Disentangling Levodopa Exposure and Freezing of Gait

Parkinson disease presents clinicians with challenging motor complications, among which freezing of gait remains one of the most debilitating. Freezing of gait causes an abrupt inability to generate effective forward stepping, frequently precipitating falls and worsening quality of life. For several years, movement disorder specialists debated whether dopaminergic therapy might paradoxically trigger or exacerbate these motor blocks. Consequently, this concern provoked substantial anxiety among treating physicians, who feared that long-term levodopa might accelerate axial disability. However, establishing true causality in observational registries is notoriously difficult due to confounding by indication. Clinicians typically prescribe higher doses of levodopa to patients with greater disease severity and longer disease duration. Therefore, patients who develop severe motor problems naturally receive more intensive levodopa regimens. To resolve this therapeutic controversy, a landmark multicentric study evaluated whether levodopa independently promotes episodic gait freezing or whether underlying neurodegeneration drives the phenomenon. By examining data from thousands of well-characterized individuals, researchers isolated the true biological contributors. Ultimately, their findings offer critical clarity for neurological practice.

Methodological Design of the NS-Park and PPMI Cohort Analyses

To evaluate the impact of dopaminergic therapy, researchers conducted a comparative analysis across two prospective cohorts: the French NS-Park registry and the Parkinson's Progression Markers Initiative (PPMI). The investigation included 25,602 patients from NS-Park and 1,441 participants from PPMI. Furthermore, the protocol required participants to have completed at least two comprehensive follow-up evaluations. Investigators created three nested subcohorts based on disease duration, previous levodopa exposure, and baseline motor status. The team quantified freezing episodes using the Movement Disorder Society-Unified Parkinson's Disease Rating Scale (MDS-UPDRS) Part II Item 2.13. Additionally, researchers applied mixed-effects ordered logistic regression to track longitudinal changes in freezing severity. They also used Kaplan-Meier estimates and Cox proportional hazards regression to analyze incident motor blocks. Importantly, all multivariable models adjusted for overall disease duration, Hoehn and Yahr clinical stage, and MDS-UPDRS Part III motor scores. Consequently, this rigorous methodological framework controlled for confounding by disease progression. Therefore, the analytic design provided an exceptionally reliable assessment of pharmacological exposure versus natural disease evolution.

Statistical Findings: Unadjusted Associations Versus Multivariable Models

The study clearly demonstrated how unadjusted observational analyses can produce misleading impressions of pharmacotherapy risks. In raw baseline comparisons within the large NS-Park cohort, levodopa use correlated with greater freezing severity and higher incidence. However, this alarming signal completely disappeared in the meticulously observed PPMI cohort. Moreover, when investigators applied multivariable mixed-effects ordinal logistic regression, the apparent levodopa hazard evaporated across both populations. Specifically, the adjusted odds ratio for levodopa status in NS-Park dropped to 1.84, with a non-significant confidence interval of 0.94 to 3.61. Similarly, the PPMI cohort demonstrated an adjusted odds ratio of 0.76, with a confidence interval of 0.38 to 1.51. Thus, neither cohort showed a statistically significant link between levodopa exposure and freezing severity after multivariable adjustment. In addition, prospective survival analyses confirmed that future levodopa exposure did not trigger de novo freezing episodes among previously unaffected patients. These rigorous data illustrate that initial clinical impressions often confuse therapeutic exposure with disease progression. Ultimately, the apparent adverse signal observed in routine practice represents an artifact of confounding by disease severity.

Disease Progression as the Primary Driver of Motor Freezing

While levodopa status showed no meaningful detrimental effect, clinical indicators of neurodegenerative advancement emerged as overwhelming predictors of motor arrest. In the adjusted regression models, disease duration showed a massive independent association with freezing severity across both cohorts. Specifically, longer disease duration yielded an odds ratio of 4.40 in NS-Park and 6.97 in PPMI. Furthermore, advanced Hoehn and Yahr staging strongly correlated with symptom burden, demonstrating an odds ratio of 6.80 in NS-Park. Similarly, objective motor impairment measured by MDS-UPDRS Part III scores predicted freezing severity, with odds ratios of 1.85 in NS-Park and 2.00 in PPMI. In addition, survival models revealed that incident motor blocks were driven entirely by clinical progression. Advanced Hoehn and Yahr stage nearly doubled the risk of new-onset freezing, generating hazard ratios of 1.94 in NS-Park and 1.78 in PPMI. Likewise, higher MDS-UPDRS Part III scores significantly predicted incident freezing across both populations. Consequently, clinicians must recognize that axial motor deterioration reflects widespread degenerating pathways beyond the nigrostriatal system. Therefore, progressive neurodegeneration, rather than dopaminergic treatment, explains the emergence of episodic gait arrest.

Clinical Implications and Multidisciplinary Management

These robust cohort findings deliver immense reassurance to neurologists and geriatricians managing Parkinson disease. Historically, unfounded fears regarding medication-induced motor complications led some practitioners to delay levodopa initiation or restrict doses prematurely. However, withholding necessary dopaminergic therapy deprives patients of essential mobility and elevates fall risks due to worsened rigidity. Because the data confirm that levodopa does not accelerate motor freezing, clinicians should not restrict its rational use. Instead, physicians should optimize dopaminergic schedules to eliminate off-state freezing episodes, which frequently resolve with adequate striatal stimulation. Conversely, if motor blocks persist during optimal on-states, clinicians must recognize that non-dopaminergic networks are likely responsible. For instance, cholinergic degeneration in the pedunculopontine nucleus and frontal executive deficits contribute heavily to refractory gait disturbances. In such scenarios, clinicians should emphasize multidisciplinary non-pharmacological interventions rather than escalating medication doses. Specifically, physical therapists can implement external auditory rhythms or visual cues to overcome motor arrest. Furthermore, clinicians must eliminate unnecessary sedative drugs and recommend home safety modifications. Ultimately, combining rational pharmacotherapy with specialized physical rehabilitation maximizes patient independence and functional mobility.

Does long-term levodopa therapy directly cause freezing of gait in Parkinson disease?

Current high-quality evidence confirms that levodopa therapy does not independently cause or accelerate freezing of gait. Although unadjusted observational data suggested a potential link, rigorous multivariable analyses across large cohorts demonstrate that the association disappears after controlling for underlying disease duration, Hoehn and Yahr stage, and overall motor severity.

What are the primary clinical predictors of freezing of gait in Parkinson patients?

The primary predictors of freezing of gait are advanced disease duration, higher Hoehn and Yahr clinical staging, and elevated MDS-UPDRS Part III motor examination scores. These variables directly reflect widespread neurodegeneration across dopaminergic and non-dopaminergic networks, which impairs supraspinal locomotor integration and causes episodic motor blocks.

How should clinicians differentiate between off-state and on-state freezing of gait?

Clinicians differentiate between freezing types by assessing symptom timing relative to medication doses. Off-state freezing occurs during trough plasma concentrations and typically improves with dopaminergic optimization. In contrast, on-state freezing persists or worsens during peak medication effect, necessitating sensory cueing, physical rehabilitation, and multidisciplinary non-dopaminergic management strategies.

Disclaimer: This content is for informational and educational purposes only and should not be considered medical advice. Always consult a qualified healthcare professional regarding any medical condition or treatment. Refer to the latest local and national guidelines for clinical practice.

References

1. Barbosa R et al. Levodopa Use and Characteristics of Freezing-of-Gait in Patients With Parkinson Disease in the NS-Park and PPMI Cohorts. Neurology. 2026 Oct 13. doi: 10.1212/WNL.0000000000218515. PMID: 42715510.


2. Jansen J, Capato T, Nonnekes J, et al. Exploring the levodopa-paradox of freezing of gait in dopaminergic medication-naïve Parkinson's disease populations. NPJ Parkinsons Dis. 2023;9(1):132.


3. Nonnekes J, Bloem BR. Freezing of gait before the introduction of levodopa. Lancet Neurol. 2019;18(4):333-334.