Inhaled Therapy Adherence and Outcomes to Kaftrio in Cystic Fibrosis

Aim of the study:

To understand the impact of adherence to preventative inhaled therapies on outcomes in people with CF (pwCF) treated with Ivacaftor/Tezacaftor/Elexacaftor.

Background:

Cystic Fibrosis (CF) is a multi-system, life-long genetic condition affecting around 10,500 people in the UK with a median predicted survival of 49 years. The gene mutations responsible for CF result in dysfunctional cystic fibrosis transmembrane conductance regulator (CFTR), which acts as a chloride channel. Impaired chloride secretion in CF leads to sticky and thick secretions for a range of epithelial tissues such as the airways and pancreatic duct. This results in malabsorption, suboptimal nutritional status, recurrent respiratory infections, progressive lung damage and respiratory failure.

Standard treatment for CF, until recently, has focused on managing the symptoms and consequences of CF. Inhaled therapies, specifically inhaled antibiotics, mucolytics and osmotics, are particularly important to reduce exacerbations and maintain lung health in people with CF.

Precision medicines, in the form of highly effective CFTR modulators, first became available around a decade ago for those with CF caused by a gating mutation, comprising around 5% of the CF population. Further developments now see highly effective modulators suitable for around 90% of people with CF (pwCF). This is changing the landscape for many pwCF with rapid and transformative improvements in health. This rapid change creates some uncertainty about the ongoing need for established treatment regimens.

The landmark randomised controlled trial evaluating the efficacy of Ivacaftor among pwCF and the G551D gene mutation found a 10.6% between-group difference in forced expiratory volume in one second (FEV1) over twelve months in favour of Ivacaftor. The between-group FEV1 difference was 9.9% even among adults ≥18 years in the RCT (Randomised Controlled Trial). Real world data from the UK CF registry found a between-group FEV1 difference of only 6% over twelve months. This efficacy-effectiveness gap is unlikely to be fully explained by differences in people's characteristics compared to the RCT because Ivacaftor was licensed in the UK only for people with at least one G551D gene mutation during the period of evaluation (from 2012 to 2013).

Preventative inhaled therapies were continued during the randomised controlled trial whereas there is real world evidence of declining inhaled therapies prescription and use following Ivacaftor initiation. This provides a potential explanation as to why the real-world effectiveness of Ivacaftor was only around 60% of the efficacy observed in the randomised controlled trial where participants are likely to have maintained high level of adherence to inhaled therapies.

Ivacaftor/Tezacaftor/Elexacaftor is a combination modulator commissioned in the UK in the summer of 2020 and is suitable for around 90% of the CF population. The first key randomised controlled trial involved 403 people with a F508del mutation and a minimal function mutation. After 24 weeks of treatment, people who took Ivacaftor/Tezacaftor/Elexacaftor had an average increase in percentage point of forced expiratory volume in the first second (ppFEV1) of 13.9 percentage points compared with a reduction of 0.4 percentage points in people who took placebo. In another study involving 107 people with two F508del mutations, people who took Ivacaftor/Tezacaftor/Elexacaftor had an average increase in ppFEV1 of 10.4 percentage points compared with an increase of 0.4 percentage points in people who took a combination of Ivacaftor and Tezacaftor alone.

Data is not yet available to identify whether the efficacy effectiveness gap seen for Ivacaftor also exists for Ivacaftor/Tezacaftor/Elexacaftor. There is a unique opportunity within the CFHealthHub learning health system to answer this important research question owing to the unique dataset of objective inhaled therapy adherence data available only within CFHealthHub.

Plan of the investigation This will be an observational study designed to emulate a target trial; the target trial being a comparison of Ivacaftor/Tezacaftor/Elexacaftor with usual care.

Methodology Adults with cystic fibrosis, who attend one of the CF centres within, or previously participating in, the CFHealthHub learning health system (REC ref: 17/LO/0032, IRAS 216782) and who are, or have ever been, prescribed Ivacaftor/Tezacaftor/Elexacaftor will be identified by their local centre. Data will also be analysed for people with CF from these centres who do not participate in CFHealthHub but who consent to share their data with the CF Registry (REC ref: 07/Q0104/01, IRAS 35220).

The CFHealthHub digital Learning Health System (ISRCTN14464661) consists of seventeen adult CF centres in the UK (with around 1,500 adults involved) where the metric for improvement is medication adherence. Electronic data capture of preventative inhaled therapy is undertaken using a bluetooth enabled nebuliser device (eTrack® or Ineb®) which connects to a platform accessible by people with CF and by their care team to view the inhaled preventative therapy adherence data. CFHealthHub is conducted as a trials within cohorts study and, therefore, at the point of entry, people have the option to consent to pseudo-anonymised data within CFHealthHub to be used for future research and also to be included for selection in future research studies which have been ethically approved.

The UK Cystic Fibrosis Registry (Doi: 10.1093/ije/dyx196) is a UK based, secure centralised database, sponsored and managed by the Cystic Fibrosis Trust. Over 99% of people with CF across every UK specialist centre and clinic have consented to their data being submitted to the CF registry by their CF care teams. When consenting to the CF registry people have the option to consent to anonymised data being shared with researchers both in the UK and in other countries.

For people consented to both CFHealthHub and the CF registry, data from CFHealthHub, the CF registry and routine clinical data will be collected by the local centre on a standardised spreadsheet. Data will include: month and year of birth; sex; ethnicity; height; pancreatic and diabetic status; FEV1; body mass index; date of starting and stopping (if applicable) Ivacaftor/Tezacaftor/Elexacaftor; Pseudomonas status; medications; comorbidities; healthcare utilisation (oral and intravenous antibiotic days); adherence to Ivacaftor/Tezacaftor/Elexacaftor (measured by medication possession ratio); and effective adherence to preventative inhaled therapies.

Data will be collected and analysed from twelve months before Ivacaftor/Tezacaftor/Elexacaftor initiation and, annually, for five years after initiation. For those consented only to the CF registry, data will be collected as above but without inhaled therapy adherence data.

Design This is a cohort observational study utilising data collected as part of routine clinical practice and/or ongoing approved research projects. This design is being used to explore the effect of starting precision medicines on adherence to preventative inhaled therapy adherence and the place of preventative inhaled therapy adherence as a potential target to optimise outcomes for people with CF taking Ivacaftor/Tezacaftor/Elexacaftor.

Sample size A corresponding trial powered for an effect size of 8% FEV1 difference between control and treatment arm (based on a similar efficacy-effectiveness gap to that observed for ivacaftor) would require n=50 participants in each arm to achieve 90% power at 5% significance. Given an estimated pool of 150 control participants, and 1300 treated participants this should easily be reached.

For consideration of different effect sizes among different adherence, we note that a corresponding trial powered for an effect size of 3% FEV1 difference between high and low adherers would require n=400 participants in each arm to achieve 90% power at 5% significance.

In addition, these numbers demonstrate that a loss of availability of 20% of participants (i.e., reduction to 120 control patients and 1040 treated patients) would be tolerable.

Analysis FEV1, body mass index, age at starting ivacaftor, CFTR genotype, sex, Pseudomonas status, medications, comorbidities, healthcare utilisation, adherence to Ivacaftor/Tezacaftor/Elexacaftor (measured by medication possession ratio) and effective adherence to preventative inhaled therapies (where the person is consented to CFHealthHub) will be analysed from 12 months before Ivacaftor/Tezacaftor/Elexacaftor initiation and at six months after Kaftrio initiation, then 12-monthly intervals following that.

For the people consented to and with data within CFHealthHub, the inhaled therapy adherence data will be explored, including determinants of adherence, and comparisons will be performed among those with different levels of adherence. FEV1 and BMI data will be presented as change from baseline to determine the difference between real-world effectiveness and the efficacy observed in clinical trials. The highest value recorded in the six months prior to starting Ivacaftor/Tezacaftor/Elexacaftor will be used as the baseline. This is an approach taken in previous literature assessing the longitudinal effects of Ivacaftor. Regression analysis will be used to look at the association between MPR (Medication Possession Ratio) and clinical outcomes and the association between adherence to preventative inhaled therapies and clinical outcomes.

The approach will follow the target trial emulation approach. Time zero will be the date of Ivacaftor/Tezacaftor/Elexacaftor initiation among those eligible for Ivacaftor/Tezacaftor/Elexacaftor (treated group). Time zero for eligible controls (i.e., those ineligible for any CFTR modulators) will have their time zero drawn at random from the list of start dates for treated group. This approach of selecting time zero for controls is similar to the Salford Lung Study. As a sensitivity analysis, time zero will be set as 21st Aug 2020 for all participants since that is the date of Ivacaftor/Tezacaftor/Elexacaftor announcement by the UK CF Trust.

The outcome is best FEV1 in the defined study period, with mean FEV1 over that period as a sensitivity analysis. A previous analysis by the epidemiologic study of cystic fibrosis group has identified average FEV1 over a six-month period as potentially a more relevant marker of lung health compared to best FEV1.

Inverse probability weighting will be conducted to achieve exchangeability between treated and control groups at baseline. Identified confounders are: age; gender; baseline FEV1; and prior year IV days. Baseline FEV1 is the best FEV1 in a six-month period prior to time zero, with mean FEV1 in that period as a sensitivity analysis.

Analysis 1: this is a comparison of the best FEV1 over the six month period a month following time zero (then in 12-monthly intervals until Year 5) in a linear model of the weighted population.

Analysis 2: the efficacy-effectiveness gap is evaluated as the difference in the FEV1 effect size identified in analysis one, and the FEV1 effect size identified in the trial9 (which was 14.3% for those with a single copy of F508del mutation in comparison to placebo).

Analysis 3: this is a descriptive analysis in which adherence to preventative inhaled therapy is described in both treatment and control groups. It will be identified whether adherence decreases more in the treatment groups, and if there are specific individuals in the treatment group whose adherence drops substantially.

Analysis 4: a per-protocol trial will be emulated, in which individuals are only considered for the treatment group when their adherence remains sufficient on all medications, including Ivacaftor/Tezacaftor/Elexacaftor. Exploratory analyses of the health outcomes among people with differing levels of adherence will also be performed.