Included Studies
New clinical evidence available since the previous CDR review of Duodopa for PD comprised 20 trials.
Double-Blind, Randomized Controlled Trial
Study 001/002 (N = 71) was a DB, double-dummy, active-controlled, multi-centre, multinational, phase III superiority randomized controlled trial (RCT) that recruited patients from North America (excluding Canada). The study objective was to evaluate the efficacy and safety of LCIG for the treatment of patients with advanced levodopa-responsive PD who do not have satisfactory control of severe, debilitating motor fluctuations and hyper/dyskinesia despite optimized treatment with available combinations of Parkinson medicinal products, and for whom the benefits of this treatment may outweigh the risks associated with the insertion and long-term use of the PEG-J tube required for administration. Patients were randomized to a 1:1 ratio of optimally titrated LCIG (20 mg/mL levodopa and 5 mg/mL carbidopa monohydrate solution) in addition to placebo immediate-release (IR) OLC capsules or optimally titrated IR OLC 100 mg/25 mg capsules in addition to placebo LCIG. The primary efficacy outcome was change from baseline to final visit (week 12) in the mean number of “off” hours recorded in the Parkinson disease home diary (PDHD) during the three consecutive days prior to study visit, normalized to a 16-hour waking day. The predefined key secondary outcome was change from baseline to final visit (week 12) in the mean number of normalized (16-hour waking day) “on” hours without troublesome dyskinesia (defined as a composite of “on” time without dyskinesia and “on” time with non-troublesome dyskinesia). Other secondary outcomes included change from baseline in the 39-item Parkinson’s Disease Questionnaire (PDQ-39) summary index score, the Clinician Global Impression – Improvement (CGI-I) score, the Unified Parkinson’s Disease Rating Scale (UPDRS) Part II (activities of daily living [ADL] subscore) and Part III (motor subscore), the Zarit Burden Interview (ZBI) score, and the EuroQol 5-Dimensions 3-Levels (EQ-5D-3L) summary index score.
Although some methodological limitations were highlighted, no major limitations were identified in Study 001/002. Some of the noted limitations included differences in PD severity between treatment groups at baseline, the potential for unblinding due to non-compliance with active IR OLC capsules in the comparator group, and the lack of information about the adequacy of caregiver report as a proxy for patient-reported “off” time.
Open-Label, Non-Comparative Study (Study 004)
Study 004 (N = 354) was a non-comparative, multinational, multi-centre, open-label, long-term safety study that recruited patients from North America (including Canada) and western Europe. Patients included in this study were not previously treated with LCIG in Study 001/002. The study objective was to evaluate the safety of LCIG for the treatment of patients with advanced levodopa-responsive PD over a 54-week period. LCIG was delivered as an aqueous solution containing 20 mg/mL levodopa and 5 mg/mL carbidopa monohydrate packaged in 100 mL cassettes administered as a morning bolus dose followed by continuous infusion at a constant rate for the remainder of each patient’s waking day (approximately 16 hours) with additional rescue doses during the day, if clinically indicated. The primary objective was to evaluate the long-term safety of LCIG based on adverse events (AEs), device complications, and number of completers. All AEs were considered as treatment-emergent adverse events (TEAEs), defined as those that began or worsened from the time of nasojejunal (NJ) tube insertion until 30 days after PEG-J removal. Long-term efficacy (as measured by “off” time, “on” time with and without troublesome dyskinesia, and UPDRS) and quality of life (QoL) (as measured by PDQ-39, EQ-5D-3L, EuroQol visual analogue scale [EQ VAS], and CGI-I) were evaluated as secondary end points. Key limitations of Study 004 include its open-label and non-comparative study design.
Additional Open-Label, Non-Comparative Studies
Patients who completed Study 001/002 had the option to enrol in an optional 12-month open-label safety extension study (Study 003, N = 62). Furthermore, patients who completed either Study 003 or Study 004 were able to enrol in Study 005 (N = 262) for up to five years of follow-up. Study 003 and Study 005 are both summarized in Appendix 6. A total of 16 other prospective, open-label, non-comparative trials were also identified in the CDR systematic review.10–25 The sample sizes of these trials ranged between nine and 375 enrolled patients; follow-ups ranged between four and 36 months. Treatment with LCIG was administered to all patients. Change from baseline to the last study visits was evaluated for relevant outcomes. Key limitations of the trials (including Study 003 and Study 005) include their open-label and non-comparative study design.
Efficacy
Double-Blind, Randomized Controlled Trial
Compared with IR OLC, LCIG was associated with a statistically significant reduction in daily normalized “off” time at week 12 (the primary outcome) completed through the PDHD. The adjusted least squares mean difference (LSMD) in change from baseline was −1.91 hours (95% CI, −3.05 to −0.76; P = 0.0015) in favour of LCIG. Furthermore, results of the sensitivity analyses (using mixed-effects models for repeated measures to impute data and sensitivity analyses with varying covariates requested by the FDA) were mostly consistent with the primary analysis for this outcome. Given that the benefit associated with treatment with LCIG exceeds the reported minimal clinically important difference (MCID) (−1.00 hours), the improvement in “off” time reported in Study 001/002 would be considered clinically meaningful.
The evaluation of “off” time as the primary end point in Study 001/002 was supported by the evaluation of a key secondary end point (adjusted for multiple statistical testing), normalized “on” time without troublesome dyskinesia (a composite of “on” time without dyskinesia and “on” time with non-troublesome dyskinesia). Overall, LCIG was also associated with a statistically significant improvement in daily normalized “on” time without troublesome dyskinesia at week 12. The adjusted LSMD in change from baseline was 1.86 hours (95% CI, 0.56 to 3.17; P = 0.0059) in favour of the LCIG. When looking at the two components of the composite separately, it appeared that the results were driven primarily by the increase in “on” time without dyskinesia (adjusted LSMD in change from baseline: 2.28 hours [95% CI, 0.47 to 4.09; P = 0.0142]), since the change in “on” time with non-troublesome dyskinesia was not statistically significant (−0.73 [95% CI, −2.22 to 0.76; P = 0.3294]). Overall, the result of the key secondary outcome was also consistent with the primary analysis; however, no MCID was identified for the change in “on” time. No adjustments for multiple statistical testing were made for the individual components of this end point.
Other secondary outcome measures (adjusted for multiple statistical testing) included change from baseline in PDQ-39 summary index score, CGI-I score, UPDRS Part II (ADL subscore), UPDRS Part III (motor subscore), EQ-5D-3L summary index score, and the ZBI score. Compared with IR OLC, LCIG was associated with a statistically significant and clinically meaningful reduction in favour of the study drug for both the PDQ-39 summary index score (adjusted LSMD in change from baseline was −7.0 [95% CI, −12.6 to −1.4; P = 0.0155] compared with an MCID of −1.6) and the UPDRS Part II score (adjusted LSMD in change from baseline was −3.0 [95% CI, −5.3 to −0.8; P = 0.0086] compared with an MCID of −2.3). The results for the CGI-I score were also statistically significant in favour of the LCIG (adjusted LSMD in change from baseline was −0.7 [95% CI, −1.4 to −0.1; P = 0.0258]); however, no MCID was identified. Therefore, the clinical meaningfulness of the change in CGI-I remains unclear. No statistically significant differences between treatments were reported for the UPDRS Part III (adjusted LSMD in change from baseline was 1.4 [95% CI, −2.8 to 5.6; P = 0.5020]).
Study 001/002 also evaluated health-related quality of life (HRQoL) using the EQ-5D-3L summary index score; the adjusted LSMD in change from baseline was 0.07 [95% CI, −0.01 to 0.15; P = 0.0670]). For caregiver burden using the ZBI score, the adjusted LSMD in change from baseline was −4.5 [95% CI, −10.7 to 1.7; P = 0.1501]). These end points should be considered exploratory despite being part of the testing hierarchy given that they were evaluated subsequent to failure of a prior end point in the hierarchy (UPDRS Part III); therefore the clinical importance of these changes also remains unclear.
Changes in the mean daily normalized “off” time and “on” time without troublesome dyskinesia at week 54 compared with baseline in Study 004 were −4.4 hours (2.9, P < 0.001) and 4.8 hours (3.4, P < 0.001), respectively. The mean change in “on” time with troublesome dyskinesia was −0.4 (2.8, P = 0.023). Changes in the mean UPDRS Part II score, PDQ-39 summary index score, EQ-5D summary index score, and EQ VAS were −4.4 (6.5, P < 0.001), −6.9 (14.1, P < 0.001), 0.064 (0.203, P < 0.001), and 14.0 (24.8, P < 0.001), respectively. Efficacy outcomes in Study 004 were not adjusted for multiple statistical comparisons.
Harms
Overall, 95% and 100% of patients experienced AEs in the LCIG + placebo (PBO) IR OLC capsules group and the PBO LCIG + IR OLC capsules group, respectively. The frequencies of AEs were relatively similar across treatment groups. The most common AEs were falls (11% versus 12%), atelectasis (8% versus 0%), anxiety (8% versus 3%), confusional state (8% versus 3%), oedema peripheral (8% versus 0%), oropharyngeal pain (8% versus 0%), and upper respiratory tract infection (8% versus 0%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively. Fewer patients experienced serious adverse events (SAEs) in the LCIG + PBO IR OLC capsules group compared with the PBO LCIG + IR OLC capsules group (14% versus 21%, respectively). The most common SAEs were confusional state (5% versus 0%) and pneumonia (0% versus 6%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively.
Overall, one patient (3%) and two patients (6%), respectively, withdrew due to AEs in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups. The most common reasons were hallucination and psychotic disorder (3% versus 0%, each) and peritonitis, post-procedural complication, and post-procedural discharge (0% versus 3%, each) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively. No deaths were reported in Study 001/002.
Most patients experienced device-related complications across both treatment groups (92% and 85% in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively.) Overall, 76% compared with 79% of patients and 57% compared with 56% of patients experienced long-term complications of PEG-J and risks of PEG-J insertion in the LCIG + PBO IR OLC capsules group and PBO LCIG + IR OLC capsules group, respectively. The most common long-term complications of PEG-J were complication of device insertion (57% versus 44%), procedural pain (30% versus 35%), and incision-site erythema (19% versus 12%), while the most common risks of PEG-J insertion were abdominal pain (51% versus 32%) and pneumoperitoneum (11% versus 3%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively.
In general, a similar number of patients (70% compared with 71%) experienced gastrointestinal (GI) AEs, the most common being nausea (30% versus 21%), constipation (22% versus 21%), and flatulence (16% versus 12%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively. More patients experienced psychiatric disorders in the LCIG + PBO IR OLC capsules group compared with the PBO LCIG + IR OLC capsules group (46% compared with 29%). The most common were depression (11% versus 3%), insomnia (11% versus 12%), anxiety (8% versus 3%), and confusional state (8% versus 3%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively. A total of 3% and 9% of patients experienced polyneuropathy and associated signs and symptoms, the most common reason being balance disorder (3% compared with 6%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively. Fewer patients experienced nervous system disorders in the LCIG + PBO IR OLC capsules group compared with the PBO LCIG + IR OLC capsules group (30% compared with 47%). The most common were dyskinesia (14% versus 12%), dizziness (8% versus 6%), and headache (8% versus 12%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively. In general, a similar number of patients (22% compared with 27%) experienced vascular disorders, the most common being orthostatic hypotension (14% versus 24%) and hypertension (8% versus 0%) in the LCIG + PBO IR OLC capsules and PBO LCIG + IR OLC capsules groups, respectively.
Generally, AEs (including serious and non-serious) reported in Study 001/002 were consistent with the known AE profile of levodopa/carbidopa (e.g., depression, anxiety, confusion) and PD patients who have undergone the PEG-J procedure.
The primary objectives of Study 003 (N = 62), Study 004 (N = 354), and Study 005 (N = 262) as well as 16 other prospective, open-label, non-comparative trials were to evaluate the long-term safety of LCIG. Overall, the safety profile was generally consistent with that identified in Study 001/002, with no new safety signals identified following up to 60 months of treatment.
Potential Place in Therapy1
Since its introduction into clinical practice almost 50 years ago, levodopa remains the most effective treatment for the motor manifestations of PD. However, levodopa is only a symptomatic treatment. It does not slow the underlying neurodegenerative process in PD, and the number of functioning nigrostriatal pathway neurons continues to decline. As the number of remaining functioning nigrostriatal neurons falls, the midbrain’s ability to convert levodopa to dopamine (and thereby stimulate the striatum) becomes increasingly impaired. Clinically, this decline in the nigrostriatal neuron population is experienced by patients as a gradual transition from the initial months or years in which levodopa produces a sustained, continuous improvement in motor function to a state in which individual doses of levodopa produce increasingly shorter periods of improvement that wear off quickly. As PD advances, patients increasingly alternate between “on” periods, when they are mobile, and “off” periods, when they are immobile. Generally, the fluctuation between the “on” and “off” states can be related to when individual doses of levodopa are administered. To some extent, these fluctuations can be minimized by spacing levodopa doses closer together and using additional drugs, such as sustained-release levodopa preparations, drugs that inhibit the metabolism of levodopa, or direct dopamine agonists (the latter generally have a longer duration of action than levodopa, but are also generally less effective). For relatively rapid relief of fluctuations, particularly those that occur unpredictably, injectable apomorphine is another option. Although patients can learn to adapt to these fluctuations to some extent, the fluctuations can be unpredictable, severe, and have a major impact on patients’ abilities to carry out ADL.
A key limitation of oral pharmacological strategies for managing “on” and “off” fluctuations is the suboptimal absorption of the drug from the GI tract. This is particularly problematic for levodopa, which must compete with other small amino acids to gain access to the same small amino-acid transporter in the gastric mucosa in order to pass from the lumen of the stomach into the bloodstream. LCIG alleviates this problem because levodopa is delivered directly to the jejunum by the use of a jejunostomy tube. In addition to improving the reliability of levodopa absorption, the jejunostomy tube allows the patient to absorb levodopa at a more or less constant rate. This implies that levodopa can be delivered to the brain at a relatively constant rate, which is presumed to more closely resemble the normal physiological state. In patients treated with LCIG, it is generally possible to reduce or even discontinue the oral medications the patient was previously receiving.
The patient most likely to benefit from LCIG is one with moderately advanced levodopa-responsive PD (disabled, but ambulatory and at least semi-independent), whose waking hours are characterized by frequent fluctuations between the “on” and “off” states despite receiving optimized therapy with existing drugs. Identification of such patients would be part of routine neurological follow-up. In some centres, where neurosurgical expertise is available, deep brain stimulation might be considered an option in such patients. In some instances, the patient’s wishes or general medical condition may make either deep brain stimulation or jejunostomy tube placement impossible, and the only option may be to continue on optimized oral medication. The potential benefits of LCIG would need to be weighed against the inconvenience and potential complications of insertion and living with a jejunostomy tube and infusion pump. The jejunostomy tube insertion requires collaboration with an endoscopist (gastroenterologist or surgeon), implying added cost to the health care system, and follow-up of patients requires some expertise with the maintenance of the infusion pump. Otherwise, the use of Duodopa would not require any new or specific diagnostic testing. At follow-up visits, the patient’s functional status would be assessed to ensure that LCIG is still providing benefit. (If in doubt, the infusion rate could be reduced and the impact on function observed directly, usually during a day-long clinic visit.)
