Inhaled Levodopa for the Management of OFF Episodes in Patients with Parkinson’s Disease: A Network Meta-analysis

Abstract

Introduction

In the course of their disease, most patients with Parkinson’s disease (PD) will experience OFF episodes, during which symptoms worsen despite symptomatic treatment. Current on-demand treatments for OFF episodes include inhaled levodopa and subcutaneous and sublingual apomorphine, however no head-to-head comparison of these treatments is available. We performed a network meta-analysis (NMA) to provide robust comparative evidence for on-demand OFF episode treatments.

Methods

Randomized controlled trials assessing OFF episode treatments in patients with PD were identified in a systematic literature review. A feasibility assessment was conducted considering OFF time, Unified Parkinson's Disease Rating Scale (UPDRS) Part III scores, Patient Global Impression of Change score (PGI-C), and safety outcomes. An NMA was carried out using a random effects model.

Results

Twenty-one trials were identified in the systematic literature review, 11 of which were included in the feasibility assessment and deemed suitable to be included in the NMA. No statistically significant difference in OFF time was observed between patients receiving inhaled levodopa and those receiving subcutaneous apomorphine, and no statistically significant difference in UPDRS Part III scores or probability of PGI-C score improvements, all-cause discontinuation, or adverse events (AEs) were observed between patients receiving inhaled levodopa, subcutaneous apomorphine, or sublingual apomorphine. Subcutaneous apomorphine had significantly higher probability of treatment discontinuation due to adverse events compared to inhaled levodopa (Log odds ratio 20.712; 95% credible interval 1.855, 55.991).

Conclusion

Inhaled levodopa demonstrated no statistically significant difference in efficacy with subcutaneous or sublingual apomorphine, however inhaled levodopa had a lower probability of treatment discontinuation due to AEs than subcutaneous apomorphine. These data highlight that inhaled levodopa is a suitable non-invasive and well-tolerated treatment for OFF episodes.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s40120-026-00991-3.

Key Summary Points

Why carry out this study?

Most patients with Parkinson’s disease will experience OFF episodes during the course of their disease

On-demand therapies to treat OFF episodes include inhaled levodopa, and subcutaneous and sublingual apomorphine

The study aimed to provide comparative evidence for on-demand treatments of OFF episodes in Parkinson’s disease

What was learned from the study?

The efficacy of inhaled levodopa was demonstrated to be not significantly different from that of subcutaneous and sublingual apomorphine while the rate of discontinuation due to adverse events was significantly lower for inhaled levodopa compared with subcutaneous apomorphine

Inhaled levodopa is an alternative treatment for patients seeking a non-invasive, well-tolerated, on-demand treatment option for managing OFF episodes in Parkinson’s disease

Introduction

Parkinson’s disease (PD) is a chronic, progressive, and disabling neurodegenerative disorder characterized by progressively worsening motor and non-motor symptoms, including bradykinesia and at least one additional characteristic, namely, tremor or muscle rigidity [1]. While treatments are available that limit motor symptoms, motor fluctuations (variations in a person’s ability to move) occur as the effects of treatments wear off. In patients with PD, periods of the day are commonly categorized as either ON or OFF time. During ON episodes, motor symptoms are generally well controlled, but ON time can be further subcategorized based on the presence and severity of dyskinesia: ON without dyskinesia, ON with non-troublesome dyskinesia, and ON with troublesome dyskinesia. Conversely, during OFF episodes, the therapeutic effects of treatment wane and motor symptoms return or worsen [2]. As PD progresses, nearly all patients will experience OFF episodes, which are associated with a significant burden on the patient, affecting activities of daily living and quality of life, in addition to affecting the caregiver and increasing the healthcare burden [2–5].

Current treatment of OFF episodes includes dose escalation of base levodopa therapy or on-demand treatments such as apomorphine and fast-acting levodopa formulations. Apomorphine is a dopamine receptor agonist that stimulates D1 and D2 receptors in the brain, thereby helping to improve motor function [6]. Apomorphine is available as subcutaneous injection or sublingual tablets. In phase 3 clinical trials, subcutaneous apomorphine has been shown to improve Unified Parkinson's Disease Rating Scale (UPDRS) Part III scores and OFF episodes in addition to non-motor features of PD compared with placebo [7, 8]. However, apomorphine is associated with challenging dose titration requirements and adverse events (AEs) such as nausea and vomiting [9].

Fast-acting levodopa formulations include rapid-release levodopa and dispersible levodopa, but both are associated with limitations in terms of absorption and bioavailability [9–11]. Inhaled levodopa is an alternative treatment used for OFF episodes, offering a systemic administration via the pulmonary route that allows fast absorption and avoids variability in gastrointestinal absorption [10, 12, 13]. In the phase 3 SPAN-PD trial, significant improvements in UPDRS Part III scores at 10, 20, 30, and 60 min and improved resolution of OFF states to ON states within 60 min were reported with inhaled levodopa compared with placebo [10]. Inhaled levodopa has also been shown to numerically improve Patient Global Impression of Change (PGI-C) scores compared with placebo (although the results were not statistically significant) while safety and tolerability are in line with base levodopa therapy, with the exception of pulmonary and respiratory AEs. These events are associated with the pulmonary route of administration and tend to be transient and non-serious. Inhaled levodopa thus provides fast-acting symptom relief by a simple means of administration which addresses an unmet need in patients requiring on-demand treatment for OFF episodes [9].

While randomized placebo-controlled trials are useful to determine true treatment effects, assessments of comparative efficacy versus other treatments are important for clinicians and decision-makers to determine optimal treatment for patients. Head-to-head clinical trials are the gold standard, but there are currently no direct head-to-head randomized controlled trials (RCTs) comparing inhaled levodopa with apomorphine. Naïve comparisons carry substantial limitations that can introduce bias and either under- or overestimate treatment effects, whereas indirect treatment comparisons can be used and are an accepted means of determining comparative efficacy. The authors of a matched-adjusted indirect comparison (MAIC) using individual patient-level data of OFF episodes for sublingual apomorphine and inhaled levodopa reported that sublingual apomorphine significantly improved UPDRS Part III scores at 60 min and significantly reduced total daily OFF time compared with inhaled levodopa, while no statistically significant difference between sublingual apomorphine and inhaled levodopa was reported in UPDRS Part III scores at 30 min [3]. This MAIC analysis was constrained by heterogeneous outcome measures and scale conversions, reliance on subjective patient diaries for ON/OFF data, and the omission of AE assessments [3]; in addition, the analysis did not consider all available on-demand treatments, which restricts the generalizability of the findings.

To better understand the real-world place in therapy of on-demand treatments for OFF episodes in PD, we performed a network meta-analysis (NMA) with the aim to provide robust comparative evidence for the safety and efficacy of inhaled levodopa compared to other on-demand treatment options.

Methods

Search Strategy and Selection Criteria

A systematic search was conducted in Embase, MEDLINE, Embase Classic, CENTRAL, and Cochrane Clinical Answers to identify clinical trials that assessed the efficacy and safety of treatments for intermittent OFF episodes in patients with PD. Gray literature searches of recent congresses were performed and supplementary hand searches of bibliographies of systematic reviews were also conducted (details in the Electronic Supplementary Material). Only English-language studies were included, and no date limits were applied. Full search terms are provided in ESM Tables S1 and S2.

Following deduplication, titles and abstracts were screened independently by two reviewers according to the PICOS criteria presented in ESM Table  S3. Discrepancies were resolved through discussion or adjudicated by a third reviewer. Full-text articles of potentially relevant studies underwent second-pass screening using the same criteria. Clinical trials were included if they reported clinical outcomes of the efficacy and safety of treatments for intermittent OFF episode treatments in patients with PD.

Data Extraction, Quality Assessment, and Feasibility Analysis

Data extraction was conducted by one reviewer, and a second reviewer verified accuracy and completeness. Extracted variables included study design, population characteristics, interventions, comparators, outcomes (covering efficacy measures such as change in OFF-time, motor fluctuations, and wearing-off; tolerability assessed by treatment discontinuation due to all causes; and safety evaluated by incidence of AEs, discontinuation due to AEs, dyskinesia, hallucination, and orthostatic hypotension), follow-up duration, and results. Risk of bias was assessed for each included study using an adapted Cochrane tool evaluating five domains: randomization, deviations from intended interventions, missing data, outcome measurement, and selective reporting (details in the ESM files). Discrepancies in data extraction or quality assessment were resolved by consensus or third-party adjudication.

A feasibility assessment was performed on the identified RCTs to assess whether networks could be formed. It should be noted that trials which did not include a comparator, such as single arm trials, were not included in the feasibility assessment or analysis.

Statistical Analysis

An NMA was performed for all endpoints that were deemed feasible in the feasibility assessment. To ensure adequate network connectivity and statistical power, data for inhaled levodopa were pooled across different dosages. Placebo efficacy was estimated by averaging placebo arms, following Dias et al. [14]. In cases where uncertainty measures were missing, standard error-to-mean ratios were imputed using data from other studies evaluating the same intervention.

Indirect comparisons were estimated using a random effects model to account for heterogeneity in treatment effects across studies. Multi-arm trials were adjusted assuming equal between-trial variance across relative effects. All analyses were conducted using R (ver. 4.4 or above; R Foundation for Statistical Computing, Vienna, Austria) via RStudio using the following packages: multinma, dplyr, tidyr, wakefield, ggplot2, and sandwich. A fixed effects model was also used as a sensitivity analysis, and model fit was assessed using the deviance information criterion (DIC), with a difference of ≥ 3 points indicating superior fit. Residual deviance was also evaluated, with values approximating the number of data points considered to be indicative of adequate fit. Convergence diagnostics included the Gelman-Rubin statistic and visual inspection of trace, density, and autocorrelation plots. Appropriate burn-in periods, iteration counts, and thinning intervals were applied, which for all variables in both fixed and random effects models was a burn-in of 10,000 iterations, retaining 5000 samples with a thinning interval of 10.

Continuous outcomes (change in OFF time and change in UPDRS Part III score) were analyzed using the identity link function while improvement in PGI-C, all-cause treatment discontinuation, and rates of AEs were assessed as binary outcomes considering the logit function. A continuity correction was applied for binary endpoints. Change in OFF time was the key efficacy endpoint analyzed given its direct clinical relevance for on-demand therapies for PD. An additional sensitivity analysis was also performed—excluding data from Katzenschlager et al. [17] evaluation of continuous subcutaneous apomorphine infusion due to potential clinical heterogeneity regarding on-demand rescue therapies (details are provided in the ESM files). All analyses were performed according to guidance provided in the National Institute of Health and Care Excellence (NICE) Decision Support Unit’s Technical Support Document [14].

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors and thus did not require ethics approval.

Results

Summary of Systematic Search Results

A total of 345 articles were retrieved in the literature searches, and following screening and eligibility assessment, 58 publications corresponding to 21 distinct clinical trials were included (Fig. 1). The overall risk of bias assessment is shown in Table 1, and risk of bias by domains is shown in ESM Table  S4.

Fig. 1.

PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) diagram of the systematic literature review. The 58 full text publications initially included were grouped into 21 unique trials based on National Clinical Trial (NCT) number. ^aStocchi et al. [32], Hauser et al. [33], Hui et al. [34], Isaacson et al. [35], Graham et al. [36], Katzenschlager et al. [37], Ondo et al. [38]; ^bFarbman et al. [39]; ^cDiamond et al. [40], Stocchi et al. [41]. EED Economic Evaluation Database, HTA Health Technology Assessment, NHS National Health Service

Table 1.

Characteristics of patients involved in and overall risk of bias of the 11 randomized clinical trials included in the feasibility assessment

Number	Main publication (full paper)	NCT number/publication	Treatment arm	Age, years (SD)	Male patients, n (%)	OFF-time in h/day, mean (SD)	Mean levodopa dose (mg/day)	Overall RoB
1	Olanow et al. 2020 [16]	NCT02469090	Sublingual apomorphine 10–35 mg (N = 54)	62.9 (9.79)	37 (69)	NR	NR	Some concern
			Placebo (N = 55)	62.5 (8.12)	31 (56)
2	Katzenschlager et al. 2018 [17]	NCT02006121	Apomorphine hydrochloride injection 3-8 mg (N = 53)	63.6 (9.3)	34 (64.2)	6.69 (2.33)	NR	Low risk
			Placebo (N = 53)	63 (8.3)	32 (60.4)	6.76 (2.51)
3	LeWitt et al. 2018 [20]	NCT01617135	Inhaled levodopa 25–50 mg (N = 43)	62 (8.4)	25 (58.1)	1.5 (NR)—self-report 2.2 (NR)—diary data	853	Some concern
			Placebo (N = 43)	62.7 (9.1)	32 (74.4)	1.6 (NR)—self-report 1.8 (NR)—diary data	687
4	LeWitt et al. 2016 [18]	NCT01777555	Inhaled levodopa 35–50 mg (N = 43)	62 (8.4)	25 (58)	NR	687	Some concern
			Placebo (N = 43)	62.7 (9.1)	32 (74)	NR	853
5	LeWitt et al. 2019 [19]	NCT02240030	Inhaled levodopa 84 mg (N = 114)	63.5 (7.97)	83 (73)	5.35 (2.26)	818.6	Some concern
			Inhaled levodopa 60 mg (N = 113)	63.9 (9.24)	80 (71)	5.6 (1.9)	822.7
			Placebo (N = 112)	62.6 (8.83)	86 (77)	5.59 (2.25)	841.4
6	Agbo et al. 2021 [21]	NCT03292016	Apomorphine hydrochloride injection and sublingual apomorphine 15–30 mg (N = 8)	67.3 (8.7)	6 (75)	NR	NR	Some concern
7	Dewey et al. 2001 [8]	Dewey et al. 2001 [8]	Apomorphine hydrochloride 5.4 mg (N = 20)	66 (2)	12 (60)	5.9 (0.5)	776	Some concern
			Placebo (N = 9)	62 (4)	8 (89)	5.9 (0.8)	819
8	Muller et al. 2004 [22]	Muller et al. 2004 [22]	Previously untreated with subcutaneous apomorphine 2-6 mg (N = 11)	57.9 (9.7)	6 (55)	NR	NR	Some concern
			Previously treated with subcutaneous apomorphine 2-6 mg (N = 11)	59.5 (6.9)	10 (9.1)
			Placebo (N = 11)	55.3 (13.5)	6 (55)
9	Nomoto et al. 2015 [23]	Nomoto et al. 2015 [23]	Apomorphine hydrochloride 3.4 mg (N = 10)	57.7 (11.4)	4 (40.0)	6.58 (2.38)	NR	Some concern
			Placebo (N = 6)	58.8 (5.3)	1 (16.7)	7.75 (3.84)
10	Pfeiffer et al. 2007 [24]	Pfeiffer et al. 2007 [24]	Subcutaneous apomorphine 1.5–10 mg (N = 35)	64.8 (1.5)	25 (71.4)	NR	NR	Low risk
			Placebo (N = 13)	66.9 (3)	8 (61.5)
11	Van Laar et al. 1993 [15]	Van Laar et al. 1993 [15]	Subcutaneous apomorphine and placebo 2.7 mg (N = 27)	54.2 (NR)	3 (60%)	NR	NR	Some concern

NCT National Clinical Trial, NR not reported, RoB risk of bias, SD standard deviation

The feasibility assessment was conducted on 11 RCTs [8, 15–24]; ten clinical trials were excluded due to lack of control group (Fig. 1). Based on the structure and connectivity of the available evidence, the feasibility assessment revealed that NMAs were feasible for change in OFF-time, change in UPDRS Part III scores, improvement in PGI-C score, all-cause discontinuation rate, AE rate, and rate of treatment discontinuation due to AEs (Fig. 2; ESM Table S5). It should be noted that the evidence networks for all outcomes were largely structured around placebo-controlled trials and it was not possible to incorporate dispersible levodopa into the networks due to the absence of connected trials.

Fig. 2.

Networks of studies for all feasible endpoints. Treatments are color coded, with yellow representing placebo; orange, inhaled levodopa; green, subcutaneous apomorphine; purple, sublingual apomorphine. a Change in OFF-time, b change in UPDRS III scores, c improvement in PGI-C, d all-cause treatment discontinuation, e AEs, f treatment discontinuation due to AEs. AEs Adverse events, APO apomorphine, PGI-C Patient Global Impression of Change score, PLC placebo, S.C. subcutaneous, S.L. sublingual, UPDRS Unified Parkinson's Disease Rating Scale

A total of 855 patients with PD experiencing motor complications were included in the RCTs considered to be eligible for the NMA based on the feasibility assessment. Among these, 165 patients were randomized to subcutaneous apomorphine, 58 to sublingual apomorphine, 290 to inhaled levodopa, and 342 to placebo.

At the time of enrollment, the mean age of participants was 61.9 years. Most studies had a higher proportion of men than women. Detailed patient characteristics are provided in Table 1.

Efficacy Outcomes

Pairwise comparison showed both inhaled levodopa and subcutaneous apomorphine tended towards being more effective than placebo in change in OFF time, with standardized mean differences (SMD) of − 0.133 (95% credible interval [CrI] − 3.052, 2.889) and − 0.908 (95% CrI: − 3.657, 1.837), respectively (Fig. 3a); However, the credible intervals were wide and results were not statistically significant. When inhaled levodopa and subcutaneous apomorphine were compared, no statistically significant difference was observed, with a SMD of 0.740 and 95% CrI overlapping 0 (95% CrI − 2.015, 3.484).

Fig. 3.

Pairwise treatment comparison. a Cells shown in pink are SMDs (95% CrI) for change in OFF time and cells shown in blue are SMDs (95% CrI) for change in UPDRS Part III scores. SMD values < 0 indicate greater effect. Values represent the row-defining treatment compared with the column-defining treatment. Bold font indicates statistically significant results. b Cells shown in pink are Log ORs (95% CrI) for improvement in PGI-C and cells shown in blue are Log ORs (95% CrI) for all-cause discontinuation. Log OR values < 0 indicate lower risk. Values represent the row-defining treatment compared with the column-defining treatment. Bold font indicates statistically significant results. c Cells shown in pink are Log ORs (95% CrI) for AEs and cells shown in blue are Log ORs (95% CrI) for treatment discontinuation due to AEs. Log OR values < 0 indicate lower risk. Values represent the row-defining treatment compared with the column-defining treatment. Bold font indicates significant results. AEs Adverse events, APO apomorphine, CrI credible interval, INL inhaled levodopa, NA not available, OR odds ratio, PGI-C Patient Global Impression of Change score, PLC placebo, S.C. subcutaneous, S.L. sublingual, SMD standardized mean difference, UPDRS Unified Parkinson's Disease Rating Scale

For reduction in UPDRS Part III scores, inhaled levodopa and sublingual apomorphine showed numerical but non-statistically significant improvements in pairwise comparisons versus placebo, with SMD of − 0.695 and wide credible intervals (95% CrI − 3.030, 1.649) and − 1.664 (95% CrI: − 4.204, 0.895). Subcutaneous apomorphine was the only treatment to show statistically significant improvements in UPDRS Part III scores compared with placebo, with SMD of − 3.609 (95% CrI − 5.789, − 1.557; Fig. 3a). When treatments were compared, no statistically significant differences were observed between inhaled levodopa and subcutaneous or sublingual apomorphine, with SMDs of 1.989 (95% CrI − 0.453, 4.522) and 0.963 (95% CrI − 1.664, 3.581), respectively. No statistically significant difference was observed between the subcutaneous and sublingual apomorphine (SMD − 0.583; 95% CrI − 3.231, 1.975).

Forest plots for absolute outcomes for change in OFF time and UPDRS Part III scores are shown in ESM Fig. S1.

Patient-Reported Outcomes

In pairwise comparisons, improvement in PGI-C scores was assessed by Log odds ratios (OR). Inhaled levodopa, subcutaneous apomorphine and sublingual apomorphine showed numerical improvements versus placebo with Log ORs of 0.470 (95% CrI: − 4.220, 5.242), 1.057 (95% CrI: − 3.761, 5.883) and 0.373 (95% CrI: − 4.440, 5.037), respectively, though not statistically significant and with wide credible intervals (Fig. 3b). When comparing between treatments, no statistically significant differences were observed for improvement in PGI-C between subcutaneous or sublingual apomorphine and inhaled levodopa with Log ORs of − 0.587 (95% CrI: − 7.262, 6.163) and 0.098 (95% CrI: − 6.520, 6.831), respectively. No statistically significant difference was observed between the subcutaneous and sublingual apomorphine (Log OR 0.685; 95% CrI: − 5.971, 7.538).

A forest plot for absolute outcomes for improvements in PGI-C is shown in ESM Fig. S1c.

Safety Outcomes

In pairwise comparisons, all-cause treatment discontinuation, rates of AEs, and treatment discontinuation due to AEs were assessed by Log ORs. For all-cause treatment discontinuation, inhaled levodopa, subcutaneous apomorphine, and sublingual apomorphine tended towards being more effective than placebo, with Log ORs of − 0.065 (95% CrI − 1.748, 1.630), 0.449 (95% CrI − 1.153, 3.331), and 1.024 (95% CrI − 1.449, 3.540), respectively. However, the differences were not statistically significant (Fig. 3b), and no statistically significant differences were observed between treatments, with Log ORs between inhaled levodopa and subcutaneous apomorphine and sublingual apomorphine of − 0.513 (95% CrI − 3.801, 1.640) and − 1.089 (95% CrI − 4.149, 1.851), respectively, and Log OR between subcutaneous and sublingual apomorphine of − 0.576 (95% CrI − 3.371, 3.368). In the random effects base case model that assessed absolute outcomes, no statistically significant difference was observed in all-cause treatment discontinuation between inhaled levodopa (mean probability 0.222; 95% CrI 0.002, 0.912), subcutaneous apomorphine (mean probability 0.365; 95% CrI 0.009, 0.998), and sublingual apomorphine (mean probability 0.602; 95% CrI: 0.005, 0.999) (ESM Fig. S1).

The results were similar for rates of AEs, with treatment with inhaled levodopa, subcutaneous apomorphine, and sublingual apomorphine tending towards being more effective than placebo, with Log ORs of 0.061 (95% CrI − 0.776, 0.961), 0.315 (95% CrI − 0.586, 1.079), and 1.013 (95% CrI − 0.521, 2.538), respectively, although the differences were not statistically significant (Fig. 3c). Similarly, no statistically significant differences for AEs were observed between treatments, with Log ORs for inhaled levodopa versus subcutaneous apomorphine and sublingual apomorphine of − 0.254 (95% CrI − 1.391, 1.013) and − 0.952 (95% CrI − 2.696, 0.830), respectively, and Log OR for subcutaneous apomorphine versus sublingual apomorphine of − 0.698 (95% CrI − 2.505, 0.977). In the random effects base case model assessing absolute rates of AEs, no statistically significant differences in AE rates were observed between inhaled levodopa (mean probability 0.556; 95% CrI 0.073, 0.964), subcutaneous apomorphine (mean probability 0.659; 95% CrI 0.119, 0.976), and sublingual apomorphine (mean probability 0.836; 95% CrI 0.171, 0.998) ) (ESM Fig. S1).

For treatment discontinuation due to AEs, pairwise comparisons showed no statistically significant differences for inhaled levodopa and sublingual apomorphine compared to placebo, with Log ORs of − 0.120 (95% CrI − 1.396, 1.105) and 0.615 (95% CrI − 1.499 2.651), respectively. However, subcutaneous apomorphine resulted in a significantly higher probability of treatment discontinuation due to AEs than placebo (Log OR 20.593; 95% CrI 1.755, 55.614). In the pairwise comparison between treatments, inhaled levodopa had significantly lower probability of treatment discontinuation due to AEs than subcutaneous apomorphine, with Log OR of − 20.712 (95% CrI − 55.991, − 1.855). Subcutaneous apomorphine also had a significantly higher probability of treatment discontinuation due to AEs than sublingual apomorphine (Log OR 19.978; 95% CrI 0.959, 54.909), although no statistically significant difference was observed between inhaled levodopa and sublingual apomorphine (Log OR − 0.734; 95% CrI − 3.201, 1.689). In the random effects base case model assessing absolute outcomes, subcutaneous apomorphine also resulted in significantly higher probability of AEs leading to treatment discontinuation (mean probability 0.980; 95% CrI: 0.684, 1) compared to inhaled levodopa (mean probabilities 0.075; 95% CrI 0.001, 0.458) and placebo (mean probability 0.254; 95% CrI 0.001, 0.966) (ESM Fig. S1).

Discussion

The analysis presented here shows that in patients with OFF episodes in PD, inhaled levodopa was not statistically significantly different from subcutaneous and sublingual apomorphine in terms of the treatment of motor fluctuations, including OFF-time and change in UPDRS Part III scores. Improvement in PGI-C, all-cause treatment discontinuation, and rates of AEs were also not statistically significantly different between inhaled levodopa and subcutaneous and sublingual apomorphine. However, inhaled levodopa did show statistically significant lower rates of treatment discontinuation due to AEs compared with subcutaneous apomorphine, although the difference was not statistically significant compared to sublingual apomorphine. While statistically significant differences were not shown for most of the outcomes in this study, this should not be interpreted as equivalence, and the analysis performed cannot provide evidence of comparability or non-inferiority between the on-demand treatments.

Inhaled levodopa is the first and only licensed, inhaled on-demand therapy for patients with OFF episodes in PD [7]. Oral treatments, such as dispersible levodopa, which is sometimes used for the treatment of OFF episodes, have limitations regarding absorption and bioavailability. Studies have shown that levodopa plasma levels vary greatly depending on how it is taken and the motility of the gastrointestinal system [9–11]. For example, gastroparesis, which is common in patients with PD due to autonomic dysfunction, delays gastric emptying and disrupts levodopa absorption [9, 25]. High protein meals have also been shown to reduce levodopa absorption [9, 11]. Rapid-release oral levodopa is also available, but is not licensed for use as an on-demand treatment.

Pulmonary delivery of the treatment bypasses the gastrointestinal tract and provides rapid absorption into the systemic circulation. It also alleviates the need for additional pills and invasive administration techniques in line with patient preferences [12]. Inhaled levodopa has shown significant improvements in efficacy outcomes compared with placebo, as well as in patient-reported outcomes, with clinically meaningful improvements in PGI-C in phase 3 clinical trials and an open-label extension study, highlighting the improvements in quality of life, which as shown in the analysis presented here are not significantly different to improvements with apomorphine treatment [19, 26].

From a safety perspective, all-cause discontinuation and rates of AEs were not statistically significantly different for inhaled levodopa and apomorphine. However, while rates of treatment discontinuation due to AEs were not significantly different for inhaled levodopa and sublingual apomorphine, inhaled levodopa was associated with significantly lower rates of discontinuation due to AEs compared with subcutaneous apomorphine.

Phase 3 clinical trials have shown that subcutaneous apomorphine improves UPDRS Part III scores and OFF episodes as well as non-motor features of PD [8, 23, 24]. AEs associated with apomorphine include nausea, vomiting, orthostatic hypotension, somnolence, and dyskinesia, in addition to injection-site reactions. Many patients discontinue apomorphine primarily due to AEs. A prospective study of patients with PD receiving on-demand subcutaneous apomorphine reported that > 60% discontinued treatment, with a mean time to discontinuation of 10.3 months [27]. Another study reported the rate of discontinuation for subcutaneous apomorphine as 15.8% within the first 3 months [28]. Rates of discontinuation for sublingual apomorphine have been reported to be lower, with 17.3% of patients discontinuing treatment over a 2-year period in a long-term, open label study [29]. In contrast, the safety profile of inhaled levodopa is similar to that associated with treatment the patients will already be receiving, with the addition of pulmonary-related AEs. Cough is the most commonly reported AE with inhaled levodopa, followed by throat irritation, upper respiratory tract infection, and nausea; however, data from an open-label extension study suggest that 71% of patients remained on inhaled levodopa at 12 months after the start of treatment [26].

Subcutaneous apomorphine also requires careful dose titration to determine the most effective and tolerable dose. The titration period can be complex and require medical supervision, while mitigation of AEs may further slow the titration period. This titration period, in addition to the requirement for an injection pen, results in the treatment not being widely accepted by patients, with reports of 93% of prescribers identifying patient reluctance to be focused around the use of subcutaneous apomorphine, primarily from apprehension around self-injection (79%) but also due to technical challenges (14%) [30].

The aim of treatment with sublingual apomorphine is—as with inhaled levodopa—to bypass the gastrointestinal tract and deliver apomorphine systemically via absorption from the oral mucosa [7]. However compared with subcutaneous apomorphine, sublingual apomorphine has lower peak drug concentration, longer time to maximal concentration, longer elimination half-life, and approximately sevenfold lower bioavailability [7]. Some limitations of current treatments can be overcome by applying a patient-focused approach to choosing an appropriate on-demand treatment. When delivered via the pulmonary route, inhaled levodopa overcomes a number of barriers to absorption, bioavailability, dose titration, and compliance. The analysis presented here, while not able to show comparability or non-inferiority, does show that in the studies analyzed there is no statistically significant difference between inhaled levodopa and other on-demand treatments.

The analysis presented here has several important limitations, including the inclusion of only RCTs in the analysis and limited real-world evidence on on-demand treatments of OFF episodes, as well as additional limitations regarding bias and confounding factors. Searches were also limited to English-language publications, introducing a potential language bias.

The analysis did not consider differences in study design which may affect the results. For example, in terms of interventions, different doses of inhaled levodopa in the trials were pooled together; dose-specific sensitivity analyses were not performed as the number of available studies and dose-stratified observations for inhaled levodopa could not support meaningful subgroup analyses within the NMA framework. In addition, the inclusion of Katzenschlager et al.’s evaluation of continuous subcutaneous apomorphine infusion may introduce clinical heterogeneity and could affect the transitivity assumption underlying the NMA [17]. However, subcutaneous apomorphine infusion, while not a traditional on-demand therapy, is used to treat OFF episodes in patients who require multiple apomorphine injections per day, and the NMA feasibility assessment did not identify substantial differences in the study compared with others included in the study. The sensitivity analysis excluding data for subcutaneous apomorphine infusion showed similar results to the base-case analysis for OFF-time and AEs leading to treatment discontinuation, but subcutaneous apomorphine showed statistically significant improvements in UPDRS Part III compared with inhaled levodopa while inhaled levodopa showed statistically significant improvements in all-cause treatment discontinuation compared with subcutaneous apomorphine.

Regarding outcomes, UPDRS Part III and Movement Disorder Society UPDRS Part III, both measures of motor impairment, were also treated as directly comparable continuous outcomes in the analysis, with no formal conversion applied between the two scales. The severity and nature of AEs were not considered in this analysis, and only the occurrence of any AE was used. In the analysis, all-cause discontinuation, AEs, and treatment discontinuation due to AEs were considered to be binary endpoints using the probability of events occurring in order to account for differences between studies. Study durations and follow-up also varied.

All of the trials included in the networks had relatively small sample sizes, and networks were sparse, focusing around placebo, leading to greater uncertainty in the estimated endpoints. This is particularly the case for sublingual apomorphine where only 54 patients were included from the Olanow et al. trial [16], while data from Katzenschlager et al. were included in the analysis to increase available data on subcutaneous apomorphine, despite these authors reporting outcomes for continuous apomorphine infusion [17]. The limited study sizes and sparse networks highlight the lack of data on the use of on-demand treatments for OFF episodes and the need for further evidence generation. The findings of this study are further limited by wide credible intervals that often overlap the null value, resulting in a lack of statistical significance between comparators. Moreover, the analysis of AEs leading to treatment discontinuation returned particularly wide credible intervals as the small sample size was also compounded by a very low number of events (sometimes zero-events were reported in the trial). These factors should be taken into account when drawing conclusions from the study.

Nevertheless, despite the associated limitations, the study does demonstrate strong methodological rigor through the use of a systematic search of the literature and use of an NMA. The systematic review of the literature followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, and the Cochrane Handbook for Systematic Reviews of Interventions in addition to the NICE guidelines for technology appraisals to ensure methodological quality and transparency. The NMA included all identified trials that were deemed feasible, which all included publicly available data, and followed the NICE Decision Support Unit’s Technical Support Document [14], applying a previously published approach [31].

Conclusion

The efficacy of inhaled levodopa was demonstrated to be not significantly different to that of subcutaneous and sublingual apomorphine across multiple endpoints, including reduction in OFF-time, improvement in motor function (UPDRS-III), and patient-reported outcomes (PGI-C). All-cause discontinuation and rate of AEs were not different between treatments; however, the probability of treatment discontinuation due to AEs was notably higher for subcutaneous apomorphine compared to inhaled levodopa. Given the limitations of the evidence base and wide credible intervals, these data should be interpreted with caution, however the lower rate of discontinuation due to AEs coupled with non-significant differences in efficacy highlights inhaled levodopa as an alternative for patients seeking a non-invasive, well-tolerated, on-demand treatment option for managing OFF episodes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Study conception and design: Omar Madani, Chetan Mistry, Stefano Perni, Ubong Silas, Carmen Walter, Georg-Alexander Pietsch. Data analysis and investigation: Omar Madani, Chetan Mistry, Rushab Shah, Grace Chattey, Leandro Guerra Primo, Stefano Perni. Writing original draft preparation: Omar Madani, Rushab Shah, Hanne C Gadeberg. Writing review and editing: Chetan Mistry, Grace Chattey, Hanne C Gadeberg, Stefano Perni, Ubong Silas, Carmen Walter, Georg-Alexander Pietsch, Günter Höglinger. Supervision: Leandro Guerra Primo, Hanne C Gadeberg, Stefano Perni, Ubong Silas, Günter Höglinger. All authors read and approved of the final version of the manuscript.

Funding

The study was funded by Merz Therapeutics GmbH. The journal’s Rapid Service Fee was funded by Merz Therapeutics GmbH (Frankfurt am Main, Germany).

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Omar Madani, Chetan Mistry, Rushab Shah, Grace Chattey, Leandro Guerra Primo, Hanne C Gadeberg, Stefano Perni and Günter Höglinger have no competing interest other than the study being funded by Merz Therapeutics GmbH. Ubong Silas, Carmen Walter and Georg-Alexander Pietsch are employees of Merz Therapeutics GmbH.

Ethical Approval

This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.