Parity and Risk of Cardiovascular Disease in Women over 45 Years in the United States: National Health and Nutrition Examination Survey 2007

Abstract

Objective:

Current research results on the association between parity and cardiovascular disease (CVD) risk are inconclusive. The purpose of this study was to examine the relationship between parity and risk of CVD in women over 45 years of age.

Materials and Methods:

Data were from the National Health and Nutrition Examination Survey for 2007–2018. A total of 8,882 women ≥45 years of age were included. The exposure of parity referred to the number of live births, and the outcome variable was the occurrence of CVD events, including stroke, heart attack, coronary heart disease (CHD), heart failure, and angina. We performed logistic regression to calculate the unadjusted and adjusted odds ratios (ORs; 95% confidence intervals [CIs]) controlling for confounding factors.

Results:

Among 8,882 women, the mean age was 62.4 ± 10.8 years, with a range of 45 to 80 years. The weighted prevalence of CVD, stroke, heart attack, CHD, heart failure, and angina in parous women were significantly higher than those of nulliparous women (p < 0.05). After adjusting for demographic factors, CVD risk factors, reproductive factors, the ORs of CVD with parity 1–2, 3–4, and 5+ were 1.85 (95% CI: 1.29–2.64), 1.70 (95% CI: 1.15–2.50), and 1.92 (95% CI: 1.28–2.88), respectively. The odds of stroke and heart attack were also significantly positively related to increasing parity compared with nulliparity. However, compared with nulliparity, parity of 3 was associated with a slightly lower risk of CVD.

Conclusions:

Our findings indicated that parity was significantly positively associated with CVD, giving birth to three children associated with a slightly lower risk compared with nulliparity. Further cohort studies are warranted to confirm the findings.

Introduction

Most countries around the world are facing the harsh reality of significantly dropping fertility rates and declining willingness to bear children.^1
–3 Among women 15–44 years of age in the United States, the birth rate has dropped from 70‰ in 2007 to 55.8‰ in 2020, a drop of about 20% in 13 years.⁴ In Thailand, the average number of births per woman during her reproductive lifespan has decreased significantly every year, and is expected to continue to decline to 1.3 by 2040.⁵ The birth rate in China has dropped astonishingly from 37‰ in 1950 to 8.52‰ in 2020 with a drop of 77% among women 15–49 years of age, and the average number of births per woman during her reproductive lifespan declined from 6.0 in the 1960s to 1.3 in 2020.⁶ Faced with such changes in the global fertility rate, the relationship between giving birth and women's future health is an issue worthy of attention. Whether giving birth will influence the development of chronic conditions in later life, and the number of births most associated with healthy outcomes are controversial.

Cardiovascular disease (CVD) is the leading cause of morbidity and mortality worldwide.⁷ For women, in addition to traditional CVD risk factors, some reproductive factors may also affect the development of maternal CVD in later life.⁸ Pregnancy is related to cardiometabolic changes such as weight gain and dyslipidemia.^9,10 Physiological changes related to pregnancy may make women susceptible to CVD through inflammation and endothelial function.^11,12 So, a broader agenda has been warranted to integrate reproductive health into women's CVD research.

The association of nulliparity and parity with CVD has been controversial.^13
–15 Several previous studies found that nulliparous women had the lowest risk of CVD,^16
–18 while others indicated that nulliparity increased women's CVD risk compared with women with parity of 1 or 2.^19
–21 Some population-based studies did not find a significant association between parity and CVD in women.^13,14,22 Therefore, the purpose of the study was to use National Health and Nutrition Examination Survey (NHANES) data to examine the association between parity and CVD in women over 45 years of age in a large and representative population in the United States. Our hypothesis was that parous women 45 years of age or older would be likely to have an increased risk of CVD.

Materials and Methods

Data source and study population

NHANES is a nationally representative cross-sectional set of health surveys conducted by the National Center for Health Statistics (NCHS) of the Centers for Disease Control and Prevention (CDC).²³ The CDC's NCHS Ethics Review Board approved NHANES data collection from the participants. It utilized a stratified, multistage probability sampling design for selection of noninstitutionalized American civilians. The survey includes assessments of diet, general health, physical examinations, and lifestyle factors. The datasets are from public-use files. We analyzed data from six 2-year NHANES datasets for the period 2007–2018. The research population was women 45 years of age or older at the time of interview. There were 10,379 women over 45 years of age in the dataset. We excluded 121 women with the response of “don't know” or “refused” or missing values for the conditions of stroke, heart attack, heart failure, coronary heart disease (CHD), and angina.

The datasets have the variable of age at these CVD events and age at the last live birth. We also excluded those participants who had these conditions of CVDs before their last live births and nulliparous women who had these CVD conditions before the age of 29 years (the average age of last live birth, n = 69). Then, we imputed missing values of parity with the sum of the number of vaginal and cesarean deliveries, and deleted cases of unknown parity, excluding the participants whose history of pregnancy was the response, “don't know” or “refused” or missing (n = 1307). Thus, a total of 8,882 participants were included in the present analyses.

Dependent and independent variables

The primary outcomes were stroke, heart attack, CHD, heart failure, and angina pectoris based on the self-reported responses to the questions “Has a health care provider or physician ever told you that you had a stroke/heart attack/CHD/heart failure/angina?”. If any of the above diseases were reported, we considered the participant to have CVD. The independent variable of parity referred to the number of live births. Parity was determined by responses to several questions, including “Have you ever been pregnant?” “Are you pregnant now?” and “How many live births have you ever had?” We categorized parity as follows: if participants were never pregnant or the number of live births was zero, we set the value of parity as “nulliparity”; if the participant had at least one live birth, they were parous women. For parous women, the number of live births was categorized as 1–2, 3–4, and 5+ of parity, respectively.²⁴ We also consider parity as continuous independent variable including 0, 1, 2, 3, 4, and 5+ to explore the association between parity and CVD.

Potential confounding variables

The potential covariates included age at the interview, race, marital status, education level, family monthly poverty level index (FMPLI), body mass index (BMI), smoking, alcohol drinking, physical activity, healthy eating, family history of heart attack, hypertension, diabetes mellitus, age at menarche, age at menopause, hysterectomy, bilateral oophorectomy, ever taking birth control pills, and taking female hormones.^8,10
–12 Age at the interview, BMI, physical activity, age at menarche, and age at menopause were continuous variables. Height and weight were measured in the mobile examination centers by trained health technicians. Body weight was measured using a digital weight scale, and standing height was measured using a stadiometer with a fixed vertical backplate and adjustable headpiece. BMI was weight (kilogram) divided by height (meters) squared, rounded to one decimal place. We quantified physical activity by metabolic equivalent scores (METs).

The METs was the sum of the minutes of vigorous work-related/leisure-time activity and moderate vigorous work-related/leisure-time activity each week. The weight coefficients of vigorous activity and moderate activity were 8 and 4, respectively. Then, we calculated the average METs/hour per week as the value of physical activity variable.²⁵ The classification of race/ethnicity was non-Hispanic White, non-Hispanic Black, Hispanic and others, including Asian and Multiracial. Education level included high school or below and college/university or above. We divided marital status into married/living with partner, widowed/divorced/separated, and never married. Smoking was based on the two questions: “Have you smoked at least 100 cigarettes in your life?” and “Do you currently smoke?”.

We grouped smoking status into three categories: now smokers referred to participants smoking at least 100 cigarettes in life, and still smoking either every day or some days; former smokers referred to those smoking at least 100 cigarettes in life but did not smoke currently; never smoking referred to smoking less than 100 cigarettes in life. Alcohol drinking was based on the responses to the question “Have you had at least 12 alcohol drinks per year?” We defined the variable of healthy diet based on the question “How healthy is the diet,” including five different responses “excellent,” “very good,” “good,” “fair,” and “poor.” Family history of heart attack was determined from the question “Have your close relative, including father, mother, sisters, or brothers, ever had heart attack?”

The systolic and diastolic blood pressures (BP) of participants were measured by physicians up to four times after sitting and resting for 5 minutes. We calculated the average of systolic and diastolic BP of the second, third, and fourth times, discarding the BP value of the first time. We defined hypertension based on the questions, “Has a health care provider or physician ever told you that you had high blood pressure?” and “Are you taking prescription for hypertension,” and the average systolic BP was ≥140 mmHg or the average diastolic BP was ≥90mmHg. We defined diabetes mellitus by the questions, “Has a health care provider or physician ever told you that you had diabetes?” and “Are you taking antidiabetic medication/insulin?” and the fasting plasma glucose ≥126 mg/dL.²⁶ A history of high cholesterol, hysterectomy, bilateral oophorectomy, taking female hormones, taking birth control pills, all came from self-reports by the survey participants.

Statistical analyses

We followed NHANES analysis guidelines for merging data from 2007 to 2018 and applying sample weights.²⁷ We performed univariate analysis for descriptive statistics, producing weighted proportions (%). We used the Rao–Scott Chi-square test to compare weighted proportions among different parity levels. We used t-tests to compare the mean values of continuous variables between two groups, applying analysis of variance for comparing the mean values in multiple groups. We conducted multivariable analyses using logistic regression accounting for survey weights to obtain unadjusted and adjusted odds ratio (OR) values and 95% confidence intervals (CIs) treating CVD as the dependent variable and parity as the independent variable. We tried all potential covariates, and then compared the models with goodness-of-fit tests and model selection criteria (Akaike Information Criterion [AIC] and log-likelihood).

When performing the goodness-of-fit test, we listed four models. Model 1 was unadjusted, and model 2 was the full model, including all the covariates of age, race, education, marital status, FMPLI, smoking, alcohol drinking, BMI, physical activity, healthy diet, high blood pressure, cholesterol, diabetes mellitus, age at menarche, age at menopause, hysterectomy, oophorectomy, taking birth control pills, and taking female hormones. Model 3 consisted of significant variables in model 2 and added interaction term parity*race. Model 4 was the final model, omitting nonsignificant covariates in model 3. To obtain the optimal model, we took advantage of AIC, log-likelihood, and R-squared (R ²) criteria. Model 4 indicated a better fit model with the largest log-likelihood and R ² and the smallest AIC. All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). A two-tailed p-value of <0.05 was considered having statistical significance.

Results

Basic characteristics of the study population

The study population consisted of 8,882 women over the age of 45 years, including 1,035 nulliparous women, 3,667 women with parity of 1–2, 3,017 women with a parity of 3–4, and 1,163 women with a parity of 5 or above. The mean age was 62.4 ± 10.8 years, with a range from 45 to 80 years. The weighted prevalence of nulliparity in women over 45 years was 13.0%, which could be representative of ∼7,501,386 females at the same age in the United States.

Distributions of race, marital status, education, and FMPLI differed significantly by parity level with nulliparous women and women in parity level 1–2 being more likely to have never been married, have a college education or higher and have a FMPLI >1.85 than women in other parity levels. Proportions of smoking, alcohol drinking, hypertension, diabetes mellitus, ever taking birth controls, taking female hormones, and having had hysterectomy differed significantly among the different parity groups. The mean of BMI, physical activity, age at menarche, and age at menopause were also statistically different in the groups. Proportions with high cholesterol or having had a bilateral oophorectomy or family history of heart attacks did not differ significantly among parity groups (Table 1).

Table 1.

Characteristics of Women 45 Years of Age and Over by Parity in National Health and Nutrition Examination Survey 2007–2018

Characteristics	Nulliparity weighted %	Parity 1–2 weighted %	Parity 3–4 weighted %	Parity 5+ weighted %
Total	1,035	3,667	3,017	1,163
Age (years)^a,b	59.4 ± 0.4	59.3 ± 0.2	62.2 ± 0.3	67.7 ± 0.5
Race^b
NH-White	76.2	76.7	72.4	54.2
NH-Black	10.3	9.4	10.6	17.3
Hispanic	6.6	7.4	11.6	22.1
Others	6.9	6.6	5.3	6.3
Marital status^b
Married/living with partner	51.9	64.2	61.4	50.0
Widowed/divorced/separated	23.5	31.9	36.2	47.0
Never married	24.7	3.9	2.4	3.0
Education^b
High school or less	28.5	35.1	44.9	68.0
College or above	71.5	64.9	55.1	32.0
FMPLI^b
FMPLI ≤1.3	18.5	16.6	23.2	43.2
FMPLI 1.3–1.85	9.8	10.9	14.6	19.8
FMPLI >1.85	71.7	72.5	62.2	37.0
Alcohol drinking^b	57.4	56.4	48.2	38.5
Smoking status^c
Now smoker	12.3	15.8	13.9	14.1
Former smoker	29.2	26.7	25.0	21.2
Never smoking	58.5	57.5	61.1	64.7
BMI (kg/m²)^a,c	29.9 ± 0.4	29.4 ± 0.2	29.8 ± 0.2	30.3 ± 0.2
Physical activity^a	11.0 ± 0.9	13.5 ± 0.6	11.9 ± 0.6	8.8 ± 0.6
Hypertension^b	51.4	52.0	58.5	66.2
High cholesterol	48.0	45.3	47.8	48.5
Diabetes mellitus^b	12.9	13.5	17.8	26.8
Birth control pills^b	65.8	79.4	69.1	54.7
Female hormones^c	31.7	35.4	32.5	25.3
Oophorectomy	20.0	17.8	20.9	19.1
Hysterectomy^b	27.7	33.0	37.6	39.4
Age at menopause^a,c	46.1 ± 0.4	45.8 ± 0.2	45.5 ± 0.3	46.2 ± 0.3
Age at menarche^a,b	12.8 ± 0.1	12.7 ± 0.1	12.8 ± 0.1	13.0 ± 0.1
Family history of heart attack	14.4	17.3	16.6	14.9
Heart failure^b	2.0	3.0	3.5	7.6
CHD^b	1.8	3.2	4.9	5.9
Heart attack^b	1.6	3.1	4.2	7.4
Angina^b	1.5	2.3	3.4	5.1
Stroke^b	2.3	3.9	5.4	9.0
CVD^b	6.5	10.2	13.0	20.2

The p value came from the analysis of variance or Rao–Scott Chi-square test among different subgroups. Physical activity indicated MET score/hour per week.

Means ± SDs.

p < 0.0001.

p < 0.05.

Alcohol drink, had at least 12 alcohol drinks per year; CHD, coronary heart disease; CVD, cardiovascular disease; FMPLI, family monthly poverty level index; NH, non-Hispanic; SD, standard deviation.

Comparison of the CVD prevalence between nulliparous and parous women

The unweighted and weighted prevalence of CVD in parous women were all significantly higher than those of nulliparous women. Specifically, the weighted prevalence of CVD, stroke, heart attack, CHD, heart failure, and angina were 12.1%, 4.9%, 3.9%, 4.0%, 3.6%, and 3.0%, respectively, in women with parity 1+, while the corresponding prevalence were 6.5%, 2.3%, 1. 6%, 1.8%, 2.0%, and 1.5% in nulliparous women (Table 2).

Table 2.

The Cardiovascular Disease Prevalence by Parity for Women ≥45 Years of Age in the United States, National Health and Nutrition Examination Survey 2007–2018

	Nulliparous women, n (%)		Parous women, n (%)		p
	Unweighted, n = 1,035	Weighted, n = 7,501,386	Unweighted, n = 7,847	Weighted, n = 50,134,514	p
CVD	92 (8.9)	484,992 (6.5)	1,114 (14.2)	6,055,050 (12.1)	<0.0001
Stroke	38 (3.7)	172,733 (2.3)	469 (6.0)	2,466,898 (4.9)	<0.0001
Heart attack	23 (2.2)	116,990 (1.6)	334 (4.3)	1,930,638 (3.9)	<0.0001
CHD	22 (2.1)	131,442 (1.8)	322 (4.1)	2,021,406 (4.0)	0.0030
Heart failure	27 (2.6)	150,336 (2.0)	350 (4.5)	1,811,628 (3.6)	0.0360
Angina	23 (2.2)	113,998 (1.5)	257 (3.3)	1,483,005 (3.0)	0.0123

The Rao–Scott Chi-square p value was performed for comparing the weighted prevalence between parity groups.

Association between parity and CVD

The final model 4 included the covariates of age, race, FMPLI, smoking, BMI, physical activity, family history of heart attack, healthy diet, high BP, high cholesterol, diabetes mellitus, and age at menopause, omitting the interaction of parity and race that was not significant in CVD. The adjusted association of CVD with parity 1–2, 3–4, and 5+ compared with nulliparity were 1.85 (95% CI: 1.29–2.64), 1.70 (95% CI: 1.15–2.50), and 1.92 (95% CI: 1.28–2.88), respectively. Model 4 indicated that stroke and heart attack were also significantly related to parity compared with nulliparity after controlling for all the confounders. However, no significant relation was observed between parity and CHD, heart failure, and angina after adjusting for covariates (Table 3).

Table 3.

The Unadjusted and Adjusted Odds Ratios of Cardiovascular Diseases by Parity, National Health and Nutrition Examination Survey 2007–2018

	Nulliparity	Parity 1–2 OR (95% CI)	Parity 3–4 OR (95% CI)	Parity 5+ OR (95% CI)
CVD (n = 1,206)
Model 1	1.00	1.64 (1.24–2.16)	2.16 (1.60–2.90)	3.67 (2.68–5.02)
Model 2	1.00	1.89 (1.30–2.75)	1.72 (1.15–2.58)	1.95 (1.25–3.05)
Model 3	1.00	2.16 (1.32–3.53)	1.88 (1.09–3.22)	2.32 (1.34–4.02)
Model 4	1.00	1.85 (1.29–2.64)	1.70 (1.15–2.50)	1.92 (1.28–2.88)
Stroke (n = 507)
Model 1	1.00	1.74 (1.15–2.62)	2.43 (1.56–3.79)	4.17 (2.65–6.57)
Model 2	1.00	1.66 (0.99–2.80)	1.90 (1.06–3.41)	2.02 (1.08–3.78)
Model 3	1.00	2.85 (1.44–5.66)	3.33 (1.40–7.92)	4.39 (1.94–9.92)
Model 4	1.00	1.73 (1.07–2.80)	1.99 (1.17–3.39)	2.13 (1.20–3.75)
Heart attack (n = 357)
Model 1	1.00	2.00 (1.22–3.27)	2.74 (1.59–4.73)	5.02 (2.81–8.94)
Model 2	1.00	2.90 (1.42–5.93)	2.54 (1.15–5.62)	3.43 (1.45–8.11)
Model 3	1.00	3.98 (1.50–10.55)	4.03 (1.44–11.29)	5.90 (1.95–17.90)
Model 4	1.00	2.90 (1.52–5.54)	2.63 (1.31–5.29)	3.53 (1.58–7.90)
CHD (n = 344)
Model 1	1.00	1.85 (0.98–3.50)	2.86 (1.61–5.09)	3.49 (1.75–6.96)
Model 2	1.00	1.86 (0.85–4.09)	1.89 (0.92–3.87)	1.86 (0.80–4.30)
Model 3	1.00	1.61 (0.74–3.53)	1.56 (0.72–3.35)	1.51 (0.62–3.67)
Model 4	1.00	1.70 (0.83–3.45)	1.73 (0.88–3.41)	1.70 (0.79–3.68)
Heart failure (n = 377)
Model 1	1.00	1.51 (0.80–2.84)	1.80 (0.97–3.33)	4.03 (2.21–7.36)
Model 2	1.00	2.04 (0.86–4.84)	1.60 (0.65–3.94)	2.07 (0.82–5.23)
Model 3	1.00	1.92 (0.75–4.90)	1.43 (0.53–3.85)	1.86 (0.68–5.11)
Model 4	1.00	1.84 (0.86–3.93)	1.50 (0.67–3.36)	1.99 (0.89–4.45)
Angina (n = 280)
Model 1	1.00	1.53 (0.85–2.76)	2.28 (1.29–4.03)	3.52 (1.82–6.80)
Model 2	1.00	1.19 (0.60–2.37)	1.39 (0.73–2.66)	1.67 (0.74–3.80)
Model 3	1.00	1.54 (0.66–3.58)	1.91 (0.88–4.16)	2.03 (0.65–6.36)
Model 4	1.00	1.32 (0.67–2.61)	1.57 (0.83–2.97)	1.87 (0.81–4.32)

Model 1 was unadjusted. Model 2 adjusted for age, race, education, marital status, FMPLI, smoke, alcohol drinking, BMI, physical activity, healthy diet, high blood pressure, cholesterol, diabetes mellitus, age at menarche, age at menopause, hysterectomy, oophorectomy, taking birth control pills and female hormones. Model 3 consisted of significant variables in model 2 and added the interaction term of parity^*race. Model 4 was the final model, omitting nonsignificant variables in model 3.

BMI, body mass index; CI, confidence interval; OR, odds ratio.

Figure 1 presents the unadjusted and adjusted associations of CVD from model 4 with individual parity levels instead of grouping parity 1–2, 3–4, and 5+ presented in Table 3. Compared with nulliparity, the unadjusted ORs for CVD were similar for parity of 1–3, but were higher for parity of 4 and 5+. After adjustment for all the covariates in model 4, adjusted ORs were significant for all levels of parity compared with nulliparous women: 1.86 (95% CI: 1.21–2.86), 1.85 (95% CI: 1.27–2.69), 1.56 (95% CI: 1.05–2.32), 1.97 (95% CI: 1.28–3.03), and 1.94 (95% CI: 1.29–2.90).

FIG. 1.

The unadjusted and adjusted odds ratios between parity and CVD in women ≥45 years of age in the United States, NHANES 2007–2018. The adjusted confounders were model 4 covariates. CVD, cardiovascular disease; NHANES, National Health and Nutrition Examination Survey.

Discussion

We found that among women over 45 years of age in the United States, the weighted prevalence of CVD, stroke, heart attack, CHD, heart failure, and angina were lower in nulliparous women than those of parous women (Table 2). Women with parity 1–2, 3–4, and 5+ were more likely to suffer from stroke, heart attack, and CVD than nulliparous women in later life (Table 3). Compared with nulliparous women, the adjusted OR between CVD and parity 3 was slightly lower than other parous women after controlling for the potential confounders (Fig. 1).

Research on the association of CVD with nulliparity has yielded conflicting findings.¹¹ The results of Healthy Aging and Body Composition Study in Pittsburgh were consistent with our findings, reporting nulliparous older women with the average of 74 years had a lower CVD prevalence than parous women (18.0% vs. 30.2%).¹⁶ In contrast, Parikh et al. reported that nulliparity conferred a moderate risk of CVD for women.²¹ Hannaford et al. demonstrated that nulliparous women (mean age 56 years) were at higher risk of hypertension or stroke than parous women.²⁰

Our findings agree with those of multiple previous studies. Catov et al. reported that compared with nulliparity, the CVD prevalence of women with parity 1+ was 1.95-fold higher (95% CI: 1.03–3.7), parity 5+ having 2.27-fold (95% CI: 1.00 5.15) after adjustment for confounders.¹⁶ One study integrating the data of the Framingham Heart Study and NHANES National Epidemic Follow-up Study showed women with six or more pregnancies had a slightly increased risk of CVD after controlling for age, CVD risk factors, and education level.²⁸

However, some research found no association of parity with CVD risk factors. A longitudinal study from the United Kingdom showed that there was no statistical association between parity and CVD risk factors (hypertension, type 2 diabetes, and BMI) after controlling for behavioral and lifestyle variables.¹⁴ Chung et al. also reported that the association of parity with CVD risk factors (systolic and diastolic pressure, fasting plasma glucose, and insulin) had no significant association after controlling for covariates.¹³ This inconsistency may be due to the differences in the average age of study populations in the studies.

Compared with nulliparous women, the CVD risk of parity 3 (OR:1.56; 95% CI: 1.05–2.32) was slightly lower than that for other parous women (Fig. 1). This was different from other research findings. The cohort studies from Sweden and Iran indicated a J-shaped relationship between parity and CVD, with parity of 2 having the lowest risk of CVD in parous women.^10,21 Another study from Britain also reported a similar J-shaped association between parity and CHD, women with parity 2 having the lowest prevalence of CHD. For those women with more than two children, each additional parity was associated with 30% greater age-adjusted odds of developing CHD (OR:1.30; 95% CI: 1.17–1.44).¹⁵ So, further studies about U.S. women are needed to confirm the findings.

The specific risk factors for women to suffer from CVD might include pregnancy and reproductive-related endocrine disorders .²⁹ Female pregnancy can cause changes in the hormone environment, glucose and fat metabolism, and mild inflammation.^30,31 These changes may promote the development of CVD in later life.^32,33 In the normal third trimester of pregnancy, diastolic and systolic functions of the heart may be slightly impaired. Additionally, the insulin level of pregnant women could increase due to assisting in fetal development.^34,35 In the case of increased number of pregnancies and parity, all these changes may elevate the risk of CVD in women in the future.

In addition to biological mechanisms related to giving birth, the interpretation for the association between parity and CVD also involved factors related to raising children.³⁶ Some researchers have explored the differences in the association of parity with CVD between mothers and fathers. For example, Ness et al. reported the number of children was not associated with father's CVD in later life, although there was significant relationship between parity and maternal CVD.³⁷ This implied that the association between parity and maternal CVD may be from the biological mechanisms involved in pregnancy and giving birth. We were unable to examine the association between parity and paternal CVD because the National Health and Nutrition Examination Survey data do not include data on the number of children for fathers.

If we could examine the relationship between the number of children and paternal CVD, we may be able to better explain whether the association between parity and CVD was derived from parity-related physiological mechanisms or residual effects of socioeconomic factors. However, no matter what the cause of the relationship between parity and CVD, parous women may be high-risk populations of CVD compared with nulliparous women.

The strengths of this study included its representativeness of U.S. women and large sample size. The outcome variable was based on the study design where we only included participants in which stroke, heart attack, CHD, heart failure, and angina occurred after the last live birth, which could support the temporal relationship between parity and CVD. This could offset the shortcomings of cross-sectional data regarding causal identification to a certain extent. Furthermore, we performed the weighted analysis of the data, which made the results more applicable to the U.S. women.³⁸ Regarding the shortcomings of this study, information concerning family history of CVD only involved heart attack, not including stroke, heart failure, and other CVD conditions. Recall bias may exist because the data were self-reported from the questionnaire. In addition, due to the small sample size of nulliparous women with a history of pregnancy, we did not extract them as a separate subgroup during the analysis. This restricted us from comparing them to participants without any history of pregnancy.

Our findings suggest that women who gave births of more than four children may have an increased risk of developing CVD in later life in the United States. This may imply that preventive screening measures for CVD are necessary to favor the subpopulation. Of course, more research is imperative to further corroborate our findings, as well as more research to examine the underlying theoretical mechanism for the association. In addition, we would like to emphasize that although the results show that nulliparous women have the lowest risk of CVD, we discourage women from childbearing because nulliparity may have other adverse physical and mental health effects on women in later life.

Conclusion

Our study used a representative U.S. population-based sample to explore the association between parity and CVD in women over 45 years of age. There were strong positive associations between increased parity and CVD, stroke, heart attack, CHD, heart failure, and angina. After controlling for potential confounding variables, increased parity was statistically positively associated with CVD, stroke, and heart attack. Among all participants, nulliparous women had the lowest risk of CVD. In parous women, giving birth to three children seemed to have a slightly lower risk of CVD. Further cohort studies are warranted to confirm the findings.

Footnotes

Acknowledgment

The authors would like to thank Centers for Disease Control and Prevention, United States, for NHANES data.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

References

Guo

, Gietel-Basten

, Gu

. The lowest fertility rates in the world? Evidence from the 2015 Chinese 1% sample census. China Popul Dev Stud, 2019; 2:245–258.

Rotella

, Varnum

MEW

, Sng

, et al. Increasing population densities predict decreasing fertility rates over time: A 174-nation investigation. Am Psychol, 2021; 76:933–946.

Sommer

. Women, demography, and politics: How lower fertility rates lead to democracy. Demography, 2018; 55:559–586.

Kearney

, Levine

, Pardue

. The puzzle of falling US birth rates since the great recession. J Econ Perspect, 2022; 36:151–176.

Peek

, Im-Em

, Tangthanaseth

. The state of Thailand's population 2015: Features of Thai families in the era of low fertility and longevity. Bangkok, Thailand: The Office of the National Economic and Social Development Board of Thailand, 2015.

National Health Commission of the People's Republic of China. China Health Statistics Yearbook. Beijing, China: National Bureau of Statistics, 2020.

Woodward

. Cardiovascular disease and the female disadvantage. Int J Environ Res Public Health, 2019; 16:1165.

Peters

, Woodward

. Women's reproductive factors and incident cardiovascular disease in the UK Biobank. Heart, 2018; 104:1069–1075.

Cain

, Salemi

, Tanner

, et al. Pregnancy as a window to future health: Maternal placental syndromes and short-term cardiovascular outcomes. Am J Obstet Gynecol, 2016; 215:484.e1–484.e14.

10.

Moazzeni

, Toreyhi

, Asgari

, et al. Number of parity/live birth(s) and cardiovascular disease among Iranian women and men: Results of over 15 years of follow-up. BMC Pregnancy Childbirth, 2021; 21:1–11.

11.

, Ruan

, Lu

, et al. Parity and risk of maternal cardiovascular disease: A dose-response meta-analysis of cohort studies. Eur J Prev Cardiol, 2019; 26:592–602.

12.

Ogunmoroti

, Osibogun

, Kolade

, et al. Multiparity is associated with poorer cardiovascular health among women from the Multi-Ethnic Study of Atherosclerosis. Am J Obstet Gynecol, 2019; 221:631.e1–631.e16.

13.

Chung

, Kim

, Usen

. Associations between parity, obesity, and cardiovascular risk factors among middle-aged women. J Womens Health (Larchmt), 2016; 25:818–825.

14.

Hardy

, Lawlor

, Black

, et al. Number of children and coronary heart disease risk factors in men and women from a British birth cohort. BJOG, 2007; 114:721–730.

15.

Lawlor

, Emberson

, Ebrahim

, 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.

16.

Catov

, Newman

, Sutton-Tyrrell

, et al. Parity and cardiovascular disease risk among older women: How do pregnancy complications mediate the association?. Ann Epidemiol, 2008; 18:873–879.

17.

Humphries

, Westendorp

, Bots

, et al. Parity and carotid artery atherosclerosis in elderly women: The Rotterdam Study. Stroke, 2001; 32:2259–2264.

18.

Palmer

, Rosenberg

, Shapiro

. Reproductive factors and risk of myocardial infarction. Am J Epidemiol, 1992; 136:408–416.

19.

Colditz

, Willett

, Stampfer

, 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.

20.

Hannaford

, Ferry

, Hirsch

. Cardiovascular sequelae of toxaemia of pregnancy. Heart, 1997; 77:154–158.

21.

Parikh

, Cnattingius

, Dickman

, et al. Parity and risk of later-life maternal cardiovascular disease. Am Heart J, 2010; 159:215–221.e6.

22.

De Kleijn

, van der Schouw

, van der Graaf

. Reproductive history and cardiovascular disease risk in postmenopausal women: A review of the literature. Maturitas, 1999; 33:7–36.

23.

Krittanawong

, Kumar

, Wang

, et al. Sleep duration and cardiovascular health in a representative community population (from NHANES, 2005 to 2016). Am J Cardiol, 2020; 127:149–155.

24.

Yehuda

, Wang

, Pak

, et al. Parity and risk of thyroid autoimmunity based on the NHANES (2001–2002, 2007–2008, 2009–2010, and 2011–2012). J Clin Endocrinol Metab, 2017; 102:3437–3442.

25.

Zhao

, Wu

, Feng

, et al. Physical activity levels and diabetes prevalence in US adults: Findings from NHANES 2015–2016. Diabetes Ther, 2020; 11:1303–1316.

26.

American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care, 2013; 36 Suppl 1:S67–S74.

27.

Centers for Disease Control and Prevention. NHANES Survey Methods and Analytic Guidelines. Centers for Disease Control and Prevention. Available at: wwwn.cdc.gov/nchs/nhanes/analyticguidelines.aspx Accessed December 2020 .

28.

Ness

, Harris

, Cobb

, et al. Number of pregnancies and the subsequent risk of cardiovascular disease. N Engl J Med, 1993; 328:1528–1533.

29.

Appelman

, van Rijn

, Ten Haaf

, et al. Sex differences in cardiovascular risk factors and disease prevention. Atherosclerosis, 2015; 241:211–218.

30.

Chung

, Leinwand

. Pregnancy as a cardiac stress model. Cardiovasc Res, 2014; 101:561–570.

31.

Toescu

, Nuttall

, Martin

, et al. Oxidative stress and normal pregnancy. Clin Endocrinol (Oxf), 2002; 57:609–613.

32.

Rebholz

, Jones

, Burke

, D'Alessio

, et al. Multiparity leads to obesity and inflammation in mothers and obesity in male offspring. Am J Physiol Endocrinol Metab, 2012; 302:E449–E457.

33.

Rich-Edwards

, Fraser

, Lawlor

, et al. Pregnancy characteristics and women's future cardiovascular health: An underused opportunity to improve women's health?. Epidemiol Rev, 2014; 36:57–70.

34.

Barbour

, McCurdy

, Hernandez

, et al. Cellular mechanisms for insulin resistance in normal pregnancy and gestational diabetes. Diabetes Care, 2007; 30 Suppl 2:S112–S119.

35.

Schannwell

, Zimmermann

, Schneppenheim

, et al. Left ventricular hypertrophy and diastolic dysfunction in healthy pregnant women. Cardiology, 2002; 97:73–78.

36.

Chen

, Jiang

, Lu

, et al. More children, more happiness? Fewer and better births, a happy life: The relationship of parity and maternal cardiovascular disease risk. Eur J Prev Cardiol, 2019; 26:589–591.

37.

Ness

, Cobb

, Harris

, et al. Does number of children increase the rate of coronary heart disease in men?. Epidemiology, 1995; 6:442–445.

38.

Hamdan

, Scott

, Kaye

, et al. Association between parity and dental caries experience among US women: Findings from the National Health and Nutrition Examination Survey. J Contemp Dent Pract, 2021; 22:933–938.

Parity and Risk of Cardiovascular Disease in Women over 45 Years in the United States: National Health and Nutrition Examination Survey 2007–2018

Abstract

Objective:

Materials and Methods:

Results:

Conclusions:

Introduction

Materials and Methods

Data source and study population

Dependent and independent variables

Potential confounding variables

Statistical analyses

Results

Basic characteristics of the study population

Comparison of the CVD prevalence between nulliparous and parous women

Association between parity and CVD

Discussion

Conclusion

Footnotes

Acknowledgment

Author Disclosure Statement

Funding Information

References