Baseline and post-bronchodilator interrupter resistance and spirometry in asthmatic children

In children unable to perform reliable spirometry, the interrupter resistance (R(int) ) technique for assessing respiratory resistance is easy to perform. However, few data are available on the possibility to use R(int) as a surrogate for spirometry. We aimed at comparing R(int) and spirometry at baseline and after bronchodilator administration in a large population of asthmatic children. We collected retrospectively R(int) and spirometry results measured in 695 children [median age 7.8 (range 4.8-13.9) years] referred to our lab for routine assessment of asthma disease. Correlations between R(int) and spirometry were studied using data expressed as z-scores. Receiver operator characteristic curves for the baseline R(int) value (z-score) and the bronchodilator effect (percentage predicted value and z-score) were generated to assess diagnostic performance. At baseline, the relationship between raw values of R(int) and FEV(1) was not linear. Despite a highly significant inverse correlation between R(int) and all of the spirometry indices (FEV(1) , FVC, FEV(1) /FVC, FEF(25-75%) ; P < 0.0001), R(int) could detect baseline obstruction (FEV(1) z-score ≤ -2) with only 42% sensitivity and 95% specificity. Post-bronchodilator changes in R(int) and FEV(1) were inversely correlated (rhô = -0.50, P < 0.0001), and R(int) (≥35% predicted value decrease) detected FEV(1) reversibility (>12% baseline increase) with 70% sensitivity and 69% specificity (AUC = 0.79). R(int) measurements fitted a one-compartment model that explained the relationship between flows and airway resistance. We found that R(int) had poor sensitivity to detect baseline obstruction, but fairly good sensitivity and specificity to detect reversibility. However, in order to implement asthma guidelines for children unable to produce reliable spirometry, bronchodilator response measured by R(int) should be systematically studied and further assessed in conjunction with clinical outcomes.