Physician decision making and clinical outcomes with laboratory polysomnography or limited channel sleep studies for obstructive sleep apnea: a randomized trial

Abstract

Background:

The clinical utility of limited-channel sleep studies (which are increasingly conducted at home) versus laboratory polysomnography (PSG) for diagnosing obstructive sleep apnea (OSA) is unclear.

Objective:

To compare patient outcomes after PSG versus limited-channel studies.

Design:

Multicenter, randomized, noninferiority study. (Australian New Zealand Clinical Trials Registry: ACTRN12611000926932).

Setting:

7 academic sleep centers.

Participants:

Patients (n = 406) aged 25 to 80 years with suspected OSA.

Intervention:

Sleep study information disclosed to sleep physicians comprised level 1 (L1) PSG data (n = 135); level 3 (L3), which included airflow, thoracoabdominal bands, body position, electrocardiography, and oxygen saturation (n = 136); or level 4 (L4), which included oxygen saturation and heart rate (n = 135).

Measurements:

The primary outcome was change in Functional Outcomes of Sleep Questionnaire (FOSQ) score at 4 months. Secondary outcomes included the Epworth Sleepiness Scale (ESS), the Sleep Apnea Symptoms Questionnaire (SASQ), continuous positive airway pressure (CPAP) compliance, and physician decision making.

Results:

Change in FOSQ score was not inferior for L3 (mean difference [MD], 0.01 [95% CI, -0.47 to 0.49; P = 0.96]) or L4 (MD, -0.46 [CI, -0.94 to 0.02; P = 0.058]) versus L1 (noninferiority margin [NIM], -1.0). Compared with L1, change in ESS score was not inferior for L3 (MD, 0.08 [CI, -0.98 to 1.13; P = 0.89]) but was inconclusive for L4 (MD, 1.30 [CI, 0.26 to 2.35; P = 0.015]) (NIM, 2.0). For L4 versus L1, there was less improvement in SASQ score (-17.8 vs. -24.7; P = 0.018), less CPAP use (4.5 vs. 5.3 hours per night; P = 0.04), and lower physician diagnostic confidence (P = 0.003).

Limitation:

Limited-channel studies were simulated by extracting laboratory PSG data and were not done in the home.

Conclusion:

The results support manually scored L3 testing in routine practice. Poorer outcomes with L4 testing may relate, in part, to reduced physician confidence.