ABSTRACT
Our understanding of the pathophysiology of obstructive sleep apnea (OSA) is constantly evolving, and therefore it is not surprising that the approach to diagnostic tests should also be evolving. In the early days of OSA, patients presented with hypersomnolence and were investigated in neurophysiology sleep laboratories, usually run by experienced neurologists or psychiatrists. These laboratories had a long history of sleep investigation, centered on the EEG and its varying patterns during sleep. Routine sleep studies in the 1960s consisted of EEG, EMG, and EOG monitoring in order to stage sleep, with particular interest in rapid eye movement (REM) sleep and its latency. Indeed in some laboratories a reduction in REM latency was part of the diagnosis of endogenous depression, and many sleep studies were carried out in the investigation of this disorder (1). It is not surprising therefore that when the first comprehensive reports of OSA began to appear they came from neurophysiology units with great expertise in sleep staging and polysomnography (PSG) (24). Emphasis was placed on neurophysiological monitoring, demonstrating poor sleep and the predominance of apneas during REM sleep. Simple signal transducers, such as oro-nasal thermistors and ribcage/abdominal mercury strain gauges, were
458 Stradling and Davies
added to polysomnography to document the irregular breathing that was usually dramatically abnormal. Since the obstructive apneas were so obvious, these were the first abnormalities to be defined, counted, and then referred to as the apnea index (the number of apneas per hour of sleep). At the time it seemed obvious that these were the primary abnormality that caused the patient’s presenting symptom of hypersomnolence. However, there was no way to prove that all the patient’s symptoms were due only to the apneas, although none of a group of 20 normal subjects (aged 40-60 years) had 4 apneas per hour (5). In the absence of anything better, or significant dissent, the standard way to investigate these patients became full polysomnography with EEG, EMG, EOG, oro-nasal airflow, ribcage, abdominal monitoring, with SaO2 added later following the wider availability of reliable pulse oximeters. No analysis was ever made as to whether this approach accurately defined the disease in a quantitative way that could produce useful thresholds for predicting likely therapeutic responses and benefits, or indeed adverse consequences. However, this approach identified thousands of patients who were helped initially by tracheotomy and subsequently, since 1981, nasal continuous positive airway pressure (CPAP) (6). This technique became established in the literature in a way that was perhaps not justified by its origins, based as they were on face validity, rather than a full understanding of the condition. It became the gold standard of investigation against which to assess other techniques, which would fall by the wayside if they could not mimic an apnea index derived from polysomnography. Full polysomnography was defended on a variety of grounds, including the importance of proving sleep and demonstrating REM periods, but, as pointed out by Guilleminault and Partinen in 1990 (7), reimbursement arrangements related to this expensive investigation may have played their part in its persistence.
