Does the number of hypopnoeas influence therapy in patients with obstructive sleep apnoea?

The definition of hypopnoea varies widely between sleep laboratories. This study sought to evaluate the impact upon therapeutic decision of the most commonly used one, which defines hypopnoea as a reduction in airflow of 50% (or more) accompanied by a decreased blood oxygen saturation of 4% (or more) (Agency for Health Care Policy and Research). We studied 109 consecutive patients (99 male) with an apnoea-hypopnoea index (AHI) greater than 5. The same investigator evaluated each patient twice. First, the hypopnoea index (HI) was unavailable to him by study design, and judgment was based upon the number of apnoeas and the clinical situation of the patient. Thereafter, he evaluated all patients again, in random order, this time taking into account the AHI and the clinical condition of the patient. Therapeutic decision was compared at the end of the study. Mean age was 54 +/- 11 years (+/- SD) and body mass index 33 +/- 6 kg m-2. Most patients had normal spirometry and day-time arterial blood gases. Mean apnoea index (AI) was 35 +/- 24 h-1. Therapeutic decision based on the AI lead to conservative treatment (diet and sleep hygiene) in 32 cases and to active treatment in 77 patients [76 continuous positive airway pressure (CPAP) and one surgery]. The AHI was 40 +/- 24 h-1. The consideration of the number of hypopnoeas lead to a therapeutic change only in four patients (3.7%) (chi 2 = 0.36, P > 0.3). These four patients had an AI of 15, 7, 3, and 9 h-1 and an AHI of 22, 10, 46, and 17 h-1, respectively. Combined with their clinical data, this increase was considered a requirement for active treatment (CPAP). The most commonly used definition of hypopnoea has a small impact upon therapeutic decision in patients with obstructive sleep apnoea syndrome (OSAS). We suggest the need to develop new definitions of hypopnoea that incorporate other variables (neurophysiological, cardiovascular) of potential clinical relevance. Our results also show that the outcome of any new definition has to be assessed rigorously in terms of its clinical impact.