Background: There is few data on factors that may be related to overall mortality in type 2 diabetes beyond cardiovascular and renal complications. Little is known about cognitive dys-function and depression as potentially life-shortening conditions in diabetes, especially in dif-ferent cultural contexts.
Aim: To identify clinical, laboratory and psychological predictors of the 5-year overall mor-tality in an unselected sample of type 2 diabetic patients in the Moscow region.
Methods: 178 consecutive type 2 patients were recruited into the study (F / M = 145/33, aged 37 to 82 years, with diabetes duration of 0,5 to 30 years). In addition to the standard clinical, instrumental and laboratory assessments of diabetes, its complications and major concomitant disorders, all patients were tested for depression (Hamilton depression scale), cognitive dys-function (locally adapted cognitive battery), brain ischemia (Hachinski ischemic score) and diabetes-dependent quality of life (Ru-ADDQoL). No study-specific intervention was planned except routine clinical management of patients by their primary doctors. After 5 years all patients were contacted by phone and/or mail to assess their vital status (dead or alive). Multiple logistic regression analysis was performed to find predictors of the 5-year overall mortality among baseline measurements, with results presented as odds ratio (OR) and 95% confidence intervals (CI). Given the exploratory nature of the study, results were considered to be statistically significant at alfa <0.1. 
Results: At 5 years, 150 (84%) patients were alive and 15 (8,4%) were dead; no information on the vital status of the rest 13 (7,3%) could be obtained. 165 patients with confirmed out-comes were included into the regression analysis. The factors that were independently associ-ated with the death outcome were as follows: male gender (OR 11.2 [95% CI 2.0-62.0]; p=0.01), age (OR 1.3 [95% CI 1.1; 1.4]; p< 0.0005), cognitive dysfunction score (OR 1.1 [1.0-1.2]; 0.04), HDS score (OR 1,1 [1.0-1.3]; 0.03), and chronic heart failure (CHF) (OR 1.8 [0.9-3.5]; 0.07). Predicted probabilities, obtained from this logistic model (with the use of marginal means), showed that male and female patients had a mean predicted probability to die within the next 5 years of 22,9% and 6,7%, respectively. The highest score of cognitive dysfunction was associated with 25% predicted probability of death, whereas the lowest score with 2%. The corresponding predicted probabilities for the highest and lowest depression scores were 26% and 2%, respectively. 5-year predicted probability of death in patients without CHF was 6,7%, with CHF NYHA class I, 9,8%, class II 13,6%, class III 18,2%, and class IV 23,5%. All other parameters, such as duration of diabetes, HbA1c, lipid levels, insulin therapy, history of severe hypoglycemia or diabetic ketoacidosis, late diabetic complications, arterial hypertension, cardiovascular and cerebrovascular disease, brain is-chemia score and diabetes-dependent quality of life, did not show any independent association with the outcomes. 
Discussion and conclusion: Cognitive dysfunction and depression have a small but signifi-cant negative impact on the 5-year survival. As “predictors” of death in our study sample, they appeared to outact the glucose control and any diabetes-related and cardiovascular com-plications, except CHF. The absence of an independent association between other cardiovas-cular disorders and mortality could be related to the small sample and small number of out-comes, which is a major limitation of our study. The influence of age on survival is self-evident, while the strongest probability of death for male diabetic patients is in line with the known much shorter life expectancy for men in Russia, compared to women. These results underscore the importance of diagnosis and management of depression and/or cognitive dys-function in type 2 diabetic patients. Whether their higher contribution to mortality, compared to the standard parameters of diabetes care, is universal or specific to the country and its healthcare system, requires further and larger studies to verify the model and the candidate predictors.

Factors associated with the 5-year overall mortality in type 2 diabetes in the Moscow region: a prospective cohort study