Hypertension and diabetes mellitus are independently associated with a high rate of target organ complications, which is particularly accentuated in the Negroid race. The aims of this study were to evaluate the mortality associated with diabetes mellitus and concurrent hypertension and diabetes mellitus in indigenous Africans, and to identify and assess the factors that are predictive of intra-hospital mortality in Non-Insulin Dependent Diabetes Mellitus (NIDDM) diabetic Nigerians. The subsequent impact of the modification of these risk factors was also evaluated. A prospective study of 51 hypertensive-diabetic (Non-Insulin Dependent Diabetics, NIDDM) Nigerians (32 males, 19 females) over a 15-month period, from 1996 to 1997 was undertaken. The mean admission blood pressures were 170/102+/-35/22 mm Hg, with a body mass index (BMI) of 25.4+/-10.2 kg/m(2). A total of 54 normotensive (BP<130/85 mm Hg)-NIDDM diabetic Nigerians (30 men, 24 women), who were concurrently admitted in the hospital, were compared to the hypertensive-NIDDM. The total mortality of all the NIDDM diabetics, as well as the mortality rates in normotensives and hypertensive-diabetics, were computed. The causes of death and associated complications were noted. Predictive indices of intrahospital mortality were statistically evaluated by comparison of proportions, chi(2) test, Fischer's exact test, logistic regression, and analysis of variance (ANOVA).Over-all mortality rate among all the diabetics (both normotensive and hypertensive) was 26.6% (28/105), which was significantly higher than the crude death rate on the Internal Medicine service of 17.8% (P=0.006) or the non-obstetric crude death rate in the hospital of 10.96% (P=0.001) Among the hypertensive-NIDDM patients (n=51) the mortality rate was 31.4% (16 deaths/51 patients). This was slightly higher than the value of 22.2% (12 deaths/54 patients) seen in normotensive -NIDDM patients. The mortality rate among the male diabetics (23/63 patients) 46.6% was significantly higher than female mortality rate of 11.6% (5/43). The 95% Confidence interval for the difference in mortality rates being 16.9% to 53.3% (P<0.0001, z=3.57). The impact of gender remained significant by the chi(2) test, chi(2)=7.17, P=0.007. 50% of the deaths in hypertensive-diabetics had associated stroke (8/16), while none of the 12 deaths among the normotensive-diabetics was stroke-related (P=0.008, Fisher's exact test). The case fatality rate for stroke in hypertensive-NIDDM men (7/7.9) was significantly greater than in hypertensive-NIDDM women (1.0/7) (P=0.04,by Fisher's exact test). Male gender, presence of Hyperosmolar Non-Ketotic Coma (HONK) (P<0. 05), associated stroke (P<0.01) and a Glasgow coma < or =10 (P<0.01) were found to be poor prognostic indices for mortality in hypertensive-NIDDM. Aggressive anti-platelet, (aspirin) anti-hypertensive, and strict glycemic control, instituted early and intensively, especially in male hypertensive-NIDDM Nigerians have resulted in reduction in the mortality rates from 26.6% in 1997 to 12.6% in 1999 [P=0.05, 95% CI -26.9% to -1.3%]. The prognosis in 1999 of hospitalized African diabetics is still dismal. Hypertensive-NIDDM represents a higher risk group for intra-hospital mortality in black Africans. Male patients appear to have significantly (P<0.001) enhanced risk, especially with thrombotic/stroke-related deaths (P<0.01). However, intervention measures can reduce the mortality rate considerably, even in developing countries. The mechanisms of the apparent male mortality excess require elucidation.