Frequency of levodopa-related dyskinesias and motor fluctuations as estimated from the cumulative literature

There is no clear consensus regarding the frequency (and hence, the risk), of dyskinesias or motor fluctuations during chronic levodopa therapy for Parkinson's disease (PD). Multiple clinical series have tabulated these frequencies since the advent of levodopa over 30 years ago. We were interested in determining: (1) the aggregate frequency figures in the existing literature; and (2) how clinical series from the early levodopa era, which included patients with longer durations of parkinsonism, compare to more recent (modern era) series. We searched MEDLINE for all English language publications reporting the cumulative frequency of levodopa-induced dyskinesias or motor fluctuations during discrete intervals of treatment. This generated 2,478 publications spanning 1966 through September 2000. Papers with appropriate titles or abstracts were reviewed; reference lists from published clinical series were a source of additional papers for review. This ultimately yielded 74 publications with adequate data, relating to 112 intervals of levodopa treatment. Series that included patients with PD-onset well before levodopa availability (pre-levodopa era) were separately analyzed from all subsequent series. Series were grouped by duration of levodopa therapy and the median frequencies of dyskinesias and motor fluctuations were tabulated for each group. The data were analyzed both with and without adjustment for the number (N) in each series. Among series containing pre-levodopa era patients, the median dyskinesia frequency was already 50% by 5-6 months of treatment. This contrasts with the modern era series where dyskinesias were reported later in treatment. The median dyskinesia frequency was slightly less than 40% by 4-6 years of levodopa therapy among modern era patients. Motor fluctuations (wearing-off) were not tabulated in most of the early levodopa series. Among modern era reports, motor fluctuations were nil during the first year of levodopa therapy but were experienced by approximately 40% of patients by 4-6 years of treatment. Similar results were found when the analyses were restricted to only prospective studies where levodopa motor complications were targeted outcome measures. The conclusions reached were: (1) patients from the pre-levodopa era experienced dyskinesias much earlier during levodopa treatment than modern era patients, perhaps because of longer durations of pre-existing PD; (2) in the present era, patients treated with levodopa therapy for 4-6 years have approximately a 40% likelihood of experiencing motor fluctuations and a risk of dyskinesias just short of 40%; and (3) these findings represent incident data and the prevalence of clinically important morbidity may be substantially less.