Unmonitored titration of therapeutic continuous positive airway pressure (CPAP) using autotitrating CPAP devices is being used for patients with sleep apnea as a lower cost alternative to formal CPAP titration in a sleep laboratory. However, this strategy cannot detect persistent nocturnal desaturations despite CPAP therapy.
In an attempt to identify predictors of patients who might require supplemental oxygen therapy while on CPAP, we identified 69 patients with significant nocturnal desaturations (defined as oxygen saturation < 90% for >10% of study) out of 483 patients consecutively diagnosed with sleep apnea on unattended cardiopulmonary sleep testing between July 2003 and March 2004.
Patients with nocturnal desaturations were predominantly male with a mean age of 67, and mean body-mass index of 36.3 (range 26.6 to 61.0). Congestive heart failure was present in 11.6%, and chronic obstructive pulmonary disease in 13%. During cardiopulmonary sleep testing, mean apnea-hypopnea index (AHI) was 40.7 (range 2.6 to 95.4), mean oxygen desaturation index (ODI) was 41.8 (range 1.3 to 101), mean oxygen saturation was 90.1 + 2.7 %, mean nadir oxygen saturation was 70.1 + 12.1 %, and mean percent of study with oxygen saturation < 90% was 34.8 + 21.9 %. Mean therapeutic CPAP pressures were 11.6 + 3.1 cm H2O, with mean AHI on CPAP of 4.35 + 6.84. Of these 69 patients, 13 (18.8%) had persistent significant desaturations when restudied with overnight oximetry while on therapeutic CPAP. Logistic regression analysis showed three variables that were significantly and independently associated with persistent desaturation on CPAP: 1) COPD; 2) the percent of time with oxygen saturation < 90% during cardiopulmonary study; and 3) percent improvement in AHI on CPAP. However, none of the variables evaluated accurately predicted which individual patients would have significant nocturnal desaturations on CPAP therapy.
We therefore caution practitioners to consider overnight oximetry studies for sleep apnea patients with nocturnal desaturations on diagnostic studies who have had unattended CPAP titration.
J.B. Rubins, None.