Ivacaftor: Five‐year outcomes in the West of Scotland cystic fibrosis population

Ivacaftor has shown to be effective in patients with cystic fibrosis (CF) with a G551D mutation.


| INTRODUCTION
Ivacaftor is licensed for the treatment of cystic fibrosis (CF) caused by specific gene mutations including G551D. 1 Efficacy was reported in two randomised, double blind, placebo controlled trials in adolescent/adult ≥12 years (STRIVE), in paediatric patients ≥6 years (ENVISION), and one open label follow-up study of participants from both trials (PERSIST). [2][3][4] STRIVE and ENVISION showed improvement in lung function, pulmonary exacerbations, and other clinical outcomes at 48 weeks, compared with placebo. 2,3 In PERSIST, improvement was maintained at 144 weeks with no additional safety concerns. 4 We report 5-year real-world outcomes of West of Scotland (WoS) patients on ivacaftor.

| MATERIALS AND METHODS
Patients ≥6 years under the care of the CF adolescent/ adult and paediatric teams in National Health Service Greater Glasgow and Clyde (NHSGGC) and CF paediatric team in NHS Lanarkshire with at least one G551D mutation treated with ivacaftor since 2012/2013 were identified. Data were collected retrospectively in 2018, at four time-points (baseline and at year 1, 2, and 5 posttreatment) using electronic patient records for forced expiratory volume in 1 s (FEV 1 ) and weight using the value closest to the 1 year mark, height, age, gender, smoking status, CFTR mutation class, date of treatment initiation, treatment interruption and reasons, hospital admissions related to pulmonary exacerbations requiring treatment with intravenous (IV) antibiotics, liver function tests (LFTs), and adverse drug reactions (ADRs) reported in clinical letters as suspected to be ivacaftor related. Body mass index (BMI) was calculated. In the paediatric cohort, standard deviation scores (SDS or zscores) were calculated using the LMS method of Cole et al for weight, height and BMI, using UK 1990 growth reference data, based on age and gender. 5,6 The Global Lung Function Initiative (GLI) equation was used to calculate per cent predicted FEV 1 . Data were anonymised and entered into an Excel spreadsheet for analysis. All data were tested for normality using the Anderson-Darling test; parametric tests were subsequently performed. Changes over time were investigated using repeated measures analysis of variance. If differences were found, they were analysed using pairwise comparisons between time-point with a Bonferroni correction.
Analyses were performed using Minitab (version 18) at a 5% significance level.

| RESULTS
Thirty-two patients were identified. One patient was excluded due to missing data at three data collection time-points including baseline. Two patients deceased resulting in a reduced sample size at year 5. Table 1 details WoS cohorts' baseline characteristics.
Statistically significant increases from baseline were observed in mean per cent predicted FEV 1 at year 1 (which was maintained at years 2 and 5) and BMI over 5 years in our adolescent/adult cohort ( Table 2). In the paediatric cohort, the statistically significant improvement in mean per cent predicted FEV 1 observed in year 1 and 2 was not sustained in year 5. A nonstatistically significant increase in paediatric BMI z-scores was observed over 5 years.
The number of patients and episodes of pulmonary exacerbations requiring IV antibiotics and time spent in hospital reduced from baseline over 5 years in both cohorts (Table 3).

| DISCUSSION
Improvement in mean per cent predicted FEV 1 was observed. The improvement of 7.5% from baseline within the adolescent/adult WoS cohort at year 1 and 6.1% at year 2 was statistically significant but lower than the improvement of 10.1% in STRIVE at year 1 and 9.1% in PERSIST at year 2. 2,4 This difference may be due to treatment in the real world and inclusion of four patients with a per cent predicted FEV 1 <40% at baseline, an exclusion criteria in STRIVE. Within our paediatric cohort, the improvement from baseline of 12.8% at year 1 was statistically significant and greater than the improvement of 10.7% in ENVISION. 3 The improvement of 9.2% at year 2 was statistically significant and approximately equivalent to the improvement of 9% in PERSIST. 4 Dryden et al reported a lower mean change in per cent predicted FEV 1 at year 1, in paediatric patients compared with clinical trials. 7 Hurbet et al found a lower increase in per cent predicted FEV 1 from baseline at year 1 and 2, compared with clinical trials in patients ≥6 years with the adolescent group showing the highest response. 8 At year 5, an improvement in per cent predicted FEV 1 , from baseline was seen in both WoS cohorts; this was not statistically significant in the paediatric cohort possibly due to our small sample size. Two adolescent/adult patients deceased prior to year 5, which may have impacted our results at year 5. Mitchell et al showed that after 3 months of treatment in adults FEV 1 declined and was not significantly different from baseline at year 5. 9 Robson et al reported improvement at year 5 in paediatrics compared with year 1. 10 A statistically significant increase in WoS adolescent/ adult BMI was observed. The increase in paediatric BMI z-score was nonstatistically significant, possibly due to the small sample size. At year 2, the mean change in BMI from baseline was greater by 0.8 and 1.5 kg/m 2 within our adolescent/adult and paediatric cohorts, respectively, compared with PERSIST. 4 Increased BMI may be desirable; however, care is needed to avoid associated adverse outcomes.
Over 5 years, there was a decrease in the number of patients and pulmonary exacerbations requiring IV antibiotics, and duration of hospital admissions. Within the adolescent/adult cohort, the largest reduction, from baseline, was at year 1. Mitchell et al showed a significant sustained reduction in IV antibiotic use, over 5 years, in adults. 9 Our paediatric cohort showed a sustained decrease over 5 years. This is important as pulmonary exacerbations can accelerate the rate of lung function decline.
Adherence to ivacaftor, in our cohorts, was encouraged through dialogue and delivery monitoring. There was no suggestion of nonadherence in our paediatric cohort. The literature reports non-adherence rates of 61-80%. 11,12 Limitations include the following: Small sample size, gender bias (5 year paediatric male data was unavailable), and lack of a comparator group.

| CONCLUSIONS
Our findings support ivacaftor's effectiveness. 13 Work is needed to assess quality of life and effectiveness with a comparator group.

AUTHOR CONTRIBUTIONS
Yasmin Al-Din designed the research study, collected the data, analysed the data, and wrote paper. Dr Carol Dryden designed the research study, collected the data, analysed the data, and contributed to the content of the paper. Dr Gordon Macgregor designed the research study and contributed to the content of the paper. Dr David Young designed the statistical analysis to the research study and provided the statistics to the research study. Cristina Coelho designed the research study and contributed to the content of the paper.