Sweden.
PII: S0002-8703(09)00897-7 doi:10.1016/j.ahj.2009.11.017 © 2010 Mosby, Inc. All rights reserved. | |
|
FULL TEXT00897-7/journalimage?img=PIIS0002870309008977.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0002-8703&ishighres=false&allhighres=false&free=yes','journalimage');)
