QJM Advance Access originally published online on May 7, 2008
QJM 2008 101(7):575-582; doi:10.1093/qjmed/hcn056
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The impact of comorbidity burden on the cardiovascular risk in the Peripheral Arteriopathy and Cardiovascular Events study
From the Department of Clinical Medicine and Cardiovascular and Immunological Sciences, University of Naples ‘Federico II’, Naples, Italy
Address correspondence to Gregorio Brevetti, MD, via G. Iannelli 45/A 80131 Napoli, Italy. email: brevetti{at}unina.it
Received 23 May 2007 and in revised form 16 November 2007
 |
Summary |
|---|
 Top Summary IntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
Background: A comprehensive evaluation of comorbidity is important in predicting outcome of patients affected by a chronic disease because of the role of competing risk.
Aim: To assess the prognostic impact of the Cumulative Illness Rating Scale (CIRS) on the cardiovascular risk of subjects participating in the Peripheral Arteriopathy and Cardiovascular Events (PACE) study.
Design: Prospective study.
Methods: The study included 60 patients with peripheral arterial disease (PAD) and 163 no-PAD subjects. CIRS-illness severity (IS) score and CIRS-comorbidity index (CI) were calculated.
Results: After a 42-month follow-up, 18/223 participants had a myocardial infarction or stroke. These subjects had a higher CIRS-IS score (1.99 ± 0.52 vs. 1.71 ± 0.37, P = 0.003) and a higher CIRS-CI (4.00 ± 2.81 vs. 2.65 ± 1.85, P = 0.005) vs. the 205 subjects without event. However, the significant association of CIRS scores with the outcome disappeared when conditions considered to be ‘concordant’ with the endpoint were excluded from the calculation of the scores. Importantly, among the 163 no-PAD subjects CIRS scores did not differ between those with and without an event. Conversely, in the 60 PAD patients, the CIRS-IS score calculated excluding the ‘concordant’ conditions was associated with an increased cardiovascular risk (RR = 4.03, 95% confidence interval (CI) 1.05–15.37, P = 0.042) after adjustment for potential confounders. The corresponding RR for the CIRS-CI was 1.43 (95% CI 1.03–1.98, P = 0.032). Furthermore, both CIRS scores improved the predictive value of ankle/brachial index, which is the most powerful prognostic indicator in PAD.
Conclusions: Our findings indicate that overall comorbidity, and not only cardiovascular comorbidity, must be considered for prediction of myocardial infarction and stroke in PAD.
|
Introduction |
|---|
|
TopSummaryIntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
As people age, they are more likely to develop chronic medical conditions,1 which can alter both the efficacy of therapies and the course of the primary disease. Indeed, comorbidity has been shown to be associated with age, disability, health service use and survival.2–5 Therefore, an objective and comprehensive evaluation of comorbid conditions is a major issue in the management of elderly people and for prediction of outcome. This is particularly true in chronic diseases in which long-term survival increases the chance that competing risk will confound predictions.
Peripheral arterial disease (PAD) is a major manifestation of atherosclerosis and thus is often associated with vascular lesions in other arterial districts. Therefore, previous studies concerning the natural history of PAD have focused on comorbidities such as coronary or cerebrovascular diseases.6–8 Consequently, the impact on outcome of other comorbid conditions associated with PAD, namely pulmonary, renal and endocrine/metabolic disease,9–11 may be underestimated. There are a number of comorbidity indices that identify and summarize comorbid burden. The Cumulative Illness Rating Scale (CIRS) is one of the few standardized instruments for the rating of medical problems by organ system,12,13 and is able to predict outcome in a variety of conditions.12,14–18
In the Peripheral Arteriopathy and Cardiovascular Events (PACE) study, the first survey in Italy to assess the prevalence and the natural history of symptomatic PAD in the general population,19,20 we used the CIRS to gauge comorbidity. In the present study we examined the follow-up data of 223 subjects (60 were affected by symptomatic PAD) for risk of myocardial infarction and stroke. We hypothesized that the greater the comorbidity burden, the higher the cardiovascular risk even when conditions considered to be ‘concordant’ with the endpoint were excluded from the calculation of the CIRS scores.
|
Materials and methods |
|---|
|
TopSummaryIntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
Participant identification
The details of the PACE study, conducted in five villages of a well-defined area of the Campania region in southern Italy, are reported elsewhere.19 Briefly, from the lists of seven general practitioners (GPs), we identified all subjects aged 40–80 years (n = 4352). GPs excluded 200 patients for the reasons reported in Figure 1. The Rose questionnaire, which identifies leg symptoms that occur during walking, was mailed to each subject with an explanatory letter. Of these, 287 did not answer. Of the remaining 3865 subjects, patients reporting pain in the calf that began while walking and did not disappear while continuing to walk, and those with other complaints of the calf, foot, thigh or buttock (not of the knee) showing the ischemic pattern described previously, regardless of whether the remaining Rose criteria for claudication were met (n = 760), were diagnosed with ‘possible’ PAD. To confirm the PAD diagnosis, ‘possible’ symptomatic PAD cases underwent Doppler examination with assessment of the ankle/brachial index (ABI) and flow velocity in the femoral and posterior tibial arteries, which were measured by trained physicians as previously described.19
|
Of the 760 subjects positive to the Rose questionnaire, 12 died before the vascular examination, 65 moved and 89 did not keep their appointments. Age, sex, risk factors and prevalence of concomitant cardiovascular disease did not differ between these 166 individuals and those who underwent the vascular examination. Symptomatic PAD, defined by an ABI <0.90 or reduced flow velocity in at least one leg,21 was present in 1.6% (60/3699 subjects) of the whole population.19 For each patient affected by PAD, three subjects negative to the Rose questionnaire and matched for sex and age (±2 years) were selected randomly from the alphabetical listing of each GP. These 180 no-PAD subjects underwent the same investigations as the PAD patients. Baseline characteristics were obtained from the subjects’ electronic medical records. The software system, used by approximately 8000 Italian GPs, encodes all diagnostics records according to the ninth edition of the International Classification of Disease (ICD-9), and prescription records according to the Anatomical Therapeutical Chemical Classification system. Hypertension was diagnosed if systolic arterial pressure exceeded 140 mmHg or diastolic arterial pressure exceeded 90 mmHg, or if the patient used antihypertensive drugs. Hypercholesterolemia was diagnosed in case of plasma total cholesterol >200 mg/dl, LDL cholesterol >130 mg/dl, HDL cholesterol <35 mg/dl, or if the patient used lipid lowering drugs. Diabetes mellitus was diagnosed if plasma fasting glucose exceeded 126 mg/dl, or if the patient used hypoglycemic drugs. Smokers were defined ‘former’ and ‘current’ smokers.
The GPs completed the CIRS. This rating scale consists of 14 items covering: the heart, hypertension, vascular and respiratory disorders, a combined eye-ear nose-throat item, the upper and lower gastrointestinal system, the hepatobiliary system, the kidney, genitourinary diseases, musculoskeletal diseases, endocrine/metabolic disorders, the neurological system, and behavioral-psychiatric disorders. Each single item was rated by the GP based upon the clinical data available according to the following algorithm: 1 = no, 2 = mild, 3 = moderate, 4 = severe, 5 = life-threatening.9 After completion of the CIRS, two summary measures were constructed. First, the overall illness severity (IS) was represented by the mean of the 14 CIRS items (CIRS-IS). A comorbidity index (CIRS-CI) was computed by counting the number of items for which moderate to severe pathology was reported (scores 3, 4 or 5). Therefore, the CIRS-CI can also be considered the number of clinically relevant concomitant diseases. To verify the prognostic impact of non-vascular comorbidities, the two indices were also calculated excluding the five CIRS items which may be considered ‘concordant’ with the endpoint of the study, which was myocardial infarction and stroke. Accordingly we analyzed the results also excluding from the calculation of the CIRS scores the following items: heart, hypertension, vascular, endocrine/metabolic and neurological. These conditions were considered ‘concordant’ because they represent parts of the same overall pathophysiologic risk profile and are more likely to be the focus of the same disease management plan.22
All participants gave their written informed consent to the study, which was approved by our institutional ethics committee.
Prospective follow-up
GPs were contacted to determine the occurrence of events at 6-month intervals. The occurrence of fatal and non-fatal myocardial infarction and stroke was prospectively assessed. Information about these events was extracted from the patients’ electronic medical records which are updated by the GP on the basis of hospital records and death certificates. Two cardiologists at our department performed a blinded review of the events to verify agreement between the diagnosis obtained from the hospital record, the death certificate and the ICD code. For patients experiencing more than one event, only the first was considered in the analysis.
Statistical analysis
Continuous variables were normally distributed and were expressed as mean ± SD. Comparisons of baseline characteristics between subjects with and without cardiovascular events at follow-up were made by t-test for unpaired samples or 
2 test, as appropriate. Associations of follow-up events with CIRS scores, used as continuous variables, were assessed by crude event rates and Cox proportional hazard models adjusted for the variables that met the entry criteria at the univariate analysis (P < 0.1).23 The comparison of the predictive values of ABI and CIRS was done by calculating the z-ratio for the significance of the difference between two independent proportions.
|
Results |
|---|
|
TopSummaryIntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
Figure 1 shows the flow-diagram of the study and the reasons why some subjects were excluded. The baseline characteristics of the study population are detailed elsewhere.19 Follow-up data were obtained for all PAD patients. Conversely, 13 no-PAD subjects could not be contacted after the baseline examination. Compared with the other no-PAD subjects, these individuals did not differ appreciably for age, sex, risk factors and comorbidity.
During a median follow-up period of 42 months, of the 223 participants, 8 (3.6%) had a myocardial infarction (5 were fatal) and 10 (4.5%) had a stroke (7 were fatal). Table 1 shows baseline characteristics of subjects with and without cardiovascular events at follow-up. Subjects who had a myocardial infarction or stroke at follow-up had a higher prevalence of PAD, diabetes mellitus and previous stroke than subjects without events. Furthermore, the 18 patients who experienced an event had higher CIRS-IS and CIRS-CI scores than the remaining 205 subjects.
|
At univariate analysis, CIRS-IS was associated with an increased risk of developing an event [crude relative risk (cRR) 5.45, 95% confidence interval (CI) 1.82–16.34, P = 0.002]. A similar result was observed for CIRS-CI (cRR 1.31, 95% CI 1.09–1.58, P = 0.004). However, these significant associations disappeared after the exclusion of ‘concordant’ conditions from the calculation of the comorbidity scores.
Importantly, while in the no-PAD group the comorbidity scores were similar in subjects with and without events at follow-up, in the PAD group, patients who experienced an event showed higher scores for both CIRS-IS and CIRS-CI (Figure 2). In no-PAD subjects CIRS-IS was 1.67 ± 0.36 in those without and 1.66 ± 0.33 (P = 0.918) in those with an event (n = 6). The corresponding CIRS-CI were 2.45 ± 1.81 vs. 2.00 ± 1.10 (P = 0.552). Conversely, in the PAD group, compared to patients without event, those with an event at follow-up (n = 12) had higher CIRS-IS (1.82 ± 0.36 vs. 2.16 ± 0.53, P = 0.011) and CIRS-CI (3.23 ± 1.85 vs. 5.00 ± 2.89, P = 0.011). Noteworthy, these latter findings remained substantially unchanged when the ‘concordant’ conditions were excluded from the composite scores. Actually, CIRS-IS was 1.61 ± 0.39 in PAD patients without and 1.91 ± 0.57 (P = 0.038) in those with an event. The corresponding values for CIRS-CI were 1.31 ± 1.31 vs. 2.25 ± 2.00 (P = 0.052). Therefore, we calculated the predictive role of CIRS in PAD patients (mean age 69.0 ± 8.4 years, males 71.6%). Compared to patients without events, those with an event at follow-up had a lower ABI (0.66 ± 0.2 vs. 0.53 ± 0.2, P = 0.035) and tended to have a greater prevalence of diabetes mellitus. Conversely, no group differences were observed in other classic cardiovascular risk factors and previous myocardial infarction and stroke. The use of cardiovascular medication was similar in the two groups. As illustrated in Table 2, at univariate analysis, both CIRS scores were associated with a higher risk of developing an event also when ‘concordant’ conditions were excluded. This latter finding did not change at the multivariate analysis including ABI and diabetes mellitus, which were the only variables that met the entry criteria (P < 0.1).
|
|
Because ABI is currently the most powerful predictor of cardiovascular risk in PAD,24 we investigated whether CIRS scores add to its predictive value. The incidence of ischemic events in the four subgroups of patients stratified according to the median value of ABI and the median values of CIRS-IS and CIRS-CI is shown in Figures 3 and 4, respectively. At the Wilcoxon test, group differences were statistically significant (P = 0.017 for CIRS-IS and P = 0.022 for CIRS-CI), and the highest incidence of events occurred in patients with both ABI < median and CIRS scores > median.
|
|
For more practical purposes, we grouped the CIRS-CI scores according to Boulos et al.15 to create groupings that might have distinct risks, and then calculated the relative risk estimates. After correction for ABI and diabetes, PAD patients with a CIRS-CI
7 had a higher risk of a cardiovascular event than those with a CIRS-CI between 0 and 4 (RR = 9.45, 95% CI 1.93–46.23, P = 0.006). Moreover, they also had a higher risk than all subjects with a score <7 (RR = 7.77, 95% CI 1.95–30.90, P = 0.004). Patients with a score of 5 and 6 had a 2.05-fold increased risk (95% CI 0.45–9.30, P = 0.351) compared with those that had a lower score. Notably, a CIRS-CI 7 had a positive predictive value of 66.7%, higher than that of ABI < median, which was 37.9%, although the difference was not statistically significant (P = 0.195). The corresponding negative predictive values were 85.2% and 96.8% (P = 0.095). |
Discussion |
|---|
|
TopSummaryIntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
Comorbidity measured by CIRS has been shown to be an independent predictor of mortality in older people.12,14,15,17,18 In the present study, which included 223 subjects with a mean age of 67 years, we assessed the impact of CIRS on the risk of future myocardial infarction and stroke, which are major causes of death in the elderly.25 The univariate analysis showed that in the entire study population, both CIRS scores were higher in subjects with than in those without an event at follow-up. However, these significant associations disappeared after the exclusion of conditions considered to be ‘concordant’ with the endpoint. In the 163 subjects without PAD, CIRS-IS and CIRS-CI were similar in those with and without event, thus showing no impact on outcome. This is in contrast with previous studies which, however, had different endpoints and included very elderly populations.12,14–17
Contrary to the results we obtained in no-PAD subjects, comorbidity was independently associated with an increased risk of future ischemic events in PAD patients. PAD is strictly associated with aging,26,27 which entails an increased risk of developing other chronic medical conditions.1 Furthermore, the limited physical activity due to intermittent claudication may promote development of untoward conditions, such as insulin resistance, metabolic syndrome and depression, which are associated with exercise restriction,28–30 and portend a high cardiovascular risk.31–33 Thus, not surprisingly, PAD patients suffer from an increased comorbidity burden, as shown in previous reports9,10 and in the present one (Figure 2). Thus far, however, comorbidity analysis of these patients has focused on a limited set of diseases, namely coronary artery and cerebrovascular disease.6–8 This is particularly unfortunate, because other comorbid conditions, by virtue of their competing risk effect, can confound our ability to evaluate the natural history of the disease. In our series, both CIRS-IS and CIRS-CI were significant predictors for incident myocardial infarction and stroke, independently from other factors that may influence the cardiovascular risk. Actually, the incidence of cardiovascular events was independent of traditional risk factors and history of previous myocardial infarction or stroke, the prevalence of which was similar in PAD patients with and without events at follow-up. Furthermore, the significant associations between CIRS scores and cardiovascular risk were maintained also when the comorbidities ‘concordant’ with the endpoint were excluded from the calculation of CIRS scores. This was likely due to the fact that in the PAD subgroup the rate of cardiovascular risk factors and history of myocardial infarction and stroke was similar among patients who at follow-up had myocardial infarction or stroke and those who had not. This finding emphasizes the importance that the burden of other comorbid conditions has in predicting the cardiovascular risk of affected individuals. In addition to being unrelated to risk factors and previous cardiovascular disease, the predictive value of CIRS scores was independent of ABI, which is the most powerful prognostic indicator in PAD.24 Even more important, survival curves showed that CIRS scores above the median were associated with a worse prognosis both in patients with ABI > median and in those with ABI < median. This indicates that combined measurement of CIRS scores and ABI provides a better stratification of PAD patients than ABI alone.
|
Conclusion and implications |
|---|
|
TopSummaryIntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
This study reports that cumulative comorbidity burden measured by CIRS has little effect, if any, on the incidence of myocardial infarction and stroke in a geriatric, but not very old no-PAD population. Conversely, CIRS has a profound prognostic impact on the cardiovascular risk of PAD patients. Even more important, both CIRS scores improved the prognostic value of ABI. These findings have important implications. First, they suggest that overall comorbidity and not only cardiovascular comorbidity must be taken into account in PAD clinical trials. Second, to draw correct conclusions about cardiovascular risk and the process of care, the physician should take account of prognostically important differences in patient's characteristics. Finally, our data may have implications for the counseling of individual patients. They indicate that physicians should investigate the entire spectrum of comorbid illness of their PAD patients and suggest that a CIRS-CI
7 may be used to identify those who need more aggressive therapy. This could be especially important for GPs, who rarely measure the ABI (from 0% to 25% of cases) when pedal pulses are absent,34 because this procedure requires a specific, although not expensive, device, training of observers, and is time-consuming. Conversely, a family physician can complete the CIRS within few minutes. Conflict of interest: None declared.
|
References |
|---|
|
TopSummaryIntroductionMaterials and methodsResultsDiscussionConclusion and implicationsReferences |
|---|
1. van den Akker M, Buntinx F, Metsemakers JF, Roos S, Knottnerus JA. Multimorbidity in general practice: prevalence, incidence, and determinants of co-occurring chronic and recurrent diseases. J Clin Epidemiol (1998) 51:367–75.[CrossRef][Web of Science][Medline]
2. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis (1987) 40:373–83.[CrossRef][Web of Science][Medline]
3. Albertsen PC, Fryback DG, Storer BE, Kolon TF, Fine J. The impact of co-morbidity on life expectancy among men with localized prostate cancer. JAMA (1996) 156:127–32.
4. Dodds TA, Martin DP, Stolov WC, Deyo RA. A validation of the functional independent measurement and its performance among rehabilitation inpatients. Arch Phys Med Rehabil (1993) 74:531–36.[CrossRef][Web of Science][Medline]
5. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol (1992) 45:613–19.[CrossRef][Web of Science][Medline]
6. Meijer WT, Hoes AW, Rutgers D, Bots ML, Hofman A, Grobbee DE. Peripheral arterial disease in the elderly: the Rotterdam Study. Arterioscler Thromb Vasc Biol (1998) 18:185–92.
7. Fowkes FG, Housley E, Cawood EH, Macintyre CC, Ruckley CV, Prescott RJ. Edinburgh Artery Study: prevalence of asymptomatic and symptomatic peripheral arterial disease in the general population. Int J Epidemiol (1991) 20:384–92.
8. Stoffers HE, Rinkens PE, Kester AD, Kaiser V, Knottnerus JA. The prevalence of asymptomatic and unrecognized peripheral arterial occlusive disease. Int J Epidemiol (1996) 25:282–90.
9. Kugler C, Rudofsky G. The role of comorbidity burden for patients with symptomatic peripheral arterial disease (PAD). A quantitative approach. Int Angiol (2003) 22:290–301.[Web of Science][Medline]
10. Heidrich H. Frequency of non-vascular accompanying diseases in patients with peripheral arterial disease. Vasa (2004) 33:155–8.[CrossRef][Web of Science][Medline]
11. McDermott MM, Guralnik JM, Ferrucci L, Criqui MH, Greenland P, Tian L, et al. Functional decline in lower-extremity peripheral arterial disease: associations with comorbidity, gender, and race. J Vasc Surg (2005) 42:1131–7.[CrossRef][Web of Science][Medline]
12. Parmelee PA, Thuras PD, Katz IR, Lawton MP. Validation of the Cumulative Illness Rating Scale in a geriatric residential population. J Am Geriatr Soc (1995) 43:130–7.[Web of Science][Medline]
13. Conwell Y, Forbes NT, Cox C, Caine ED. Validation of a measure of physical illness burden at autopsy: the Cumulative Illness Rating Scale. J Am Geriatr Soc (1993) 41:38–41.[Web of Science][Medline]
14. Nagaratnam N, Gayagay G Jr. Validation of the Cumulative Illness Rating Scale (CIRS) in hospitalized nonagenarians. Arch Gerontol Geriatr (2007) 44:29–36.[CrossRef][Web of Science][Medline]
15. Boulos DL, Groome PA, Brundage MD, Siemens DR, Mackillop WJ, Heaton JP, et al. Predictive validity of five comorbidity indices in prostate carcinoma patients treated with curative intent. Cancer (2006) 106:1804–14.[CrossRef][Web of Science][Medline]
16. Giaquinto S, Palma E, Maiolo I, Piro MT, Roncacci S, Sciarra A, et al. Importance and evaluation of comorbidity in rehabilitation. Disabil Rehabil (2001) 23:296–9.[CrossRef][Web of Science][Medline]
17. Bo M, Cacello E, Chiggia F, Corsinovi L, Bosco F. Predictive factors of clinical outcome in older surgical patients. Arch Gerontol Geriatr (2007) 44:215–24.[CrossRef][Web of Science][Medline]
18. Extermann M. Measuring comorbidity in older cancer patients. Eur J Cancer (2000) 36:453–71.[CrossRef][Web of Science][Medline]
19. Brevetti G, Oliva G, Silvestro A, Scopacasa F, Chiariello M. Peripheral Arteriopathy and Cardiovascular Events (PACE) Study Group. Prevalence, risk factors and cardiovascular comorbidity of symptomatic peripheral arterial disease in Italy. Atherosclerosis (2004) 175:131–8.[CrossRef][Web of Science][Medline]
20. Brevetti G, Schiano V, Verdoliva S, Silvestro A, Sirico G, De Maio J, et al. Peripheral arterial disease and cardiovascular risk in Italy. Results of the Peripheral Arteriopathy and Cardiovascular Events (PACE) study. J Cardiovasc Med (2006) 7:608–13.[Web of Science]
21. Fronek A, Coel M, Berstein EF. Quantitative ultrasonographic studies of lower extremity flow velocities in health and disease. Circulation (1976) 53:957–60.
22. Piette JD, Kerr EA. The impact of comorbid chronic conditions on diabetes care. Diabetes Care (2006) 29:725–31.
23. Cox DR. Regression and life tables. J R Stat Soc (1972) 34:187–220.
24. Newman AB, Shemanski L, Manolio TA, Cushman M, Mittelmark M, Polak JF, et al. Ankle-arm index as a predictor of cardiovascular disease and mortality in the Cardiovascular Health Study. The Cardiovascular Health Study Group. Arterioscler Thromb Vasc Biol (1999) 19:538–45.
25. Aronow WS. Clinical causes of death of 2372 older persons in a nursing home during 15-year follow-up. J Am Med Dir Assoc (2000) 1:95–6.[Medline]
26. Kannel WB. The demographics of claudication and the aging of the American population. Vasc Med (1996) 1:60–4.[Medline]
27. Kannel WB, Skinner JJ Jr, Schwartz MJ, Shurtleff D. Intermittent claudication. Incidence in the Framingham Study. Circulation (1970) 41:875–83.
28. Pan XR, Li GW, Hu YH, Wang JX, Yang WY, An ZX, et al. Effects of diet and exercise in preventing NIDDM in people with impaired glucose tolerance. The Da Qing IGT and Diabetes Study. Diabetes Care (1997) 20:537–44.[Abstract]
29. Park YW, Zhu S, Palaniappan L, Heshka S, Carnethon MR, Heymsfield SB. The metabolic syndrome: prevalence and associated risk factor findings in the US population from the Third National Health and Nutrition Examination Survey, 1988-1994. Arch Intern Med (2003) 163:427–36.
30. Blumenthal JA, Babyak MA, Moore KA, Craighead WE, Herman S, Khatri P, et al. Effects of exercise training on older patients with major depression. Arch Intern Med (1999) 159:2349–56.
31. Bonora E, Kiechl S, Willeit J, Oberhollenzer F, Egger G, Meigs JB, et al. Insulin resistance as estimated by homeostasis model assessment predicts incident symptomatic cardiovascular disease in caucasian subjects from the general population: the Bruneck study. Diabetes Care (2007) 30:318–24.
32. Onat A, Ceyhan K, Basar O, Erer B, Toprak S, Sansoy V. Metabolic syndrome: major impact on coronary risk in a population with low cholesterol levels–a prospective and cross-sectional evaluation. Atherosclerosis (2002) 165:285–92.[CrossRef][Web of Science][Medline]
33. Barefoot JC, Schroll M. Symptoms of depression, acute myocardial infarction, and total mortality in a community sample. Circulation (1996) 93:1976–80.
34. McLafferty RB, Dunnington GL, Mattos MA, Markwell SJ, Ramsey DE, Henretta JP, et al. Factors affecting the diagnosis of peripheral vascular disease before vascular surgery referral. J Vasc Surg (2000) 31:870–9.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||