Abdominal Adiposity Correlates with Adenotonsillectomy Outcome in Obese Adolescents with Severe Obstructive Sleep Apnea

Table 3

Multivariate regression analysis. Either regular linear regression or binary logistic regression was performed, depending on the type of outcome variable in the statistical model. For logistic regression, the odds ratios (OR) and their 𝑃 values are reported; while for regular linear regression, the parameter estimates and their 𝑃 values are reported.

	Response to AT (Y/N)	% of AHI change after AT
Waist size predictive model variables adjusted	OR, 𝑃 value	Parameter estimate, 𝑃 value
	(by multivariate logistic regression)	(by multiple linear regression)
(A) Demographic
Waist circumference	1.23, P≤ 0.0 𝟓^∗	−1.92, P ≤ 0.0 𝟏^∗∗
Age	0.78, 𝑃 = 0.54	2.18, 𝑃 = 0.49
Gender
Female^#
Male	1.34, 𝑃 = 0.42	−3.3, 𝑃 = 0.82
Ethnicity
White^#
Other	1.33, 𝑃 = 0.33	−17, 𝑃 = 0.23
(B) Obesity parameters
Waist circumference	1.25, P≤ 0.0 𝟓^∗	−1.35, P≤ 0.0 𝟓^∗
BMI	0.94, 𝑃 = 0.68	−1.32, 𝑃 = 0.34
Neck size	0.89, 𝑃 = 0.60	0.63, 𝑃 = 0.86
(C) Upper airway variables
Waist circumference	1.18, P≤ 0.0 𝟓^∗	−1.65, P ≤ 0.0 𝟏^∗∗
Mallampati score	1.43, 𝑃 = 0.8	−16.8, 𝑃 = 0.12
Tonsilar size	0.55, 𝑃 = 0.64	4.1, 𝑃 = 0.7
(D) OSA variables
Waist circumference	1.33, P≤ 0.0 𝟓^∗	−1.96, P ≤ 0.0 𝟏^∗∗
Obstructive AHI	0.96, 𝑃 = 0.29	0.30, 𝑃 = 0.34
SaO2 nocturnal nadir	1.04, 𝑃 = 0.78	9.2, 𝑃 = 0.18
Snoring severity	0.29, 𝑃 = 0.34	−0.53, 𝑃 = 0.61

Note: #reference level.
(∗∗0𝑥0𝑒𝑓88P ≤ 0.01,∗0𝑥0𝑒𝑓88P ≤ 0.05 ).