In schizophrenics cardiovascular autonomic reactivity (CAR) can be used as an indicator of autonomic arousal. Using a standardized autonomic test battery (modified according to Ewing and Clarke) we prospectively compared the CAR between 46 actually ill schizophrenics (diagnosis according to DSM-III-R) treated with either haloperidol (n = 26) or clozapine (n = 20) and 30 well-matched healthy volunteers. Multivariate analysis demonstrated a significant effect of neuroleptic medication (haloperidol vs. clozapine) on heart rate and diastolic blood pressure under resting conditions as well as on the heart rate variance (30:15 ratio, deep-breathing, Valsalva) and blood pressure tests (sustained handgrip, Schellong). In addition a positive treatment response (using predefined outcome criteria of the Brief Psychiatric Rating Scale) was independently associated with lower resting heart rates and less impaired 30:15 ratios under neuroleptic medication. Our data indicate that clozapine treatment was associated with a substantial impairment of CAR, which can be explained by the drug's anticholinergic properties in combination with an increase in norepinephrine outflow. The greater heart rate variability in responders might be due to an early neuroleptic-induced decrease of sympathetic activity in the autonomic nervous system, which may precede clinical improvement. Our findings are discussed in relation to neuroleptic-induced changes in plasma catecholamine levels suggested to be useful biological markers in predicting treatment outcome.