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

Table 3

Study demographics and mathematical equations from Asian countries. Definitions: AHI = apnea-hypopnea index, APAP = automatic positive airway pressure, BMeH = the angle between a line from point B to the menton and from the menton to the hyoid bone, BMI = body mass index (in kg/m²), CPAP = continuous positive airway pressure, cwp = centimeters of water pressure, DI = desaturation index, ESS = Epworth Sleepiness Scale [71], HMD = hyoid-mental distance, LFC = lower face cage, NC = neck circumference, NSBa = cranial base flexure (cephalometrics), PAP = positive airway pressure, Peff = effective pressure, PnCPAP = optimal nasal continuous positive airway pressure, RDI = respiratory disturbance index, saO₂ = saturation of oxygen, spO₂ = oxygen saturation, TG = tongue area, and uFTP = updated Friedman’s tongue position.

Study group, year, and country	Number of pts; Development (D); Validation (V)	Mean age	Mean BMI	Mean AHI	Mathematical equation	Accuracy of formula
Lai et al. (2015), Taiwan [56]	D = 129	46.2 ± 11.0	27.1 ± 3.6	43.3 ± 22.5	1.01 uFTP + 0.74 HMD + 0.059 AHI − 1.603	No prospectively validation testing performed
Ito et al. (2014), Japan [66]	D = 66; V = 66	—	25.1 (21.2, 30.4)	33.9 (19.5, 59.9)	1.000 + 0.043 × AHI + 9.699 × TG/LFC	Equation accounted for 28% of the total variance in PnCPAP
Wu et al. (2014), Taiwan [57]	D = 57; V = 30	53.3 ± 13.1	28.1 ± 3.5	53.6 ± 18.3	6.380 + 0.033 × AHI − 0.068 × SaO2 nadir; + 0.171 × BMI	Within ±1 cwp in 30%, within ±2 cwp in 56.7%, and within ±3 cwp in 86.7% of validation pts
Luo et al. (2013), China [35]	D = 51	48.0 ± 11.3	28.0 ± 4.1	54.3 ± 18.9	0.05 × AHI + 0.15 × BMI + 0.066 × NC − 1.712	, ; calculated pressure; 7.7 ± 1.4 cwp versus 7.3 ± 1.5 cwp. CPAP titration was compared to APAP, and APAP pressure was higher
Lee et al. (2013),; Republic of Korea [31]	Group 1: D = 178; Group 2: V = 178	51.7 ± 10.6	26.3 ± 3.6	40.1 ± 29.0	6.656 × 0.156 × BMI − 0.071 × minimal spO2% + 0.041 × RDI + 0.094 × ESS	Equation accounted for 38.9% of the total variance. Predicted the titration CPAP pressure in 38.8%
Choi et al. (2010), Republic of; Korea [19]	D/V = 202	44.8 ± 8.5	27.6 ± 3.4	36.6 ± 25.1	0.681 + (0.205 × BMI) + (0.040 × AHI)	Equation accounted for 42% variance for optimal CPAP
Akahoshi et al. (2009), Japan [13]	D = 170; V = 110	52.9 ± 12.4	27.8 ± 4.7	50.1 ± 18.8	27.78 + (0.041 × BMeH) + (0.141 × BMI); + (0.040 × AHI) − (0.312 × mean SaO2)	Optimal CPAP (9.5 ± 3.0 and 9.2 ± 2.1 cwp) similar to calculated pressure respectively. Equation accounted for 47% of variance
Chuang et al. (2008),; Taiwan [20]	D = 418; V = 124	49 ± 12	28.4 ± 4.4	58 ± 23	1.98 + 0.184 × BMI + 0.01 × AHI + 0.016 × DI	Successful prediction ±2 cwp of the effective pressure in 84% of study group and 73% in validated group. Equation accounted for 28% of the total variance
Lin et al.; (2003), Taiwan [32]	D/V = 121	49.2 ± 12.6	28.3 ± 4.0	53.8 ± 23.6	0.52 + 0.174 × BMI + 0.042 × AHI	Successful prediction ±2 cwp of the effective pressure in 86% in validation patients
Akashiba et al. (2001), Japan [14]	D/V = 27	51.5 ± 9.6	28.1 ± 2.7	54.7 ± 22.6	42.036 − 0.209 × mean SaO2 − 0.099 × NSBa	Equation accounts for 57.5% of the total variance