Pihtili · Sleep medicine 2025 · multi-center cross-sectional registry study · n=7130

Sex differences in clinical and polysomnographic features of obstructive sleep apnea: The Turkish sleep apnea database (TURKAPNE) cohort.

Cited 6 times in the scientific literature.

Level 4 - case-series / case-control

Multi-center cross-sectional registry study

PubMed 39721358 · doi:10.1016/j.sleep.2024.12.018 · record verified 2026-08-28

What was done

Cross-sectional analysis of 7,130 adult patients (2,259 women) across 34 clinical centers in the ongoing Turkish Sleep Apnea Database (TURKAPNE) registry. Obstructive sleep apnea (OSA) was diagnosed and categorized by polysomnography using apnea-hypopnea index (AHI) cut-offs of ≥5 (mild), ≥15 (moderate), and ≥30 events/hour (severe). Clinical symptoms, polysomnographic metrics, and comorbidities were compared by sex.

What was found

OSA was confirmed in 6,323 participants (70.2% male, 29.8% female). Women with OSA were older (56.7 ± 11.9 vs. 49.5 ± 11.3 years; p < 0.001) and had higher BMI (34.3 ± 7.2 vs. 31.4 ± 5.6 kg/m²; p < 0.001), but lower mean AHI (29.8 ± 24.1 vs. 36.8 ± 26.2 events/h; p < 0.001) than men. Loud snoring and witnessed apnea were more frequent in men, whereas women more often reported insomnia, headache, and mood changes. Polysomnography showed women had significantly less total sleep time, lower sleep efficiency, and longer sleep latency (p < 0.001 each). Comorbidities (diabetes, hypertension, asthma, psychiatric disorders, hypothyroidism) and drug use were significantly more common in women independent of age and obesity (p < 0.05 each).

Why it matters

This study highlights that female OSA patients present with distinct, often atypical clinical features and higher comorbidity loads despite lower AHI severity. Recognizing sex-specific symptom profiles can reduce underdiagnosis and delayed clinical care in women.

Limits

The study is cross-sectional and registry-based, preventing causal conclusions or longitudinal tracking. The cohort consists of clinic-referred patients in Turkey, introducing potential referral bias, and specific numerical effect sizes (e.g., odds ratios for adjusted comorbidity models) are omitted in the abstract.

Cited by