The relation between sleep and weight in a suburban sleep center: observations and speculations on apnea and weight

The relationship between obstructive sleep apnea (OSA) and body weight is not clearly established. In order to describe the relationship of weight and OSA severity seen in a suburban sleep center, an observational review was performed of initial diagnostic polysomnograms (PSGs) ordered on patients with American Academy of Sleep Medicine (AASM) symptomatic indications.

Initial, full-night diagnostic or initial split-night (diagnostic portion) PSGs performed for any indication on patients >18 years old were retrospectively reviewed for a two year period. All studies were performed following AASM guidelines. PSG data were reviewed for the presence and severity of apnea (no OSA - apnea hypopnea index (AHI) <5, mild - AHI 5-14, moderate - AHI 15-29, severe - AHI 30-59, and very severe - AHI >60). Data were reviewed from 629 PSGs (37% females and 63% males) of which 450 met the criteria for apnea. Studies were classified by apnea severity (196 mild, 103 moderate, 91 severe apnea and 60 with very severe apnea) and weight (body mass index (BMI)). Of those with apnea, and BMIs <25, severe or very severe apnea occurred in 22% (10/45). Three individuals with BMIs <20 had apnea, one severe. Of those with BMIs ≥40, one (1.6%) did not have apnea and 52% (31/60) had AHI >30.

The profile of this nonrandom series, tested because they were suspected of having a disorder of sleep, provides guidelines for physicians in their approach to symptomatic patients. Individuals with a normal BMI can have apnea, including severe apnea. Severe obesity (BMI >40) is almost always associated with apnea when symptoms are present. Obesity increases the severity of the diagnosed apnea. Excessive weight should be an indication for testing, but normal weight should not exclude individuals with appropriate symptoms. Obesity, while a major contributing factor to severity, is not the etiological cause of OSA in the majority of these patients.