The current study enrolled women aged ≥19 years who participated in the fifth KNHNES for hypertension measurement in the first, second, and third year (2010-2012). Figure 1 presents the selection criteria of the study population. We excluded women who were diagnosed with hypertension before the first delivery or underwent bilateral oophorectomy or hysterectomy; pregnant women; and women with malignancy, stroke, myocardial infarction, angina pectoris, or renal failure. Finally, we enrolled, in total, 8,890 women in the current study.

Parity is defined as the number of times a female has given birth counting multiple births as one.¹³⁾ In the current study, we calculated the sum of spontaneous deliveries and the frequency of cesarean sections on the basis of a questionnaire survey.

Hypertension was defined as hypertension diagnosed by a physician, the current use of antihypertensive drugs, or a final systolic blood pressure (SBP) of >140 mm Hg or a final diastolic blood pressure (DBP) of >90 mm Hg.

According to the KNHNES, blood pressure was measured using a standard sphygmomanometer at rest for more than 10 minutes; it was measured at the upper arm 3 times at a 5-minute interval. Blood pressure was also measured in the second and third blood pressure monitoring sessions, and the average of the measurements was taken as the final SBP and DBP.

Natural menopause was defined as the lack of menstruation for consecutive 12 months without physiological or pathological causes.¹⁾ During menopause , hormon levels alter dramatically. so menopause affect blood pressure greatly.^4,5) Accordingly, in the current study, women who responded that they experienced natural menopause to a questionnaire survey were classified into the group of women who achieved natural menopause. On the basis of these criteria, our subjects were classified into the premenopausal group and the postmenopausal group.

The following items were evaluated through an interview: parity, a family history of hypertension, a past history of treatment with antihypertensive drugs, a past history of treatment with oral contraceptives, a past history of estrogen therapy, and a past history of menopause. In addition, a past history of smoking and drinking and the pattern of exercise were evaluated through a self-administered questionnaire study. Regarding the nutritional status, the daily sodium intake was evaluated through measurement of the types and the amount of ingested food 1 day before the current study through a 24-hour recall. Thus, we calculated the sum of the daily sodium intake from the diet of the subjects over a 1-day period. Although we determined the sodium intake 1 day before the current study, we presumed that the intake might not differ from their daily intake. Therefore, sodium intake was considered a variable in the current study. Regarding the smoking status, the subjects were classified into current smokers, past smokers, and non-smokers. Subjects were regarded as drinkers if they consumed alcohol more than once a month, and they were regarded as non-drinkers if they consumed alcohol less than once a month. The pattern of exercise was evaluated using the International Physical Activity Questionnaire. The group of subjects who performed regular exercise comprised those who performed a single session of vigorous physical activity for >20 minutes for >3 times a week or those who performed a single session of a moderate-level physical activity for >30 minutes for >5 days a week. The total amount of household income was adjusted using the number of households, and the subjects were classified into the following 4 groups: inferior, middle inferior, middle superior, and superior. In terms of the years of education, the subjects were classified into the following groups: <9 years, 9-12 years, 12-16 years, and >16 years. Subjects were considered to be on contraceptive and estrogen medication if they were administered these therapies more than once a month. Physical measurements were recorded using a mobile health check-up system. The height and weight were measured in units of 0.1 cm and 0.1 kg, respectively. The body mass index (BMI) was calculated using the formula: (weight in kg)/(height in m²).

The postmenopausal period is characterized by remarkable alterations in the blood pressure due to alterations in the endocrine system.⁸⁾ All the analyses were performed by classification of the subjects into the premenopausal group and the postmenopausal group. To analyze the baseline and general characteristics of the subjects, we expressed continuous variables as mean±standard error (SE) by using the independent sample t-test. Categorical variables were expressed as frequency (percentage, %) using the χ² test. In addition, regarding the correlation between parity and hypertension, the premenopausal group was compared with the postmenopausal group by using a multiple regression analysis. To avoid the effect of the use of antihypertensive drugs on the results of the multiple regression analysis, we performed the analyses by using tables that included data of subjects with a past history of antihypertensive medication and tables that excluded data of such subjects. At the time, the statistical significance was tested using the following 4 models: (1) a model where only the age was adjusted according to the adjusted variables (model 1); (2) a model where the age, diabetes, dyslipidemia, BMI, a family history of hypertension, and a past history of antihypertensive medication were adjusted (model 2); (3) a model where smoking, drinking, exercise, daily sodium intake, a past history of oral contraceptive therapy, a past history of estrogen therapy, and the age of menarche were adjusted in model 2 (model 3); and (4) a model where years of education and the level of income were adjusted in model 3 (model 4). Moreover, to identify a correlation between parity and hypertension, we calculated the odds ratio (OR) and 95% confidence intervals (CI) by using a logistic regression analysis. For the logistic regression analysis, we used the same 4 abovementioned models that we used for the multiple regression analysis. Statistical analysis was performed using the IBM SPSS ver. 19.0 (IBM Co., Armonk, NY, USA). A P-value of <0.05 was considered statistically significant.

As shown in Table 1, this study included 4,907 premenopausal women and 3,983 postmenopausal women. The premenopausal women comprised 4,537 (92.5%) non-hypertensive women and 370 (7.5%) hypertensive women. The mean±SE age of the non-hypertensive and hypertensive premenopausal women was 35.59±0.13 years and 45.09±0.35 years, respectively (P<0.001). The group of postmenopausal women comprised 1,914 (48.1%) non-hypertensive women and 2,069 (51.9%) hypertensive women. The mean±SE age of the non-hypertensive and hypertensive postmenopausal women was 61.16±0.20 years and 67.47±0.20 years, respectively (P<0.001). The non-hypertensive premenopausal women significantly differed from the hypertensive premenopausal women in terms of the following variables: mean BMI (P<0.001), parity (P<0.001), mean age at menarche (P<0.001), years of education (P<0.001), a past history of oral contraceptive therapy (P=0.028), a past history of estrogen therapy (P=0.010), a family history of hypertension (P<0.001), a diagnosis of diabetes (P<0.001), and a diagnosis of dyslipidemia (P<0.001). The non-hypertensive postmenopausal women significantly differed from the hypertensive postmenopausal women in terms of the following variables: mean BMI (P<0.001), parity (P<0.001), mean age at menarche (P<0.001), years of education (P<0.001), a past history of oral contraceptive therapy (P=0.028), a past history of estrogen therapy (P=0.009), a family history of hypertension (P<0.001), a diagnosis of diabetes (P<0.001), and a diagnosis of dyslipidemia (P<0.001).

Tables 2 and 3 show the correlation between parity and blood pressure. As shown in Table 2, analysis of the data of the premenopausal women, under all the models (models 1, 2, 3, and 4), showed an inverse correlation between parity and SBP (β=-0.074 [P<0.001]), β=-0.081 [P<0.001], β=-0.080 [P<0.001], and β=-0.091 [P<0.001], respectively) and between parity and DBP (β=-0.055 [P=0.002], β=-0.064 [P<0.001], β=-0.060 [P=0.001], and β=-0.069 [P<0.001], respectively). Analysis of the data of the postmenopausal women, under model 1-3, showed no statistically significant correlation between parity and hypertension (P>0.05). In model 4, however, an inverse correlation was observed between parity and SBP (β=-0.044 [P=0.028]). In the postmenopausal women, however, no significant correlation was observed between parity and DBP. As shown in Table 3, analysis of the data of the premenopausal women, under all the models (models 1, 2, 3, and 4), showed that parity was inversely correlated with both SBP and DBP. Analysis of the data of postmenopuasal women showed that parity was inversly correlated with SBP and DBP in model 4 only (β=-0.059 [P=0.022] and β=-0.054 [P=0.044], respectively).

Table 4 presents the correlation between parity and the prevalence of hypertension. Among postmenopausal women with >1 deliveries and 2> deliveries, as compared with nonparous women, according to model 4, the OR was 0.55 (95% CI, 0.310-0.985; P=0.044) and OR was 0.55 [95% CI, 0.307-0.985]; P=0.045). This finding suggests that the prevalence of hypertension was significantly lower, by approximately 45%, among postmenopausal women with >1 deliveries and >2 delivaries than among nonparous women. Among postmenopausal women with a past history of >3 deliveries, as compared with nonparous women, the OR was 0.55 (95% CI, 0.293-1.018; P=0.057).