Levodopa-induced dyskinesia in dementia with Lewy bodies and Parkinson disease with dementia

Abstract

Objective

To investigate the frequency of levodopa-induced dyskinesia in dementia with Lewy bodies (DLBs) and Parkinson disease with dementia (PDD) and compare these frequencies with patients with incident Parkinson disease (PD) through a population-based cohort study.

Methods

We identified all patients with DLB, PDD, and PD without dementia in a 1991–2010 population-based parkinsonism-incident cohort, in Olmsted County, Minnesota. We abstracted information about levodopa-induced dyskinesia. We compared patients with DLB and PDD with dyskinesia with patients with PD from the same cohort.

Results

Levodopa use and dyskinesia data were available for 141/143 (98.6%) patients with a diagnosis of either DLB or PDD; 87 (61.7%), treated with levodopa. Dyskinesia was documented in 12.6% (8 DLB and 3 PDD) of levodopa-treated patients. Among these patients, median parkinsonism diagnosis age was 74 years (range: 64–80 years); 63.6%, male. The median interval from levodopa initiation to dyskinesia onset was 2 years (range: 3 months–4 years); the median daily levodopa dosage was 600 mg (range: 50–1,600 mg). Dyskinesia severity led to levodopa adjustments in 5 patients, and all improved. Patients with dyskinesia were diagnosed with parkinsonism at a significantly younger age compared with patients without dyskinesia (p < 0.001). Levodopa dosage was unrelated to increased risk of dyskinesias among DLB and PDD. In contrast, 30.1% of levodopa-treated patients with PD developed dyskinesia. In age-, sex-, and levodopa dosage–adjusted models, Patients with DLB and PDD each had lower odds of developing dyskinesia than patients with PD (odds ratio = 0.42, 95% CI 0.21–0.88; p = 0.02).

Conclusions

The dyskinesia risk for levodopa-treated patients with DLB or PDD was substantially less than for levodopa-treated patients with PD.

Lewy body parkinsonism may be complicated by dementia, either early (dementia with Lewy bodies, DLB) or later (Parkinson disease with dementia, PDD). The arbitrary distinction between DLB and PDD has been based on the onset time of cognitive decline compared with the motor symptoms.¹ For all these disorders, alpha-synuclein constitutes a major component of the Lewy neuropathology and is presumed to be a pathogenic factor; thus, these conditions are defined as synucleinopathies.²

Levodopa continues to be the primary symptomatic therapy for the parkinsonian symptoms of all Lewy body disorders.³ However, over time, levodopa dyskinesias or motor fluctuations may develop.⁴ Concerns about these levodopa complications have erroneously led some physicians to postpone the introduction of levodopa. Although the frequency and risks of dyskinesias in PD have previously been addressed,^5,6 they have not been adequately studied in DLB and PDD.

The aim of this investigation was to assess the frequency and characteristics of levodopa-induced dyskinesias among patients with a diagnosis of DLB or PDD, using our incident cohort of all parkinsonism types identified among Olmsted County, MN, residents from 1991 to 2010. We compared the frequency and characteristics of levodopa-induced dyskinesias among patients with DLB and PDD to patients with incident PD from the same cohort, which has been previously reported.⁵ We tabulated and compared potential risk factors, medication doses, and dosage adjustments.

Methods

We identified 669 incident cases of parkinsonism in Olmsted County, MN, from 1991 to 2010 using the unique infrastructure of the Rochester Epidemiology Project (REP). Details regarding the identification of the cases and the clinical diagnostic criteria are reported elsewhere.²

Case identification

We reviewed the medical records of all patients who fulfilled the criteria for the clinical diagnosis of DLB and PDD and were residents of Olmsted County.^2,7,8 We abstracted information about levodopa treatment, sex, age at onset of parkinsonism, time between the initial levodopa administration and onset of dyskinesia, dyskinesia severity, therapy adjustments (levodopa dosage changes or amantadine use), and duration of follow-up from their medical records. A movement disorders specialist (R.S.) reviewed all the identified cases to ascertain the final diagnosis and the presence of dyskinesia. We recorded dyskinesia only if specifically mentioned in the clinical records. Pure dystonia (e.g., dystonic toe curling or foot inversion) reflective of a levodopa-off state was not counted as dyskinesia. We graded dyskinesias as mild, moderate, or severe according to the physician's assessment recorded in the records, as previously reported.⁵ A movement disorder specialist (R.S.) reviewed all the cases to confirm severity, according to the signs reported in the patient medical records.²

Statistical analysis

Wilcoxon rank-sum and Fisher exact tests were used to compare characteristics between patients with DLB/PDD who developed dyskinesia and those who did not. We used logistic regression models to determine the association between levodopa dosage in tertiles and dyskinesia among the patients with DLB and PDD. Model 1 was unadjusted. Model 2 was adjusted for age and sex. We also used logistic regression to determine the odds of developing dyskinesias for patients with DLB/PDD compared with patients with PD. Model 1 adjusted for levodopa dosage, and model 2 adjusted for levodopa dosage, age, and sex. All analyses were completed using Stata version 13.0 (Stata Corp, College Station, TX).

Standard protocol approvals, registrations, and patient consents

The Mayo Clinic and Olmsted Medical Center Institutional Review Boards approved this study, and participating patients (or their legally authorized representatives) provided written consent to use their medical information.

Data availability

All the relevant data have been shared and published in this article; data regarding case ascertainment of parkinsonism and methodology on case identification have been previously published.²

Results

Demographics

This incidence cohort included 143 patients with a clinical diagnosis of DLB or PDD; 87 were diagnosed with DLB and 56 with PDD. Information about levodopa use and dyskinesia was available in 141/143 (98.6%). Among the 141 cases, 87 (61.7%) were treated with levodopa (54 DLB and 33 PDD). The median age at diagnosis of DLB/PDD was 77 years (range: 33–98 years). Among those treated with levodopa, the median duration of treatment was 6 years (range: 1 month–21 years). Levodopa was lowered or stopped in 5 patients because of severe hallucinations/delirium (3.5% of the cases: 2 with DLB and 3 with PDD).

The comparison group of patients with PD (nondemented) was from the same incidence cohort, and their data have previously been reported.⁵ Among these patients with PD, the levodopa dosage was documented in 247/309 (79.9%), and dyskinesia occurrence in these 247 patients was tabulated for comparison to the DLB and PDD cases. None of these PD patients were initially treated with a dopamine agonist.

Dyskinesia frequency, severity, and duration of treatment

Among the 87 patients with DLB and PDD treated with levodopa, 11 (12.6%) developed dyskinesia at any time after levodopa therapy initiation. The breakdown was 8/54 (14.8%) for DLB and 3/33 (9.0%) for PDD. The median age at diagnosis of DLB/PDD for the levodopa-treated patients was 77 years (range: 61–98 years), and the median follow-up time after diagnosis was 6 years (range: 1 month–15 years). The median levodopa treatment duration was 6 years (range: 1 month–14 years); the median maximum daily levodopa dosage was 600 mg (range: 50–1,600 mg). The median time between levodopa initiation and dyskinesia onset was 2 years (range: 3 months–4 years). None of the patients with dyskinesia were treated with dopamine agonists. Two patients were treated with dopamine agonists along with levodopa after the clinical diagnosis was made; neither developed dyskinesias.

Among the 11 patients with PDD/DLB with dyskinesia, 7 (63.6%) were male; the frequency of dyskinesia did not differ by sex (p = 1.00). Patients with dyskinesia were diagnosed with parkinsonism at a significantly younger age (median 74 years, range: 64–80 years) compared with patients without dyskinesia (median 78 years range: 61–98 years; p < 0.001). There was no difference in levodopa therapy duration between dyskinetic and nondyskinetic patients (p = 0.09). Dyskinetic patients had significantly longer median follow-up time after diagnosis compared with nondyskinetic PDD/DLB patients (p = 0.03).

Levodopa dosage was not associated with increased odds of developing dyskinesias even after adjusting for age and sex in these DLB/PDD cases (table 1). We observed a 2-fold increased odds of developing dyskinesias in the third tertile of levodopa dosage, but this was not significant. Among the 11 dyskinetic patients, 8 had mild dyskinesia, 1 had moderate dyskinesia, and the severity was not reported for 2. Levodopa dosage was adjusted in 5 dyskinetic patients, with symptomatic improvement or resolution of dyskinesias in all.

Table 1.

Association between maximum levodopa dosage and odds of developing dyskinesia among patients with PDD and DLB, with and without adjustment for age and sex

Comparison analysis between DLB/PDD and PD

In this incidence cohort, 247 patients with PD without dementia were treated with levodopa. Among these, 84 (30.1%) developed dyskinesia of any severity at any time during the disease course.⁵ These served as the comparison group.

In models adjusted for levodopa dosage (odds ratio [OR] 0.34, 95% confidence interval [CI] 0.17–0.69, p = 0.003) and models adjusted for age, sex, and levodopa dosage (OR 0.42, 95% CI 0.21–0.88, p = 0.02), patients with DLB/PDD had significantly lower odds of developing dyskinesia compared with patients with PD. The median maximum daily levodopa dosage among patients with DLB/PDD (600 mg, range: 50–1,600 mg) was significantly lower than the daily levodopa dosage among PD patients (700 mg, range: 150–2,700 mg; p < 0.001). Patients with DLB/PDD who developed dyskinesia were diagnosed at a significantly older age compared with PD dyskinetic patients (74 vs 68 years, respectively, p < 0.001). The median time from levodopa therapy initiation to dyskinesia onset for patients with DLB/PDD was 2 years (range: 3 months–4 years), whereas for patients with PD, it was 4 years (range: 2 months–20 years; p = 0.003). There was a statistically significant difference between the median follow-up time after levodopa therapy initiation between patients with DLB/PDD and PD (p < 0.001). No difference in terms of median follow-up time after symptom onset was observed between dyskinetic DLB/PDD and dyskinetic PD patients (p = 0.35). Results of the comparison analysis are summarized in table 2.

Table 2.

Comparison of characteristics associated with dyskinesia for patients with DLB/PDD parkinsonism vs Parkinson disease

Postmortem examinations

Brain autopsy was available in 3 dyskinetic patients with a clinical diagnosis of DLB. The median time from dyskinesia onset to death was 3 years (range: 2–8 years). The pathologic diagnoses were consistent with the clinical diagnoses. For further details about the pathologic characteristics of our cohort, please refer our previous reports.^5,7

Discussion

Optimized levodopa therapy is, by far, the most efficacious symptomatic treatment of dopamine deficiency symptoms and signs. Some authors have advised delaying levodopa initiation and limiting the dosage due to fear of dyskinesia (and response fluctuations). The parkinsonism of DLB and PDD has a predominantly dopamine deficiency basis and typically benefits from levodopa therapy. Hence, the dyskinesia risk among levodopa-treated patients with DLB/PDD is very relevant to practice.

The dyskinesia risk in our incidence cohort of patients with PDD/DLB was only 12.6%, with dyskinesia characterized as mild except for 1 patient with moderate dyskinesia. Clinicians in our community are aware that a small reduction of each levodopa dosage typically resolves or markedly improves dyskinesia, and this proved to be the case among the patients with PDD/DLB where the dosage was reduced. Few studies have been conducted on the frequency of dyskinesia in this group of patients, and results were often influenced by the relatively small sample size or the use of levodopa challenge tests, which do not reflect the everyday experience with chronic dosing. In any case, the findings in this study are consistent with previous reports. Thus, dyskinesias were reported in 16% of the cases in a cohort of patients with DLB.⁹ Also, the dyskinesia frequency was reported to be lower in DLB compared with both PDD and PD.¹⁰

PDD/DLB dyskinesia frequency was substantially less than documented among patients with PD from the same incidence cohort (12.6% vs 30.1%).⁵ Several factors may have contributed, including the levodopa dosage, age, duration of treatment, and length of follow-up. In the statistical analysis, the levodopa dosage and age were adjusted, but the small number of PDD/DLB dyskinetic cases likely challenged the power of this analysis. Although not statistically significant, the third tertile of levodopa dosage was associated with a 2-fold increase in the odds of developing dyskinesias. Levodopa dosage is a well-known risk factor for dyskinesia.¹¹

Clinicians may choose to limit levodopa dosages among patients with PDD/DLB to avoid psychosis.¹² However, in our cohort, only a small number of patients (3.5%) needed to discontinue levodopa treatment due to delirium or psychosis.

Another factor that may have resulted in a lower frequency of dyskinesia among the PDD/DLB cases is the degree of Lewy neuropathologic progression. PDD/DLB Lewy pathology is presumably more widespread than in PD, with a heavier Lewy body pathology burden in the striatum compared with PD.¹³ This proliferation of Lewy pathology may obviously extend into basal ganglia neuronal circuits wired in series with the dopaminergic nigrostriatal system. Whereas pallidotomy is known to abolish levodopa dyskinesia, a Lewy-neurodegenerative lesion of the pallidum or related circuits due to a heavier burden of α-synuclein pathology in DLB/PDD compared to PD may lead to the same outcome.¹⁴ Note also that dyskinesia risk substantially diminishes progressively with age,^11,15 as potential result of the progressive loss of neuropil attending normal aging.^16,17 This has sometimes been referred to as an “autopallidotomy” effect.

Our study has several strengths. First, we used the well-defined population-based incident cohort of patients with parkinsonism of the record-linked system of the REP. The population-based nature of this cohort is less likely to be affected by selection bias than many case series. Second, the vast majority of patients have been treated and followed by Mayo Clinic neurologists, who have very similar training and practice approaches that favor optimizing levodopa dosage. Because of the geography of Olmsted County, the Mayo Clinic is the only referring center in the area with a neurology department with movement disorders specialists. Therefore, all patients were seen at least once by a movement disorders specialist. Thus, patients with DLB and PDD and motor complications are usually likely to be followed by 1 neurologist in the Movement Disorders Division of the Mayo Clinic. Although few cases had postmortem neuropathologic examination, notably, all 3 deceased, dyskinetic patients with complete autopsy showed concordance between the clinical and the pathologic diagnoses.

This study also has limitations. First, we acknowledge that mild or infrequent levodopa-induced dyskinesia might not have been reported in the medical records. This may have led to an underestimation of the prevalence of dyskinesia in our cohort. Second, patients were seen as part of the routine clinical practice without standardization of the recorded data. Third, we could not gather data about the severity and progression of atypical parkinsonisms assessed with the Unified Parkinson Disease Rating Scale or the Hoehn and Yahr stage. Fourth, the median levodopa treatment duration of only 6 years may have influenced our low dyskinesia frequency report. Fifth, it was not possible to compare dyskinesia characteristics between patients with DLB and PDD due to the relatively small number of dyskinetic patients present in these groups. Last, our study is based on a relatively small sample size, and it is possible that the results may change in a larger population; however, our results are consistent with previous literature, and our study population is one of the largest explored.

In conclusion, our study showed that only 12.6% of patients with a clinical diagnosis of either DLB or PDD developed levodopa-induced dyskinesias. This was substantially less than the dyskinesia frequency among patients with PD from this cohort. When present, dyskinesias were mild and easily managed with levodopa dosage reductions. These data suggest that dyskinesia risk should not discourage clinicians from prescribing levodopa when appropriate for treating parkinsonism in patients with PDD or DLB.

Appendix. Authors

Study funding

The work was supported by the National Institute on Aging of the National Institutes of Health (R01 AG034676) and by the Mayo Foundation for Medical Education and Research.

Disclosure

P. Turcano and C.D. Stang report no disclosures. J.H. Bower receives funding from the Parkinson's Disease Foundation, Inc. J.E. Ahlskog receives royalties from Oxford University Press for 3 recently published books. B.F. Boeve receives funding from the National Institute on Aging, the National Institute of Neurological Disorders and Stroke, the Lewy Body Dementia Association, GE Healthcare, Axovant Sciences, Inc., and Biogen. M.M. Mielke receives funding from the NIH and unrestricted research grants from Biogen and Lundbeck. She has consulted for Lysosomal Therapeutics, Inc, and Eli Lilly R. Savica receives funding from the National Institute on Aging, the National Institute of Neurological Disorders and Stroke, and the Parkinson's Disease Foundation, Inc. Full disclosure form information provided by the authors is available with the full text of this article at Neurology.org/cp.