In order to compare the beta blockers bisoprolol and diazepam in the treatment of cardiac neurosis, 40 patients (16 males and 24 females, mean age: 39 +/- 11 years) were examined in a double-blind, crossover study. Following a 4-week placebo period, patients were randomized to receive either bisoprolol 10 mg daily or diazepam 5 mg twice daily for 4 weeks. After a second 4-week washout period on placebo, patients were switched to the alternative regimen for a further 4 weeks. At the end of the placebo periods and during each phase of treatment, the following parameters were evaluated: somatic symptoms by self-assessment questionnaire, anxiety state by Hamilton rating scale, reaction time to both acoustic and visual stimuli, blood pressure, and heart rate. Both treatments were effective in reducing somatic symptoms of cardiac neurosis, but bisoprolol was significantly more effective than diazepam (p less than 0.01). On the contrary, diazepam was superior to bisoprolol in improving the Hamilton scale related to psychic symptoms. Only diazepam prolonged reaction times. Both treatments were well tolerated; however, 12 patients complained of drowsiness and nine of sedation under diazepam. In conclusion, bisoprolol appeared to be as effective as diazepam in the treatment of cardiac neurosis, but with better effects on somatic symptoms and without affecting patients' psychomotor performance.