Background Antihypertensive medication use in Canada and the US has varied during the 1990s while the incidence rates of ESRD have increased in both countries. We analyzed variations in use of the four predominant antihypertensive categories retrospectively to determine if ESRD variation was associated with variations in antihypertensive medication use. Methods Data fusion technique assessing market share of antihypertensive medications and annual changes in ESRD incidence was used. The predictor variables were market share percentage of total national cardiovascular expenditure for four medication groups (diuretics (D), beta-blockers (B), calcium channel blockers (C) or ACE-inhibitors (A)) in each country (data from IMS Health) for the years 1990 to 2001 inclusive. Data were time-lagged by 4 years, consistent with our previous analysis of these databases. Multiple regression analysis was done using MODSTAT (all rights reserved). Coefficients of correlation were calculated to show relationships existing between predictor variables and the dependent variable. Coefficients for each predictor variable are also calculated. Drugs associated with the highest predictor coefficient carry the highest association with increased ESRD, and those with the lowest predictor coefficient carry the lowest association. Drugs with a negative coefficient are named protective, since increased use would be associated with lower ESRD incidence. Results The Canadian database showed a coefficient of 0.95568, indicating that 91.33 percent of the variation in ESRD incidence in Canada from 1990–1996 could be explained by drug variations (p=0.037). From highest to lowest, the drugs had the following coefficients of prediction: diuretics (26.6), CCBs (3.3), ACE-Is (0.9), BB (-12.8). The US database showed a coefficient 0.91856, indicating that 84.37 percent of the variation in ESRD incidence in the US from 1990–1997 could be explained by drug variations (p=0.05). From highest to lowest, the drugs had the following coefficients of prediction: diuretics (6.7), CCBs (0.99), ACE Inhibitors (-1.97), Beta-Blockers (-2.9). Conclusions Retrospective analysis of antihypertensive medication use shows that variability of ESRD incidence can be predicted knowing the market share of the major drug categories four years earlier. In both analyses, diuretics were identified with the highest predictive coefficient associated with ESRD, CCBs were essentially neutral, ACE inhibitors (US) and beta-blockers (both) were associated with a negative predictive coefficient, implying renoprotection. In both databases the order of coefficients for the drugs were DCAB. In view of recent JNC 7 and ALLHAT recommendations, the use of diuretics may need re-evaluation.