THE links between depression and coronary heart

Transcription

1 Journal of Gerontology: MEDICAL SCIENCES 2005, Vol. 60A, No. 1, Copyright 2005 by The Gerontological Society of America Depressive Symptoms and Development of Coronary Heart Disease Events: The Italian Longitudinal Study on Aging Chiara Marzari, 1 Stefania Maggi, 1 Enzo Manzato, 1 Carla Destro, 1 Marianna Noale, 1 Daniele Bianchi, 1 Nadia Minicuci, 1 Gino Farchi, 2 Marzia Baldereschi, 3 Antonio Di Carlo, 3 Gaetano Crepaldi, 1 and the Italian Longitudinal Study on Aging Working Group 1 National Research Council, Aging Branch, Institute of Neuroscience, Padova, Italy. 2 Istituto Superiore di Sanità, Rome, Italy. 3 National Research Council, Institute of Neuroscience, Florence, Italy. Background. Studies on the association between depressive symptomatology (DS) and cardiovascular events and mortality in elderly persons have yielded contradictory findings. To address this issue, the authors assessed DS and an extensive array of sociodemographic, behavioral, and biological variables in the largest population-based sample of older Italians ever studied and analyzed their association with coronary heart disease (CHD) morbidity and total number of deaths. Methods. This prospective, community-based cohort study included a sample of 5632 Italians, 65 years and older, who were recruited from the demographic registries of eight municipalities in Italy. Depressive symptomatology was assessed using the Geriatric Depression Scale, and a score 10 was used to indicate the presence of DS. All traditional cardiovascular disease risk factors were assessed at baseline, through questionnaires, blood tests, and physical examinations. The outcomes were CHD fatal and nonfatal events and total number of deaths. The association of the predictive variables with the outcomes was assessed using different Cox models. Results. Baseline DS was associated with a higher incidence of fatal and nonfatal CHD events (hazard ratio [HR], 1.66; 95% confidence interval [CI], ) and with cardiovascular mortality in men (HR, 2.49; 95% CI, ) and with total mortality in men (HR, 2.02; 95% CI, ) and women (HR, 1.43; 95% CI, ) at the 4-year follow-up assessment. This association was observed after adjusting for a vast array of potential confounding variables, including major chronic conditions. Conclusions. Depressive symptomatology confers an increased risk for CHD in men and for total mortality in men and women but is not explained by health behaviors, social isolation, or biological or clinical determinants. THE links between depression and coronary heart disease (CHD) and mortality continue to be debated, with some studies showing that depression is an independent risk factor for CHD and for mortality (1 11) and others finding no association (12 14). Furthermore, no studies have assessed this association in the Italian elderly, despite the notably higher prevalence rates of depressive symptomatology (DS), after age 65 years, in this population compared with persons from other countries (15). Surely the causal association can be supported only if an extensive health assessment that controls for the most important biological, behavioral, and socioeconomic predictors of CHD is considered. A recent meta-analysis (16) showed that depression is associated with the development of CHD in initially healthy persons. However, in only 2 of 11 studies were the data stratified by sex. Given that the prevalence rates of depression and its correlates differ in men and women (15), we believe that analyses stratified by sex could be more informative. In this article, we report the results of a 4-year study that prospectively investigated the relationship between DS and subsequent risks for CHD and mortality in a cohort of older Italians. The clinical evaluation and the determination of a comprehensive range of potential risk factors at baseline and follow-up allowed us to assess the association between DS and CHD incident events and the association with 4-year mortality. MATERIALS AND METHODS Participants The Italian Longitudinal Study on Aging (ILSA) has been described in detail elsewhere (17). Briefly, a random sample of 5632 persons aged 65 to 84 years, including both community-dwelling and institutionalized persons, stratified by age and sex using an equal-allocation strategy, was identified from the demographic lists of the registry offices of 8 municipalities. Eighty-eight persons of each sex in four age groups (65 to 69 years, 70 to 74 years, 75 to 79 years, and 80 to 84 years) were included in the study sample. The baseline survey, performed in 1992, had two phases. The first was a screening phase, administered to all participants, that included a personal interview on selfreported conditions, results of laboratory analysis of a blood 85

2 86 MARZARI ET AL. sample obtained after an overnight fast, and findings from physical and electrocardiographic examinations. The second phase, administered to participants who screened positive in the first phase, consisted of the clinical confirmation of suspected cases of cardiovascular diseases, diabetes, Parkinson s disease, stroke, dementia, and peripheral neuropathy by a specialist (internist or neurologist) through a visit and the review of medical records. Only basic demographic data for nonrespondents were collected from a proxy; the respondents themselves provided all the data presented here. Incident and recurrent cases of nonfatal and fatal diseases in the sample were identified through a complete second interview and clinical examination in Copies of official death certificates were obtained and causes of death were classified according to ICD 9. The objectives of the follow-up survey were: (a) to estimate the incidence rates for the diseases being studied; (b) to calculate the hazard ratio (HR) for all risk factors; and (c) to assess the natural history of the diseases. Measurements A comprehensive questionnaire was administered to all participants to obtain information on sociodemographic characteristics, living arrangements, family composition, physical functioning (including activities of daily living [ADL] and instrumental ADLs), and health behaviors, such as smoking habits and alcohol consumption. Activities of Daily Living. The ADL scale (18), which includes eating, continence, transferring in and out of bed, toileting, dressing, and bathing, allows the analysis of the level of dependency in many basic activities of daily life. An ADL item score varies between 1 and 3. A score of 1 indicates that the person is completely independent, whereas a score of 3 indicates total dependence. We used an adjusted total score varying from 0.33 to 1, obtained by dividing the total score of the answered items by the sum of scores of the same items in the hypothesis of complete dependency. A score of 0.33 means that the person is completely independent. A score between 0.33 and 0.56 indicates a dependency for at most two ADLs. A score between 0.56 and 0.78 indicates a dependency for at most three ADLs. And a score between 0.78 and 1 indicates a dependency for four or more ADLs. Instrumental Activities of Daily Living. The instrumental ADL scale (19) allows the determination of the levels at which an older person functions in performing the more sophisticated tasks of everyday life, such as using the telephone, shopping for groceries or personal items, preparing meals, performing light or heavy housework, doing laundry, using public or private transportation, managing medications, and managing money. The adjusted total score took into consideration the fact that some items had not been answered by the persons interviewed because they were not applicable (i.e., many not-applicable values came from questions such as preparing meals, performing housework, and doing laundry, which most of the men answered that they never performed, and thus the information on the ability in that item was lost). We considered an adjusted score varying between 0.24 and 1, obtained by dividing the total score of the answered items by the sum of scores of the same items in the hypothesis of complete dependency. A score less than 0.26 indicates that the person is completely independent in instrumental ADLs, whereas a score between 0.26 and 0.5 indicates a mild dependency. A score between 0.5 and 0.75 indicates a moderate disability, and between 0.75 and 1 a complete dependency for instrumental ADLs. Blood determinations. Plasma lipid (total and highdensity lipoprotein cholesterol and triglycerides) and glucose levels were measured using standard enzymatic methods. The fibrinogen level was determined by electrophoresis, and blood cell counts were measured using an automatic counter. Low-density lipoprotein cholesterol was calculated using the Friedewald equation (20). Diagnosis of Health Conditions The diagnostic criteria used to define the prevalence rates of the investigated conditions included a screening phase and clinical confirmation of the disease. The criteria for the diagnosis and the prevalence rates were previously published (21). Briefly, myocardial infarction and angina were considered evidence of CHD. Myocardial infarction was assessed in two phases. In phase I (screening), a positive result of a questionnaire (Rose questionnaire or self-reported diagnosis) or findings of a diagnostic electrocardiogram were used. In phase II (medical confirmation), clinical records and findings of diagnostic electrocardiogram or documented hospital discharge diagnosis or physician diagnosis of myocardial infarction were reviewed. The participant s physician or the ILSA internist confirmed the diagnosis. Angina pectoris was assessed in two phases. In phase I (screening), a positive finding of a questionnaire (selfreported diagnosis or treatment with nitrates beta blockers, or calcium-channel blockers, or coronary artery bypass surgery) or positive responses to a Rose questionnaire was used. In phase II (medical confirmation), clinical records (coronary angiography showing 70% obstruction of any coronary artery or ST depression.1 mm on exercise testing) were reviewed. The participant s physician or the ILSA internist confirmed the diagnosis. Depressive symptomatology. The Italian version of the Geriatric Depression Scale-30 items was administered to all participants (22). We used the following scores to define DS: 10 to 19 ¼ mild DS and 20 to 30 ¼ severe DS. Participants with a total score from 0 to 9 were considered normal. Mortality. Mortality was ascertained for an average of 4 years from the baseline assessment. If a participant died, we obtained a copy of the official death certificate. We classified the causes of death according to the ICD-9 codes (23). Statistical Analysis The major focus of our analysis is the relation between DS and CHD, fatal and nonfatal events, in those participants

3 DEPRESSION AND CHD 87 who in the prevalence wave were free of CHD, as described previously, after controlling for other known sociodemographic, biological, and behavioral risk factors. We investigated an association of covariates in the group that developed a CHD event and in the control group using the chi-square test or the exact test. We evaluated the comparison of group mean values using the Student s t test; if the F test indicated that the assumption of homoschedasticity of variances was violated, then we used the Satterwhite approximation to compare means. We applied Cox proportional hazards regression models to determine the effect of covariates on time to CHD. We assessed the assumption of proportionality through the analysis of Schoenfeld residuals of the covariates introduced in the model. We performed a preliminary proportional hazards regression model with stepwise selection to determine how many variables were related to the outcome, and then we conducted a separate analysis for men and women, using the Cox model, to study the effect of each previously selected covariate on the outcome of interest, that is the development of a CHD event or mortality. We calculated HRs and 95% confidence intervals, adjusted for confounding variables, to estimate the association of each covariate with the outcomes. We performed a second analysis on the complete ILSA sample (n ¼ 5632) with a Cox proportional hazards model to estimate the HR on CHD mortality, by sex. Covariates introduced in the model are the same as in the previous analysis. Covariates We introduced in the model some covariates, including sociodemographic and biological variables that are considered traditional risk factors for CHD development and DS. In particular, we included the following variables: Age stratified in classes of 10 years (65 to 74 years vs 75 to 84 years) Marital status (married vs not married) Years of schooling ( 5vs.5) Smoking status (smoker or ex-smoker vs nonsmoker) Drinking status (drinker or ex-drinker vs nondrinker) Depressive symptomatology (DS present vs DS absent) ADL disability level (none vs mild, moderate, or severe disability) Fibrinogen ( 433 vs.433 mg/dl) Platelets ( 450 vs /mm 3 ) Total Cholesterol ( 240 vs.240 mg/dl) Triglycerides ( 165 vs.165 mg/dl) Glycemia ( 115 vs.115 mg/dl) Body mass index ( 23.9 vs.23.9 for women and 25 vs.25 kg/m 2 for men) Arrhythmia Hypertension Diabetes Congestive heart failure Claudication Stroke Dementia The metabolic and blood variables were dichotomized 0 1 on the basis of the range of normality. All the covariates introduced in the model for men satisfy the proportional hazard assumption, so in the proportional hazards regression procedure we introduced no variables in the strata to correct the estimation of all the other variables. For women, we used the extension of the Cox proportional hazards model for timedependent variables instead, because the diagnosis of stroke does not satisfy that assumption. This new variable is associated with the outcome so in the final Cox model we introduced the time interaction to estimate the relative risk for stroke. Appropriate weights were used to correct for oversampling in the oldest age groups. Analyses were performed using the SAS software package (24). RESULTS Of the 5632 persons originally sampled, 3882 (69%) were free of CHD disease at baseline. Of these 3882 persons free of CHD at baseline, 2830 had information about DS. Among them, 26.5% of men and 39.6% of women had mild DS, whereas 3.9% and 11.9%, respectively, had severe DS. Table 1 shows the sociodemographic, functional status, and metabolic characteristics according to DS at baseline in women and men. With regard to sociodemographic factors, men were more often married, had a higher education level, and were living with someone else in all three groups (normal, mild DS, and severe DS) compared with women. With respect to behavioral factors, a significantly higher percentage of men compared with women were smokers and drank alcohol. Among the diseases considered, intermittent claudication and stroke occurred more frequently among men than women in all three groups, whereas hypertension was more common among women. Finally, body mass index and the blood levels of platelets, cholesterol, highdensity lipoprotein cholesterol, and low-density lipoprotein cholesterol and T3 were significantly lower among men in all three groups. Disability in ADLs was more frequent among men than women with mild DS (35.8% vs 27.1%) and with severe DS (57.1% vs 38.8%). For 1116 (29%) participants, the clinical evaluation for CHD events at follow-up was not available, so our sample for this analysis includes 2766 participants. Table 2 shows the sociodemographic, functional status, and metabolic characteristics in the group in whom a CHD event developed between the baseline evaluation and the 4-year follow-up, and in the control group, by sex. For men, smoking habits and ADL and severe instrumental ADL disability levels were significantly higher among participants in whom CHD developed. Diabetes was also more prevalent among women in whom a CHD event developed (20.4% vs 11.5%) and among men (19.2% vs 10.6%); 14.6% of women and 14.3% of men who had a new CHD event have a previous diagnosis of stroke, and these percentages were significantly greater compared with 5.2% of women and 6.5% of men in whom the disease did not develop. The prevalence rates of DS were significantly greater in both sexes among persons in whom CHD developed (51.7% vs 38.4% in women and 32.6% vs 6% in men with mild DS and 13.8% vs 11.3% in women and 6.9% vs 3% in men with severe DS). For women, we found other significant differences in the prevalence rates of congestive heart failure (11.5% in the

4 88 MARZARI ET AL. Table 1. Distribution of Sociodemographic, Functional Status, and Metabolic Variables, According to DS (Subjects Free of CHD at Baseline) in the ILSA Cohort n ¼ 668 No DS Mild DS Severe DS n ¼ 1010 n ¼ 547 n ¼ 384 n ¼ 165 n ¼ 56 Mean Age Marital status, %,.0001,.0001,.0001 Married years of school, % Living arrangement, %,.0001, Alone With 1 person With 2 or more persons Smoking, % , , ,.0001 Drinking, % , , Arrhythmia, % , Hypertension, % Congestive heart failure Diabetes, % Claudicatio intermittens, % Stroke, % ADL disability level, % None Mild Moderate Severe ADL (dichotomized), % IADL disability level, %, None Mild Moderate Severe BMI, kg/m 2 (Female 20 24; Male 21 25) , , Fibrinogen, mg/dl , , Platelets ([times]1000/mm 3 ) , , Cholesterol, mg/dl , , HDL cholesterol, mg/dl , LDL cholesterol, mg , , Triglycerides, mg/dl Glycemia, mg/dl HbA1c T3, ng/ml T4, ng/ml Tsh, lu/ml Note: DS¼ Depressive symptomatology; CHD ¼ coronary heart disease; ILSA ¼ Italian Longitudinal Study on Aging; ADL ¼ Activities of Daily Living; IADL ¼ Instrumental Activities of Daily Living; BMI ¼ body mass index; HDL ¼ high-density lipoprotein; LDL ¼ low-density lipoprotein. group with a CHD event vs 5.4% in the control group) and in the mean values of fibrinogen (394 mg/dl vs 361 mg/dl) and cholesterol (239.5 mg/dl in the group that developed incident CHD and mg/dl in the control group). Table 3 shows the incidence rates of CHD events for women (24.3&) and for men (22.5&). The difference between these rates is not significant according to the test for the comparison of incidence rates. On the contrary, mortality rates for all causes in the subsample considered are statistically different for men and women (24.2& for women and 36.3& for men). Table 3 also presents the mortality rates for diseases of the circulatory system and for all causes for the entire ILSA sample (5632 persons). The difference between the rate for women (18&) and for men (24.9&) is statistically significant for mortality rates for diseases of the circulatory system and for total mortality (47& vs 70&). Table 4 presents the estimates of the HRs for the incidence of CHD events using the Cox proportional hazards model, including all covariates listed above. A high body mass index is a risk factor only for women (HR, 2.25; 95% confidence interval [CI], 1.03 to 4.90), whereas DS (HR, 1.66; 95% CI, 1.06 to 2.60), stroke (HR, 1.92; 95% CI, 1.01 to 3.62), and congestive heart failure (HR, 4.48; 95% CI, 1.89 to 10.56) are risk factors only for men. Diabetes is a risk factor for both men (HR, 1.82; 95% CI, 1.06 to 3.14) and women (HR, 1.88; 9% CI., 1.05 to 3.38). Table 5 shows estimates of the HR for total mortality and for cardiovascular disease mortality among women. Age,

5 DEPRESSION AND CHD 89 Table 2. Distribution of Sociodemographic, Functional Status, and Metabolic Variables at Baseline, Stratified by the Development of CHD at follow-up in the ILSA Cohort Characteristic CHD Event N ¼ 104 No CHD Event N ¼ 1262 CHD Event N ¼ 99 No CHD Event N ¼ 1301 Mean age Marital status, %.5.3 Married years of school, % Living arrangement, %.8.8 Alone With 1 person With 2 or more persons Smoking, % Drinking, % Arrhythmia, % Hypertension, % Congestive heart failure, % Diabetes, % Claudicatio intermittens, % Stroke, % ADL disability level, % None Mild Moderate Severe ADL (dichotomized), % IADL disability level, % None Mild Moderate Severe DS, % No DS Mild DS Severe DS Body Mass Index, kg/m 2 (Female 20 24; Male 21 25) Fibrinogen, mg/dl Platelets, [times] 1000/mm Cholesterol, mg/dl HDL-Cholesterol, mg/dl LDL-Cholesterol, mg/dl Triglycerides, mg/dl Glycemia, mg/dl HbA1 C T3, ng/ml T4, ng/ml Tsh, lu/ml Note: CHD¼ Coronary heart disease; ILSA ¼ Italian Longitudinal Study on Aging; ADL ¼ Activities of Daily Living; IADL ¼ Instrumental Activities of Daily Living; DS ¼ depressive symptomatology; HDL ¼ high-density lipoprotein; LDL ¼ low-density lipoprotein. ADL disability, myocardial infarction, and diabetes are risk factors in both models, whereas DS (HR, 1.43; 95% CI, 1.04 to 1.95) is a risk factor only for total mortality. Stroke and congestive heart failure are risk factors for cardiovascular disease mortality. Table 6 presents the same estimates for men, showing that age, congestive heart failure, and DS (HR, 2.02; 95% CI, 1.58 to 2.58 in the first model and HR, 2.49; 95% CI, 1.60 to 3.87 in the second model) are risk factors for all-cause mortality and for cardiovascular disease mortality. Furthermore, we performed a Cox analysis in the subsample of participants who were free of CHD in the baseline wave (n ¼ 3882) to evaluate risk factors for allcause mortality in this group. Our results show that DS is a risk factor for both women (HR, 1.68; 95% CI, 1.17 to 2.41) and men (HR, 1.63; 95% CI, 1.22 to 2.15), as in the complete ILSA sample. Finally, we performed all analyses by stratifying DS by severity (mild and severe) and we found that the associations remained significant in each stratum. Furthermore, we

6 90 MARZARI ET AL. Table 3. Incidence Rates of CHD Event and Mortality in the ILSA Cohort Test ID CHD event (n ¼ 2766) p,.09 Incident cases Incidence rate (& person year) Mortality for all causes (n ¼ 2766) p,.05 No. of deaths Annual mortality rate (&) Mortality for diseases of circulatory system (n ¼ 5632) p,.05 No. of deaths Annual mortality rate (&) Mortality for all causes (n ¼ 5632) p,.05 No. of deaths Annual mortality rate (&) Note: CHD ¼ Coronary heart disease; ILSA ¼ Italian Longitudinal Study on Aging. performed the analyses stratified by geographical location (northern, central, and southern Italy) and found no significant differences across strata. Therefore, we present the data for the overall sample to increase the statistical power of our analyses. Conclusions Depressive symptomatology has been implicated as a factor in the cause of CHD in the general population (1 13) and in the prognosis of CHD patients (25 36), but seldom have studies focused on the elderly. This study has shown that DS is associated with the development of CHD events and with total and cardiovascular disease mortality in men free of the disease at baseline and with total mortality in women older than 65 years. We could not present data separately for community-dwelling and institutionalized persons because only about 1% of our sample was institutionalized. This, indeed, represents the percentage of institutionalized persons generally reported for the Italian population. Therefore, DS seems to be a stronger risk factor in men for both fatal and nonfatal CHD events. This could be a result of selected risk factors more common among depressed men compared with controls. However, we have adjusted for most known behavioral (smoking and drinking habits), physical (physical functioning, selected chronic diseases, and biomarkers), and sociodemographic factors Table 4. Association of Predictive Variables With CHD Incident Events in the ILSA Cohort Body mass index 2.25 ( ) Depressive symptomatology 1.66 ( ) Diabetes 1.88 ( ) 1.82 ( ) Stroke 1.92 ( ) Congestive heart failure 4.48 ( ) Note: CHD ¼ Coronary heart disease; ILSA ¼ Italian Longitudinal Study on Aging; HR ¼ hazard ratio; CI ¼ confidence interval. Table 5. Association of Predictive Variables with Mortality in in the ILSA Cohort Mortality for All Causes (N ¼ 345) Mortality for Diseases of Circulatory System (N ¼ 132) Age 3.76 ( ) 3.41 ( ) ADL disability 1.90 ( ) 3.19 ( ) Arrhythmia 1.48 ( ) Angina 0.66 ( ) Myocardial infarction 1.32 ( ) 2.22 ( ) Diabetes 1.93 ( ) 1.30 ( ) Congestive heart failure 2.38 ( ) Stroke 7.39 ( ) DS 1.43 ( ) Note: ILSA ¼ Italian Longitudinal Study on Aging; HR ¼ hazard ratio; CI ¼ confidence interval; ADL ¼ Activities of Daily Living; DS ¼ depressive symptomatology. (education, marital status, and living arrangements), but DS remained an independent predictor. Lack of family and social support among persons with DS has been found to be associated with poor health status, but persons living alone did not show any increased risk for death and CHD events compared with those living with other family members in our cohort. We could not control for physical activity, which is probably lower among persons with DS, so we cannot rule out that the association between DS and CHD is partly mediated through lack of physical activity. However, in a previous community study (37), adjusting for exercise level did not change this association. Psychophysiologic research has indicated some new mechanisms that could explain the association between DS and CHD, such as the higher frequency of modifications of the autonomic nervous system and of the hypothalamicpituitary-adrenal axis among persons with DS. These modifications affect the cardiovascular system through the release of catecholamines and corticosteroids, which might cause immunologic, hemodynamic, and metabolic changes (38 43). We have included as predictors the presence of hypertension, arrhythmia, glycemia, and other metabolic indexes, as well as coagulation factors, but they did not affect the predictive value of DS. In addition, DS could be a marker of other diseases that are associated with the Table 6. Association of Predictive Variables with Mortality in in the ILSA Cohort Mortality for All Causes (N ¼ 523) Mortality for Diseases of Circulatory System (N ¼ 185) Age 2.47 ( ) 4.59 ( ) Claudication 1.64 ( ) Diabetes 1.27 ( ) Congestive heart failure 2.19 ( ) 2.01 ( ) Stroke 2.78 ( ) Fibrinogen 1.71 ( ) DS 2.02 ( ) 2.49 ( ) Note: ILSA ¼ Italian Longitudinal Study on Aging; HR ¼ hazard ratio; CI ¼ confidence interval; DS ¼ depressive symptomatology.

7 DEPRESSION AND CHD 91 development of CHD and mortality, and for this reason we adjusted for all major chronic conditions assessed in the ILSA, but still DS remained an independent predictor. We could not control for the presence of cancer, and this could explain the association with total mortality, given that previous studies have shown an association (44). However, other studies did not support such findings (45). Some studies have reported that the increased risk for CHD among depressed persons results from antidepressive treatment, which could cause arrhythmia and orthostatic hypotension, both precipitating factors for myocardial infarction, and could increase insulin resistance in non insulin-dependent diabetes (46). Other studies (37,47) found no direct association, and our findings support this hypothesis. In the ILSA cohort, only 1% of participants with DS were taking antidepressants, and their exclusion did not change the risk for CHD and mortality among participants with DS. We found a significant association of DS for CHD events among men but not among women, which corresponds with findings of some previous studies (3,6 9). The sex-specific association we found seems to suggest that other factors, beyond the traditional ones, could explain it. A potential explanation for these findings could be that women free of CHD events at baseline were more likely to be widows, less educated, and living alone; have higher prevalence rates of hypertension, congestive heart failure, diabetes, and physical disability; and have higher cholesterol and triglycerides levels compared with men. Therefore, they could have been at higher risk for the development of CHD because of a cluster of traditional risk factors, probably a more powerful predictor than DS alone. Furthermore, the presence of subclinical diseases or other conditions not investigated in the ILSA, such as cancer, could explain the association with total mortality. However, other explanations could be considered (48). and men may psychologically define events differently. A growing literature now indicates that negative emotions might be the pathways through which loss of control could lead to psychophysiologic processes that might finally result in CHD. Among Italian men, more than among women, restricted capacity of acting in the working and social environment is likely to represent a risk factor for DS, for giving up and for disengaging from preventive health and social behaviors. This could lead, finally, to CHD and death. Several limitations of our study deserve comment. First, we used a psychometric scale and we did not have a clinical assessment of the participants depression. However, several studies have shown a high correlation between the score in the scales and the clinical diagnosis (22,49). Second, we had information on the development of CHD events in only 71% of our original cohort at follow-up, which could affect our analyses, if those persons lost to follow-up differ from the rest in terms of incidence of CHD. To address this issue, at baseline we assessed the distribution of the major determinants of CHD, including DS, in the participants lost to follow-up and found no major differences compared with the rest of the cohort. Third, although we measured many traditional risk factors for CHD, the association between DS and CHD events and mortality may have been affected by unmeasured variables, such as physical activity or some chronic conditions that were not included in our assessment. Our study has some unique strengths. It is the first study to analyze the association of DS and CHD and mortality in a large sample of older Italians, with a prospective design and a relatively long follow-up period. Furthermore, we performed a clinical evaluation for the diagnosis of CHD and other major chronic conditions and we did not use selfreported information, which might lead to misclassification of the conditions. Finally, the use of death certificates allowed a more reliable evaluation of the causes of deaths, compared with causes reported by a proxy. As a final point, we controlled for most of the traditional behavioral, demographic, biological, and physical risk factors for CHD and mortality. In conclusion, we found that DS is an independent predictor of total mortality in both men and women and for CHD events in men. These results deserve careful consideration, particularly given the high prevalence rates of DS in both Italian men and women. These rates are, indeed, significantly higher than in other populations. Previous studies, performed mainly in Northern American and Northern European older populations, reported rates that varied between 6% and 11% among men and between 10% and 20% among women, whereas in our sample these rates were about 30% and 40%, respectively (15). Depression is a treatable condition, but very few older persons receive appropriate treatment, which might not only improve their quality of life but also reduce their mortality rates. ACKNOWLEDGMENTS This study was supported by joint research grants from the CNR Targeted Project on Aging from 1991 through 1995 and by two grants from the Italian Ministry of Health (D.L. 502/92, 1998): Epidemiologia dell anziano (ISS-Roma) and Previsione del fabbisogno sanitario dell anziano (Regione Toscana). The Italian Longitudinal Study on Aging Working Group: S. Maggi, N. Minicuci, A. Di Carlo, M. Baldereschi, Italian National Research Council; L. Candelise, E. Scarpini, University of Milan; P. Carbonin, Università Cattolica del Sacro Cuore, Rome; G. Farchi, E. Scafato, S. Brescianini, Istituto Superiore di Sanità, Rome; F. Grigoletto, E. Perissinotto, L. Battistin, M. Bressan, G. Enzi, G. Bortolan, University of Padua; C. Loeb, Italian National Research Council, Genoa; C. Gandolfo, University of Genoa; N. Canal, M. Franceschi, San Raffaele Institute, Milan; A. Ghetti, R. Vergassola, Health Area 10, Florence; D. Inzitari, University of Florence; S. Bonaiuto, F. Fini, A. Vesprini, G. Cruciani, INRCA Fermo; A. Capurso, P. Livrea, V. Lepore, University of Bari; L. Motta, G. Carnazzo, P. Bentivegna, University of Catania; F. Rengo, University of Naples, Naples, Italy. Address correspondence to Stefania Maggi, MD, MPH, CNR Center on Aging, c/o Clinica Medica 1, University of Padua, Via Giustiniani, 2, Padova, Italy. smaggi@unipd.it REFERENCES 1. Stansfeld SA, Fuhrer R, Shipley MJ, et al. Psychological distress as a risk factor for a coronary disease in the Whitehall II Study. Int J Epidemiol. 2002;31: Ariyo AA, Haan M, Tangen CM, et al. Depressive symptoms and risks of coronary heart disease and mortality in elderly Americans. Circulation. 2000;102: Ferketich AK, Schwartzbaum JA, Frid DJ, et al. Depression as an antecedent to heart disease among women and men in the NHANES I Study. Arch Intern Med. 2000;160:1261.

8 92 MARZARI ET AL. 4. des de Leon CF, Krumholz HM, Seeman TS, et al. Depression and risk of coronary heart disease in elderly men and women. Arch Intern Med. 1998;158: Bankier B, Littman AB. Psychiatric disorders and coronary heart disease in women: a still neglected topic: review of literature from 1971 to Psychother Psychosom. 2002;71: Hippisley-Cox J, Fielding K, Pringle M. Depression as a risk factor for ischaemic heart disease in men: population based case-control study. BMJ. 1998;316: Ford DE, Mead LA, Chang PP, et al. Depression is a risk factor for coronary artery disease in men: the precursors study. Arch Intern Med. 1998;158: Penninx BWJH, Geerlings SW, Deeg DJH, et al. Minor and major depression and the risk of death in older persons. Arch Gen Psychiatry. 1999;56: Penninx BWJH, Guaralnik JM, des de Leon CF, et al. Cardiovascular events and mortality in newly and chronically depressed persons.70 years of age. Am J Cardiol. 1998;81: Schultz R, Scott R, Beach D, et al. Association between depression and mortality in older adults. Arch Intern Med. 2000;160: Blazer DG, Hybles CF, Pieper CF. The Association of depression and mortality in elderly persons: a case for multiple, independent pathways. J Gerontol Biol Sci Med Sci. 2001;56A:M505 M Thomas C, Kelman HR, Kennedy GJ, et al. Depressive symptoms and mortality in elderly persons. J Gerontol Biol Sci Med Sci. 1992;47: S80 S Vogt T, Pope C, Mullolly J, et al. tal health status as a predicator of morbidity and mortality: a 15-year follow up of a members of a health maintenance organization. Am J Public Health. 1994;84: Shah A. Can depression and depressive symptoms predict mortality at 18-month follow-up in acutely medically ill inpatients over the age of 80 years? Int J Geriatr Psychiatry. 1998;279: Minicuci N, Maggi S, Pavan M, et al. Prevalence rate and correlates of depressive symptoms in old individuals. The Veneto Study. J Gerontol Biol Sci Med Sci. 2002;57:M55 M Rugulies R. Depression as a predictor for coronary heart disease. Am J Prev Med. 2002;23: Maggi S, Zucchetto M, Grigoletto F, et al. for the ILSA Group. The Italian Longitudinal Study on Aging (ILSA): design and methods. Aging Clin Exp Res. 1994;6: Katz S. Studies of illness in the aged. The Index of ADL. A standardized measure of biological and physiological function. JAMA. 1963;185: Lawton MP, Brody EM. Assessment of older people: self-maintaining and instrumental activities of daily living. J Gerontol Biol Sci Med Sci. 1982;37: Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18: The Italian Longitudinal Study on Aging Working Group. Prevalence of chronic diseases in older Italians: comparing self-reported and clinical diagnoses. Int J Epidemiol. 1997;26: Yesavage JA. Geriatric Depression Scale. Psychopharmacol Bull. 1988;24: World Health Organization. International Classification of Diseases, 9th Revision (ICD-9). Geneva, Switzerland: WHO; SAS/STAT User s Guide, 4th ed., vol. 1 2, version 6. Chicago, IL: SAS Institute; Frasure-Smith N, Lesperance F, Tlajic M. Depression and 18-month prognosis after myocardial infarction. Circulation. 1995;91: Letho S, Koukkunen H, Hintikka J, et al. Depression after coronary heart disease events. Scand Cardiovasc J. 2000;34: Musselman D, Evans D, Nemeroff C. The relationship of depression to cardiovascular disease. Arch Gen Psychiatry. 1998;55: Lespérance F, Frasuer-Smith N. Depression in patients with cardiac disease: a practical review. J Psychosoms Res. 2000;48: Hemingway H, Marmot M. Psychosocial factors in the aetiology and prognosis of coronary heart disease: systematic review of prospective cohort studies. BMJ. 1999;138: Kaufmann MW, Fitzgibbons JP, Sussman EJ, et al. Relation between myocardial infarction, depression, hostility, and death. Am Heart J. 1999;138: Glassman AH, Shapiro PA. Depression and the course of coronary artery disease. Am J Psychiatry. 1998;155: Frasure-Smith N, Lesperance F, Juneau M, et al. Gender, depression, and one year prognosis after myocardial infarction. Psychosom Med. 1999;61: Welin C, Lappas G, Wilhelmsen L. Independent importance of psychosocial factors for prognosis after myocardial infarction. J Intern Med. 2000;247: Carney RM, Saunders RD, Freedland KE, et al. Association of depression with reduced heart rate variability in coronary artery disease. Am J Cardiol. 1995;76: Barefoot JC, Helms MJ, Mark DB, et al. Depression and long-term mortality risk in patients with coronary heart disease. Am J Cardiol. 1996;78: Vaccarino V, Kasl SV, Abramson J, et al. Depressive symptoms and risk of functional decline and death in patients with heart failure. JAm Coll Cardiol. 2001;38: Pratt LA, Ford DE, Crum RM, et al. Depression, psychotropic medication, and risk of myocardial infarction: prospective data from the Baltimore ECA follow-up. Circulation. 1996;94: Jonas BS, Franks P, Ingram DD. Are symptoms of anxiety and depression risk factors for hypertension? Longitudinal evidence from the National Health and Nutrition Examination Survey I epidemiologic follow up study. Arch Fam Med. 1997;6: Schneiderman N, Skyler J. Insulin metabolism, sympathetic nervous system regulation and coronary heart disease. In: Orth-Gomer K, Schneiderman N eds. Behavioral Medicine Approaches to Cardiovascular Disease Prevention. Mahwah NJ: Lawrence Erlbaum, 1996: Lane D, Carrol D, Lip GYH. Psychology in coronary care. Q J Med. 1999;92: Sapolsky RM, Romero LM, Munck AU. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocrine Reviews. 2000;21: Sher L. Effects of psychological factors on the development of cardiovascular pathology: role of the immune system and infection. Med Hypotheses. 1999;53: Carney RM, Freedland KE, Sheline YI, et al. Depression and coronary heart disease: a review for cardiologist. Clin Cardiol. 1997;20: Penninx BW, Guralnik JM, Pahor M, et al. Chronically depressed mood and cancer risk in older persons. J Natl Cancer Inst. 1998;90: Roberts FD, Newcomb PA, Trentham-Dietz A, et al. Self-reported stress and risk of breast cancer. Cancer. 1996;77: Hippsley-Cox J, Pringle M, Hammersley V, et al. Antidepressants as a risk factor for ischaemic heart disease: case-control study in primary care. BMJ. 2001;323: Larson SL, Owens PL, Ford D, et al. Depressive disorder, dysthymia, and risk of stroke. Thirteen-year follow up from the Baltimore Epidemiologic Catchment area study. Stroke. 2001;32: Sewick MA, Rolih C, Pahor M. Gender differences in morbidity and mortality related to depression: a review of the literature. Aging Clin Exper Res. 2000;12: Hautzinger M, Bailer M, Worall H, et al. Beck-depressions-inventar. Bern: Huber; Received June 23, 2003 Accepted July 31, 2003 Decision Editor: John E. Morley, MB, BCh