sagepub.com/journalsPermissions.nav DOI: 10.1177/0898264309349422 http://jah.sagepub.com
Disability, and 13-Year Mortality
in Older French Adults
Mohamed Berraho, MD,
1,2Chakib Nejjari, MD, PhD,
1Chantal Raherison, MD, PhD,
3Youness El Achhab, PhD,
1Nabil Tachfouti, MD,
1Zineb Serhier, MD,
1Jean François Dartigues, MD, PhD
4and Pascale Barberger-Gateau, MD, PhD
4Abstract
Objective: To investigate the relationship between mortality and BMI in older people, taking into account other established mortality risk factors.
Methods: A total of 3,646 French community dwellers aged 65 years and older from PAQUID cohort study were included. Cox proportional- hazards analysis was used to assess association between BMI and mortality.
Results: Death occurred in 54.1% of the cohort more than 13 years: 68.99%
of the underweight (BMI <19), 52.13% of the obese (BMI >30), 51.66% of the overweight (BMI 25-30), and 51.79% of the reference participants (BMI
1Fez University, Morocco
2Equipe “Epidémiologie de la Prévention des Cancers” INSERM 897, ISPED, Bordeaux, France
3University Victor Segalen Bordeaux 2, ISPED
4INSERM, U897, Bordeaux, F-33076 France; Univ Victor Segalen Bordeaux 2, Bordeaux, F-33076 France
Corresponding Author:
Mohamed Berraho, Department of Epidemiology, Clinical Research and Community Health, Faculty of Medicine, Fez University, BP 1893; Km 2.2 Route Sidi Hrazem, Fez Morocco Email: maberraho@yahoo.fr
Berraho et al. 69
22-25) died. The relative risk of death as a function of BMI, adjusted for gender and age, formed a U-shaped pattern, with larger risks associated with lower BMI (<22.0) and for BMI of 25.0 to 30.0 and BMI ≥30. (BMI 22.0-24.9 was the reference.) After adjustment for demographic factors, smoking history, and comorbidity, increased mortality risk persisted in underweight older people, BMI <18.5 and BMI 18.5-22 (respectively, HR = 1.45, 95% CI 1.17-1.78; HR = 1.27, 95% CI 1.12-1.43) compared with reference. Overweight (BMI 25-29.9) and obesity (≥30) were not associated with increased mortality compared with the reference category (respectively, HR = 0.98, 95% IC 0.88-1.10; HR = 1.06, 95% IC 0.89-1.27). Similar relationships persisted for disabled participant.
For nondisabled participant disability did not alter the associations for BMI of 25.0 and higher but for BMI less than 22.0, the risks become insignificantly different from those for the reference group. Discussion: BMI below 22 kg/
m
2is a risk factor for 13-year mortality in older people, but our findings suggest that overweight and obesity may not be associated to mortality after adjustment for established mortality risk factors.
Keywords
body mass index, mortality, elderly, morbidity, disability
The number and proportion of elderly persons is increasing all over the world.
In addition to the presence of frequent and important polypathology, this population is characterized by a high disability prevalence (Cornoni-Huntley et al., 1991; Larrieu et al., 2004).
The prevalence of overweight and obesity is increasing even in elderly subjects. For example, the prevalence of obesity in American older adults aged 60 and more was 32.0% in 2000 (McGee, 2005). For French older adults, the overall prevalence of obesity was 10.8% in the Three-City Study in the same year (Larrieu et al., 2004).
Given the increasing incentive to struggle against obesity in all age groups,
there is increasing interest today in achieving better knowledge of the associa-
tion between BMI and mortality. Indeed, maintaining optimal BMI may
contribute to healthier individuals at middle age is well established (McGee,
2005; Stevens et al., 1998). However, controversy still persists as to whether
the relationship between BMI and mortality at younger ages persists at the
older age (Bender, Jöckel, Trautner, Spraul, & Berger, 1999; Calle, Thun,
Petrelli, Rodriguez, & Heath, 1999; Cornoni-Huntley et al., 1991; DeVore,
1993; Diehr et al., 1998; Grabowski & Ellis, 2001; Harris et al., 1998; Stevens
et al., 1998; Takala, Mattila, & Ryynänen, 1994).
Regarding older people, there are still many unresolved questions related to this association. First, very few data focus on the relationship between weight and longevity for old age. Most studies documenting a relationship between BMI and increased mortality have been performed in populations between adolescence and middle age. However, as people grow older, they gain fat mass and lose muscle mass (Williamson, 1993). In addition, the defi- nition of BMI categories in elderly people was controversial (Heiat, Vaccarino, & Krumholz, 2001). Moreover, most studies conducted for older persons focused on short-term mortality (Deschamps et al., 2002; Diehr et al., 1998; Losonczy et al., 1995; Takala et al., 1994).
Second, among the studies that suggest associations between obesity, underweight, and mortality in older people, some have been criticized particu- larly for being inappropriate in controlling comorbidity and smoking status (Kushner, 1993). Reverse causation owing to preexisting chronic disease and inadequate control for smoking status can distort the true relationship between body weight and risk of death because chronic illness and smoking are associ- ated with both decreased BMI and an increased risk of death (Willett, Dietz,
& Colditz, 1999).
The purpose of this study is to investigate the relationship between mortal- ity and body mass index (BMI) in older people. We examined the relationship between BMI at entry into a cohort of elderly people, the Personnes Agées QUID (PAQUID) study (Dartigues et al., 1991), and mortality during the 13-year follow-up, taking into consideration disability, morbidity, and smoking status.
Method
The general methodology of the PAQUID study has been described elsewhere
(Dartigues et al., 1991). A total of 3,777 subjects aged 65 years and older
were recruited in two regions of France, Gironde, and Dordogne. The sample
was representative of age and gender distribution of elderly community dwell-
ers of the area (Dartigues et al., 1991). The main variables of interest in this
analysis were mortality and BMI. Every death occurring during the follow-up
was systematically recorded. The BMI, at baseline, was calculated as self-
reported weight (kg) divided by self-reported height (m) squared, which we
divided into 6 six grades: less than 18.5 (underweight), 18.5 to 24.9 (normal
weight), 25.0 to 29.9 (overweight), 30.0 to 34.9 (Grade 1 obesity), 35.0 to
39.9 (Grade 2 obesity), and 40.0 and higher (Grade 3 obesity). For the current
analyses, the normal-weight category was divided into two categories: from
Berraho et al. 71
18.5 to 21.9 (Grade 1 normal weight) and from 22.0 to 24.9 (Grade 2 normal weight).
This cutoff was chosen to provide more conservative estimates (because for smaller cutoffs, the risks for the above-normal BMI categories were even smaller) and to adjust to the available sample size. Obesity was subsequently grouped into a single category of BMI of 30.0 and higher because of the small sample sizes in each obesity class. The considered thresholds of BMI were those recommended by the World Health Organization and the U.S. federal guidelines (Flegal, Graubard, Williamson, & Gail, 2005; WHO Expert Consultation, 2004). BMI was considered a categorical variable since the risk of mortality associated with BMI was not supposed to be linear (Dartigues et al., 1991).
Disability was measured with a hierarchical index which aggregates three domains of disability in a single measure: mobility (Rosow and Breslau scale), Activities of Daily Living (ADL) and Instrumental Activities of Daily Living (IADL; Barberger-Gateau, Rainville, Letenneur, & Dartigues, 2000).
Functional assessment variables used in this hierarchical index included five ADL items from the Katz scale (bathing, dressing, going to toilet, trans- fer, and feeding) and five IADL items from the Lawton scale for both genders (ability to use telephone, shopping, mode of transportation, responsibility for own medication, and ability to handle finances). Three household activities were added when assessing women: food preparation, housekeeping, and doing laundry; also added were three items from the Rosow and Breslau scale to assess gross mobility: walk up and down to second floor, walk half a mile, and do heavy work around the house. For each of these three domains a participant was considered “dependent” if he or she could not perform at least one activity of the domain without a given level of help. Thresholds for identifying the need for help on ADL and IADL items were those initially defined by the authors of each scale.
Level of dyspnea at baseline was obtained by a direct question using the Fletcher scale (Fletcher, Peto, Tinker, & Speizer, 1976). A participant was considered dyspneic if he or she scored 3, 4, or 5.
Smoking status was recorded as follows: former smokers who recently quit smoking (<10 years); former smokers who quit more than 10 years ago, current smokers with consumption under 20 packs per year, current smokers with a consumption of 20 packs a year or more, and never-smokers.
Comorbidity included antecedent of ischemic heart disease, antecedent of
stroke, hypertension, and diabetes. In addition, we considered the number
of medications currently taken at the time of the baseline interview as an
indicator of the severity of comorbidity.
Statistical Analysis
First, we described the distribution of BMI classes at entry into the cohort.
By univariate analysis, we compared 13-year mortality rates according to BMI classes and 13-year survival curves, which were obtained using the Kaplan–Meier methods.
To adjust to potential confounding factors, we performed a multivariate analysis by the Cox proportional-hazards model. Since age is a major factor of mortality, we used age as a time scale for survival analysis. The dependent variable was the time of death more than 13 years of follow-up; BMI of 22.0 to 24.9 was taken as the reference category. In addition to age which was already taken into account in the Cox model, and to address potential biases, relative risks and 95% confidence intervals were adjusted for gender, smok- ing status, dyspnea, and comorbidity. In a second step, we further stratified the analysis by disability level to examine whether the relationship between BMI and mortality might be mediated by disability.
To mitigate the impact of reverse, causation sensitivity analysis were con- ducted. We exclude the deaths in the first 3 years of follow-up.
Statistical analysis was performed with SAS software 9.1.
Results
The study sample included 3,646 subjects aged 65 and over, after exclusion of 131 (3.4%) subjects with missing BMI; 57.35% of participants were women. Mean age was 75.25 (SD 6.78) years.
A general description of the sample according to BMI category at baseline is presented in Table 1. With the exception of antecedent of stroke, partici- pants in each category of BMI differed significantly for all the characteristics examined here. Women were overrepresented in classes of lower BMI. Mean age decreased with increasing BMI. The prevalence of dyspnea was highest among obese persons. The prevalence of diabetes tended to strongly decrease with increasing BMI, particularly in overweight or obese older persons.
Never-smokers were more frequent in categories of low BMI, whereas there
were more former smokers among normal or overweight (including obese)
older persons. The association between BMI and the number of medications
followed a U-shaped curve. The proportion of fully independent persons was
the highest for the overweight older persons (BMI range 25-30), and that of
severely disabled persons in BMI class <18.5. Obese older persons had the
highest rate of mobility disability.
Table 1. Characteristics of the Sample According to BMI Category(N = 3,646) < 18.518.5-2222-2525-30≥30Totalp N (%)158 (4.33)749 (20.54)1,226 (33.63)1,208 (33.13)305 (8.37)3,646 (100.0) % Death (13 years of follow-up)66.9959.5551.7951.6652.1354.11.000 % Women76.5872.1054.8148.7655.412091 (57.35).000 Mean age (SD) 79.28 (7.46)76.73 (7.21)75.31 (6.69)74.29 (6.24)73.05 (6.04)75.25 (6.78).000 Physical activities13.2920.4322.3921.3613.1620.48.000 Smoking status (%) Current smokers (≥20 packs year)5.065.074.493.154.284.17.000 Current smokers (<20 packs year)5.705.345.065.975.265.46 Former smokers (<10 years)6.966.8110.2812.9411.1810.38 Former smokers (≥10 years)7.5910.4117.5422.3920.3917.49 Never-smokers74.6872.3662.6455.5658.8862.50 Disability Nondisabled9.6821.2928.5430.0820.9326.11.000 Disabled Rosow37.4242.9946.7946.0548.5045.50 Disabled Rosow + IADL44.5231.2721.1319.5027.2424.20 Disabled Rosow + IADL + ADL8.394.453.544.373.324.19 Dyspnea (%)27.5623.1219.8026.0036.3924.27.000 Diabetes (%)6.334.817.6010.6920.009.03.000 Hypertension65.6169.6175.6282.0687.5477.09.000 Ischemic heart disease antecedent19.7518.6119.3124.0526.5621.36.000 Stroke antecedent8.866.285.957.044.266.36.16 Mean number of medication (SD) 4.44 (2.66)3.99 (2.79)3.81 (2.76)4.10 (2.76)4.39 (2.88)4.02 (2.8).000 Source: PAQUID baseline, 1988-1989.
Mean BMI was 24.56 (SD 3.90) kg/m
2. Men had slightly, but statistically significant, higher mean BMI than women (25.24, SD 3.58 vs. 24.07 SD 4.06; p = .000). The overall prevalence of obesity (BMI ≥30) was 8.37% and similar in both genders.
The prevalence of underweight (BMI <18.5) was much higher in women than in men (5.79% vs. 2.38%; p = .000). BMI was significantly associated with age, mean age declining with BMI (p = .000).
Of the 3,646 individuals, 1,973 (54.11%) died over the 13 years of follow- up. The univariate analysis showed a U-shaped relationship with BMI. Death occurred in 68.99% of the underweight (BMI <18.5), 52.13% of the obese (BMI ≥30), and 51.66% of the overweight (BMI 25-30).
In Figure 1, we calculated unadjusted Kaplan Meier survival curves across different categories of BMI. Mortality was high in the thin group (BMI <18.5 and BMI 18.5-21.9).
To address the issue of potential confounding, sensitivity analyses were conducted for subjects who did not die within the first 3 years of follow-up.
As they could include many deaths due to medical conditions that were already present at baseline, we compared the BMI hazard ratios from this subgroup combined with hazard ratios from the full sample. Indeed, the results of the present study are virtually identical whether or not deaths in the first 3 years are retained or excluded (results not shown).
Figure 1. Kaplan Meier survival estimates, by BMI category
Berraho et al. 75
First, a basic model of the relative risk (RR) of death according to BMI category adjusted only for age and gender was estimated (Figure 2; Table 2, Model 1). The relative risk of death as a function of BMI, adjusted for gender and age, formed a U-shaped pattern with increased risk of mortality in the lower BMI category (<18.5 and 18.5-29.9) (HR = 1.44, 95% CI 1.17-1.77, and HR = 1.25, 95% CI 1.10-1.41, respectively), and obese category (HR 1.28, 95% CI 1.08-1.53) (Table 2, Model 1). For larger BMI (≥25), the risks were greater except for a BMI of 25.0-29.9, for which the risk was the same as for the reference group.
Adjustment of these risks for various traditionally considered confounders did not result in changes in the originally estimated risks except for a BMI of 30.0 and higher, for which the risk was the same as for the reference group (Table 2, Model 2). Similar relationships persisted for disabled participant (Table 2, Model 3). Test for interaction between BMI and disability was sta- tistically significant (p = .038).
Table 2. Multivariate Cox Models of the Relative Risk of Death According to Selected BMI Categories
BMI Kg/m2
<18.5 18.5-21.9 22.0-24.9 25.0-29.9 ≥30 Model Relative risk (95% confidence interval)
Model 1a 1.44* 1.25* 1 1.06 1.28*
N = 3,646 (1.17-1.77) (1.10-1.41) (0.95-1.18) (1.08-1.53)
Model 2b 1.45* 1.27* 1 0.98 1.06
N = 3,555 (1.17-1.78) (1.12-1.43) (0.88-1.10) (0.89-1.27)
Model 3c 1.43* 1.30* 1 0.95 1.06
(Disabled) (1.15-1.77) (1.14-1.49) (0.84-1.08) (0.87-1.28) N = 1,614
Model 4d 0.88 0.95 1 1.12 0.87
(Nondisabled) (0.32-2.42) (0.67-1.35) (0.86-1.46) (0.51-1.48) N = 923
Note: BMI = body mass index. The reference category is BMI 22.0-24.9.
a. Adjusted for gender and age.
b. Adjusted as Model 1 plus physical activities, smoking status, and comorbidity (diabetes, dyspnea, hypertension, ischemic heart disease antecedent, stroke antecedent, and number of medications consumed by participant).
c. Adjusted as Model 2 among disabled participant.
d. Adjusted as Model 2 among nondisabled participant.
*p > .05.
Conversely, for nondisabled individuals (Table 2, Model 4) with a BMI less than 22.0, the risks become not significantly different from those for the reference group.
Discussion
Our findings suggest that the relationship between mortality and BMI for older people is different than for adults. Overweight and obesity appear not to be a risk factor for 13-year mortality in a sample of community-dwelling older (≥65) individuals participating in the PAQUID cohort when adjusting for multiple potential confounders.
Contrarily, underweight (BMI <22.0) appears to be a more serious prob- lem for such individuals. Our data are consistent with most previous studies that have evaluated mortality rates among elderly patients, demonstrating higher mortality among individuals with low BMI (Flodin, Svensson, &
Cederholm, 2000; Miyazaki et al., 2002; Sergi et al., 2005; Taylor & Ostbye,
2001); that is not to say that low BMI itself causes death. Low BMI may be
Figure 2. Adjusted hazard ratio (with 95% confidence interval) of death, adjusted for gender and age in inclusionBerraho et al. 77
the result of malnutrition, as a consequence of other health problems which influence mortality (Garry, Goodwin, Hunt, Hooper, & Leonard, 1982; Landi et al., 1999) such use increases the delay of recuperation from the illnesses, decreases the immune capacities, and increases susceptibility to infections (Sullivan, 1995). Low BMI may also be a marker of frailty in elderly people, again reflecting other underlying health problems, precipitating susceptibility to illness and mortality (Heiat et al., 2001; Hogan, MacKnight, & Bergman, 2003; Visscher et al., 2000).
Given the importance of connections between obesity, disability, and mortality (Al Snih et al., 2007; Ferrucci & Alley, 2007), the analyses focused on differences in mortality-risk BMI patterns in disabled and non- disabled individuals. For nondisabled individuals, the relationship between BMI and mortality disappeared for small BMI values (Table 2, Model 4).
Disabled individuals still have high-risk mortality for BMI lower than 25.0 (Table 2, Model 3).
The difference in patterns between disabled and nondisabled persons implies that nondisabled individuals have a narrower range of tolerable devia- tions from optimal weight, corresponding to the minimum of the BMI pattern of the RR of death than disabled individuals. Therefore, health and well-being of older persons should be of concern when attempting to develop recommen- dations of optimal body weight for them (Kulminski et al., 2008).
Obesity showed no association with mortality, despite the relationship of morbidity and other factors to weight at inclusion. In the literature, controversy has persisted as to whether the relationship of obesity to increased mortality at younger ages persists into later life. Some studies (Calle et al., 1999; Cornoni- Huntley et al., 1991; Deschamps et al., 2002; Harris et al., 1998) have observed higher mortality for obese older adults, while other studies (Bender et al., 1999; Stevens et al., 1998) observed a mitigated effect relative to younger obese individuals. Still others (Diehr et al., 1998; Losonczy et al., 1995; Takala et al., 1994) found no significant relationship in the older adults population, and a final set of studies (DeVore, 1993; Grabowski & Ellis, 2001) showed a negative (or protective) effect of obesity on mortality for older adults. A recent study conducted in older men even found a paradoxical association between obesity and lower mortality risk, obesity being associated with a substantially lower mortality risk in a clinical population of non–heart-failure veterans (McAuley, Myers, Abella, & Froelicher, 2007).
The health hazards of being obese in later life may be masked by other
health risks associated with aging (Hall et al., 2000). At age 65 and over,
selection can favor disabled individuals with excessive weight, because they
are in a better position (because of prior selection, e.g., disabled individuals
who have survived can be more robust than nondisabled individuals because only the strongest individuals with disabilities survive) and have more reserves (because of excessive weight) to cope with additional stressful factors (e.g., disability and its related conditions).
The relationship between overweight and mortality is controversial. Our data suggest that, among older people, a BMI (25 to 29.9 kg/m
2) does not represent a risk factor for mortality. This result fits some previous studies (Bender et al., 1999; Dorn, Schisterman, Winkelstein, & Trevisan, 1997;
Landi et al., 1999) but not others (Deschamps et al., 2002; Katzmarzyk, Craig, & Bouchard, 2001), which have identified increased mortality associ- ated with high BMI. This discrepancy may be attributed to differences in exclusion criteria, cutoff points for categories of BMI, or management of variables during analysis (Landi et al., 1999). This situation can also be explained by selective survival of overweight individuals from middle to old age, that BMI is not a good measure of body weight and that overweight might be a protective factor (Allison, Faith, Heo, & Kotle, 1997; Al Snih et al., 2007; Elia, 2001; Takala et al., 1994).
The relationship between BMI and mortality for older adults is different from the relationship found in younger populations. It is possible that the middle-aged individuals whose health is sensitive to their weight, perhaps because of genetic or environmental factors, are less likely to survive into old age. This would result in less susceptibility among older people (survivors) to the health problems associated with being overweight (Diehr et al., 1998). It is also possible that, in old age, obesity provides a nutritional reserve to the individual in times of illness or trauma, and individuals with a higher BMI are more likely to survive acute illness (Potter, Klipstein, Reilly, & Roberts, 1995). Another reason for differences between older and younger adults is that the nature of the disease and treatment processes that affect the middle- aged and older populations changes. The excess of chronic disease in the higher weight categories tends to decrease as lighter weight people eventually contract these diseases (Van Itallie & Lew, 1990). The total disease burden increases as the population ages, which leads to different interactions with the health care system, an increase in hospitalizations, and changes in patterns of medication use and lifestyle (Diehr et al., 1998).
The results of the present study are virtually identical whether or not deaths in the first 3 years are retained or excluded. Allison and colleagues have con- cluded, based on a meta-analysis and simulation studies (Allison et al., 1997;
Allison, Faith, Heo, Townsend-Butterworth, & Williamson, 1999; Allison,
Heo, et al., 1999), that eliminating early deaths has very little impact on the
BMI–mortality relationship.
Berraho et al. 79
The main limit of our survey results from the recruiting mode of the sub- jects. In the PAQUID cohort, the participants in the survey were aged 65 years and more and lived at home at baseline. It is demonstrated that among adult subjects, obesity is associated with increased mortality and a more elevated morbidity capable of leading to institutionalization (Field et al., 2001).
However; individuals were kept in the PAQUID cohort even if they were insti- tutionalized during the follow-up, so that the proportion of persons living in institutions in the sample progressively reached that of the general population.
The BMI calculated from self-reported weight and height seems valid. In fact, the similarity between self-reported and measured weight was evaluated in a subsample of the PAQUID cohort and found to be good (Dartigues et al., 1991). Stevens and coauthors also reported that, in general, the self-reported weight and height are not different from those measured (Stevens et al., 1998).
In our study, we did not have the information about waist circumference.
Thus we could not take account abdominal obesity in our analysis. Another limitation is the single measurement of BMI at baseline; we could not take into account fluctuations thereafter.
Despite the limitations of this research, we concluded that BMI may be a useful predictor of death. A lower BMI appears to be an independent predic- tor of shortened survival in older adults whereas overweight and obesity do not. Our results also suggested that the combination of simple indicators of health, as disability and BMI, can provide important information to identify and target individuals with more elevated mortality risk among the older adult population.
Acknowledgment
The authors are grateful to Dr. Jean-François Tessier for his continuous support, encouragement, and assistance.
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interests with respect to their authorship or the publication of this article.
Funding
The authors declared that they received no financial support for their research and/or authorship of this article.
References
Allison, D. B., Faith, M. S., Heo, M., & Kotle, D. P. (1997). Hypothesis concerning the U-shaped relation between body mass index and mortality. American Journal of Epidemiology, 146, 339-349.
Allison, D. B., Faith, M. S., Heo, M., Townsend-Butterworth, D., & Williamson, D. F.
(1999). Meta-analysis of the effect of excluding early deaths on the estimated rela- tionship between body mass index and mortality. Obesity Research, 7, 342-354.
Allison, D. B., Heo, M., Flanders, D. W., Faith, M. S., Carpenter, K. M., & Williamson, D. F. (1999). Simulation study of the effects of excluding early deaths on risk factor- mortality analyses in the presence of confounding due to occult disease: The exam- ple of body mass index. Annals of Epidemiology, 9, 132-142.
Al Snih, S., Ottenbacher, K. J., Markides, K. S., Kuo, Y. F., Eschbach, K., & Goodwin, J. S.
(2007). The effect of obesity on disability vs mortality in older Americans. Archives of Internal Medicine, 167, 774-780.
Barberger-Gateau, P., Rainville, C., Letenneur, L., & Dartigues, J. F. (2000). A hier- archical model of domains of disablement in the elderly: A longitudinal approach.
Disability and Rehabilitation, 22, 308-317.
Bender, R., Jöckel, K. H., Trautner, C., Spraul, M., & Berger, M. (1999). Effect of age on excess mortality in obesity. Journal of the American Medical Association, 281, 1498-1504.
Calle, E. E., Thun, M. J., Petrelli, J. M., Rodriguez, C., & Heath, C. W. (1999). Body mass index and mortality in a prospective cohort of U.S. adults. New England Journal of Medicine, 341, 1097-1105.
Cornoni-Huntley, J. C., Harris, T. B., Everett, D. F., Albanes, D., Micozzi, M. S., Miles, T. P., et al. (1991). An overview of body weight of older persons, including the impact on mortality. The National Health and Nutrition Examination Survey I–
Epidemiologic Follow-up Study. Journal of Clinical Epidemiology, 44, 743-753.
Dartigues, J. F., Gagnon, M., Michel, P., Letenneur, L., Commenges, D., Barberger- Gateau, P., et al. (1991). The Paquid research program on the epidemiology of dementia: Methods and initial results. Revue neurologique, 147, 225-230.
Deschamps, V., Astier, X., Ferry, M., Rainfray, M., Emeriau, J. P., & Barberger- Gateau, P. (2002). Nutritional status of healthy elderly persons living in Dordogne, France, and relation with mortality and cognitive or functional decline. European Journal of Clinical Nutrition, 56, 305-312.
DeVore, P. A. (1993). Assessment of nutritional status and obesity in elderly patients as seen in general medical practice. Southern Medical Journal, 86, 1008-1010.
Diehr, P., Bild, D. E., Harris, T. B., Duxbury, A., Siscovick, D., & Rossi, M. (1998).
Body mass index and mortality in nonsmoking older adults: The Cardiovascular Health Study. American Journal of Public Health, 88, 623-629.
Dorn, J. M., Schisterman, E. F., Winkelstein, W., Jr., & Trevisan, M. (1997). Body mass index and mortality in a general population sample of men and women: The Buffalo Health Study. American Journal of Epidemiology, 146, 919-931.
Elia, M. (2001). Obesity in the elderly. Obesity Research, 9 (Suppl. 4), 244S-248S.
Ferrucci, L., & Alley, D. (2007). Obesity, disability, and mortality: A puzzling link.
Archives of Internal Medicine, 167, 750-751.
Berraho et al. 81
Field, A. E., Coakley, E. H., Must, A., Spadano, J. L., Laird, N., Dietz, W. H, et al.
(2001). Impact of overweight on the risk of developing common chronic diseases during a 10-year period. Archives of Internal Medicine, 161, 1581-1586.
Flegal, K. M., Graubard, B. I., Williamson, D. F., & Gail, M. H. (2005). Excess deaths associated with underweight, overweight, and obesity. Journal of the American Medical Association, 293, 1861-1867.
Fletcher, C. M., Peto, R., Tinker, C., & Speizer, F. E. (1976). The natural history of chronic bronchitis: An eight year follow-up study of working men in London.
Oxford, UK: Oxford University Press.
Flodin, L., Svensson, S., & Cederholm, T. (2000). Body mass index as a predictor of 1 year mortality in geriatric patients. Clinical Nutrition, 19, 121-125.
Garry, P. J., Goodwin, J. S., Hunt, W. C., Hooper, E. M., & Leonard, A. G. (1982).
Nutritional status in a healthy elderly population: Dietary and supplemental intakes.
American Journal of Clinical Nutrition, 36, 319-331.
Grabowski, D. C., & Ellis, J. E. (2001). High body mass index does not predict mor- tality in older people: Analysis of the longitudinal study of aging. Journal of the American Geriatrics Society, 49, 968-979.
Hall, D. M., Xu, L., Drake, V. J., Oberley, L. W., Oberley, T. D., Moseley, P. L., et al.
(2000). Aging reduces adaptive capacity and stress protein expression in the liver after heat stress. Journal of Applied Physiology, 89, 749-759.
Harris, T., Cook, E. F., Garrison, R., Higgins, M., Kannel, W., & Goldman, L. (1988).
Body mass index and mortality among nonsmoking older persons: The Framingham Heart Study. Journal of the American Medical Association, 259, 1520-1524.
Heiat, A., Vaccarino, V., & Krumholz, H. M. (2001). An evidence-based assessment of federal guidelines for overweight and obesity as they apply to elderly persons.
Archives of Internal Medicine, 161, 1194-1203.
Hogan, D. B., MacKnight, C., & Bergman, H. (2003) Models, definitions, and criteria of frailty. Aging Clinical and Experimental Research, 15, S3-29.
Katzmarzyk, P. T., Craig, C. L., & Bouchard, C. (2001) Original article underweight, overweight and obesity: Relationships with mortality in the 13-year follow-up of the Canada Fitness Survey. Journal of Clinical Epidemiology, 54, 916-920.
Kulminski, A. M., Arbeev, K. G., Kulminskaya, I. V., Ukraintseva, S. V., Land, K., Akushevich, I., et al. (2008). Body mass index and nine-year mortality in disabled and nondisabled older U.S. individuals. Journal of the American Geriatrics Society, 56, 105-110.
Kushner, R. F. ( 1993). Body weight and mortality. Nutrition Reviews, 51, 127-136.
Landi, F., Zuccala, G., Gambassi, G., Antonelli Incalzi, R., Manigrasso, L., Pagano, F., et al. (1999). Body mass index and mortality among older people living in the com- munity. Journal of the American Geriatrics Society, 47, 1072-1076.
Larrieu, S., Pérès, K., Letenneur, L., Berr, C., Dartigues, J. F., Ritchie, K., et al. (2004).
Relationship between body mass index and different domains of disability in older persons: The 3C study. International Journal of Obesity and Related Metabolic Disorders, 28, 1555-1560.
Losonczy, K. G., Harris, T. B., Cornoni-Huntley, J., Simonsick, E. M., Wallace, R. B., Cook, N. R., et al. (1995). Does weight loss from middle age to old age explain the inverse weight mortality relation in old age? American Journal of Epidemiology, 141, 312-321.
McAuley, P., Myers, J., Abella, J., & Froelicher, V. (2007). Body mass, fitness and sur- vival in veteran patients: Another obesity paradox? American Journal of Medicine, 120, 518-524.
McGee, D. L. (2005). Diverse populations collaboration. Body mass index and mortality:
A meta-analysis based on person-level data from twenty-six observational studies.
Annals of Epidemiology, 15, 87-97.
Miyazaki, M., Babazono, A., Ishii, T., Sugie, T., Momose, Y., Iwahashi, M., et al.
(2002). Effects of low body mass index and smoking on all-cause mortality among middle-aged and elderly Japanese. Journal of Epidemiology, 12, 40-44.
Potter, J., Klipstein, K., Reilly, J. J., & Roberts, M. (1995). The nutritional status and clinical course of acute admissions to a geriatric unit. Age Ageing, 24, 131-136.
Sergi, G., Perissinotto, E., Pisent, C., Buja, A., Maggi, S., Coin, A., et al. (2005). An adequate threshold for body mass index to detect underweight condition in elderly persons: The Italian Longitudinal Study on Aging (ILSA). Journal of Gerontology, 60, 866-871.
Stevens, J., Cai, J., Pamuk, E. R., Williamson, D. F., Thun, M. J., & Wood, J. L. (1998).
The effect of age on the association between body-mass index and mortality. New England Journal of Medicine, 338, 1-7.
Sullivan, D. H. (1995). Impact of nutritional status on health outcomes of nursing home residents. Journal of the American Geriatrics Society, 43, 195-196.
Takala, J. K., Mattila, K. J., & Ryynänen, O. P. (1994). Overweight, underweight and mor- tality among the aged. Scandinavian Journal of Primary Health Care, 12, 244-248.
Taylor, D. H., Jr., & Ostbye, T. (2001). The effect of middle- and old-age body mass index on short-term mortality in older people. Journal of the American Geriatrics Society, 49, 1319-1326.
Van Itallie, T. B., & Lew, E. A. (1990). Health implications of overweight in the elderly.
Progress in Clinical and Biological Research, 326, 89-108.
Visscher, T. L., Seidell, J. C., Menotti, A., Blackburn, H., Nissinen, A., Feskens, E. J., et al. (2000). Underweight and overweight in relation to mortality among men aged 40-59 and 50-69 years: The Seven Countries Study. American Journal of Epidemi- ology, 151, 660-666.
Berraho et al. 83
WHO Expert Consultation. (2004). Appropriate body-mass index for Asian popula- tions and its application for policy and intervention strategies. Lancet, 363, 157-163.
Willett, W. C., Dietz, W. H., & Colditz, G. A. (1999). Guidelines for healthy weight.
New England Journal of Medicine, 341, 427-434.
Williamson, D. F. (1993). Descriptive epidemiology of body weight and weight change in U.S. adults. Annals of Internal Medicine, 119, 646-649.
