A screening tool for obstructive sleep apnoea (OSA) is useful in low-income countries where it may be difficult to access sleep recordings. The objective of this study was to assess the performance of six screening scores compared with objective sleep recording. This analysis is based on the “Benin Sleep and Society” (BeSAS) study, in which respiratory polygraphy (PG) was performed using a type III device and OSA screening questionnaires (STOP, STOP-Bang, Berlin, NOSAS [≥8 and ≥5), No-Apnea, GOAL) were administered to participants. PG-defined OSA severity categories were defined according to the apnoea-hypopnoea index (AHI): mild (AHI 5 to <15/h), moderate (AHI 15 to <30/h) or severe (AHI ≥30/h), and these were compared to score findings. A total of 1810 subjects (mean age 45.4±14.6 years; 64.3% women) were included. For moderate to severe OSA, the area under the receiver operating characteristic curve was greatest for GOAL and No-Apnea (0.70, CI 0.67-0.73; 0.70, CI 0.68-0.73), followed by NoSAS 5 (0.69, CI 0.65-0.73). The highest sensitivity values were for NoSAS 5 (0.73, CI 0.67-0.79), No-Apnea (0.72, CI 0.66-0.78), and GOAL (0.69, CI 0.67-0.79), while NoSAS 8 had the highest specificity (0.91, CI 0.90-0.93), followed by Berlin (0.88, CI 0.87-0.90) and GOAL (0.71, CI 0.69-0.73). All scores performed poorly with respect to the positive predictive value (PPV), which was highest with NoSAS 8 (0.38, CI 0.31-0.44). Although overall performance was modest for all instruments, NoSAS 8 showed the highest PPV among the evaluated scores. These findings suggest that NoSAS 8 may be a useful option to support OSA risk stratification in resource-constrained settings, but should be interpreted with caution. • NoSAS8 had highest specificity (91%) for moderate-to-severe OSA. • NoSAS8 showed highest positive predictive value (38%). • GOAL and No-Apnea showed highest sensitivity (73% and 72%). • First validation of multiple OSA scores in an African general population. • Underscores need for resource-adapted OSA screening in low-income settings
Wachinou et al. (Fri,) studied this question.