Combined Effect of Famine Exposure and Obesity Parameters on Hypertension in the Midaged and Older Adult: A Population-Based Cross-Sectional Study

Objectives Undernutrition early in life may increase the incidence of adverse effects on adult health. The relations between undernutrition and obesity parameters (body mass index (BMI) and WC (waist circle)) and hypertension were often contradictory. Our study is aimed at identifying the combined effects of famine exposure and obesity parameters on hypertension in middle-aged and older Chinese. Design A population-based cross-sectional study. Setting. Data were selected from the China Health and Retirement Longitudinal Study Wave2011 (CHARLS Wave2011). Participants. The sample included 12945 individuals aged 45 to 96. Main Outcome Measurements. The study analyzed data from 12945 middle-aged and older Chinese selected from CHARLS Wave2011. Differences between baseline characteristics and famine exposure/BMI levels/WC levels were evaluated using the t-, Chi-square- (χ2-), and F-test. Then, the difference in the prevalence of hypertension between baseline characteristics was estimated by the t- and χ2-test. Finally, multivariable-adjusted logistic regression models were used to explore the associations of famine exposure and obesity parameters with odds of prevalence of hypertension. Results Among the 12945 participants, 1548 (11.96%) participants had been exposed to the Chinese famine during the fetal group, whereas 5101 (39.41%) participants and 4362 (33.70%) participants had been exposed to the famine during childhood and adolescence/adult group, respectively. Regarding the participants with BMI levels, 3746 (28.94%) were overweight, and 1465 (11.32%) were obese, whereas 5345 (41.29%) of the participants with WC levels were obese, respectively. Furthermore, 1920 (31.17%) had hypertension in males and 2233 (32.91%) in females. In multivariable-adjusted models, famine exposure and obesity parameters were related with prevalence of hypertension independently in total populations ((1) model threec, famine exposure with prevalence of hypertension: the fatal-exposed vs. no-exposed group (OR1.27; 95% CI 1.08, 1.49); childhood-exposed vs. no-exposed group (OR1.64; 95% CI 1.44, 1.87); the adolescence/adult-exposed vs. no-exposed group (OR3.06; 95% CI 2.68, 3.50); P for trend < 0.001; (2) model threee, famine exposure with prevalence of hypertension: the fatal-exposed vs. no-exposed group (OR1.25; 95% CI 1.06, 1.47); childhood-exposed vs. no-exposed group (OR1.52; 95% CI 1.34, 1.73); the adolescence/adult-exposed vs. no-exposed group (OR2.66; 95% CI 2.33, 3.03); P for trend < 0.001; (3) model threeg, BMI levels with prevalence of hypertension: overweight vs. normal (OR1.75; 95% CI 1.60, 1.91); obesity vs. normal (OR2.79; 95% CI 2.48, 3.15); P for trend < 0.001; (4) WC levels with prevalence of hypertension: overweight vs. normal (OR1.42; 95% CI 1.36, 1.48)). When stratified by sex, results in both males and females were mostly similar to those in the total population. In general, interaction analysis in the multivariable-adjusted model, compared with the combination of normal BMI/WC levels and no-exposed famine group, all groups trended towards higher odds of prevalence of hypertension (the greatest increase in odds, adolescence/adult-exposed group with obesity in BMI levels: (OR8.13; 95% CI 6.18, 10.71); adolescence/adult-exposed group with obesity in WC levels: (OR6.36; 95% CI 5.22, 7.75); P for interaction < 0.001). When stratified by sex, the results in both males and females were also similar to those in the total population. Conclusion Our data support a strongly positive combined effect of famine exposure and obesity parameters on hypertension in middle-aged and elderly Chinese.


Introduction
Hypertension or elevated blood pressure (BP) is a severe medical condition that significantly increases the risks of cardiovascular diseases (CVD) as well as other chronic diseases [1][2][3], such as congenital heart disease, heart failure, heart attack, peripheral vascular disease, stroke, and vascular disease. Hypertension modifiable risk factors [4][5][6][7][8] include age, sex, life stress, excessive drinking, high-salt diet, being overweight or obese, saturated fat and trans fats, tobacco use, low intake of fruits and vegetables, lack of physical activity, family history, low diet in vitamin D, advanced age, and coexisting diseases. Though the etiology of hypertension is complex, it was known as one of the strongest risk factors was overweight or obesity. Thus, increased body mass index (BMI) or centrally located body fat (especially waist circle (WC)) increases the risk of hypertension. In addition to known and probable risk factors for hypertension, early life malnutrition may also have an effect on hypertension.
It was hypothesized that early developmental adaptions in response to malnutrition in early life, which are the main factor for short-term survival, have adverse cardiovascular outcomes [9,10]. Historical famine exposure has provided a unique and natural opportunity to identify the hypothesis. Several previous studies [11][12][13][14][15][16][17][18][19][20][21][22][23] have provided evidence to support the relationship between famine exposure and increased risk of hypertension. Most studies [11,13,14,[16][17][18][19][20][21][22][23] have found that exposure to famine in early life could increase the risk of hypertension/BP in adulthood. Furthermore, exposure to famine has more deleterious effects on adult health for females than males [12,15]. However, other studies [13,24,25] found no relationship between famine exposure and hypertension. Therefore, the relationship between famine exposure in early life and the risk of hypertension/BP needs to be further investigated. Moreover, studies also provided that malnutrition in early life [26][27][28][29][30][31] was more positively correlated with obesity among adults in late life. Generally speaking, it is not completed understood association and interaction analysis between famine exposure and obesity parameters (BMI and WC) and hypertension in the midage and older adult.
Given the limitations of previous studies, our study analyzed data from the China Health and Retirement Longitudinal Study Wave2011 (CHARLS Wave2011) and is aimed at exploring the individual and combined effects of famine exposure and obesity parameters on hypertension after adjustment for potential confounding variables.

Study Design and
Setting. Data from the China Health and Retirement Longitudinal Study Wave2011 (CHARLS Wave2011) were used in our study. The CHARLS was a nationally representative longitudinal study conducted by the China Centre for Economic Research at Peking University [32]. In the CHARLS Wave2011, 13107 individuals were recruited for the baseline, after excluding participants with missing data, 12945 individuals were included in our study. All data are openly published as microdata at http://charls .pku.edu.cn/index/zh-cn.html with no direct contact with individuals. The Ethics Committee of the China Centre for Economic Research at Peking University approved the study; all individuals have provided informed consent before the data collection.
2.6. Statistical Analysis. Analyses were conducted using SPSS software, version 22.0 (IBM SPSS, Armonk, NY, USA). The data are presented as mean ± SD unless indicated otherwise. Means and SD (continuous data) were used to measure the continuous variable (age), and count and percentage were used to describe categorical variables (sex, education, marital status, living place, drinking status, smoking habit, eating habit, social activities, the experience of traumatic events, taking physical activity or exercise, famine exposure, BMI levels, WC levels, and hypertension categories). Betweengroup differences according to hypertension (hypertension and no-hypertension) were evaluated by the chi-square test (categorical data). Differences between baseline characteristics (sex, education, marital status, living place, drinking status, smoking habit, eating habit, social activities, the experience of traumatic events, and taking physical activity or exercise) and categories of famine exposure groups/BMI levels/WC levels were also evaluated using the chi-square test (categorical data). Age between groups was used by tor F-test. Logistic regression models were used to compute ORs with accompanying 95% CIs as estimates of associations of BMI/WC levels and exposure groups separately and in combination, with the prevalence of hypertension. Two-tailed P < 5% was considered to indicate statistical significance. The proportions on the characteristics were statistically different between the WC groups except for age and marital status. Table 2 shows the characteristics of study participants categorized by blood pressure status. Of the participants, 1920 (31.17%) reported having hypertension in male and 2233 (32.91%) in the female. Significant differences were observed in age, sex, education, marital status, living place, drinking status, the experience of traumatic events, famine groups, BMI levels, and WC groups (P < 0:05) between participants with and without hypertension. Table 3 shows the separate associations of famine exposure, BMI, and central obesity with the prevalence of hypertension. Firstly, after controlling for confounding factors including age, education, marital status, living place, smoking status, drinking status, eating habit, social activities, the experience of traumatic events, taking physical activity or exercise, and famine exposure in a multivariable logistic regression model three, higher odds of prevalence of hypertension in the total population were observed with increasing levels of BMI (overweight vs. normal: 1.75 (95% CI 1.60, 1.91); obesity vs. normal: 2.79 (95% CI 2.48, 3.15)) and WC (overweight vs. normal: 1.42 (95% CI 1.36, 1.48)) independently of famine groups only (BMI, P for trend < 0:001). When stratified by sex, the results of model three in both males and females were mostly similar to those in the total population. Secondly, after controlling for confounding factors including age, education, marital status, living place, smoking status, drinking status, eating habit, social activities, the experience of traumatic events, taking physical activity or exercise, and BMI in a multivariable logistic regression model three, higher odds of prevalence of hypertension in the total population were observed with famineexposed groups (fatal-exposed group vs. nonexposed group: 1.27 (95% CI 1.08, 1.49); childhood-exposed group vs. nonexposed group: 1.64 (95% CI 1.44, 1.87); adolescence/adultexposed group vs. nonexposed group: 3.06 (95% CI 2.68, 3.50), P for trend < 0:001) independently of BMI only (P 3 BioMed Research International Table  1: Characteristics of participants in the cross-sectional study by level of famine exposure, BMI, and central obesity (N = 12945).

Characteristics
Famine exposure P No-exposed Fetal-exposed Childhoodexposed Adolescence/ adult-exposed   BioMed Research International for trend < 0:001). When stratified by sex, the results of model three in both males and females were mostly similar to those in the total population. Lastly, after controlling for confounding factors including age, education, marital status, living place, smoking status, drinking status, eating habit, social activities, the experience of traumatic events, taking physical activity or exercise, and WC in a multivariable logistic regression model three, higher odds of prevalence of hypertension in the total population were observed with famine-exposed groups (fatal-exposed group vs. nonexposed group: 1.25 (95% CI 1.06, 1.47); childhood-exposed group vs. nonexposed group: 1.52 (95% CI 1.34, 1.73); adolescence/adult-exposed group vs. nonexposed group: 2.66 (95% CI 2.33, 3.03), P for trend < 0:001) independently of WC only (P for trend < 0:001). When stratified by sex, the results of model three in both males and females were mostly similar to those in the total population. Table 4 shows the combined associations of obesity parameters and famine exposure with the prevalence of hypertension in males. Compared with the combination of the normal BMI/WC level and no-exposed famine group, all groups trended towards higher odds of prevalence of hypertension except the obesity; furthermore, in multivariable model one, the greatest increase in odds was observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 7.38; 95% CI 4.81, 11.32; adolescence/adultexposed group and obesity in WC: OR 6.13; 95% CI 4. 54, 8.26). And similarly, in multivariable-adjusted model two, the highest odds of prevalence of hypertension were observed for the adolescence/adult exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 6.87; 95% CI 4.47, 10.57; adolescence/adult-exposed group and obesity in WC: OR 5.75; 95% CI 4.26, 7.77). Additionally, in multivariable-adjusted model three, the highest odds of prevalence of hypertension were observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 7.30; 95% CI 4.74, 11.25; adolescence/adult-exposed group and obesity in WC: OR 6.68; 95% CI 4.92, 9.07). Finally, combined associations of obesity parameters and famine exposure with the prevalence of hypertension were observed in males (Pinteraction = <0:001). Table 5 shows the combined associations of obesity parameters and famine exposure with the prevalence of hypertension in females. Compared with the combination of the normal BMI/WC level and no-exposed famine group, all groups trended towards higher odds of prevalence of hypertension; furthermore, in multivariable model one, the greatest increase in odds was observed for the adolescence/a-dult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 10.38; 95% CI 7.26, 14.48; adolescence/adult-exposed group and obesity in WC: OR 7.59; 95% CI 5. 86, 9.84). And similarly, in multivariable-adjusted model two, the highest odds of prevalence of hypertension were observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 8.88; 95% CI 6.18, 12.75; adolescence/adult-exposed group and obesity in WC: OR 6.58; 95% CI 5.05, 8.58). Additionally, in multivariable-adjusted model three, the highest odds of prevalence of hypertension were observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 8.89; 95% CI 6.19, 12.78; adolescence/adult-exposed group and obesity in WC: OR 6.59; 95% CI 5.05, 8.59). Finally, combined associations of obesity parameters and famine exposure with the prevalence of hypertension were observed in females (Pinteraction < 0:001). Table 6 shows the combined associations of obesity parameters and famine exposure with the prevalence of hypertension in the total population. Compared with the combination of the normal BMI/WC level and no-exposed famine group, all groups trended towards higher odds of prevalence of hypertension; furthermore, in multivariable model one, the greatest increase in odds was observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 9.04; 95% CI 6.89, 11.86; adolescence/adult-exposed group and obesity in WC: OR 7.05; 95% CI 5.80, 8.56). And similarly, in multivariable-adjusted model two, the highest odds of prevalence of hypertension were observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 7.94; 95% CI 6.03, 10.44; adolescence/adultexposed group and obesity in WC: OR 6.29; 95% CI 5.16, 7.66). Additionally, in multivariable-adjusted model three, the highest odds of prevalence of hypertension were observed for the adolescence/adult-exposed group and obesity combination (adolescence/adult-exposed group and obesity in BMI: OR 8.13; 95% CI 6.18, 10.71; adolescence/adultexposed group and obesity in WC: OR 6.36; 95% CI 5.22, 7.75). Finally, combined associations of obesity parameters and famine exposure with the prevalence of hypertension were observed in the total population (Pinteraction < 0:001).

Discussion
In the study, we found that the individuals who had been exposed to famine in early life had an increased risk of     11 BioMed Research International hypertension in adults. After adjustment for the full set of potential confounders, including age, education, marital status, living place, smoking status, drinking status, eating habit, social activities, the experience of traumatic events, taking physical activity or exercise, and obesity parameters (BMI or WC), the associations still can be found in males and females. Additionally, our study found that there were linear trends in the associations of BMI with hypertension. After adjustment for observed potential confounders, the associations still existed both in males and females. In summary, our study supports a strongly positive combined effect of famine exposure and obesity parameters on hypertension in middle-aged and elderly Chinese. When stratified by sex, similar results were found with respect to the association.
The Chinese famine of 1959-61 caused over 30 million excess deaths [42]. A large number of such studies have explored the associations of famine exposure during early life with the risk of hypertension in adults, and there were no consistent associations observed for these studies. Therefore, our study attempted to explore the associations between obesity parameters (BMI or WC) and hypertension based on a national study from CHARLS2011. In conclusion, the findings from our study support a strongly positive combined effect of famine exposure and obesity parameters (BMI or WC) on hypertension in middle-aged and elderly Chinese. Both nutrition intervention for exposure to the famine in early life and weight control in later life may be required to substantially reduce the risk of hypertension in later life.
The effect of the worst famine to hypertension may be masked, however, by a selection effect of survivors who might be healthier than the frail members more likely to survive. The finding is in line with Darwin's theory of survival of the fittest [43]. Individuals who were exposed to famine in early life should decrease the risk of hypertension in adults. However, this was not observed in our research. The reason for the inconsistency may be due to the environmental changes. When facing the later "rich" environment, the risk of hypertension may be increased.
Our results are partly in line with several previous studies. Although the Dutch famine and the Leningrad siege study [44][45][46] have generally agreed that early-life exposure to famine was not associated with hypertension, most current published research findings [11,12,[14][15][16][17][18][19][20][21][22][23] in China indicated that exposure to famine in early life increased the risk of hypertension. However, it was found there was no association between the Chinese famine and hypertension risk in Chongqing [24]. Such discrepancies between these studies may be a result of methodological differences in definitions of the different sample selection effect and famine exposure groups. Additionally, these studies have been criticized for not being adjusted to the confounding bias of age. To control the age confounding, we categorized the famine exposure into four exposure groups based on the birth year and we also combined the no-exposure as the reference group to identify the effect of the fetal-exposed group, childhood-exposed group, and adolescence/adult-exposed group. Our study found that early famine exposure was associated with an increased risk of hypertension. The sex difference of early life famine exposure and hypertension were common in other studies [15,17]. Furthermore, exposure to famine during early life exerted more deleterious effects on females than males. This could be explained by the fact the female may suffer more than males during early life, because of the dominance of a patriarchal mentality in China [47]. The potential mechanisms of the associations between famine exposure in early life and the increased risk of hypertension in later life were still not fully understood. Animal experiments [48,49] have proved that malnutrition in early life could result in elevated BP in later life. Additionally, epigenetic might play a role in the association between famine exposure in early life and hypertension in late life [50,51].
Though so many studies have explored the association analysis between famine exposure/obesity parameters and BP, there were only two studies that explored the combined effect of obesity parameters on the relation between famine exposure and hypertension. Yu et al. [19] found that interactions between famine and obesity on hypertension prevalence risk were not observed. In contrast, Li et al. [11] reported that a stronger interaction between obesity and famine exposure with regard to BP among individuals who were exposed to famine during fetal life and had a western dietary pattern in adults was observed. Interestingly, our data support a strongly positive combined effect of famine exposure and obesity parameters on hypertension in middle-aged and elderly Chinese. The difference between our research and others may be due to the different populations, different definitions of famine exposure cohort, and 12 BioMed Research International different confounding variables by controlling. The individuals in our study were midaged and elderly Chinese, where the mean age at recruitment was older than in Yu et al.' s study, and the level of socioeconomic development also made some contribution to ontogenetic development. In addition, the participants were similar in China and its socioeconomic background, and this phenomenon could be explained by the cumulative effect. Several limitations have to be taken into account as well. Firstly, selection bias was to be considered: famine may weed out the frail members and leave the healthier ones. Secondly, individual famine exposure data have not been collected. Lastly, not all families were equally affected by famine exposure. However, our study provided a large data that could be explored further in the combined effect of famine exposure and obesity parameters (BMI or WC) on hypertension. Moreover, a significant strength of the study is the large sample of 12945 middle-aged and older Chinese. Another strength is the analytical method of controlling for a number of confounders.

Conclusions
Our data support a strongly positive combined effect of famine exposure and obesity parameters on hypertension in middle-aged and elderly Chinese. Both nutrition intervention for exposure to the famine in early life and weight reduction in later life may be required to substantially reduce the risk of hypertension in later life.

Data Availability
All data are openly published as microdata at http://charls .pku.edu.cn/index/zh-cn.html with no direct contact with all participants.

Consent
Completion of all author declaration and consent to publish form is required.

Disclosure
The authors declare that they have no potential conflict of interest relevant to the study.