Famotidine, a Histamine H₂ Receptor Antagonist, Does Not Reduce Levodopa‐Induced Dyskinesia in Parkinson's Disease: A Proof‐of‐Concept Study

Abstract

The neural mechanisms underlying levodopa‐induced dyskinesia (LID) in Parkinson's disease (PD) may involve histamine (H₂) receptors on striatopallidal pathways. We recently demonstrated that the clinically available oral histamine H₂ receptor antagonist (H₂ RA), famotidine, can reduce l‐dopa‐induced chorea in MPTP‐lesioned macaques. We hypothesized that famotidine may be useful in the treatment of LID in PD patients. We performed a proof‐of‐concept, double‐blind, randomized, multiple cross‐over (4×) trial. Seven PD subjects with bothersome dyskinesia were randomized to oral famotidine 80, 120, and 160 mg/day and placebo. Each subject was randomized to receive each of the four treatment phases for 14 days followed by a 7‐day wash‐out period between each treatment phase. The primary outcome measure was change in the Unified Dyskinesia Rating Scale (UDysRS; part III) between placebo and famotidine. Secondary outcomes were UDysRS (parts I and II), Global Impression of Change, Lang‐Fahn Activities of Daily Living Dyskinesia Scale, Unified Parkinson's Disease Rating part III, and adverse events (AEs). Outcomes were evaluated pre‐ and post‐treatment per dose and analyzed using a mixed‐effects linear model. There was no significant effect of famotidine treatment on any of the primary or secondary outcome measures compared to placebo (each dose and all doses combined). There were no significant AEs. Even though the sample size of the current study is limited, famotidine seems to be safe in patients with PD and LID, but showed no potential as an antidyskinetic agent.

Levodopa‐induced dyskinesia (LID) remains a problem in advanced Parkinson's disease (PD). Targeting nondopaminergic systems is an option for reducing dyskinesia, without worsening motor symptoms.1 The central histaminergic system is involved in diverse biological functions, including thermoregulation, eating, and sleep; a role in motor activity is suggested by strong histaminergic innervation of the basal ganglia. Histamine H₂ receptors are highly expressed in the striatum, particularly on the gamma‐aminobutyric acid (GABA)ergic striatopallidal and striatonigral pathways.2 Histamine H₂ stimulation modulates acetylcholine (Ach) release.3, 4 Previous studies have demonstrated that blocking the action of ACh with anticholinergic agents can induce chorea.5 We propose that histamine H₂ receptor stimulation decreases ACh in the striatum and increases activity of the direct striatal output pathway, a key component of the neural mechanisms underlying dyskinesia.6, 7 We have recently shown that the clinically available oral selective histamine H₂ receptor antagonist (H₂RA), famotidine, can reduce dyskinesia in the MPTP‐lesioned marmoset model of PD.8

Famotidine oral therapy is licensed for use in various gastrointestinal (GI) conditions, including prevention of duodenal ulcer recurrence (20 mg/day), treatment of duodenal ulcer and gastric ulcers (40 mg/day), treatment and maintenance of remission in GI esophageal reflux disease (40 mg/day and up to 80 mg/day for esophageal erosions), and in treatment of pathological hypersecretory conditions (e.g., Zollinger‐Ellison syndrome, up to a maximum of 800 mg/day). Famotidine has been shown to be safe and well tolerated.9, 10 Furthermore, famotidine has the advantage of being a clinically available selective H₂RA, that crosses the blood–brain barrier,11 for which preclinical and clinical data support its experimental use in clinical studies for PD. Thus, we performed a phase IIa study to determine safety and preliminary efficacy of the H₂RA, famotidine, in the treatment of LID in patients with PD.

Methods

Subjects with idiopathic PD, according to the UK Parkinson Disease Society Brain Bank criteria,12 were recruited from the Movement Disorder Clinic at Toronto Western Hospital (Toronto, Ontario, Canada) between July 2011 and July 2013. Subjects, male and female <80 years of age, were on stable PD medications (amantadine included) for at least 1 month, and reported stable bothersome dyskinesia (International Parkinson and Movement Disorder Society [MDS]‐UPDRS item 4.2 score of 2 or greater). Exclusion criteria were previous surgery for PD, H & Y score of 5 when OFF medication, history of moderate‐to‐severe renal impairment, dementia (defined by a score <25 in the Montreal Cognitive Assessment), or known hypersensitivity to lactose, famotidine, or other H₂RAs. The local research ethics committee approved the study, and all subjects gave informed written consent. The study was conducted according to good clinical practice standards. The trial was registered at clincialtrials.gov (NCT01937078).

We performed a phase IIa, randomized, double‐blind, placebo‐controlled, multiple‐dose, multiple cross‐over study. Each subject completed 4‐ × 2‐week treatment periods, each separated by a 1‐week wash‐out period. Treatment periods were randomized and included three periods of famotidine (80, 120, and 160 mg/day) and one period with placebo. Eligible subjects received encapsulated tablets of famotidine (40 mg) or matching placebo according to randomization tables generated by the hospital pharmacy. Placebo capsules were identical in appearance, smell, and taste. At the start of each treatment period, regardless of the target dose, there was an up‐titration schedule for 3 days, starting at 40 mg/day on day 1, 40 mg twice‐daily on day 2, and 80 mg twice‐daily on day 3 (i.e., 2 capsules twice‐daily, with varying ratios of active famotidine to placebo capsules), followed by a 12‐day maintenance phase of the target dose taken twice‐daily. The elimination half‐life of famotidine is 2.5 to 3.5 hours. As such, treatment duration of 2 weeks was deemed sufficient to ensure consistent therapeutic plasma levels of the drug, following the described titration regimen (drug levels were not measured). If side effects were encountered, the dose of study drug was decreased, initially by 1 capsule in the morning and then 1 capsule evening; if still intolerant, the dose was reduced further to 1 capsule only in the morning. In order to assess medication compliance, the number of capsules dispensed to and returned by each subject were counted and documented. At the end of each treatment phase, subjects stopped drug without down‐titration. The wash‐out period between treatment phases was 7 days (which is >5 times the elimination half‐life required to completely wash out).

Outcome Measures

The primary outcome measure was the change in Unified Dyskinesia Rating Scale (UDysRS)13 (part III/Impairment) between placebo and famotidine (for each dose and for all doses combined), evaluated by a rater, blinded to the drug assignment or visit. To ensure peak‐dose LID, the subject was assessed 1 hour after the last antiparkinsonian medication and at the same time of day for each visit. Secondary outcome measures included subject‐rated assessments of dyskinesia using the UDysRS parts I and II, MDS‐UPDRS part 4.1 and 4.2, and Lang‐Fahn Activities of Daily Living Dyskinesia scale (LFADLDS) as well as patient‐rated clinical global impression (CGI) of change in dyskinesia. We used the blinded‐rated MDS‐UPDRS part III to measure parkinsonian disability. The frequency, severity, nature, and duration of any adverse events (AEs) were assessed at each visit and by telephone after 7 days of treatment. Secondary outcome measures (apart from MDS‐UPDRS part III) and AEs were assessed at each visit by the primary investigator.

Statistical Analysis

Continuous data were described as median (interquartile range) and mean (standard deviation; SD), and categorical data were described as counts expressed in absolute values and frequency. The primary outcome and the secondary outcomes (score change in UDysRS parts I and II, Disability, and LFADLDS) that were considered as continuous variables were analyzed with a mixed‐effects linear model, using PROC MIXED in SAS 9.3 (SAS Institute Inc., Cary, NC). Treatment dose (80, 120, and 160 mg/day and placebo assigned 0 mg/day), timing of assessment within each treatment period (end vs. start of treatment period), and interaction between dose and timing of assessment were considered fixed predictors. A random effect was allowed for intercepts, grouped by study participant. To compare treatment effects for the various outcome measures, the estimates of the interaction between dose and timing of assessment were compared, taking the placebo estimate as the comparator. Sequence of treatment periods was incorporated in the resulting model as a random factor to test for a period effect, and models were comparable as measured by Akaike's information criterion (data not shown). Secondary outcomes that were considered categorical variables were further classified considering the change from start to end of treatment period in improved (at least 1‐point decrease), unchanged or worsened (at least 1‐point increase), and then tested using the chi‐squared test or Fischer's exact test, when applicable. Currently, a clinically meaningful change in UDysRS‐Impairment is not known; thus, a formal power calculation was not possible. Statistical significance was set at P < 0.05. All analyses were performed using SAS 9.3 (SAS Institute Inc.).

Results

Forty‐three PD patients were assessed for eligibility and 7 were randomized (CONSORT Fig. 1). All randomized subjects completed the study protocol, with the exception of a missed visit for 1 study participant. All participants completed the titration protocol for all treatment periods. See Table 1 for full details on baseline demographics of all randomized study participants.

Figure 1.

Consort diagram.

Table 1.

Demographic and clinical characteristics of study participants at baseline

Sex M/F	3M/4F
Age (y), mean (SD)	66.3 (9.2)
Duration of PD (y), mean (SD)	6.3 (6.4)
Total LEDD (mg/day), mean (SD)	501.5 (460.5)
Amantadine use	N = 2
Percentage ON time with dyskinesia mean (SD)	51.8 (30.3)
UDysRS part I, median (interquartile range)	18 (10)
UDysRS part II, median (interquartile range)	2.0 (7.7)
UDysRS part III Impairment, median (interquartile range)	10.0 (5.0)
UDysRS part IV Disability, median (interquartile range)	6.0 (4.0)
LFADLS, median (interquartile range)	10.0 (9.0)
MDS‐UPRS part III ON, median (interquartile range)	17.0 (9.0)
CGI‐dyskinesia severity
Mildly ill	N = 5
Moderately ill	N = 2

M, male; F, female; LEDD, l‐dopa equivalent daily dose.19

There was no difference in change scores in the primary outcome measure (UDysRS part III/Impairment) from start to end of treatment period for each dose of famotidine compared to placebo (P = not significant [NS]) or for the combined treatment periods with famotidine compared to placebo (P = NS; see Table 2 and Fig. 2).

Table 2.

Modeling of UDysRS part III (impairment) scores according to dose, timing of assessment in each treatment period, and interaction between timing and dose

Dose (Comparator = Placebo) (mg)	Estimate (95% CI)	T Value	P Value
80	−1.5 (−5.3, 2.2)	−0.83	0.41
120	−2.4 (−6.1, 1.2)	−1.34	0.18
160	−1.8571 (−5.5, 1.81)	−1.02	0.31

The results of the interaction term are provided to show the effect of treatment in the UDysRS part III (impairment) scores from end to start of treatment period. A linear mixed model was used, with subject ID as random factor. Negative values represent improvement.

Figure 2.

Change in UDysRS part III (Impairment) (yy axis) at each treatment period (xx axis) from start to end of treatment. Negative values represent improvement. Difference between different treatment doses and placebo (treatment = 0).

In addition, there was no difference between famotidine and placebo in all secondary outcomes. There was no statistically significant change in the UDysRS‐Disability for each dose of famotidine compared to placebo (P = NS) or for the combined treatment periods of famotidine compared to placebo (P = NS). Similar results were found for UDysRS parts I and II as well as the LFADLDS. Importantly, there was no change in parkinsonian disability measured by the blinded‐rated MDS‐UPDRS‐III for each dose of famotidine compared to placebo (P = NS) or for the combined treatment periods of famotidine compared to placebo (P = NS; Table 3). For the MDS‐UPDRS items 4.1 (Daytime spent with dyskinesia), 4.2 (Functional impact of dyskinesia), time spent with ON‐dyskinesia (UDysRS part I, item 1), and the patient‐rated CGI of change in dyskinesia, there was no significant improvement, and the modal category was no change across the different doses of famotidine and placebo (Table 4).

Table 3.

Modeling of secondary outcomes measures UDysRS parts I and II, LFADLDS, and blinded‐rated MDS‐UPDRS part III scores according to dose, timing of assessment in each treatment period, and interaction between timing and dose

	Placebo			80 mg			120 mg			160 mg			All Treatment Doses
	Estimate	95% CI	P Value a	Estimate	95% CI	P Value a	Estimate	95% CI	P Value	Estimate	95% CI	P Value	Estimate	95% CI	P Value
UDysRS part III (disability)	—			−1.5	−5.3–2.2	0.41	−2.4	−6.1–1.2	0.18	−1.9	−5.5–1.8	0.3	−2.0	−4.9–1.0	0.18
UDysRS part I	—			2.8	−3.7–9.3	0.39	1.3	−5.1–7.7	0.69	0	−6.4–6.4	1.00	0.3	−3.4–3.9	0.88
UDysRS part II	—			1.5	−2.9–5.9	0.49	1.4	−2.9–5.7	0.50	0.4	−3.9–4.7	0.84	2.1	−0.3–4.5	0.09
LFADLDS	—			0	−4.0–4.0	1.00	1.6	−2.9–5.2	0.43	1.1	−2.9–5.2	0.6	0.4286	−1.9–2.3	0.71
MDS‐UPDRS‐III	–			4.1	−5.1–13.4	0.37	4.6	−4.7–13.8	0.32	3.3	−5.9–12.5	0.5	4	−3.2–11.2	0.27

The results of the interaction term are provided, because they test the effect of treatment in the outcome measure from end to start of treatment period. A linear mixed model was used, with subject ID as random factor. Negative values represent improvement.

^aEstimates and P values are in comparison of each dose level (or all treatment doses) with placebo.

Table 4.

Responder analysis for the secondary outcomes MDS‐UPDRS part 4.1 and 4.2, time spent with ON‐dyskinesia (UDysRS item 1), and patient‐rated CGI of change in dyskinesia

	Placebo		80 mg		120 mg		160 mg		All Drug Doses
	n	%	n	%	n	%	n	%	n	%
MDS‐UPDRS part 4.1: Daytime spent with dyskinesia	P = 0.37								P = 0.06
Improvement	2	28.6	0	0.0	0	0.0	1	14.3	1	5.0
No change	3	42.9	6	100.0	6	85.7	6	85.7	18	90.0
Worsening	2	28.6	0	0.0	1	14.3	0	0.0	1	5.0
MDS‐UPDRS part 4.2: Functional impact of dyskinesia	P = 0.054								P = 0.58
Improvement	2	28.6	1	16.7	6	85.7	3	42.9	10	50.0
No change	4	57.1	1	16.7	1	14.3	3	42.9	5	25.0
Worsening	1	14.3	4	66.7	0	0.0	1	14.3	5	25.0
Time spent with ON‐dyskinesia (UDysRS part I, item 1)	P = 0.53								P = 0.74
Improvement	1	14.3	1	16.7	1	14.3	0	0.0	2	10.0
No change	4	57.1	4	66.7	5	71.4	5	71.4	14	70.0
Worsening	2	28.6	1	16.7	1	14.3	2	28.6	4	20.0
Patient GCI: change in dyskinesia	P = 0.37								P = 0.69
Improvement	1	14.3	0	0.0	1	14.3	4	57.1	5	25.0
No change	6	85.7	5	83.3	5	71.4	3	42.9	13	65.0
Worsening	0	0.0	1	16.7	1	14.3	0	0.0	2	10.0

AEs

There were no serious AEs. During the study, 5 subjects reported AEs. All events occurred while taking famotidine (all doses). One subject felt dizzy and fell, resulting in a rib fracture—this may be causally related to the drug because dizziness has been reported at a rate of greater than 1% in patients on therapy with oral famotidine in controlled trials.10 One subject developed fatigue, and 1 had a brief episode of abdominal pain—these reactions have been reported in clinical trials with use of oral famotidine, but occurred under circumstances where a causal relationship could not be established.10 One subject fell, resulting in leg bruising, and 1 had mild transient possible leg cellulitis; both felt unrelated to study drug. All AEs resolved fully without change in study protocol. No AE resulted in withdrawal of a subject.

Discussion

This study has shown that the oral H₂RA, famotidine, is well tolerated by PD patients; however, there is no evidence that famotidine can reduce LID. This finding was observed across different doses of famotidine and a wide variety of outcomes measures, including objective blinded rater scores, as well as subjective ratings of dyskinesia disability with both patient and physician‐centered outcomes. The negative outcome may be a result of a number of factors, including lack of benefit for biological reasons or study design issues. There have been several drugs that showed promise at the preclinical level of study, but did not reduce LID in human subjects in phase IIa trials.14, 15 There are several possible reasons for this translational failure of famotidine. In particular, we did not measure target engagement and used potentially nonequivalent dosages. Doses used in the clinical trial were adapted from earlier clinical use in non‐PD subjects, and, although we did cover a wide range, they may not be not be entirely equivalent to the single doses of famotidine (1–30 mg/kg) used in primate study. The pathophysiology of LID in PD patients is clearly more complex than in the MPTP primate. The lack of an ongoing neurodegenerative process, combined with the relatively “pure” dopamine deficiency, in the latter means that exact correlations with changes in nondopaminergic neurotransmitter systems that occur in the human brain may not be replicated. However, the MPTP model of LID is a powerful screening tool for initial hypothesis testing. Our study also suggests that using a single selective neurotransmitter receptor as a target is one plausible explanation as to failure. For example, the most effective oral antidyskinetic agent, amantadine, has multiple receptor targets, including N‐methyl‐D‐aspartate glutamatergic, muscarinic, and dopaminergic. Thus, to significantly reduce LID in PD patients, several neurotransmitter pathways need to be implicated.

The practicalities of clinical trial design in evaluating novel drugs for LID remains challenging. Several factors play a role, including a large placebo effect, thus masking a potentially beneficial outcome, lack of recognition by patients of dyskinesia and heterogeneity of the phenomenology of LID, plus reduced availability of subjects as a result of the earlier use of STN‐DBS in an otherwise suitable candidate for drug trials. Recruitment can be a challenge because subjects experiencing dyskinesia may not report disability, often a requirement for inclusion. Alternative study designs to the classic parallel group randomized, clinical trial include “n‐of‐1” study designs that have been used for LID in PD.16, 17, 18 We attempted a pragmatic multiple cross‐over design to evaluate multiple doses in 1 subject and one placebo arm (rather than multiple in the classic n‐of‐1 study) to reduce the duration of the study, yet provide statistically robust data.

Although famotidine is unlikely to be further studied as an antidyskinetic agent in PD, it is an example of how clinically available drugs can be potentially evaluated in PD for hypothesis testing preceding expensive, lengthy development of more target‐specific drugs.

Author Roles

(1) Research Project: A. Conception, B. Organization, C. Execution; (2) Statistical Analysis: A. Design, B. Execution, C. Review and Critique; (3) Manuscript: A. Writing of the First Draft, B. Review and Critique.

T.A.M.: 1B, 1C, 2A, 2B, 2C, 3A, 3B

B.B.S.: 1A, 1B, 1C, 2A, 2B, 2C, 3B

B.S.C.: 1C, 3B

C.d.A.: 1C, 2C, 3B

A.A.D.: 1C, 2C, 3B

R.W.: 1C, 2C, 3B

T.G.: 1B, 1C, 2C, 3B

J.P.L.: 1A, 1B, 1C, 2C, 3B

S.H.F.: 1A, 1B, 1C, 2A, 2B, 2C, 3A, 3B

Disclosures

Funding Sources and Conflicts of Interest: The study was supported by a Pilot Project Grant from the Parkinson Society Canada (awarded to B.B.S. and S.H.F.). There are no conflicts of interest.

Financial Disclosures for previous 12 months: Information concerning all sources of financial support and funding for the preceding 12 mo, regardless of relationship to current manuscript, must be submitted with the following categories suggested. T.A.M. has held a consultancy with the Cure Huntington's Disease Initiative (CHDI) Foundation; has been employed by the University of Ottawa; and has been awarded grants from Parkinson Society Canada. B.B.S. has been employed with the University of Virginia and has been awarded grants from Parkinson Society Canada. B.S.C. has been employed with McMaster University and has received honoraria from the Parkinson Disease and Movement Disorder Society. C.d.A. has been employed with the Parkinson Society Canada Clinical Fellowship. A.A.D. has been employed with the Toronto Western Hospital Clinical Fellowship. R.W. has served on the advisory boards of Novartis and Abbvie; has been employed with Trinity College Dublin; has received honoraria from Lundbeck; and has been awarded grants from AbbVie and Ataxia Ireland. T.G. has been employed by the University Health Network, Toronto. J.P.L. has been employed by the University Health Network, Toronto. S.F. has served on the advisory boards of Teva, Merz, and Zambon; has been employed with the Department of Medicine Practice Plan, University Health Network; has signed contracts with Chelsea Therapeutics, Kyowa, and Avanir; has received honoraria from Teva and Novartis; has received royalties from Oxford University Press; and has been awarded grants from the National Institutes of Health, Parkinson Society Canada, the Michael J. Fox Foundation for Parkinson's Research, and the Toronto Western Research Institute (TWRI) Foundation.