Table 3

Logistic regression model predicting progression to eESLD in 18–22 year-old patients with CF

VariableOR (95% CI)P values
FEV1% predicted (baseline)0.92 (0.8911 to 0.9601)<0.001
Shannon index <10.23 (0.0593 to 0.9270)0.039
  • In the univariate analysis, covariates such as Shannon’s and Simpson’s diversity indices, FVC, body mass index (BMI), relative abundance (RA) of Pseudomonas at 50% and 75%, Simpson index >0.5 and RA of Streptococcus at 25% were significantly associated with the primary outcome and then were considered for the multivariable logistic regression model. However, in the final model, those were not significant (p>0.05) and thus removed. The final model (n=104) had good calibration (goodness of fit p=0.99) and discrimination (area under the receiver operating characteristic (ROC) curve=0.92).

  • CF, cystic fibrosis; eESLD, early end-stage lung disease.