BACKGROUND: Orthostatic hypotension is the most disabling symptom of autonomic failure. The choice of a presser agent is largely empiric, and it would be of great value to define predictors of a response. PATIENTS AND METHODS: In 35 patients with severe orthostatic hypotension due to multiple system atrophy or pure autonomic failure, we determined the effect on seated systolic blood pressure (SBP) of placebo, phenylpropanolamine (12.5 mg and 25 mg), yohimbine (5.4 mg), indomethacin (50 mg), ibuprofen (600 mg), caffeine (250 mg), and methylphenidate (5 mg). In a subgroup of patients, we compared the presser effect of midodrine (5 mg) with the effect of phenylpropanolamine (12.5 mg). RESULTS: There were no significant differences in the presser responses between patients with multiple system atrophy or pure autonomic failure. When compared with placebo, the presser response was significant for phenylpropanolamine, yohimbine, and indomethacin. In a subgroup of patients, we confirmed that this presser effect of phenylpropanolamine, yohimbine, and indomethacin corresponded to a significant increase in standing SEP. The presser responses to ibuprofen, caffeine, and methylphenidate were not significantly different from placebo. Phenylpropanolamine and midodrine elicited similar presser responses. There were no significant associations between drug response and autonomic function testing, postprandial hypotension, or plasma catecholamine levels. CONCLUSIONS: We conclude that significant increases in systolic blood pressure can be obtained in patients with orthostatic hypotension due to primary autonomic failure with phenylpropanolamine in low doses or yohimbine or indomethacin in moderate doses. The response to a presser agent cannot be predicted by autonomic function testing or plasma catecholamines. Therefore, empiric testing with a sequence of medications, based on the risk of side effects in the individual patient and the probability of a response, is a useful approach. (C) 1998 by Excerpta Medica, Inc.