Gender differences in the association between body mass index and health-related quality of life among adults:a cross-sectional study in Shandong, China

A recent pooled analysis of 1698 population-based study with 19.2 million participants from 200 countries estimated that global age-standardized mean body mass index (BMI) was in an increasing trend for both men and women from 1975 to 2014, which will become a serious public health concern if post-2000 trends continue [1]. Some previous research showed that high BMI (overweight or obese) was an important risk factor for a variety of chronic diseases [2‐5], and that may result heavy health and economic burden, such as increased mortality and morbidity and higher annual medical care cost [6‐9]. In China, the prevalence of high BMI (overweight and obesity) has been increasing and is fast approaching epidemic proportions [10]. The Report on Nutrition and Chronic Diseases of Chinese Residents (2015) revealed that the prevalence of overweight and obesity among adults aged 18 years were 30.1 and 11.9%, increased by 7.3 and 4.8 percentage compared with that in 2002 separately. Whether for young people or adults, the growth rate of high BMI in China was higher than that in the most developed countries [11]. Simultaneously, much lower BMI (underweight) remained prevalent and was also a serious problem in China [12]. There was evidence that much lower BMI (underweight) was associated with physical, functional, and psychological impairment and increased risk of chronic disease and mortality [13, 14]. Therefore, both much higher and lower BMI are likely to impair the overall health of population.

Health-related quality of life (HRQOL) has been recognized as a valid health indicator with the realization that biomedically oriented measures of a population’s well-being, such as mortality and morbidity rates, to provide a partial picture of public health needs and prevention outcomes [15, 16]. Some studies have examined the association between BMI and HRQOL in general population or specific population [17‐19].

However, the impact of BMI on HRQOL may vary by gender. Some studies have showed that much higher or lower BMI was negatively associated with HRQOL in both men and women, including decreasing physical or emotional well-being or both [18, 19], while some study indicated that at higher BMI values, men reported higher HRQOL than women; at lower BMI values, HRQOL was lower in men than women [20], and another study reported that the association between obesity and lower HRQOL remained significant for women, but not for men [21]. In the current sociocultural context, people still value physical attractiveness and thinness, particularly in women, which may contribute in higher risk of experiencing low self-esteem, negative body image, and depression for women than men. By contrast, some people regard obesity a protective factor, especially for men or patients with chronic diseases. As a few studies have reported, high BMI (overweight or obesity) was associated with slightly better HRQOL as compared to those with normal weight [22‐24]. Excess weight is paradoxically associated with a decreased risk of adverse outcomes, and the ‘obesity paradox’ has been proposed as a label for above unexpected results [25, 26]. Another study conducted by Yanbo Zhu [27] showed that those with the class I obese (25–29.9 kg/m²) had better HRQOL in both the physical and the mental domains than those with normal weight, especially in the mental well-being, which extended the “obesity paradox” to HRQOL outcomes. Therefore, there is no consistent recognition of the association between high or low BMI and HRQOL.

Most studies assessing the association between BMI and HRQOL among Chinese adults focused on obesity group or weight in age-specific groups [28‐30], or did not further explore the gender difference explicitly for the above association [12]. A better understanding of the mechanisms will help in the design of more targeted and appropriate interventions. However, the differences on BMI- HRQOL association between men and women are still inconclusive, and no similar in-depth investigations were conducted among adults from Shandong Province, where the total population ranks second in China. To fill the gap literature, the aims of our study are 1) to evaluate the association between BMI and HRQOL among adults sampled from Shandong Province, China, and 2) to further explore gender differences on association between BMI category and HRQOL among adults.

Gender differences on association between BMI and HRQOL was investigated explicitly in our study. Overall, overweight and obese adults accounted for 30.7 and 4.4%, respectively, and there was a higher proportion of overweight/obese men than women. The highest mean EQ-5D utility value and VAS score was achieved by men of obese group and by women of normal weight group, respectively. After controlling potential confounders, gender had a significant effect in the association between BMI categories and HRQOL, especially for obese group.

A lot of attention was paid to dealing with obesity-HRQOL, and the relationship between underweight and HRQOL was likely to be ignored. Nevertheless, an obvious fact in our study was that being underweight was associated with the worst HRQOL and our findings were broadly consistent with those reported in many studies [12, 17, 23, 27, 37]. Previous studies have reported that underweight was associated with increased risk of excess mortality [38], poor cognition [39, 40], and poor self-rated health [41]. A previous study based on population in five cities of China also showed that there was more percentage of underweight persons (8.64%) than obesity (3.27%) in China [29]. In view of this, the neglected underweight problems need as much attention as obesity.

Generally, overweight and obesity were negatively associated with HRQOL, a surprising but interesting finding in our study was that obesity had a negative impact on women’ HRQOL after adjusting for other influential factors including chronic diseases, but not for men. A possible explanation about gender differences in obesity was that women might be more susceptible to distress to body weight or body image than men [42, 43]. Excessive dieting may result in lower HRQOL among obese women. It could also be explained by cultural perceptions about personal body weight and more discrimination against women with excess body weight in their work-related life and social roles [44]. Considering this, we performed another Tobit regression analysis to estimate the effects of the interactions of gender by BMI on HRQOL. As the results showed, women suffered from the negative effects of obesity on HRQOL even after adjusting for demographic, life style and chronic condition variables. All of those imply that we need to pay more attention on the overweight and obese females when improving HRQOL. Awareness to challenges facing those women can help health care providers work more effectively with them to develop different plans for losing weight, adopting healthier lifestyles.

The relationship between BMI and HRQOL in our study varied by gender, especially for population with obesity. Many studies have shown that the “obesity paradox”, a “paradoxical” decrease in morbidity and mortality with increasing BMI, existed in elderly or patients with chronic conditions [24, 45‐47]. Recent studies extended the “obesity paradox” to HRQOL outcomes, suggesting that overweight and obesity may paradoxically correlate with higher HRQOL, called “obesity-HRQOL paradox” [22, 27, 48]. There has been some emerging evidence that the association between obesity and HRQOL may be stronger in women than men [29, 49, 50]. According to the available data, our finding is first to verify that this “obesity-HRQOL paradox” phenomenon appeared in Chinese male adults. Cultural and social atmosphere could well explain why obese men was positively associated with HRQOL. In Chinese traditional culture, being fat was not a symbol of unhealthiness because only the wealthier people can afford eating more and can put on more weight. A more recent Chinese saying described becoming fat during middle age as acquiring a good fortune [51]. Also, it may be a matter of special concept in China that some people regarded fat as “robust”, “affluent” state, especially for men [52]. Our results were also consistent with the results reported by Audureau et al. [53] that the significant “obesity-HRQOL paradox” were found in men for mentally oriented HRQOL domains. The increased amounts of nonfat and muscle mass, which are important contributors to fitness or overall health, might be another important concept in understanding this obesity paradox [54]. However, there is wide agreement that BMI is seriously flawed because it is unable to distinguish between lean body mass and fat mass, and BMI may underestimate the number of people for whom fatness impacts health [55]. Therefore, further studies based on other measures should be carried out to was needed to explain the phenomenon, but in this study, we confirmed that gender as a moderator had a significant effect on the relationship between BMI on HRQOL among adults. Accordingly, more targeted and appropriate interventions should be conducted, especially in women, to decrease the detrimental obesity effects associated with HRQOL. Furthermore, given that HRQOL was used as input for cost-effectiveness analyses, findings from our study are important for identifying population subgroups in which weight management programs are potentially most cost-effective.

The limitations of this study should be acknowledged. Firstly, this was a cross-sectional study which may result that the causality between BMI and HRQOL cannot be inferred despite the clear association. Secondly, our analyses were not weighted by the appropriate sampling weights. However, a series of quality and consistency tests including Myer’s Index (Myer’s index =2.99), Test of Goodness for Fit(χ² = 12.59, P>0.05), DELTA No-similarity Coefficient (Δ = 0.0915) and GINI Concentration Ratio (GINI = 0.1079) showed that our survey data had high quality and the sample had a strong population representativeness of Shandong Province. Thirdly, we did not conduct more different subgroup analyses in this study, and will have a more explicit analysis in the follow-up study. In addition, the NHSS only collected the BMI data as a measure of obesity, we could not compare BMI with other measures, such as waist-to-hip ratio and visceral measurement, to explore more accurately and comprehensively gender difference in the association. However, there are also several strengths in this study. Firstly, the data we used were sampled from the general population of the Shandong’ NHSS 2013, which gave this analysis excellent power. More importantly, we confirmed that the gender differences on relationship of BMI and HRQOL by adding the interaction term of gender by BMI into regression model. Finally, our present study verified the “obesity paradox” to HRQOL outcomes in the male adults.

In this population-based study, we found that the underweight group had the lowest HRQOL, which deserved more attention. The effect of obesity on HRQOL in adults differed by gender, and it is critical to develop effective gender-specific weight control programs. The counter-intuitive “obesity- HRQOL” phenomenon was verified in male adults. Further research based on other measures are needed to investigate the mechanisms that underpin the gendered nature of obesity prevalence.