(156) Multicenter Validation of the CF-ABLE Score as a Predictor of Outcome and Therapeutic Response in Cystic Fibrosis

Background: The CF-ABLE score is an easy-to-use prognostic scoring system, validated in a large cohort from the national CF Registry of Ireland. This weighted score ranges from 0-7 and predicts risk of poor outcome, defined as death or requirement for lung transplantation, over time using commonly available clinical parameters such as age, body mass index (BMI), forced expiratory volume in 1s (FEV₁) and exacerbation frequency. In the original study describing the score, people with CF (PWCF) with a score of ≥5 were deemed to have a 26% risk of poor outcome within 4 years. The present study aimed to evaluate real-world clinical performance of the CF-ABLE score as a clinical predictor in CF. Since clinical endpoints such as mortality and transplantation are typically difficult to satisfy in time-limited clinical trials, we also investigated the effect of double- and triple-combination CFTR modulator therapy on CF-ABLE score.

Methods: PWCF (n=611) were recruited from three large specialist CF centers in the United States and Ireland. Participants were enrolled between 2013 and 2015. PWCF were followed clinically for 4 years from the point of enrollment, unless they transitioned to a CF-ABLE score of ≥5 within 2 years – in which case the 4-year follow-up period commenced at the time the score of ≥5 was achieved. In a parallel cross-sectional analysis of 60 PWCF chosen at random from the total cohort, the relationship between sputum inflammatory mediators associated with disease progression in CF and CF-ABLE score was also evaluated. Changes in CF-ABLE scores over time in PWCF who commenced ivacaftor, double-combination therapy (DCT) with lumacaftor/ivacaftor or tezacaftor/ivacaftor, and/or triple-combination therapy (TCT) with elexacaftor/tezacaftor/ivacaftor were also recorded prospectively.

Results: In clinical practice, the CF-ABLE score performed substantially better than it did in its original derivation study, and outperformed FEV₁ alone as a predictor of outcome. A CF-ABLE cutoff score of ≥5 clearly distinguished PWCF who went on to poor outcome from those who did not (P < 0.0001). Approximately half of PWCF with a score of ≥5 died or underwent lung transplantation within 4 years. In contrast, poor outcome at 4 years was observed in less than 5% of PWCF with a score < 5. The score correlated with sputum neutrophil elastase activity, matrix metalloprotease activity, and levels of interleukin (IL)-1β, IL-6, IL-8 and inversely with IL-10 (all P < 0.05). For each mediator, the correlation with CF-ABLE score was stronger than with FEV₁ alone (all P < 0.05). In F508del/F508del PWCF, DCT did not improve FEV₁ at 24 months compared to matched controls not receiving DCT but significantly improved CF-ABLE score (P = 0.002), largely due to an effect on exacerbations. In contrast, TCT resulted in marked decreases in CF-ABLE score within 3 months (P < 0.0001), enabling PWCF with baseline scores ≥5 to return to the low-risk range.

Conclusions: The CF-ABLE score robustly identifies PWCF at increased risk of poor outcome and correlates with airway inflammation better than FEV₁ alone. The score identifies a probable effect on mortality in response to CFTR modulators early in the treatment period – an endpoint that has proven elusive in prospective clinical trials and served as an obstacle in drug reimbursement negotiations with healthcare policymakers.

Acknowledgements: This work was supported by the Elaine Galwey Memorial Research Bursary.

References: