Aim: This study aimed to analyze the Constant Phase Model (CPM) Coefficient of Determination (COD) and an index of harmonic distortion ([Formula: see text]) behavior in intravenous methacholine dose response curve. We studied the COD and [Formula: see text] behavior of Control and Lung Inflammation (OVA) groups of mice and we proposed an alternative for moments when the CPM should not be applied. Methods: 9-week female BALB/c mice were studied, 8 of the control group (23.11 ± 1.27 g) and 11 of the lung inflammation group (OVA) (21.45 ± 2.16 g). The COD values were obtained during the respiratory mechanics assessment via Forced Oscillation Technique (FOT) and the [Formula: see text] was estimated a posteriori. Both control and OVA groups were submitted to 4 doses of Methacholine (MCh) protocol. Results: A strong correlation between COD and [Formula: see text] was present at the last two doses (0.3 mg/kg: r = -0.75, p = 0.0013 and 1 mg/kg: r = -0.91; p < 0.0001) in the OVA group. Differences were found in doses of 0.3 mg/kg between control and OVA for the maximum values of Rn (Newtonian Resistance) and G (tissue viscous); and between groups at PBS and doses of 0.03, 0.1 and 0.3 mg/kg for H (Elastance). A similar behavior was observed for the analysis of Area Under the Curve with the exclusion of the 3 first measurements of each dose. However, in this scenario, the comparison with the maximum value presented a higher discriminatory capacity of the parameters associated with the parenchyma. Conclusions: During severe bronchoconstriction there is a strong negative correlation between model goodness of fit and nonlinearities levels, reinforcing that COD is a robust acceptance criterion, whether still simple and easily obtained from the ventilator. We also pointed out the area under the CPM parameters dose response curve is a useful and can be used as a complementary analysis to peak comparison following bolus injections of methacholine.
Keywords: respiratory impedance; Assessment of respiratory mechanics; constant phase model.