Impact of trial design on the estimation of drug potency and power in clinical trials of haemophilia with inhibitors
Graphical abstract
Introduction
Haemophilia is an inherited X-linked bleeding disorder caused by deficiency of coagulation factor VIII (haemophilia A) or IX (haemophilia B) which is characterized by spontaneous bleeding events primarily in muscles and joints (Bolton-Maggs and Pasi, 2003). Current first-line treatment of patients with haemophilia relies on prophylactic replacement therapy with the corresponding factor products. However, some haemophilia patients develop anti-drug antibodies (inhibitors) to factor VIII or IX (30 and 5% of the patients, respectively (Gomez et al., 2014)) which partially or completely inhibit the effect of treatment, necessitating the use of bypassing agents. At present, there is approximately 3000 inhibitor patients globally (World Federation of Hemophilia 2015).
Historically, clinical trials of haemophilia with inhibitors (HwI) have been challenged by the limited size of the patient population (Dimichele et al., 2012), making it difficult to obtain statistically meaningful endpoints. These challenges may be overcome by utilization of high-powered trial designs requiring a low number of patients to identify a significant treatment effect. Several different trial designs have previously been applied in clinical trials of haemophilia, including parallel-group studies (Ljung et al., 2013), placebo-controlled parallel-group studies (ClinicalTrials.gov, 2015) and crossover studies (ClinicalTrials.gov, 2006). In a clinical trial investigating the efficacy of two prothrombin-complex concentrates in haemophilia patients with inhibitors, Lusher et al. (1980) reported a 25% placebo effect for the treatment of acute haemarthrosis. In spite of this, most clinical trials of haemophilia do not include a placebo group (Dimichele et al., 2012). The absence of a placebo group may lead to inaccurate estimates of the drug potency (EC50) as it may be difficult to distinguish the drug effect from the placebo effect.
The primary efficacy endpoint in clinical trials of haemophilia is the occurrence of spontaneous bleeding events (event-type data), summarised as annualized bleeding rate (ABR) for statistical evaluation. Traditional statistical methods for analysis of event-type data typically correlate the average number of events for the entire observation time (e.g. ABR) to the dose or average steady-state concentration. Such approaches have several shortcomings, and do not account for that i) exposure may change over time, ii) exposure may differ between individuals, iii) the frequency of events may change over time and iv) the dynamics of the drug effect, which may ultimately result in an inaccurate estimate of the EC50 and a loss of power (Plan et al., 2011). All of these concerns may be solved by utilization of a longitudinal model-based approach, e.g. repeated time-to-event (RTTE) modelling (Cox et al., 1999; Juul et al., 2015; Plan et al., 2011; Sargent, 1998). In RTTE analysis event-type data is modelled using a hazard function which can be transformed into the survival function. The hazard, being the instantaneous rate of bleeding events, may be affected by disease progression and/or drug treatment, thereby, altering the rate at which subjects develop spontaneous bleedings over time. Recently, RTTE modelling was successfully used to describe the relationship between the exposure of emicizumab and reduction in bleeding frequency (Yoneyama et al., 2017).
During clinical drug development approximately 85% of biologics fail (Hay et al., 2014). Therefore, clinical trial simulations are increasingly being used in the pharmaceutical industry to optimize trial designs in the various phases of drug development by increasing power, choosing appropriate dosing regimens and identifying design inefficiencies (Holford et al., 2010).
The aim of this study was to perform clinical trial simulations for a hypothetical drug to i) elucidate the influence of different trial designs and study conditions on the precision and accuracy of the estimated EC50 and power, and ii) compare different statistical methods (RTTE modelling, t-test and negative binomial regression) in terms of power.
Section snippets
Baseline analysis of bleeding events in haemophilia patients with inhibitors
Data on the occurrence of bleeding events in haemophilia patients with inhibitors treated on-demand with rFVIIa were obtained from the observation period of a study by Ljung et al. (2013). In short, during the observation period, 23 patients, aged 13 to 54 years, were treated on-demand for a period of up to 16.8 weeks. Time and date of the occurrence of bleeding events was recorded. During the observation period, a total of 181 bleeding events were observed with patients having between 0 and 22
Baseline repeated time-to-event model
The baseline hazard was best described by an exponential distribution with a hazard constant of 22.6 year−1 and IIV of 77.3% coefficient of variation. To evaluate the adequacy of the developed RTTE model to predict the observed KM curves, KM VPCs were generated for the 1st, 5th, 10th, 15th, 20th and 22nd event (Fig. 3). As apparent from Fig. 3, the RTTE model was able to provide an accurate description of the observed data.
Impact of trial designs and study conditions on the estimation of drug potency
The median and 95% confidence interval of the estimated EC50 for each
Discussion
In this work, we found that crossover designs in combination with RTTE modelling can markedly reduce the required sample size and study duration, while ensuring high power and precise estimation of EC50, in clinical trials of HwI.
In recent years, application of individualized dosing to improve patient care has been a topic of interest in haemophilia (Ar et al., 2014; Björkman, 2010; Preijers et al., 2017). In continuation of this, attempts have been made to establish the link between drug
Disclosures
All authors declare no competing financial interests.
References (38)
- et al.
Haemophilias A and B
Lancet
(2003) - et al.
Simulation for clinical research trials: a theoretical outline
J. Crit. Care
(2009) - et al.
The pharmacokinetics of a B-domain truncated recombinant factor VIII, turoctocog alfa (NovoEight®), in patients with hemophilia A
J. Thromb. Haemost.
(2015) - et al.
40K glycoPEGylated, recombinant FVIIa: 3-month, double-blind, randomized trial of safety, pharmacokinetics and preliminary efficacy in hemophilia patients with inhibitors
J. Thromb. Haemost.
(2013) - et al.
In silico evaluation of limited blood sampling strategies for individualized recombinant factor IX prophylaxis in hemophilia B patients
J. Thromb. Haemost.
(2017) - et al.
Randomized comparison of prophylaxis and on-demand regimens with FEIBA NF in the treatment of haemophilia A and B with inhibitors
Haemophilia
(2014) - et al.
Methods for individualising factor VIII dosing in prophylaxis
Eur. J. Haematol.
(2014) - et al.
NONMEM Users Guides (1989–2009)
(2009) Limited blood sampling for pharmacokinetic dose tailoring of FVIII in the prophylactic treatment of haemophilia A
Haemophilia
(2010)Extreme Value Theory
(1988)
Rapid musculoskeletal ultrasound for painful episodes in adult haemophilia patients
Haemophilia
Study Comparing On-demand Treatment With Two Prophylaxis Regimens of BeneFIX in Patients With Severe Hemophilia B
Trial Investigating Safety, Pharmacokinetics and Pharmacodynamics of Concizumab Administered Subcutaneously to Haemophilia A Subjects
A population pharmacokinetic–pharmacodynamic analysis of repeated measures time-to-event pharmacodynamic responses: the antiemetic effect of ondansetron
J. Pharmacokinet. Biopharm.
Clinical trial design in haemophilia
Haemophilia
Small Clinical Trials Issues and Challenges
Key issues in inhibitor management in patients with haemophilia
Blood Transfus.
On the nature of the function expressive of the law of human mortality, and on a new mode of determining the value of life contingencies
Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci.
Clinical development success rates for investigational drugs
Nat. Biotechnol.
Cited by (3)
Bayesian approach for a 2 x 2 crossover design with repeated measures: a simulation study
2024, Acta Scientiarum - TechnologyUse of Alternative Designs and Data Sources for Pediatric Trials
2020, Statistics in Biopharmaceutical Research