Association of Olfactory Performance With Motor Decline and Age at Onset in People With Parkinson Disease and the LRRK2 G2019S Variant

Abstract

Background and ObjectivesThere is clinical and phenotypic heterogeneity in LRRK2 G2019S Parkinson disease (PD), including loss of smell. Olfactory scores have defined subgroups of LRRK2 PD at baseline. We now extend this work longitudinally to better determine features associated with olfactory classes and to gain further insight into this heterogeneity.MethodsEvaluation of 162 patients with LRRK2 PD and 198 patients with idiopathic PD (IPD) from the LRRK2 Ashkenazi Jewish Consortium was performed, with follow-up available for 92 patients with LRRK2 PD and 74 patients with IPD. Olfaction (University of Pennsylvania Smell Identification Test [UPSIT]), motor function (Unified Parkinson Disease Rating Scale), and cognition (Montreal Cognitive Assessment), as well as sleep, nonmotor, and mood, were measured. Gaussian mixture models were applied on the UPSIT percentile score to determine subgroups based on olfactory performance. Linear mixed effects models, using PD duration as the time scale, assessed the relationship between UPSIT subgroup membership and motor/cognitive change.ResultsBaseline olfaction was better in LRRK2 PD compared with IPD (mean UPSIT ± SD: 24.2 ± 8.8 vs 18.9 ± 7.6), with higher mean percentile scores (difference: 15.3 ± 11.6) (p < 0.001) and less frequent hyposmia (55.6% vs 85.4%; p < 0.001). Analysis suggested 3 classes among LRRK2 PD. Age at onset in LRRK2 PD was earlier in the worst olfaction group (group 1), compared with groups 2 and 3 (54.5 ± 11.1 vs 61.7 ± 9.3) (p = 0.012), and separately in the hyposmic group overall (55.0 ± 11.3 vs 61.7 ± 9.1) (p < 0.001). Longitudinal motor deterioration in LRRK2 PD was also significantly faster in the worst UPSIT group than the best UPSIT group (group 3 vs group 1: B = 0.31, SE = 0.35 vs B = 0.96, SE = 0.28) (rate difference = −0.65, SE = 0.29) (p = 0.03). However, olfactory group membership was not significantly associated with cognitive decline.DiscussionIn this large LRRK2 cohort with longitudinal analysis, we extend prior work demonstrating subgroups defined by olfaction in LRRK2 G2019S PD and show that the worst olfaction group has earlier age at PD onset and more rapid motor decline. This supports a subgroup of LRRK2 PD that might show more rapid change in a clinical trial of LRRK2-related agents and highlights the need to integrate careful phenotyping into allocation schema in clinical trials of LRRK2-related agents.Classification of EvidenceThis study provides Class II evidence that worse olfactory scores were associated with an earlier age at symptomatic onset and a faster rate of motor deterioration in patients with LRRK2 PD.