We compared the response of blood pressure (BP) to either K-Canrenoate (K-Can) or hydrochlorothiazide (HCTZ) in 26 mild-to-moderate essential hypertensives in a double-blind, cross-over design over 2 months each. The dose was 12.5 mg o.d. for HCTZ and 50 mg o.d. for K-Can: dosing was doubled after 1 month if seated diastolic BP was > or = 95 mmHg. Eight pts were "selective responder" to the lowest dose of HCTZ (HCTZ-R), and 6 to K-Can (K-Can-R). Seven pts had their high blood pressure controlled by the highest dose of both drugs and 4 were insensitive to both. One pt dropped out during HCTZ for low plasma K+, and 3 during K-Can (nausea and dizziness: 2 pts; plasma creatinine rise: 1 pt). All these side effects were reverted after drug withdrawal. HCTZ-R and K-Can-R differed for PRA (1.4 +/- 0.6 vs 0.8 +/- 0.4 Ang I ng/ml/h, p < 0.05) and Na-K-Cl cotransport (230 +/- 39 vs 372 +/- 24 mumolNa/L RBC/h, p < 0.01). Our data suggest the existence of a subgroup of essential hypertensives surprisingly insensitive to HCTZ, characterized by a "low" PRA and by a Na(+)-K(+)-Cl- cotransport higher than the HCTZ-R. Their selective response to K-Can suggest a peculiar pathogenetic mechanism underlying their high blood pressure.