Cross-sectional studies find an elevated prevalence of depression among subjects with diabetes mellitus (DM). The causal mechanisms and temporal sequence of this association have not been clearly delineated. This study investigated the prospective relationship between DM and depressive symptoms.
The Health, Aging, and Body Composition Study was a cohort study conducted in the metropolitan areas of Memphis, Tenn, and Pittsburgh, Pa. The analysis included 2522 community-dwelling subjects, aged 70 to 79 years, without baseline depressive symptoms. Incident depressed mood was defined as use of antidepressants at follow-up visits or presence of depressive symptoms (score ≥10 on the 10-item Center for Epidemiological Studies Depression scale). Presence of incident depressed mood at 2 consecutive annual clinic visits defined the incidence of recurrent depressed mood. Diabetes mellitus status, glycosylated hemoglobin (HbA1c) level, and DM-related comorbidities were assessed at baseline. Diabetes mellitus status was further characterized as absent, controlled (HbA1c level <7%), or uncontrolled (HbA1c level ≥7%). Discrete time survival analysis was used to estimate depressive events risk.
During a mean follow-up of 5.9 years, participants with DM had a higher age-, sex-, race-, and site-adjusted incidence of depressed mood (23.5% vs 19.0%) (P = .02) and recurrent depressed mood (8.8% vs 4.3%) (P<.001) than those without DM. Diabetes mellitus was associated with a 30% increased risk of incident depressed mood (odds ratio [OR], 1.31; 95% confidence interval [CI], 1.07-1.61), which was attenuated after adjustment for DM-related comorbidities (OR, 1.20; CI, 0.97-1.48). A stronger relationship was observed between DM and recurrent depressed mood (OR, 1.91; CI, 1.32-2.76), particularly among participants with poor glycemic control.
Among well-functioning older adults, DM is associated with increased risk of depressive symptoms.
Depression is a major health problem in late life.1 Although major depression affects only 1% to 2% of community-dwelling elderly people,2 a high proportion of older persons (12%-20%) reports significant depressive symptoms.3 Depressive disorders significantly decrease quality of life4 and have important socioeconomic consequences.2 Despite this heavy burden, depression is often underdiagnosed and undertreated.5,6 Understanding factors that promote depression is a major public health priority.7
Diabetes mellitus (DM) is highly prevalent, and its prevalence is expected to increase, particularly among older adults.8 Persons with DM are at increased risk of several poor health outcomes, including cardiovascular diseases, obesity, peripheral neuropathy, renal failure, and visual deficits. Several cross-sectional studies have found an elevated prevalence of depression among subjects with DM7 that has been related to poor glycemic control, DM-related complications, and obesity.9-13 Furthermore, DM is associated with increased risk of physical disability14,15 and cognitive impairment,16 both of which are associated in turn with depression.17-20
Presence of depressive symptoms among subjects with DM negatively affects life quality and treatment adherence and increases health care expenditure. Prospective studies of subjects with DM also suggest that the course of depression is often unfavorable, and persistent depression is common.21,22 However, the causal mechanisms and the temporal sequence of the DM-depression association have not been clearly delineated because no prospective studies have formally investigated DM as a risk factor for depression. Understanding the mechanisms underlying this relationship is essential from a clinical and public health perspective. The heavy burden of DM-related diseases and impairments makes DM among the most psychologically demanding illnesses. Moreover, neuroendocrine abnormalities associated with DM, including altered activity of the hypothalamic-pituitary-adrenocortical axis with higher levels of plasma cortisol,23,24 may be involved in the development of mood disorders.25
For these reasons, we hypothesized that DM, in part through its associated burden, may increase the risk of depressive symptoms. Using data from the Health, Aging, and Body Composition (Health ABC) Study,26 we investigated the prospective association between DM and the risk of depressive symptoms in a sample of well-functioning older persons and evaluated the extent to which DM-related diseases and impairments may mediate this relationship.
We performed a secondary analysis from the Health ABC study, a prospective cohort (n = 3075) of community-dwelling adults, aged 70 to 79 years, living in areas surrounding the Pittsburgh, Pa, and Memphis, Tenn, clinical centers. Participants were recruited from a random sample of white and all-black Medicare-eligible adults, with oversampling of black participants to provide enough statistical power in each race. Eligibility criteria included (1) no difficulty walking 1/4 of a mile, climbing 10 steps, or performing basic activities of daily living; (2) no life-threatening illness; and (3) no plans to leave the area for 3 years. The present analyses are based on 2522 participants; 471 were excluded owing to ongoing depression treatment or report of depressive symptoms at baseline (among these, 22.1% had DM), 4 owing to DM onset in childhood (age ≤20 years), and 78 owing to missing data on depression treatment (n = 15) and on glycosylated hemoglobin (HbA1c) levels (n = 63). Forty-two (1.7%) were lost to follow-up. All participants gave informed consent. The institutional review boards at each study site approved all protocols.
Dm and related comorbidities
Prevalent DM was defined as self-report of physician diagnosis or hypoglycemic medication use (known DM). Among undiagnosed participants with DM, DM was defined by a fasting plasma glucose level of 126 mg/dL or higher (≥7.0 mmol/L) or after a 75-g oral glucose tolerance test.27 Poor glycemic control was defined as an HbA1c level of 7% or higher27 (uncontrolled DM). An ordinal variable for DM control was created as follows: 1, no DM; 2, controlled DM (HbA1c level <7%); and 3, uncontrolled DM (HbA1c level ≥7%). Duration of DM was determined by an interviewer-administered questionnaire. Current use of hypoglycemic medication (oral hypoglycemic agents and insulin) was determined from drug data coded using the Iowa Drug Information System ingredient codes.28
The ankle-brachial index was calculated as the systolic blood pressure (SBP) of the ankle divided by the SBP of the arm. The ankle-brachial index was used as an indicator of atherosclerosis severity.29 Prevalence of chronic conditions (coronary heart disease, cerebrovascular disease, heart failure, hypertension, kidney disease, or retinal disease) was determined through disease algorithms using self-reported, physician-diagnosed disease information, clinic data, and medication use mirroring adjudicated diagnoses in the Cardiovascular Health Study.30
Renal function was assessed by means of cystatin C, a novel and accurate marker of renal function in the elderly population.31 Inflammatory status was assessed by interleukin 6 levels (enzyme-linked immunosorbent assay kit from R&D Systems, Minneapolis, Minn). Lower extremity performance was evaluated using 6-m gait speed. Cognitive function was assessed using the Modified Mini-Mental State Examination,32 with cognitive impairment defined as a score lower than 80.33
Baseline prevalent depression was defined by self-report of depression and/or use of antidepressant drugs. Use of an antidepressant medication was considered present when the Iowa Drug Information System code indicated an antidepressant drug, and the reason for use indicated depression or mood disorder. Depressed mood at baseline was evaluated by means of the standard Center for Epidemiologic Studies Depression Scale (CES-D),34 and by the CESD-10,35 a 10-item subset of the standard CES-D. The measurement properties of the CESD-10 have shown satisfactory test-retest correlations and good predictive accuracy compared with the standard 20-item version of the CES-D.35 Baseline prevalence of significant depressive symptoms was defined by a score above 16 on the standard CES-D or by a score above 10 on the CESD-10.
Incidence of treated depression was defined as current use of antidepressant drugs assessed at the annual clinic visits at study years 2, 3, 5, and 6 in participants with no history of depression and no prior use of antidepressants. Significant depressive symptoms during follow-up were measured at study years 3, 4, 5, and 6 by means of the CESD-10. Incidence of clinically relevant depressive symptoms was identified by a score of 10 or higher35 in participants without significant depressive symptoms at the previous follow-up visit. A first study outcome was the combined incidence of treated depression or significant depressive symptoms, globally defined as incident depressed mood. To identify the incidence of more severe depressive symptoms, a second study outcome of recurrent depressed mood, defined as the presence of treated depression or significant depressive symptoms at 2 consecutive annual clinic visits, was also considered.
The following characteristics were considered as covariates: age, sex, race, study site, level of education, marital status, and smoking habits. Information on alcohol consumption was assessed by means of a standardized questionnaire26,36; alcohol intake was categorized as follows: formerly; never or currently less than 1 drink per week; currently 1 to 7 drinks per week; and currently more than 7 drinks per week. Physical activity performed during the last week was assessed. Data on time spent on climbing stairs, walking for exercise or other purposes, aerobics, weight or circuit training, and high- and/or moderate-intensity exercise activities were obtained as well as information on the intensity level. A metabolic equivalent value was assigned to each activity/intensity combination.37 The scores of performed activities were summed and multiplied by body weight to create an overall physical activity score in kilocalories per week. Participants were categorized as physically inactive if they expended less than 200 kcal/week.38
Sample characteristics were compared according to DM status and by DM glycemic control categories using the χ2 test for proportions and the analysis of variance for continuous variables. Regression analysis was used to estimate the P values for trend across DM control categories. Age-, sex-, race-, and study site–adjusted cumulative incidences were computed from logistic regression analyses using the following equation: P = adjusted odds/(1+adjusted odds).39 Discrete time-survival analysis with logistic regression was used to estimate the association between DM and the likelihood of incident depressed mood and recurrent depressed mood during the study. This method uses logistic regression to determine the odds ratios (ORs) of incident depressed mood for participants who had not previously experienced depressive events during the study. For the statistical analysis, each participant contributed data up to the time point at which he or she first reported the outcome, died, or was lost to follow-up and was thereafter censored.40 Thus, each participant potentially contributed an observation for each 1-year follow-up interval (for a maximum of 5 for participants who completed the study without depression). Variables to be included in the models were selected on the basis of their association with DM and with the study outcome at an α level of .10. To investigate the effects of DM-related comorbidities, 4 different models were built: (1) adjusted for demographic characteristics (age, sex, race, study site, and baseline CES-D score); (2) adjusted for sociodemographic and lifestyle variables (age, sex, race, study site, baseline CES-D score, smoking habits, alcohol intake, education, and physical activity); (3) adjusted for demographic variables and DM-related comorbidities and impairments (age, sex, race, study site, baseline CES-D score, hypertension, cerebrovascular disease, ankle-brachial index, obesity, 6-m walk time, and cognitive impairment); and (4) fully adjusted. Finally, to identify independent predictors of recurrent depressed mood among participants with DM, factors associated with risk of depressed mood were identified in age-, sex-, and race-adjusted analyses. All factors associated with the outcome at an α level of .10 were then included in a multivariate logistic regression model.
A total of 23.4% of participants had DM, of whom nearly 65% had poor glycemic control. The mean duration of disease among participants with known DM was 12.8 years. About 35% of the subjects with DM were newly diagnosed as having DM at baseline (Table 1). Participants with DM had a higher prevalence of cardiovascular diseases, obesity, and cognitive impairment and slower gait speed. A significant trend across DM control categories was observed for all of these conditions (Table 2).
During a mean follow-up time of 5.9 years, 77 participants (3.1%) had incident treated depression; 472 (18.7%) experienced significant depressive symptoms; 517 (20.5%) had 1 episode of incident depressed mood (treated depression or significant depressive symptoms); and 141 (5.6%) had an episode of recurrent depressed mood. The age-, sex-, race-, and study site–adjusted cumulative incidences of these events were significantly higher among participants with DM, and there was a graded relationship between DM control and the risk of depression, with subjects having uncontrolled DM at the highest risk (P<.05) (Figure 1).
Table 3 lists the ORs for the incidence of depressed mood. Adjusting for demographic characteristics, DM was associated with 30% increased risk of depressed mood (model 1 OR, 1.31; 95% confidence interval [CI], 1.07-1.61). Further adjustment for lifestyle factors did not substantially affect this relationship (model 2 OR, 1.27; 95% CI, 1.03-1.57), whereas inclusion of DM-related clinical conditions reduced the strength of the association (model 3 OR, 1.21; 95% CI, 0.98-1.49). Among the DM-related comorbidities, obesity and gait speed had the greatest impact on the DM-depression association. The clinical conditions that remained independently associated with risk of incident depressed mood were cognitive impairment (OR, 1.71; 95% CI:1.29-2.27), gait speed (OR, 0.59; 95% CI, 0.38-0.93), and obesity (OR, 1.30; 95% CI, 1.05-1.60).
When the analyses were performed using DM control categories as the main predictor, increased risk of depressed mood was limited to participants with uncontrolled DM (OR, 1.39; 95% CI, 1.09-1.76). However, adjustment for DM-related comorbidities substantially attenuated the strength of this relationship (OR, 1.23; 95% CI, 0.96-1.58). A stronger relationship was observed among DM and the risk of recurrent depressed mood (Table 4). In the analysis adjusted for demographic characteristics, DM was associated with a 2-fold increased risk of recurrent depressed mood that was only modestly attenuated after adjustment for DM-related comorbidities and full adjustment (OR, 1.91; 95% CI, 1.32-2.76). Further adjustment for incident cardiovascular and cerebrovascular diseases during the follow-up did not affect the results (OR, 1.86; 95% CI, 1.28-2.70). The increased risk of recurrent depressed mood was mainly observed among participants with uncontrolled DM. These results were not affected after further inclusion of DM duration in the fully adjusted model.
To address the potential issue of a surveillance bias, analyses were also performed considering as outcomes the incidence of significant depressive symptoms and significant depressive symptoms at 2 consecutive annual clinic visits, without considering the treatment data. Results from these analyses were consistent with the previous results; in particular, DM was still associated with a significantly increased risk of recurrent depressive symptoms (OR, 1.76; 95% CI, 1.18-2.64) mainly observed among participants with uncontrolled DM (OR, 1.99; 95% CI, 1.26-3.15).
To further investigate the effect of DM control, we further categorized participants with uncontrolled DM into 2 categories according their median HbA1c values (cutoff, 8.1%). The risk of recurrent depressed mood significantly increased across DM control categories (Figure 2) (P<.001 for trend), and participants with HbA1c levels higher than 8.1% had the highest risk of developing recurrent depressed mood (OR, 2.63; 95% CI, 1.54-4.50).
Finally, in the analysis limited to participants with DM (Table 5), HbA1c level was directly and independently associated with increasing risk of depressed mood. Duration of DM (Table 5) and DM treatment did not have any effect on the risk of recurrent depressed mood. No significant interactions of DM with sex and race were observed (P>.10 for interaction terms).
Results from this study demonstrate that among a cohort of well-functioning older adults, DM is associated with increased risk of depressed mood over a mean follow-up time of about 6 years. In particular, older adults with DM had an almost 2-fold increased risk of developing recurrent depressed mood. The increased risk of depressed mood was mainly observed among participants with DM and poor glycemic control, and HbA1c was an independent predictor of recurrent depressed mood among subjects with DM.
An association between DM and depression has been consistently reported in the literature,7 and recent longitudinal studies suggest that depression may be a risk factor for DM.41 However, the causal direction and the underlying mechanisms are not completely understood. In fact, understanding of the causal direction of this relationship is complicated by several factors that are associated with each of these conditions, and most of the previous studies7 have been cross-sectional and do not allow conclusions regarding cause-effect relationships. In addition, it is not clear whether DM is independently associated with depression or if this relationship is mediated by sociodemographic factors and DM-related comorbidities. The lack of simultaneous adjustment for these factors is a limitation of most of the previous studies. Diabetes mellitus also has important adverse consequences on physical14,15 and cognitive function16 that might further account for the DM-depression relationship,17-20 and to our knowledge, these impairments have never been considered in previous studies.
Findings from the present study may provide new insights into the mechanisms underlying the relationship between DM and depression in older adults. To our knowledge, this is the first study to investigate the prospective relationship between DM and incidence of depressed mood. The increased risk of depressed mood was mainly observed among subjects with poor glycemic control, and no single DM-related comorbidity per se explained this association. These observations suggest that the relationship between DM and depressive symptoms may be in part explained by the global burden of comorbidities and impairments associated with DM. However, when we considered the more specific outcome of recurrent depressed mood, which is likely to indicate a more severe mood disorder, DM was associated with a significantly increased risk that was only modestly affected by DM-related diseases and impairments. Of note, among participants with DM, HbA1c level was an independent predictor of recurrent depressed mood. Since glycemic control is a strong predictor of development and progression of DM complications,42 this finding may support the role of DM-related comorbidity in explaining the DM-depression relationship.
Obesity and lower-extremity performance were the conditions that contributed the most toward explaining the studied relationship. Recent studies suggest that obesity, through psychological, sociological, and biological factors, may predict the onset of depression.11,12 Our findings seem to confirm this observation because in the present study, obesity not only explained in part the studied association but also was an independent predictor of both depressed mood and recurrent depressed mood. Diabetes mellitus has been shown to be associated with lower-extremity performance,14,15 which is a strong predictor of poor health outcomes in older persons,43,44 including depression.18 From this point of view, our results confirm the important role of physical function in determining poor health-related outcomes in older adults.
Our findings have important clinical and public health implications. Diabetes mellitus is among the most psychologically and behaviorally demanding chronic illnesses,45 and patients with DM play a major role in their own disease management. Presence of depressive symptoms reduces quality of life46 and has been associated with poorer treatment adherence and self-care regimens such as diet, exercise, and quitting smoking,47-50 which are key components of DM management. This may lead to increasing complications and severity of the disease, which in turn may further aggravate depressive symptoms, leading to a downward spiral in the patient's health status. This also results in adverse effects on public health because the worsening of medical outcomes and the increased use of health care resources associated with depressed mood lead to higher health care costs.2,51,52
Some limitations of the present study should be considered. Subjects may have had episodes of depressive symptoms between the clinic visits that were also resolved during that period. In fact, the CES-D assesses the presence of depressed mood only in the past week and so does not capture subjects who may have had an episode of depressed mood of brief duration. Data on depressive episodes prior to study entry were not available; this may limit the possibility to fully elucidate the temporal relationship of DM and onset of depression. Information on the use of nonpharmacologic depression treatment was not available. We do not have information on change over time and on severity of DM complications. Although the analysis considered a number of potential confounders, residual confounding effect cannot be completely ruled out, since total independence from confounders cannot be established in observational studies. Finally, the narrow age range of the study inclusion criteria may limit the generalizability of the results.
In conclusion, among this cohort of well-functioning older adults, DM was associated with an increased risk of depressive symptoms. As the life expectancy of older Americans increases, DM is becoming a disease of older adults. Increasing evidence underlines the importance of geriatric outcomes such as physical disability,14,15 cognitive impairment,16 falls,53 and fractures54 as DM complications. Depression is among these key geriatric outcomes that strongly impact health-related quality of life of older adults but that are often underdiagnosed and undertreated. From this point of view, our results underline the importance of a clinical approach to the patient with DM that includes an appropriate screening for early detection and treatment of depressive symptoms.
Correspondence: Cinzia Maraldi, MD, Department of Clinical and Experimental Medicine, Section of Internal Medicine, Gerontology, and Geriatrics, University of Ferrara, Via Savonarola 9, 44100, Ferrara, Italy (email@example.com).
Accepted for Publication: January 10, 2007.
Author Contributions: Dr Maraldi had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Maraldi, Volpato, Kritchevsky, and Pahor. Acquisition of data: Simonsick, Kritchevsky, and Pahor. Analysis and interpretation of data: Maraldi, Volpato, Penninx, Yaffe, Strotmeyer, Cesari, Kritchevsky, Perry, and Ayonayon. Drafting of the manuscript: Maraldi, Volpato, and Penninx. Critical revision of the manuscript for important intellectual content: Penninx, Yaffe, Simonsick, Strotmeyer, Cesari, Kritchevsky, Perry, Ayonayon, and Pahor. Statistical analysis: Maraldi, Volpato, and Strotmeyer. Administrative, technical, and material support: Simonsick and Ayonayon. Study supervision: Cesari and Pahor.
Financial Disclosure: None reported.
Funding/Support: This research was supported by the Intramural Research program of the National Institutes of Health (NIH), National Institute on Aging (NIA), by NIA contracts N01-AG-6-2106, N01-AG-6-2101, and N01-AG-6-2103, and by grant R01 HL72972 from the NIH National Heart, Lung, and Blood Institute.
et al. Depressive symptoms and the cost of health services in HMO patients aged 65 years and older: a 4-year prospective study. JAMA
1997;2771618- 1623PubMedGoogle ScholarCrossref
NE The spectrum of quality-of-life impairments in recurrent geriatric depression. J Gerontol A Biol Sci Med Sci
2002;57M134- M137PubMedGoogle ScholarCrossref
YW Depression in medical outpatients: underrecognition and misdiagnosis. Arch Intern Med
1990;1501083- 1088PubMedGoogle ScholarCrossref
MWvan den Brink
Wvan de Willige
G Recognition, management, and course of anxiety and depression in general practice. Arch Gen Psychiatry
1991;48700- 706PubMedGoogle ScholarCrossref
PJ The prevalence of comorbid depression in adults with diabetes: a meta-analysis. Diabetes Care
2001;241069- 1078PubMedGoogle ScholarCrossref
et al. Projection of diabetes burden through 2050: impact of changing demography and disease prevalence in the U.S. Diabetes Care
2001;241936- 1940PubMedGoogle ScholarCrossref
RE Depression and poor glycemic control: a meta-analytic review of the literature. Diabetes Care
2000;23934- 942PubMedGoogle ScholarCrossref
PJ Association of depression and diabetes complications: a meta-analysis. Psychosom Med
2001;63619- 630PubMedGoogle ScholarCrossref
GA Prospective association between obesity and depression: evidence from the Alameda County Study. Int J Obes Relat Metab Disord
2003;27514- 521PubMedGoogle ScholarCrossref
et al. Depression and glycemic control in Hispanic primary care patients with diabetes. J Gen Intern Med
2005;20460- 466PubMedGoogle ScholarCrossref
et al. Progression of lower-extremity disability in older women with diabetes: the Women's Health and Aging Study. Diabetes Care
2003;2670- 75PubMedGoogle ScholarCrossref
et al. Diabetes and physical disability among older U.S. adults. Diabetes Care
2000;231272- 1277PubMedGoogle ScholarCrossref
et al. Is diabetes associated with cognitive impairment and cognitive decline among older women? Arch Intern Med
2000;160174- 180PubMedGoogle ScholarCrossref
M Clinically defined vascular depression. Am J Psychiatry
1997;154562- 565PubMedGoogle Scholar
C The emergence of depressive symptoms in late life: the importance of declining health and increasing disability. J Community Health
1990;1593- 104PubMedGoogle ScholarCrossref
RGvan der Mast
RC Temporal relation between depression and cognitive impairment in old age: prospective population based study. BMJ
RE Depression in adults with diabetes: results of 5-yr follow-up study. Diabetes Care
1988;11605- 612PubMedGoogle ScholarCrossref
R Hypothalamic arousal, insulin resistance and type 2 diabetes mellitus. Diabet Med
1999;16373- 383PubMedGoogle ScholarCrossref
JF Hypercortisolism in diabetes mellitus. Diabetes Care
1987;10662- 664PubMedGoogle Scholar
J Coexisting dysregulations of both the sympathoadrenal system and hypothalamic-pituitary-adrenal-axis in melancholia. J Neural Transm Gen Sect
1991;85195- 210PubMedGoogle ScholarCrossref
et al. Impact of inflammation on the relationship among alcohol consumption, mortality, and cardiac events: the Health, Aging, and Body Composition Study. Arch Intern Med
2006;1661490- 1497PubMedGoogle ScholarCrossref
American Diabetes Association, Standards of medical care for patients with diabetes mellitus. Diabetes Care
S33- S50PubMedGoogle ScholarCrossref
P Drug data coding and analysis in epidemiologic studies. Eur J Epidemiol
1994;10405- 411PubMedGoogle ScholarCrossref
et al. Ankle-arm index as a marker of atherosclerosis in the Cardiovascular Health Study. Circulation
1993;88837- 845PubMedGoogle ScholarCrossref
et al. Cystatin C and mortality risk in the elderly: the Health, Aging, and Body Composition Study. J Am Soc Nephrol
2006;17254- 261PubMedGoogle ScholarCrossref
HC The Modified Mini-Mental State (3MS) Examination. J Clin Psychiatry
1987;48314- 318PubMedGoogle Scholar
et al. Chronic kidney disease and cognitive impairment in the elderly: the Health, Aging, and Body Composition Study. J Am Soc Nephrol
2005;162127- 2133PubMedGoogle ScholarCrossref
LS The CES-D scale: a self-report depression scale for research in the general population. Appl Psychol Meas
1977;1385- 401Google ScholarCrossref
DL Screening for depression in well older adults: evaluation of a short form of the CES-D (Center for Epidemiologic Studies Depression Scale). Am J Prev Med
1994;1077- 84PubMedGoogle Scholar
et al. Relationship of alcohol intake with inflammatory markers and plasminogen activator inhibitor-1 in well-functioning older adults: the Health, Aging, and Body Composition Study. Circulation
2004;109607- 612PubMedGoogle ScholarCrossref
et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc
1993;2571- 80PubMedGoogle ScholarCrossref
JE Physical activity and coronary heart disease in women: is “no pain, no gain” passe? JAMA
2001;2851447- 1454PubMedGoogle ScholarCrossref
FJ Epidemiology: Beyond the Basics. Boston, Mass Jones & Bartlett Publishers2000;
JF Statistical Models and Methods for Lifetime Data. Hoboken, NJ John Wiley & Sons Inc1981;
F Depression as a risk factor for the onset of type 2 diabetes mellitus: a meta-analysis. Diabetologia
2006;49837- 845PubMedGoogle ScholarCrossref
et al. Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ
2000;321405- 412PubMedGoogle ScholarCrossref
RB Lower-extremity function in persons over the age of 70 years as a predictor of subsequent disability. N Engl J Med
1995;332556- 561PubMedGoogle ScholarCrossref
JM Lower extremity performance in nondisabled older persons as a predictor of subsequent hospitalization. J Gerontol A Biol Sci Med Sci
2000;55M691- M697PubMedGoogle ScholarCrossref
DH Diabetes, depression, and quality of life: a population study. Diabetes Care
2004;271066- 1070PubMedGoogle ScholarCrossref
PL Depression and the dynamics of smoking: a national perspective. JAMA
1990;2641541- 1545PubMedGoogle ScholarCrossref
TL Physiological and psychological variables predict compliance to prescribed exercise therapy in patients recovering from myocardial infarction. Psychosom Med
1982;44519- 527PubMedGoogle ScholarCrossref
IB The relationship of depressive symptoms to symptom reporting, self-care and glucose control in diabetes. Gen Hosp Psychiatry
2003;25246- 252PubMedGoogle ScholarCrossref
TW Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence. Arch Intern Med
2000;1602101- 2107PubMedGoogle ScholarCrossref
JE Depression and diabetes: impact of depressive symptoms on adherence, function, and costs. Arch Intern Med
2000;1603278- 3285PubMedGoogle ScholarCrossref
K Comorbid depression is associated with increased health care use and expenditures in individuals with diabetes. Diabetes Care
2002;25464- 470PubMedGoogle ScholarCrossref
JM Risk factors for falls in older disabled women with diabetes: the Women's Health and Aging Study. J Gerontol A Biol Sci Med Sci
2005;601539- 1545PubMedGoogle ScholarCrossref
et al. Nontraumatic fracture risk with diabetes mellitus and impaired fasting glucose in older white and black adults: the Health, Aging, and Body Composition Study. Arch Intern Med
2005;1651612- 1617PubMedGoogle ScholarCrossref