Abstract
Background: The diagnosis of different parkinsonian syndromes and the ability to predict long-term drug efficacy constitute important clinical issues. Design: Motor responses to the acute administration of levodopa and apomorphine were analyzed in a series of 134 parkinsonian patients, including 83 patients with a clinical diagnosis of idiopathic Parkinson’s disease (PD), 28 patients with multiple-system atrophy (MSA), 6 with progressive supranuclear palsy, and 17 with an unclassified parkinsonian syndrome. Methods: The patients received oral levodopa/carbidopa (250/25 mg) and subcutaneous apomorphine (1.5, 3 and 4.5 mg). Clinical variations of the Unified Parkinson’s Disease Rating Scale (UPDRS) motor score were evaluated 1 h following levodopa administration or 20 min following apomorphine. The motor improvement produced by each acute challenge was matched with the clinical diagnosis and with the response to chronic levodopa treatment. The diagnosis was verified by repeated clinical assessments or by autopsy in 2 cases. A receiver operating characteristics curve was plotted comparing PD vs. non-PD, PD vs. MSA and chronic responders vs. nonresponders. Cutoff threshold improvement was defined as the value closest to the crossing point for 80% sensitivity and 80% specificity, corresponding to the best trade-off for a predictive evaluation. Results: UPDRS motor score improvement was on average higher in PD than in non-PD patients (levodopa: 29.8 vs. 12.2%; apomorphine 1.5 mg: 27.1 vs. 10.5%; apomorphine 3 mg: 27.7 vs. 9.7%; apomorphine 4.5 mg: 28.8 vs. 11.8%; p < 0.01 with Student’s t test). When PD patients were compared to non-PD patients, levodopa challenge had the best diagnostic accuracy with a threshold improvement of 16%. Apomorphine had the best diagnostic accuracy with a threshold improvement of 13.5% for 1.5 mg, 13% for 3 mg, and 16% for 4.5 mg. This meant that patients improving at least 16% in all tests had the highest probability of having PD. When PD patients were compared to MSA patients, levodopa acute challenge had the best diagnostic accuracy with a threshold improvement of 17%. Apomorphine had the best diagnostic accuracy with an improvement of 13% for 1.5 mg, 15% for 3 mg, and 18% for 4.5 mg. This meant that patients improving at least 18% in all tests had the highest probability of having PD rather than MSA. When patients who responded to chronic levodopa treatment were compared to those who did not, acute challenge with levodopa had the best predictive accuracy with a threshold improvement of 14.5%. Apomorphine had the best predictive accuracy with an improvement of 13% for 1.5 mg, and 14% for 3 and 4.5 mg. This meant that patients improving at least 14.5% in all tests had the highest probability of responding to chronic treatment. Conclusion: A good agreement was found between acute challenges with levodopa/carbidopa and apomorphine, and the use of both improved the reliability of the test. Different threshold improvements after acute challenges would support a diagnosis of PD or the exclusion of MSA, and would have a predictive value for subsequent response to chronic levodopa therapy.