Cox regression analysis: associations of polysomnographic parameters and incident MACEs
| PSG parameters | Univariable model | Multivariable model* |
| HR (95% CI) | HR (95% CI) | |
| Total dataset (n=1860) | ||
| AHI (46.9 vs 8.0) | 1.24 (1.05 to 1.45) | 0.95 (0.76 to 1.17) |
| REM-AHI (54.4 vs 11.1) | 1.19 (0.98 to 1.44) | 1.03 (0.80 to 1.31) |
| 3% ODI (41.3 vs 4.9) | 1.24 (1.07 to 1.44) | 1 (0.82 to 1.22) |
| Mean AHD (26.8 vs 18.9) | 1.02 (0.88 to 1.18) | 0.96 (0.82 to 1.13) |
| TST90† (36.4 vs 0.7) | 1.10 (1.05 to 1.16) | 1.41 (1.10 to 1.81) |
| Mean SpO2 (96.0 vs 93.7) | 0.89 (0.83 to 0.96) | 0.99 (0.89 to 1.10) |
| Min SpO2 (87.0 vs 71.0) | 0.86 (0.75 to 0.97) | 0.99 (0.85 to 1.15) |
| Arousal index (30.9 vs 11.8) | 1.18 (1.06 to 1.32) | 1.03 (0.90 to 1.17) |
| Nocturnal MHR (70.2 vs 58.5) | 1.21 (1.05 to 1.40) | 1.27 (1.10 to 1.46) |
| Wake MHR (75.7 vs 63.1) | 1.24 (1.07 to 1.42) | 1.30 (1.12 to 1.50) |
| Untreated subset (n=1567) | ||
| AHI (39.4 vs 6.3) | 1.31 (1.12 to 1.52) | 0.99 (0.81 to 1.22) |
| REM-AHI (51.0 vs 8.9) | 1.27 (1.04 to 1.56) | 1.06 (0.82 to 1.38) |
| 3% ODI (34.0 vs 3.8) | 1.29 (1.12 to 1.47) | 1.03 (0.86 to 1.24) |
| Mean AHD (26.5 vs 18.7) | 0.97 (0.83 to 1.14) | 0.93 (0.78 to 1.11) |
| TST90† (25.4 vs 0.3) | 1.09 (1.05 to 1.13) | 1.46 (1.14 to 1.86) |
| Mean SpO2 (96.2 vs 9.0) | 0.89 (0.82 to 0.95) | 0.99 (0.89 to 1.11) |
| Min SpO2 (88.0 vs 74.0) | 0.82 (0.73 to 0.92) | 0.96 (0.83 to 1.10) |
| Arousal index (27.4 vs 11.0) | 1.23 (1.11 to 1.37) | 1.08 (0.95 to 1.22) |
| Nocturnal MHR (70.0 vs 58.6) | 1.29 (1.12 to 1.49) | 1.33 (1.15 to 1.54) |
| Wake MHR (75.7 vs 63.0) | 1.33 (1.15 to 1.54) | 1.38 (1.19 to 1.61) |
These PSG parameters were modelled separately. The PSG variables are compared by 75th versus 25th percentiles.
Bold font indicates statistical significance.
*Multivariable models adjusted for age, gender, body mass index, neck circumference, waist circumference, body weight change, Epworth sleepiness scale, total sleep time, cigarette smoking, alcohol use, physical activity, prior hypertension, prior cardiovascular disease, prior hyperlipidaemia and prior diabetes.
†TST90 with non-linearity was transformed using restricted cubic spline transformation with three knots at default 0.1, 0.5 and 0.9 quantiles.
AHD, apnoea–hypopnoea duration; AHI, apnoea–hypopnoea index; BMI, body mass index; HR, hazard ratio; MACEs, major adverse cardiovascular events; MHR, mean heart rate; ODI, oxygen desaturation index; OSA, obstructive sleep apnoea; PSG, polysomnography; REM-AHI, apnoea–hypopnoea index during rapid eye movement sleep; SpO2, peripheral capillary oxygen saturation; TST90, sleep time with oxygen saturation below 90%; TST, total sleep time.