Effect of leukotriene receptor antagonist use on the future risk of Parkinson's disease in older patients with asthma

Abstract

Current treatments for Parkinson's disease focus on symptom management, with no therapies yet demonstrated to slow disease progression. Leukotriene receptor antagonists, widely used for asthma, have shown potential neuroprotective effects for Parkinson's disease in preclinical studies, but have also been associated with an elevated risk of neuropsychiatric events and sleep disorders. We assessed the effect of leukotriene receptor antagonist treatment on the risk of Parkinson's disease, neuropsychiatric events, and sleep disorders in patients with asthma aged over 50 years. We conducted a cohort study using the UK Clinical Practice Research Datalink between January 2000 and December 2020. The study emulated sequential target trials (n = 140) using observational data, comparing leukotriene receptor antagonist treatment to no leukotriene receptor antagonist treatment among patients aged 50-84 years with asthma. The primary outcome was the risk of incident Parkinson's disease, and the secondary outcomes were neuropsychiatric events (anxiety, depression, and psychosis), and sleep disorders. Propensity score matching was employed to minimize confounding. We used pooled logistic regression models to calculate risk ratios as observational analogues of intention-to-treat and per protocol effects. A total of 97 049 matched pairs were included in the analysis, with 573 Parkinson's disease cases observed in the leukotriene receptor antagonist group and 537 in the nonleukotriene receptor antagonist group over a median follow-up of 5.9 years and 5.7 years, respectively. No significant difference in Parkinson's disease risk was observed between the two groups in either the intention-to-treat analysis [10-year risk ratio: 1.09; 95% confidence interval (CI), 0.94-1.26] or the per protocol analysis (10-year risk ratio: 0.95; 95% CI, 0.75-1.16). However, there was a higher risk of depression (intention-to-treat effect: 10-year risk ratio: 1.12; 95% CI, 1.07-1.16; number-needed-to-harm = 93; per protocol effect: 10-year risk ratio: 1.15; 95% CI, 1.08-1.22; number-needed-to-harm = 75) and sleep disorders (intention-to-treat effect: 10-year risk ratio: 1.14; 95% CI, 1.11-1.19; number-needed-to-harm = 77; per protocol effect: 10-year risk ratio: 1.12; 95% CI, 1.06-1.19; number-needed-to-harm = 88) with leukotriene receptor antagonist treatment. No clear effect was observed for anxiety or psychosis. Leukotriene receptor antagonist treatment was not associated with an altered risk of Parkinson's disease among people aged 50-84 years with asthma but was linked to a higher incidence of neuropsychiatric events.

Keywords: Parkinson’s disease; asthma; leukotriene receptor antagonist; target trial emulation.

Graphical Abstract

Figure 1

Selection of patients from clinical practice research datalink (CPRD) for trial emulation. A total of 1 270 839 unique patients from the CPRD met the eligibility criteria. Under the sequential trial design, these patients contributed to 140 nested trials, resulting in 37 268 844 patient-trials. Of these, 97 054 patient-trials were eligible for the leukotriene receptor antagonist (LTRA) initiation arm and 37 171 790 for the no LTRA initiation arm. After propensity score matching, 97 049 patient-trials initiating LTRA were matched to 97 049 patient-trials not initiating LTRA. Numbers in parentheses represent unique individuals in each group. The number of LTRA initiators and noninitiators does not sum to the total number of eligible individuals because some contributed to both arms across different emulated trials.

Figure 2

Cumulative incidence for Parkinson’s disease with leukotriene receptor antagonist (LTRA) treatment and no LTRA treatment among patients with asthma. (A) intention-to-treat effect, 10-year risk ratio: 1.09 (95% CI, 0.94–1.26); (B) per protocol effect, 10-year risk ratio: 0.95 (95% CI, 0.75–1.16). The number of patients (n) included in each analysis under each treatment group was labelled in each panel. Cumulative incidence curves were constructed using pooled logistic regression models to estimate discrete-time hazards at 3-month intervals. Each data point represents the estimated cumulative risk at the corresponding time point. 95% CIs were calculated using nonparametric bootstrapping with 300 resamplings. No formal test statistic or P-value is reported due to the use of nonparametric bootstrap-based inference.