A modified GAP model for East-Asian populations with idiopathic pulmonary fibrosis

Abstract

Background: The easy-to-calculate gender, age, and lung physiology (GAP) model shows good predictive and discriminative performance in the prognosis of idiopathic pulmonary fibrosis (IPF). However, the GAP model was not effective in predicting the prognosis accurately in previous Japanese and Korean IPF cohort studies. Therefore, we developed a modified GAP model for the East-Asian populations by weighing the GAP variables. The validity of the modified GAP model was subsequently evaluated in East-Asian IPF patients.

Methods: The derivation cohort comprised 326 patients with IPF. Weights of the variables were adjusted on the basis of coefficients derived from Cox regression models. The total points were distributed to the three stages of the disease so that the number of patients included in each stage was appropriate. The validity of the modified model was analyzed in another Japanese cohort of 117 patients with IPF and a nationwide cohort of Korean patients with IPF.

Results: Predicted survival rates differed significantly in the derivation cohort using the modified GAP model for each stage of IPF (log-rank test: stage I vs. stage II, p < 0.001; stage II vs. stage III, p < 0.001). Model performance improved according to Harrell's C-index (at three years: 0.696 in the original GAP model to 0.738 in the modified model). The performance of the modified model was validated in the Japanese validation and Korean national cohorts.

Conclusions: Our modification of the original GAP model showed improved performance in East-Asian IPF patient populations.