点击显示 收起
the Klinik für Innere Medizin B (B.W., D.R., S.B.F.), Institut für Epidemiologie und Sozialmedizin (H.V., J.L., U.J.), Klinik für Stomatologie und Zahnheilkunde (C.S.), and Klinik für Neurologie (C.K.), Universitt Greifswald, Greifswald, Germany.
Abstract
Background and Purpose— Metabolic and hormonal changes associated with pregnancy and childbirth are assumed to contribute to the development of cardiovascular disease among women. We analyzed the association of parity with common carotid intima-media thickness (IMT), which has a predictive value of subsequent myocardial infarction and stroke.
Methods— The Study of Health in Pomerania (SHIP), an epidemiological study of the general population in the northeast of Germany, included 1195 women aged 45 to 79 years. Mean and maximum far-wall IMT of the common carotid arteries were assessed by high-resolution ultrasound. All women were comprehensively characterized as to their reproductive history as well as to socioeconomic, behavioral, and biological risk factors.
Results— There was a U-shaped association between the number of children (from 0 to 4) and mean and maximum IMT. Nulliparous women had the highest age-adjusted mean (0.81 mm [95% CI, 0.78 to 0.84]) and maximum IMT (1.04 mm [95% CI, 1.00 to 1.09]), and women with single parity the lowest (mean IMT, 0.73 [95% CI, 0.72 to 0.74]; maximum IMT, 0.91 mm [95% CI, 0.89 to 0.93]; P<0.001 versus nulliparity for both parameters). Stepwise multivariate adjustment for socioeconomic factors, lifestyle variables, and biological variables attenuated the magnitude of this association yet significance remained.
Conclusions— Nulliparity and higher number of children are associated with increased carotid IMT. These findings add support to the hypothesis of a link between the reproductive history of women and cardiovascular disease.
Key Words: atherosclerosis epidemiology
Introduction
Pregnancy and childbirth are accompanied by profound changes in metabolism as well as in endocrine regulation and activity. Not only do these changes concern altered blood lipoprotein levels,1–4 increased insulin concentrations,5 and enhanced generation of reactive oxygen species,6 but also structural and functional alterations of cardiovascular regulation.7,8 Therefore, the hypothesis of a relation between the childbearing experience of women and cardiovascular risk has a great deal of biological plausibility.
Several studies have investigated this issue, yet produced equivocal results. Some investigations reported a modestly higher mortality9,10 and a higher incidence of coronary heart disease (CHD)11 and ischemic stroke12 with increasing number of births. But others found no association13 or even described opposite effects with nulliparity or lower parity rates conferring higher cardiovascular risk.14
Investigations on the putative link of parity with cardiovascular disease are hampered by a low incidence of cardiovascular end points among young childbearing women. However, markers of generalized atherosclerosis such as carotid intima-media thickness (IMT), which have a predictive value of subsequent myocardial infarction and stroke,15,16 may offer an alternative. To date, only 1 study17 has addressed the link between parity and carotid atherosclerosis among elderly women. Therefore, this issue has not yet attracted adequate attention. We undertook to further investigate this relationship and examined the association of parity with carotid IMT among female participants of the Study of Health in Pomerania (SHIP), an epidemiological study in a general population sample.
Methods
Study Population
SHIP18 is a cross-sectional study of the population in West Pomerania, the northeast coastal region of Germany. From the total population of 212 157 people living in the study area, a random sample of subjects aged 20 to 79 years was drawn. The sample was selected using population registries and performed by a 2-stage cluster sampling method that was adopted from the World Health Organization Monitoring Trends Determinants in Cardiovascular Disease (WHO MONICA) Project of Augsburg, Germany.19 In a first step, 3 cities (17 076 to 65,977 inhabitants) and 12 towns (1516 to 3044 inhabitants) and villages (<1500 inhabitants) were selected. Additionally included were 17 randomly selected communities out of the overall 97. In a second step, subjects were drawn at random from each of the selected communities, proportional to the population size of each community and stratified by age and gender. Finally, 7008 subjects with German citizenship and main residency in the study area were sampled. The net sample (without migrated or deceased persons) comprised 6267 eligible subjects. Selected persons received a maximum of 3 written invitations. In case of nonresponse, letters were followed by phone calls and home visits. The final sample comprised 4310 participants (2193 women). Ultrasound investigation of the carotid arteries was performed for 1248 women (out of a total of 1267 women aged 45 to 79 years). A further 53 women with incomplete data were excluded. This resulted in a final sample of 1195 women who were available for the current analysis. The study was approved by the ethics committee of the University of Greifswald, and all participants gave informed written consent.
Sociodemographic characteristics and reproductive as well as medical histories were assessed by computer-aided face-to-face interviews. Systolic and diastolic blood pressure (BP) were measured 3x in seated subjects after a 5-minute rest period, with each reading being followed by a further rest period of 3 minutes. Mean BP was calculated. Hypertension was defined as a systolic BP of 140 mm Hg, a diastolic BP of 90 mm Hg, or self-reported use of antihypertensive medication. Study participants were assigned to 1 of 3 smoking categories (never, former, current). The physical activity was assessed by 4 questions concerning the frequency and intensity of sport during summer and winter, and the study participants were assigned to 1 of 2 categories (active, nonactive). Alcohol consumption was quantified in grams of pure alcohol per day. Diabetes was defined as self-reported physician diagnosis of diabetes or serum hemoglobin A1C (HbA1C) of >7.0%. Height and weight were measured for calculation of body mass index (BMI). Nonfasting blood samples were taken and analyzed in a central laboratory. HbA1C was determined per high-performance liquid chromatography (Bio-Rad Diamat). High-density lipoproteins (HDLs) and low-density lipoproteins (LDLs) were measured photometrically (Boehringer Mannheim). Blood samples were analyzed in a central laboratory that participated semiannually in the official German tests for quality assurance. In addition, duplicate blood samples were collected and measured for internal quality assurance tests every week.
Ultrasound Measurements
The ultrasound protocol has been described previously.20 In brief, certified medical assistants examined the extracranial carotid arteries bilaterally with B-mode ultrasound using a 5-MHz linear array transducer and a high-resolution instrument (Diasonics VST Gateway). For measurement of carotid IMT, scans through the axis of the distal straight portion (1 cm in length) of both common carotid arteries were digitized and recorded for subsequent offline analysis. Certified readers calculated the mean far-wall IMT by averaging the 10 consecutive measurement points (in 1-mm steps) from the bulb of both sides. The IMT was defined as the distance between the characteristic echoes from the lumen–intima and media–adventitia interfaces. Maximal IMT was defined as the greatest value of those measurement points of both sides. Reproducibility studies for IMT measurements have been performed between paired measurements of sonographers and readers. All measurements of intrareader, intrasonographer, inter-reader, and intersonographer variations revealed coefficients of variations of 2.6%, Spearman correlation coefficients of >0.90, and differences in mean IMT (2 SD) of <1.0% (<10.0%).
Statistical Analyses
Comparisons between parity groups were made using 1-way ANOVA for continuous variables, and 2 test for categorical variables. The association of parity with IMT was analyzed by linear regression and ANCOVA. The following variables were considered as potential confounders: age, hypertension, diabetes, smoking status, BMI, alcohol consumption, HDL and LDL cholesterol, hormone replacement therapy, age at menopause, family income, marital status, primary education, vocation, physical exercise, and use of oral contraceptives. There is evidence from previous studies5,17 that some of these potential confounders, notably lipid profiles (HDL and LDL cholesterol) and diabetes, may not represent confounders but actually causal mediators. Thus, to create a causality-based model, these variables were not included as cofounders into the multivariate analyses. Collinearity was assessed by using the tolerance and variation–inflation factor (VIF). Collinearity was found if the tolerance was <0.2 and the VIF >5, respectively. A P value of <0.05 was considered statistically significant. All analyzes were performed with SPSS software (version 11.0; SPSS).
Results
The number of children per women ranged from 0 to a maximum of 14 (n=1). To create study groups with adequate numbers of subjects, women with 4 children were combined into 1 group. The clinical characteristics of all study participants are outlined in Table 1. Compared with women with 2 children, those with <2 and those with >2 children were slightly older, had marginally higher systolic BP values and, accordingly, there was a greater proportion of subjects with hypertension. The proportion of diabetics was lowest in the group of childless women as well as among those with 2 children. Women with 1 or 2 children had somewhat lower LDL cholesterol and BMI levels, whereas their HDL cholesterol levels were minimally higher. Women with 1 or 2 children also tended to have higher levels of primary education. Moreover, these 2 groups had a lower proportion of subjects without completed vocational training, and greater proportions of subjects with higher vocational training. In contrast, the proportion of subjects with college or university graduation fell across the parity groups with nulliparous women having the highest and multiparous women (4 children) having the lowest proportion. Nulliparous women showed the greatest proportion of singles, whereas the proportion of married/divorced women was greatest among those with 1 to 3 children. The proportion of widowed women increased with higher number of children. Women with 1 or 2 children had higher household incomes, demonstrated a greater degree of physical activity, and less often pursued a sedentary lifestyle.
As demonstrated in Tables 2 and 3, the association between parity and crude mean and maximum carotid IMT values tended to show a U-shaped relationship (P<0.001 for quadratic terms). Women with 1 and 2 children exhibited the lowest mean as well as maximum carotid IMT values, whereas nulliparity and higher number of children (>2) were associated with higher IMT values. To separately assess the impact of different sets of potential confounders, 3 consecutive models of multivariate adjustment were created. The first model included age and socioeconomic variables (primary education, vocation, marital status, and per capita household income) as covariates, the second included lifestyle variables (BMI, smoking, alcohol consumption, and degree of physical activity). The third model was obtained through introduction of biological confounders (hypertension, hormone replacement therapy, age at menopause, and use of oral contraceptives). As also shown in Tables 2 and 3, the relation between parity and mean as well as maximum IMT obtained after age adjustment was nearly entirely maintained after additional adjustment for socioeconomic factors and lifestyle variables. Full adjustment, which also included biological variables, somewhat attenuated the association for the group with 4 parities yet not nulliparity.
Discussion
In this study, we observed an association of parity with carotid IMT in which women with a single child demonstrated the lowest IMT values, whereas multiparous and, in particular, nulliparous women showed higher IMT values. Because carotid IMT has predictive value for myocardial infarction and stroke,15,16 our findings suggest an association of the childbearing history of women with cardiovascular disease, and this would be consistent with several previous examinations. A study among women from 2 epidemiological surveys in the United States by Ness et al9 found an increased number of cardiovascular end points among women with multiple pregnancies, whereas a further study by Qureshi et al12 observed higher rates of ischemic stroke among multiparous women. A study by Green et al10 that included married women from a census registry observed a trend toward higher cardiovascular mortality among multiparous women but also an increased all-cause mortality among nulliparous women.
A principal finding of our study was that the relation between number of children and carotid IMT pursued a U-shaped pattern. The relation was paralleled by similar associations with conventional cardiovascular risk factors: age, systolic BP, lipid profiles, diabetes, hypertension, and BMI. This pattern is in contrast to 2 other investigations in which cardiovascular risk increased with greater number of children and nulliparity conferred the lowest risk: Kritz-Silverstein et al5 observed an increasing risk for diabetes with higher number of children. Similar findings were raised by Humphries et al17 in the Rotterdam study, which described a trend toward higher BMI values, increasing insulin resistance, and higher total/HDL cholesterol ratios with higher number of children. Furthermore, an association was reported between the number of children and carotid atherosclerosis as well as carotid IMT, both of which increased linearly with higher numbers of children. Thus, it might be assumed that the association between parity and carotid IMT in our study is solely attributable to the specific distribution of risk factors. However, because the association was maintained even after adjustment, we hypothesize that some modest yet independent proatherogenic effect of parity might be involved. This assumption would be consistent with findings from the Rotterdam study, although the association described there was not U-shaped but more linear. However, this finding is only seemingly in contradiction to ours. First, albeit less pronounced than in our examination, nulliparous women in the Rotterdam study also exhibited increased carotid IMT values. Secondly, whereas nulliparous women made up a mere 8.5% in our investigation, the latter had a higher proportion of 20% in the Rotterdam study, and these showed particularly favorable socioeconomic characteristics. This suggests that many childless women in the Rotterdam study were socially well situated and deliberately childless. In contrast, most women in our study had grown up in the former East Germany, where childbearing was strongly encouraged and generously subsidized and where childlessness was hence uncommon. Thus, it seems justified to presume that in our study, the group of childless women was, for 1 part, made up of well-educated, more affluent women who are voluntarily childless, and for the other part, of women who, for reasons to be determined, were incapable of begetting children. These women may, for example, have been affected by polycystic ovarian syndrome (PCOS), which has a prevalence of 4% to 7% among women21 and is a major cause of infertility.22 PCOS is characterized by insulin resistance, obesity, hypertension,23,24 and accelerated development of atherosclerosis.25 Although speculative, PCOS might be an explanation for increased IMT values among childless women in our study. Furthermore, it is established that women who experience pregnancy-related illnesses and complications, such as (pre)eclampsia or miscarriages and who consequently tend to remain childless, have an increased proneness to develop conditions like manifest hypertension26 or diabetes27 in later adult life. This could also account for higher IMT values in this group. Thus, it is reasonable to speculate that the U-shaped relation between parity and carotid IMT in our study may be the result of 2 opposite trends: 1 in which women with inherent proneness for cardiovascular disease, such as PCOS, remain nulliparous, and another in which increasing number of births promote the development of atherosclerosis.
Because the number of children is also linked to socioeconomic conditions, possible impact by such factors deserves attention. A recent study by Lawlor et al11 investigated the prevalence of CHD in relation to the number of children, to risk factors, and to potential socioeconomic confounders in a cohort of women and men. In this study, a similar U-shaped relationship was found between parity and prevalent CHD, which was lowest among subjects with 2 children. This association was nearly alike in both sexes and was markedly reduced in magnitude after adjustment. This finding led to the conclusion that the association is largely determined by adverse social factors rather than by biological or metabolic factors. However, a residual independent association of parity with CHD among women, yet not among men, was also found, and this is consistent with our findings. In our study, the association between parity and carotid IMT remained even after adjustment for socioeconomic and lifestyle variables. However, it was reduced in magnitude and significance after further adjustment for potential biological confounders, notably hypertension, diabetes, lipids, and hormonal influences, was performed. However, this observation comes as no surprise, given that the association between parity and IMT was paralleled by similar distributions of lipid profiles and prevalent diabetes as well as hypertension. And although it contradicts the findings by Lawlor et al,11 it seems justified to speculate that the association between parity and IMT, as observed in our study, is to some extent mediated through biological factors (at the very least, more so than through socioeconomic determinants or lifestyle factors. However, the number of conceivable socioeconomic and lifestyle factors of influence is huge. Lawlor et al11 had elegantly solved this problem by using an approach in which the association between parity and CHD was assessed separately among women and men. It remains, of course, uncertain how far diverse socioeconomic and lifestyle conditions have been adequately and comprehensively reflected by the items that have been selected as potential confounders in our study. This represents a significant limitation of our study. A further limitation is the cross-sectional design of this study, which did not allow us to assess the time course of the relation between parity and carotid IMT. In addition, because of a possibly higher mortality among multiparous women, this study design might contribute to survival bias. However, survival bias could at most have lessened the strength of the association between parity and carotid IMT. Moreover, the study sample was largely made up by women from the former communist East Germany and consequently shows specific sociodemographic features. These may limit the capability of our findings to be extrapolated to other countries. Finally, miscarriages, stillbirths, and abortions may also have an effect, yet they were not considered in the present analyses. However, this seems justified given that a normal pregnancy takes 9 months; although the duration of these conditions may vary tremendously from only a few weeks to nearly 9 months, and this renders their impact on cardiovascular disease difficult to assess.
In summary, this study identified a relation between parity and carotid IMT in which nulliparity and higher numbers of children were associated with increased carotid IMT values. Because carotid IMT constitutes an established marker of generalized atherosclerosis, this relation might have implications for the development of cardiovascular disease among women. The association seems to be partly mediated through conventional cardiovascular risk factors and socioeconomic determinants. However, an independent effect of parity might also be involved. The nature of such an effect remains to be determined and requires further research.
References
Ordovas J, Pocovi M, Grande F. Plasma lipids and cholesterol esterification rate during pregnancy. Obstet Gynecol. 1984; 63: 20–25.
de Alvarez RR, Gaiser DF, Simkins DM, Smith EK, Bratvold GE. Serial studies of serum lipids in normal human pregnancy. Am J Obstet Gynecol. 1959; 77: 743–757.
Fahraeus L, Larsson-Cohn U, Wallentin L. Plasma lipoproteins including high-density lipoprotein subfractions during normal pregnancy. Obstet Gynecol. 1985; 66: 468–472.
Lewis D, Funkhouser D, Racqynski J, Sidney S, Bild D, Howard B. Adverse effect of pregnancy on high-density lipoprotein (HDL) cholesterol in young adult women. Am J Epidemiol. 1996; 144: 247–254.
Kritz-Silverstein D, Barrett-Connor E, Wingard DL. The effect of parity on the later development of non-insulin-dependent diabetes mellitus or impaired glucose tolerance. N Engl J Med. 1989; 321: 1214–1219.
Toescu V, Nuttall SL, Martin U, Kendall MJ, Dunne F. Oxidative stress and normal pregnancy. Clin Endocrinol. 2002; 57: 609–613.
Sadaniantz A, St. Laurent L, Parisi A. Long-term effects of multiple pregnancies on cardiac dimensions and systolic and diastolic function. Am J Obstet Gynecol. 1996; 174: 1061–1064.
Clapp J, Capeless E. Cardiovascular function before, during, and after the first and subsequent pregnancies. Am J Cardiol. 1997; 80: 1469–1473.
Ness R, Harris T, Cobb J, Flegal K, Kelsey J, Balanger A, Stunkard A, D’Agostino R. Number of pregnancies and risk of cardiovascular disease. N Engl J Med. 1993; 328: 1528–1533.
Green A, Beral V, Moser K. Mortality in women in relation to their childbearing history. BMJ. 1988; 297: 391–395.
Lawlor DA, Emberson JR, Ebrahim S, Whincup PH, Wannamethee SG, Walker M, Smith GD. Is the association between parity and coronary heart disease due to biological effects of pregnancy or adverse lifestyle risk factors associated with child-rearing Circulation. 2003; 107: 1260–1264.
Qureshi A, Giles W, Croft J, Stern B. Number of pregnancies and risk for stroke and stroke subtypes. Arch Neurol. 1997; 54: 203–206.
Steenland K, Lally C, Thun M. Parity and coronary heart disease among women in the American Cancer Society CPSII population. Epidemiology. 1996; 7: 641–643.
Colditz GA, Willett WC, Stampfer MJ, Rosner B, Speizer FE, Hennekens CH. 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.
O’Leary DH, Polak JF, Kronmal RA, Manolio TA, Burke GL, Wolfson SK Jr. Carotid-artery intima and media thickness as a risk factor for myocardial infarction and stroke in older adults. Cardiovascular Health Study Collaborative Research Group. N Engl J Med. 1999; 340: 14–22.
Touboul PJ, Elbaz A, Koller C, Lucas C, Adrai V, Chedru F, Amarenco P. Common carotid artery intima-media thickness and brain infarction: the Etude du Profil Genetique de l’Infarctus Cerebral (GENIC) case-control study. The GENIC Investigators. Circulation. 2000; 102: 313–318.
Humphries KH, Westendorp IC, Bots ML, Spinelli JJ, Carere RG, Hofman A, Witteman JC. Parity and carotid artery atherosclerosis in elderly women: the Rotterdam Study. Stroke. 2001; 32: 2259–2264.
John U, Greiner B, Hensel E, Ludemann J, Piek M, Sauer S, Adam C, Born G, Alte D, Greiser E, Haertel U, Hense HW, Haerting J, Willich S, Kessler C. Study of Health In Pomerania (SHIP): a health examination survey in an east German region: objectives and design. Soz Praventivmed. 2001; 46: 186–194.
Keil U, Stieber J, Doring A, Chambless L, Hartel U, Filipiak B, Hense HW, Tietze M, Gostomzyk JG. The cardiovascular risk factor profile in the study area Augsburg. Results from the first MONICA survey 1984/85. Acta Med Scand Suppl. 1988; 728: 119–128.
Wolff B, Volzke H, Ludemann J, Robinson D, Vogelgesang D, Staudt A, Kessler C, Dahm JB, John U, Felix SB. Association between high serum ferritin levels and carotid atherosclerosis in the Study of Health in Pomerania (SHIP). Stroke. 2004; 35: 453–457.
Knochenhauer ES, Key TJ, Kahsar-Miller M, Waggoner W, Boots LR, Azziz R. Prevalence of the polycystic ovary syndrome in unselected black and white women of the southeastern United States: a prospective study. J Clin Endocrinol Metab. 1998; 83: 3078–3082.
Costello MF, Eden JA. A systematic review of the reproductive system effects of metformin in patients with polycystic ovary syndrome. Fertil Steril. 2003; 79: 1–13.
Peppard HR, Marfori J, Iuorno MJ, Nestler JE. Prevalence of polycystic ovary syndrome among premenopausal women with type 2 diabetes. Diabetes Care. 2001; 24: 1050–1052.
Lobo RA, Carmina E. The importance of diagnosing the polycystic ovary syndrome. Ann Intern Med. 2000; 132: 989–993.
Martikainen H, Ruokonen A, Hartikainen AL, Jarvelin MR. Polycystic ovary syndrome and hyperandrogenism as a risk factor for cardiovascular disease. Int J Circumpolar Health. 1998; 57: 133–137.
He S, Silveira A, Hamsten A, Blomback M, Bremme K. Haemostatic, endothelial and lipoprotein parameters and blood pressure levels in women with a history of pre-eclampsia. Thromb Haemost. 1999; 81: 538–542.
Peters RK, Kjos SL, Xiang A, Buchanan TA. Long-term diabetogenic effect of single pregnancy in women with previous gestational diabetes mellitus. Lancet. 1996; 347: 227–230.