Mathematical Equations to Predict Positive Airway Pressures for Obstructive Sleep Apnea: A Systematic Review

Table 2

Study demographics and mathematical equations for studies from non-Asian countries. Definitions: %IBW = percent of ideal body weight, AHI = apnea-hypopnea index, BMI = body mass index (in kg/m²), CPAP = continuous positive airway pressure, cwp = centimeters of water pressure, ESS = Epworth Sleepiness Scale [71], HI = health index, Msat = mean saturation, nadir SaO₂ = lowest oxygen saturation, NC = neck circumference, NR = not reported, Nsat = nadir saturation, ODI = oxygen desaturation index, PAP = positive airway pressure, Peff = effective pressure, Ppred = predicted pressure, pts = patients; RDI = respiratory disturbance index, spO₂ = oxygen saturation, SSS = snoring severity score, USA = United States of America, and , , , and = coefficients based on artificial neural networks.

Study group, year, and country	Number of pts; Development (D); Validation (V)	Mean age	Mean BMI	Mean AHI	Mathematical equation	Accuracy of formula
Basoglu and Tasbakan (2012), Turkey [15]	D = 250; V = 130	52.3 ± 10.3	32.3 ± 5.3	56.7 ± 22.8	0.148 × NC + (0.038 × ODI)	Within ±3 cwp in 96.2% of pts
Tofts et al. (2012), USA [58]	D/V = 310	NR	NR	NR	5.55 + 0.05327 (HI) + 0.03276 (NC) + 0.03422 (AHI Crude) + 0.0005568 (AHI Supine); + 0.001110 (AHI REM) + 0.01301 (RDI)	Predicts 30% of the variability while being within ±2 cwp 74% of the time
Schiza et al. (2011), Greece [44]	D/V = 1111	54.6 ± 10.67	34.35 ± 6.03	41.5 ± 20.16	Men: 5.16 + (0.003 × smoking in pack years); + (0.054 × BMI) + (0.016 × AHI) − 0.403; Women: 5.16 + (0.003 × smoking in pack years); + (0.054 × BMI) + (0.016 × AHI) − 0.806	Within ±2 cwp of the effective pressure in 95% of pts
Anees (2010), USA [63]	V = 27	—	—	—	0.086 × BMI + 0.029 × SSS + 5.989	Within ±2 cwp of CPAP titration in 44%; 2 cwp higher than predicted in 37%
El Solh et al. (2007), USA [59]	D = 311; V = 98	49.6 ± 12.4	35 (34–37)	33 (28–38)		Correlation coefficients between the titration study and predicted pressure was 0.86. The equation underestimated optimal pressures
Loredo et al. (2007), USA [34]	D/V = 76	47.6 ± 9.8	31.3 ± 5.4	55.5 ± 31.3 RDI	30.8 + 0.03 × RDI − 0.05 × NSAT − 0.2 × MSAT	Equation predicted 67% of the variance for Peff
Skomro et al. (2007), Canada [47]	D/V = 183	51 ± 11	37 ± 8	46 ± 33	6.2 × [BMI × 0.11]	Empiric CPAP pressure was suboptimal in 40% pts
Torre-Bouscoulet et al. (2007/2009), Mexico [50, 51]	D/V = 100	49 ± 11	34 ± 4	≥30	Men: (BMI × 0.09) + (ODI × 0.01) − (mean SpO2 × 0.06) + 11.9; Women: (BMI × 0.07) + (ESS × 0.1) + (ODI; × 0.02) + 5.33	Poor agreement between 95% APAP pressures and predictive equations. Equations were not successful
Panagou et al. (2005), Greece [64]	D = 26	51 ± 11.2	—	—	4.95 + (0.18 × AHI) − (0.133 × DI)	No validation testing
Stradling et al. (2004), Britain [48, 49]	D = 101; V = 30	49.0 ± 10.5	36.5 ± 6.5		(0.048 × 4% saO2 dips/h) + (0.128 × NC) + 2.1	Considerable PAP variations from night to night. Similar results for APAP trial and CPAP titration
Rowley et al. (2005), USA [43]	D = 224; V = 192	50.5 ± 9.5	40.6 ± 8.8	32.0 ± 26.4	(0.16 × BMI) + (0.13 × NC) + (0.04 × AHI) − 5.12	Equation improves success rate of in lab titrations but equations were not as helpful for prescribing CPAP
Sériès (2000), Canada [45]	D/V = 40	—	—	46.1 ± 26.3	0.193 × BMI + 0.077 × NC + 0.02 × AHI − 0.611	Home APAP titration successfully predict fixed CPAP in 95% of pts
Nahmias et al. (1995) [65]	D/V = 40	—	—	37.7 ± 3.3	8.7 + 0.028 ×  %IBW + 0.015 × RDI − 0.071; × nadir SaO2	In 73% of patients, the equation predicted CPAP ≥ CPAP titration study
Teschler et al. (1995), Germany [60]	D = 77; V = 180	55 ± 10	30.9 ± 5.8	38 ± 21	1.95 + 0.80 × sex factor (men = 2, women = 1); + 0.09 × BMI + 0.01 × NC + 0.03 × AHI	Mean CPAP pressure was 9.1 ± 2.0 cwp and the predicted was 8.4 ± 3.6 cwp. In 51% of pts the difference was greater than ±1 cwp
Hoheisel and Teschler (1994), Germany [61]	—	—	—	—	(0.07 × NC) + (0.02 × BMI) + (0.03 × AHI) + 3.2	Calculations were made to attempt to improve home prescriptions
Miljeteig and Hoffstein (1993), Canada [38]	D = 208; V = 129	50 ± 11	34 ± 8	50 ± 31	−5.12 + 0.13 × BMI + 0.16 × NC + 0.04 × AHI	Within ±2 cwp in 75% of pts. The equation accounted for 76% of the variability in CPAP