Study: Efficacy and Tolerability of a New Formulation of Pancrelipase Delayed-Release Capsules in Children Aged 7 to 11 Years with Exocrine Pancreatic Insufficiency and Cystic Fibrosis: a Multicenter, Randomized, Double-Blind, Placebo-Controlled, Two-Period Crossover, Superiority Study
Authors: Gavin R. Graff, MD; Karen Maguiness, MS, RD, CSP; John McNamara, MD; Ronald Morton, MD; David Boyd, PharmD; Katrin Beckmann, MSc; and Djenane Bennett, BSN
Group members: Neha Ohri, MD; Joshua Obuch, MD; Sharon Seifert, PharmD
1. Background and setting:
Cystic fibrosis (CF) is a hereditary condition that is transmitted in an autosomal recessive manner and affects individuals worldwide. In the US, approximately 14,000 children under the age of 18 and 27,000 adults over the age of 18 are affected with CF as of 2013 (1). CF is the most common life threatening genetic mutation among Caucasians in the US. The primary abnormality is a mutation in the CF transmembrane conductance regulator (CFTR) gene leading to dysfunction in the translated epithelial chloride channel. This mutation causes abnormal chloride (and other electrolyte) transport across membranes. The CFTR chloride channel plays a key role in the hydration of exocrine gland and epithelial surface secretions, helping to maintain mucous fluidity (2). The increased mucous viscosity is especially detrimental for pulmonary function, as the mucus clogs airways and traps bacteria leading to infections, extensive lung damage and potentially, respiratory failure. While impaired lung function remains the characteristic feature of CF, the CFTR gene is expressed in leukocytes, osteoblasts, islet cells, and proximal renal tubule cells, and is involved in immunoregulation, bone mass accrual, and tubular protein reabsorption (2). Importantly, exocrine pancreatic dysfunction, including exocrine pancreatic insufficiency (EPI), is seen in up to 85% of CF patients as well (3).
EPI is defined as the lack of digestive enzymes secreted by the pancreas due to disease affecting pancreatic secretion or from chronic pancreatitis (3). This leads to maldigestion, with resultant malabsorption of fat, protein, fat-soluble vitamins, and ultimately to malnutrition. Symptoms of EPI, including bloating, fatty/greasy stools, abdominal pain, and failure to gain weight, frequently begin in newborns and increase throughout infancy. In attempts to mitigate the resultant malnutrition and associated complications, approximately 90% of patients with EPI due to CF are treated with pancreatic enzyme replacement therapy (PERT) to maintain adequate nutrition (4). This can promote improved growth, development, and pulmonary health in children, as well as weight maintenance in adults.
Pancrelipase delayed-release capsules were used for several decades in patients with EPI without an explicit FDA indication. However, in 2004 the FDA notified manufacturers of pancreatic insufficiency products that these drugs now needed FDA approval in order to remain on the market. The FDA decided to require approval of new drug applications (NDAs) for all pancreatic extract drug products after reviewing data that showed substantial variations among currently marketed products (12). Following this mandate, several efficacy and safety trials of PERT in patients with EPI were conducted (5–11). A formulation of pancrelipase delayed-release capsule was approved by the FDA in 2009 for the treatment of EPI due to CF in patients ≥12 years following a randomized, double-blind, placebo controlled crossover study, which found that the new pancrelipase delayed-release 24,000-lipase unit capsule formulation improved fat digestion better than placebo (13).
In the aforementioned study, 32 patients were randomized (1:1) to placebo or pancrelipase. Patients were first provided their assigned intervention for 5 days, followed by a 3-14 day washout, then a crossover was performed and patients received the alternate intervention for a final 5 days. Patients were followed for safety analysis for 3-14 days after completion of the study. The results showed that the pancrelipase capsule was associated with significantly higher coefficient of fat absorption (CFA) and coefficient of nitrogen absorption (CNA) values compared to placebo (CFA: 88.6 vs 49.6, respectively [P < 0.001]; CNA: 85.1 vs 49.9 [P < 0.001]). Reported clinical symptoms were decreased, and the formulation was well tolerated although statistical comparison of symptoms was not performed.
Given the positive results of the above study, two follow-up studies of pancrealipase were performed in patients under the age of 12. In 2010, Graff, et.al, conducted one of these studies which involved children with EPI due to CF aged 7 to 11 years old, and compared the efficacy of a new formulation of pancrelipase delayed-release 12,000–lipase unit capsules to placebo using the coefficient of fat absorption as the primary endpoint. This multicenter, randomized, double-blind crossover superiority trial is the focus of the following sections.
2. Synopsis of study:
a) Inclusion and Exclusion Criteria
Inclusion Criteria:
Aged 7-11 years old
Confirmed CF by two different sweat or genetic tests
Confirmed malabsorption with CFA <70%
Taking a stable PERT regimen x 3 months
No major respiratory events x 1 month
Stable body weight
Able to swallow capsules
Able to be on a diet that would require medication.
Exclusion Criteria:
Significant “medical condition(s)” that could limit participation
Recently undergone major surgery (other than appendectomy)
BMI <10%
HIV infection
Crohn’s or cancer of the GI tract
Allergy to PERT
b) Study Treatment: Patients received either pancrelipase 12,000–lipase unit capsules or identical placebo capsules. The number of capsules to be consumed was calculated to provide a target dose of 4,000 lipase units/g of dietary fat intake, per CF consensus statements. Each patient received an individualized, prospectively designed diet containing ≥40% of calories derived from fat. With ≥40% of each patient’s total calorie intake derived from fat, no minimum daily dietary fat requirement was implemented (3).
c) Study Measurements: Evaluation of eligibility occurred at visit 1. Patients continued their usual PERT for up to 14 days while eligibility was confirmed. At visit 2, which took place on day 1 of the first crossover period (baseline), patients were randomized to 1 of 2 treatment sequences—pancrelipase followed by placebo, or placebo followed by pancrelipase. Each treatment was taken for 5 days. Visit 3 took place at the end of the first crossover period (day 6 or 7) and included a physical examination, measurement of vital signs, and safety assessments. This visit was followed by a washout period of 3 to 14 days, during which patients took their previous PERT. Patients entered the second crossover period at visit 4. The procedures and timing of this period were identical to those of the first crossover period. Visit 5 marked the end of the second period and included a physical examination, measurement of vital signs, and safety assessments. A safety follow-up call was made 5 to 7 days after visit 5 (Table 1). During both crossover periods, patients stayed in a dedicated research facility and were allowed to return home during washout periods (3).
The primary efficacy outcome, CFA, was assessed by analyzing stool samples. For accuracy of stool collection, patients were administered two 250-mg doses of blue food dye 72 hours apart, marking the beginning and end of each stool collection period (days 2 and 5 of each crossover period). Stool collections were performed during each crossover period, beginning after the appearance of the first marker and ending with the stool containing the second marker. Dietary recording took place during each treatment period and total daily fat intake was determined from each patient’s dietary intake during both 72-hour stool-collection periods. The CFA was calculated based on fat intake and excretion using the following equation: CFA (%) = 100 ([grams fat intake – grams fat excretion]/grams fat intake). Stool fat was measured in the stool samples collected between the 2 dye markers in both crossover periods. Stool fat was determined by the gravimetric method (3).
Table 1. Study visit procedures
Visit 1 / Visit 2 / 5 days of active or placebo treatment / Visit 3 / 5 days of active or placebo treatment / Visit 4 / Visit 5Evaluation of eligibility / X
Patients randomized to sequence (activeplacebo or placeboactive) / X
Patients start taking either active or placebo treatment for 5 days / X / X
Measurement of CFA% from collected stool / X / X
Physical examination, measurement of vital signs, and safety assessments / X / X
d) Statistical Design: This study had a crossover design. There are certain advantages and disadvantages to using this kind of study design, as illustrated in Table 2.
Table 2. Crossover study design considerations
Primary outcome: The primary outcome was measured using the coefficient of fat absorption (CFA%). In a normal, healthy adult the CFA should be 100% as all fat should be absorbed by the small intestine with the aid of pancreatic enzymes and bile salts.
Secondary outcomes: Secondary outcomes included the CNA%, calculated using the following equation: CNA (%) = 100 ([grams nitrogen intake – grams nitrogen excretion]/ grams nitrogen intake). Nitrogen intake was calculated based on the protein intake recorded in patient’s dietary diaries during the two 72-hour stool-collection periods. Clinical symptoms were assessed, as were tolerability measures (vital signs, physical examinations, laboratory safety tests, and adverse events).
Probability model: Continuous, as CFA% is measured from 0-100% absorption
Functional: The difference in CFA% in the treatment group (ϴ1) or placebo (ϴ0)
Contrast: Treatment (risk) difference, ϴ1- ϴ0 = ϴ
Hypothesis: The alternative hypothesis was that active treatment was superior to placebo. The null hypothesis was that there was no difference between the 2 groups. In the notation below, ϴϕ= 0 and ϴ+ = effect size of 1.
Sample size: The study authors calculated a sample size of 16 using 95% power, effect size of 1.0, and two-tailed alpha of 0.05. They planned to enroll 18 subjects to allow for drop outs. The sample size calculation carried out by the authors used t-scores rather than z-scores, as the authors anticipated that they would be enrolling a small number of patients. When the sample size calculation is carried out using z-scores, as in the following equation, , and substituting an effect size of 1.0 for the variance, , the sample size is 13 rather than 16. Due to the crossover design of the study, N signifies the total number of subjects (i.e. each group will have N/2 subjects), as opposed to a 2-group comparison study where N denotes the number of subjects per group. Effect size has no scientific interpretation; you should discuss potential inference upon trial completion given a sample size of 16 (see my lecture notes). This requies an approximate standard deviation (from table 2 in the Graff paper which gives a standard error of 3.8, so sd = 3.8*4 = 15.2).
Analysis plan: The primary outcome was analyzed using ANOVA. An ANOVA model was used because it is capable of handling the carryover effect that may occur in crossover studies. This model provided an estimate of the treatment difference, along with a 95% confidence interval and P value for testing the null hypothesis. All clinical symptom assessments and tolerability variables were summarized using standard descriptive methods.
3. Study Implementation and Conduct:
a) Randomization procedures: The randomization scheme was generated by the Global Clinical Supplies Office of Solvay Pharmaceuticals B.V., Weesp, The Netherlands. Patients were assigned to a treatment sequence via a centralized electronic interactive voice response system (IVRS). If an emergency unblinding was required as a result of an adverse event (AE), the IVRS was to be used. The investigators were required to assess the potential relationship between the AE and study treatment prior to breaking the treatment. A 24-hour telephone number was available in case there was a need for emergency unblinding of a treatment code and the investigator had no access to the unblinding procedures (3).
b) Procedures for blinding: Each study site was supplied with blinded, packaged study medication. The appearance, shape, smell, and taste of the active-treatment and placebo capsules were identical and were packaged to ensure correct dosing and maintenance of blinding (3).
References:
1. Accessed 11.3.15
2. Haller W, Ledder O, Lewindon PJ, Couper R, Gaskin KJ, Oliver M. Cystic fibrosis: An update for clinicians. Part 1: Nutrition and gastrointestinal complications. J Gastroenterol Hepatol. 2014;29(7):1344-55
3. Graff GR, Maguiness K, et al. Efficacy and tolerability of a new formulation of pancrelipase delayed-release capsules in children aged 7 to 11 years with exocrine pancreatic insufficiency and cystic fibrosis: a multicenter, randomized, double-blind, placebo-controlled, two-period crossover, superiority study. Clin Ther. 2010 Jan;32(1):89-103.
4. FitzSimmons SC. The changing epidemiology of cystic fibrosis. J Pediatr. 1993;122:1–9.
5. Cystic Fibrosis Mutation Database. Cited 28 Feb 2014. Available from URL:
6. Boyle MP, De Boeck K. A new era in the treatment of cystic fibrosis: correction of the underlying CFTR defect. Lancet Respir.Med. 2013; 1: 158–63.
7. Cystic Fibrosis Australia. Cystic Fibrosis in Australia 2012: 15th Annual Report from the Australian Cystic Fibrosis Data Registry. 2013. Cited 28 Feb 2014. Available from URL: .pdf
8. Knowles MR, Drumm M. The influence of genetics on cystic fibrosis phenotypes. Cold Spring Harb. Perspect. Med. 2012; 2: a009548.
9. Accurso FJ, Rowe SM, Clancy JP et al. Effect of VX-770 in persons with cystic fibrosis and the G551D-CFTR mutation. N. Engl. J. Med. 2010; 363: 1991–2003.
10. Ong T, Ramsey BW. Modifying disease in cystic fibrosis: current and future therapies on the horizon. Curr. Opin. Pulm. Med. 2013;19: 645–51.
11. De Lisle RC, Borowitz D. The cystic fibrosis intestine. Cold Spring Harb. Perspect. Med. 2013; 3: a009753.
12. US Food and Drug Administration. FDA requires pancreatic extract manufacturers to submit marketing applications. http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/2004/ucm108289.htm
13. Trapnell BC, Maguiness K, Graff GR, et al. Efficacy and safety of Creon 24,000 in subjects with exocrine pancreatic insufficiency due to cystic fibrosis. J Cyst Fibros. 2009;8: 370–377