This is a retrospective national cohort study that aims to assess the risk of developing pancreatic cancer in patients with type 2 diabetes mellitus (T2DM) who are being treated with Glucagon-like Peptide-1 receptor agonists. We exclude patients with multiple risk factors for pancreatic cancer from our study. We retrospectively follow the patients for 7 years after the initiation of treatment. Current evidence is controversial about their use and the possible risk of pancreatic disease. We aim to revoke the association of pancreatic cancer with their use in support of their continued use due to their beneficial cardiovascular and renal effects.
Background: GLP-1 RAs are widely used for T2DM treatment due to their cardiorenal and metabolic benefits. This study examines the risk of pancreatic cancer with GLP-1 RA use in patients with T2DM. Methods: We analyzed TriNetX’s deidentified research database using the U.S. Collaborative Network comprising 62 healthcare organizations across the U.S.A. Patients with T2DM were split into two cohorts: one receiving GLP-1 RAs, and one not receiving GLP-1 RAs. We excluded patients with known risk factors for pancreatic cancer, including pancreatic cysts, a personal or family history of BRCA1, BRCA2, CDKN2A, KRAS, MEN1, MLH1, MSH2, NOTCH1, PALB2, PMS2, and PRSS1S genes, family history of pancreatic cancer, and VHL syndrome. Using a 1:1 propensity score-matching model based on baseline characteristics and comorbidities, we created comparable cohorts. We then compared the rate of pancreatic cancer between the two cohorts at a 7-year interval. Results: Out of 7,146,015 identified patients with T2DM, 10.3% were on a GLP-1 RA and 89.7% were not. Post-PSM, 721,110 patients were in each group. Patients on GLP-1 RAs had a 0.1% risk compared to a 0.2% risk of pancreatic cancer in the 7-year timeframe. Conclusion: The use of GLP-1 RAs in patients with type 2 diabetes mellitus (T2DM) does not appear to substantially elevate the risk of pancreatic cancer; in fact, it may potentially exert a protective effect.
Pancreatic cancer poses a significant health concern globally due to its aggressive nature and poor prognosis. Notably, patients with type 2 diabetes mellitus (T2DM) have garnered attention as a potentially high-risk group for developing pancreatic cancer. Glucagon-like Peptide-1 receptor agonists (GLP-1 RAs) have emerged as a cornerstone in the management of type 2 diabetes mellitus (T2DM), offering glycemic control with favorable cardiovascular outcomes. However, recent observations and clinical studies have raised concerns regarding the potential association between GLP-1 RA use and the risk of pancreatic cancer. This association has prompted extensive investigation into the safety profile of these medications, aiming to elucidate the underlying mechanisms and assess the true magnitude of risk. Understanding the implications of GLP-1 RA therapy on pancreatic health is crucial for clinicians and patients alike, guiding informed decision-making in diabetes management. Therefore, our study aims to assess the risk of pancreatic cancer in patients with T2DM who are being treated with GLP-1 RAs.
The Institution Board Review Committee at the Charleston Area Medical Center has approved this study (IRB Number: 23-956). Written informed consent from patients was waived due to the deidentified nature of the TriNetX clinical database. The TriNetX (Cambridge, MA, USA) database is a global federal research network that combines real-time data with electronic medical records. This study was conducted using the TriNetX database through the US Collaborative Network, which comprises 63 Healthcare Organizations (HCOs) from the United States of America. Patients with type 2 diabetes mellitus were identified and divided into patients receiving GLP-1 RAs and those who are not. We excluded patients with pancreatic cysts, specific genes associated with pancreatic cancer, which include BRCA1, BRCA2, CDKN2A, KRAS, MEN1, MLH1 MSH2, NOTCH1, PALB2, PMS2, and PRSS1S, a family history of the previously mentioned genes, a family history of pancreatic cancer, and Von Hippel–Lindau syndrome. We also excluded patients above the age of 90. The prior exclusion criteria were identified with specific ICD-10 and genetic codes. A list of all codes used in our study is highlighted in the Supplementary Material. We compared the rate of pancreatic cancer between the two cohorts over 7 years.
Adult patients aged ≥18 years with type 2 diabetes mellitus in real-time data were identified. Patients with type 2 diabetes mellitus who were included in our study were divided into two cohorts: the first cohort included patients who were being treated with a GLP-1 RA and the second cohort included patients who were not being treated with a GLP-1 RA. Following this, propensity score matching (PSM) of both cohorts to ensure successful and effective balancing was completed. TriNetX conducts logistic regression analysis to obtain propensity scores for each cohort. PSM was performed using patients’ baseline demographics, comorbidities, lab values, and medications received. Demographics included in PSM included age at index, race, gender, and body mass index (BMI). Comorbidities included were tobacco use, history of chronic pancreatitis, prior acute pancreatitis, and alcohol use disorder. Lab values included were hemoglobin A1C levels. Medications used in PSM were metformin, insulin, and dipeptidyl peptidase-4 (DPP-4) inhibitors. A study flow diagram is shown below in Figure 1.
Following propensity score matching, an analysis of the outcomes was conducted. Cumulative Incidence curves and log-rank tests were employed to investigate the correlation of GLP-1 RA use and the risk of pancreatic cancer between groups. Risk ratios (RRs) with their respective 95% confidence intervals (CIs) were calculated for the outcomes. Statistical significance was determined at a p-value of less than 0.05. The statistical analyses were conducted using the TriNetX platform.
We identified 7,146,016 patients with type 2 diabetes mellitus. Of those, 10.3% (n = 736,015) were being treated with a GLP-1 RA. The remaining 89.7% (n = 6,410,000) were not being treated with a GLP-1 RA. The mean age in the GLP-1 RA group was 54.4 with a standard deviation (SD) of 12.8 as compared to 60.9 with an SD of 15.3 in the non-GLP-1 RA receivers. More than half the patients receiving GLP-1 RAs were females 52.6% as compared to 47.3% in the other group. In total, 61.7% of the GLP-1 RA receivers were white as compared to 57.8% of the non-receivers. The mean body mass index (BMI) ratio in the GLP-1 RA group was 35.3 with an SD of 6.8 with a mean body weight of 228.5 with an SD of 59.6, as compared to a BMI of 30.8 with an SD of 7.2 and a body weight of 200 with an SD of 55.4 in the group that was not treated with a GLP-1 RA. Of the GLP-1 RA group, 0.5% had a history of acute pancreatitis and 0.1% had chronic pancreatitis as compared to 0.9% and 0.4%, respectively, in the other group. In total, 2.1% of the GLP-1 RA group were smokers and 0.2% used alcohol as compared to 1.9% and 0.5%, respectively. The mean hemoglobin A1C in the GLP-1 RA group was 7.9 with an SD of 2.2 as compared to 7.3 with an SD of 2.1 in the other group. Of the GLP-1 RA group, 39.5% were on metformin, 24.1% were on insulin, and 7% were on dipeptidyl peptidase-4 (DPP-4) inhibitors. On the other hand, of the group that did not receive GLP-1 RAs, 19.8% were on metformin, 24.1% were on insulin, and 3.1% were on DPP-4 inhibitors.
After propensity score matching (PSM), our cohort consisted of 1,442,220 patients. It was divided evenly into two groups: those who received GLP-1 RA and those who did not. There was no statistically significant difference between the two groups in the above-mentioned variables. In the GLP-1 RA group, the mean age was 54.4 with an SD of 12.8. More than half the cohort were females (52.6%) and more than half the cohort was white 61.7%. The mean BMI was 35.3 with an SD of 6.8, and the mean weight was 228.5 with an SD of 59.6. In total, 0.5% had a history of acute pancreatitis and 0.1% had chronic pancreatitis. In the GLP-1 RA group, 2.1% were smokers, and 0.2% used alcohol. The mean hemoglobin A1C level was 7.9 with an SD of 2.2. In terms of medications, 39.5% were on metformin, 24.1% were on insulin, and 7% were on DPP-4 inhibitors. A full list of PSM components before and after matching is shown below in Table 1.
After PSM, we compared the rate of pancreatic cancer between both groups after 7 years of therapy. After 7 years of therapy, the cancer risk between the two groups was noted to be lower in the GLP-1 RA group. In the GLP-1 RA group, it was 0.14% as compared to 0.2% (p-value < 0.0001). The risk ratio of developing pancreatic cancer in the GLP-1 RA group was 0.69 with a 95% confidence interval of (0.639, 0.752). Consequently, patients who received GLP-1 RA had a calculated pancreatic cancer risk reduction of 31%, highlighting a possible pancreatic-protective effect of GLP-1 RA. A summary of results, log-rank tests, and hazard ratios is highlighted below in Table 2 with the cumulative event curve in Figure 2 highlighting the difference.
In the United States of America (U.S.A.), the estimated number of cases of new pancreatic cancer in 2024 was more than 65,000 cases. The incidence continues to rise at a rate of up to 1% per year, and it is expected to be the second leading cause of cancer death in the U.S.A. by 2030. Unfortunately, there is no effective screening method for pancreatic cancer, making its 5-year survival rate in 2020 up to 10%, which is higher than previously reported in 2000 at approximately 5%. Pancreatic cancer is usually diagnosed at the age of 71 in the U.S.A., with a slightly higher incidence in women compared to men. At the time of presentation, half the patients already have metastatic disease, with only 10–15% having surgery-amendable disease. In 2019, the United States Preventive Services Task Force (USPSTF) continued to recommend against screening for pancreatic cancer as there was no evidence that screening improves disease-related morbidity or mortality and that the risks of screening outweigh the benefits.
Risk factors for pancreatic cancer can be divided into modifiable and inherited. Some of the strongest modifiable risk factors include smoking, alcohol use, and chronic pancreatitis. Pathogenic germline gene variants associated with pancreatic cancer were found in up to 9.7% of these cases.
Smoking is the most common modifiable risk factor, identified in up to 35% of these patients with pancreatic cancer. The pathogenic mechanism involves mutations in the KRAS and p53 genes. Additionally, smoking causes a state
