Applicable predictive factors extracted from peak flow trajectory for the prediction of asthma exacerbation.

BACKGROUND: Real-time asthma exacerbation prediction and acute asthma attack detection are essential for patients with severe asthma. Peak expiratory flow (PEF) exhibits a potential for use in long-term asthma self-monitoring. However, the method for processing PEF calculations remains to be clarified.

OBJECTIVE: Present research was conducted to develop clinically applicable novel exacerbation predictors calculated using PEF records.

METHODS: Previously proposed exacerbation predictors, including the slope of PEF, percentage predicted PEF, percentage best PEF, the highest PEF over the lowest PEF within specific periods, and PEF coefficient of variation, as well as a novel indicator delta PEF moving average (ΔMA), defined as the difference between 14-day and 3-day average PEF values along with MA adjusted for PEF reference (%ΔMA), were verified using the Hokkaido-based Investigative Cohort Analysis for Refractory Asthma data of 127 patients with severe asthma from whom 73,503 PEF observations were obtained. Receiver operating characteristic curves for all predictors were drawn and the corresponding areas under the curve (AUCs) were computed. Regression analysis for MA and % MA were conducted.

RESULTS: The most outstanding performance was demonstrated by ΔMA and %ΔMA, with AUC values of 0.659 and 0.665 in the univariate model, respectively. When multivariate models incorporated with random intercepts for individual participants, the AUC for ΔMA and %ΔMA soared to 0.907 and 0.919, respectively.

CONCLUSION: The MA and % MA are valuable indicators that should be considered when deriving predictors from the PEF trajectory for monitoring exacerbations in patients with severe asthma.