Abstract Introduction Single-night home sleep apnea test (HSAT) remains the standard method to diagnose obstructive sleep apnea (OSA). However, night-to-night variability can introduce diagnostic inconsistencies. How much these variations are influenced by specific oxygen desaturation criteria used is still unclear, highlighting the need for further clarification on their impact on measurement stability. Methods We identified 4,197 patients who completed multi-night HSAT monitoring (mean 2.86 nights, range 2-7), during which both 3% and 4% oxygen desaturation indices were recorded concurrently. For each night, we calculated the sleep-adjusted apnea hypopnea index (sAHI). Night-to-night variability for each patient was evaluated using the coefficient of variation (CV). We then compared variability measures between the 3% and 4% desaturation thresholds and examined how these differences might influence diagnostic reliability. Results A total of 4,197 patients (mean age 54.3±13.8 years) recorded 12,008 HSAT nights. Significant night-to-night variability in sAHI was observed for both desaturation criteria. Using the 3% desaturation threshold, the mean coefficient of variation (CV) was 25.41% (median 21.37%). Similarly, the variability observed using the 4% desaturation threshold had a mean CV of 30.20% (median 25.84%). Furthermore, when determining OSA severity from a single night of testing, 32.31% of patients were misclassified compared to their multi-night average severity using the 3% criterion. A misclassification rate of 32.57% was also observed with the 4% criterion. The statistically insignificant difference in misclassification rates between the two scoring methods (0.26 percentage points) indicates that both desaturation thresholds are comparably influenced by physiologic variability across nights. Conclusion Home sleep apnea testing shows considerable night-to-night fluctuation, regardless of whether a 3% or 4% desaturation threshold is applied. Nearly one-third of patients would receive an incorrect severity classification if diagnosis were solely based on a single night of testing. These results highlight the value of multi-night testing for improving diagnostic accuracy, avoiding treatment decisions based on unrepresentative data. The comparable variability seen with both desaturation criteria suggests that the challenge lies in the inherent physiologic variability of sleep apnea rather than in the choice of the threshold itself. Support (if any)
Jain et al. (Fri,) studied this question.