0713 Variability of Software Scored PLM's

Introduction Periodic leg movements have been a source of debate as to their clinical significance. Interpretation varies between no clinical significance (Mahowald MW 2001) to a relationship to Heart disease ( Skomo R 2009). Routine technician scoring of PLM’s is rare outside of academic institutions. Insurances, including Medicare and Medicaid, will not approve a repeat study only for evaluation of PLM’s. The following brief study was initiated to review automated software scoring of periodic leg movements based on one sleep study per a patient (Respironics Alice G3). Methods Data was obtained from 56 sequential sleep studies with ~240 min of sleep (225 – 434, median 330). “PSG” represents all studies without titration. There were more females (17) than males (8), Ages (24 – 70, median 49.5). This data set is small and not a normal distribution. A technician scored minutes of sleep and AHI, (.4-96.7, median 5.4) while PLM's were autoscored (G3). Results There was significant variability in the software generated PLMI, Range 765 (4 –769) median 75.5. • There wasn’t a useful correlation between AGE and PLMI (r2 = .019); thus, of little predictive value in an individual patient’s single test. • The PLMI was not predicted nor differentiated by gender. • There wasn’t a useful correlation between PLMI being related to AHI (r2 = .006), also of little predictive value in an individual patient’s single test. Conclusion Most clinicians can rely only on one sleep study, one measure, of PLMI. This brief study suggests that PLM’s scored by software varies greatly when obtained from a single study. Software scored PLMI results must be used with caution in making diagnoses such as PLMD or criteria for RLS. A clinician might have to reject the reported PLMI result. Support (if any)