PubMed1998
We estimated the accuracy of a modified commonly used set of clinical diagnostic criteria for the diagnosis of multiple system atrophy (MSA) by retrospectively applying the criteria to the features recorded by six neurologists who had evaluated 105 autopsy-confirmed cases (16 MSA and 89 non-MSA disorders). Cases were abstracted from the records of the patients' first visit to an academic center, and were presented as clinical vignettes to six neurologists, each of whom recorded the main clinical features of the presented clinical vignette on a standardized form. Sensitivity and positive predictive values were chosen as validity outcome measures and were calculated by comparing the applied diagnostic criteria to the neuropathologic information. Of note, most MSA patients in this study (mainly those with Shy-Drager type) had not received levodopa therapy since the primary neurologists often had not perceived a need to administer this treatment. The validity of the retrospectively applied criteria for the diagnosis of possible MSA (sensitivity: median, 53%, range, 50-69%; positive predictive value: 30%, 28-39%) and probable MSA (sensitivity: 44%, 31-60%; positive predictive value: 68%, 54-80%) at the first visit was suboptimal. The best, still not perfect, accuracy for this set of diagnostic criteria was obtained when six out of eight features (sporadic adult onset, dysautonomia, parkinsonism, pyramidal signs, cerebellar signs, no levodopa response, no cognitive dysfunction, or no downward gaze supranuclear palsy) were present (median sensitivity, 59%; range, 50-75%; positive predictive value: 67%, 53-83%). This is the first study to validate criteria for the clinical diagnosis of MSA. Our data suggest that it is difficult to achieve an early and accurate clinical diagnosis of this disorder. The probability of correctly diagnosing MSA increases when at least six features of this modified set of criteria are present or when requiring the set for probable MSA.