The present multicentre study, conducted in patients with angina pectoris on chronic therapy with continuous 24-hour transdermal nitroglycerin was aimed at investigating: (1) the actual incidence of nitrate tolerance; (2) the clinical features that characterize those who do not develop tolerance; (3) whether the evidence of clinical and ergometer benefits at 1-month assessment predicts their long-term maintenance. Eligible patients (n = 110; average age 56 +/- 5 yrs) had stable angina pectoris (78 effort angina, 32 mixed angina) with symptom-limited, reproducible cycloergometer tests and were responsive to s.l. nitroglycerin. After a 7-day washout period and a placebo run-in week, all patients were assigned to continuous 10 mg/24-hour nitroglycerin patches (T0). Exercise tests were performed again after 1 month (T1) and 6 months (T6) of treatment. At T1, ergometric data in each patient were compared with those at T0 and showed an increase (> or = 15%) in exercise tolerance in only 61/110 patients: 38 (49%; Group A) of the 78 patients with effort angina and 23 (72%; Group B) of the 32 patients with mixed angina (p < 0.05). Those patients with no significant change in exercise tolerance were assigned to conventional antianginal therapy and were excluded from the study. At T6, both group A and B patients maintained the favourable effects on total exercise duration, time and maximum workload at ischemic threshold and maximal ST depression recorded at T1 vs T0. The weekly attacks of angina and nitroglycerin s.I. consumption decreased significantly in both Group A and Group B patients from the beginning of therapy and throughout the study. These results show that: 1) tolerance to continuous 24-hour transdermal nitroglycerin therapy is not constant phenomenon, but occurs only in a subgroup of patients; 2) patients with mixed angina are more resistant to develop tolerance than those patients with effort angina; 3) the 1-month evidence of clinical and ergometric benefits predicts their maintenance during long-term treatment, as well. These results allow one to hypothesize that the loss of nitrate efficacy on venous pooling capability, but not on arterial tone, might constitute the more frequent cause of tolerance.