Main Article Content
Methods: Participants were 420 consecutive individuals (49% men) with type 2 diabetes receiving chronic care in two main referral centres in the two major cities of Cameroon. Logistic regression models were used to compute the odd ratio (OR) and 95% confidence interval (95% CI) for a standard deviation (SD) higher level of SBP (25 mmHg), DBP (13), PP (18) and MAP (16) with nephropathy risk. Discrimination was assessed and compared with c-statistics and relative integrated discrimination improvement (RIDI, %).
Results: The adjusted OR (95% CI) for nephropathy with each SD higher BP variable were: 1.45 (1.15-1.84) for SBP, 1.33 (1.06-1.66) for DBP, 1.35 (1.06-1.71) for PP and 1.42 (1.13-1.78) for MAP. C-statistic comparison showed no difference in discrimination of models with each of the BP variables (p-values B 0.69 for c-statistics comparison). However, RIDI statistic always showed and enhancement in models discrimination when other BP variables were replaced with SBP, although such an enhancement was marginal for MAP. Using BP combination modestly improved models? discrimination.
Conclusion: SBP was the best predictor of prevalent nephropathy in this population, while DBP was the less effective. This may have implication for kidney disease risk stratification and protection.