Abstract
Background:
To date clinical trials of cystic fibrosis transmembrane conductance regulator (CFTR) modulator therapies have focused on outcomes that can be captured under relatively short follow-up periods. The effectiveness of CFTR modulator therapies on survival has not been fully described.
Research Question:
What is the population-level treatment effect of initiating any CFTR modulator therapy compared to no initiation on the risk of death in people with cystic fibrosis?
Study Design and Methods:
We conducted a retrospective cohort study using United States Cystic Fibrosis Foundation Patient Registry data between 2012 and 2022. We included individuals who met eligibility criteria for a CFTR modulator prescription (based on age and genotype of 2020 regulatory approvals and no prior use). At every clinic or telehealth visit we assessed eligibility and classified treatment status based on registry records. We included individuals in the analysis multiple times if eligibility was met at multiple visits. We followed individuals from each visit until death (outcome), lung transplantation, loss to follow-up, or administrative censoring. We censored untreated observations if they later initiated CFTR modulator therapy. We estimated 8-year risk curves using Kaplan-Meier and hazard ratios from Cox regression models. We used inverse probability weighting to balance baseline covariates and account for non-random censoring.
Results:
We included 25,103 individuals and 178,835 visits; 18,056 individuals initiated therapy. The 8-year risk difference of death due to initiating CFTR modulator therapy was −7.2% (95% CI: −9.6, −4.9). The hazard for death was 66% lower after CFTR modulator initiation than without use (HR=0.34; 95% CI:0.28,0.41).
Interpretation:
CFTR modulator therapies reduce the risk of death among people living with cystic fibrosis over 8-years. This improved understanding will allow individuals living with CF and clinical providers to have reasonable expectations about survival and appropriately allocate care and resources aligned with increased life expectancy.
Keywords: cystic fibrosis, cystic fibrosis transmembrane conductance regulator, death, registry
Introduction
While there have been numerous advancements in cystic fibrosis (CF) care and management over recent decades, Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) modulators are genetically-targeted therapies that revolutionized treatment and vastly improved the prognosis for people living with CF.1–4 CFTR modulator therapies help to fix CFTR protein dysfunction caused by CFTR genetic variants, through potentiation of chloride channels (ivacaftor), correction of cellular folding and protein movement (lumacaftor and tezacaftor), and amplification of protein production (elexacaftor).2 Since regulatory approval of the first CFTR modulator therapy in 2012 (ivacaftor) for which less than 10% of individuals with CF were eligible, four combination CFTR modulator therapies have been developed and regulatory approvals expanded to genetic variants and age groups beyond initial indications.2,5–9
Elexacaftor/tezacaftor/ivacaftor was first approved in October 2019 for children ages 12 and older and indicated for the most common genetic variant (at least one copy of F508del), which is present in approximately 80% of people with CF.2,5,6 Therefore, as of 2021, almost 90% of people with CF across all ages and genotypes were eligible to receive a prescription for CFTR modulator therapy.2,5,6
To date clinical trials of CFTR modulator therapies have focused on pulmonary, nutritional, and quality of life outcomes that can be captured within relatively short follow-up periods.5,6,10 But short follow-up periods and low numbers of deaths in clinical trials made it difficult to assess therapeutic effects of CFTR modulator therapies on survival.10 Trial results show that CFTR modulator therapies markedly improve lung functioning and reduce pulmonary exacerbations, suggesting subsequent improvements in life expectancy.10 It is also postulated that initiation of CFTR modulator therapies at younger ages prior to significant disease progression may improve long-term disease trajectories.11 Also due to limited follow-up data, most observational studies examining death outcomes thus far have focused on the effectiveness of ivacaftor, the CFTR modulator therapy with the earliest approval but for which only 10% of individuals with CF are eligible.5,10,12–15 Registry reports and a post-marketing study of elexacaftor/tezacaftor/ivacaftor demonstrate decreased mortality rates over time, however these analyses were descriptive and causal inferences cannot be drawn.16,17 Modeling and simulation studies predict the effect of CFTR modulator therapies on long-term outcomes, such as life expectancy.10,18–20 Beyond these studies, only indirect evidence for CFTR modulator therapies through surrogate markers for death exist.10
Since people with CF are living longer than ever before and life expectancy is anticipated to continue increasing, understanding the direct effect of CFTR modulator therapies on survival is crucial for CF healthcare providers to strategize how to meet the changing clinical needs of aging CF patients and for people with CF to have reasonable expectations about their own survival.21,22 Therefore, we aimed to estimate the average treatment effect of initiating any CFTR modulator therapy compared to not initiating any CFTR modulator therapy (and remaining solely on other standards of care) on risk of death from any cause.
Study Design and Methods
Study Design
We conducted a retrospective cohort study using data from the United States (US) Cystic Fibrosis Foundation patient registry from 2012 through 2022. All individuals provided consent or assent to participate in the registry.23 The Johns Hopkins Bloomberg School of Public Health Institutional Review Board approved this analysis of the registry data (IRB No. 24592). We report this study according to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.24
Data Source
The US CF Foundation patient registry includes patients with CF seen within the nationwide network of 130 accredited CF clinical care facilities in the US between 1986 and the present.23 Demographic, CF diagnosis, anthropometric, laboratory, pulmonary functioning, CF-related complications, therapeutics, and hospitalization data are collected for these patients. Since 2010, health care providers in the nationwide network enter these data into an online portal for each clinic visit, hospitalization, home intravenous treatment, telehealth visit or other encounter.
Participants
We included individuals who could have received an initial prescription for any CFTR modulator therapy (based on age and genotype of regulatory approvals as of 2020 and no prior use). We allowed for the inclusion of historic controls by expanding the 2020 regulatory criteria to all time after the date of initial CFTR modulator-specific regulatory approvals, under the assumption that if the 2020 regulatory approvals were in place earlier these individuals would have been eligible for therapy. For example, ivacaftor was initially approved on January 31, 2012 for people age six or older with one genetic variant (G551D) but by December 2020 approvals expanded to people age four months or older with one of 97 genetic variants. We considered individuals eligible for ivacaftor if they were at least 4 months old as of January 31, 2012 with one of the 97 approved genetic variants. We assessed CFTR modulator eligibility at every clinic or telehealth visit more than 30 days apart between January 31, 2012 (ivacaftor’s initial approval) and December 31, 2020. We included individuals in the analysis multiple times if eligibility was met at multiple visits (e-Figure 1).25
Exposure
We estimated the average treatment effect of initiation of any CFTR modulator therapy compared to no CFTR modulator therapy use (receiving only other standards of care), on the risk of death from any cause under the hypothetical scenario where there are no lung transplantations. Treatment status was assigned at each clinic or telehealth visit (baseline for unique observations) based on prescriptions in registry records (e-Figure 2A). We classified observations as treated when individuals received their first CFTR modulator prescription at the visit where eligibility was assessed. We classified observations as untreated when individuals did not receive a CFTR modulator prescription at that baseline visit. Due to high numbers of individuals transitioning from untreated to treated, we censored records classified as untreated when individuals later initiated CFTR modulator therapy. We included records classified as treated regardless of whether individuals discontinued treatment; thus the hypothetical intervention corresponding to the treatment arm is treatment initiation rather than initiation and persistence. Our results will underestimate the etiologic effect of CFTR modulator therapies to the degree that patients do not adhere to treatment.26 We describe treatment cessation in sensitivity analyses.
Outcome
We followed individuals from each baseline clinic or telehealth visit until death (outcome), lung transplantation, loss to follow-up, or administrative censoring on December 31, 2022 (e-Figure 2B). Loss to follow-up was defined as at least one year since the last clinic or telehealth visit.
Statistical Analysis
For each treatment group we estimated 8-year cumulative incidence (risk) curves non-parametrically as the complement of the Kaplan-Meier survival estimator. From these curves, we estimated risk differences and risk ratios at 1-year, 5-years, and 8-years from the start of follow-up for the CFTR modulator therapy group compared to no CFTR modulator therapy group. We also estimated overall hazard ratios (HRs) from Cox proportional hazards models.
For all analyses, we used inverse probability of treatment weighting to balance measured baseline covariates between treatment groups. Baseline covariates were prognostic factors assessed at the visit where eligibility was assessed and included: age, gender, race, ethnicity, genotype, geographic region, education, insurance coverage, smoking status (individual and household), chronic comorbidities, pulmonary functioning, prior infections with Burkholderia cepacia and other Burkholderia species, number of hospitalizations within the past year, and calendar period. Genotype was characterized by presence of the delta F508 variant and class severity. Education was defined as the maximum level of education among the patient and parents. Pulmonary functioning was assessed by the Global Lung Initiative (GLI) percent predicted forced expiratory volume in one second (FEV1), recent pulmonary exacerbations, and use of oxygen therapy or a ventilation machine. Chronic comorbidities include CF-related diabetes, liver disease, pancreatitis, anxiety, and depression. Calendar period was used to capture shifting availability of different CFTR modulator therapies as well as clinical care changes during the COVID-19 pandemic. Covariates were defined based on data from encounter and annual files collected at CF clinical care center visits, which are recommended quarterly.2 We assessed covariate balance between treatment groups before and after weighting with standardized mean differences and a cutoff value of 0.1, visualized with a Love plot.
We used inverse probability of censoring weights updated at quarterly intervals to adjust for selection bias induced by non-independent censoring due to lung transplantation, loss to follow-up, or initiation of treatment in the untreated observations. Covariates for the censoring weights included all the time-fixed baseline covariates as well as time-varying covariates. The following variables were time-updated using annual-level data: insurance coverage, education, geographic region, smoking status (individual and household), oxygen therapy, use of a ventilation machine, and number of recent hospitalizations. The following variables were time-updated using encounter-level data: age, GLI percent predicted FEV1, chronic comorbidities, and prior infections with Burkholderia species.
We multiply imputed variables missing at baseline and missing time-varying values of GLI percent predicted FEV1. For all other time-varying missing values we carried the last recorded observation forward. We generated forty imputed data sets with chained equations using all exposure, outcome, and covariate variables described above, except number of hospitalizations and calendar period. We conducted analyses within each imputed data set, then used Rubin’s rule to combine point estimates across the forty imputations.27
For the Kaplan-Meier based risk estimates, standard errors were estimated as the standard deviation of estimates across 200 bootstrap re-samples of the data within each of the forty imputations. For the Cox proportional hazard models, we used a robust sandwich variance estimator to account for lack of independent observations.
All analyses were conducted using SAS 9.4 (SAS, Cary, NC). Covariate balance was assessed using the cobalt package in R studio version 4.4.0.28
Secondary Analyses
We tested whether the effect of CFTR modulator therapy was modified by age of initiation and gender. Age of initiation was categorized as: <6 years, 6–11 years, 12–17 years, 18–29 years, and 30+ years. We used a likelihood ratio test to compare Cox proportional hazards regression models with and without an interaction term between the effect modifier and CFTR modulator initiation. If the interaction was significant, then we conducted stratified analyses.
Sensitivity Analyses
To aid interpretation of our results, we described CFTR modulator therapy discontinuation. We wanted to quantify how our estimate of effect of CFTR modulator therapy initiation might differ from the effect of CFTR modulator therapy persistence. We did not estimate the effect of CFTR modulator persistence directly, because it involves a potentially untenable hypothetical intervention to maintain patients on therapy even if contraindicated due to intolerance or side effects. We defined discontinuation as no prescription for at least 120 days before the last recorded encounter date under the assumption that individuals may have extra pills on hand due to not taking as prescribed. We also assessed the time individuals were on CFTR modulator therapies from their first recorded prescription to 30 days after their last recorded prescription.
Results
Our study included 25,103 individuals who attended 178,835 clinic or telehealth visits where they were eligible to receive a CFTR modulator prescription. Individuals contributed a median (interquartile range (IQR)) of 4 (2–11) visits to the analysis (Figure 1). The study sample was 52% female, 95% white, 92% had at least one copy of the delta F508 variant, and 98% had a severe genetic variant (Class I, II, or III) (Table 1). On their first eligible visit, most individuals did not smoke and 12% reported smoking within their household. The maximum education level was college graduate or higher for 52% of patient families.
Figure 1.
Flow diagram for selection of clinic or telehealth visits where individuals in the US Cystic Fibrosis Foundation patient registry were eligible to initiate CFTR modulator therapy between 2012–2020
Table 1.
Baseline characteristics at first (individuals) and each clinic or telehealth visit in the US Cystic Fibrosis Foundation patient registry between January 2012 and December 2020 when individuals could have initiated CFTR modulator therapy
| Total Individuals (n=25,103) | Total eligible visits (n=178,835) | Initiated CFTR modulator (n=18,506) | Did not Initiate CFTR modulator (n=160,329) | ||
|---|---|---|---|---|---|
| Demographics | |||||
| Male Gender, n(%) | 13,098 (52) | 89,212 (50) | 9,729 (53) | 79,483 (50) | |
| Age, median(IQR) | 18.5 (9.0–29.0) | 14.4 (6.4–27.6) | 19.4 (11.6–30.0) | 13.8 (6.1–27.2) | |
| Race, n(%) | White | 23,957 (95) | 170,163 (95) | 17,740 (96) | 152,423 (95) |
| Black | 819 (3) | 6,066 (3) | 536 (3) | 5,530 (3) | |
| Other* | 409 (2) | 3,580 (2) | 275 (1) | 3,305 (2) | |
| Hispanic, n(%) | 1,779 (7) | 15,136 (9) | 1,232 (7) | 13,904 (9) | |
| Missing | 909 | 6,828 | |||
| Region | Northeast | 4,569 (18) | 34,055 (19) | 3,335 (18) | 30,720 (19) |
| Southeast | 6,889 (27) | 47,283 (26) | 5,231 (28) | 41,951 (26) | |
| Midwest | 6,667 (27) | 49,335 (28) | 4,864 (26) | 44,471 (28) | |
| West | 4,471 (18) | 30,372 (17) | 3,266 (18) | 27,106 (17) | |
| Southwest | 2,507 (10) | 16,081 (10) | 1,810 (10) | 16,081 (10) | |
| Person who smokes, n(%) | Ever | 341 (1) | 2,103 (1) | 181 (1) | 1,922 (1) |
| Missing | 1,124 | 6,102 | |||
| Smoking in Household, n(%) | 2,437 (12) | 18,498 (12) | 1,640 (11) | 16,858 (12) | |
| Missing | 4,611 | 27,840 | |||
| Education**, n(%) | Less than High School | 737 (4) | 5,952 (4) | 473 (3) | 5,479 (4) |
| High School diploma or equivalent | 4,063 (21) | 28,972 (21) | 2,990 (20) | 25,982 (21) | |
| Some College | 4,507 (23) | 31,530 (23) | 3,352 (23) | 28,178 (23) | |
| College Graduate | 7,488 (38) | 52,094 (38) | 5,802 (39) | 46,292 (38) | |
| Masters/Doctoral level | 2,748 (14) | 20,192 (15) | 2,266 (15) | 17,926 (14) | |
| Missing | 5,560 | 40,095 | |||
| Public Insurance, n(%) | 9,311 (37) | 71,201 (40) | 6,634 (36) | 64,567 (41) | |
| Missing | 113 | 1,053 | |||
| CF Severity/Progression | |||||
| Severe Variant (class I, II, or III), n(%) | 24,544 (98) | 171,862 (96) | 18,232 (99) | 153,630 (96) | |
| Delta F508 variant, n(%) | Homozygous | 13,614 (54) | 96,840 (54) | 10,385 (56) | 86,455 (54) |
| Heterozygous | 9,736 (38) | 59,192 (33) | 6,917 (38) | 52,275 (33) | |
| GLI predicted FEV1, median (IQR) | 77.7 (55.3–94.3) | 83.4 (61.5–98.2) | 80.8 (58.7–95.8) | 83.8 (61.8–98.5) | |
| Missing | 6,790 | 53,569 | |||
| Oxygen Therapy, n(%) | 2,562 (10) | 16,234 (9) | 1,767 (10) | 14,467 (9) | |
| Missing | 456 | 3,207 | |||
| Ventilation Machine, n(%) | 638 (3) | 4,527 (3) | 452 (2) | 4,075 (3) | |
| Missing | 592 | 3,568 | |||
| Prior Burkholderia cepacia, n(%) | 1,338 (5) | 7,602 (4) | 1,184 (6) | 6,418 (4) | |
| Prior Other Burkholderia species, n(%) | 945 (4) | 5,586 (3) | 862 (5) | 4,724 (3) | |
| BMI z-score, median (IQR) | 0.1 (−0.6–0.7) | 0.2 (−0.5–0.9) | 0.1 (−0.5–0.8) | 0.2 (−0.5–0.9) | |
| Missing | 2,976 | 20,390 | |||
| Pulmonary Exacerbation***, n(%) | 2,529 (10) | 16,564 (9) | 1,309 (7) | 15,255 (10) | |
| Hospitalizations prior year, | 0 | 17,882 (71) | 127,654 (71) | 13,839 (75) | 113,815 (71) |
| n(%) | 1–4 | 6,756 (27) | 47,659 (27) | 4,426 (24) | 43,233 (27) |
| 5+ | 465 (2) | 3,522 (2) | 241 (1) | 3,281 (2) | |
| Comorbidities | |||||
| CF-related diabetes, n(%) | 5128 (20) | 27,073 (15) | 4,099 (22) | 22,974 (14) | |
| Liver Disease, n(%) | 1288 (5) | 7,937 (4) | 995 (5) | 6,942 (4) | |
| Pancreatitis, n(%) | 357 (1) | 2,765 (2) | 135 (1) | 2,630 (2) | |
| Anxiety, n(%) | 1,915 (8) | 12,884 (7) | 1,935 (10) | 10,949 (7) | |
| Depression, n(%) | 2,841 (11) | 18,100 (10) | 2,501 (14) | 15,599 (10) | |
| Temporality | |||||
| Calendar period, n(%) | Jan 31, 2012 – Jul 1, 2015 | 4,405 (18) | 33,960 (20) | 1,249 (7) | 33,960 (21) |
| Jul 2, 2015 – Feb 11, 2018 | 13,376 (53) | 93,435 (52) | 7,731 (42) | 85,704 (53) | |
| Feb 12, 2018 – Oct 21, 2019 | 1,474 (6) | 30,400 (17) | 3,761 (20) | 26,639 (17) | |
| Oct 21, 2019 – Dec 31, 2020 | 5,848 (23) | 19,791 (11) | 5,765 (31) | 14,026 (9) |
Other race includes: American Indian or Alaska Native, Asian, and Native Hawaiian or Pacific Islander
Maximum education for mother/father/patient
Moderate or severe exacerbation in prior year
Ultimately, 18,506 individuals (74%) initiated CFTR modulator therapy at an eligible visit. Compared to patient-visits at which therapy was not initiated, patients were more likely to initiate at a visit when they had a higher comorbidity burden, prior Burkholderia species infection, and private insurance coverage. However, after applying inverse probability of treatment weighting, all measured baseline covariates were balanced between treatment groups (e-Figure 3).
Overall, there were 3,797 deaths (15% of individuals) during follow-up. We censored 6,226 observations due to loss to follow-up (3.5%), 1,849 due to lung transplantation (1.0%), 3,940 due to administrative end of follow-up (2.2%), and 137,075 due to non-CFTR modulator users initiating therapy (77%).
We estimated that the 1-year, 5-year, and 8-year risks of death were 0.4% (95% CI: 0.1, 0.6), 2.5% (95% CI: 2.2, 2.8), and 4.1% (95% CI: 3.5, 4.8) following CFTR modulator therapy initiation compared to 1.0% (95% CI: 0.7, 1.4), 8.1% (95% CI: 6.5, 9.7), and 11.4% (95% CI: 9.2, 13.5) without therapy (Table 2, with full risk curves from one imputation presented in Figure 2). On average, if everyone had initiated CFTR modulator therapy, the 1-year risk of death would be 0.7% (95% CI: −1.1, −0.2) lower, the 5-year risk of death would be 5.6% (−7.2, −3.9) lower, and the 8-year risk of death would be 7.2% (95% CI: −9.6, −4.9) lower compared to if no one had initiated therapy (Table 2). The corresponding risk ratios are also presented in Table 2. Though there is possible departure from proportional hazards (e-Figure 4), the overall hazard for death over 8 years was 66% lower after CFTR modulator therapy initiation than without CFTR modulator therapy (HR=0.34; 95% CI:0.27,0.42).
Table 2.
Risks, risk differences, and risk ratios based on the Kaplan-Meier estimator for death at 1-year, 5-years, and 8-years for individuals in the US Cystic Fibrosis Foundation patient registry who initiated and did not initiate CFTR modulator therapy
| Risk (%) (95% CI*) | Risk Difference (95% CI*) | Risk Ratio (95% CI*) | ||
|---|---|---|---|---|
| Initiated CFTR modulator | Never initiated CFTR modulator | |||
| 1-year follow-up | 0.39 (0.14, 0.64) | 1.04 (0.66, 1.43) | −0.65 (−1.08, −0.22) | 0.38 (0.27, 0.52) |
| 5-years follow-up | 2.48 (2.18, 2.79) | 8.05 (6.45, 9.65) | −5.57 (−7.24, −3.90) | 0.31 (0.24, 0.39) |
| 8-years follow-up | 4.14 (3.47, 4.81) | 11.35 (9.21, 13.50) | −7.21 (−9.56, −4.86) | 0.37 (0.28, 0.48) |
CI= confidence interval based on 200 bootstraps within 40 imputations combined using Rubin’s rule
Figure 2.
Inverse probability weighted cumulative risk functions based on the Kaplan-Meier estimator for death (from a single imputation) for individuals in the US Cystic Fibrosis Patient Registry who did and did not initiate CFTR modulator therapy between January 31, 2012 and December 31, 2020 with follow-up through 2022
We observed differences of effect across strata for age of initiation but not for gender. Results are not reported for those who initiated CFTR modulator therapies under the age of 12 due to few deaths occurring within the early years of life. The hazard ratio for death associated with CFTR modulator therapy initiation was 0.51 (95% CI: 0.37, 0.71) for those initiating between ages 12–17 years, 0.28 (95% CI: 0.20, 0.39) for those initiating between 18–29 years, and 0.30 (95% CI: 0.23, 0.39) for those initiating at age 30 or older. The 1-year, 5-year, and 8-year risks, risk differences, and risk ratios by age category are presented in e-Table 1. The hazard ratio for death associated with CFTR modulator therapy initiation was 0.31 (95% CI: 0.22, 0.44) for males and 0.35 (95% CI: 0.30, 0.42) for females. The 1-year, 5-year, and 8-year risks, risk differences, and risk ratios by gender are presented in e-Table 2.
Among those who initiated CFTR modulator therapy, only 1,527 (6.7%) discontinued use during follow-up. Median time on therapy was 1.9 years (IQR: 0.6–3.8) for those who discontinued and 4.2 years (IQR: 2.8–6.3) for those who remained on therapy.
Discussion and Interpretation
Among individuals living with CF in the US from 2012 through 2020, those who initiated CFTR modulator therapy had a 7% lower risk of death than those who did not initiate over an 8-year follow-up period. Our results are consistent with existing studies suggesting that CFTR modulator therapies lead to drastically improved survival for people living with CF. Our study has improved upon prior studies by estimating effectiveness over a much longer follow-up period and with greater precision and validity.10 Observational studies have focused on ivacaftor due to its earlier regulatory approval, and using US and UK patient registry data showed between 60–78% reduction in death.5,10,12–15 However, these studies were limited by cross-sectional assessments and only 2–5 years of follow-up,12–14 as well as selecting a comparator group with genotypes potentially unresponsive to ivacaftor.15 Furthermore, as fewer than 10% of individuals with CF have ivacaftor-responsive genotypes, previous results may lack generalizability. Our study includes a broader study sample than prior observational studies and most clinical trials,10 including people under the age of six and having a wider range of lung function. Our study is the largest registry-based cohort with the longest longitudinal follow-up examining the real-world effectiveness of CFTR modulator therapies on death to date.
We estimated the average effectiveness of all four currently approved CFTR modulator therapies. However, individual drugs may be differentially effective. Ivacaftor and elexacaftor/tezacaftor/ivacaftor demonstrated the largest improvement in lung functioning in trials;5,10 they may also be more effective at improving survival. Our risk estimates are influenced by the timing of regulatory approvals for each CFTR modulator therapy, with 8-year estimates driven by ivacaftor, 5-year estimates adding the effects of lumacaftor/ivacaftor and tezacaftor/ivacaftor, and 1-year estimates also integrating the effect of elexacaftor/tezacaftor/ivacaftor. Additionally, the effect of each CFTR modulator therapy may differ at the individual level due to specific combinations of genetic variants,29 dose, food intake and content (CFTR modulators are lipophilic), body weight, age, pancreatic and hepatic functioning, and concomitant medication use.30 While we observed lower effectiveness of CFTR modulator therapies for those initiating between the ages of 12–17, we believe that this was likely due to lower therapeutic adherence during adolescence which has been observed in other disease areas31 due to transitions to independent medication management, and higher social embarrassment, disease denial, and risk-taking behaviors.32 We cannot be certain about individual adherence since the registry captures provider-reported prescription records, but we believe that effectiveness rather than efficacy better reflects the real-world population effect. We did not observe a meaningful difference in CFTR modulator therapeutic effect on death by gender, however prior studies suggest sex-specific CFTR modulator effects for other outcomes.33,34 Despite some heterogeneity in magnitude of effects, overall CFTR modulator therapies are highly effective in delaying death in people living with CF.
The US CF Foundation Patient Registry contains high quality data, but as with any observational study, our analyses were limited by data availability. We accounted for the confounding effects of socioeconomic status through inclusion of familial education level, public insurance coverage, and region in inverse probability of treatment weights. However, since CFTR modulator therapies are extremely expensive, and factors such as income and co-pays are not fully measured, there is the potential for residual confounding and overestimation of therapeutic effectiveness. Additionally, we captured disease severity by genetic variant class, GLI predicted FEV1, ventilation machine or oxygen therapy use, mycobacterial infections, and recent pulmonary exacerbations and hospitalizations. But residual confounding due to other signs and symptoms of severe disease not captured may lead to an underestimation of therapeutic effectiveness while residual confounding due to anticipated high therapeutic adherence may lead to an overestimation of therapeutic effectiveness. Finally, while there are over 1,900 CFTR genetic variants that cause CF, our assessment of genetic eligibility for CFTR modulator therapies was limited to the 272 variants captured in the registry.
In conclusion, CFTR modulator therapy use drastically reduces the risk of death among people living with CF. A better understanding of the effect of CFTR modulator therapies on death will allow individuals living with CF, their families, and clinical providers to have reasonable expectations about their survival and appropriately plan and allocate care and resources as needs change with increased life expectancy. Although about 90% of individuals with CF are now eligible for CFTR modulator therapies, they remain inaccessible to some due to unresponsive genetic variants, lack of tolerance, and high costs, but new treatments are under development that will hopefully benefit them.5,35–37 Furthermore, hopefully demonstration of broader effectiveness of CFTR modulator therapies will motivate expanded access through reduced costs and wider insurance coverage,38 and further reduce CF mortality through increased uptake.
Supplementary Material
Take Home Points.
Study Question:
What is the average treatment effect of initiating any CFTR modulator therapy compared to no initiation on the risk of death in people with cystic fibrosis?
Results:
The 8-year risk of death would be 63% lower (RR=0.37; 95% CI: 0.28, 0.48) if everyone initiated a CFTR modulator therapy compared to if no one initiated CFTR modulator therapy.
Interpretation:
CFTR modulator therapies reduce the risk of death among people living with cystic fibrosis over 8-years.
Acknowledgements:
The project described was supported by Grant Number T32 AG000247 from the National Institute on Aging, National Institutes of Health.
The authors would like to thank the Cystic Fibrosis Foundation for the use of CF Foundation Patient Registry data to conduct this study. Additionally, we would like to thank the patients, care providers, and clinic coordinators at CF centers throughout the United States for their contributions to the CF Foundation Patient Registry.
Abbreviations:
- BMI
Body mass index
- CF
Cystic fibrosis
- CFTR
Cystic fibrosis transmembrane conductance regulator
- CI
Confidence intervals
- FEV1
Forced expiratory volume in one second
- GLI
Global lung initiative
- HR
Hazard ratio
- IQR
Interquartile range
- US
United States
Footnotes
Conflicts of Interest: None.
Ethics Approval: This study was determined to be exempt category 4 research by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board (IRB No. 24592).
Consent to participate: All participants provided consent or assent to participate in the US Cystic Fibrosis Foundation Patient Registry.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Burgel P-R, Burnet E, Regard L, Martin C. The Changing Epidemiology of Cystic Fibrosis. Chest 2023; 163: 89–99. [DOI] [PubMed] [Google Scholar]
- 2.Dickinson KM, Collaco JM. Cystic Fibrosis. Pediatr Rev 2021; 42: 55–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Stephenson AL, Stanojevic S, Sykes J, Burgel P-R. The changing epidemiology and demography of cystic fibrosis. Presse Médicale 2017; 46: e87–95. [DOI] [PubMed] [Google Scholar]
- 4.Barr HL, Bihouee T, Zwitserloot AM. A year in review: Real world evidence, functional monitoring and emerging therapeutics in 2021. J Cyst Fibros 2022; 21: 191–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.King JA, Nichols A-L, Bentley S, Carr SB, Davies JC. An Update on CFTR Modulators as New Therapies for Cystic Fibrosis. Pediatr Drugs 2022; 24: 321–33. [DOI] [PubMed] [Google Scholar]
- 6.Goetz DM, Savant AP. Review of CFTR modulators 2020. Pediatr Pulmonol 2021; 56: 3595–606. [DOI] [PubMed] [Google Scholar]
- 7.Regard L, Martin C, Burnet E, Da Silva J, Burgel P-R. CFTR Modulators in People with Cystic Fibrosis: Real-World Evidence in France. Cells 2022; 11: 1769. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Costa E, Girotti S, Pauro F, Leufkens HGM, Cipolli M. The impact of FDA and EMA regulatory decision-making process on the access to CFTR modulators for the treatment of cystic fibrosis. Orphanet J Rare Dis 2022; 17: 188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Keating C, Yonker LM, Vermeulen F, et al. Vanzacaftor–tezacaftor–deutivacaftor versus elexacaftor–tezacaftor–ivacaftor in individuals with cystic fibrosis aged 12 years and older (SKYLINE Trials VX20–121-102 and VX20–121-103): results from two randomised, active-controlled, phase 3 trials. Lancet Respir Med 2025; 13: 256–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Balfour-Lynn IM, King JA. CFTR modulator therapies – Effect on life expectancy in people with cystic fibrosis. Paediatr Respir Rev 2022; 42: 3–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Nichols DP, Donaldson SH, Frederick CA, et al. PROMISE: Working with the CF community to understand emerging clinical and research needs for those treated with highly effective CFTR modulator therapy. J Cyst Fibros 2021; 20: 205–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Bessonova L, Volkova N, Higgins M, et al. Data from the US and UK cystic fibrosis registries support disease modification by CFTR modulation with ivacaftor. Thorax 2018; 73: 731–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Higgins M, Volkova N, Moy K, Marshall BC, Bilton D. Real-World Outcomes Among Patients with Cystic Fibrosis Treated with Ivacaftor: 2012–2016 Experience. Pulm Ther 2020; 6: 141–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Volkova N, Moy K, Evans J, et al. Disease progression in patients with cystic fibrosis treated with ivacaftor: Data from national US and UK registries. J Cyst Fibros 2020; 19: 68–79. [DOI] [PubMed] [Google Scholar]
- 15.Merlo CA, Thorat T, DerSarkissian M, et al. Long-term impact of ivacaftor on mortality rate and health outcomes in people with cystic fibrosis. Thorax 2024; : thorax-2023–220558. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bower JK, Volkova N, Ahluwalia N, et al. Real-world safety and effectiveness of elexacaftor/tezacaftor/ivacaftor in people with cystic fibrosis: Interim results of a long-term registry-based study. J Cyst Fibros 2023; 22: 730–7. [DOI] [PubMed] [Google Scholar]
- 17.Cystic Fibrosis Foundation. Cystic Fibrosis Foundation Patient Registry 2023 Annual Data Report. Bethesda,MD; 2024. [Google Scholar]
- 18.Dilokthornsakul P, Hansen RN, Campbell JD. Forecasting US ivacaftor outcomes and cost in cystic fibrosis patients with the G551D mutation. Eur Respir J 2016; 47: 1697–705. [DOI] [PubMed] [Google Scholar]
- 19.Rubin JL, O’Callaghan L, Pelligra C, et al. Modeling long-term health outcomes of patients with cystic fibrosis homozygous for F508del-CFTR treated with lumacaftor/ivacaftor. Ther Adv Respir Dis 2019; 13: 1753466618820186. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lopez A, Daly C, Vega-Hernandez G, MacGregor G, Rubin JL. Elexacaftor/tezacaftor/ivacaftor projected survival and long-term health outcomes in people with cystic fibrosis homozygous for F508del. J Cyst Fibros 2023; 22: 607–14. [DOI] [PubMed] [Google Scholar]
- 21.Sala MA, Vitale KM, Prickett M. Looking toward the future: Approaching care of the aging CF patient. Pediatr Pulmonol 2022; 57. DOI: 10.1002/ppul.25745. [DOI] [PubMed] [Google Scholar]
- 22.Blankenship S, Landis AR, Harrison Williams E, et al. What the future holds: cystic fibrosis and aging. Front Med 2024; 10: 1340388. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Knapp EA, Fink AK, Goss CH, et al. The Cystic Fibrosis Foundation Patient Registry. Design and Methods of a National Observational Disease Registry. Ann Am Thorac Soc 2016; 13: 1173–9. [DOI] [PubMed] [Google Scholar]
- 24.Von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for reporting observational studies. Int J Surg 2014; 12: 1495–9. [Google Scholar]
- 25.Hernán MA, Robins JM. Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available: Table 1. Am J Epidemiol 2016; 183: 758–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Hansen CME, Breukelman AJ, Van Den Bemt PMLA, Zwitserloot AM, Van Dijk L, Van Boven JFM. Medication adherence to CFTR modulators in patients with cystic fibrosis: a systematic review. Eur Respir Rev 2024; 33: 240060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Rubin DB. Multiple Imputation After 18+ Years. J Am Stat Assoc 1996; 91: 473–89. [Google Scholar]
- 28.Greifer N cobalt: Covariate Balance Table and Plots. 2024. R package version 4.5.5. [Google Scholar]
- 29.Sosnay PR, Siklosi KR, Van Goor F, et al. Defining the disease liability of variants in the cystic fibrosis transmembrane conductance regulator gene. Nat Genet 2013; 45: 1160–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Van Der Meer R, Wilms EB, Heijerman HGM. CFTR Modulators: Does One Dose Fit All? J Pers Med 2021; 11: 458. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Foster C, Ayers S, Fidler S. Antiretroviral adherence for adolescents growing up with HIV: understanding real life, drug delivery and forgiveness. Ther Adv Infect Dis 2020; 7: 2049936120920177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Sawicki GS, Tiddens H. Managing treatment complexity in cystic fibrosis: Challenges and Opportunities: Addressing Complexity in Cystic Fibrosis. Pediatr Pulmonol 2012; 47: 523–33. [DOI] [PubMed] [Google Scholar]
- 33.Wang A, Lee M, Keller A, et al. Sex differences in outcomes of people with cystic fibrosis treated with elexacaftor/tezacaftor/ivacaftor. J Cyst Fibros 2024; 23: 91–8. [DOI] [PubMed] [Google Scholar]
- 34.Secunda KE, Guimbellot JS, Jovanovic B, et al. Females with Cystic Fibrosis Demonstrate a Differential Response Profile to Ivacaftor Compared with Males. Am J Respir Crit Care Med 2020; 201: 996–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Allen L, Allen L, Carr SB, et al. Future therapies for cystic fibrosis. Nat Commun 2023; 14: 693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Despotes KA, Donaldson SH. Current state of CFTR modulators for treatment of Cystic Fibrosis. Curr Opin Pharmacol 2022; 65: 102239. [DOI] [PubMed] [Google Scholar]
- 37.Desai M, Hine C, Whitehouse JL, Brownlee K, Charman SC, Nagakumar P. Who are the 10%? - Non eligibility of cystic fibrosis (CF) patients for highly effective modulator therapies. Respir Med 2022; 199: 106878. [DOI] [PubMed] [Google Scholar]
- 38.Seyoum S, Regenstein M, Nolan L. Cost, coverage, and the underuse of medications among people with CF. Himmelfarb Health Sciences Library: Health Policy and Management Issue Brief. Milliken Institute School of Public Health, The George Washington University; 2020. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.