Introduction Ivacaftor was the first therapy licensed to address the underlying defect in cystic fibrosis (CF). The improvements in lung function, nutritional status and pulmonary exacerbations in patients carrying a Gly551Asp mutation were greater than previously seen in clinical trials for other therapies. Limited data are available regarding long-term outcomes and adherence to ivacaftor outside clinical trials.
Methods We conducted a 5-year single-centre retrospective study of people with CF carrying the Gly551Asp mutation who received ivacaftor. Clinical outcome data were extracted from medical notes and databases. Drug delivery data were used to assess medicine possession ratio (MPR).
Results 35 people were included. After commencing ivacaftor, FEV1 improved by 9.6% (SE±1.59%) predicted by 6 months. Thereafter, FEV1 declined, and at 5 years had returned to pre-ivacaftor baseline. Ivacaftor did not alter annual rate of FEV1 decline (1.57% pre vs 1.82% post, p=0.74). Body mass index (BMI) increased for 4 years. There was a significant reduction in inpatient and total intravenous antibiotic days sustained over 5 years. MPR remained high but declined over time (−2.5±0.9% per year, p=0.007). FEV1 was better maintained in patients with higher MPRs.
Conclusion The addition of ivacaftor provides acute benefits for people with the Gly551Asp mutation and established lung disease. We report a sustained reduction in intravenous antibiotic use but following acute improvement in lung function, decline continues, and patients will continue to require medical observation and optimisation. Strategies to maintain high adherence should be a priority to prolong the benefits of ivacaftor.
- cystic fibrosis
Data availability statement
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
Statistics from Altmetric.com
What is the key question?
What is the long-term adherence to and effects of ivacaftor on lung function, weight and healthcare utilisation in people with cystic fibrosis and the Gly551Asp mutation?
What is the bottom line?
In this single-centre study, ivacaftor therapy leads to a sustained reduction in hospitalisation and intravenous antibiotic utilisation and an increase in lung function; however, rate of lung function decline is unaltered. Adherence to ivacaftor measured using medicine possession ratio is very high but declines over time.
Why read on?
The recent development of triple cystic fibrosis transmembrane conductance regulator (CFTR) modulation means that up to 90% of people with CF will be eligible for ‘highly effective’ CFTR modulator therapy. To fully understand the impact, this will have both on health and healthcare provision, it is essential that we measure the longitudinal outcomes and adherence to these therapies.
Cystic fibrosis (CF) is an autosomal recessive, life-shortening multisystem disease characterised by lung, pancreatic, and hepatobiliary involvement. The underlying defect is a mutation in the CF transmembrane conductance regulator (CFTR) protein, which regulates the movement of chloride ions across the apical membrane of epithelial cells. In the lungs, this leads to inspissated mucous and resultant airway infection, inflammation, bronchiectasis and ultimately respiratory failure.
Ivacaftor was the first drug licensed of a new class of targeted therapies for CF-termed CFTR modulators. These drugs are named due to their ability to directly interact with the dysfunctional CFTR protein formed due to mutations in the CFTR gene. Ivacaftor was initially evaluated in people with CF carrying one copy of the Gly551Asp mutation. Ivacaftor is specifically termed a CFTR ‘potentiator’ due to its ability to increase the probability that the CFTR channel is open and conducting ions. In patients with the Gly551Asp mutation, CFTR protein is normally situated but the channel exists in a closed state. This leads to little functional activity of the protein and confers a potentially severe clinical phenotype.
The initial phase 3 trial results were greatly heralded as a milestone in CF care, demonstrating dramatic acute improvements in lung function (a 10% increase in FEV1), a reduction in pulmonary exacerbations and improved quality of life in patients carrying a Gly551Asp mutation.1–3 Ivacaftor was approved in the USA in 2011 and in Europe in 2012 for people with CF carrying this mutation.
There are an estimated 10 655 people diagnosed with CF in the UK, 60% of whom are over the age of 16.4 Manchester Adult Cystic Fibrosis Centre is the second largest adult CF centre in the UK, providing care for 464 people across the North West of England. Ivacaftor was initially prescribed at our centre on a compassionate basis in 2012 for patients with severe lung disease and since 2013 has been routinely prescribed to patients with the Gly551Asp or other qualifying gating CFTR mutation.
The benefits of ivacaftor have been shown to be maintained for at least 4 years, from extension studies of the original clinical trials5 and real-life clinical data from CF registries.6 7 Adherence to ivacaftor in clinical trials was reported to very high at between 89% and 91%;2 however, electronic monitoring data suggest adherence may be much lower in clinical practice.8
The aim of this study was to investigate longer term clinical outcomes in patients with a Gly551Asp mutation treated with ivacaftor at our centre as well as looking at adherence to treatment.6 As effective CFTR modulator therapies become available for a greater proportion of people with CF, it is essential for clinical teams not only to understand the longitudinal effects of these therapies on healthcare systems including rates of hospital admissions but also to identify issues that may need to be specifically monitored including weight gain or suboptimal adherence.
We conducted a retrospective analysis on patients receiving ivacaftor therapy for the Gly551Asp mutation over a 5-year period in a single large adult centre to better understand these issues.
All patients with CF carrying at least one Gly551Asp mutation who had taken ivacaftor at any time since its introduction were identified from the Manchester Adult Cystic Fibrosis Centre’s patient database, including deceased patients and patients who had started ivacaftor in paediatric services before transitioning to the adult centre.
Data were collected from patients’ clinical records from 2 years prior to starting ivacaftor and for up to 5 years following initiation. Pre-ivacaftor measures were limited to 2 years due to difficulties obtaining data prior to this from patients who had transitioned from paediatric care at other centres across the North West of England.
The following demographics were recorded at baseline: age at starting ivacaftor, FEV1, BMI, CFTR genotype, sex, infecting pathogen, medications, sweat chloride and comorbidities.
FEV1 values were collected from all stable clinic appointments as well as the highest FEV1 reading during exacerbations (requiring oral or IV antibiotics). Best quarterly values for FEV1 and weight were used in the analysis to minimise the effects of short-term illnesses. All documented weight and height measurements from clinic appointments and hospital admissions were recorded.
Days of intravenous antibiotics, oral antibiotics and days in hospital per annum were recorded as a measure of healthcare utilisation for the year prior to starting ivacaftor and for 5 years of therapy.
Adherence (medicine possession ratio)
Patients received ivacaftor through a homecare delivery company which contact patients monthly to monitor stock levels and fulfil orders when further supply is needed. If an individual is not ready for delivery, the company will ask how much ivacaftor they have left and arrange a second call to arrange delivery to ensure continuous provision. A 28-day supply of ivacaftor is provided on each delivery. Using the time interval between deliveries, we calculated the medicine possession ratio (MPR):
Quarterly, yearly and overall MPRs were calculated for each patient. Periods of ivacaftor dose alterations, for example, dose reductions due to drug interactions or temporary cessations due to pregnancy were noted and MPRs were adjusted according to the new prescription.
FEV1 and BMI data are presented as change from baseline. The highest value recorded in the 3 months prior to starting ivacaftor was used as baseline. Generalised estimating equations with estimated marginal means were used to model trends over time. Data were adjusted for age, baseline FEV1 and gender. Statistical analysis was performed in R9 using the ‘geepack’10 and ‘emmeans’11 packages. Where data from multiple time points were analysed, the Tukey method was used for p value adjustment. A paired t-test was used for comparison of sweat chloride concentration before and after starting ivacaftor. FEV1 rate of decline was calculated using the slope of the longitudinal model.
Log transformation was used for analysis of positively skewed healthcare utilisation data. MPR data which had a strong negative skew were not easily corrected through transformation, therefore, raw data were used in generalised estimating equations. For data visualisation purposes, healthcare utilisation and MPR data are displayed as mean and SE; however, median and interquartile data are also included in the text due to the skew of the data.
Subjects were included in analysis until their last recorded visit. Individuals without FEV1 or BMI data available in the 3 months prior to starting ivacaftor were not included in individual FEV1/BMI change from baseline analyses. Quarterly data were not complete for some patients as patients did not always have a clinical encounter every 3 months.
For dynamic analysis of MPR versus FEV1 and BMI data in the months and years after starting ivacaftor, three additional patients were included who did not have weight and lung function data available from the 3 months before starting ivacaftor. In these patients, baseline was assigned as the most recent clinical data prior to starting ivacaftor.
Regression analysis using generalised estimating equations was used to look at the association between MPR and clinical outcomes.
There were 35 people with CF with at least one Gly551Asp mutation who had taken ivacaftor. Five patients died before 5 years of therapy were completed. Four patients had been taking ivacaftor for less than 5 years at the time of data collection, and data collection was stopped for two patients after they enrolled in clinical trials (see flowchart, figure 1). Three patients were under the age of 18 years at the time of commencing ivacaftor (14–17 years). Baseline demographics of the study population are shown in table 1.
Sweat chloride data from pre-ivacaftor and 3 months post-ivacaftor therapy were available in 25 patients. Mean sweat chloride prior to therapy was 110.7 mmol/L, SD ±19.2, falling to 55.1±22.4 mmol/L post-ivacaftor (p<0.001).
Thirty-two patients had baseline FEV1 data available from a clinic appointment in the 3 months prior to starting ivacaftor and were included in FEV1 analysis. Four patients had FEV1 data available for every quarter post-ivacaftor. Twenty-seven patients had at least two measures recorded each year after starting ivacaftor and before censoring. Mean baseline FEV1 was 58.4% predicted, SE ±4.01%. Mean absolute change from baseline was 9%±1.32% (SE) predicted at 3 months (p<0.001) and 9.62%±1.59% predicted at 6 months (p<0.001, figure 2A). After 6 months, FEV1 declined, and at 60 months, absolute change from baseline was 1%±2.14% predicted (p=1). Rate of decline in FEV1 was 1.82±0.45 (SE) predicted per annum over the 5 years postivacaftor. Preivacaftor, annual rate of decline in the 2 years prior to therapy was 1.57%±1.31% predicted. On a group level, there was no significant difference in rate of decline pre and post-ivacaftor (p=0.74).
Absolute change in FEV1 in patients with mild (FEV1 >70% predicted), moderate (FEV1 41%–70% predicted) and severe (FEV1 ≤40% predicted) lung disease is illustrated in figure 2B. Absolute change in FEV1 at 3 months appears lower in the severe group compared with those with milder lung disease; however, the difference in rate of decline is not statistically significant.
Body mass index
Thirty-three patients who had baseline weight and height data available from a clinic visit in the 3 months prior to starting ivacaftor were included in analysis. Six patients had BMI data available for every quarter post-ivacaftor. Twenty-four patients had at least two measures recorded each year after starting therapy and before censoring. Average BMI increased at a rate of 0.2 kg/m2±0.08 (SE) per year (p=0.019, figure 3A). BMI did not peak until 51 months (change from baseline=2.11±0.55 kg/m2 (SE) after which there was a decline to +1.33±0.43 kg/m2 at 60 months. BMI increased after starting ivacaftor in patients with mild to moderate lung disease (FEV1 41%–70% and >70% predicted, figure 3B). In patients with severe lung disease (FEV1 <40%), BMI was maintained but did not increase.
Days of home intravenous antibiotics, oral antibiotics, inpatient days and total intravenous antibiotics for the year pre-ivacaftor and 5 years post-ivacaftor are shown in figure 4. Baseline data were available for 34 patients who were subsequently included in analysis. In the year prior to starting ivacaftor, patients spent a mean of 27.3±SE 6.1 (median 14, IQR 0–40) days on intravenous antibiotics (home and inpatient IV antibiotics). This fell to 11.9±SE 4.8 (median 0, IQR 0–12.5) days in the first year of ivacaftor and remained low through to year 5, where mean days were 12.4±SE 5.6 (median 0, IQR 0–9.8). Patients spent a mean of 23±SE 6.8 (median 6, IQR 0–35.3) days as inpatients pre-ivacaftor, 9.2±SE 4.2 (median 0, IQR 0–7.3) days during the first year of ivacaftor and 4.6±SE 1.7 (median 0, IQR 0–3.5) days in the fifth year of treatment. Inpatient days and intravenous antibiotic days were significantly lower than baseline for years 1, 2, 3 and 5 (figure 4). The same pattern was not seen for oral antibiotics: patients took oral antibiotics for a mean of 18.6±3.7 (median 14, IQR 0–28) days in the year pre-ivacaftor, 19.8±SE 3.6 (median 14, IQR 0–28) days in year 1 and 20±SE 4.5 (median 14.5, IQR 0–33.3) days in year 5. Estimated marginal means for days in hospital were lower in patients with mild lung disease (FEV1 >70%) compared with those with moderate disease (FEV1 41%–70%, p=0.002).
Medicine possession ratio
MPR data were available for 33 patients for up to 5 years. Three patients with severe lung disease started taking ivacaftor on a compassionate basis before the drug became widely available in 2013. Data were available for one of these patients only due to a differing supply chain. Data at 60 months were available for 26 patients.
Mean MPR declined from 99.6%±0.3% (SE) at 3 months to 87.5%±4.5% at 60 months (figure 5A). Median MPR was 100% (IQR 100%–100%) at 3 months of therapy and 100% (78%–100%) at 60 months. Mean MPR stayed above 80% throughout the study and median MPR above 90%. Rate of decline was 2.5%±0.9% (SE) per year (p=0.007). Rate of decline in MPR was 4%±1.9% per year in patients ≤25 years and 1.5%±0.6% per year in patients >25 years (figure 5B); however, the difference in rate of decline did not reach statistical significance (p=0.17).
Effect of MPR on clinical outcomes
Patients with a higher overall MPR (MPR calculated across the whole 5 years) had a greater FEV1 change from baseline at 60 months (p=0.036, table 2). Higher annual MPR was associated with a reduced rate of decline in FEV1 from the previous year (p=0.006). There was a statistically significant negative association between overall MPR and BMI; however, there was no relationship between overall MPR and BMI change from baseline at 60 months or annual MPR with BMI change from the previous year.
This study describes the MPR and clinical response to ivacaftor over 5 years in people with CF and a Gly551Asp mutation. We report a prolonged improvement in FEV1 following initiation of ivacaftor but a continuing decline in lung function over time. There was progressive weight gain over 4 years of therapy. Rescue antibiotic treatment pattern changed for intravenous but not oral antibiotics. This is the first study to report long-term adherence to ivacaftor. MPR was very high but began to decline over time. There was an association between MPR and clinical outcomes with the highest adherers having better-maintained lung function. These findings, in keeping with larger annual registry studies, have the potential to inform future healthcare provision for CF services.
The initial lung function improvement on commencing ivacaftor resembles the original clinical trial findings (9.6% improvement from baseline in this study at 6 months vs 10.4% in the clinical trial setting2). In the longer term, we saw a steeper decline in FEV1 than reported in the ivacaftor clinical trial extension studies.5 12 Sawicki and colleagues compared clinical trial patients from the PERSIST5 extension study to matched Phe508del patients and found a reduced rate of decline in FEV1 in the patients taking ivacaftor.12 Over 3 years, annual rate of decline was 0.91% predicted. We saw a steeper rate of decline of 1.82% per annum in our patients, with a return to baseline lung function by 5 years. A steeper rate of decline was also seen in an analysis of Irish CF registry data by Kirwan and colleagues who reported an FEV1 decline of 1.74% per annum over 3 years in 35 adults. This Irish study population closely resembles that of Manchester with similar average FEV1, BMI and genotype distribution.6 A recent large UK and US registry study compared outcomes in children and adults treated with ivacaftor with a non-ivacaftor comparator group.7 Two hundred and forty-seven UK patients were followed for 4 years after which mean FEV1 was 4.9% higher than baseline. In 635 US patients who were followed for an additional year, mean FEV1 had fallen to 0.7% below baseline at 5 years. This decline was reported to be much smaller than in the comparator group (8.3%); however, this did not factor in the relative contribution of the acute rise in lung function witnessed. This study did not separate adults and children, hence age was lower than in our study and patients had a higher baseline FEV1. The CFTR genotypes in both US and UK populations were present in similar proportions to our study. Kirwan et al looked at patients with paediatrics taking ivacaftor in a separate analysis. In contrast to adults, lung function stabilised or even improved.6 The ongoing decline in FEV1 in adults despite ivacaftor may reflect the irreversible lung damage that has already occurred by the time they reach adulthood. Whether patients who start taking ivacaftor early in life will avoid this decline will become apparent as these individuals enter adulthood.
Intravenous antibiotic utilisation and time in hospital remained lower than baseline for 5 years, consistent with US and UK registry data;7 however, there was no reduction in oral antibiotic prescriptions. The reasons for this are unclear, it is plausible that instead of admitting patients to hospital for intravenous antibiotics, there may be an increased utilisation of oral antibiotics to treat pulmonary exacerbations. It is unclear whether this reflects a true change in exacerbations, patient choice, or whether clinician decisions are influenced knowing the patient is taking ivacaftor. There is evidence to suggest that exacerbations and rate of return to baseline lung function are similar in both ivacaftor-treated patients and other patients with CF.13 Further evidence is needed to address whether strategies for treating pulmonary exacerbations in people receiving modulator therapy may need to be altered.14
Another effect of ivacaftor is promoting weight gain. The patients in this study gained weight but unlike FEV1, weight gain was gradual and continued up to 4 years. Weight gain is not always desirable in patients with a normal baseline BMI, and patients may need advice to avoid excessive weight gain. Weight did not increase in the subgroup of patients with severe lung disease, but importantly their weight was maintained over the 5 years.
Can the benefits of ivacaftor be extended by improving adherence? As new life-prolonging treatments are discovered, adherence becomes an important consideration as patients are faced with increasingly complex treatment regimens. Adherence to CF treatments including inhaled therapies15–19 and physiotherapy20 has been shown to be variable and often suboptimal. Higher adherence to CF therapies is associated with improved outcomes including reduced courses of intravenous antibiotics,19 admissions to hospital17 and length of inpatient stay.16 We expected that a highly anticipated and expensive drug such as ivacaftor should be adhered to, especially as it is in tablet form and easy to take. Mean MPR for ivacaftor remained at least 80% throughout the study, and median adherence stayed above 90% suggesting excellent adherence. By contrast, median MPRs for inhaled therapies have been reported between 49% and 70%16 19 and 76% for oral azithromycin.19 The high MPR to ivacaftor may result from the relative ease of administration of an oral agent and the effectiveness of the drug, with patients feeling much better in the months following initiation. Frequent contact from the delivery company may act as a prompt for patients to adhere to treatment and makes it easy for patients to receive a regular supply of ivacaftor. In our cohort, a small number of poor adherers were identified from the MPR data. Although not meeting statistical significance, there was a drop in MPR in younger patients after 2 years. Our results suggest an association between MPR and clinical outcomes, with better-maintained lung function in patients with the highest MPRs. Lower adherence may contribute to the greater rate of FEV1 decline seen in young adults21 and could be a target for future study. Many CF centres use a home delivery service to supply expensive drugs such as ivacaftor; therefore, MPR data are easily obtainable and could become a useful tool to identify low adherers and offer strategies to improve adherence.
The MPR results for ivacaftor are encouraging when considering adherence to other CFTR modulators. This includes the recently developed triple combination CFTR modulator elexacaftor/tezacaftor/ivacaftor, which has the potential to broaden eligibility for highly effective modulator therapy to approximately 90% of people with CF. We acknowledge that using MPR as a measure of adherence is limited and may overestimate adherence; however, our data suggest an association between MPR and FEV1, supporting its use as a surrogate measure of adherence in this case. MPR is the most readily available form of adherence data available to clinicians and can be used to look at adherence trends over months and years.
This study is limited by its size and observational design. Due to the observational nature of the study, there are missing data where patients moved from other centres or did not have regular clinic visits. This meant that not all outcomes were available for all 35 patients, which may lead to issues around generalisability and selection bias. This study is also limited by the age of the population and may not inform what will occur with early initiation of these therapies in childhood. However, the majority of patients with CF are currently adults, and this information is important for modelling future care.
In this study of 35 people with CF with the Gly551Asp mutation, ivacaftor therapy resulted in improved lung function, and a reduction in hospitalisations and intravenous antibiotic utilisation. After an acute improvement, lung function continued to decline over time at similar rates to preivacaftor levels. Adherence to ivacaftor measured by MPR was very high; however, fell over time particularly in younger adults. Our data suggest that higher adherence is associated with better-preserved lung function; therefore, the development and utilisation of strategies to improve adherence should be a priority. With triple CFTR modulator therapy likely to make treatment available to the majority of patients with CF, further work to study the long-term effects of CFTR modulators is essential.
Data availability statement
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.
The local ethics committee approved the study as an audit of practice.
The authors would like to thank Michael Dooney for his help in acquiring the MPR data.
Contributors RMM—study design, data collection, statistical analysis and writing of the manuscript. AJ—study design, manuscript writing. PF—statistical advice. KS—statistical advice. PB—study design, manuscript writing.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests PJB has received consultancy fees, lecture fees from Vertex pharmaceuticals; their institution has received a grant to organise an educational event and PJB is the local principal investigator on trials sponsored by Vertex pharmaceuticals. RMM, AMJ, KS and PF have no conflicts of interest to declare.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.