Cost-Effectiveness of Extended-Release Carbidopa-Levodopa for Advanced Parkinson's Disease
Existing levodopa therapies for Parkinson’s disease often provide inadequate symptom control after prolonged treatment. This study assessed the cost-effectiveness of an extended-release carbidopa-levodopa formulation (IPX066).
Objectives: Prolonged treatment of Parkinson’s disease (PD) with immediate-release (IR) carbidopa-levodopa (CD-LD) often leads to motor fluctuations, including dyskinesia and reemergence of PD symptoms (“off” time). Adding entacapone (E) to IR CD-LD (to create CL+E) improved efficacy; however, recent clinical studies demonstrated improved efficacy with extended-release carbidopa-levodopa (IPX066) compared with CL+E. This study evaluated the long-term cost-effectiveness of IPX066 versus branded and generic CL+E from the US payer perspective.
Study Design: State-transition Markov analysis.
Methods: A Markov model (6-month cycle, 3% annual discount rate) was developed to simulate PD progression over 5 years through 3 health states: ≤25% “off” time, >25% “off” time, and death. Dosing and efficacy data were from a phase 3 study comparing IPX066 with CL+E (administered either as IR CD-LD [Sinemet 25/100 mg] + entacapone [Comtan 200 mg] or CL+E combination tablet [branded]) and an open-label extension study. Health state utilities were from published literature. Total direct costs and quality-adjusted life-years (QALYs) were evaluated. One-way sensitivity analyses evaluated input uncertainty ranges and identified inputs that most affected the results.
Results: Total 5-year costs with IPX066, branded CL+E, and generic CL+E therapy were $68,703, $91,949, and $79,332, respectively. Intervention-related costs for IPX066 treatment were less than those for branded CL+E and slightly more than those for generic CL+E. The analysis showed cost savings with IPX066 of $166,044 per QALY gained, compared against branded CL+E, and $75,920 per QALY gained against generic CL+E. One-way sensitivity analyses demonstrated dominance with IPX066 in the majority of scenarios.
Conclusions: IPX066 was cost-effective compared with both branded and generic CL+E and was expected to lower total 5-year costs and increase QALYs for patients with PD.
Am J Pharm Benefits. 2017;9(1):23-29
Parkinson’s disease (PD) is a progressive, chronic disease impacting approximately 1 million individuals in the United States. The prevalence of PD substantially increases with age, and given the growing aging population in the United States, the prevalence of the disease will continue to increase.1 The annual economic burden associated with PD in the United States was estimated to exceed $14.4 billion in 2010, with 44% due to direct medical costs.2 Prescription drugs account for 17% of direct medical costs, while nursing home care is the largest component, accounting for approximately 57% of direct medical costs.2
The cardinal symptoms of PD are resting tremor, bradykinesia, postural instability, and rigidity, which result from the death of dopamine-containing neurons in the substantia nigra. The debilitating symptoms markedly decrease the ability of patients with PD to perform activities of daily living.3,4 In addition, dementia, depression, pneumonia, chronic pain, and other serious comorbidities are commonly associated with this disease.5 Treatments for PD are only symptomatic; currently, there are no treatments that slow the progression of the disease or reverse neuronal degeneration.
Immediate-release (IR) carbidopa-levodopa (CD-LD) is the standard of care for patients with advanced PD. However, the effectiveness of IR CD-LD diminishes with long-term treatment, which is associated with development of motor fluctuations, including dyskinesias (uncontrolled movements) and the reemergence of PD symptoms (“wearing off”).6 The catechol-o-methyl transferase inhibitor, entacapone (E), has been shown to increase the duration of therapeutic levodopa plasma concentrations when given after an individual dose of IR CD-LD, resulting in subsequent clinical improvement in “on” time and “off” time compared with IR CD-LD alone.6,7 Longer-term evaluations, however, have failed to provide evidence of improvement in the onset of dyskinesia with IR CD-LD + E (CL+E) compared with IR CD-LD without entacapone.8 Given the conflicting evidence of clinical superiority and the higher incidence of adverse events with CL+E,6-8 alternative therapies are needed to treat a growing population of individuals with PD.
IPX066 (Impax Laboratories, Inc, Hayward, California) is an extended-release oral formulation of CD-LD designed to address some of the limitations of the IR formulation by prolonging therapeutic LD plasma concentrations while maintaining the onset of effect seen with IR CD-LD. Improved efficacy of IPX066 over CL+E was demonstrated in a phase 3 randomized clinical trial, where a statistically significant reduction in “off” time was observed in patients during treatment with IPX066.9
The objective of this study was to develop a Markov model to evaluate the cost-effectiveness of IPX066 compared with both branded (Novartis Pharmaceutical Corporation, Basel, Switzerland) and generic CL+E, utilizing clinical inputs from 2 clinical trials9,10 and published literature.
A Markov model with 6-month cycles was developed to simulate progressions through PD health states and mortality for 3 hypothetical cohorts of patients with PD and end-of-dose motor fluctuations: 1) intervention with IPX066, 2) branded CL+E, and 3) generic CL+E. Institutional review board approval was not necessary for this study. Figure 1 depicts a simplified presentation of the model. Six-month cycles were selected to closely approximate the prior economic evaluations in this drug class.11,12
PD health states were defined as the percentage of “off” time per waking hours per day: ≤25% versus >25%. All patients were assumed to start in the health state of >25% “off” time.11,12 As patients aged in the 5-year model, they could remain in the >25% “off” time health state, transition to an improved health state (≤25% “off” time), or die. The probability of death was assumed to be independent of PD health state and was based on the latest available US life table from the CDC National Vital Statistics System.13 Based on the total time spent in the different health states, the model estimated the expected costs and quality of life associated with each treatment to estimate the relative cost-effectiveness.
Clinical inputs. Clinical inputs were based on data from 2 clinical trials: a phase 3 trial comparing the efficacy and safety of IPX066 with CL+E (ASCEND-PD)9 and an open-label extension study of IPX066 in patients with early or advanced PD.10 The ASCEND-PD trial enrolled patients with advanced PD who had at least 2.5 hours of “off” time per day (mean = 6.0 hours/day), using a double-blind, randomized, crossover design.9 Following a 6-week dose conversion period, during which all patients were converted from IR CD-LD (Sinemet 25/100 mg [Merck & Co, Inc; Whitehouse Station, New Jersey]) + entacapone (Comtan 200 mg [Orion Pharma; Espoo, Finland]) or the CL+E combination tablet (branded) to treatment with IPX066, each double-blind treatment period with IPX066 or CL+E was 2 weeks, separated by a 1-week washout period with open-label IPX066. The median dosage of levodopa from IPX066 for patients in the trial was 1495 mg per day, and the median dosage of levodopa from CL+E was 600 mg per day.9 The second trial used for model inputs was a 9-month open-label extension study that enrolled both early and advanced PD patients10 who had each completed a previous controlled, randomized clinical study of IPX066.14-16 The median dose of levodopa for IPX066 reached 1450 mg per day in patients with advanced PD.10 The assumption that patients were aged 60 years on entry to the model was based on the average age (mean = 65 years) of the randomized population in the phase 3 clinical trial of IPX066 versus CL+E.9
Transition probabilities. The probability of clinical improvement within the first 6 months (cycle 1) was assumed to be similar to data provided in the literature.11 Transitions to the better, ≤25% “off” time health state, were modeled only during the first cycle. After the first cycle of 6 months, PD can deteriorate from ≤25% “off” time to >25% “off” time or remain at ≤25% “off” time; the probability of deterioration was the same across the 3 treatment arms.11 The calculated transition probabilities estimated that the improvement in “off” time plateaued at week 10.9 Due to the chronic, progressive nature of PD, it was assumed that the probability of a patient in the ≤25% “off” time category transitioning to the >25% “off” time category after week 10 would be 1 percentage point every 2 weeks.
For IPX066, transition probabilities were derived from study IPX066-B09-03 (NCT01096186).10 At each visit, patients were administered the Unified Parkinson’s Disease Rating Scale, part of which queries the patient’s average proportion of the day in the “off” state. This was categorized into >25% or ≤25% “off” time per day. Shift tables were created indicating the proportion of individuals who either remained in the current state or switched to the alternative “off” time state after approximately 6 months.
CL+E transition probabilities were derived from study IPX066-B09-06 (NCT01130493).9 Patient PD diary data were collected for 3 consecutive days prior to each clinic visit for the assessment of treatment efficacy; thus, due to the trial design, after randomization, only 1 transition probability was available. Transitions were calculated between conversion period and the end of treatment for a total of 2 weeks, and, as above, shift tables were created to indicate the proportion of individuals who either remained in the current state or switched to the alternative “off” time state. Six-month transition probabilities for improvement to ≤25% “off” time in the first cycle were derived from the 2-week data, assuming that all patients began in >25% “off” time. Transition probabilities used in the model are shown in Table 1.
Utility measures. The utility measures for each of the health states in the Markov model were based on the literature.12 These utility estimates were determined using the standard gamble method in a US population. Utility is the degree of preference for a particular health state, where 1 is perfect health and 0 is death. In this analysis, and in a previous Markov model,12 the health states were indicated by time spent in the “off” state (>25% “off” time and ≤25% “off” time), with patients having lower utility in the >25% “off” time health state. The cutoff values for “off” time health states were utilized based on a prospective economic assessment of 40 outpatients with PD, which observed a significant decrease in utility estimates and a substantial increase in cost of care in patients experiencing >25% “off” time per day.17 After the first cycle, patients did not experience a disutility for having transitioned into the >25% “off” state (deterioration of health states). The utility values used in the model are shown in Table 1.
Cost inputs and utilization of medical resources. The analysis was conducted from the perspective of the third-party payer and included direct medical costs and pharmacy costs related to the treatment of PD. Assumptions of the costs associated with the 2 health states were based on the literature.11 All reported costs were inflated to 2014 values using the medical component of the Consumer Price Index.18 All costs were discounted in the base case at 3% annually, based on the real, riskless rate recommended by Gold et al.19 Medical costs included those associated with physician or other healthcare providers, diagnostic procedures, equipment, hospitalizations, and other institutional/facility care costs (Table 1). Pharmacy costs were calculated based on mean daily dosage of IPX066 and CL+E obtained from published trials9,10 and multiplied by the manufacturer’s recommended 2014 Wholesale Acquisition Cost (WAC) for IPX066 ($2.30) and by the WAC price for CL+E ($3.03 and $4.81 for generic and branded, respectively) as referenced from Red Book Average Wholesale Price 2014 (Micromedex Healthcare Series Red Book, version 5.1 [Thomson Reuters (Healthcare) Inc; Montvale, New Jersey]). The 6-month costs of IPX066 and generic and branded CL+E are shown in Table 1.
The key outcomes of interest were total costs for 6 months of treatment with IPX066, as well as for branded and generic CL+E, and incremental cost per quality-adjusted life-year (QALY) gained for IPX066 versus both branded and generic CL+E. The incremental cost-effectiveness ratio (ICER) was calculated for IPX066 relative to (separate) branded and generic CL+E. ICERs were calculated as the difference in cumulative discounted total direct costs of IPX066 and CL+E treatments divided by the difference in cumulative discounted QALYs.
Base-Case Analysis and Sensitivity Analysis
The decision-analytic Markov model was programmed and analyzed using TreeAge Pro 2015 (TreeAge Software, Inc, Williamstown, Massachusetts). In the base-case analysis, the model estimates incremental cost per QALY gained for the 5-year follow-up period of transitioning between “off” states for IPX066 versus generic and branded CL+E. The willingness-to-pay measure was defined at $50,000 per QALY. Multiple one-way sensitivity analyses were performed to measure the impact of uncertainty in individual parameters on the base-case results (Table 1).
The economic evaluation of IPX066 relative to CL+E is presented in Table 2. The average total costs per patient treated with CL+E branded and CL+E generic for the 5-year timeframe were $91,949 and $79,332, respectively; whereas that for IPX066 was $68,703, resulting in incremental cost-savings of $23,246 and $10,629 with IPX066 over branded and generic CL+E, respectively. On average, IPX066 was associated with a 5% increase in QALYs compared with either branded or generic CL+E (2.90 vs 2.76 QALYs) (Table 2). The corresponding ICERs were −$166,044 and −$75,920 per QALY gained for IPX066 versus branded CL+E and generic CL+E, respectively.
Utility inputs were varied based on their literature-reported ranges.12 The other base-case inputs were varied in a uniform distribution by 25% (Table 1). The sensitivity analysis revealed that the total 5-year cost of IPX066 (base case: $68,703) was most impacted by changes in direct medical costs of the >25% “off” time health state, followed by IPX066 drug cost, then the probability of transition from poor health state (>25% “off” time) to improved health state (≤25% “off time), and then the probability of the reverse transition (Figure 2 [a]). This is in contrast to the ICERs for IPX066 versus branded CL+E (Figure 2 [b]) and for IPX066 versus generic CL+E (Figure 2 [c]), where the ICERs were most affected by the utility of the >25% “off” time health state. Figure 3 shows the effect of varying 6-month pharmacy costs of IPX066 on the ICERs. The model estimated that IPX066 is no longer cost-effective relative to generic CL+E and branded CL+E at 6-month pharmacy costs of greater than approximately $5150 and $6650, respectively (Figure 3).
The current 5-year Markov model compared the cost-effectiveness of treating patients with advanced PD with either IPX066 or branded or generic CL+E. Utilizing patient-level data from 2 clinical trials, the current study demonstrated that IPX066 may result in cost savings, along with improved effectiveness, over CL+E, from a US payer’s perspective. Due to the debilitating and costly nature of this disease, the potential for cost savings and improved management of PD are noteworthy. The main driver underlying the cost-effectiveness of IPX066 over CL+E in this study was the utility of the improved health state. This resulted from the greater efficacy seen with IPX066 (a lower percentage of “off” time and higher percentage of “on” time without worsening of troublesome dyskinesia).9 In addition, IPX066 may be favored over CL+E in terms of patient-reported treatment preference.9 Multiple one-way sensitivity analyses demonstrated the robustness of the Markov model, with the dominance of IPX066 being maintained in nearly all analyses.
Given that the ICER of IPX066, versus branded CL+E, was sensitive to health-state utilities, the results of this analysis may vary based on the real-world applicability of these literature-based health-state utility estimates.12 As such, further real-world studies should be conducted to confirm the utility benefit of IPX066 over CL+E. The sensitivity analyses also indicated that future fluctuations in drug pricing for IPX066 and branded and generic CL+E may change the cost-effectiveness outlook over time. However, given that IPX066 demonstrated improved efficacy relative to CL+E,10 even at the same price, IPX066 will likely result in lower costs and improved clinical outcomes.
Other cost-effectiveness analyses for PD treatments utilized similar Markov model designs; 6-month cycles were selected to closely approximate the prior economic evaluations in this drug class, and the health-state utility parameters were based on the percentage per day spent in “off” time.11,12 Our results generally coincide with previous studies, which propose that treatments that reduce the percentage of “off” time would improve the cost of care for PD.12,20,21 Similarly, other models were also sensitive to increases in the amount of “off” time per day.12 However, some differences were seen between our results and those previously published.
A previous cost-effectiveness analysis of branded CL+E in the United Kingdom (Findley et al ) reported higher costs than did our model (£34,200 in 2003 currency [approximately $71,286 in 2014 currency18,21,22] vs $34,095 in the current analysis) and lower overall QALYs gained (1.70 vs 2.76) over 5 years21; however, differences in the structure of the respective health systems may have contributed to these differences. Palmer et al estimated the 5-year cost-effectiveness of adding entacapone to standard treatment (levodopa) and found the ICER to be in the accepted range for healthcare interventions.11 The Palmer et al trial demonstrated that, compared with standard treatment (levodopa), CL+E resulted in an ICER of $9327 per QALY gained.11 Comparatively, the current analysis demonstrated that IPX066 results in an incremental cost of $23,691 per QALY gained compared with $28,734 per QALY gained with generic CL+E, and $33,315 per QALY gained with branded CL+E. Although Palmer et al calculated a slightly lower QALY for CL+E (2.59) compared with that calculated in our analysis (2.76), they were similar. Further, the CL+E transition probability in our analysis corresponds with the range calculated in the Palmer (2002) article (0.336 within 0.208-0.464).11
Differences from previously published reports may stem from the potential limitations in our model. The 6-month transition probabilities for CL+E treatment were based on 2-week treatment data.9 Only 1 transition probability was available from the Stocchi et al trial, which corresponded to a 2-week period of treatment. As a result, all transition probabilities for CL+E were calculated based on a 2-week period. In clinical practice, the actual probability of transitioning from 1 health state to another may vary for PD patients taking CL+E. Additionally, 5-month data were counted as 6 months as an assumption in the transition probabilities calculations for IPX066. The probability of transitioning to a different health state after the 5-month timeframe for patients taking IPX066, therefore, may not be accurately represented in this analysis.
Furthermore, it was assumed that all patients began in the >25% “off” time category before taking the medications, which will not be true for all patients with PD. It was assumed that a decrease in “off” time may only occur in the first 6-month cycle of PD therapy, based on the existing literature.11 Also, the model relies on a somewhat dated reference for the probability of transitioning to a poorer health state during CL+E therapy.11 The 25% “off” time transition data for standard-of-care treatment should be reevaluated in future studies in order to validate the model.
Moreover, in this analysis, mortality was assumed to be independent of PD health states, which may not be an accurate representation. In particular, for patients with a more advanced presentation of the disease, the average lifespan for a person with PD may be impacted by associated complications.23 Additionally, because the data used in this analysis were based on only 2 clinical trials, the favorable results for IPX066 may be limited only to patients who match the study populations found in the 2 clinical trials. In both trials, the majority of patients were male and white.9,10 Lastly, the results of this analysis took the US payer perspective, not incorporating indirect costs, such as patient productivity, so that must also be considered.
Further validation of the model inputs and long-term follow-up through 6 months for CL+E effectiveness would be useful to confirm the results from our model.
IPX066, an extended-release form of CD-LD, is likely to be a cost-effective alternative to CL+E for the treatment of advanced PD in the United States. Based on preliminary results, the model indicates that IPX066 is likely to dominate both branded and generic CL+E, even at a greater 6-month total cost. Further validation of the model inputs, as well as longer-term data describing the transition probabilities of health states for CL+E beyond the current follow-up for CL+E effectiveness, are needed.
The authors thank Melissa Frasco, PhD, previously of Quorum Consulting, Inc, and Tiffany Yu, of Quorum Consulting, Inc, for their contributions to the manuscript development.
Author Affiliations: Quorum Consulting, Inc (RJGA, AL), San Francisco, CA; Department of Environmental Medicine and Public Health, Icahn School of Medicine at Mount Sinai (RJGA), New York, NY; Impax Laboratories, Inc (NRR, SC), Hayward, CA.
Source of Funding: These analyses were supported by Impax Laboratories, Inc. Impax was involved in the interpretation of the study results, and in the writing, review, and approval of the manuscript for submission.
Author Disclosures: Dr Arnold and Mr Layton are employees of Quorum Consulting, Inc, which received a grant from Impax Pharmaceuticals. At the time of the writing of this manuscript, Dr Rustay was an employee of Impax Laboratories, Inc, and held Impax stock. At the time of the writing of this manuscript, Dr Chen was an employee of Impax Laboratories, Inc.
Authorship Information: Concept and design (RJGA, SC); acquisition of data (AL); analysis and interpretation of data (RJGA, AL, NRR); drafting of the manuscript (RJGA, SC, NRR); critical revision of the manuscript for important intellectual content (RJGA, SC, NRR); statistical analysis (RJGA, AL); obtaining funding (RJGA, AL); and supervision (RJGA, AL).
Address Correspondence to: Renée J.G. Arnold, PharmD, RPh, 180 Sansome St, 10th Fl, San Francisco, CA 94104. E-mail: firstname.lastname@example.org.
- Pringsheim T, Jette N, Frolkis A, Steeves TD. The prevalence of Parkinson’s disease: a systematic review and meta-analysis. Mov Disord. 2014;29(13):1583-1590. doi: 10.1002/mds.25945.
- Kowal SL, Dall TM, Chakrabarti R, Storm MV, Jain A. The current and projected economic burden of Parkinson’s disease in the United States. Mov Disord. 2013;28(3):311-318. doi: 10.1002/mds.25292.
- Dams J, Klotsche J, Bornschein B, et al. Mapping the EQ-5D index by UPDRS and PDQ-8 in patients with Parkinson’s disease. Health Qual Life Outcomes. 2013;11:35. doi: 10.1186/1477-7525-11-35.
- Martínez-Martín P, Rodríguez-Blázquez C, Forjaz MJ, et al. Relationship between the MDS-UPDRS domains and the health-related quality of life of Parkinson’s disease patients. Eur J Neurol. 2014;21(3):519-524. doi: 10.1111/ene.12349.
- Ryoo HL, Pierrotti D, Joyce JN. Dopamine D3 receptor is decreased and D2 receptor is elevated in the striatum of Parkinson’s disease. Mov Disord. 1998;13(5):788-797.
- Boiko AN, Batysheva TT, Minaeva NG, et al. Use of the new levodopa agent Stalevo (levodopa/carbidopa/entacapone) in the treatment of Parkinson’s disease in out-patient clinical practice (the START-M open trial). Neurosci Behav Physiol. 2008;38(9):933-936. doi: 10.1007/s11055-008-9085-3.
- Myllylä V, Haapaniemi T, Kaakkola S, et al. Patient satisfaction with switching to Stalevo: an open-label evaluation in PD patients experiencing wearing-off (Simcom Study). Acta Neurol Scand. 2006;114(3):181-186.
- Hauser RA, Panisset M, Abbruzzese G, et al. Double-blind trial of levodopa/carbidopa/entacapone versus levodopa/carbidopa in early Parkinson’s disease. Mov Disord. 2009;24(4):541-550. doi: 10.1002/mds.22343.
- Stocchi F, Hsu A, Khanna S, et al. Comparison of IPX066 with carbidopa-levodopa plus entacapone in advanced PD patients. Parkinsonism Relat Disord. 2014;20(12):1335-1340. doi: 10.1016/j.parkreldis.2014.08.004.
- Waters CH, Nausieda P, Dzyak L, et al. Long-term treatment with extended-release carbidopa—levodopa (IPX066) in early and advanced Parkinson’s disease: a 9-month open-label extension trial. CNS Drugs. 2015;29(4):341-350. doi: 10.1007/s40263-015-0242-2.
- Palmer CS, Nuijten MJ, Schmier JK, Subedi P, Snyder EH. Cost effectiveness of treatment of Parkinson’s disease with entacapone in the United States. Pharmacoeconomics. 2002;20(9):617-628.
- Nuijten MJ, van Iperen P, Palmer C, van Hilten BJ, Snyder E. Cost-effectiveness analysis of entacapone in Parkinson’s disease: a Markov process analysis. Value Health. 2001;4(4):316-328.
- Arias E. United States life tables, 2009. CDC website. https://www.cdc.gov/nchs/data/nvsr/nvsr62/nvsr62_07.pdf. 2009. Accessed August 14, 2014.
- Hauser RA, Ellenbogen AL, Metman LV, et al. Crossover comparison of IPX066 and a standard levodopa formulation in advanced Parkinson’s disease. Mov Disord. 2011;26(12):2246-2252. doi: 10.1002/mds.23861.
- Hauser RA, Hsu A, Kell S, et al; IPX066 ADVANCE-PD Investigators. Extended-release carbidopa-levodopa (IPX066) compared with immediate-release carbidopa-levodopa in patients with Parkinson’s disease and motor fluctuations: a phase 3 randomised, double-blind trial. Lancet Neurol. 2013;12(4):346-356. doi: 10.1016/S1474-4422(13)70025-5.
- Pahwa R, Lyons KE, Hauser RA, et al; APEX-PD Investigators. Randomized trial of IPX066, carbidopa/levodopa extended release, in early Parkinson’s disease. Parkinsonism Relat Disord. 2014;20(2):142-148. doi: 10.1016/j.parkreldis.2013.08.017.
- Dodel RC, Singer M, Köhne-Volland R, et al. The economic impact of Parkinson’s disease. an estimation based on a 3-month prospective analysis. Pharmacoeconomics. 1998;14(3):299-312.
- Consumer price index . US Bureau of Labor Statistics website. http://www.bls.gov/cpi/. Accessed February 2014.
- Weinstein M, Siegel J, Gold M, Kamlet M, Russell L. Recommendations of the panel on cost-effectiveness in health and medicine. JAMA. 1996;276(15):1253-1258.
- LePen C, Wait S, Moutard-Martin F, Dujardin M, Ziegler M. Cost of illness and disease severity in a cohort of French patients with Parkinson’s disease. Pharmacoeconomics. 1999;16(1):59-69.
- Findley LJ, Lees A, Apajasalo M, Pitkänen A, Turunen H. Cost-effectiveness of levodopa/carbidopa/entacapone (Stalevo) compared to standard care in UK Parkinson’s disease patients with wearing-off. Curr Med Res Opin. 2005;21(7):1005-1014.
- Treasury reporting rates of exchange as of March 31, 2016. US Department of the Treasury, Bureau of the Fiscal Service website. https://www.fiscal.treasury.gov/fsreports/rpt/treasRptRateExch/itin-03-31-2016.pdf. Accessed May 9, 2016.
- Frandsen R, Kjellberg J, Ibsen R, Jennum P. Morbidity in early Parkinson’s disease and prior to diagnosis. Brain Behav. 2014;4(3):446-452. doi: 10.1002/brb3.228.