Abstract
Background
Diabetes has been proposed to be a risk factor for and a consequence of pancreatic cancer (PC). The relationship between recent-onset diabetes and PC is not well understood, and data in minorities are sparse. We examined the relationships between recent-onset diabetes and PC incidence in African Americans and Latinos in the Multiethnic Cohort.
Methods
A total of 48 995 African Americans and Latinos without prior diabetes and cancer at baseline (1993–1996) were included in the study. Questionnaires, Medicare data, and California hospital discharge files were used to identify new diabetes diagnoses. Cox regressions were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for cancer associated with diabetes and with diabetes duration.
Results
A total of 15 833 (32.3%) participants developed diabetes between baseline and 2013. A total of 408 incident PC cases were identified during follow-up. Diabetes was associated with PC (HRage75 = 2.39, 95% CI = 1.91 to 2.98). Individuals with recent-onset diabetes (within three or fewer years of PC diagnosis) had a greater risk compared with those with long-term diabetes across all ages. The HRage75 for recent-onset diabetes was 4.08 (95% CI = 2.76 to 6.03) in Latinos and 3.38 (95% CI = 2.30 to 4.98) in African Americans.
Conclusions
Diabetes was associated with a more than twofold higher risk of PC in African Americans and Latinos, but recent-onset diabetes was associated with a 2.3-fold greater increase in risk of PC than long-standing diabetes. Our findings support the hypothesis that recent-onset diabetes is a manifestation of PC and that long-standing diabetes is a risk factor for this malignancy.
Pancreatic cancer is one of the most fatal cancers, with an overall five-year survival rate of only 8% (1). The reason for this dismal prognosis is that the vast majority of pancreatic cancer patients (80%) are diagnosed at a late stage. Identification of high-risk individuals and ability to detect pancreatic cancer earlier might help improve patient outcomes (2).
Diabetes mellitus has been consistently associated with pancreatic cancer in previous epidemiological studies, with a twofold higher risk of developing pancreatic cancer among diabetes patients (3–5). The diabetes–pancreatic cancer relationship, however, is complex and is thought to be bi-directional (6–9); that is, diabetes has been proposed to be both a risk factor for and a consequence of pancreatic cancer. The prevalence of diabetes among pancreatic cancer patients is unusually high relative to other cancers (10,11). Most importantly, the majority (>50%) of the diabetes patients with pancreatic cancer are diagnosed with diabetes less than three years before the cancer diagnosis (12,13). It has been shown, among pancreatic cancer patients undergoing pancreaticoduodenectomy, that over half of patients with recent-onset diabetes have their diabetes paradoxically resolved postoperatively, whereas no effect is observed in those with long-standing diabetes (14). Based on these findings, recent-onset diabetes is thought to be an early manifestation of asymptomatic pancreatic cancer, and it has been proposed as a potential early detection marker for this fatal cancer (15–18).
Previous epidemiologic studies of a temporal association between diabetes and pancreatic cancer show that patients with recent-onset diabetes have greater risk of pancreatic cancer compared with those with long-standing diabetes (4,19). The relationship between recent-onset diabetes and pancreatic cancer incidence in US minorities, however, has not been examined in population-based prospective studies (3–5).
In the United States, African Americans have the highest incidence of pancreatic cancer, whereas the incidence among Latinos is relatively low and closer to that of whites (20). Both African Americans and Latinos are at high risk of developing diabetes (21), which cannot be solely attributed to greater obesity rates in these populations (22,23). Here we examined the association between recent-onset diabetes and incidence of pancreatic cancer in a prospective study of African Americans and Latinos, two minority populations with high diabetes risk but with contrasting rates of pancreatic cancer.
Methods
Study Population
The details of the Multiethnic Cohort Study (MEC) and the baseline characteristics have been previously published (24). Briefly, the MEC is an ongoing, population-based, prospective cohort study with more than 215 000 men and women from Hawaii and California (mainly Los Angeles County) assembled between 1993 and 1996. Potential participants, age 45 to 75 years at recruitment, were identified primarily through their Department of Motor Vehicles driver license, voter registration, and Health Care Financing Administration data files. All participants returned a self-administered baseline questionnaire that sought information about demographic and anthropometric measures, family history of cancer, personal history of medical conditions, and lifestyle factors, including diet, physical activity, smoking history, and medication use.
For the current study, we included only the California component of the MEC because almost all African Americans and Latinos in the cohort are from California, and state hospital discharge data are only available for linkage for California participants through the Office of Statewide Health Planning and Development (OSHPD). Figure 1 shows the criteria we used to establish the analytic cohort. We excluded participants with a diagnosis of cancer except for nonmelanoma skin cancer (n = 6120) or diabetes (n = 12 782) before cohort entry. We excluded prevalent cancer cases because their lifestyle factors and other medical conditions were likely to be influenced by their cancer diagnosis and treatment. We also excluded participants who were younger than 50 years (n = 8503) and those with missing covariate data and/or invalid dietary data based on macronutrient intake as a marker of quality for the questionnaire (n = 2153). As a result, data on 48 995 participants (20 403 African Americans and 28 592 Latinos) were included in the current analysis. Informed consent was provided by study participants, and the institutional review boards at the University of Hawaii and the University of Southern California approved the study protocol.
Figure 1.
Selection of study population, 1993–2013. MEC = Multiethnic Cohort Study.
Incident Diabetes Identification
We used a combination of follow-up questionnaires, Centers for Medicare and Medicaid Services Medicare Chronic Condition files, and California OSHPD hospital discharge files to identify new diabetes diagnoses in cohort participants after cohort entry. Self-reported diabetes was recorded and updated in each of the subsequent follow-up questionnaires (1999–2001, 2003–2008, 2011–2013). Participants were asked whether they had ever been told by a physician that they have diabetes mellitus (yes/no), and they were additionally asked when their diabetes had been diagnosed relative to the most recent survey. In addition to these questionnaires, we identified incident diabetes via Medicare Chronic Condition files (one inpatient, skilled nursing facility or home health agency claim with diabetes ICD-9 CM 249, 250, 357.2, 362, 366.41, or at least 2 outpatient or carrier claims with diabetes ICD-9 CM between 1999 and 2014) or California OSHPD hospital discharge files (hospitalization with any of the first five diagnosis codes with diabetes ICD-9 CM between 1993 and 2014). Out of 15 833 incident diabetes cases, 6531 were identified via self-report and 9302 via Medicare and/or California OSHPD hospital discharge files; 79% of the self-reported cases were confirmed by Medicare and/or California OSHPD hospital discharge files. The earliest date of diabetes discovery from the above sources was considered the date of diagnosis.
Outcome Identification
Incident cancer cases in the cohort were identified via annual linkage to the California State Cancer Registry, which is part of the National Cancer Institute’s Surveillance, Epidemiology, and End Results program. In addition to our primary interest in pancreatic cancer (ICD-O-3 site codes C25.0–C29.9), we also included invasive breast (C50.0–C50.9), colorectal (C18.0–C18.9, C26.0, C19.9, C20.9), and prostate cancers (C61.9) as comparisons. Deaths in the cohort were ascertained through annual linkage to state death certificate files in California and periodic linkage to the National Death Index. The end point ascertainment was complete through December 31, 2013. Based on linkages to Medicare and other sources, the out-migration in the MEC from California is 2.5%; therefore, the passive follow-up for outcomes should be virtually complete.
Statistical Analysis
The prevalence of diabetes and percentage with recent-onset diabetes, defined as preceding the cancer diagnosis by three years or less, were compared between cancer cases by site using the χ2 test. Hazard rate ratios (HRs) and 95% confidence intervals (CIs) for cancer incidence associated with diabetes were calculated with Cox proportional hazards models.
Incident diabetes was categorized as yes/no, and duration of diabetes was dichotomized into recent-onset (three or fewer years) and long-standing (more than three years) diabetes. Incident diabetes status and diabetes duration were time-dependent variables. Sex and ethnicity were entered as strata variables in the pancreatic cancer model. Each case of pancreatic cancer defined a risk set within a stratum consisting of that case and all other cohort members who were alive, on study, and free of pancreatic cancer at the age of diagnosis for the index case. The age (in days) of diagnosis of pancreatic cancer for the index case was compared with the age of diagnosis of diabetes for each member of the risk set, and the difference between these two ages formed the time-dependent diabetes duration variable (duration was set to zero if the difference was negative or if a risk set member was never diagnosed with diabetes). The proportional hazards assumption was tested by adding a time-dependent interaction term between diabetes and age. The proportionality assumption for pancreatic cancer analysis was not met, and to account for this, we modeled the effect of time-dependent diabetes status and diabetes duration as log-linear in attained age (ie, age of the index case in each risk set) and provided hazard ratios for specific ages. The proportional hazards assumption was met for the other cancer types. Cox models were adjusted for body mass index (BMI <25, 25 to <30, or ≥30 kg/m2), cigarette smoking status (never, former, or current), and alcohol intake (0, <12, or ≥12 g of ethanol/d), collected from the baseline questionnaire. Red meat intake was included into the colorectal and pancreatic cancer models. For breast cancer, we also adjusted for parity (zero, one, two, three, and four or more). All P values were two-sided, and a P value of less than .05 was considered statistically significant. Statistical analyses were performed with SAS 9.4 (SAS Institute, Inc., Cary NC). PROC PHREG in SAS was used to conduct the time-dependent analysis.
Results
The baseline characteristics of study participants by incident diabetes status are shown in Table 1. Participants who developed diabetes during follow up (n = 15 833, 32.3%) had higher BMI at baseline compared with those who did not develop diabetes (n = 33 162, 67.7%). There were no differences between cohort members with and without diabetes in terms of race/ethnicity, sex, education level, smoking status, and alcohol consumption.
Table 1.
Baseline characteristics of participants by incident diabetes status*
| Characteristics | No diabetes | Diabetes | Total |
|---|---|---|---|
| (n = 33 162) | (n = 15 833) | (n = 48 995) | |
| No. (%) | No. (%) | No. (%) | |
| Race | |||
| African American | 14 115 (42.6) | 6288 (39.7) | 20 403 (41.6) |
| Latino | 19 047 (57.4) | 9545 (60.3) | 28 592 (58.4) |
| Sex | |||
| Men | 14 742 (44.5) | 6741 (42.6) | 21 483 (43.8) |
| Women | 18 420 (55.5) | 9092 (57.4) | 27 512 (56.2) |
| Education | |||
| ≤ high school | 18 742 (56.5) | 9432 (59.6) | 28 174 (57.5) |
| Vocational/some college | 8880 (26.8) | 3988 (25.2) | 12 868 (26.3) |
| College or higher | 4968 (15) | 2074 (13.1) | 7042 (14.4) |
| BMI, kg/m2 | |||
| <25 | 10 906 (32.9) | 3098 (19.6) | 14 004 (28.6) |
| 25–< 30 | 14 821 (44.7) | 7091 (44.8) | 21 912 (44.7) |
| ≥30 | 6767 (20.4) | 5390 (34) | 12 157 (24.8) |
| Smoking status | |||
| Never | 14 443 (43.6) | 6734 (42.5) | 21 177 (43.2) |
| Past | 12 145 (36.6) | 5960 (37.6) | 18 105 (37) |
| Current | 5727 (17.3) | 2644 (16.7) | 8371 (17.1) |
| Alcohol intake, g/d | |||
| Nondrinker | 16 319 (49.2) | 8485 (53.6) | 24 804 (50.6) |
| ≤12 | 11 268 (34) | 5069 (32.0) | 16 337 (33.3) |
| >12 | 5575 (16.8) | 2279 (14.4) | 7854 (16) |
BMI = body mass index.
During an average follow-up of 14 years, 128 participants developed pancreatic cancer among those with incident diabetes (age-adjusted incidence rate = 72.3 per 100 000) (data not shown). Among participants without incident diabetes, 280 participants developed pancreatic cancer (age-adjusted incidence rate = 36.0 per 100 000). The median age of pancreatic cancer diagnosis (range) was 76.7 (55.3–92.3) years in this analysis.
Figure 2 shows the presence of diabetes among incident cases of pancreatic cancer (n = 408), colorectal cancer (n = 1398), prostate cancer (n = 3163), and breast cancer (n = 1509) cases in the MEC. The frequency of incident diabetes was statistically significantly higher (P < .001) among participants who developed pancreatic cancer (31.4%) during the follow-up period compared with those who developed colorectal cancer (16.4%), breast cancer (14.5%), or prostate cancer (11.7%). The presence of recent-onset diabetes was strikingly higher among pancreatic cancer cases (16.4%) (P < .001) compared with those with colorectal (6.7%), breast (5.3%), and prostate (5.5%) cancer.
Figure 2.
Presence of incident diabetes in pancreatic (n = 408), breast (n = 1509), colorectal (n = 1398), and prostate (n = 3163) cancers in African Americans and Latinos in the Multiethnic Cohort, 1993–2013. Duration of diabetes was defined as the time period between the dates of diabetes discovery and cancer diagnosis. Recent-onset diabetes was defined as preceding the cancer diagnosis by three years or less.
In pancreatic cancer with diabetes, 52.3% of cases developed diabetes in the 36 months preceding the pancreatic cancer diagnosis. Figure 3 shows the distribution of diabetes duration among pancreatic cancer cases with diabetes; the median duration of diabetes (range) was 34.4 (0.07–193.4) months.
Figure 3.
Distribution of diabetes duration in pancreatic cancer cases (n = 128) in African Americans and Latinos in the Multiethnic Cohort, 1993–2013. Duration of diabetes was defined as the time period between the dates of diabetes discovery and cancer diagnosis. The median duration of diabetes (range) was 34.4 (0.07–193.4) months.
The associations of diabetes with pancreatic cancer at attained age 65, 75, and 85 years are shown in Table 2. Diabetes overall was associated with pancreatic cancer at age 65 years (HR = 4.60, 95% CI = 3.00 to 7.05) and age 75 years (HR = 2.39, 95% CI = 1.91 to 2.98). At age 75 years (closest to the median age of pancreatic cancer diagnosis in this study), the association for recent-onset diabetes (three or fewer years) was 2.3-fold greater (HR = 3.71, 95% CI = 2.83 to 4.88) than for long-standing diabetes (more than three years; HR = 1.61, 95% CI = 1.18 to 2.21). The association between recent-onset diabetes and pancreatic cancer was statistically significant at all ages examined, but declined with increasing age. At age 65 years, having recent-onset diabetes corresponded to an estimated hazard ratio of 6.56 compared with no diabetes. By age 75 and 85 years, the hazard ratios for recent-onset diabetes dropped to 3.71 and 2.26, respectively.
Table 2.
Association between incident diabetes and pancreatic cancer by sex and race/ethnicity in the Multiethnic Cohort
| Variable | All HR* (95% CI) | Men HR* (95% CI) | Women HR* (95% CI) | African American HR* (95% CI) | Latino HR* (95% CI) |
|---|---|---|---|---|---|
| At attained age 65 y | |||||
| Diabetes | |||||
| No | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Yes | 4.60 (3.00 to 7.05) | 4.53 (2.46 to 8.36) | 4.69 (2.59 to 8.51) | 3.38 (1.78 to 6.42) | 5.99 (3.36 to 10.69) |
| Duration ≤3 y | 6.56 (4.01 to 10.72) | 5.93 (2.92 to 12.04) | 7.27 (3.67 to 14.41) | 5.49 (2.65 to 11.39) | 7.70 (3.94 to 15.03) |
| Duration >3 y | 2.58 (1.33 to 4.98) | 2.95 (1.14 to 7.62) | 2.28 (0.91 to 5.72) | 1.48 (0.51 to 4.28) | 3.94 (1.66 to 9.34) |
| At attained age 75 y | |||||
| Diabetes | |||||
| No | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Yes | 2.39 (1.91 to 2.98) | 2.28 (1.63 to 3.18) | 2.48 (1.83 to 3.34) | 2.03 (1.47 to 2.82) | 2.76 (2.03 to 3.76) |
| Duration ≤3 y | 3.71 (2.83 to 4.88) | 3.22 (2.09 to 4.96) | 4.12 (2.88 to 5.89) | 3.38 (2.30 to 4.98) | 4.08 (2.76 to 6.03) |
| Duration >3 y | 1.61 (1.18 to 2.21) | 1.72 (1.09 to 2.70) | 1.52 (0.98 to 2.36) | 1.19 (0.72 to 1.98) | 2.05 (1.36 to 3.08) |
| At attained age 85 y | |||||
| Diabetes | |||||
| No | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) | 1.00 (ref) |
| Yes | 1.35 (0.97 to 1.87) | 1.25 (0.73 to 2.12) | 1.41 (0.93 to 2.15) | 1.30 (0.83 to 2.03) | 1.40 (0.86 to 2.29) |
| Duration ≤3 y | 2.26 (1.44 to 3.53) | 1.88 (0.90 to 3.94) | 2.51 (1.42 to 4.41) | 2.21 (1.22 to 4.03) | 2.34 (1.19 to 4.59) |
| Duration >3 y | 1.07 (0.71 to 1.62) | 1.07 (0.55 to 2.08) | 1.07 (0.63 to 1.81) | 0.98 (0.56 to 1.74) | 1.16 (0.63 to 2.11) |
Based on Cox proportional hazards models, with age as the time metric and sex and race/ethnicity as strata variables, and adjusted for BMI, smoking status, alcohol intake, and red meat intake. Diabetes and duration of diabetes were entered as time-dependent variables, as main effects, and as interaction terms with attained age. CI = confidence interval; HR = hazard ratio.
An association between recent-onset diabetes and pancreatic cancer at attained age 75 years was observed in men (HR = 3.22, 95% CI = 2.09 to 4.96) and women (HR = 4.12, 95% CI = 2.88 to 5.89) (Table 2). An association was observed in African Americans (HRage75 = 3.38, 95% CI = 2.30 to 4.98) and Latinos (HR age75 = 4.08, 95% CI = 2.76 to 6.03). At age 75 years, long-standing diabetes was statistically significantly associated with pancreatic cancer in Latinos (HR = 2.05, 95% CI = 1.36 to 3.08) but not in African Americans (HR = 1.19, 95% CI = 0.72 to 1.98).
As a comparison, we examined the association of diabetes with prostate, breast, and colorectal cancers in the MEC (Supplementary Table 1, available online). No association was observed between recent-onset diabetes or long-standing diabetes with colorectal and breast cancer. Diabetes was inversely associated with prostate cancer (HR = 0.85, 95% CI = 0.76 to 0.96); however, this inverse association was limited to long-standing diabetes (HR = 0.74, 95% CI = 0.64 to 0.86).
Discussion
In this large population-based cohort study, diabetes was associated with an approximately twofold increased risk of pancreatic cancer. When stratifying by disease duration, individuals with recent-onset diabetes had the highest risk of developing pancreatic cancer. Importantly, we demonstrated that the association of recent-onset diabetes with pancreatic cancer incidence was evident in African Americans and Latinos, two understudied minority populations with high risk of diabetes but different pancreatic cancer rates.
Previous epidemiological studies have reported a positive association between diabetes and pancreatic cancer (3,5). Results from meta-analyses show that diabetes overall is associated with an approximately twofold increased risk of pancreatic cancer (summary RRs ranged from 1.82 to 1.94) (3,4). Besides a few Veterans Administration (VA)–based studies (25,26), prospective studies with substantial numbers of African American and Latino participants are nonexistent, but the results of the current study in African Americans and Latinos agree with published findings in whites.
The association between diabetes and pancreatic cancer was strong during the first three years after diabetes diagnosis. Individuals with recent-onset diabetes had a greater risk of pancreatic cancer compared with those who had long-term diabetes, consistent with those of previous studies (3–5,27). In our study, the association was evident in men and women and in African Americans and Latinos. This striking relationship was unique to pancreatic cancer and was not seen in breast, prostate, or colorectal cancer. A clinical study using the Mayo Clinic database with 500 cancer patients also reported similar findings (10). Specifically, the study found that the prevalence of recent-onset diabetes was higher among pancreatic cancer patients compared with those other common cancers, that is, prostate, breast, and colorectal cancers. It should be noted that in our current analysis, we only considered incident diabetes, and thus the overall prevalence of diabetes and the associations reported for these cancers are likely to be underestimated.
The potential mechanisms for pancreatic cancer–induced diabetes have been reviewed (28,29). Phenomena that occur close to pancreatic cancer diagnosis that might aggravate glucose intolerance or cause diabetes include progressive pancreatic tissue destruction due to tumor growth and cachexia. However, hyperglycemia and diabetes often occur months to years before diagnosis, and they resolve after cancer resection (14,30). The most likely mechanism is a paraneoplastic phenomenon caused by tumor-secreted products (29). One such factor is adrenomedullin, which is overexpressed in pancreatic cancer, is secreted by pancreatic cancer cell lines, and has been shown to inhibit insulin secretion in beta cells (29). Interestingly, plasma adrenomedullin levels were found to be higher in pancreatic cancer cases compared with healthy controls and were even higher in pancreatic cancer–induced diabetes (31).
In our study, we also found that long-term diabetes was modestly but statistically significantly associated with pancreatic cancer. Previous studies have suggested that diabetes contributes to the pathogenesis of pancreatic cancer independent of obesity by as-yet unclear mechanism(s) (32). Many mediators of a pro-carcinogenic effect of diabetes on pancreatic cancer development have been proposed, including insulin-like growth factor-1 (IGF-1), insulin resistance and the compensatory increased levels of insulin, and sustained hyperglycemia (33). We did not observe a statistically significant association between long-term diabetes and pancreatic cancer in African Americans. This may be due to a higher number of long-term diabetes cases in African Americans being excluded when prevalent diabetes was removed from the analyses, African Americans having died of diabetes complications or other causes at a higher rate (in the MEC, the death rate in African Americans with diabetes was 36% higher than in Latinos), and/or to biologic differences. Szczepaniak et al. showed that among adults with mild obesity and without diabetes, African Americans have the highest compensatory insulin secretion and lowest pancreatic triglyceride levels compared with whites and Hispanics (34). But overall, our results support a modest causal relationship between long-term diabetes and pancreatic cancer.
Our study has numerous strengths, including its prospective nature, inclusion of minority populations, and ability to analyze other cancers for comparison and to assess detection bias and adjust for potential confounders. The major limitation of this study is that diabetes identification is based on self-report, Medicare claims, and hospital discharge records, which may have led to under-reporting and imprecise date of onset. The underdiagnosis of diabetes is likely to occur in any population-based study like the MEC. If there had been a high proportion of undiagnosed diabetes among noncases, the associations in our study would have been attenuated. African Americans and Latinos in the MEC may not be representative of the US population, and this may affect the generalizability of our results. However, the distribution of education level and marital status in the cohort generally resembles that reported by the US Census in LA County for the same age and ethnic groups (24).
In conclusion, our findings support the hypothesis that recent-onset diabetes in pancreatic cancer is a manifestation of developing pancreatic cancer. Studies to elucidate the potential mechanisms underlying pancreatic cancer–associated diabetes need to be conducted. Our work suggests that patients with recent-onset diabetes who go on to develop pancreatic cancer represent a high-risk population of patients who can be studied for additional risk predictors and may be targeted for development of the liquid biopsy assays that are needed in the field for earlier diagnosis.
Funding
This study was supported by the National Cancer Institute (RO1CA209798 to VWS and UO1CA164973 to LLM). V. W. Setiawan is supported by an American Cancer Society Research Scholar Grant (RSG-16-250-01-CPHPS).
Notes
Affiliations of authors: Department of Preventive Medicine (VWS, DOS, CAH, KRM) and Norris Comprehensive Cancer Center (VWS, JP, CAH), Keck School of Medicine of University of Southern California, Los Angeles, CA; Department of Internal Medicine, Mayo Clinic College of Medicine, Rochester, MN (STC); Epidemiology Program, University of Hawaii Cancer Center, Honolulu, HI (GM, LLM, LRW); Division of Gastroenterology, Departments of Medicine, Cedars-Sinai Medical Center and Department of Veterans Affairs, Los Angeles, CA (SJP).
The study funders had no role in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; or the decision to submit the manuscript for publication.
Supplementary Material
References
- 1. American Cancer Society. Cancer Facts and Figures 2017. Atlanta: American Cancer Society; 2017. [Google Scholar]
- 2. Chari ST, Kelly K, Hollingsworth MA, et al. Early detection of sporadic pancreatic cancer: Summative review. Pancreas. 2015;445:693–712. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Ben Q, Xu M, Ning X, et al. Diabetes mellitus and risk of pancreatic cancer: A meta-analysis of cohort studies. Eur J Cancer. 2011;4713:1928–1937. [DOI] [PubMed] [Google Scholar]
- 4. Huxley R, Ansary-Moghaddam A, Berrington de Gonzalez A, et al. Type-II diabetes and pancreatic cancer: A meta-analysis of 36 studies. Br J Cancer. 2005;9211:2076–2083. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Tan J, You Y, Guo F, et al. Association of elevated risk of pancreatic cancer in diabetic patients: A systematic review and meta-analysis. Oncol Lett. 2017;133:1247–1255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Hart PA, Chari ST.. Diabetes mellitus and pancreatic cancer: Why the association matters? Pancreas. 2013;428:1207–1209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Andersen DK, Andren-Sandberg A, Duell EJ, et al. Pancreatitis-diabetes-pancreatic cancer: Summary of an NIDDK-NCI workshop. Pancreas. 2013;428:1227–1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Salvatore T, Marfella R, Rizzo MR, et al. Pancreatic cancer and diabetes: A two-way relationship in the perspective of diabetologist. Int J Surg. 2015;21(Suppl 1):S72–S77. [DOI] [PubMed] [Google Scholar]
- 9. Li J, Cao G, Ma Q, et al. The bidirectional interation between pancreatic cancer and diabetes. World J Surg Oncol. 2012;10:171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Aggarwal G, Kamada P, Chari ST.. Prevalence of diabetes mellitus in pancreatic cancer compared to common cancers. Pancreas. 2013;422:198–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Dankner R, Boffetta P, Balicer RD, et al. Time-dependent risk of cancer after a diabetes diagnosis in a cohort of 2.3 million adults. Am J Epidemiol. 2016;18312:1098–1106. [DOI] [PubMed] [Google Scholar]
- 12. Aggarwal G, Rabe KG, Petersen GM, et al. New-onset diabetes in pancreatic cancer: A study in the primary care setting. Pancreatology. 2012;122:156–161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chari ST, Leibson CL, Rabe KG, et al. Pancreatic cancer-associated diabetes mellitus: Prevalence and temporal association with diagnosis of cancer. Gastroenterology. 2008;1341:95–101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Pannala R, Leirness JB, Bamlet WR, et al. Prevalence and clinical profile of pancreatic cancer-associated diabetes mellitus. Gastroenterology. 2008;1344:981–987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Noy A, Bilezikian JP.. Clinical review 63: Diabetes and pancreatic cancer: Clues to the early diagnosis of pancreatic malignancy. J Clin Endocrinol Metab. 1994;795:1223–1231. [DOI] [PubMed] [Google Scholar]
- 16. Andersen DK. Diabetes and cancer: Placing the association in perspective. Curr Opin Endocrinol Diabetes Obes. 2013;202:81–86. [DOI] [PubMed] [Google Scholar]
- 17. Risch HA, Yu H, Lu L, et al. Detectable symptomatology preceding the diagnosis of pancreatic cancer and absolute risk of pancreatic cancer diagnosis. Am J Epidemiol. 2015;1821:26–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Green RC Jr, Baggenstoss AH, Sprague RG.. Diabetes mellitus in association with primary carcinoma of the pancreas. Diabetes. 1958;74:308–311. [DOI] [PubMed] [Google Scholar]
- 19. Li D, Tang H, Hassan MM, et al. Diabetes and risk of pancreatic cancer: A pooled analysis of three large case-control studies. Cancer Causes Control. 2011;222:189–197. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Ashktorab H, Kupfer SS, Brim H, et al. Racial disparity in gastrointestinal cancer risk. Gastroenterology. 2017;1534:910–923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Menke A, Casagrande S, Geiss L, et al. Prevalence of and trends in diabetes among adults in the United States, 1988-2012. JAMA. 2015;31410:1021–1029. [DOI] [PubMed] [Google Scholar]
- 22. Bonham GS, Brock DB.. The relationship of diabetes with race, sex, and obesity. Am J Clin Nutr. 1985;414:776–783. [DOI] [PubMed] [Google Scholar]
- 23. O'Brien TR, Flanders WD, Decoufle P, et al. Are racial differences in the prevalence of diabetes in adults explained by differences in obesity? JAMA. 1989;26211:1485–1488. [PubMed] [Google Scholar]
- 24. Kolonel LN, Henderson BE, Hankin JH, et al. A multiethnic cohort in Hawaii and Los Angeles: Baseline characteristics. Am J Epidemiol. 2000;1514:346–357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Gupta S, Vittinghoff E, Bertenthal D, et al. New-onset diabetes and pancreatic cancer. Clin Gastroenterol Hepatol. 2006;411:1366–1372; quiz 1301. [DOI] [PubMed] [Google Scholar]
- 26. Jamal MM, Yoon EJ, Vega KJ, et al. Diabetes mellitus as a risk factor for gastrointestinal cancer among American veterans. World J Gastroenterol. 2009;1542:5274–5278. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ben Q, Cai Q, Li Z, et al. The relationship between new-onset diabetes mellitus and pancreatic cancer risk: A case-control study. Eur J Cancer. 2011;472:248–254. [DOI] [PubMed] [Google Scholar]
- 28. Hart PA, Bellin MD, Andersen DK, et al. Type 3c (pancreatogenic) diabetes mellitus secondary to chronic pancreatitis and pancreatic cancer. Lancet Gastroenterol Hepatol. 2016;13:226–237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Sah RP, Nagpal SJ, Mukhopadhyay D, et al. New insights into pancreatic cancer-induced paraneoplastic diabetes. Nat Rev Gastroenterol Hepatol. 2013;107:423–433. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Permert J, Ihse I, Jorfeldt L, et al. Improved glucose metabolism after subtotal pancreatectomy for pancreatic cancer. Br J Surg. 1993;808:1047–1050. [DOI] [PubMed] [Google Scholar]
- 31. Aggarwal G, Ramachandran V, Javeed N, et al. Adrenomedullin is up-regulated in patients with pancreatic cancer and causes insulin resistance in beta cells and mice. Gastroenterology. 2012;1436:1510–1517, e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Song S, Wang B, Zhang X, et al. Long-term diabetes mellitus is associated with an increased risk of pancreatic cancer: A meta-analysis. PLoS One. 2015;107:e0134321. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Eibl G, Cruz-Monserrate Z, Korc M, et al. Diabetes mellitus and obesity as risk factors for pancreatic cancer. J Acad Nutr Diet. 2018;1184:555–567. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Szczepaniak LS, Victor RG, Mathur R, et al. Pancreatic steatosis and its relationship to beta-cell dysfunction in humans: Racial and ethnic variations. Diabetes Care. 2012;3511:2377–2383. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.