Heterogeneity in Estimates of Incidence and Prevalence of Narcolepsy: A Systematic Review and Meta-Regression Analysis

Abstract

Introduction: Narcolepsy is a chronic neurological disorder. The diagnostic criteria of narcolepsy evolve from clinical symptoms to molecular biomarkers, along with the understanding of its clinical nature and pathogenesis. Estimates of incidence and prevalence of narcolepsy vary between studies, while the contribution of changing diagnostic criteria to the variation remains unclear. We aimed to explore sources of heterogeneity in estimates of incidence and prevalence, with a particular focus on diagnostic criteria. Methods: We searched 5 databases for observational studies on the incidence or prevalence of narcolepsy published before October 14, 2021. Subgroup analyses and meta-regression were used to assess the impact of diagnostic criteria on incidence/prevalence of narcolepsy after adjusting for age group, region, study period, vaccination status, index date, and type of narcolepsy. Results: 35 studies were selected from 2,833 articles. The estimates of incidence and prevalence were wide-ranging with high heterogeneity (incidence I2 = 99.8%; prevalence I2 = 99.7%), from 0.06–6.56 per 100,000 person-years for incidence and 1.05–79.40 per 100,000 population for prevalence, respectively. Totally 10 diagnostic criteria were used, including the 1st revised edition of International Classification of Diseases (ICSD-1), ICSD-2, ICSD-3, the 8th revision of International Classification of Diseases (ICD-8), ICD-9, ICD-10, Brighton collaboration case definition (Brighton), Mayo classification, the Ullanlinna Narcolepsy Scale, and clinical symptoms with the multiple sleep latency test. ICD tended to provide higher estimates of incidence/prevalence than Brighton (incidence odds ratio [OR] 1.38, [95% CI 1.02, 1.86]; prevalence OR 1.50, [95% CI 1.04, 2.39]). No significant difference was found in estimates of two rates between ICSD and Brighton. Incidence was higher for children than adults (OR 1.61, [95% CI 1.25, 2.07]), and for individuals vaccinated with Pandemrix than those unvaccinated (OR 6.49, [95% CI 3.86, 10.91]). Conclusions: Estimates of incidence/prevalence of narcolepsy could not be pooled reliably with substantial heterogeneity. Incidence/prevalence studies using ICSD and Brighton provided lower estimates than studies using ICD and other criteria. Diagnostic criteria should be standardized when comparing or pooling the incidence/prevalence to understand the epidemiology of narcolepsy. Future studies are needed to focus on the at-risk population for the etiology investigation of narcolepsy.

Prevalence and incidence of narcolepsy symptoms in the US general population

“…The accuracy of narcolepsy prevalence estimates can be affected not only by the type of sample used, but also by differences in diagnostic criteria, potentially resulting in overestimation. A meta-regression analysis by Wang et al [ 13 ] revealed that having stringent diagnostic criteria are essential to determine the true prevalence of narcolepsy, while the use of non-specific criteria can lead to overestimates.…”