Association between parity and obesity patterns in a middle-aged and older Chinese population: a cross-sectional analysi
- PDF / 503,345 Bytes
- 8 Pages / 595.276 x 790.866 pts Page_size
- 69 Downloads / 152 Views
RESEARCH
Open Access
Association between parity and obesity patterns in a middle-aged and older Chinese population: a cross-sectional analysis in the Tongji-Dongfeng cohort study Wending Li1, Yi Wang1, Lijun Shen1,2, Lulu Song2, Hui Li2, Bingqing Liu2, Jing Yuan3 and Youjie Wang2,3*
Abstract Background: Higher parity has been implicated as a risk factor for obesity of women. The objective of the study was to examine whether parity was associated with general obesity or abdominal obesity, or both, among middle-aged and older Chinese women. Methods: A total of 12,829 Chinese women (mean age: 64.8 years) with at least one live birth were selected from the Dongfeng–Tongji Cohort Study (phase II). We used body mass index to assess general obesity, and waist-to-hip ratio (WHR), waist-to-height ratio (WHtR) and waist circumference (WC) to assess abdominal obesity. We used multivariate linear and logistic regression models to investigate the association between parity and obesity. Results: The values of all four obesity measures increased with the greater number of live births (P for trend 140 mmHg or diastolic blood pressure >90 mmHg. Statistical analysis
We summarized numerical data as means ± standard deviation (SD) and presented categorical variables as percentages. We used analysis of variance (ANOVA) or
Li et al. Nutrition & Metabolism (2016) 13:72
Page 3 of 8
χ2 test to test the difference among parity groups. We used four hierarchical models to estimate the effect and the risk of increased parity on obesity in both linear and logistic regression. Model 1 examined the relationship between parity and obesity without adjustment for any covariates. Model 2 included age plus parity. Model 3 included the variables in Model 2 plus diabetes and hypertension. Model 4 included the variables in Model 3 plus education level, marital status, physical activity, smoking status (current or passive smoker), current alcohol drinker and current tea drinker, use of contraceptives, hormone replacement therapy, menopause status and abortion. In general linear regression, we calculated the variance inflation factor (VIF) to detect possible multi-collinearity during modeling. We carried out statistical analysis of the data using SPSS statistical software (version 18.0, IBM, Inc.).
Results Table 1 presents the descriptive characteristics of the study population. Women with higher parity were more likely to be older, less educated, doing less physical exercise, married or widowed, and current or previous smokers. We also found prevalence of diabetes mellitus, hypertension or menopause to increase with parity. Multiparous women tended to show a lower prevalence of abortion, passive smoking, having a habit of drinking alcohol or tea, or having used contraceptives or hormone replacement therapy. The age-adjusted mean values of the four obesity measurements according to parity are shown in Table 2. The mean values of BMI, WC, WHtR and WHR showed an increasing trend with higher parities (P for trend
Data Loading...
