For dopaminergic therapies formulated and/or delivered in efforts to approach or attain CDD, numerous studies have evaluated the impact on motor fluctuations and dyskinesia. Both acute effects and the ability to delay the onset of such complications have been assessed. The impact on nonmotor function has not yet been evaluated extensively.
Prevention of motor complications in early PD
The most recent practice guidelines for treating early PD are those of a consensus document published jointly in 2006 by the European Federation of Neurological Societies (EFNS) and the Movement Disorder Society (MDS) European Section (Horstink et al.
2006). Based on available evidence, this panel judged ropinirole and pramipexole, in their IR formulations, to be effective as monotherapy both for motor-symptom control and for prevention of levodopa-associated motor complications, especially among younger patients, in whom such complications are thought to be more likely. For other dopaminergic therapies, data either were lacking, as in the case of COMT inhibitors and the MAO-B inhibitor rasagiline, or did not support efficacy for motor-complication prevention, as in the case of levodopa CR and the MAO-B inhibitor selegiline.
Since then, long-term (6-year) data from the CALM-PD study of levodopa versus pramipexole IR as initial PD pharmacotherapy have found dopaminergic motor complications (encompassing “wearing-off”, “ON–OFF” effects, or dyskinesias) to be more likely for levodopa than for the dopamine agonist (Parkinson Study Group CALM Cohort Investigators
2009). However, disabling dyskinesias were uncommon in both treatment groups. Long-term (6.5-year) data from an open-label extension (Hauser et al.
2009a) of the TEMPO study (Parkinson Study Group
2004) of early- versus delayed-start rasagiline have failed to demonstrate a difference in the median time to development of motor complications (or time to addition of levodopa) in patients who started receiving rasagiline 6 months earlier in their PD than did the delayed-start group. Data have also been reported concerning the efficacy of COMT inhibitors in early PD. In the 39-week, double-blind FIRST-STEP study (Hauser et al.
2009b), the incidence of “wearing-off” and dyskinesia did not differentiate between groups receiving levodopa/carbidopa/entacapone (LCE) or only levodopa/carbidopa three times daily, despite the superiority of LCE on efficacy measures including the study’s primary outcome, the sum of motor and ADL scores in the UPDRS (but not on measures including the motor score alone). STRIDE-PD (Stocchi et al.
2010), a large-scale, double-blind study designed specifically to evaluate the capacity of LCE to delay dyskinesia, has also failed to find such benefit. For treatment lasting up to 208 weeks, the risk of dyskinesia was actually higher in the study’s LCE group than in its levodopa/carbidopa group. Although the mean dosage of levodopa was highly similar across these groups, the estimated bioavailability of levodopa was significantly heightened by the entacapone in LCE, conceivably hastening dyskinesia (
p < 0.001). The investigators also hypothesized that the study’s four-times-daily dosing of LCE (at 3.5-h intervals) might not have achieved CDS.
Two recent meta-analyses (Baker et al.
2009; Stowe et al.
2008) support the 2006 EFNS/MDS recommendations on dopamine agonists for prevention of motor complications. For oral IR dopamine-agonist treatment compared with levodopa, one study (Stowe et al.
2008) reported an odds ratio (OR) of 0.51 (95 % CI: 0.43, 0.59;
p < 0.00001) for risk of dyskinesia and 0.75 (95 % CI: 0.63, 0.90;
p = 0.002) for risk of motor fluctuations. In the other study (Baker et al.
2009),while the UPDRS motor scores demonstrated that patients receiving dopamine agonists had a significantly inferior response compared with patients receiving levodopa, based on a >4-point higher ADL score (weighted mean difference, 4.69; 95 % CI: 3.76, 5.61;
p < 0.0001), the ORs for risk of dyskinesia and “wearing off” were 0.36 (95 % CI: 0.22, 0.60;
p < 0.0001) and 0.52 (95 % CI: 0.40, 0.66;
p < 0.0001), respectively. Both meta-analyses found that dopamine agonists conferred an increased risk for somnolence, dizziness, nausea, and hallucinations. In a recent double-blind study of adding ropinirole PR in levodopa-treated patients with early PD (Watts et al.
2010), onset of dyskinesia was significantly delayed in the ropinirole PR group compared with a group receiving additional levodopa (hazard ratio, 6.46;
p < 0.001). In a head-to-head comparison between transdermal rotigotine and oral ropinirole IR in early PD (Giladi et al.
2007), the patch failed to demonstrate noninferiority on UPDRS motor plus ADL scores. Differences in rates of motor complications were not assessed.
Treatment of motor complications in advanced PD
In a literature review published in 2006 (Pahwa et al.
2006), a subcommittee of the American Academy of Neurology (AAN) concluded that entacapone and rasagiline had established their efficacy, and that pramipexole, ropinirole, and tolcapone were “probably” effective, for reducing “OFF” time in PD patients with motor fluctuations. For “OFF” time reduction, levodopa/carbidopa CR was not considered to be more effective than the IR formulation. Continuous-infusion therapies (i.e., apomorphine or LCIG) were not included in the analyses.
More recent meta-analyses of treatments adjunctive to levodopa in advanced PD have confirmed the capacities of oral IR dopamine agonists, COMT inhibitors, and MAO-B inhibitors to reduce “OFF” time and UPDRS scores. For all three drug classes, the improvements were at the expense of an increase in dyskinesia, compared with placebo (Stowe et al.
2010; Talati et al.
2009). By indirect comparisons of the three classes, IR dopamine agonists appeared to be the most effective for reducing “OFF” time (at −1.54 h/day vs. −0.83 for COMT inhibitors and −0.93 for MAO-B inhibitors), but the agonists (and the COMT inhibitors) carried a higher risk for dyskinesias than did the MAO-B inhibitors (Stowe et al.
2010).
In recent studies with placebo control (Pahwa et al.
2007; Schapira et al.
2011), slow-release oral formulations of ropinirole and pramipexole, taken adjunctive to levodopa, have also shown efficacy for reducing “OFF” time and UPDRS scores. For both agents, “ON” time without troublesome dyskinesia was significantly increased. However, the incidence of dyskinesia as an adverse event was higher (at 13 % for ropinirole PR vs. 3 % for placebo; and at 17 % for pramipexole ER vs. 8 % for placebo, a difference not tested statistically in either trial). In a recent 24-week study comparing ropinirole formulations (Stocchi et al.
2011), the slow-release form had a response rate (defined by ≥20 % reduction in “OFF” time; adjusted OR, 1.82; 95 % CI: 1.16, 2.86;
p = 0.009) and a capacity to reduce UPDRS motor scores (adjusted mean change from baseline for ropinirole PR vs IR, −10.2 vs. −7.9, respectively;
p = 0.022) significantly greater than those for the IR form. The incidence of adverse events was numerically higher among recipients of the slow-release form (who also reached higher agonist dosage, but lower levodopa dosage, than in the IR arm). In an analogous comparison (Schapira et al.
2011), pramipexole ER showed capacities to reduce “OFF” time and UPDRS motor plus ADL scores resembling those of pramipexole IR. The incidence of adverse events was numerically lower for the ER form (at similar agonist dosage and lower levodopa dosage than in the IR arm).
In each of two studies with placebo control (LeWitt et al.
2007; Poewe et al.
2007), transdermal rotigotine, adjunctive to levodopa, has shown efficacy for motor complications in advanced PD, as evidenced by significant reductions in “OFF” time (
p ≤ 0.0031) and increases in “ON” time without troublesome dyskinesia (
p ≤ 0.0078). Incidence rates for dyskinesia as an adverse event were higher for rotigotine than for placebo (but were not tested statistically). In one of these studies (Poewe et al.
2007), a third treatment arm permitted head-to-head comparisons between the rotigotine patch and pramipexole IR. Improvements in “OFF” time, “ON” time without troublesome dyskinesias, and other outcomes showed no significant differences between these two treatments.
No double-blind studies have yet been reported for chronic treatment with apomorphine. However, in 2004 a review of 11 long-term, uncontrolled, open-label studies of subcutaneous apomorphine infusion (Deleu et al.
2004) reported a 60 % mean reduction in “OFF” time (range 50–80 % across studies) and also improvement in dyskinesia, which, however, required a mean of 12 months (range 0.5–50 months) to reach its maximum. Two further studies of subcutaneous apomorphine infusion have since been reported. In a prospective study (Katzenschlager et al.
2005), 12 patients with motor fluctuations and disabling dyskinesias received apomorphine infusion for 6 months. “OFF” time was reduced from baseline by 38 % (
p < 0.05) and dyskinesia duration by 40 % (
p < 0.01). In four of the 12 patients, oral medication could be discontinued, and within this apomorphine monotherapy group, there was a significantly greater decrease in “OFF” time and dyskinesia severity and duration than in the polytherapy group (for “OFF” time, 64 vs. 18 %, respectively,
p < 0.05; for dyskinesia severity and duration expressed as centimeters on a visual analogue scale, 8.6 vs. 19.8, respectively,
p < 0.05). Skin nodules were reported in 11 of 12 patients, including two patients with skin changes and inflammatory reactions requiring rotation of the infusion site. A retrospective analysis (García-Ruiz et al.
2008) of 82 patients who tolerated subcutaneous apomorphine infusion for at least 3 months (mean, 20 months) also identified significant decreases from baseline in “OFF” time (mean baseline vs. last follow-up visit, 6.64 vs. 1.36 h/waking day,
p < 0.0001) and dyskinesia severity score (1.65 vs. 1.15, respectively,
p < 0.0006). In this study, 68 % of patients reported treatment-related skin nodules.
For LCIG, findings of a double-blind, double-dummy trial comparing intrajejunal gel infusion and oral administration of levodopa-carbidopa in IR form have now been presented. For 12 weeks, PD patients selected for having motor complications underwent active LCIG infusion and took placebo IR capsules or took active IR capsules and underwent placebo gel infusion (Olanow et al.
2011). Among 66 study completers (93 % of the 71 randomized subjects), decrease in “OFF” time and increase in “ON” time without troublesome dyskinesia favored LCIG by means of −1.91 and +1.86 h/d, respectively, while “ON” time with troublesome dyskinesia showed no significant change (Olanow et al.
2012). Significant global, functional (ADL), and quality-of-life LCIG benefits were also identified (Kieburtz et al.
2012).
Several previous, open-label studies had already evaluated LCIG during chronic treatment lasting up to 7 years (Table
1) (Antonini et al.
2007,
2008,
2010b; Eggert et al.
2008; Isacson et al.
2008; Nilsson et al.
2001). Although variations in trial design preclude any summarization of the numerical findings, measures of “OFF” time improved in all six studies, with significance achieved in all five studies in which statistical testing was performed; dyskinesia measures also improved in all six studies, with statistical significance in four of the five studies with statistical testing. In addition, a randomized crossover trial (Nyholm et al.
2005) has compared 3 weeks of individually optimized conventional treatment with 3 weeks of nasoduodenal LCIG infusion. The conventional treatment was oral levodopa/carbidopa in optional combinations with oral dopamine agonists, COMT-inhibitors, MAO-B inhibitors, amantadine, or subcutaneous apomorphine (by injections or infusion). Motor tasks were videotaped every 30 min for 8 h on 2 days during the second and third week of each treatment for rating of motor function by neurologists blinded to treatment identity. (To blind the conventional therapies, a dummy nasogastric tube was emplaced.) Of 24 enrolled patients, 20 completed conventional therapy and 21 completed LCIG. The mean percentage of ratings falling within the predefined range for a clinically desirable “ON” state was significantly greater during LCIG than during conventional therapy (at 90.7 vs. 74.5 %, respectively,
p < 0.01), and the mean percentage of ratings of “OFF” state was significantly lower (at 1.8 vs. 19.2 %, respectively,
p < 0.01). For ratings of “ON” with moderate-to-severe dyskinesia, LCIG and conventional therapy showed no significant difference. Of the four patients who received subcutaneous apomorphine infusion as part of their conventional therapy, two were rated as being in an “ON” state all or nearly all of the time on both conventional treatment and LCIG, but the other two had substantial improvements on LCIG. In one patient, the proportion of “ON” ratings improved from 56 to 94 %, and in the other from 68 to 100 %.
Table 1
“OFF” time and dyskinesia outcomes in prospective studies of long-term LCIG therapy
| 9/6a
| 7 years | “OFF” time duration decreased at 3–8 months and at 4–7 yearsb. Dyskinesia duration decreased at 3–8 months and further decreased at 4–7 years |
| 9/7 | 1 year | “OFF” time and disabling-dyskinesia duration significantly decreased |
| 22/17 | 2 years | “OFF”-time duration and severity (by UPDRS part IV) significantly decreased |
| 13/11 | 6 monthsc
| “OFF” time and “ON” time with disabling dyskinesia significantly reduced as percent of patient’s day |
| 14/12 | 6 months | Proportion of patients with reduced “OFF” time significantly greater; proportion with reduced dyskinesia numerically greater |
| 15/4 | 3 years | “OFF” time duration and dyskinesia severity (by UPDRS part IV) significantly decreased |
The most common adverse events leading to discontinuation of chronic LCIG treatment have been related to medication delivery, including problems with tubing displacement or occlusion, the pump system, and stoma infections (Antonini et al.
2007,
2008,
2010b; Eggert et al.
2008; Nilsson et al.
2001). To monitor long-term safety, close observation is warranted, as guided by recent reports of neuropathy associated with homocysteine elevation and at least functional vitamin B12 and/or B6 deficiency (Klostermann et al.
2012).
Treatment of nonmotor symptoms
In general, published management recommendations for nonmotor PD symptoms have not attempted to differentiate between symptoms intrinsic to PD and those arising as complications of PD therapy. A contributory problem may be that in clinical studies of dopaminergic options, improvement of nonmotor PD symptoms, levodopa-related or not, has seldom been either a primary endpoint or a means for comparing treatments with potentially different abilities to produce CDD. In a double-blind, 12-week, parallel-group study of patients with early PD (Fung et al.
2009), oral LCE was superior to levodopa/carbidopa for improving quality of life, as rated by total score on the eight-item PDQ (PDQ-8) (Jenkinson et al.
1997b) and by subscores for the PDQ-8 nonmotor domains of depression, personal relationships, communication, and stigma. In a small single-dose, crossover study of oral levodopa/carbidopa CR versus IR (Kulisevsky et al.
2007), the seven PD patients with “wearing-off” showed significant mood improvement after IR treatment, while the seven nonfluctuating patients showed no mood improvement after receiving either drug. Plasma levodopa concentration correlated with anxiety level but not with mood.
Transdermal rotigotine has recently been assessed versus placebo in a large, double-blind trial (Trenkwalder et al.
2011a) in PD patients with unsatisfactory early morning motor-symptom control. On the PDSS-2 (Trenkwalder et al.
2011b), the mean 12-week change in total score (a coprimary outcome) was significantly improved in the rotigotine group compared with placebo from baseline to end of maintenance [least squares (LS) mean treatment difference, −4.26; 95 % CI: −6.08, −2.45;
p < 0.0001]. Difficulty in falling asleep and feeling tired and sleepy in the morning were among the ten items showing significant improvement (among the instrument’s total of 15). Significant improvement versus placebo was also documented for mean change in total score on the NMSS (Martinez-Martin et al.
2009) from baseline to end of treatment (LS mean treatment difference, −6.65; 95 % CI: −11.99, −1.31;
p = 0.015), with significant changes on the sleep/fatigue and the mood/cognition domains (LS mean treatment differences of −2.03; 95 % CI: −3.31, −0.75;
p = 0.002 and −3.40; 95 % CI: −5.22, −1.58;
p = 0.0003, respectively), and on the BDI-II depressive-symptomatology scale (Visser et al.
2006) from baseline to end of treatment (LS mean treatment difference, −2.01; 95 % CI: −3.55, −0.47;
p = 0.011).
Improvement of nonmotor PD symptoms by subcutaneous apomorphine infusion has been a primary endpoint in two small prospective trials (Martinez-Martin et al.
2010; Reuter et al.
1999). In six patients with refractory nocturnal symptoms (Reuter et al.
1999), continuous overnight infusion (with placebo control in three of the subjects) was associated with decreases in awakenings, dystonia, pain, and nocturia. In an open-label trial (Martinez-Martin et al.
2010), 17 patients with advanced PD received subcutaneous apomorphine infusion and 17 received conventional therapy. At approximately 6 months, NMSS and PDQ-8 total scores showed significant improvement from baseline in the apomorphine group but did not change in the conventional-therapy group.
Improvement of nonmotor PD symptoms by intraduodenal LCIG infusion was the primary endpoint in a prospective, open-label trial conducted in 22 patients with daily motor fluctuations and dyskinesia refractory to optimized conventional therapy with oral medications, transdermal rotigotine, or subcutaneous apomorphine infusion (Honig et al.
2009). After discontinuation of the conventional therapy (except for nighttime oral dosing with levodopa CR or a long-acting dopamine agonist) and its replacement by LCIG for a mean of 6.7 months, mean change in NMSS total score showed significant improvement from baseline (−50.5,
p = 0.0001). Of the nine NMSS domains, six were significantly improved [cardiovascular (−2.41,
p = 0.0004), sleep/fatigue (−11.32,
p = 0.0001), attention/memory (−3.27,
p = 0.002), gastrointestinal (−6.23,
p = 0.0003), urinary (−6.64,
p = 0.002), and miscellaneous (−7.73,
p = 0.0004)]. The change in total NMSS score was correlated with changes on measures of motor function [UPDRS motor-complication score (−5.91,
p = 0.0000), UPDRS dyskinesia subscore (−3.7,
p = 0.0001), and “OFF” time as a proportion of the waking day (
r = 0.54,
p < 0.01)]. On the PDSS and PDQ-8, mean improvement in total score was also significant (+28.51;
p = 0.002 and −23.4;
p = 0.0003, respectively).