In healthy individuals, the cystic fibrosis transmembrane conductance regulator (CFTR) maintains an alkaline environment in the pancreatic duct, which neutralizes acidic chyme and delivers proenzymes to the duodenal lumen (1–3). In persons with CF, CFTR malfunction results in increased pancreas juice viscosity and acidity leading to ductal obstruction, chronic inflammation, and progressive acinar cell destruction (1,4,5). Ultimately, as acinar cells are damaged, the pancreas parenchyma becomes fibrotic, resulting in pancreas exocrine and endocrine insufficiency (1,4,5). This process may occur in utero among individuals with severe CFTR gene mutations, but individuals with class V or mild class IV mutations retain apical CFTR activity and may have recurrent acute pancreatitis (AP) until eventually developing chronic pancreatitis over their lifetime (6–8). Accordingly, this subgroup of patients with pancreas-sufficient cystic fibrosis (PS-CF) having recurrent AP may be appropriate for therapy to decrease episodes of AP to reduce or delay the progression to exocrine pancreatic insufficiency (i.e., pancreas-insufficient cystic fibrosis [PI-CF]) and/or chronic pancreatitis.
Treatment of CF has evolved to personalized CFTR modulator therapy targeting multiple protein activity defects including synthesis, transport, trafficking, and function of CFTR and continues to progress (1,9–12). Among patients with PS-CF and recurrent AP, use of CFTR modulators leads to a reduction in the rate of subsequent AP (13–15). In addition, improved weight and pancreas exocrine function have been described (16–18). Among patients with PI-CF, AP is thought to be rare due to insufficient enzyme production to allow autoactivation within the pancreatic ducts. Therefore, AP in patients with PI-CF is hypothesized to occur at early stages of pancreatic insufficiency, when steatorrhea has occurred, but fecal elastase is still detectable. For example, 1 study reported a mean fecal elastase of 97 mcg/g (normal > 200 mcg/g) among patients with PI-CF who developed AP (19). CFTR modulators may affect pancreatic exocrine function in patients with PI-CF. One study showed an increase of fecal elastase from 83 to 465 mcg/g with CFTR modulator therapy (20). Thus, in patients with low fecal elastase, treatment with CFTR modulators may increase pancreas enzyme production. This could potentially explain the increased rate of AP among patients with PI-CF treated with CFTR modulators (20–24).
No large studies have investigated the incidence of AP among patients treated with CFTR modulators over time. Therefore, we sought to use a large database to assess the rate of AP among patients with CF treated with CFTR modulators. We hypothesized that patients with PS-CF would experience a reduction in AP hospitalizations while on CFTR modulator therapy. We also sought to study the effect of CFTR modulators on patients with PI-CF, hypothesizing that the impact of these medications would be less significant.
MATERIALS AND METHODS
A retrospective analysis was performed using the MarketScan database from 2012 to 2018, which includes an opportunity sampling of deidentified patients in the United States. MarketScan includes more than 32 billion service records of over 200 million individuals, providing inpatient and outpatient health care and medication costs and also compliance measures (25). MarketScan also reports time-sensitive data that allow recruitment to a study cohort subsequent to an intervention, thus providing a robust temporal association compared with other claims databases. The Ohio State University Wexner Medical Center Institutional Review Board deemed this study exempt.
All patients with CF were considered eligible for inclusion, as determined by International Classification of Diseases (ICD) version 9 and 10 codes (see Supplementary Table 1, https://links.lww.com/AJG/C245). To support ongoing use of CFTR modulators, patients were required to have a minimum of 1-year pharmaceutical coverage in the data set before the date of enrollment. Patients with less than 3 months of follow-up were excluded. Other exclusion criteria included history of pancreas transplant or resection (Figure 1).
Outcomes of interest
The primary outcome of interest were hospitalizations for AP. A secondary outcome was to evaluate medication use characteristics.
Definition of variables
Study definitions and corresponding ICD codes are presented in Supplementary Table 1, https://links.lww.com/AJG/C245. The diagnosis of CF was based on 1 inpatient diagnosis code or on 2 outpatient codes, as has been performed previously to improve diagnostic accuracy (26,27). Pancreas insufficiency was defined by prescription of any pancreas enzyme replacement therapy during the preceding 12 months. Patients who did not receive pancreas enzyme replacement therapy were deemed to be pancreas sufficient. Patients were considered to be on CFTR modulator therapy when it was prescribed and filled, and patients were considered to be off of CFTR modulator therapy when no medications were prescribed or filled for 60 days. After the 60-day period without CFTR modulators, patients were allowed to crossover to the no CFTR modulator cohort. Thus, the time from the last CFTR modulator dose to 60 days without CFTR modulators was considered a washout period, and these data were not included in the analysis. Varying time periods (10, 30, 60, and 90 days) were used for this washout period, which all produced similar findings (Supplementary Table 2, https://links.lww.com/AJG/C245). Other variables included age, sex, type of insurance, region of residence, and presence of comorbidities defined by the Charlson Comorbidity Index (CCI) (28,29). Readmissions for AP within 30 days were considered to be part of the index AP and were therefore considered to be the same episode, rather than counted as a different AP event.
Demographics and medication use characteristics were displayed with descriptive statistics such as means with SDs or medians with interquartile ranges and ranges for continuous variables and frequencies and percentages for categorical variables. For continuous variables that followed normality assumption (such as age), the 2-sample t test was used for the bivariate relationships. For continuous variables that displayed distribution skewness, the Wilcoxon rank-sum test was used. The χ2 test (or the Fisher exact test when appropriate) was used for the bivariate relationships among categorical variables.
Patient follow-up time was divided into time on and off CFTR modulators. AP admissions were totaled separately for periods of time when on and off CFTR modulators. Generalized linear models with a log-link and Poisson distribution were used to compare AP rates by CFTR modulator status. The outcome was number of distinct AP events with an offset for the log of the patient-years, so that the outcome was a rate, and rate ratios were estimated between groups. The correlation between data from time on and off of CFTR modulators within the same patient was accounted for in the models. Separate models were fit to patients with PI-CF and PS-CF, including unadjusted models followed by multivariable models. The multivariable model for all patients adjusted for age (as a continuous variable), sex, CCI, and geographic region, whereas the multivariable model for patients with previous CFTR modulator use only adjusted for age (due to the smaller sample size). Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC). Significance was defined as a 2-sided alpha <0.05.
A total of 10,417 patients were included, of which 1,795 (17.2%) had some previous CFTR modulator use (Table 1). Patients with CFTR modulator use were younger and more likely to be male compared with those without CFTR modulator use. Insurance plan types and region of residence were different between groups (Table 1). Patients with CFTR modulator use had a lower CCI (0.4 vs 0.7, P < 0.0001). A higher percentage of patients with CFTR modulator use were pancreas insufficient (84.5% vs 59.6%, P < 0.0001) compared with those without CFTR modulator use. Among patients with CFTR modulator use, the most commonly used agent was lumacaftor/ivacaftor, followed by ivacaftor alone. Treated patients received CFTR modulators 51% of the time within the study period (SD 34%).
Among the 1,795 patients with some CFTR modulator use during the study period, 279 (15.5%) were PS-CF (Table 2). Patients with PS-CF were older and more likely to be female compared with PI-CF. Insurance plan types were different between groups, but region of residence was no different (Table 2). Patients with PS-CF had a higher CCI (0.6 vs 0.3, P < 0.0001) compared with PI-CF. Patients with PS-CF spent on average less percentage of the study period on CFTR modulators than did patients with PI-CF (42% vs 53%, P < 0.0001). The most commonly used agent among patients with PS-CF was ivacaftor (61%), whereas lumacaftor/ivacaftor was used most often in PI-CF (51%) (Table 2).
Incidence of acute pancreatitis
During the study period, there were a total of 240 AP hospitalizations in the whole study population, 31 of which occurred among patients with some previous CFTR modulator use. In the full study cohort of 10,417 patients 145 AP admissions occurred among 3,759 patients with PS-CF, of which 2 (1.4%) occurred during CFTR modulator use (Table 3). Also, among 10,417 patients there were 95 AP admissions among 6,658 patients with PI-CF, of which 5 (5.3%) occurred during CFTR modulator use (Table 3). A multivariable model adjusted for age, sex, CCI, and geographic region found that the rate ratio for AP among patients with PS-CF was 0.32 (95% confidence interval [CI] 0.10, 0.99, P = 0.05) and that the rate ratio for PI-CF was 0.37 (95% CI 0.15, 0.92, P = 0.03) (Table 3).
Among 1,795 patients with some CFTR modulator use during the study period, 13 AP admissions occurred among 279 patients with PS-CF, of which 2 (15.4%) occurred during CFTR modulator use (Table 4). Also, among 1,795 patients with some CFTR modulator use, there were 18 AP admissions among 1,516 patients with PI-CF, of which 5 (27.8%) occurred during CFTR modulator use (Table 4). A multivariable model adjusting for age found that the rate ratio of AP for patients with PS-CF was 0.36 (95% CI 0.13, 1.01, P = 0.05) and that the rate ratio for PI-CF was 0.51 (95% CI 0.20, 1.30, P = 0.16) (Table 4).
Prediction of acute pancreatitis
The multivariable models were then used to predict the rate of AP hospitalizations per 1,000 patient-years of follow-up based on CFTR modulator use and pancreas sufficiency (Tables 5 and 6). In the full cohort, untreated patients with PI-CF had an estimated 1.76 (95% CI 0.98, 3.17) AP hospitalizations per 1,000 patient-years, and untreated patients with PS-CF had an estimated 10.20 (95% CI 6.19, 16.81) AP hospitalizations per 1,000 patient-years (Table 5).
The highest and lowest rates were seen in the model restricted to patients with some previous CFTR modulator exposure of 1,795 patients AP rates ranged from 12.93 (95% CI 3.47, 48.14) admissions per 1,000 patient-years among patients with PS-CF who were no longer using CFTR modulators to 0.34 (95% CI 0.08, 1.43) admissions per 1,000 patient-years among patients with PI-CF during CFTR modulator use (Table 5).
Using a large administrative database study, we found that the use of CFTR modulators in patients with CF was associated with a significant reduction in the incidence of hospitalizations for AP. Although the relative reduction in AP hospitalizations was greater among patients with PS-CF, the effect was also significant in those with PI-CF. In a subset analysis of patients who had some CFTR modulator use, the relative reduction in AP hospitalizations remained similar for PS-CF. These findings support further evaluation of the use of CFTR modulator therapies to reduce AP admissions in patients with CF.
The CFTR ion channel is primarily expressed on the apical plasma membrane. Through Cl- and HCO3- secretion, it regulates the amount and volume of pancreatic ductal secretions (1–3). In addition, CFTR participates in complex epithelial cell signaling and mitochondrial function pathways, and possibly also acinar cell function (30–32). CFTR passes through many checkpoints before delivery to the apical plasma membrane, and CFTR mutations are categorized broadly into 6 classes based on the effect of the mutation (33,34). Cystic fibrosis (CF) severity is generally determined by mutation class. However, disease modifiers such as bacterial colonization and toxin exposure may influence the course of the disease among patients with some residual CFTR function (5,35,36).
Although AP is more common in patients with PS-CF compared with patients with PI-CF, we identified that 58 (0.9%) patients with PI-CF had at least 1 episode of AP. This is an important finding and suggests that AP among patients with PI-CF may not be as rare as previously described in the literature. AP in patients with PI-CF treated with CFTR modulators is hypothesized to occur due to an increase in pancreas enzyme production without improvement in pancreatic ductal fluid, tipping the balance toward ductal obstruction, premature activation of trypsinogen, and AP (24). Our study demonstrated that patients with PI-CF exhibited a reduced risk of developing AP while on CFTR modulators. One possible hypothesis for this is that CFTR modulators may result in increased bicarbonate secretion, which would reduce the risk of premature autoactivation of trypsinogen within the pancreatic duct (37).
The previously reported incidence of AP among untreated patients with PS-CF of between 1% and 2% was confirmed in this study (7,19). It is notable that the PS-CF cohort exhibited a decreased risk of AP while on CFTR modulator therapy, as this could ultimately lead to prolonged exocrine function, improved nutritional status, and reduced risk of developing, or prolonging the progression to, chronic pancreatitis and exocrine pancreatic insufficiency (11,16–18). CFTR modulator therapy is expensive, with 1 study suggesting that a 40% decline in medication cost would be needed for the therapy to become cost-effective, assuming each quality-adjusted life year to be valued at $500,000 (38). However, the cost of CFTR modulators may be more justifiable if they are proven to reduce hospitalizations for AP, reduce the need for pancreas enzyme replacement therapy, and alter the natural history of CF. Further cost-benefit analyses are warranted and should take into account the degree of baseline exocrine pancreatic function.
A reduction in AP among patients with PS-CF has been previously reported. In the largest single-center study to date, there were 13 episodes of AP among 15 patients during the 2 years leading up to CFTR modulator initiation, but only 5 episodes of AP in available follow-up (45.3 patient-years) after initiating CFTR modulators (13). The rate ratio in the above study was 0.25, which is lower than our study, where we observed a rate ratio of 0.33 in the complete cohort. The previous study included only patients with PS-CF with previous AP, suggesting greater benefit from CFTR modulators compared with patients with PS-CF without previous AP. In our study, patients with PS-CF received CFTR modulators for shorter periods compared to patients with PI-CF, yet AP occurred much less frequently in patients with PS-CF while on modulator therapy. This suggests a major protective effect of the therapy when acinar cell function is still preserved.
Our study has several limitations inherent to the use of an administrative database, including reliance on ICD coding and possible bias due to insurance type. Although ICD coding validation studies in CF are lacking, CF is a very specific diagnosis with major clinical implications, so the likelihood of accurate coding for identification of CF is high. The diagnosis of pancreas insufficiency relied on prescription of pancreas enzyme therapy rather than laboratory data, but this definition is commonly used in practice and research because assessing pancreas function is not uniform (39). This assumption carries the risk of misclassifying patients with PS-CF as PI-CF if pancreas enzyme treatment was prescribed during the study period. However, approximately 2/3 of the study population was categorized as PI-CF, which suggests that if any miscategorization was present, it erred on the side of an overly specific allocation to PI-CF. The low incidence of AP in the cohort and inability to determine AP hospitalizations previous to the study period could result in a type II error. Last, genetic mutations were not included in the data set, so a secondary analysis of AP incidence according to mutation classifications was not possible.
Despite these potential limitations, our study has many strengths. The use of a large database that allows extraction of more granular data allowed for a robust characterization of AP in a large cohort of patients with CF. We were able to study the effect of CFTR modulators on the occurrence of AP in patients with CF with or without pancreas insufficiency in a time-sensitive manner, which approximates a cause-and-effect association, especially because we could compare the effect of CFTR modulators by assessing the risk of AP in treated and untreated patients with CF. In addition, the use of a time-varying covariate analysis allowed for considerable patient-years of follow-up, which highlighted the effect of CFTR modulator therapy on the risk of AP. This approach also allowed patients to serve as their own controls simulating a crossover study, partially controlling for genotype and propensity to develop AP, which would be otherwise challenging to perform outside of a large database or longitudinal patient registry. Even the subgroup analysis of patients with previous CFTR modulator use included 6,248 patient-years of follow-up, which is multiple times greater than any previously published study evaluating the incidence of AP in this population.
In this large database analysis, we demonstrated that the use of CFTR modulators is associated with a significant reduction in the rate of subsequent AP hospitalizations among patients with CF. Patients with CF could benefit from therapy with CFTR modulators to reduce the risk of AP, and this includes patients who already have pancreas insufficiency. A prospective study of CFTR modulator therapy among persons with CF to determine pancreas outcomes including AP and exocrine and endocrine function as well as a cost analysis should be performed.
CONFLICTS OF INTEREST
Guarantor of the article: Luis F. Lara, MD.
Specific author contributions: M.L.R.: conception of the work, initial draft of the manuscript, final approval, and agreement to be accountable. Y.G.: data analysis, critical review of the manuscript, final approval, and agreement to be accountable. L.A.S., M.R.W., S.E.K., S.S.L., G.I.P., S.G.K., P.P.S., P.A.H., D.L.C., and L.F.L: conception of the work, critical review of the manuscript, final approval, and agreement to be accountable. K.P.: conception of the work, data analysis, critical review of the manuscript, final approval, and agreement to be accountable.
Financial support: The project described was supported by Award Number UL1TR002733 from the National Center for Advancing Translational Sciences. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Center for Advancing Translational Sciences or the National Institutes of Health.
Potential competing interests: None to report.
WHAT IS KNOWN
- ✓ Individuals with cystic fibrosis (CF) may experience acute pancreatitis.
- ✓ Cystic fibrosis transmembrane conductance regulator modulators affect the risk of pancreatitis in individuals with CF.
WHAT IS NEW HERE
- ✓ Acute pancreatitis occurs in pancreas-sufficient and -insufficient patients with CF.
- ✓ Cystic fibrosis transmembrane conductance regulator modulator use is associated with a reduction in pancreatitis hospitalizations among patients with CF.
1. Hegyi P, Wilschanski M, Muallem S, et al. CFTR: A new horizon in the pathomechanism and treatment of pancreatitis. Rev Physiol Biochem Pharmacol 2016,170:37–66.
2. Madácsy T, Pallagi P, Maleth J. Cystic fibrosis of the pancreas: The role of CFTR channel in the regulation of intracellular Ca(2+) signaling and mitochondrial function in the exocrine pancreas. Front Physiol 2018,9:1585.
3. Pallagi P, Hegyi P, Rakonczay Z Jr. The physiology and pathophysiology of pancreatic ductal secretion: The background for clinicians. Pancreas 2015,44:1211–33.
4. Shwachman H, Lebenthal E, Khaw KT. Recurrent acute pancreatitis in patients with cystic fibrosis with normal pancreatic enzymes. Pediatrics 1975,55:86–95.
5. Ooi CY, Durie PR. Cystic fibrosis transmembrane conductance regulator (CFTR) gene mutations in pancreatitis. J Cyst Fibros 2012,11:355–62.
6. Frulloni L, Castellani C, Bovo P, et al. Natural history of pancreatitis associated with cystic fibrosis gene mutations. Dig Liver Dis 2003,35:179–85.
7. Gooding I, Bradley E, Puleston J, et al. Symptomatic pancreatitis in patients with cystic fibrosis. Am J Gastroenterol 2009,104:1519–23.
8. Baldwin C, Zerofsky M, Sathe M, et al. Acute recurrent and chronic pancreatitis as initial manifestations of cystic fibrosis and cystic fibrosis transmembrane conductance regulator-related disorders. Pancreas 2019,48:888–93.
9. Wilschanski M, Novak I. The cystic fibrosis of exocrine pancreas. Cold Spring Harb Perspect Med 2013,3:a009746.
10. Chaudary N. Triplet CFTR modulators: Future prospects for treatment of cystic fibrosis. Ther Clin Risk Manag 2018,14:2375–83.
11. Lommatzsch ST, Taylor-Cousar JL. The combination of tezacaftor and ivacaftor in the treatment of patients with cystic fibrosis: Clinical evidence and future prospects in cystic fibrosis therapy. Ther Adv Respir Dis 2019,13:1753466619844424.
12. Clancy JP, Jain M. Personalized medicine in cystic fibrosis: Dawning of a new era. Am J Respir Crit Care Med 2012,186:593–7.
13. Akshintala VS, Kamal A, Faghih M, et al. Cystic fibrosis transmembrane conductance regulator modulators reduce the risk of recurrent acute pancreatitis among adult patients with pancreas sufficient cystic fibrosis. Pancreatology 2019,19:1023–6.
14. Carrion A, Borowitz DS, Freedman SD, et al. Reduction of recurrence risk of pancreatitis in cystic fibrosis with ivacaftor: Case series. J Pediatr Gastroenterol Nutr 2018,66:451–4.
15. Cimbalo C, De Gregorio F, Castaldo A, et al. Effect of ivacaftor on comorbidities in a patient with cystic fibrosis and pancreatic sufficiency. Ital J Pediatr 2019,45(Suppl 1):8
16. Taylor-Cousar JL, Munck A, McKone EF, et al. Tezacaftor-Ivacaftor in patients with cystic fibrosis homozygous for Phe508del. N Engl J Med 2017,377:2013–23.
17. Rowe SM, Daines C, Ringshausen FC, et al. Tezacaftor-ivacaftor in residual-function heterozygotes with cystic fibrosis. N Engl J Med 2017,377:2024–35.
18. Graeber SY, Dopfer C, Naehrlich L, et al. Effects of lumacaftor-ivacaftor therapy on cystic fibrosis transmembrane conductance regulator function in Phe508del homozygous patients with cystic fibrosis. Am J Respir Crit Care Med 2018,197:1433–42.
19. De Boeck K, Weren M, Proesmans M, et al. Pancreatitis among patients with cystic fibrosis: Correlation with pancreatic status and genotype. Pediatrics 2005,115:e463-9.
20. Hamilton JL, Zobell JT, Robson J. Pancreatic insufficiency converted to pancreatic sufficiency with ivacaftor. Pediatr Pulmonol 2019,54:1654.
21. Petrocheilou A, Kaditis AG, Loukou I. Pancreatitis in A patient with cystic fibrosis taking ivacaftor. Children (Basel) 2020,7:E1–e3.
22. Johns JD, Rowe SM. The effect of CFTR modulators on a cystic fibrosis patient presenting with recurrent pancreatitis in the absence of respiratory symptoms: A case report. BMC Gastroenterol 2019,19:123–e4.
23. Kounis I, Lévy P, Rebours V. Ivacaftor CFTR potentiator therapy is efficient for pancreatic manifestations in cystic fibrosis. Am J Gastroenterol 2018,113:1058–9.
24. Megalaa R, Gopalareddy V, Champion E, et al. Time for a gut check: Pancreatic sufficiency resulting from CFTR modulator use. Pediatr Pulmonol 2019,54:E16–E8.
25. IBM Watson Health. IBM MarketScan Research Databases for Health Services Researchers. Somers: NY, 2019. Accessed August 2020.
26. Grosse SD, Do TQN, Vu M, et al. Healthcare expenditures for privately insured US patients with cystic fibrosis, 2010-2016. Pediatr Pulmonol 2018,53:1611–8.
27. Miller AC, Comellas AP, Hornick DB, et al. Cystic fibrosis carriers are at increased risk for a wide range of cystic fibrosis-related conditions. Proc Natl Acad Sci USA 2020,117:1621–7.
28. Quan H, Sundararajan V, Halfon P, et al. Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. Med Care 2005,43:1130–9.
29. Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: Development and validation. J Chronic Dis 1987,40:373–83.
30. Freedman SD, Kern HF, Scheele GA. Pancreatic acinar cell dysfunction in CFTR(-/-) mice is associated with impairments in luminal pH and endocytosis. Gastroenterology 2001,121:950–7.
31. Kerem B, Rommens JM, Buchanan JA, et al. Identification of the cystic fibrosis gene: Genetic analysis. Science 1989,245:1073–80.
32. Riordan JR, Rommens JM, Kerem B, et al. Identification of the cystic fibrosis gene: Cloning and characterization of complementary DNA. Science 1989,245:1066–73.
33. Gibson-Corley KN, Meyerholz DK, Engelhardt JF. Pancreatic pathophysiology in cystic fibrosis. J Pathol 2016,238:311–20.
34. Kristidis P, Bozon D, Corey M, et al. Genetic determination of exocrine pancreatic function in cystic fibrosis. Am J Hum Genet 1992,50:1178–84.
35. Maléth J, Balázs A, Pallagi P, et al. Alcohol disrupts levels and function of the cystic fibrosis transmembrane conductance regulator to promote development of pancreatitis. Gastroenterology 2015,148:427–39.e16.
36. Kadiyala V, Lee LS, Banks PA, et al. Cigarette smoking impairs pancreatic duct cell bicarbonate secretion. Jop 2013,14:31–8.
37. Gelfond D, Heltshe S, Ma C, et al. Impact of CFTR modulation on intestinal pH, motility, and clinical outcomes in patients with cystic fibrosis and the G551D mutation. Clin Transl Gastroenterol 2017,8:e81.
38. Balk EM, Trikalinos TA, Mickle K, et al. Modulator Treatments for Cystic Fibrosis: Effectiveness and Value. Final Evidence Report and Meeting Summary. Institute for Clinical and Economic Review, 2018, Public Meeting.
39. Hart PA, Conwell DL. Challenges and updates in the management of exocrine pancreatic insufficiency. Pancreas 2016,45:1–4.