How much sleep apnea is too much? Authors' response

Abstract

Our study [1] had two main purposes: first, to examine the occurrence of sleep-disordered breathing (SDB) in patients with chronic fatigue syndrome (CFS), and then to exclude patients with this confound in order to evaluate the structure of sleep in a group of patients with pure CFS to look for relations with the symptoms of fatigue and sleepiness. In order to do this, we had to define an upper limit for the occurrence of respiratory events that might explain the daytime symptoms. Martinez and Cassol are correct that the currently accepted cutpoint for abnormality proposed by the American Academy of Sleep Medicine (AASM) and Centers for Medicare & Medicaid Services (CMS) in the US is an apnea-hypopnea index (AHI) ≥ 5 per hour but less than 15 per hour – but only if symptoms such as EDS are present. However, the AHI is an insensitive measure of the severity of mild respiratory events (such as respiratory event-related arousals (RERAs)) as it requires either a 3 or 4% desaturation for the definition of hypopnea and this is frequently not present in respiratory events that cause arousal and may contribute to sleep disruption (RERA). For this reason, our cutpoint for SDB is based on a respiratory disturbance index (RDI) of 18 events per hour and not the AHI. The RDI includes RERAs along with the AHI, although we used a slightly different definition of hypopneas beginning with any event showing flow limitation rather than the current amplitude based entry criterion. The rationale for this is from an earlier paper by our group [2] that shows that the sensitivity and specificity for explaining EDS in a group composed of symptomatic but mild SDB and non-complaining controls were greater using the RDI criterion than the AHI criterion. Since that paper, it has become customary to use a cut-off of 15 events per hour for RDI in many applications. It is important to note that for RDI, using 5 events per hour will include essentially 80% of the normal population. Furthermore, as pointed out in our current paper [1], although we chose 18 events per hour as a cut-off to exclude three subjects (one with CFS and two normal subjects) when defining the final group for our sleep analysis who might have SDB, in fact the highest RDIs we saw in the non-SDB group were 10.4, 10.8, 10.1, and 9.5 events per hour; we feel these numbers are within a range seen in many normal subjects without EDS [2]. Thus, our cut-off of 18 is not effectively different from the more recently accepted number of 15 events per hour. And after excluding those subjects with high RDI values, there was no significant difference in the AHI of patients with only CFS, those with CFS and fibromyalgia and healthy controls (1.7 ± 1.6, 2.8 ± 2.6, and 2.2 ± 1.7, respectively). In fact, the above logic is not fundamentally different from the choice made by Martinez and Cassol in their own work [3-6], where they define SDB as present in fibromyalgia when the AHI <5 and when more than 10 RERAs per hour occurred (5 + 10 events = 15 per hour, similar to our cut-off of 18 per hour).