Efficacy and safety of ataluren in patients with nonsense-mutation cystic fibrosis not receiving chronic inhaled aminoglycosides: The international, randomized, double-blind, placebo-controlled Ataluren Confirmatory Trial in Cystic Fibrosis (ACT CF)

Abstract

Background:

Ataluren was developed for potential treatment of nonsense-mutation cystic fibrosis (CF). A previous phase 3 ataluren study failed to meet its primary efficacy endpoint, but post-hoc analyses suggested that aminoglycosides may have interfered with ataluren’s action. Thus, this subsequent trial (NCT02139306) was designed to assess the efficacy and safety of ataluren in patients with nonsense-mutation CF not receiving aminoglycosides.

Methods:

Eligible subjects with nonsense-mutation CF (aged ≥6 years; percent predicted (pp) FEV₁ ≥40 and ≤90) from 75 sites in 16 countries were randomly assigned in double-blinded fashion to receive oral ataluren or matching placebo thrice daily for 48 weeks. The primary endpoint was absolute change in average ppFEV₁ from baseline to the average of Weeks 40 and 48.

Findings:

279 subjects were enrolled; 138 subjects in the ataluren arm and 136 in the placebo arm were evaluable for efficacy. Absolute ppFEV₁ change from baseline did not differ significantly between the ataluren and placebo groups at Week 40 (−0.8 vs −1.8) or Week 48 (−1.7 vs −2.4). Average ppFEV₁ treatment difference from baseline to Weeks 40 and 48 was 0.6 (95% CI −1.3, 2.5; p = 0.54). Pulmonary exacerbation rate per 48 weeks was not significantly different (ataluren 0.95 vs placebo 1.13; rate ratio p = 0.40). Safety was similar between groups. No life-threatening adverse events or deaths were reported.

Interpretation:

Neither ppFEV₁ change nor pulmonary exacerbation rate over 48 weeks were statistically different between ataluren and placebo groups. Development of a nonsense-mutation CF therapy remains elusive.

1. Introduction

More than 2000 different mutations of the cystic fibrosis transmembrane conductance regulator (CFTR) gene have been identified within the global CF population, of which ~8% are nonsense mutations resulting in premature termination of CFTR protein synthesis [1] occurring in ~10% of CF cases worldwide [2,3]. These CFTR nonsense mutations are categorized as ‘severe’, in that little to no protein product is produced by the ribosome, and patients carrying two of these types of mutations have very poor prognoses [4–8]. Drug developers have made remarkable progress in developing small molecule therapies that increase mutant CFTR protein function in persons with CF, but these advances have largely been limited to individuals with mutations resulting in complete mutant CFTR proteins that are improperly processed and/or catalytically diminished. Pharmacologic approaches to increasing CFTR protein activity in individuals with premature termination of CFTR protein synthesis from mRNA must necessarily target nucleic acids, either by changing the coding sequence of the CFTR gene or its mRNA or by modulating the translation of mRNA by the ribosome to ‘read-through’ the premature termination codon.

Ataluren (PTC124) is an oral agent that has been shown to allow ribosomes to read through premature termination codons [9–12] and produce full-length, functional CFTR protein [12,13]. A randomized, double-blind, placebo-controlled 48-week multinational trial of ataluren in 238 persons with CF and at least one nonsense CFTR mutation (NCT00803205) showed no statistically significant difference in change from baseline in percent predicted forced expiratory volume in 1 s (ppFEV₁) or rate of pulmonary exacerbations between treatment arms [14]. However, a post hoc subgroup analysis of study data suggested that subjects not receiving chronic inhaled tobramycin as a concomitant medication experienced ataluren-associated benefits, including a 5.7% difference in ppFEV₁ change from baseline and 40% fewer pulmonary exacerbations [14]. Affinity of aminoglycosides such as tobramycin for both the prokaryotic [15] and eukaryotic [16] ribosome supports a hypothesis that the activity of ataluren, which binds to the eukaryotic ribosome to facilitate stop codon readthrough [17], may be inhibited by intracellular tobramycin. This Phase 3 clinical trial (NCT02139306, the Ataluren Confirmatory Trial in Cystic Fibrosis; ACT CF) was designed to evaluate the efficacy and safety of ataluren in persons with CF and nonsense CFTR mutations not receiving concomitant inhaled aminoglycosides.

2. Methods

2.1. Study oversight

The protocol was reviewed and approved by the ethics committee or institutional review board of each participating institution. Written informed consent was obtained from patients or their custodians prior to patient screening. External oversight of the study was provided by a Study Steering Committee and an independent Data Safety Monitoring Committee.

2.2. Study design and participants

This randomized, double-blind, placebo-controlled, phase 3 trial was performed between August 2014 and November 2016 at 75 sites in 16 countries in North America, South America, Europe, and Israel. Included subjects were ≥6 years old with a body weight ≥16 kg, a sweat chloride >60 mEq/L, and documentation of the presence of a nonsense mutation in at least 1 allele of the CFTR gene. They also had the ability to perform a valid, reproducible spirometry tests, a screening ppFEV₁ ≥ 40 and ≤ 90, confirmed screening laboratory values within pre-specified central laboratory ranges, and a resting oxygen saturation by pulse oximetry of ≥92% on room air. Sexually active subjects had to be willing to abstain from sexual intercourse or employ a barrier or medical method of contraception during the study and 60-day follow-up period.

Major exclusion criteria were known hypersensitivity to any study drug ingredients or excipients, previous participation in the ataluren Phase 3 trial PTC124-GD-009-CF, changes in a chronic treatment/prophylaxis regimen for CF or for CF-related conditions within 4 weeks prior to screening or any change in acute therapy between screening and randomization, chronic use of inhaled or systemic tobramycin within 4 weeks prior to screening, exposure to another investigational drug within 4 weeks prior to screening, evidence of pulmonary exacerbation or acute upper or lower respiratory tract infection (including viral illnesses) within 3 weeks prior to screening or between screening and randomization, treatment with intravenous (IV) antibiotics within 3 weeks prior to screening, ongoing immunosuppressive therapy (other than corticosteroids), ongoing warfarin, phenytoin, or tolbutamide therapy, history of solid organ or hematological transplantation, major complications of lung disease (including massive hemoptysis, pneumothorax, or pleural effusion) within 8 weeks prior to screening, known portal hypertension, positive hepatitis B surface antigen, hepatitis C antibody test or human immunodeficiency virus (HIV) test, pregnancy or breast-feeding, smoking or a smoking history of ≥10 pack-years, and prior or ongoing medical conditions, medical history, physical findings, ECG findings, or laboratory abnormality that, in the investigator’s opinion, could have adversely affected the safety of the patient, made it unlikely that the course of treatment or follow up would be completed, or could have impaired the assessment of study results.

2.3. Randomization and masking

Eligible subjects were randomized in a 1:1 ratio to ataluren or placebo by means of interactive response technology, using a block size of 4 within the cells created by the interaction of the 3 stratification factors (current use of a chronic regimen of inhaled antibiotic [yes vs no], baseline age [<18 years versus ≥18 years], and baseline ppFEV₁ [< 65 vs ≥65]). Subsequently enrolled family members (e.g., siblings) were assigned to the same treatment group as the first family member enrolled to minimize unblinding potential and to reduce the chance of inadvertent dosing errors. The random allocation sequence was generated by the contract research organization. Patients, medical and ancillary staff, the study investigators, and the sponsor were masked to treatment assignment. Only designated personnel at the contract research organization had access to treatment assignments.

2.4. Procedures

The active study medication ataluren, a powder for oral suspension, was dosed thrice daily at 10, 10, 20 mg/kg (total daily dose 40 mg/kg) every day during the treatment period as morning, mid-day, and evening doses, respectively, for 48 weeks. Placebo was identical in appearance. The primary efficacy endpoint of the study was absolute ppFEV₁ change from baseline to Week 48, defined as the average of the change at Week 40 and that at Week 48. Spirometry was performed at screening, at randomization, and every eight weeks during the 48-week study duration. All sites used a study-specific spirometer for FEV₁, FVC, and FEF_25–75 assessments following American Thoracic Society/European Respiratory Society guidelines [18,19]. At baseline and each subsequent study visit, subjects completed the Cystic Fibrosis Questionnaire-Revised (CFQ-R) [20].

2.5. Statistical analysis

A sample size of 252 subjects (126/treatment group) was chosen to detect a 3% predicted greater average change from baseline ppFEV₁ (SD = 7) among subjects randomized to receive ataluren with >90% power using a two-sided t-test at a 0.05 significance level. Assuming a 10% premature discontinuation rate, an enrollment maximum of 280 patients was set. Efficacy analyses were performed on the intent-to-treat (ITT) population, defined as those patients who had at least 1 valid post-baseline spirometry measurement.

The primary efficacy endpoint (difference in average ppFEV₁ absolute change from baseline at Weeks 40 and 48) tested in the ITT population using mixed-model repeated-measures (MMRM) analysis of covariance (ANCOVA), with an unstructured variance/covariance matrix and ppFEV₁ change from baseline (defined as the ppFEV₁ average at screening and randomization) as the dependent variable. If the unstructured variance/covariance matrix did not converge, the analysis was to be replaced with an MMRM analysis with multiple imputation. Rate of pulmonary exacerbation (PEx), defined as clinician decision to treat with (oral, inhaled, and/or IV) antibiotics in the presence of ≥4 Fuchs sign and symptom criteria [21] was tested by a negative binomial regression model with independent variables including treatment group, 3 stratification factors, and baseline ppFEV₁. Maximum likelihood methods were used to fit the model. The analysis was performed using the generalized linear model as implemented in SAS by the GENMOD procedure [22]. All analyses were 2-sided at the 0.05 significance level, with analyses and data tabulations performed using SAS (Version 9.1 or higher).

2.6. Role of the funding source

The study sponsor oversaw trial management, data collection, statistical analyses, and the writing and review of the clinical study report. The corresponding author had full access to all data in the study and had final responsibility for the decision to submit for publication.

3. Results

In all, 279 subjects were randomly assigned to the ataluren treatment arm (N = 140) or to the placebo arm (N = 139) (Fig. 1). Treatment groups were generally well-balanced at randomization with respect to presence of CFTR nonsense mutation alleles and other demographic characteristics (Table 1). Use of concomitant CF medications, and notably chronic inhaled antibiotics, was somewhat more frequent among subjects randomized to receive ataluren: 86 ataluren subjects (61.4%) were receiving chronic inhaled antibiotics at randomization versus only 66 placebo subjects (47.5%)(Table 1). Five patients did not have a valid post-baseline spirometry measurement, resulting in an ITT population of 274 subjects, 138 subjects in the ataluren arm and 136 in the placebo arm. Thirteen subjects withdrew from the ataluren treatment group and 14 withdrew from the placebo group; 252 subjects completed the study (Fig. 1).

Fig. 1.

Subject disposition. Subjects include in the intent-to-treat population were required to have at least one spirometry measure at Week 8.

Table 1.

Subject demographics at randomization.

	Ataluren (N = 140)	Placebo (N = 139)	Total (N = 279)
Mean Age, years (SD)	22.0 (11.00)	22.0 (10.44)	22.0 (10.70)
Min, Max, years	6, 52	7, 52	6, 52
Age Categories, n (%)
≥6 – < 12 years	25 (17.9)	19 (13.7)	44 (15.8)
≥12 – <18 years	34 (24.3)	37 (26.6)	71 (25.4)
≥18 years	81 (57.9)	83 (59.7)	164 (58.8)
Females, n (%)	59 (42.1)	74 (53.2)	133 (47.7)
Race category, n (%)
Caucasian	136 (97.1)	133 (95.7)	269 (96.4)
Asian	1 (0.7)	4 (2.9)	5 (1.8)
Hispanic	3 (2.1)	2 (1.4)	5 (1.8)
CFTR nonsense mutation category, n (%)
G542X	42 (30.4)	43 (31.6)	85 (31.0)
W1282X	33 (23.9)	34 (25.0)	67 (24.5)
R553X	21 (15.2)	17 (12.5)	38 (13.9)
R1162X	9 (6.5)	13 (9.6)	22 (8.0)
Other nonsense mutations	37 (26.8)	31 (22.8)	68 (24.8)
Mean ppFEV1 at Screening (SD)	63.5 (13.46)	62.2 (14.34)	62.9 (13.90)
Min, Max	40.0, 88.0	40.0, 90.0	40.0, 90.0
Baseline ppFEV1 Categories, n (%)
40 to <65	78 (55.7)	78 (56.1)	156 (55.9)
65 to 90	62 (44.3)	61 (43.9)	123 (44.1)
Mean Sweat Chloride, mEq/L (SD)	100.7 (2.87)	100.0 (2.99)	100.4 (2.94)
Min, Max, mEq/L	95, 120	93, 107	93, 120
Pancreatic Insufficiency, n (%)	115 (82.1)	124 (89.2)	239 (85.7)
P. aeruginosa Infection History, n (%)	113 (80.7)	110 (79.1)	223 (79.9)
P aeruginosa in Sputum at Screening, n (%)	37 (26.4)	29 (20.9)	66 (23.7)
History of Diabetes, n (%)	28 (20.0)	34 (24.5)	62 (22.2)
CF Medications used at Randomization, n (%)
Bronchodilators	122 (87.1)	119 (85.6)	241 (86.4)
Dornase Alfa	102 (72.9)	97 (69.8)	199 (71.3)
Chronic Azithromycin	76 (54.3)	71 (51.1)	147 (52.7)
Inhaled Hypertonic Saline	72 (51.4)	73 (52.5)	145 (52.0)
Inhaled Glucocorticoids	30 (21.4)	24 (17.3)	54 (19.4)
Chronic Inhaled Antibiotic Use at randomizationa, n (%)	86 (61.4)	66 (47.5)	152 (54.5)
Colistimethateb	60 (42.9)	49 (35.3)	109 (39.1)
Aztreonamb	33 (23.6)	22 (15.8)	55 (19.7)
Tobramycinb	0	0	0
Otherb	2 (1.4)	3 (2.2)	5 (1.8)

^aSubjects who received inhaled antibiotics (excluding aminoglycosides) up to 56 days prior to their randomization date.

^bSubjects who took multiple inhaled antibiotics prior to randomization are counted once for each inhaled antibiotic received.

3.1. Primary endpoint

At baseline, mean ppFEV₁ for the ITT population was 63.2 for the ataluren group and 61.6 for the placebo group. Both treatment groups experienced a mean loss in lung function during the 48-week study (Fig. 2). LS means for ppFEV₁ change from baseline to Week 40 and Week 48 averages (MMRM modeling of the ITT Population) were −1.396 for the ataluren group and −1.992 for the placebo group. The MMRM-based average treatment difference between ataluren and placebo groups in ppFEV₁ change from baseline to Weeks 40 and 48 was 0.597 (95% CI −1.3, 2.5; p = 0.53; Table 2.

Fig. 2.

Modeled Mean ppFEV₁ Change in from Baseline over 48 Weeks (ITT Population). Plotted values are observed means; bars represent ±SEM. Covariates were baseline ppFEV₁, treatment, visit, treatment-by-visit interaction, ppFEV₁-by-visit interaction, and the stratification factors of baseline inhaled antibiotics (yes vs no), baseline age (<18 vs ≥18 years), and baseline ppFEV₁ (40 to <65% vs ≥65 to 90%). Abbreviations: ppFEV₁ = percent predicted forced expiratory volume in 1 s, SEM = standard error of the mean.

Table 2.

Primary and secondary efficacy results.

Efficacy endpoint	Endpoint hierarchy	LS mean	Difference (ataluren minus placebo)	Std error	95% CI	p-Value
ppFEV1 change from baseline a	Primary
Ataluren		−1.396		0.6996	−2.7735, −0.0180
Placebo		−1.992		0.6777	−3.3271, −0.6576
Ataluren vs Placebo			0.597	0.957	−1.2881, 2.4813	0.5336
PEx rate, events/48 weeks	1st Secondary
Ataluren		0.950 (1.4038)b		0.1195	0.7136, 1.1862
Placebo		1.127 (2.5241)b		0.2164	0.6986, 1.5547
Ataluren/Placebo		0.8567c				0.4008
BMI change from baseline, kg/m2	2nd Secondary
Ataluren		0.296		0.093	0.1126, 0.4789
Placebo		0.361		0.0938	0.1759, 0.5455
Ataluren vs Placebo			−0.065	0.1312	−0.3233, 0.1934	0.6208
Respiratory HRQL change from baseline	3rd Secondary
Ataluren		−0.76		1.3691	−3.4566, 1.9364
Placebo		−1.032		1.3733	−3.7368, 1.6728
Ataluren vs Placebo			0.272	1.93	−3.5292, 4.0731	0.8881

ppFEV₁, forced expiratory volume in 1 s; PEx, pulmonary exacerbation; BMI, body mass index; HQRL, health-related quality of life.

^aMean of Week 40 and 48 minus baseline.

^bMean (SD).

^cRatio of ataluren to placebo rate.

3.2. Secondary endpoints

Mean (SD) protocol-defined pulmonary exacerbation rates per 48 weeks were 0.95 (1.404) and 1.13 (2.524) for ataluren and placebo subjects, respectively (Table 2). The 48-week ataluren/placebo exacerbation rate ratio of 0.86 was not statistically significant (p = 0.40 by negative binomial regression). Both treatment groups experienced modest BMI increases during the study (Table 2). At Week 48, the MMRM-modeled mean BMI change from baseline was 0.30 kg/m² for the ataluren group and 0.36 kg/m² for the placebo group, a difference (ataluren minus placebo) of −0.06 kg/m² (95% CI −0.32, 0.19; p = 0.62). Both treatment groups experienced a net decline in their respiratory quality of life during the study (Table 2). At Week 48, the MMRM-modeled mean change in the respiratory domain of the CFQ-R from baseline was −0.76 for the ataluren group and −1.03 for the placebo group, a difference (ataluren minus placebo) of 0.272 (95% CI −3.53, 4.07; p = 0.89).

3.3. Safety and study drug compliance

The as-treated population comprised 279 patients: 140 subjects treated at least once with ataluren and 139 subjects treated at least once with placebo. The mean [range] duration of study drug treatment was 45.39 weeks [0.3–53.3 weeks] among the 140 ataluren subjects and 44.68 weeks [1.1–54.9 weeks] among the 139 placebo subjects. Median compliance rate, based on study drug accountability, was 95.7% for ataluren and 94.9% for placebo.

Ataluren was generally well-tolerated, with the numbers of subjects experiencing treatment-emergent adverse events (TEAEs) similar between treatment groups (Table 3). However, more TEAEs were reported among subjects receiving ataluren (905) than among those receiving placebo (768). No clinically significant differences in vital sign measurements were observed between treatment groups and no clinically meaningful abnormalities were identified based on physical examinations, ECGs, and renal ultrasound. The most common TEAEs (≥15% in either treatment group) were infective pulmonary exacerbation of CF, cough, viral upper respiratory tract infection, and upper respiratory tract infection (Table 3). There was a somewhat higher incidence of gastrointestinal disorders (diarrhea, abdominal pain, nausea, and vomiting) in the ataluren group (31 subjects, 22.1%) than in the placebo group (22 subjects, 15.8%). Three subjects receiving ataluren (2.1%) and four subjects receiving placebo (2.9%) experienced TEAEs that led to discontinuation.

Table 3.

Summary of treatment-emergent adverse events occurring in ≥5% of subjects; as-treated population.

System organ class Preferred terma	Ataluren (N = 140) n (%)	Placebo (N = 139) n (%)	Total (N = 279) n (%)
Number of Subjects with at Least One Such Treatment Emergent Adverse Eventb	125 (89.3)	127 (91.4)	252 (90.3)
INFECTIONS AND INFESTATIONS	116 (82.9)	119 (85.6)	235 (84.2)
Infective pulmonary exacerbation of cystic fibrosis	84 (60.0)	92 (66.2)	176 (63.1)
Viral upper respiratory tract infection	27 (19.3)	21 (15.1)	48 (17.2)
Upper respiratory tract infection	21 (15.0)	21 (15.1)	42 (15.1)
Sinusitis	16 (11.4)	13 (9.4)	29 (10.4)
Nasopharyngitis	10 (7.1)	8 (5.8)	18 (6.5)
Rhinitis	10 (7.1)	8 (5.8)	18 (6.5)
Influenza	7 (5.0)	8 (5.8)	15 (5.4)
Pharyngitis	7 (5.0)	2 (1.4)	9 (3.2)
Pseudomonas infection	6 (4.3)	9 (6.5)	15 (5.4)
Staphylococcal infection	4 (2.9)	8 (5.8)	12 (4.3)
RESPIRATORY, THORACIC AND MEDIASTINAL DISORDERS	38 (27.1)	35 (25.2)	73 (26.2)
Cough	26 (18.6)	25 (18.0)	51 (18.3)
Haemoptysis	14 (10.0)	11 (7.9)	25 (9.0)
GASTROINTESTINAL DISORDERS	31 (22.1)	22 (15.8)	53 (19.0)
Diarrhoea	13 (9.3)	12 (8.6)	25 (9.0)
Abdominal pain	11 (7.9)	5 (3.6)	16 (5.7)
Nausea	11 (7.9)	10 (7.2)	21 (7.5)
Vomiting	8 (5.7)	4 (2.9)	12 (4.3)

Abbreviations: MedDRA = Medical Dictionary for Regulatory Activities; N = total number of patients; n = number of patients.

^aAdverse Events were coded using MedDRA, Version 17.0.

^bA treatment-emergent adverse event is defined as an adverse event that occurs or worsens in the period extending from the day of a subject’s first dose of study drug to 4 weeks after the last dose of study drug in this study. The total number of AEs counts all treatment-emergent AEs for subjects. At each level of subject summarization, a subject is counted once if the subject reported one or more events.

Across both treatment arms, 63.1% of subjects experienced Grade 2 (moderate) TEAEs, 17.2% experienced Grade 3 (severe) TEAEs, and 15.8% experienced Grade 1 (mild) TEAEs. No subjects experienced Grade 4 (life-threatening) TEAEs and there were no subject deaths during the study. Grade 3 infective pulmonary exacerbations of CF occurred in 5 subjects receiving ataluren (3.6%) and 22 subjects receiving placebo (15.8%). Two subjects receiving ataluren (1.4%) experienced Grade 3 nephrolithiasis, which was not observed among subjects receiving placebo. One ataluren subject experienced nephrolithiasis and acute renal failure, the latter associated with dehydration.

4. Discussion

A previous double-blinded, placebo-controlled phase 3 study of ataluren failed to meet its primary efficacy endpoint of change in FEV₁ over 48 weeks, but post hoc analyses suggested that study participants not receiving concomitant inhaled aminoglycosides experienced ataluren-associated FEV₁ and exacerbation benefits during the study [14]. The goal of the current study was to prospectively assess ataluren’s efficacy in CF subjects with at least one nonsense CFTR mutation not concomitantly receiving inhaled aminoglycosides, again using improvement in pulmonary function and reduction in pulmonary exacerbation rate as primary and secondary efficacy endpoints.

Neither ppFEV₁ change from baseline nor pulmonary exacerbation rate over 48 weeks were statistically different between ataluren and placebo groups in this trial, suggesting that results observed from the previous post hoc subgroup analyses occurred by chance. Both treatment groups experienced a net loss in ppFEV₁ over the course of the study, reflecting the severity of CF lung disease associated with CFTR nonsense mutations. However, subjects receiving ataluren had an increased TEAE incidence.

Further development of ataluren for treating persons with nonsense-mutation CFTR has been halted [23]. There may be several reasons why ataluren failed to be associated with clinical benefit in this study. In addition to the possibility of poor ataluren bioavailability in the lung, mRNA transcripts carrying nonsense mutations are subjected to nonsense mediated decay (NMD), a surveillance mechanism which detects and degrades transcripts carrying nonsense mutations, preventing the synthesis of truncated proteins [24] and protecting the cell from potentially deleterious truncated proteins [25]. This process may also regulate the response to read-though treatment with ataluren [26]. Inhibition of the NMD pathway has been associated with upregulated the CFTR RNA, protein, and surface-localized protein in cells homozygous for the W1282X mutation [27], and may be a viable approach to augment the efficacy of read-through [28]. Inefficient NMD may also activate the unfolded protein response (UPR), which restores cellular homeostasis by activating chaperons and foldases in the endoplasmic reticulum, inhibiting new substrate translation [29]. Because translation attenuation also inhibits NMD by a feedback mechanism, NMD and UPR may modulate response to ataluren read-through treatment. Observations of UPR differences in patients correlating with readthrough response appear to support this suggestion [30]. Finally, although successful readthrough with ataluren would allow translation elongation to continue in the correct reading frame, elongation is achieved by insertion of near-cognate tRNAs at the nonsense codon, which could result in a non-functional “neo-formed” protein [31] or a protein with reduced CFTR activity requiring a CFTR modulator for restored activity [32].

Although negative in outcome, the results of this study should be of value for the CF community, in that they demonstrated what we all should accept: that post hoc subgroup analyses, no matter how rational and well-justified, are hypothesis forming with results unverified in the absence of a prospective test. It has been suggested that ataluren’s initial discovery in cell-based firefly luciferase assays may have been an artifact resulting from direct interaction of ataluren with the reporter system, as use of a different luciferase reporter did not confirm ataluren’s activity with respect to nonsense codon suppression [33]. In retrospect, these observations in addition to the lack of sweat chloride changes associated with ataluren exposure in the previous Phase 3 study (including in the subgroup not receiving aminoglycosides) should have raised suspicions as to the probability of success of the current study. Similarly, developers of future stop codon read-through candidates might place greater weight on observing sweat chloride changes early in clinical development before proceeding to large pivotal trials.