Design: Two prospective multicentre cohorts, adjusted mixed-effects and Cox models
Population: 27,043 people with Parkinson's disease (NS-Park and PPMI)
Primary outcome: Freezing-of-gait severity and incidence (MDS-UPDRS 2.13)
Effect: Adjusted OR 1.84 (0.94–3.61) and 0.76 (0.38–1.51); H&Y stage HR 1.78–1.94

A Neurology study (9 September) examined two prospective cohorts — 25,602 patients in the French NS-Park network and 1,441 in the Parkinson's Progression Markers Initiative — to test the idea that levodopa exposure contributes to freezing of gait.

Unadjusted, levodopa use was associated with more freezing in NS-Park. After adjusting for disease duration, Hoehn and Yahr stage and motor score, the association disappeared (NS-Park OR 1.84, 95% CI 0.94–3.61; PPMI OR 0.76, 0.38–1.51). Future levodopa exposure did not predict new freezing. What did was disease stage (HR about 1.8–1.9 per Hoehn and Yahr stage) and motor severity.

The practical reading: freezing tracks how advanced the disease is, not the drug. That matters because fear of this effect is one reason levodopa is delayed.

Do not delay or under-dose levodopa for fear of causing freezing
Ask specifically about freezing at each review as disease stage advances
Distinguish OFF-period freezing, which may improve with better dopaminergic cover, from ON-freezing
Refer for physiotherapy with cueing strategies once freezing appears

Why it matters

It removes one of the reasons clinicians hold back levodopa in early Parkinson's disease.

The statistics, in plain English

The NS-Park adjusted odds ratio of 1.84 has an interval from 0.94 to 3.61, crossing 1, and PPMI pointed the other way. Together they give no consistent signal. Residual confounding could still hide a small effect in either direction, which the authors acknowledge.