Impact of hepatitis virus infection on inpatient outcomes of acute pancreatitis: A population-based study.

Acute pancreatitis (AP) is inflammation of the pancreas, characterized clinically by acute abdominal pain and elevated levels of pancreatic enzymes in the blood. Incidence is about 36.4 in 100,000 population annually, with women affected more than men (7.9% vs 5.7%, respectively). [ 1 ] Incidence of AP is increasing in the US and Europe, making it the leading cause of hospitalization due to gastrointestinal disorders. [ 1 ] Gallstones (40–70%) and excess or prolonged alcohol consumption (30%) are the most frequent causes of AP, followed by hypertriglyceridemia, [ 2 ] certain drugs (e.g., azathioprine, valproic acid, ACE-inhibitors), [ 3 ] infections, trauma and gene mutations and polymorphisms. [ 2 , 4 ] In about 10% of AP cases, miscellaneous causes are responsible, including infection by parasites (e.g., Ascaris lumbricoides ), bacteria (e.g., Leptospira and Mycoplasma ), and viruses (e.g., mumps-virus, herpes simplex, varicella-zoster, coxsackievirus and hepatitis viruses), [ 5 ] with each organism causing AP through different mechanisms. [ 4 ] However, to date, studies on the viral causes of AP are limited. More than half of hepatitis cases in the United States are caused by the presence of a specific hepatitis virus. The most common causes are hepatitis A virus (HAV), hepatitis B virus (HBV) and hepatitis C virus (HCV), any of which may result in acute disease with symptoms such as nausea, abdominal pain, fatigue and jaundice, leading to significant morbidity and mortality worldwide. [ 6 ] Acute infection may lead to chronic infection and sequelae, sometimes progressing to cirrhosis and hepatocellular carcinoma. [ 6 , 7 ] In addition, hepatitis D virus and hepatitis E virus (HEV) have been found to cause hepatitis sporadically, primarily in countries where the viruses are endemic. [ 7 ] The surveillance report of the Centers for Disease Control and Prevention reported incidence of HAV (0.4 per 100,000 population), HBV (1.1 per 100,000 population), and HCV (0.8 per 100,000 population) in 2015, but the absence of symptoms in many cases precludes accurate statistics of incidence. [ 8 ] While hepatitis viruses infect mainly hepatocytes, tissues of other organs such as the pancreas have also been found to carry antigens for hepatitis viruses A, B, C and non-A/non-B viruses. [ 9 ] In particular, AP is associated with fulminant hepatic failure and hepatitis virus antigens have been found in results of histological and serological examination of pancreatic tissue at autopsy. [ 10 ] A study of the frequency and characteristics of pancreatitis in acute viral hepatitis found that AP occurs in 5.65% of non-fulminant acute hepatitis-infected patients and that it is usually mild and both diseases can be managed conservatively; among 124 patients, AP was associated with HAV in 2 patients, HBV in 1 patient and HEV in 4 patients. [ 11 ] However, no previous population-based study has yet evaluated the impact of hepatitis virus infection on outcomes of AP. Since greater understanding of the role of viral hepatitis in patients with AP may help risk stratification and optimization of treatment, we conducted this study to investigate the prevalence and associations of hepatitis virus infection with outcomes of AP using a nationally representative large cohort.

Conceptualization: Hsin-Wu Lai, Hung-Chang Hung. Data curation: Hsin-Wu Lai, Chun-Che Lin. Formal analysis: Hung-Chang Hung. Methodology: Hung-Chang Hung. Supervision: Hsin-Wu Lai. Writing – original draft: Hsin-Wu Lai, Chun-Che Lin. Writing – review & editing: Hsin-Wu Lai, Chun-Che Lin.

Data for this study were extracted from the 2005 to 2018 Nationwide Inpatient Sample (NIS), a database developed by the Healthcare Cost and Utilization Project (HCUP) in the United States that is maintained by the Agency for Healthcare Research and Quality. [ 12 ] The NIS database represents a 20% sample of inpatient admissions from 45 states and 1051 hospitals that participated in collecting patient data at discharge. Principal and secondary diagnoses, principal and secondary procedures, admission date and diagnosis, discharge status, patient demographic data, and length of stay are included for each inpatient. Statistical sample weights that allow generalized estimates of national case volumes are also provided in the NIS. This was a population-based, retrospective study analyzing data from the NIS database during 2005 to 2018. The certificate number HCUP-72GWV89J8 was obtained from HCUP-NIS. The study conforms to the data-use agreement of the NIS. [ 12 ] Because the NIS received signed informed consent from all included patients before the data collection, and the patients were deidentified in the database, signed informed consent was waived for the present analysis. Accordingly, further approval was also exempt by the institutional review board of our institution. Inclusion criteria were: all patients in the 2005-2018 NIS with principal diagnosis of AP, identified through the International Classification of Disease ninth and tenth revision (ICD-9 and ICD-10) (ICD-9: 577.0; ICD-10: K85). Patients with history of pancreatic cancer (ICD9: 157; ICD10: C25) were excluded. We also excluded patients without complete data on age, gender, study outcomes, and sample weights. For further analysis, the study population was further divided into patients with or without hepatitis virus infection (ICD9: 070; ICD10: B15-19). Main outcomes were: death/discharge against medical advice (DAMA); prolonged length of stay (LOS), defined as LOS > 75th percentile of the cohort; life-threatening complications, i.e., ischemia/infarction of the intestine (ICD-9: 557; ICD-10: K55.0, K55.1, K55.9), portal vein thrombosis (ICD-9: 452; ICD-10: I81), acute organ failure including respiratory, renal and hepatic failure (ICD-9: 518.81, 584, 570; ICD-10: J96.0, N17, K72.0), systemic inflammatory response syndrome (SIRS) (ICD-9: 995.9, 998.02, 038, 785.52; ICD-10: A40, A41, R65, T81.12), hypovolemic shock (ICD-9: 785.50, 276.52; ICD-10: R57.1), and abdominal compartment syndrome (ICD-9: 729.73; ICD-10: M79.A3). Demographic characteristics including patients’ age, gender, race, year of admission, and primary payer (insurance status). Clinical characteristics including etiology of AP (i.e., gallstone/biliary, alcohol abuse, prior endoscopic retrograde cholangiopancreatography, hypercalcemia, and disorders of lipid metabolism), type of viral hepatitis, and medical comorbidities (obesity, ischemic heart disease, metabolic syndrome, diabetes mellitus [DM], and chronic kidney disease [CKD]) were extracted from the database using relevant ICD diagnostic codes. Data of hospital characteristics including hospital bed size, region, and location/teaching status were also identified as covariates in the analyses. Categorical variables were summarized as frequency and percentage, and also were assessed using the χ 2 test. Numerical variables were presented as mean and standardized error. Logistic regression analysis was performed to estimate the associations between study variables and death/DAMA, prolonged LOS, and occurrence of life-threatening complication. Significant variables having a P value < .05 in the univariate analysis entered the multivariable model, including age, gender, race, year of admission, insurance status, etiology of AP, medical comorbidities, hospital bed size, hospital region, and hospital location/teaching status. All p values were two-sided and a P < .05 was regarded as a statistical significance. Considering the complex sampling design of the HCUP-NIS data, all statistical analyses were performed using PROC SURVEY procedures in the statistical software package SAS software version 9.4 (SAS Institute Inc., Cary, NC). All results reported are weighted unless noted otherwise.

Figure 1 shows the flowchart of study selection process. A total of 781,347 patients with the diagnosis of AP were identified in the 2005 to 2018 NIS database. After excluding a diagnosis of pancreatic cancer, missing data for age, gender, study outcomes and data lacking sample weights, a total of 775,416 patients (representing 3,787,043 subjects in the entire US population) hospitalized for AP comprised the analytic cohort (Fig. 1 ). Flow chart of patients’ selection. HCUP = Healthcare Cost and Utilization Project, NIS = Nationwide Inpatient Sample. Demographic and clinical characteristics of the study cohort are summarized in Table 1 . Among the study cohort, 26,407 subjects (3.4%) had been diagnosed hepatitis virus infection, whereas 749,009 (96.6%) had not. Mean age of the subjects was 51.4 years. In the cohort, 52.2% were males and 47.8% females, and most subjects (64.7%) were Whites. In addition, 13.5% patients had obesity, 11.7% had ischemic heart disease, 25.9% had DM and 7.4% had CKD. Among the patients with hepatitis virus infection, 23,857 (90.3%) had hepatitis C infection. Compared to patients who had no viral hepatitis, patients with viral hepatitis were younger (50.9 vs 51.4 years, P < .001) and had higher percentages of male (63.2% vs 51.9%, P < .001), Black race (27.8% vs 16.2%, P < .001), alcohol abuse (52.3% vs 27.8%, P < .001), and CKD (10.4% vs 7.3%, P < .001). In contrast, patients without viral hepatitis had lower percentages of admissions during 2015 to 2018 (28.0% vs 30.5%, P < .001) and insurance covered by private including Health Maintenance Organization (16.7% vs 33.9%, P < .001). With regard to hospital characteristics, higher percentages of patients with viral hepatitis admitted to hospitals of higher bed size (56.7% vs 53.6%, P < .001), urban-teaching hospitals (55.1% vs 46.9%, P < .001), whereas lower percentages of patients with viral hepatitis admitted to hospitals located in the Midwest (17.8% vs 22.6%, P < .001). The frequencies of nearly all outcomes in patients with viral hepatitis were significantly higher as compared with those in patients without viral hepatitis (all P < .001), except for ischemia/ infarction of the intestine, hypovolemic shock and abdominal compartment syndrome (Table 1 ). Characteristics of study population. Continuous variables are presented as mean ± SE, categorical variables are presented as unweighted counts (weighted percentage). P value < .05 were showed in bold. AP = acute pancreatitis, CKD = chronic kidney disease, DAMA = discharge against medical advice, DM = diabetes mellitus, ERCP = endoscopic retrograde cholangiopancreatography, LOS = length of stay, SIRS = systemic inflammatory response syndrome. Excluded In-hospital mortality patients. LOS > 5 days. Associations between death/DAMA and viral hepatitis in AP patients are summarized in Table 2 . After adjusting for relevant variables in the multivariable analysis, hepatitis virus infection was significantly associated with increased odds of death/DAMA (aOR = 1.33, 95% CI = 1.26–1.40, P < .001) and prolonged LOS (aOR = 1.12, 95% CI = 1.09–1.16, P < .001) (Table 2 ). Associations between viral hepatitis, death/AMA and prolonged LOS in hospitalized patients with AP. P value < .05 is shown in bold. AP = acute pancreatitis, CKD = chronic kidney disease, DAMA = discharge against medical advice, DM = diabetes mellitus, ERCP = endoscopic retrograde cholangiopancreatography, LOS = length of stay. Excluded in-hospital mortality patients. LOS > 5 days. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Associations between life-threatening complication and viral hepatitis in AP patients are summarized in Table 3 . Multivariable analyses revelated that hepatitis virus infection was not independently associated with ischemia/ infarction of the intestine, portal vein thrombosis, hypovolemic shock, or abdominal compartment syndrome. In contrast, individuals with hepatitis virus infection were significantly more likely to have acute organ failure (aOR = 1.06, 95% CI = 1.01–1.12, P = .030) (Table 3 ). Associations between viral hepatitis and life-threatening complications in hospitalized patients with AP. P value < .05 is shown in bold. aOR = adjusted odds ratio, AP = acute pancreatitis, CI = confidence interval, CKD = chronic kidney disease, DM = diabetes mellitus, ERCP = endoscopic retrograde cholangiopancreatography. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, race, year of admission, insurance status, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, disorders of lipid metabolism, ischemic heart disease, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, year of admission, insurance status, alcohol abuse, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, alcohol abuse, hypercalcemia, obesity, ischemic heart disease, hospital bed size, hospital region and hospital location/teaching status. Table 4 shows the results of stratified associations between study outcomes and hepatitis virus infection. Age and status of alcohol abuse did not modify the associations between viral hepatitis, death/DAMA and prolonged LOS. However, after adjustment, viral hepatitis was significantly associated with life-threatening complications among patients < 60 years (aOR = 1.10, 95% CI = 1.04–1.16) but not ≥60 years. Also, after adjustment, viral hepatitis was significantly associated with life-threatening complications (aOR = 1.08, 95% CI = 1.01–1.16) among patients without alcohol abuse (Table 4 ). Stratified associations between viral hepatitis and inpatient outcomes of AP. P value < .05 is shown in bold. AMA = against medical advice, aOR = adjusted odds ratio, AP = acute pancreatitis, CI = confidence interval, CKD = chronic kidney disease, DM = diabetes mellitus, ERCP = endoscopic retrograde cholangiopancreatography, LOS = length of stay. Adjusted for gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, hospital bed size, hospital region and hospital location/teaching status. Adjusted for gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for gender, race, year of admission, insurance status, gallstone/biliary, alcohol abuse, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status. Adjusted for age group, gender, race, year of admission, insurance status, gallstone/biliary, ERCP, hypercalcemia, disorders of lipid metabolism, obesity, ischemic heart disease, DM, CKD, hospital bed size, hospital region and hospital location/teaching status.

Results of the present study show that hepatitis virus infection independently predicts worse inpatient outcomes among patients admitted for AP. In specific, hepatitis virus infection is independently associated with greater risk for death/DAMA, prolonged LOS, and acute organ failure. These findings may help risk stratification and development of proper strategies in the management of patients with AP. Previous studies have also reported increased deaths and organ failure in AP patients with viral hepatitis. Viruses are, in fact, the largest group of infectious organisms associated with AP, and HBV is most often the causative hepatitis virus, especially in those who have undergone liver transplantation. [ 4 , 5 ] In a meta-analysis of AP patients, risk of death was found to double when severity was marked by the presence of infected pancreatic necrosis and organ failure; when either is present, the diagnosis is critical AP and when both are present the mortality rate is as high as 43%. [ 13 ] In AP that develops in conjunction with fulminant viral hepatitis, mortality is contingent upon severity of the hepatitis infection; while AP with HAV responds to conservative measures, HCV and HEV are associated with necrotizing pancreatitis, multiple organ failure and death. [ 14 ] DAMA, in particular, has been linked with higher rates of 90-day mortality and 30-day readmission in general population who admitted to the hospitals. [ 15 ] Also, it is believed that DAMA patients often readmitted in critical condition and were more likely to have organ failure leading to death. [ 15 , 16 ] Post-discharge interventions is needed for 1% to 2% of DAMA patients to help reduce the high risk of adverse outcomes. [ 16 ] Although readmission may be important when evaluating the short-term outcome of AP, such data are not available from the NIS dataset. In the present study, nevertheless, we combined deaths and DAMA as a composite outcome, and found hepatitis virus infection independently predicted higher risk for deaths/DAMA. In our analysis, after adjustment, hepatitis viral infection did not significantly raise the risk for ischemia/infarction of the intestine or portal vein thrombosis. Specifically, colonic necrosis and splenic vein thrombosis are noted complications of AP, [ 17 , 18 ] and colonic involvement especially is associated with increased mortality due to erosion or inflammation of the large bowel and progression to life threatening necrosis, hemorrhage or perforation. [ 17 ] The predictive factors identified for splenic vein thrombosis in severe AP are alcoholism, smoking and male gender; however, in that study, viral causes were listed with other possible causes of AP but were not associated with severe AP. [ 18 ] Other authors have also reported colonic stenosis following severe AP. [ 18 , 19 ] Aswani et al [ 19 ] reviewed venous complications of AP, noting that splenic vein thrombosis may lead to splenic complications such as rupture or subscapular hematoma, and that in portal vein thrombosis, cavernous transformation of the portal vein occurring within three weeks after obstruction is expected to maintain hepatopetal flow in about 70% to 100% of patients. In the present study, however, no significant associations were found between colonic necrosis or splenic vein thrombosis in AP patients with hepatitis virus infection compared to those without viral hepatitis. Among the subtypes hepatitis virus in the present study, HCV was predominant in AP patients, followed by HBV and HAV. In the United States, 2.4 million people are estimated to be living with hepatitis C. [ 20 ] On the other hand, inactive carriers are the largest group (300 million) among patients infected with HBV, and the diagnostic workup can mask positive HBV status because it is based on negative HBeAg and positive anti-HBeAg, with normal alanine aminotransferase levels and only slight liver fibrosis. [ 21 ] In a study of the frequency and characteristics of pancreatitis in acute viral hepatitis, HAV and HBV were present among 5% of patients with AP, and a small percentage had HEV. [ 11 ] While none of the hepatitis virus-infected AP patients in the NIS database had HEV, at least 14 case studies of HEV infection associated with AP have been reported in the literature, mainly describing young adults in areas endemic for the virus. [ 22 ] A review study identified 55 cases of HEV-associated AP, among which about one-fifth were severe and overall mortality was 3.2%. [ 23 ] Kamar et al [ 24 ] reported that HEV, despite its association with AP, was also associated with extrahepatic disorders, including hematologic disorders and autoimmune myocarditis and thyroiditis. Our analysis revealed that hepatitis virus infection is independently associated with greater risk for acute organ failure. A previous systematic review summarized the data on pancreatitis associated with viral hepatitis. Pancreatitis was reported as severe in 32.88% of cases. Moreover, the review documented that respiratory system was affected in 2.74% of patients, and 6.85% experienced renal failure, while 5.48% experienced a multiorgan dysfunction syndrome. [ 25 ] Most cases of acute pancreatitis due to hepatitis viruses had been reported in association with acute liver failure. [ 11 ] In addition, about 60% of all acute liver failure patients fulfill the criteria for SIRS irrespective of presence or absence of infection and often contributes towards multi organ failure. [ 26 ] This partly addressed the potential mechanisms on the greater risk of acute organ failure posed by hepatitis virus infection among patients with AP. An important part of our strategy for this study was using the comprehensive, nationally representative patient data in the 2005 to 2018 NIS, the largest inpatient sample in the United States containing data from more than 1000 participating hospitals in the United States. Nevertheless, this study had several limitations. First, the study variables were identified based on medical records of inpatient admissions using ICD-9 diagnostic codes from discharge data. Therefore, the time sequence of individual exposure to hepatitis virus infection and outcomes is not totally clear, that is, it is not known in every case whether the hepatitis virus was active during hospitalization. Second, the diagnosis relied on the accuracy of hospital coders who assign the codes based on discharge data and pathology reports. Nevertheless, errors in ICD coding are reported to be limited. Unmeasured confounders such as lifestyle, behavior, environmental exposure, family medical history and clinical lab data were not included in the NIS database and could not be accounted for. Because the NIS database includes discharge data only, the present study also could not evaluate long-term health status of the patient population. Long-term prospective study is required to confirm results of the present study and to further explore associations between AP outcomes in the presence of acute and chronic viral hepatitis. Lab data and severity of viral hepatitis were not available in the NIS, thus could not be controlled for.

This nationally representative large study found hepatitis virus infection independently predicts worse inpatient outcomes in patients admitted for AP. The findings may help risk stratification and development of proper strategies in the management of patients suffered from AP.