New-onset diabetes mellitus and risk of pancreatic cancer: A systematic review and meta analysis.

e16778 Background: Several epidemiologic studies have illustrated a general association between diabetes mellitus and pancreatic cancer (PC). Very few, however, described a temporal association between the new onset of diabetes mellitus, particularly within the first 2 years of diagnosis, and the development of PC. We sought to systematically review the literature and perform the first meta-analysis to study the risk of PC among patients with new-onset diabetes mellitus (NOD). Methods: A systematic literature search was conducted in PubMed, MEDLINE, EMBASE, and Cochrane databases from inception through November 2019 to identify studies evaluating the relationship between NOD and PC. NOD was defined as diabetes mellitus diagnosed within 3 years or less prior to diagnosis of PC. Studies were examined and relevant data was extracted and analyzed using Comprehensive Meta-Analysis software. A random effect meta-analysis approach was used to pool the data and odds ratio was used to calculate the overall effect estimate. Results: A total of 283 articles were identified by the search but only 4 relevant studies met our inclusion criteria and they examined a total of 42055 cases and 2402213 controls. The odds ratio of pancreatic cancer development among patients with NOD was 3.021 (CI 2.192- 4.165) compared to controls with significant heterogeneity (I2= 90%). Meta regression evaluating relevant variables such as mean age, race and smoking were not able to explain the heterogeneity. A meta-regression model accounting for sample size as an independent factor showed that sample size was inversely correlated to the OR of PC in patients with NOD and was able to explain 93% of the noted heterogeneity between studies. Conclusions: Our meta-analysis has shown a significant temporal association between NOD and pancreatic cancer. Diabetes mellitus, being mostly diagnosed in an outpatient setting by primary care physicians (PCPs), can be the first clue to detect pancreatic cancer at an early stage, and thus contribute to more favorable outcomes. These findings might guide PCPs to consider pancreatic cancer in the differential diagnosis in high risk patients. Further studies are needed to discern the role of NOD in pancreatic cancer screening. Attention is required to the possible inflationary effect of low sample sizes on the odds of developing PC after NOD.