- Top of page
- Research Methods and Procedures
Objective: To examine relationships of BMI with health-related quality of life in adults 65 years and older.
Research Methods and Procedures: In 1996, a health survey was mailed to all surviving participants ≥ 65 years old from the Chicago Heart Association Detection Project in Industry Study (1967 to 1973). The response rate was 60%, and the sample included 3981 male and 3099 female respondents. BMI (kilograms per meter squared) was classified into four groups: underweight (<18.5), normal weight (18.5 to 24.9), overweight (25.0 to 29.9), and obese (≥30.0). Main outcome measures were Health Status Questionnaire-12 scores (ranging from 0 to 100) assessing eight domains: health perception, physical functioning, role limitations-physical, bodily pain, energy/fatigue, social functioning, role limitations-mental, and mental health. The higher the score, the better the outcome.
Results: With adjustment for age, race, education, smoking, and alcohol intake, obesity was associated with lower health perception and poorer physical and social functioning (women only) but not impaired mental health. Overweight was associated with impaired physical well-being among women only. Both underweight men and women reported impairment in physical, social, and mental well-being. For example, multivariable-adjusted health perception domain scores for women were 50.8 (underweight), 62.7 (normal weight), 60.5 (overweight), and 52.1 (obese), respectively. Associations weakened but remained significant with further adjustment for comorbidities.
Discussion: Compared with normal-weight people, both underweight and obese older adults reported impaired quality of life, particularly worse physical functioning and physical well-being. These results reinforce the importance of normal body weight in older age.
- Top of page
- Research Methods and Procedures
Obesity is a major public health problem due to its increasing prevalence (1) and its associations with higher morbidity and mortality from multiple diseases (2, 3). Many cross-sectional studies have also documented that obesity is related to impaired physical functioning—one aspect of health-related quality of life (HRQoL)1 (4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16). However, previous findings on obesity and social well-being and mental health, two other aspects of HRQoL, have been inconsistent (5, 7, 8, 11). Although some studies have shown that obese persons have impaired social and mental well-being (7, 11), others reported no such associations (5, 8). It is also not clear whether observed associations of obesity with HRQoL are independent of co-existing chronic diseases (7, 8, 16), important determinants of HRQoL (17). Furthermore, most studies on overweight and obesity and quality of life have relied on data from young, middle-aged, or mixed age group populations. Research on persons 65 years of age and older (65+) is limited and has examined only physical functioning (13, 14), except one recent study of a Spanish population 60 years of age and older that included all aspects of HRQoL (15).
At present, ∼7% of the world population is 65+, and it is projected to rise to 12% by 2030, and in the U.S., it is projected to rise from 12% (35 million) to 20% (71 million) (18). In this era of population aging, the quality of life of older adults is of particular importance at both the societal and individual levels. We investigated, among men and women ages 65+ in the U.S., the cross-sectional associations of BMI with all three aspects of HRQoL (physical, social, and mental well-being) and studied whether any associations can be explained by presence of chronic conditions.
- Top of page
- Research Methods and Procedures
In 1996, average age was 74.3 ± 6.0 years for women and 72.8 ± 5.9 for men, and average BMI was 25.1 ± 4.6 kg/m2 for women and 26.4 ± 3.7 kg/m2 for men. In general, there was an inverse association between BMI and age, education, smoking, and alcohol intake among women (Table 1). Similar patterns were observed for men except for alcohol.
Table 1. . Characteristics of 3099 women and 3981 men ages 65+ by BMI categories, 1996
|Variable‡ (units)||<18.5||18.5 to <25||25 to <30||≥30.0||<18.5||18.5 to <25||25 to <30||≥30.0|
|N (%)||146 (4.7)||1499 (48.4)||1010 (32.6)||444 (14.3)||39 (1.0)||1366 (34.3)||1923 (48.3)||653 (16.4)|
|BMI (kg/m2)||17.2 (0.9)||22.1 (1.8)||27.1 (1.3)||33.2*** (3.0)||17.4 (1.5)||22.8 (1.4)||27.0 (1.4)||32.4*** (2.2)|
|Age (years)||77.1 (6.1)||75.1 (6.2)||73.6 (5.8)||72.7*** (5.2)||76.4 (7.4)||74.1 (6.3)||72.4 (5.7)||70.9*** (4.8)|
|Race (% black)||0||3.5||4.7||5.6**||10.3||2.2||2.7||3.7**|
|Education (years)||13.0 (2.6)||13.1 (2.7)||12.8 (2.6)||12.2*** (2.8)||15.1 (3.4)||14.7 (3.0)||14.4 (3.2)||13.9*** (2.8)|
|Smoking (%)|| || || || || || || || |
|Alcohol (mL/d)||8.4 (18.1)||6.0 (12.6)||4.2 (10.2)||4.2*** (13.0)||4.8 (9.8)||14.9 (23.7)||15.7 (25.5)||13.7* (24.4)|
For both men and women, age-adjusted prevalence of chronic diseases differed significantly across BMI groups (Table 2). Compared with normal-weight persons, obese men and women had higher prevalence of most chronic diseases. Underweight individuals, especially underweight men (n = 39), also had more, although not significantly more, comorbidities.
Table 2. . Age-adjusted prevalence (%) of chronic diseases for 3099 women and 3981 men ages 65+ by BMI categories, 1996
|Diseases||<18.5 n = 146||18.5 to <25 n = 1499||25 to <30 n = 1010||≥30.0 n = 444||<18.5 n = 39||18.5 to <25 n = 1366||25 to <30 n = 1923||≥30.0 n = 653|
|Any disease (out of 30 categories)¶||86.6||85.8||92.3***||94.8***||84.3||85.7||86.8||93.3***|
|Mean number of diseases,¶ if any||2.9||2.7||2.9*||3.5***||3.0||2.8||2.9||3.1***|
Figure 1 shows the Model I-adjusted prevalence of adverse responses for eight items in HSQ-12 for men and women separately. For seven of eight items, obese and underweight women had higher prevalence of adverse conditions compared with normal-weight women (Figure 1A). Results for underweight and obese men were similar for physical health items but not for social or mental health, whereas outcomes for overweight men compared favorably with those for normal-weight men (Figure 1B).
Figure 1. Adjusted prevalence (percentage) of adverse outcomes for eight HSQ-12 items by BMI categories in 1996 for 3099 women (A) and for 3981 men (B), ≥65 years of age. *p < 0.05, †p < 0.01, and ‡p < 0.001 compared with the reference group (normal weight; BMI 18.5 to 24.9 kg/m2). Adjusted for age, race, education, smoking, and alcohol intake. HP is Health Perception (“fair” or “poor”); PF is Physical Functioning (represented by one item: “limited a lot” in walking several blocks); RP is Role limitations due to Physical health (“quite a bit” or “could not do”); BP is Bodily Pain (“severe” or “very severe”); EF is Energy/Fatigue (having a lot of energy “a little” or “none of the time”); SF is Social Functioning (“interfered quite a bit” or “extremely”); RM is Role limitations due to Mental health (“quite a bit” or “extremely”); MH is Mental Health (represented by one item: “feeling downhearted and blue most or all of the time”).
Download figure to PowerPoint
Patterns for HSQ-12 items were also evident in the eight domain scores with the same set of covariates (Table 3, Model I). In general, additional adjustments for major diseases only partially attenuated associations of BMI with HRQoL measures (Table 3, Model II). Similar results were observed when the number of comorbid conditions, ranging from 0 to 14, was used to adjust for chronic illnesses (data not shown). Virtually identical patterns were also obtained from analyses using weight categories with adjustment for height. For example, Model I-adjusted HSQ-12 summary scores were 567, 574, 557, 524, and 420 of a total of 800 for women with weight <130, 130 to 149, 150 to 169, 170 to 199, and ≥200 lbs, respectively.
Table 3. . Adjusted mean scores for HSQ-12 domains and mean summary score for 3099 women and 3981 men ages 65+ by BMI categories, 1996
| || ||BMI|
| || ||Women||Men|
|HSQ-12 Domain†||Model‡||<18.5 (n = 146)||18.5 to <25 (n = 1499)||25 to <30 (n = 1010)||≥30.0 (n = 444)||<18.5 (n = 39)||18.5 to <25 (n = 1366)||25 to <30 (n = 1923)||≥30.0 (n = 653)|
- Top of page
- Research Methods and Procedures
The epidemic of obesity across all ages (1) coupled with rapid aging of the population (18) means that the proportion and number of obese older persons will be unprecedented. This study of 7080 men and women ages 65+ showed that obesity was strongly associated with lower HRQoL for both men and women in the health perception domain and physical domains including physical functioning, role limitations-physical, bodily pain, and energy/fatigue. Impaired social functioning was observed among obese women only. No statistically significant differences in mental health domains (role limitations-mental and mental health) were observed between obese persons and other individuals. Among women only, overweight was related, although to a smaller degree than obesity, to impaired physical well-being. All domains except bodily pain were adversely influenced by being underweight for both older men and women. All of these associations were independent of age, race, education, smoking, and alcohol intake. They were attenuated but not eliminated with further adjustment for comorbidities.
Previous research on BMI and HRQoL among older adults is very limited (13, 14, 15), and most studies have included people from mixed age groups (for review, see (9, 10, 12). A small clinical study of 88 older women (mean age 71 ± 5 years) reported that higher BMI was significantly associated with poorer upper and lower body function in performance tests of 18 physical tasks (13). In a study of Spanish persons 60 years of age and older (n = 3605), prevalence of suboptimal physical functioning was higher among both obese men and obese women (BMI ≥ 30 kg/m2) (15). Our finding of a strong association between BMI and physical well-being is consistent with these reports and several other studies on young and middle-aged adults in different countries (4, 5, 6, 7, 8, 11).
The lack of association between obesity and mental health was consistent with some (5, 8) but not all (7, 11) previous studies of primarily middle-aged adults. The Spanish study described above found no relation between obesity and mental health among women but a positive relation among men (15). However, longitudinal findings on BMI measured in midlife and HRQoL in older age from the same population as in the present study found a dose-response increase not only in impairment of physical domains but also, although less strongly, in impairment of social functioning and mental health domains (25). One possible explanation for the different cross-sectional and longitudinal findings is that some individuals may have gained weight in the last few years and were classified in the obese group; thus, duration of exposure may have been too short for the adverse effects of obesity to become evident, especially in the social and mental domains. In our sample, 47% of obese women and 43% of obese men reported gaining more than 10 lbs in the last 5 years, compared with 8% and 18% of normal-weight women and men, respectively.
The relationship between BMI and HRQoL is not linear; both underweight and obese older adults had considerably more chronic diseases and poorer HRQoL than those with normal weight. We also found a significant association between overweight and poor physical functioning among women but not among men. This is consistent with cross-sectional studies in young and middle-aged adults (8, 11) but in contrast to longitudinal findings from the same cohort where overweight in middle age was associated with lower physical well-being in older age for both women and men, although the association was weaker for men (25). It is plausible that the healthy BMI range for men may be different from that for women. More in-depth studies on potential gender differentials and on weight changes over time are needed.
Most chronic diseases are important determinants of HRQoL (17). In addition, obesity and underweight are related to increased risks (2, 3) of many chronic diseases that are highly prevalent among older adults (26). Therefore, it is important to examine whether comorbidities can explain the association between BMI and HRQoL. In this study, additional adjustment for comorbidities only partially attenuated the association of BMI with HRQoL, which is consistent with most previous studies (7, 8, 16). For example, a study of 155, 989 U.S. adults 18 years and older using self-reported height and weight found that adjustment for joint pain and obesity-related comorbidities diminished but did not eliminate the association between obesity and quality of life as determined by general health status and number of unhealthy days in the past month (16). On the other hand, some underweight persons may have lost weight due to underlying illness (e.g., 51.4% of underweight individuals reported losing 10 lbs or more during the past 5 years, compared with 28.4%, 23.7%, and 24.2% of normal-weight, overweight, and obese participants, respectively); thus, the relationships among underweight, HRQoL, and chronic disease are more complex.
One important limitation of this study is that BMI was computed from height measured in middle age and self-reported categorical weight instead of from current objective measurements. However, average height shrinkage has been reported to be only ∼2 cm from middle age (mean age 48 years) to old age (mean age 73 years) (27, 28), translating to a 0.6-unit difference in BMI for a person with mean height (1.7 m) and weight (75 kg). Thus, it is unlikely that our four-group BMI classifications were much affected by this phenomenon. Moreover, despite use of categorical weight, the high correlation between measured and self-reported BMI in our validation study lends support to the credibility of our findings. Another validation study from the Nurses’ Health Study also showed that self-reported weight and measured weight were highly correlated (Spearman r = 0.96) (29). Nevertheless, it has been documented that obese people tend to underreport and underweight persons overreport their weight (30), which is likely to yield underestimates of associations of BMI with HRQoL. In addition, BMI is only a proxy measure for obesity status and body composition. Direct measures of body composition (e.g., fat mass and fat-free mass) and fat distribution (e.g., waist circumference) may provide more precise estimates of the association between obesity and HRQoL than BMI per se (14).
Another limitation is that the study population was drawn from a follow-up study of participants originally from the Chicago area, with a response rate of ∼60%; thus, the results may not be generalizable. However, 37% of the respondents had moved out of the Chicago area, with residences in all 50 U.S. states in 1996. Further, our response rate of 60% compares favorably with similar mail surveys of older persons (31). The estimates are likely to be conservative due to worse outcomes and lower response rates among obese people (54.7%) compared with normal-weight individuals (60.5%). Finally, the number of underweight participants, especially men, was small. Nevertheless, patterns distinct from the other groups were observed for these men and women.
To our knowledge, this is the first study in older American men and women on relationships between all dimensions of HRQoL and BMI classified using contemporary guidelines for identification of obesity. Our main findings are that extremes of body weight (underweight and obesity) are associated with lower health perception and poorer physical functioning. The results reinforce the importance of normal body weight in older age (2).