Patient disposition flowchart.
Distribution of the therapeutically equivalent dose (or maximum dose achieved) by treatment group. One placebo-treated patient dropped out after receiving the 0.6-mL dose, declaring a lack of effect. The average (±SEM) apomorphine hydrochloride dose was 5.4 ± 0.5 mg. The distribution was statistically significantly different between the groups (P<.001, Mantel-Haenszel χ2 test). For the active apomorphine group, milligrams are 10 times the milliliter values.
Regression analysis of therapeutically equivalent single doses of apomorphine hydrochloride and oral levodopa. The regression equation is: apomorphine dose in milligrams = 0.0062 (levodopa AM dose+4.0 mg); R2 = 0.05, P = .35. The regression equation for the ampomorphine dose vs the total daily levodopa dose (not plotted) is: apomorphine dose in milligrams = 0.0015 (levodopa daily dose+4.2 mg); R2 = 0.05, P = .32. The 20 data points are reflected with 3 superimposed outcomes.
Correlation of inpatient vs outpatient response. Linear regression of outpatient-inpatient response on inpatient response shows a statistically significant correlation, R2 = 0.54, P<.001. Categorical analysis is more appropriate since outpatient scores cluster around a single score of 100% relief. With a 2:1 preponderance of apomorphine, the 67th percentile is selected to categorize responses as successful or not. The Fisher exact test of the resulting contingency table is statistically significant at P<.001. UPDRS indicates United Parkinson Disease Rating Scale.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Dewey, Jr RB, Hutton JT, LeWitt PA, Factor SA. A Randomized, Double-blind, Placebo-Controlled Trial of Subcutaneously Injected Apomorphine for Parkinsonian Off-State Events. Arch Neurol. 2001;58(9):1385–1392. doi:10.1001/archneur.58.9.1385
To assess the safety and efficacy of subcutaneous apomorphine hydrochloride administration for off-state (poor motor function) periods in patients with Parkinson disease with motor fluctuations under both inpatient titration and outpatient therapeutic conditions.
Patients and Methods
Twenty-nine patients had advanced Parkinson disease with 2 hours or more off time despite aggressive oral therapy. Patients randomly received titrated doses of subcutaneous apomorphine hydrochloride (2-10 mg, n = 20) or pH-matched vehicle placebo (n = 9) during an inpatient and 1-month outpatient phase. A change in the United Parkinson Disease Rating Scale motor score 20 minutes after inpatient dosing during a practically defined off-state event and the percentage of injections successfully aborting off-state events were the primary inpatient and outpatient efficacy factors.
The average (SEM) levodopa equivalent dose of apomorphine hydrochloride was 5.4 ± 0.5 mg and the mean placebo dose was 1.0 mL. Mean inpatient United Parkinson Disease Rating Scale motor scores were reduced by 23.9 and 0.1 points (62% and 1%) by apomorphine treatment and placebo, respectively (P<.001). The mean percentage of outpatient injections resulting in successful abortion of off-state events was 95% for apomorphine and 23% for placebo (P<.001). Inpatient response was significantly correlated with and predictive of outpatient efficacy (P<.001). The levodopa dose was not predictive of the apomorphine dose requirement. Frequent adverse events included dyskinesia, yawning, and injection site reactions.
Apomorphine by intermittent subcutaneous injection is effective and safe for outpatient use to reverse off-state events that occur despite optimized oral therapy.
MOTOR fluctations are common in patients with advanced Parkinson disease (PD) whose symptoms are managed with oral levodopa and represent some of the most disabling complications of the disorder. The transition from good motor (on-state) function to that of poor motor (off-state) function occurs when brain levodopa levels fall below the threshold needed to adequately stimulate striatal dopamine receptors. Since the progression of PD is typically accompanied by progressive shortening of the clinical effect from each dose of levodopa,1 off states often occur unless strategies are used that can provide a more continuous stimulation of dopamine receptors. A number of such strategies have been developed including the use of dopamine agonists, catechol O-methyl transferase inhibitors, and controlled-release formulations of levodopa. While these strategies are often helpful, some patients continue to suffer from episodic, often unpredictable, off states.
There are no agents available in the United States that can provide a rapid reversal of an individual off state. This "rescue" property is highly desirable since off states can be very disabling to some patients precipitating immobility, panic attacks, pain, screaming, or drenching sweats.2-4
Apomorphine is a direct-acting dopamine agonist with strong D1 and D2 dopamine receptor–stimulating properties that is administered by a parenteral route (intravenously, rectally, subcutaneously, sublingually, or intranasally). It has similar efficacy to levodopa with a substantially more rapid time to onset.5 While an extensive literature exists for apomorphine therapy for PD, only a few studies have been conducted as randomized, placebo-controlled trials using subcutaneous injections of apomorphine to abort off-state events.6-15 To our knowledge, only a single study has been previously published evaluating the drug in both inpatient and outpatient settings; in the outpatient setting, all patients received active drug without a placebo comparator.7 To date there have been no published studies attempting to correlate inpatient efficacy measures with outpatient results. The present study was performed to (1) establish the efficacy and safety of subcutaneous apomorphine injections in a therapeutic setting, and (2) establish whether inpatient efficacy measures accurately predict outpatient responses when both assessments are performed in a placebo-controlled, double-blind fashion.
This prospective randomized trial was approved by institutional review boards of each of the 4 participating US sites, and all patients provided written informed consent. Enrollment was limited to patients with advanced idiopathic PD suffering from motor fluctuations. At least 2 hours of off time per day despite an optimized oral drug regimen, including levodopa and an oral dopamine agonist, was required. Exclusion criteria included atypical parkinsonism, psychosis, dementia, drug or alcohol dependency, previous stereotactic brain surgery for PD, unstable medical illnesses, or previous exposure to apomorphine. A significant improvement in the United Parkinson Disease Rating Scale (UPDRS) motor score (part III) after administration of oral levodopa was required (an expected UPDRS motor score improvement of ≥30% response was defined by protocol, but a single patient with a levodopa response of 24% was allowed to randomize).
The study was a prospective, randomized, double-blind, placebo-controlled, parallel group trial involving 2 phases. Phase 1 consisted of inpatient observation of upwardly titrated doses given to reverse a practically defined off state achieved by withholding antiparkinsonian drugs overnight. Phase 2 involved a 1-month period of outpatient observation of drug effectiveness when administered by patients or caregivers as needed for reversal of spontaneous off-state events (wearing-off or on-off). Prior to the inpatient phase, patients were observed for at least 14 days as outpatients to establish the number of off hours present per day at baseline. On the first day of the inpatient phase, all patients were tested under unblinded conditions for their response to their typical morning dose of oral levodopa (levodopa challenge).
Apomorphine hydrochloride, 10 mg/mL (Bertek Pharmaceuticals Inc, a subsidiary of Mylan Laboratories Inc, Pittsburgh, Pa), was compared with its pH 3.5–matched vehicle placebo. The study drug was packaged in sealed glass ampules, and at the time of use, was drawn up into a 1-mL insulin syringe. Upward titration of apomorphine hydrochloride or placebo was begun at 0.2 mL with 0.2-mL increments to a maximum of 1.0 mL (2-10 mg active) per dose. Titration was terminated at 1.0 mL or on demonstration of a UPDRS motor score reduction of at least 90% of that recorded during the levodopa challenge. If a dose of apomorphine or placebo had not generated a levodopa-equivalent response, a second or third dose could be tested on a single day at intervals of not less than 2 hours.
During the 4-week outpatient phase, patients continued either active apomorphine or placebo injections as previously randomized in addition to their regular oral PD medications at the same baseline dosage and frequency. Patients or caregivers could draw up several doses of the study drug into syringes at the start of each day that were carried with them wherever they went. The initial outpatient dose of apomorphine or placebo was the highest dose achieved during titration, with the option to adjust the dose once after 2 weeks. Outpatient doses were allowed up to 5 times daily as needed to reverse off-state events. Outpatient injections were not administered if the off-state event occurred within 1 hour of the last dose of oral medication or within 15 minutes of the next scheduled dose of oral drug. Trimethobenzamide hydrochloride, 250 mg thrice daily, initiated at least 3 days prior to the first possible dose of apomorphine, was used to minimize the risk of nausea.16
After overnight PD medication withdrawal, patients were evaluated in the practically defined off state using the UPDRS motor score as the primary outcome measure supplemented with the hand tapping score,5 the Webster step-seconds score,17 and the dyskinesia score.18 Motor assessments were conducted prior to dosing and were repeated when a clinical on state occurred or within 60 minutes after the levodopa challenge and 15 to 20 minutes after study drug injection. The primary efficacy factor was the predose to postdose change in UPDRS motor score assessed at the highest titration dose achieved.
During the 4-week outpatient phase, patients completed daily diaries. For each injection, patients recorded whether the off-state event was aborted, the latency to perceived response, and the presence and severity of dyskinesia and nausea and/or vomiting. Also recorded were the total awake time, time off, time on, time on with dyskinesias, time of meals, and time of oral medication ingestion.
Dyskinesia was specifically recorded by the investigator when seen following levodopa challenge or injected study drug during the inpatient phase, and by the patient (via diary entry) when occurring following study drug injection in the outpatient phase. In addition, an adverse event of dyskinesia was reported for patients developing more severe dyskinesias than prior to study. All other adverse events were queried using open-ended questions. Routine clinical laboratory tests, electrocardiograms, and physical examinations were also conducted at study enrollment and exit.
A 2:1 preponderance of active drug to placebo assignment was designed. Sample size estimation assumed postdosing UPDRS motor scores of (mean ± SEM) 27.0 ± 12.0 for placebo-treated subjects and 10.0 ± 12.0 for apomorphine-treated subjects. A sample size of 8 patients receiving placebo and 16 patients receiving apomorphine would support a .05 level of statistical significance with 87% power. A 10% to 20% dropout was expected.
Treatment assignments were randomly and blindly assigned. Using identical-appearing ampules, medications were independently packaged to support randomized assignment in the order of patient presentation. All treatment assignments were concealed until the last patient completed the study and until documentation of all decisions regarding patient qualification for analysis. With striking differences in inpatient outcome, it is impossible to assure that blinding was functionally maintained after the first dose of drug, but no aspect other than drug effect could provide any basis for a correct guess of drug identity.
The primary measure of inpatient drug effectiveness was the comparison of predose to postdose change in UPDRS motor raw scores between the highest doses of apomorphine and the highest doses of placebo. The primary measure of outpatient drug effectiveness was the percentage of off-state events aborted.
Two inpatient efficacy factors (change in UPDRS motor score and change in hand tapping score) were suitable for parametric testing using analysis of variance (ANCOVA). All other secondary inpatient efficacy measures and all outpatient measures were tested using nonparametric tests (Wilcoxon rank sum test, χ2 test, or others where noted). α Level was .05. Except where noted, means are presented as mean ± SEM.
Correlation between inpatient response to study injection (percentage of UPDRS motor score change) and levodopa dose (single morning dose and total daily dose) was executed by linear regression (levodopa controlled-release doses multiplied by 0.75 for pooling with immediate-release doses). Correlation between inpatient response (percentage of UPDRS motor score change) and outpatient response (percentage of off-state events aborted) was a categorical analysis with the 67th percentile demarcating the 2:1 equivalent for median success. Comparison of the efficacy of first vs last daily outpatient doses was tested for active doses only using a t test. All tests other than change in UPDRS motor raw score were declared to be secondary or exploratory; no correction for test multiplicity was imposed.
Intent-to-treat (ITT) assumptions were applied. Patients who failed to progress through the complete titration were analyzed at the highest dose achieved. For outpatient factors, analyses were determined on the basis of available observations with no adjustment or imputation for missing observations.
Thirty-two patients were enrolled in the study, 29 qualified for ITT inpatient analysis (20 apomorphine-treated patients and 9 placebo-treated patients) and 26 qualified for ITT outpatient analysis (18 apomorphine-treated patients and 8 placebo-treated patients). Two patients dropped out prior to inpatient dosing because of failure to demonstrate a significant response to the levodopa challenge, and 1 patient dropped out after signing the consent form but prior to any inpatient dosing. Three patients tested under inpatient conditions failed to progress to the outpatient evaluation. One placebo-treated patient discontinued participation in the study because of a lack of effect after the third level of inpatient dosing and 1 apomorphine-treated patient discontinued participation in the study because of adverse events (nausea and vomiting) during dose titration. One additional apomorphine-treated patient began outpatient dosing, discontinued participation in the study because of chest pain during the first week of treatment, and did not keep at least 1 outpatient evaluation visit. A flowchart of the study is shown in Figure 1.
The baseline demographics of the safety-ITT population are shown by treatment assignment in Table 1. There were no statistically significant differences between groups at baseline.
The distribution of levodopa-equivalent doses of the injected study drug (or maximum dose achieved if not therapeutically equivalent to levodopa) is shown by treatment group in Figure 2. The distribution of doses was significantly different between the apomorphine-treated and placebo-treated groups (P<.001, Mantel-Haenszel χ2 test) and was bell shaped for the apomorphine-treated group with an average dose of 5.4 ± 0.5 mg. The only placebo dose less than the maximum of 1 mL was a placebo-treated patient who terminated the trial after the 0.6-mL injection because of a lack of effect.
Table 2 lists the UPDRS motor score along with associated dyskinesia scores measured in response to oral levodopa (60-minute assessment) and blinded study drug injection (20-minute assessment). The change in UPDRS motor scores was not significantly different following oral levodopa (P = .29, ANCOVA primary analysis) but was significantly different following the study drug injection. Apomorphine treatment resulted in a change of 23.9 UPDRS points (62% improvement) while placebo produced essentially no change in UPDRS motor score (P<.001, ANCOVA primary analysis). There was no difference between the 2 groups in dyskinesias following levodopa challenge, but in the group receiving active apomorphine, dyskinesias similar to those seen with levodopa treatment were preferentially seen (P = .001). Table 3 shows similar results for the secondary efficacy measures.
At the 2-week point of the outpatient phase, doses were increased by 2 mg in 5 apomorphine-treated patients to improve efficacy (from 2, 4, 6, 8, and 8 mg) and reduced from 6 to 4 mg to control nausea in one patient assigned to apomorphine treatment. In 1 of the patients whose dose increased from 8 to 10 mg, subsequent dose reduction to 5 mg was required because of resultant confusion. Patient experiences during the outpatient phase (as recorded in diaries) are given in Table 4. With up to 5 doses per day allowed as needed, patients elected to inject an average of 2.5 doses per day. Apomorphine treatment aborted 95% of the off-state events for which it was used compared with 23% for placebo (P<.001, primary outpatient analysis). Exploratory analysis of first daily dose vs last daily dose demonstrated no significant difference in outpatient efficacy under these conditions of use. From a baseline of 6 hours of off time per day, apomorphine-treated patients demonstrated a 2-hour reduction in off time while placebo-treated patients demonstrated no reduction (P = .02). This reduction in off time was seen without a reduction in the number of discrete off-state events suffered per day.
Figure 3 shows a scatterplot for the regression of apomorphine dose on oral levodopa dose (that single morning dose that produced the effect to which apomorphine responses were matched). Despite a response ratio of 96%, the morning levodopa dose was not predictive of required apomorphine dose (R2 = 0.05, P = .35, slope = 0.006, and intercept = 4.0). Total daily levodopa dose also was not predictive of apomorphine dose (R2 = 0.05, P = .32, slope = 0.002, and intercept = 4.2).
Figure 4 shows a scatterplot with contingency table for the correlation between inpatient and outpatient efficacy results. Outpatient outcome was characterized by the invariant grouping of results about 2 distinct values (100% or 0% successful abortion of off-state events). The 67th percentile for success (2:1 randomization) was a 35% reduction in UPDRS motor score and reversal of at least 80% of outpatient off-state events. In 22 of 26 patients, inpatient and outpatient results were concordant (P<.001, Fisher exact test).
Table 5 summarizes the response to apomorphine at doses less than levodopa equivalent. The 2-mg dose of apomorphine hydrochloride, (optimal for 3 of the 20 patients) was statistically distinct from placebo (32% improvement apomorphine vs 6.3% change placebo, P = .02). The levodopa-equivalent dose of apomorphine hydrochloride produced a 62% improvement in UPDRS motor score while 2 mg below the levodopa-equivalent dose (or 2 mg in 3 patients titrated to only 2 mg) still provided 42% improvement in UPDRS motor score (P<.001 ANCOVA, apomorphine-treated vs placebo-treated patients).
Adverse events (mostly mild in severity) occurred in 85% of apomorphine-treated patients and in 89% of placebo-treated patients (Table 6). A single serious adverse event (chest pain, myocardial infarction ruled out) occurred in a patient assigned to the placebo study drug. Similar events, without hospitalization, occurred in 3 apomorphine-treated patients. Injection site complaints (including bruising, pain, skin reaction, and nodule development) were common. Yawning was reported by 40% of the apomorphine-treated group but none of the placebo-treated group (P = .03). Thirty-five percent of the apomorphine-treated patients (and no placebo-treated patients) experienced drowsiness or somnolence (P = .07). Dyskinesias were reported as an adverse event by 35% of the apomorphine-treated group and 11% of the placebo-treated group. Nausea or vomiting occurred in 30% of the apomorphine-treated patients and 11% of the placebo-treated patients. In 1 apomorphine-treated patient, nausea with vomiting at the 6-mg dose (given during the inpatient phase) was severe and resulted in discontinuation from the study. Outpatient nausea was usually mild to moderate in severity and was generally reported as an isolated event (nausea was not seen in 96% of all apomorphine-treated patients who received outpatient injections).
Uric acid demonstrated statistically significant change from the mean baseline value (+0.27 mg/dL for apomorphine, −0.34 mg/dL for placebo, P = .02) but was never outside the normal range. There were no statistically significant changes in other safety measures (blood test results, electrocardiograms, or physical examination findings).
This study was conducted in patients with significant residual off time despite aggressive attempts to control symptoms using both levodopa and oral dopamine agonists. In this setting, outpatient subcutaneous injections represent a reasonable therapeutic option for acutely reversing individual off events. This is the first clinical trial to evaluate both inpatient and outpatient use under placebo-controlled conditions, and it established the predictive nature of inpatient test responses on outpatient therapeutic responses to injected apomorphine.
Our study showed that subcutaneous apomorphine injections effectively reverse off-state events that occur in advanced PD. Outpatients or caregivers demonstrated the capacity to prepare apomorphine from glass ampules and inject it. While we did not use formal quality of life measures, most patients assigned to active drug were pleased with the effects. Patients randomized to apomorphine treatment had reduced total off time that was not derived from lessening the number of off-state events, but from shortening the duration of individual off states. Although within-day tolerance has been previously demonstrated following intravenous infusions of apomorphine,19 this phenomenon was not observed with the subcutaneous injections used in our study.
Since levodopa dosage was not predictive of apomorphine dose requirements, individual titration is required to establish the correct dose. Under inpatient observation, the average dose required to produce effects equivalent in magnitude to oral levodopa was 5.4 mg, while the average final outpatient dose was 5.8 mg. At levodopa-equivalent doses, dyskinesias were equal in magnitude to those produced by levodopa, but these were mild and generally nondisabling. The only adverse event that was significantly more common in the apomorphine-treated group was yawning. This adverse effect has been reported in patients treated with apomorphine in other studies20 but is rare with levodopa and the oral dopamine agonists.21 A review of the Physicians Desk Reference revealed that of the oral dopamine agonists and levodopa preparations available in the United States, yawning is a listed adverse effect only for ropinirole hydrochloride occurring in 3% of patients. The 40% incidence of yawning associated with apomorphine in our study represents more than a 10-fold increase compared with ropinirole. Studies of apomorphine-induced yawning in rats have indicated that stimulation of D2 dopamine receptors on paraventricular neurons of the hypothalamus leads to increased nitric oxide synthase activity and yawning behavior.22 Why levodopa, which produces a similar degree of antiparkinsonian efficacy, is associated with a much lower incidence of yawning is unknown.
Our data show that domperidone is not needed to support the use of apomorphine in patients selected and treated according to our protocol. Clinically significant nausea and vomiting was as rare in our study (only 4% of injections of active drug caused nausea) as has been previously reported in European studies of apomorphine combined with domperidone.20,23 Our data suggest that trimethobenzamide is an adequate replacement for domperidone therapy. Alternatively, it is possible that down-regulation of dopamine receptors in the area postrema had already occurred because of chronic exposure to long-acting oral dopamine agonists and that this patient group actually does not require an antinauseant.
We conclude that subcutaneously injected apomorphine rapidly and reliably reversed off-state events in patients with advanced PD in whom conventional antiparkinsonian medications had been optimized. We believe that this study confirms previous work and serves as proof of the clinical effectiveness of subcutaneous apomorphine injections when used to reverse off events.
Accepted for publication May 14, 2001.
This study was supported by Bertek Pharmaceuticals Inc, Division of Mylan Laboratories Inc.
We thank the following coinvestigators, study coordinators, and data analysis personnel for their contributions to this project. Coinvestigators: Ben Williams, MD, PhD; Bhupesh Dihenia, MD; and Eric Molho, MD. Study Coordinators: Jennifer Stanford, RN; Janice Stewart, RN, BSN, CCRC; Angela Bednarz, RN, BSN; Merena Tindall, RN; Diane Brown, RN; Sharon Evans, LPN; and Kathie Mistura, RN. Data Management, Statistical, and Editorial Support Staff: Marlene R. Pope, BS; Brenda E. VanLunen, MS; Jeffrey P. Smith, PhD, and Patrick D. McGrath, PhD.
Corresponding author: Richard B. Dewey, Jr, MD, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-9036.