Treated (n=59) | Untreated (n=45) | p | |

Age at time of symptoms (years)* | 4.7 (4.3–5.6) | 4.8 (4.1–5.5) | 0.99 |

Route of treatment (n (%)) | |||

Oral | 48 (81.4) | – | |

Inhaled | 3 (5.1) | – | |

Intravenous inpatient | 8 (13.6) | – | |

Missing visits† (n (%)) | |||

Missing baseline only | 15 (25.4) | 13 (28.9) | 0.24 |

Missing follow-up only | 9 (15.3) | 12 (26.7) | |

Missing neither baseline nor follow-up | 35 (59.3) | 20 (44.4) | |

Time intervals (days)* | |||

Baseline to symptomatic visit | 109 (84–162) | 112 (84–184) | 0.94 |

Symptomatic visit to follow-up | 72 (35–98) | 91 (43–98) | 0.33 |

Pulmonary function at symptomatic visit‡; mean (95% CI) | |||

LCI (units) (treated n=59, untreated n=45) | 10.4 (9.8 to 11.1) | 9.8 (9.1 to 10.4) | 0.001 |

zFEV (z-scores) (treated n=44, untreated n=33) | −1.4 (−1.8 to –1.0) | −0.6 (−1.0 to–0.2) | <0.001 |

ppFEV (%) (treated n=44, untreated n=33) | 79.8 (74.3 to 85.4) | 91.8 (86.0 to 97.5) | <0.001 |

Change in pulmonary function from baseline to symptomatic visit‡; mean difference (95% CI) | |||

Relative LCI change (%) (treated n=44, untreated n=32) | 23.8 (16.2 to 31.4) | 11.2 (2.4 to 19.9) | 0.02 |

Absolute LCI change (units) (treated n=44, untreated n=32) | 1.8 (1.2 to 2.3) | 0.9 (0.2 to 1.5) | 0.03 |

Relative ppFEV change (%) (treated n=34, untreated n=19)¶ | −9.3 (−14.7 to −3.9) | 3.4 (−3.5 to 10.3) | 0.002 |

Absolute zFEV change (z-scores) (treated n=34, untreated n=19)¶ | −0.7 (−1.0 to –0.3) | 0.2 (−0.3 to 0.6) | 0.003 |

Average treatment effect; mean difference (95% CI)§ | |||

Relative change in LCI (%) | −15.5 (−25.4 to −5.6) | 0.002 | |

Absolute change in zFEV (z-scores) | 0.58 (−0.30 to 1.47) | 0.20 | |

Relative change in ppFEV (%) | 1.8 (−17.2 to 20.8) | 0.19 |

*Data presented as median (IQR). P value shown for Wilcoxon rank-sum test for comparison of medians.

†P value for χ

^{2}test of association is shown.‡Data presented as estimated mean values (95% CI) as determined by linear regression using GEE models to account for multiple measurements in the same participant.

¶Mixed effects linear regression model results shown as convergence not achieved for GEE model.

§Average treatment effect estimated using propensity score matching, accounting for baseline parameter value and magnitude of worsening parameter from baseline at the time of symptoms. Results shown as mean relative change in LCI and mean absolute change in zFEV.

GEE, generalised estimating equations; LCI, lung clearance index; ppFEV, per cent-predicted values for FEV

_{1}; zFEV_{1}, z-score for FEV_{1}.