Objective: The aim of this study was to establish a new stepwise type of acute challenge test with incremental doses of levodopa/benserazide, and verify its predictive value in follow-up diagnoses and outcomes of deep-brain stimulation (DBS) in treated patients with parkinsonism. Design: Prospective cohort study. Outcomes: The optimal cutoff points for UPDRS-III improvement in these stepwise levodopa tests. In this study, we established acute challenge tests with incremental doses of levodopa/benserazide (100/25 mg, 150/37.5 mg, 200/50 mg and 300/75 mg) in treated patients with parkinsonism (n = 175). The receiver operating characteristic (ROC) curves were plotted to compare peak UPDRS-III improvement of PD patients (n = 112)with that of non-PD parkinsonism patients (n = 63). The point on the ROC curve with the highest Youden index was defined as the optimal cutoff point in motor improvement for differential diagnoses. The results of the new tests were compared with follow-up diagnoses and the outcomes of DBS. Results: The optimal cutoff points for UPDRS-III improvement with maximal Youden Indices on ROC curves from the tests, with the four incremental doses of levodopa/benserazide, were 12.2% (100125 mg), 22.3% (150/37.5 mg), 27.9% (200/50 mg) and 33.4% (300/75 mg). The test showed significant correlation with follow-up diagnosis and the outcomes of DBS (P of Kappa <0.01). Conclusions: The results suggested that the new acute stepwise levodopa challenge test is a useful tool for the diagnosis of PD. Crown Copyright (C) 2008 Published by Elsevier Ltd. All rights reserved.