Although most alcoholic subjects show little autonomic dysfunction, severe alcoholic subjects may have pathological changes in autonomic nerves. We asked if respiratory sinus arrhythmia amplitude (RSA), an index of vagal cardiac control, is decreased in alcoholism and, if so, whether the decrease is reversed with abstinence. RSA was assessed in 17 normotensive alcoholic subjects (A) at 1, 4, 12, and 24 wk of abstinence after detoxification and at similar intervals in 17 controls (C) matched for age, race, and gender. Subjects were studied in both supine and seated positions while breathing in a prescribed deep (>50% vital capacity) and slow (5-7/min) pattern. Mean heart rate (HR) was determined over 30 s from the electrocardiogram; RSA (the difference between maximum and minimum instantaneous HRs after inspiratory onset) was determined from 10 consecutive breaths. In C, both HR (supine: 61.5 +/- 2.2 beats/min; seated: 71.3 +/- 1.7 beats/min; P < 0.002) and RSA (supine: 22.5 +/- 1.0 beats/min; seated: 28.4 +/- 1.4 beats/min; P < 0.003) were higher when seated than when supine, but neither HR nor RSA varied over 24 wk. At week 1 of abstinence, HRs for A were higher than those for C (supine: 74.2 +/- 2.3 beats/min, P < 0.001; seated: 83.2 +/- 2.7 beats/min, P < 0.003), but by week 24, both seated and supine values returned to control levels. RSA in A at week 1, was only one-half that of C (supine: 11.1 +/- 1.4 beats/min, P < 0.001; seated: 14.7 +/- 1.9 beats/min, P < 0.001) and independent of body position. With maintained abstinence, RSA increased, until at week 12, it was not different from that in C. Furthermore, seated RSA values were now higher than supine RSA values. The recoveries of RSA amplitude, and of the postural effect on RSA, in abstinent alcoholic subjects, suggest that alcoholism produces a subclinical autonomic dysfunction that is rapidly reversed by abstinence.