BACKGROUND:Variability of the airflow resistance (AR) between different dry powder inhalers (DPIs) can affect patients' ability to achieve adequate peak inspiratory flow (PIF).

AIM:We investigated the susceptibility of patient PIF variability to DPI-AR to understand PIF variability in guiding DPI prescribing in chronic obstructive pulmonary disease (COPD) patients.

METHODS:We undertook a meta-analysis of original studies measuring PIF against different DPI-AR in COPD patients. DPI-AR was categorised into 5 groups, from low AR (R1) to high AR (R5). Logistic curves were fitted to the PIF data for each DPI-AR group. We report the mean PIF and the slope, k, of the logistic fit that indicates the width (i.e., variance) of the PIF distribution.

RESULTS:In 27 studies (n=2525 COPD patients), both the mean PIFs and the variances of the PIF distributions (k value, p<0.0001) were significantly different (p<0.0001) between all 5 DPI-AR groups. As AR increased, the width of the PIF distributions became narrower, with greatest variance reported for R1 [k=0.0672, mean PIF=92.3 L/min, CI 90.6-94.2 L/min] versus R5 [k=0.0811, mean PIF=46.3 L/min, CI 45.2-47.4 L/min]. Importantly, we identified DPI-AR could be simply classified into 3 clusters-R1, R2&R3, R4&R5-, based on the patient?s PIF variance.

CONCLUSION:Our work identifies PIF variance as a potential tool to inform DPI selection (using just 3 AR groups as compared to 5 AR groups), whilst maintaining COPD patients? ability to achieve adequate inhalation flows. Low AR DPIs remain challenging due to their greater variability in PIFs. Funding for this work was made possible, in part, by the U.S. FDA through contract HHSF223201710072C.