We did a cross-section study using the National Inpatient Sample (NIS) database between 2011 and 2015. The NIS database is the largest all-payer inpatient database in the United States and contains a sample of over eight million inpatient hospitalizations each year, which represents about 20% of all the discharges from all community hospitals participating in the Healthcare Cost and Utilization Project (HCUP). It excludes rehabilitation and long-term acute care hospitals. Each record of the NIS data includes up to 25 primary and secondary diagnoses in addition to up to 15 primary and secondary procedures. It also has the patient demographics, discharge status, length of stay, disease severity and comorbidity measures.
Study population, inclusion and exclusion criteriaAll adult patients (≥18 years old) from the NIS database between 2011 and 2015 were included. Using the International Classification of Diseases 9th version (ICD-9) code, we identified all records with cholangiocarcinoma using the appropriate codes. We identified patients' demographics using the Clinical Classification Software codes provided by the HCUP and appropriate ICD-9 codes. Using appropriate ICD-9 codes and Elixhauser comorbidity, we also identified the comorbidities of interest as shown in Supplementary Table 1 and 2, Supplemental digital Introduction Cholangiocarcinoma is a rare malignancy accounting for 3% of gastrointestinal cancers in the USA. While multiple risk factors for cholangiocarcinoma are established, other potential risk factors are still controversial. Herein, we used a large national database to investigate possible risk factors and associations. Method We used the National Inpatient Sample database to review all admissions between 2011 and 2015. We grouped patients based on the presence and absence of cholangiocarcinoma. Using multivariate logistic regression analysis, we assessed the association between obesity, alcohol abuse, smoking, diabetes mellitus and cholangiocarcinoma. Results Out of 30 9552 95 admissions, 20 030 had cholangiocarcinoma. Cholangiocarcinoma patients were older (67 ± 12.8 vs. 57 ± 20.6; P < 0.001) and had fewer female patients (48 vs. 59%; P < 0.001). Multivariate logistic regression analysis showed that diabetes mellitus was associated with cholangiocarcinoma (OR, 1.04; 95% CI, 1.01-1.08; P < 0.001). On the other hand, alcohol, smoking and obesity were all inversely associated with cholangiocarcinoma (OR, 0.75; 95% CI, 0.69-0.81; P < 0.001), (OR, 0.75; 95% CI, 0.71-0.79; P < 0.001) and (OR, 0.71; 95% CI, 0.67-0.75; P < 0.001), respectively. In addition, compared to Whites, Hispanic and Asian/Pacific Islander races were more associated with cholangiocarcinoma (OR, 1.27; 95% CI, 1.21-1.34) and (OR, 1.79; 95% CI, 1.67-1.92) (P < 0.001 for all), respectively, whereas African American race was inversely associated with cholangiocarcinoma (OR, 0.85; 95% CI, 0.81-0.89; P < 0.001). Conclusion Patients with a diagnosis of diabetes mellitus or from certain ethnic groups (Hispanic and Asian/Pacific Islander) ar...