Society: AGA

Introduction:
It is controversial whether glucagon-like peptide 1 (GLP-1) receptor agonists (RA) are associated with adverse pancreatic, gallbladder or biliary outcomes. Using a global aggregated electronic health record database, we assessed if starting patients with T2DM on GLP-1 RAs was associated with an increased risk of pancreaticobiliary disease.
Materials and Methods:
A retrospective query was constructed on the TriNetX platform querying 91 healthcare organizations from 12 countries. We identified all non-deceased adult patients with active records who had T2DM at the time of GLP-1 RA initiation between 2002 to 2022. The control group was patients with T2DM at the time of sodium-glucose co-transport 2 inhibitors (SGLT2i) initiation. Patients in both cohorts were excluded if they had a history of pancreatic or hepatobiliary disease or had undergone any procedure related to this system up until 1 day after the index event. Patients were also excluded if they ever had a neoplasm of any digestive organ or a history of malignant neuroendocrine tumors. Furthermore, both groups were exclusive and propensity-matched based on key factors that would predict GLP-1 RA vs. SGLT2i initiation, as well as risk factors for pancreatic and hepatobiliary disease (Table 1). We extracted the number of patients who were diagnosed with acute pancreatitis or cholelithiasis, as well as those who had a first-time diagnosis of cholecystitis, chronic pancreatitis, or underwent cholecystectomy. We combined bile duct stones, bile duct obstruction, biliary colic, and biliary fistula under a single outcome of biliary disease. Finally, we extracted the number of patients who underwent first-time endoscopic retrograde cholangiopancreatography (ERCP) as well as the number of hospitalizations or emergency visits. All outcomes were identified between 1 day after the index event and up to 3 years. Odds ratios were calculated within the TriNetX Analytics Platform with significance set at a 2-sided P value <0.05.
Results:
After propensity matching, there were 206, 717 patients on GLP-1 RA without SGLT2i and vice versa. Within 3 years of initiating GLP-1 RA, (1) patients were at increased risk of developing cholecystitis (OR 1.42; 95% CI 1.25-1.61) and cholelithiasis (1.32; 1.24-1.39), and undergoing cholecystectomy (1.41; 1.24-1.61); (2) patients also had an increased risk of emergency department visits (1.27; 1.24-1.30) and hospitalization (1.14; 1.13-1.16); these associations persisted with initial A1c <7%; however, (3) patients were not more likely to experience a biliary disease outcome or acute/chronic pancreatitis, or undergo an ERCP.
Conclusion:
GLP-1 RAs are associated with developing cholecystitis and cholelithiasis, as well as undergoing a cholecystectomy; however, they are not associated with acute or chronic pancreatitis or other biliary disease events.

<b>Table 1: Characteristics of GLP-1 RA and SGLT2i Cohort Before and After Propensity Matching</b>

Table 1: Characteristics of GLP-1 RA and SGLT2i Cohort Before and After Propensity Matching

<b>Figure 1: Comparison of Risk of Pancreatic, Gall bladder, and Hepatobiliary Complications in GLP-1 RA and SGLT2i</b>

Figure 1: Comparison of Risk of Pancreatic, Gall bladder, and Hepatobiliary Complications in GLP-1 RA and SGLT2i

Background: Acute cholangitis is caused by a bacterial infection in the biliary tract, most commonly due to choledocholithiasis and malignant strictures. Previous studies have shown a 30-day mortality in patients with acute cholangitis of 2.6-16%. The importance of early ERCP in these patients is controversial. The aims of the study were to examine the prognosis of patients with acute cholangitis and compare outcomes in those who had early ERCP within 24 hours from diagnosis and those who had ERCP undertaken later.
Methods: A prospectively maintained endoscopic database was used to identify all patients who underwent ERCP between 2010 and 2021, in a university hospital setting which has a catchment area of around 320,000 inhabitants. Medical records of patients diagnosed with cholangitis (k83.0) and calculus of bile duct with cholangitis (k80.3) according to ICD-10 diagnostic codes were reviewed. The Tokyo guidelines from 2018 were used to verify the diagnosis of acute cholangitis and to determine the severity grading. Sepsis was analysed and defined by the Sepsis-3 criteria from 2016. Early ERCP was defined as ERCP procedure undertaken within 24 hours from diagnosis.
Results: Among 2616 patients who underwent ERCP, 240 (9%) fulfilled the inclusion criteria, of whom 107 were women (45%), median age 72 years. The most common cause of acute cholangitis was gallstones in 180 (75%) of cases, 46 (19%) had a malignant cause and 14 (6%) other causes. A total of 61 (25%) of the patients underwent ERCP within 24 hours from diagnosis. Overall 30-day mortality was 3.3% (n=8), mortality was similar in the early ERCP group (4.9%) and those who had ERCP later (2.8%). Patients who underwent early ERCP were more likely to have severe acute cholangitis than those who underwent ERCP later (31% vs 18%, p=0.047) but had a shorter median hospital stay (4 vs. 6 days, p=0.006). In all 54 (23%) patients had sepsis during their hospital stay, 33% among those who had early ERCP vs. 19% among those who underwent ERCP later (p=0.033)
Conclusion: Thirty-day mortality was only 3% in patients with acute cholangitis. ERCP was undertaken in 25% of patients within 24 hours, more commonly in those with more severe symptoms. Despite increased severity in those who underwent ERCP early, this was associated with a shorter hospital stay, which suggests benefits of performing early ERCP in patients with acute severe cholangitis.

Background and Aims: Nonalcoholic fatty liver disease (NAFLD) is common in people with human immunodeficiency virus (HIV)(PWH). Identifying those with significant (F2-4) and advanced (F3-4) fibrosis (AF) is important. Noninvasive testing has been well studied in people with NAFLD without HIV. Our aim was to determine the performance of FIB-4 and NAFLD fibrosis score (NFS) in PWH and NAFLD to identify those with F2-4 and F3-4 by liver stiffness measurement (LSM).
Methods: We prospectively collected data on adults from 8 clinic sites in the NIH sponsored HIV NASH CRN. All had HIV infection on antiretroviral treatment (ART) ≥ 6 months with HIV RNA < 200 copies/mL. Those with viral hepatitis, other liver disease, excessive alcohol use or hepatic decompensation were excluded. Vibration controlled transient elastrography for LSM was performed and stage ≥2 and ≥3 defined LSM ≥8 and ≥11.4 kPa and compared to FIB-4 and NFS.
Results: 312 participants were analyzed: mean age 54.7 yrs, 71% male, 28% White, 50% Black, 20% Hispanic, with mean BMI 30.5 kg/m². Features of the metabolic syndrome were common: hyperlipidemia 39%, diabetes 25%, and hypertension 59%. The mean CD4 was 700 cells/mm³, 84% had undetectable HIV RNA and the duration of ART was 18 yrs with most taking a nonnucleoside reverse transcriptase inhibitor (90%) and an integrase inhibitor (80%). The mean LSM was 6.27 kPa. The prevalence of LSM defined stage ≥ 2 and ≥ 3 was 15% 5%, respectively. The operating characteristics [sensitivity (Sens), specificity (Spec), positive (PPV) and negative (NPV) predictive values] of FIB-4 and NFS for F2-4 and F3-4 by FIB-4 (<1.3 and >2.67) and NFS (<-1.455 and >0.675) thresholds are shown in the table. For LSM > 8kPa (F2-4), FIB-4 <1.3 and NFS <-1.455 had Sens 57% and 68%, Spec 63% and 52%, and NPV of 89% and 90%, respectively. For FIB-4 >2.67 and NFS > 0.675, the Spec increased to 97% and 93% with NPV 85% to detect LSM >8kPa. For LSM > 11.4 kPa (F3-4), FIB-4 <1.3 and NFS <-1.455 had Sens 81% and 87%, Spec 62% and 51% and NPV 98% and 99%, respectively. For FIB-4 >2.67 and NFS >1.675, the Spec increased to 97% and 94% with NPV of 96% to detect F3-4. Neither FIB-4 or NFS at either threshold had good PPV to detected F2-4 or F3-4. Both FIB-4 and NFS had similar AUROC for F2-4 (0.63 and 0.64) and F3-4 (0.78 and 0.80).
Conclusions: FIB4 and NFS have excellent Sens (>80%) at low thresholds and high Spec (>90%) at high thresholds with high (>95%) NPV at any threshold for detecting AF by LSM. Spec and NPV for FIB-4 and NFS were also high for F2-4 albeit with lower Sens. Therefore, both FIB-4 and NFS can be used as screening tools in those living with HIV to identify AF who needs further testing (LSM) and might benefit from interventions while excluding those with mild fibrosis who do not require further testing or referral to hepatologist.

Background: Metabolic liver disease and cardiovascular disease (CVD) share numerous risk factors. CVD accounts for 40% of total deaths in non-alcoholic fatty liver disease (NAFLD). Therefore, early risk stratification of patients for prevention is crucial. We evaluated the association of MRI-measured liver pathology with CVD outcomes.

Methods: Of 502,506 participants in the UK Biobank study, 33,616 had complete MRI liver data. Liver MRI biomarkers of fat (PDFF) and disease activity (cT1) were measured using LiverMultiScan®. Multivariable Cox regression was used to model associations between liver MRI metrics and CVD outcomes (new onset major CVD event, CVD hospitalization, and new onset: coronary artery disease (CAD), atrial fibrillation (AF), embolism/vascular events (VTE), heart failure (HF), and stroke) and all-cause mortality. In multivariable models, associations were adjusted for selected blood biomarkers (FIB-4 scores or liver function tests) and demographics (body mass index (BMI), age, and sex).

Results: Participants were followed up for a median of 2.5 years (range:1.1-5.2), with a median time to major CVD events of 1.4 years (range:0.01-5). Median age was 65 years (range:45-82), and median BMI was 26 kg/m2 (range:13-62). In multivariate analyses, liver disease activity, but not fat, was associated with an increased risk of major CVD events [HR(CI) 1.14(1.03-1.26), p=0.008] (figure 1), AF [1.30 (1.12-1.5), p<0.001]; HF [1.30 (1.08 - 1.58), p=0.004]; CVD hospitalisation [1.27(1.18-1.387, p<0.001] and all-cause mortality [1.19(1.02-1.38), p=0.026]. FIB-4 was only associated with HF [1.06 (1.01 - 1.10)), p=0.007]. These associations were also present at pre-determined clinical thresholds (cT1>800ms and PDFF>5%) for CVD hospitalization [1.38(1.1-1.75, p=0.005]. The risk of CVD hospitalisation was also independently associated with cT1 in individuals without metabolic syndrome [1.26(1.13-1.4), p<0.001] (figure 2).

Conclusion: LiverMultiScan® assessment of liver disease activity is associated with higher risk of future CVD hospitalization, all-cause mortality and CVD events independently from blood biomarkers. This study indicates that even in individuals without metabolic syndrome, liver disease activity is an important and directly modifiable risk factor for heart disease. Future clinical trials of liver therapeutics may consider liver cT1 as a surrogate endpoint of CVD risk.

Figure 1: Liver disease activity (cT1) is associated with any major cardiovascular events in the whole population: Multivariable model: Hazard ratios of associations for risk of any major CVD event in the whole cohort; variables treated as continuous based on Z-scores to enable comparisons across different unit scales. Significant associations shown in red

Figure 1: Liver disease activity (cT1) is associated with any major cardiovascular events in the whole population: Multivariable model: Hazard ratios of associations for risk of any major CVD event in the whole cohort; variables treated as continuous based on Z-scores to enable comparisons across different unit scales. Significant associations shown in red

Figure 2: Liver disease activity (cT1) is associated with hospitalisation due to cardiovascular reasons in population with no metabolic syndrome: Multivariable model: Hazard ratios of associations for risk of hospitalization due to cardiovascular reasons in the subset of the population with no metabolic syndrome; variables treated as continuous based on Z-scores to enable comparisons across different unit scales. Significant associations shown in red.

Figure 2: Liver disease activity (cT1) is associated with hospitalisation due to cardiovascular reasons in population with no metabolic syndrome: Multivariable model: Hazard ratios of associations for risk of hospitalization due to cardiovascular reasons in the subset of the population with no metabolic syndrome; variables treated as continuous based on Z-scores to enable comparisons across different unit scales. Significant associations shown in red.

Adriana Roca-Fernandez

Appearances