Increase in mortality associated with the presence of diabetes mellitus in Japanese-Brazilians

• Published in: Revista de saúde pública Vol. 32; no. 2; pp. 118 - 124
• Main Authors: Gimeno, S G, Ferreira, S R, Franco, L J, Iunes, M, Osiro, K
• Format: Journal Article
• Language: Portuguese; Spanish
• Published: Brazil 01-04-1998
• Subjects: Adult; Aged; Brazil - epidemiology; Creatinine - blood; Diabetes Mellitus, Type 2 - mortality; Female; Glucose Tolerance Test; Humans; Hyperlipidemias; Hypertension; Japan - ethnology; Male; Middle Aged; Obesity; Brazil; Japan
• ISSN: 0034-8910
• DOI: 10.1590/S0034-89101998000200003

As part of a study involving Japanese migrants, living in a developed city in the state of S. Paulo, Southeastern Brazil, a four-year experience of mortality among diabetic and non-diabetic subjects is described and their respective death rates are compared. In 1993, a cohort of 530 Japanese-Brazilians (236 issei or 1st generation and 294 nisei or 2nd generation) of both sexes, aged 40 from to 79 years old, were identified.
At that time, 91 (17%) were classified as non-insulin-dependent diabetic subjects (NIDDM), 90 (17%) with impaired glucose tolerance (IGT) and 349 (66%) as normal, according to WHO criteria. In 1996, families were questioned with a view detecting the deaths which had occurred among the subjects previously studied. This information, in addition to that from death certificates was used to record the date and the causes of death. Mortality rates for all causes and for specific causes (circulatory and renal diseases) were obtained for the three groups of subjects, by glucose tolerance status. Proportional hazard regression models were used to compare the mortality rates, adjusted for several covariables (gender, age, generation, hypertension, dyslipidemia, obesity and serum creatinine).
Crude mortality rate ratios for all causes and specific causes, for NIDDM, and normal subjects were 2.95 (95% CI: 1.10-7.62) and 4.57 (95% CI: 1.31-16.48), respectively. No difference was observed between the crude mortality rate ratio for IGT and normal subjects. After simultaneous adjustments for the covariates, higher mortality rates for specific causes were observed among NIDDM than in the normal subjects (mortality rates ratio: 3.86; 95% CI: 1.11-13.38). These results in Japanese-Brazilians are consistent with previous reports of increased mortality in other diabetic subjects, thus confirming the adverse effect of this metabolic disturbance on mortality among diabetic subjects.