HAS OLDER ADULTS BMI A MULTIFACTORIAL AETIOLOGY ?

Ganhão-Arranhado. S 1,2, Paúl. C 1,2 and Pinhão. S 3,4 . 1. Institute of Biomedical Sciences Abel Salazar, Universidade do Porto, Porto, Portugal. 2. CINTESIS, Center for Health Technology and Services Research, Porto, Portugal. Research and Education Unit on Ageing, UnIFai, Porto, Portugal. 3. Centro Hospital São João, E.P.E., Porto. 4. Faculdade de Ciências da Nutrição e Alimentação da Universidade do Porto. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History


Methods:-
This cross-sectional study included 337 community-dwelling elders aged 65 years or more who were attending seven day care centers in Lisbon. ICBAS-UP Ethics Committee (Proj. nº 167/2016) approved the investigation and all participants signed an informed consent.
We conducted face-to-face interviews and anthropometric measurements. The Mini Mental State Examination (MMSE) was used to assess cognitive function and baseline characteristics of the sample were surveyed using a structured questionnaire, including information about sociodemographic characteristics, chronic diseases and lifestyle variables. Health: respiratory diseases, myocardial infarction (if yes, how many times it occurred), liver diseases, high blood pressure, stroke, cancer, diabetes, depression and neuropsychological problems (yes/no), self-perceived health status (in a five point scale ranged from very good to very poor).
Lifestyle: current smoking and alcohol habits (yes/no), psychological stress in the last 3 months and acute illness (yes/no), social net and relationships.
Day care center attendance: reasons to attend and type of services provided.
Statistical Analysis:-All variables were tested for normality using Kolmogorov-Smirnov and Shapiro-Wilk tests. Descriptive analysis was performed and chi-square test was used to examine association between variables and nutritional status. Multinomial regression models were conducted to assess the significance of sociodemographic, psychosocial, health and day care center attendance factors predicting underweight and obesity. The reference category was normal range, from which the other 2 categories were compared. For statistical analysis SPSS® Statistical Package for Social Sciences (IBM®, version 23) was used and a p<0.05 was considered significant.

Results:-
Of the 337 participants, almost two-thirds (210; 62.3%) were female, aged between 66 and 99 years. The overall BMI mean was 28.1±4.4kg/m 2 . BMI mean in the female group was 27.7±4.9kg/m 2 and in the male group was 27.3±3.4kg/m 2 . There was an overall prevalence of underweight of 21.7% (n=73) and obesity of 38.3% (n=129) as shown in Fig.1. In the female group almost a quarter (51) were underweight, almost a third were (66) normal weight and 44.3% were obese. Regarding to males, less than on fifth (17.3%) were underweight, more than half (54.3%) were normal and 28.3% were obese. BMI was associated with the sociodemographic profile of participants (Table1). Between the selected variables, BMI was associated with gender (X 2 =17.386, p<0.001) and age group (X 2 =15.416, p=0.003). Underweight and obesity were more prevalent among females. It was also possible to observe that underweight was more frequent on younger age groups.

718
Of the health problems considered, respiratory diseases (X 2 =9.562, p=0.008), high blood pressure (X 2 =45.455, p<0.001) and diabetes (X 2 =10.467, p=0.005) were positively associated with BMI (Table2). Obese elders presented higher prevalence of these diseases than underweight or normal ranged elderly. Consumption of alcoholic beverages (X 2 =7.189, p=0.035) is more frequent among the obese elderly and the recent experience of psychological stress or acute illness (X 2 =6.686, p=0.035) is more prevalent among the undernourished elderly.

Discussion:-
Nutritional status is related to clinical status and to different socioeconomic parameters (Donini et al., 2013). Several factors appear to give their contribution to the nutritional status in this sample. Nutritional status was associated with sociodemographic, health and social conditions. According to Pirlich & Lochs (2001), undernutrition in the elderly is a consequence of somatic, psychological and also social problems (Pirlich & Lochs, 2001). Health has been shown to be linked to nutrition (Chen, Cheng, Chuang, & Shao, 2015) and considering WHO definition of health an individualized and careful approach to underweight and obesity is required.
In the current study, females presented higher prevalence of underweight. Similarly, in another research, underweight was significantly more prevalent among female elders compared to males (Reijnierse et al., 2015) and opposite from the results observed by others researchers (Caroline Boscatto et al., 2013).
Obese participants had higher frequencies of diabetes and high blood pressure. These findings are according some reports that mentioned that senior obesity is accompanied by chronic diseases such as high blood pressure, stroke, . It has been reported that a predominant activation of the sympathetic-adrenal-medullary system may promote undereating/hypophagia and thus decrease energy intake, leading to weight loss and consequently elderly could be underweight. Indeed, exposure to stressful situations is usually associated with anorexia and subsequent restriction of food consumption since corticotropinreleasing hormone stimulates the proopiomelanocortin neurons, which produce anorexic signals and increase thermogenesis. Neuro Peptide Y secretion stress-induced suppression is also likely to be involved in the display of anorexia under stressful conditions due to decreases on central orexigenic and anxiolytic actions (Kyrou & Chrousos, 2006).
Our findings from multinomial regression suggested that being female was one of the determinants for be underweight among seniors. Similarly, Alzahrani et al. recognised that female gender was a significant predictor for underweight, once female elders may suffer more frequently from lack of appetite and consequent weight lost when comparing to males (Alzahrani, Abdelmoneim, Sayed, & Alshamrani, 2016).
Age is recognised as one of the major risk factors for nutrition in the elderly (Ji, Meng, & Dong, 2012; Simsek et al., 2013) and in this survey we also found an association between age underweight. However, some studies stated that there is no relation between age and nutritional status (Forster & Gariballa, 2005). In contrast, among the participants, the oldest elderly presented the higher risk of underweight and the 75-84 age group was associated with a 57.6% lower risk of be underweight. It has been well documented that several aging factors contribute to poor nutritional condition on elders, namely changes in sensory abilities such as loss of vision, smell and taste and the increased requirements for certain nutrients may predict adverse outcomes on weight decline (Chwang, 2012;Hickson, 2006). No data were collected in this survey to discuss specificity of nutrients.
Other factor that contributed to underweight among the elderly in this study was inability to cook. Accordingly, recent literature on elders reported association between poor nutrition status and inability to shop, prepare and cook food (Donini et al., 2013;Hanandita & Tampubolon, 2015;Vedovato et al., 2016).

721
A noteworthy result was regarding to diabetes and high blood pressure, which were found to decrease underweight risk among participants in 66.4% and 54.2%, respectively. Given the fact that these two conditions can promote the development of metabolic syndrome and thus increase morbidity and mortality, one might speculate that close follow-up of diabetic and hypertensive elders on primary care services may promote more disease knowledge and more effectiveness on these pathologies management among this subpopulation.
Weight may have a protective role on elders and although geriatric obesity definition is still controversial and the appropriated BMI cut-off values for the elderly aren't yet defined (Cetin & Nasr, 2014;Reijnierse et al., 2015), previously published works suggested that ideal elderly BMI may increase with age (Schrock et al., 2017). Higher BMI may be a protective outcome due to promote greater reserves of body fat for stressful situations on elders and also lower rates of injury from falls (Cassell & Gleaves, 2006).Regarding to obesity, an association with gender was found and it was consistent with reports from others researchers. In fact, being a female was also a determinant for obesity among the sample. Menopause may be one of the main causes of the higher risk of overweight in female (Inelmen et al., 2003). Post-menopause period is accompanied by decreased 17βestradiol concentrations and increased visceral fat tissue (Björntorp, 2001). Also, female geriatric subjects suffer frequently from osteoarthritis, which together with changes in the body composition of female aging (more adipose tissue and reduction of free-fat mass) can lead to impaired functionality and increased weight (Inelmen et  . High blood pressure was associated with higher odds of having obesity in our sampled population. Since both obesity and hypertension are diet-related illnesses, a possible explanation is that due to economic constrains and to physiological loss of taste, elders may prefer palatable cheaper food with more sodium and sugar content as well as energy density. These foods are plenty of energy which together with sedentary lifestyle characteristic from this age group may increase risk of obesity (Dhurandhar, 2016). But other reasons are possible: with aging there is a predominant loss of skeletal muscle and a progressive increase in body fat tissue, especially over age 70 and also a central fat redistribution to the body trunk (Gill et al., 2015). Indeed, hypertension is frequent in elderly who have higher waist circumference (Cetin & Nasr, 2014). Taking into account the above, we hypothesise that probably this possible underlying centripetal adiposity may lead to high blood pressure. Although the paucity of studies about the complex relation between obesity and hypertension, there's evidence that imbalances in leptin hormone levels may exacerbate sympathetic nerve activity and increase blood pressure It is still debated the association between alcohol consumption and body weight (Fares et al., 2012;Liangpunsakul, Crabb, & Qi, 2010). Drinking alcoholic drinks was found more frequent among obese and it was the only lifestyle factor identified as a determinant for high BMI. Beyond fat (9kcal/g), ethanol has de highest energy density (7.1kcal/g) and can increase daily energy intake, and thus playing an important role on obesity, by representing "empty calories" and leading to weight gain (Liangpunsakul et al., 2010). Another hypothesis is that alcohol consumption can influence satiety hormones via central mechanisms, with effects on opioid, serotonergic and GABAergic brain pathways and therefore increasing appetite (Traversy & Chaput, 2015). However, when interpreting the results of the current research the cross-sectional design of the study should be considered, once it constrains to determine causality.
In conclusion, this research may contribute to the knowledge on elderly's weight and nutritional status by identifying factors influencing it. These data support the fact that an BMI in older adults has a multifactorial aetiology and poor nutrition status was found to be related to social and health circumstances. Thus, the evaluation of seniors should integrate a holistic approach and consider all dimensions of the concept of health and not only weight status and their classification.