The Most Commonly Used Indicator of Sleep Disor- Dered Breathing (sdb) Severity Is the Respiratory Disturbance Index (rdi), Which Is Synonymous with the Apnea Hypopnea Index (ahi) in Many

THE MOST COMMONLY USED INDICATOR OF SLEEP DISORDERED BREATHING (SDB) SEVERITY IS THE RESPIRATORY DISTURBANCE INDEX (RDI), WHICH IS SYNONYMOUS WITH THE APNEA HYPOPNEA INDEX (AHI) IN MANY REPORTS. The RDI is the total number of apneas and hypopneas per hour of sleep. In clinical populations, the RDI is used to define the presence of SDB and thus identify individuals who require treatment.1,2 In addition, for epidemiologic studies, the RDI is used to quantify SDB status and to estimate SDB prevalence.1 Usually, the RDI is derived from information obtained after a single night of polysomnography. However, if the RDI exhibits considerable night-to-night variability, estimates of disease severity based on a single night study could be misleading. Previous investigations of RDI variability are conflicting. Some have suggested that RDI derived from a single night of polysomnography is a stable estimate of SDB severity, with excellent reliability for categorizing individuals as having or not having SDB.3-7 Other reports indicate that there may be considerable night-to-night variability in the RDI, leading to substantial risk of misclassification.8-14 However, most studies have been small,10,13 and almost all have utilized laboratory polysomnography.3,5,8-13,15 Data from unattended nonlaboratory sleep monitoring in the setting of an epidemiologic study are limited.16,17 It is generally recognized that measurements of sleep architecture are subject to a “first-night effect.” The first night of sleep in a laboratory in comparison to subsequent nights is characterized by more wakefulness, a longer initial sleep latency, greater amounts of stage 1 sleep, decreased REM sleep and more sleep fragmentation.18,19 In contrast, there may be less first-night effect observed when studies are conducted in the home environment.20-22 However, most of these latter studies were performed in small numbers of individuals, many of whom had insomnia.21,22 There are few data pertaining to variability of sleep architecture indices from a general population sample. The Sleep Heart Health Study (SHHS) is a large multicenter cohort study that explores the link between SDB and cardiovascular and cerebrovascular mortality and morbidity in the general population.23 From December 1995 through February 1998, unattended nocturnal polysomnograms (NPSG) were obtained from 6441 individuals at 10 geographic sites. This paper reports on a substudy performed on a sample of SHHS participants to determine the short-term variability of 2 nights of unattended nonlaboratory NPSG data collected several months