Optimizing MSLT Specificity in Narcolepsy With Cataplexy.

Abstract

Multiple sleep onset rapid eye movement (R) periods (SOREMPs) and a mean sleep latency of ≤8 minutes on the multiple sleep latency test (MSLT) are diagnostic criteria of narcolepsy (NC), but also occur in other conditions with increased sleep pressure, including insufficient sleep syndrome (ISS), sleep-disordered breathing (SDB), or Parkinson's disease (PD). These false positives are common, may create diagnostic uncertainty, and highlight the need for complementary MSLT measures with high specificity for NC. Detailed analysis of MSLT findings in 56 NC, 83 PD, 89 SDB, and 23 ISS patients, using receiver operating characteristic curves. A positive MSLT (mean sleep latency ≤ 8.0 minutes and ≥2 SOREMPs) was found in 53 NC (95%), 1 PD (1%), 8 SDB (9%), and 12 ISS patients (52%). MSLT-based differentiation between NC and non-NC patients was best when applying a mean R latency of ≤5 minutes (sensitivity/specificity/positive predictive value [PPV]: 49%/95%/96%) or a mean percentage of sleep stage R ≥ 40% (sensitivity/specificity/PPV: 60%/100%/100%) as cutoffs. When analyzing all 252 naps with SOREMPs in isolation, the combination of both R latency of ≤5 minutes and R percentage of ≥50% yielded a sensitivity/specificity/PPV of 50%/99%/99%. In addition, a sleep stage sequence with R occurring prior to N2 was more common in NC than in non-NC (71% vs. 32%, p < .001), and in combination with R percentage of ≥50% yielded a sensitivity/specificity/PPV of 53%/96%/97%. A better characterization of R sleep by latency, duration, and sleep stage sequence facilitates detection of false positives and, hence, contributes to a higher MSLT specificity in NC.