We performed a nested case–control analysis using the global health data platform TriNetX, a deidentified web-based tool for research population cohort and feasibility queries that provides access to EMR-based clinical data from 67 healthcare organizations (HCOs) worldwide. The TriNetX database provides real-time access to longitudinal EMR information, derived from both inpatient and outpatient encounters. TriNetX performs recurring refreshers and spot checks of the data on a monthly basis, with varying data from HCOs.

We included all patients between the ages of 0 and 21 years (inclusive) between 1 January 2010 and 1 January 2020. Those with a history of solid-organ transplants were excluded due to the greater complexity of these patients and a greater likelihood of biasing results. Out of this cohort, cases were defined as patients with an ICD-9 or ICD-10 diagnosis code for CF (Supplement). Controls were defined as those patients without a diagnosis of CF.

Comparisons between the CF and non-CF cohorts were done using a chi-square test or Fisher’s exact test where appropriate. Bonferroni correction was used due to multiple comparisons, and a p-value of <0.0001 was considered significant. Demographic and clinical characteristics were summarized using descriptive statistics. Means with standard deviations or medians with interquartile ranges were used for continuous variables. Counts and percentages were used for categorical variables. GI manifestations in pediatric PwCF were categorized as intestinal, hepatobiliary, or pancreatic disease. Analyses were conducted using the TriNetX platform and SAS version 9.4.

Query of the TriNetX Research Network dataset identified a total of 81,126 PwCF, including 19,979 pediatric patients ≤21 years of age. Of these, 7,649 met our inclusion criteria for cases, compared to 22,516,240 patients without CF (controls). There was a statistically significant difference in cases versus controls for age, race, and ethnicity as seen in Table 1 .

The prevalence of GI disorders was significantly higher among children with CF ( Table 2 ). GI disorders were diagnosed in 5,556 of pediatric PwCF (72%), compared to 120,804 controls (0.54%) (p < 0.0001). Similarly, the prevalence of each GI disorder subtype—pancreatic, hepatobiliary, and intestinal—was significantly higher among children with CF. Among pancreatic disorders in pediatric PwCF, exocrine pancreatic insufficiency was the most common (31%), followed by chronic pancreatitis (19%). Among hepatobiliary disorders, liver disease unspecified (4.7%) and other specified disease of the liver (3%) were the most common. For intestinal and luminal disorders, constipation was the most common at 38%, followed by gastroesophageal reflux disease (GERD) at 30%.

Digestants were the most common GI medication class used in pediatric PwCF ( Table 2 ). Compared to 0.01% of controls, 46% of PwCF were administered digestants. This is consistent with the higher prevalence of constipation seen in pediatric PwCF. Laxative medications and gastric medications (e.g., proton pump inhibitors, antacids) were the second and third most commonly administered GI medications among pediatric PwCF, consistent with the higher prevalence of constipation and GERD in PwCF.

The prevalence of GI procedures was less than 10% in both groups ( Table 2 ). However, PwCF still had significantly higher rates of GI procedures, with the most common being esophagogastroduodenoscopy (EGD) (9.4%) and colonoscopy (2.8%) (p < 0.0001).

Our study highlights the broad and diverse diagnostic codes practitioners use to describe chronic pancreatic and liver diseases, none of which are specific for CF. Our findings of a high prevalence of chronic pancreatic disease (at 53.6%, including exocrine pancreatic insufficiency, chronic pancreatitis, and pancreatic steatorrhea, while 46.2% of patients had received a diagnosis of other specified disease of the pancreas) in pediatric PwCF are concordant with prior knowledge regarding pancreatic disease, indicating a prevalence of up to 85% and its underlying mechanisms in CF (1, 2, 14–18). This database further provides greater specificity to the different liver manifestations in CF while prior studies have previously assumed a broad prevalence of <5%–68% (19), and it also highlights the evolving landscape of CFLD as most recent clinical guidelines recommend stratifying disease as advanced CFLD (aCFLD) versus CF hepatobiliary involvement (CFHBI) (20). The variety of codes used especially in liver disease is reflective of the heterogeneity of this disease and the likely need for a CF-specific ICD-10 code which will be instrumental in future clinical trials (15, 16).

Our findings further demonstrate a high prevalence of constipation among pediatric PwCF and as such more frequent use of digestants and laxative medications. While most laxatives were oral, a surprisingly high proportion of patients needed rectal therapies to help alleviate constipation due to likely a higher severity of disease and due to the risk of distal intestinal obstruction syndrome (DIOS) in CF. Whether or not rectal therapies were more frequently used in the setting of DIOS or routinely for constipation management could not be appreciated through this study and could be a future area of investigation.

Despite GERD being the second most common intestinal diagnosis, correlating with the high use of gastric medications, there was a lower prevalence of esophageal manometry and impedance testing. There is a lack of diagnostic guidelines within the pediatric CF population for GERD, but typically, pH impedance, esophageal manometry, or endoscopy is not indicated unless the patient is resistant to therapy or has alarming symptoms (21). The use of HEMT has improved lung disease, its impact on GERD has not yet been examine. Defining the true prevalence of GERD in this population along with the prevalence of a healthy control group prior to the broad adoption of HEMT use is critical as it allows us to determine in the future if HEMT therapy allows PwCF to have a similar prevalence of GI manifestations with adequate and early therapy (13). Therefore, considering the role HEMT has on improving lung disease, we may better delineate the prevalence of a GERD diagnosis for either 1) true reflux or 2) to aid in the management of malnutrition in PwCF by administering acid blockade for presumed improved enzyme absorption.

The overall prevalence of GI procedures, although thought to be low, was comparably significantly greater in the PwCF compared to the non-CF control sample even with the Bonferroni adjustment due to such a large control sample. Despite the greater risk of sedation in PwCF, procedures were more likely to be performed on these patients again likely due to the significantly higher burden of disease. Endoscopies being more common in patients with CF over controls but still proportionally less common than controls may be secondary to multiple factors including: 1) historically, gastroenterologists were scarcely involved in the care of patients with CF as this was primarily managed by pulmonologists and pediatricians; 2) some of the GI CF manifestations especially reflux and dysmotility are widely accepted in the CF community as part of the disease; and 3) secondary to underlying lung disease and subsequently greater risk of sedation. Thus, it is likely that despite a high burden of disease, the use of procedures is less commonly felt to be needed for diagnosis.

Prior studies have shown a high GI symptom burden for these patients; thus, defining the prevalence of GI-related manifestations is important to develop measures that improve quality of life and define changes in disease state after using highly effective modulators.