The effect of propranolol on the occurrence of hypoglycaemic symptoms was assessed in insulin-dependent diabetic patients with hypoglycaemia unawareness. A double-blind, randomised parallel group study (2:1 fashion) was conducted over 4-week period. The propranolol group (n = 9) received 20 mg (week 1 and 2) and 30 mg (week 3 and 4) twice daily, and the other group (n = 5) a matched placebo for 4 weeks. Patients included had experienced at least two severe hypoglycaemic episodes (coma or seizures) during the previous year, which were characterised by a lack of adrenergic symptoms and required the assistance of another person. The mean number of hypoglycaemias during the study period was similar in both groups (placebo: 13 +/- 2 propranolol: 11 +/- 1), whereas the number of totally asymptomatic hypoglycaemias (< 0.6 g/l) was lower on propranolol than on placebo (3 +/- 1 vs 8 +/- 3, NS) and the number of symptomatic hypoglycaemias was higher (7.2 +/- 2 vs 4.6 +/- 1, NS). Subjective evaluation of treatment by the investigators showed 0/5 successes in the placebo group and 5/9 in the propranolol group (chi2 = 4.32, p = 0.038). The main advantage of propranolol over placebo was an increased incidence of sweating. The ratio [number of hypoglycaemias with sweating/total number of hypoglycaemias] being higher with propranolol (0.28 +/- 0.08 vs 0.06 +/- 0.02, p = 0.06). This pilot study suggests that beta-blockers may be useful in restoring adrenergic symptoms during hypoglycaemia in insulin-dependent diabetic patients without warning symptoms of hypoglycaemia. This beneficial effect seems to be predominantly related to an increase in hypoglycaemia-induced sweating. A larger study is needed to confirm or invalidate these preliminary results.