Screening for impaired pulmonary function using peak expiratory flow: Performance of different interpretation strategies.

BACKGROUND: Spirometry is the gold standard for diagnosis of impaired pulmonary function, but is often unavailable in resource-constrained settings. Some authors have suggested using peak expiratory flow (PEF) to screen for impaired pulmonary function when spirometry is unavailable, but with no consensus on how to define abnormally low PEF. Strategies have included cutoffs based on absolute value of PEF, PEF in percent predicted, PEF Z-score, PEF × height-2, and gender-specific cutoffs of absolute PEF. The objective of this paper is to determine the PEF interpretation strategy with the highest predictive ability for low pulmonary function, with spirometry as the gold standard. METHODS: We analyzed data on individuals aged 40-79 years in the United States National Health and Nutrition Examination Survey 2007-2012. 6,144 individuals fulfilled inclusion criteria for the main analysis. For each PEF interpretation strategy, we calculated the area under the receiver operating curve (AUC) for the detection of low pulmonary function (defined by FEV1 Z-score < -1.645, < -2, < -2.5 or < -3). RESULTS: The AUC was substantially and statistically significantly higher for PEF in percent predicted and PEF Z-score than for absolute value and PEF × height-2, including after stratification by gender. There was no difference in AUC between PEF in percent predicted and PEF Z-score. CONCLUSION: If using PEF to screen adults aged 40 years or older for impaired pulmonary function defined by low FEV1 Z-score, basing cutoffs on PEF in percent predicted or PEF Z-score may result in improved predictive ability. As percent predicted is a mathematically simpler term than Z-score, it may be preferable to use cutoffs based on PEF in percent predicted.