SEARCH

SEARCH BY CITATION

Keywords:

  • pancreatic cancer;
  • type 2 diabetes mellitus;
  • prognosis;
  • survival;
  • epidemiology

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES

BACKGROUND:

Recent evidence has identified pre-existing type 2 diabetes mellitus (T2DM) as a risk factor for the development of PAC, but relatively little is known about its effects on survival. Our aim was to determine the effect of varying durations of pre-existing T2DM on survival in patients with PAC.

METHODS:

We conducted a retrospective cohort study using The Health Improvement Network (THIN), a primary care electronic medical record database from the UK (2003-2010). The study cohort included all subjects with a diagnostic code for PAC. Subjects with a diagnostic code for T2DM before their PAC diagnosis were classified as exposed; otherwise, subjects were classified as unexposed. The primary outcome was overall survival. The analysis was performed using univariate and multivariable Cox proportional-hazards models. Additional analysis was performed to assess the effect of increasing duration of pre-existing T2DM [i.e., <90 days, 90 days to <1 year, 1 to <3 years, 3 to 5 years, >5 years] on survival.

RESULTS:

The study included 3,147 patients with PAC, with 745 patients having pre-existing T2DM and 2,402 patients without pre-existing T2DM. In the primary multivariate analysis, there was no difference in survival between those exposed and those unexposed to pre-existing T2DM (HR 1.02 [0.93, 1.12], p = 0.620). In the secondary analysis, only those patients with T2DM > 5 years duration had a significantly increased mortality (HR 1.16 [1.00, 1.33], p < 0.05).

CONCLUSIONS:

Long-term pre-existing T2DM is associated with increased mortality in patients diagnosed with PAC. Cancer 2013. © 2012 American Cancer Society.

Pancreatic adenocarcinoma (PAC) is the fourth leading cause of cancer death in the United States, and despite medical advances, the overall survival from pancreatic cancer remains poor.1 The 5-year survival rate for all stages combined was 3% in 1975 and has increased to only 6% in the past 35 years.1 This poor prognosis has motivated researchers to identify modifiable factors that may influence survival after the diagnosis of PAC. The identification of such factors would provide prognostic information as well as mechanistic insights that could help uncover new potential therapeutic targets.

One potentially important risk factor for PAC is type 2 diabetes mellitus (T2DM). Recent evidence has identified T2DM as a risk factor for PAC, with a higher risk among patients with a more recent diagnosis of diabetes.2, 3 A meta-analysis of 36 studies reported a summary odds ratio (OR) of 1.82 (95% confidence interval [CI], 1.66-1.89) of developing pancreatic cancer among patients with T2DM compared to controls.2 Potential mediators of this association are hyperinsulinemia and insulin resistance.4 However, the relationship between PAC and T2DM is complex. In many cases, T2DM may be a consequence of PAC. This is supported by the finding that the prevalence of T2DM among newly-diagnosed PAC cases is as high as 47%, and is more often “new-onset” (defined as <2 years duration) when compared to diabetic controls.5 Furthermore, surgical resection often causes resolution of diabetes among PAC patients if the preexisting T2DM was “new-onset” but not if the preexisting diabetes was long-standing (>2 years duration).5 It is plausible that the same mechanism by which T2DM may cause PAC may also accelerate PAC progression and impact survival.

Several studies have been conducted in various settings to elucidate the potential relationship between survival in PAC and preexisting T2DM.5-13 However, these studies have yielded conflicting results. Furthermore, the interpretation of some of these studies is hampered by small sample size and limited generalizability. More importantly, few studies have adequately addressed the effect of new-onset versus long-standing diabetes on survival in PAC, which is an important factor given the complex relationship and temporal associations between T2DM and the diagnosis of PAC.

The purpose of our study was to comprehensively examine the relationship between varying durations of preexisting T2DM and survival from PAC. We hypothesized that preexisting long-standing T2DM exposure would be associated with a shorter median survival after the diagnosis of PAC.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES

This study was conducted in The Health Improvement Network (THIN), which is a primary care electronic medical record database well-suited for epidemiologic research.14, 15 The database contains data of acceptable research quality on more than 7.5 million patients from 495 general practitioner offices in the United Kingdom. Patients included in THIN are similar in age, sex, and geographic characteristics to the general UK population.15 Approximately 3% of patients are lost annually because of leaving a practice or death. The data are collected from routine charting records from participating general practitioners through a software package named Vision. The records are then collected, stripped of identifying information, and aggregated at regular intervals. The database contains information on all past and current medical diagnoses, both acute and chronic (coded using British Read codes), and prescribed medications (coded using British National Formulary codes). Laboratory values, social habits, radiologic studies, and some physical examination elements are also electronically captured. The diagnoses of diabetes and pancreatic cancer, as well as information on death, have all been previously validated in THIN.16-18

The study cohort included all subjects with a diagnostic code for PAC. The exclusion criteria were a diagnosis of pancreatic cancer other than adenocarcinoma, age <40 years at the time of incident pancreatic cancer diagnosis, and a diagnosis of PAC before the start of follow-up in THIN.

To ensure the inclusion of only incident PAC, subjects were only defined as having PAC if their first appropriate diagnostic code occurred after the latter of 2 dates: 1) Vision date, which is the date that the primary care practice first instituted the Vision software package; and 2) 6 months after registration date, which is the date that the patient was registered with a particular practice. Only diagnoses entered after Vision software implementation are likely to represent actively recorded medical diseases rather than the retroactive documentation of preexisting conditions. The 6-month window after registration allows for the “purge” of prevalent medical conditions upon registration.19

Only subjects with adenocarcinoma of the pancreas were included for this study to avoid a mixed effect from differing cancer types. Subjects aged <40 years were excluded to avoid including patients with genetic or familial syndromes for pancreatic cancer (whose tumor biology may act independently or differently from those with sporadic pancreatic cancer).

Subjects with a diagnostic code for T2DM before their PAC diagnosis were classified as exposed; otherwise, subjects were classified as unexposed. A calculated duration of T2DM exposure was used to assign subgroups of exposed patients based on the duration of exposure. The duration of exposure was defined as beginning at the time of the first T2DM Read code entry and ending at the time of PAC Read code entry. Subgroups were defined by increasing durations of exposure as follows: <90 days, 90 days to <1 year, 1 to <3 years, 3 to 5 years, >5 years. The primary outcome was overall survival.

To compare the exposed and the unexposed groups, the Pearson chi-squared test was used for categorical variables and Student t test was used for continuous variables. To compare the multiple groups of T2DM duration, 1-way analysis of variance was used for continuous variables and the Pearson chi-squared test was used for categorical variables. Univariate and multivariate Cox proportional hazards models were used to conduct comparisons of survival between the exposed and unexposed groups. Follow-up time for the exposed and unexposed groups began at the time of incident PAC Read code entry and ended at either the time of death or censoring events, including withdrawal from the database, discontinuation of primary care practice enrollment in the database, or patient loss to follow-up. Nonproportionality was assessed using scaled Schoenfeld residual score testing.

In the primary analysis, we examined the following potential confounding variables that are associated with the exposure, and possibly the outcome: age at diagnosis, sex, body mass index (BMI), glomerular filtration rate (GFR) (as calculated by the Modification of Diet in Renal Disease [MDRD] equation), history of pancreatic resection, history of pancreatitis, history of smoking, and smoking status at the time of diagnosis. We also included the Charlson comorbidity index as a potential confounding variable because it is a predictor of mortality.20 All patients were assigned 2 points for having pancreatic cancer when calculating the Charlson score.

In order to examine the possibility that the survival differences seen across increasing duration of T2DM were due to potential confounders that are only relevant among patients with diabetes, we conducted a secondary analysis restricted among the varying durations of T2DM with T2DM <90 days as the reference group. Potential confounders examined in this restriction analysis included diabetes mellitus with complications, hemoglobin A1c (HbA1c), and antidiabetic medications, including metformin, sulfonylureas, thiazolidinediones, and insulin.

Whether a variable was included in the final Cox regression model depended on the extent to which it affected the unadjusted hazard ratio. In order to determine which variables to include in the multivariable models, we determined if the unadjusted hazard ratios differed from the hazard ratios adjusted for each potential confounder by more than 10% of the unadjusted hazard ratios. Those variables that create such a difference between unadjusted and adjusted odds ratios were included in the model; this approach has been shown to be superior to other methods of selecting confounders in observational studies.21

Multiple imputation was performed to impute missing values for HbA1c, BMI, and GFR according to the Gaussian normal imputation method using the following variables: the presence of diabetic complications, duration of diabetes, age, sex, pancreatitis, smoking status, and use of insulin, thiazolidinediones, or sulfonylureas.

All P values were 2-sided, and P <.05 was considered significant. Data extraction and statistical analyses were performed using Stata software, version 12.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES

After applying the inclusion and exclusion criteria, there were 3147 final subjects for analysis, comprising 745 patients in the exposed group and 2402 patients in the unexposed group.

Descriptive data are shown in Table 1. Those individuals exposed were significantly more likely to be older, male, and to have a pancreatic resection, history of pancreatitis, a history of smoking, a higher Charlson index, and a higher BMI. Those exposed were also significantly more likely to be smoking at the time of PAC diagnosis. Among subjects with T2DM, there was a statistically significant increase in the prevalence of diabetic complications, HbA1c, insulin use, thiazolidinedione use, sulfonylurea use, and metformin use with increasing duration of T2DM.

Table 1. Patient Characteristics
CharacteristicsNo T2DM (2402 Subjects)All T2DM (745 Subjects)PaT2DM Duration CategoriesPb
<90 d (89 Subjects)90 d-1 y (151 Subjects)1-3 y (151 Subjects)3-5 y (91 Subjects)>5 y (263 Subjects)
  • Abbreviations: BMI, body mass index; GFR, glomerular filtration rate; HbA1c, hemoglobin A1c; n/a, not applicable; SD, standard deviation; T2DM, type 2 diabetes mellitus.

  • a

    P value for the comparison between No T2DM and All T2DM.

  • b

    P value for the comparison between No T2DM and T2DM duration categories.

Age, y (SD)71.2 (11.6)72.3 (10.0).02069.4 (10.9)71.7 (10.6)71.0 (10.2)73.5 (10.4)73.9 (8.4)<.001
Male sex1110 (47%)414 (55%)<.00151 (57.3%)85 (56.3%)70 (46.4%)49 (53.9%)155 (58.9%)<.001
History of pancreatic resection195 (8%)42 (5%).0165 (5.6%)8 (5.3%)6 (4.0%)6 (6.6%)15 (5.7%).205
History of pancreatitis88 (4%)46 (6%).0058 (9.0%)9 (6.0%)10 (6.6%)6 (6.6%)12 (4.6%).049
History of smoking1278 (54%)506 (67%)<.00155 (61.8%)103 (68.2%)103 (68.2%)59 (64.8%)176 (66.9%)<.001
Smoking at time of diagnosis830 (35%)315 (41%)<.00139 (43.8%)64 (42.4%)60 (39.7%)33 (36.3%)113 (43.0%).026
Charlson index (SD)4.7 (1.6)5.9 (1.7)<.0015.4 (1.1)5.7 (1.4)6.0 (1.8)5.8 (1.6)6.3 (1.8)<.001
BMI, kg/m2 (SD)25.3 (4.6)26.6 (5.3)<.00126.5 (5.2)27 (5.4)26.4 (5.8)27.3 (5.4)26.6 (5.7)<.001
GFR, mL/min (SD)73.1 (22.0)73.4 (24.0).83573.1 (20.4)76.5 (22.2)76.1 (24.3)74.2 (24.6)70.6 (25.3).018
Diabetes with complicationsn/a77 (10.1%)n/a0 (0%)3 (1.8%)10 (6.6%)4 (4.4%)60 (22.8%)<.001
HbA1c (SD)n/a8.2% (2.2%)n/a9.1% (2.5%)8.0% (2.0%)7.8% (2.1%)7.7% (2.4%)8.4% (2.1%)<.001
Insulin usen/a283 (37.2%)n/a26 (29.2%)37 (24.5%)47 (31.1%)28 (30.8%)139 (52.9%)<.001
Thiazolidinedione usen/a86 (11.3%)n/a2 (2.3%)3 (2.0%)14 (9.3%)16 (17.6%)51 (19.4%)<.001
Sulfonylurea usen/a462 (60.8%)n/a49 (55.1%)78 (51.7%)81 (53.6%)54 (59.3%)193 (73.4%)<.001
Metformin usen/a440 (57.9%)n/a31 (34.8%)75 (49.7%)85 (56.3%)57 (62.6%)190 (72.2%)<.001

In the Cox proportional hazards model, 2733 deaths were observed with 739,698 patient-days of follow-up. The earliest time of failure was at day 0. The last observed failure was at day 5108. The median survival times in the exposed and unexposed groups were 96 days (range, 1-4369 days) and 109 days (range, 1-5108 days), respectively. The test of nonproportionality revealed a test score that was not significant, and therefore the assumption of proportionality was not violated.

Records were complete for all 3147 patients in the study for all variables except for BMI (504 missing, 16% overall, 19% in unexposed), GFR (604 missing, 19% overall, 23% in unexposed), and HbA1c (149 missing, 20% in exposed). We accounted for the missing data using multiple imputation as described in the Materials and Methods section.

In the unadjusted Cox proportional hazards model, T2DM exposure was associated with an increased risk of death (HR, 1.09 [95% CI = 1.00, 1.19], P <.05) (Table 2). Age, history of pancreatic resection, history of pancreatitis, and Charlson comorbidity index fit the criteria for a confounder in the primary analysis among all subjects. Patient sex was included in the final model, because it was felt to be an important covariate regardless of statistical significance. The adjusted hazard ratio for T2DM exposure was 1.02 (95% CI = 0.93, 1.12), P = .620.

Table 2. T2DM and Survival in Pancreatic Adenocarcinoma
 Crude Hazard RatioAdjusted Hazard Ratioa
  • Abbreviation: T2DM, type 2 diabetes mellitus.

  • a

    Adjusted for age, sex, history of pancreatic resection, pancreatitis, and Charlson index.

No T2DMReferenceReference
All T2DM1.09 (1.00, 1.19), P < .051.02 (0.93, 1.12), P = .620
T2DM < 90 d0.96 (0.76, 1.21), P = .7300.96 (0.76, 1.21), P = .722
T2DM 90 d to 1 y0.99 (0.83, 1.18), P = .9120.93 (0.77, 1.11), P = .401
T2DM 1 to 3 y1.02 (0.86, 1.22), P = .7880.94 (0.79, 1.12), P = .513
T2DM 3 to 5 y1.27 (1.02, 1.59), P < .051.23 (0.98, 1.54), P = .071
T2DM > 5 y1.27 (1.11, 1.45), P < .051.16 (1.00, 1.33), P < .05

In the unadjusted analysis to assess the effect of increasing duration of T2DM on survival, T2DM exposure >3 years was associated with increased mortality (HR, 1.27 [95% CI = 1.02, 1.59] P <.05 for duration 3 to 5 years, HR, 1.27 [95% CI = 1.11, 1.45] P <.05 for duration >5 years). Again, age, sex, pancreatic resection, pancreatitis, and Charlson index were included in the multivariate analysis due to their clinical importance. When the results were adjusted for these confounders, only a duration of diabetes >5 years was associated with an increase in mortality (HR, 1.16 [95% CI = 1.00, 1.33], P <.05). A Kaplan-Meier curve can be seen in Figure 1 comparing subjects without T2DM to those with >5 years of T2DM.

thumbnail image

Figure 1. Kaplan-Meier curves are shown for proportion surviving, with at-risk table given below the graph.

Download figure to PowerPoint

In the secondary analysis restricted among patients with T2DM, diabetes with complication, insulin use, and HbA1c met our predefined confounder selection criterion. In the unadjusted Cox regression model, patients with long-term preexisting DM were again found to have increased risk of mortality (Table 3). When adjusted for confounders, a trend was seen toward increased mortality, but was nonsignificant.

Table 3. Duration of T2DM and Survival in Pancreatic Adenocarcinoma
 Crude Hazard RatioAdjusted Hazard Ratioa
  • Abbreviation: T2DM, type 2 diabetes mellitus.

  • a

    Adjusted for age, sex, history of pancreatic resection, pancreatitis, Charlson index, diabetes with complications, insulin, and hemoglobin A1c.

T2DM < 90 dReferenceReference
T2DM 90 d to 1 y1.04 (0.78, 1.37), P = .7970.98 (0.67, 1.43), P = .920
T2DM 1 to 3 y1.06 (0.80, 1.41), P = .6640.86 (0.60, 1.24), P = .409
T2DM 3 to 5 y1.32 (0.97, 1.81) P = .0801.24 (0.83, 1.85), P = .299
T2DM > 5 y1.32 (1.02, 1.72) P < .051.36 (0.95, 1.94), P = .093

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES

In this large population-representative cohort from the United Kingdom, we found no significant overall survival difference between those with and without a T2DM diagnosis at the time of PAC diagnosis. However, when evaluating varying durations of exposure, a duration of preexisting T2DM >5 years was associated with a modestly increased mortality, whereas new-onset diabetes among patients with PAC was not associated with increased mortality. This duration effect appeared to be independent of factors associated with DM severity or DM medication use.

Several other population-based studies have examined the potential relationship between preexisting diabetes and PAC mortality and have yielded conflicting results.5-13 Busaidy et al found that diabetes is a significant independent risk factor for increased overall mortality (OR, 1.55 [1.15-2.07]) and disease-specific mortality (OR, 1.37 [1.00-1.89]).7 Chu et al similarly reported that preexisting T2DM increased mortality (HR, 1.55 [1.02-2.35]).8 However, the findings in these 2 studies are difficult to interpret as they were conducted only among PAC patients who underwent surgical resection, therefore limiting the generalizability of the results. Furthermore, there may have been confounding by the inclusion of T2DM cases that were diagnosed shortly before, or even after, the diagnosis of PAC. Wakasugi et al reported a significantly reduced survival among patients with PAC who had T2DM compared with those who were not diabetic and evaluated T2DM durations of both >2 years and <2 years compared with those who were not diabetic.9 However, many of the patients among the cohort of T2DM duration <2 years developed incident T2DM after hospital admission and the diagnosis of PAC.

On the other hand, several studies have reported similar findings to ours. Dandona et al found no difference in overall survival among surgically resected patients with a personal history of T2DM (HR, 0.87 [0.66-1.14]).11 Similarly, both McWilliams et al and Ganti et al reported no observed association between a personal history of T2DM and survival.6, 12 Pannala et al reported no significant differences in median survival among PAC patients with and without T2DM (259 vs 270 days; P = 0.9).5 Finally, Olowokure et al determined that patients with pancreatic cancer who also have diabetes do not have a worse overall survival than those without diabetes, and in fact have a statistically significant longer survival compared with those without diabetes (4.2 vs 3.5 months, P <.048).10 However, unlike our study, these studies did not address the issue of reverse causality by evaluating varying durations of T2DM exposure.

Some studies have, however, addressed the effect of new-onset vs. long-standing diabetes on survival by evaluating the duration of diabetes before the diagnosis of PAC. In a hospital-based study of 475 patients with PAC, Olson et al accounted for the possibility of PAC-associated T2DM by requiring that the self-reported history of T2DM be ≥3 years before the PAC diagnosis.13 They reported no significant association between survival and diabetes among both resected and nonresected patient subgroups (HR, 0.97 [0.42-2.26] and HR, 0.78 [0.47-1.30], respectively).13 In their study population of 206 resected patients with PAC, Chu et al found that subjects with new-onset T2DM (<2-years duration) had worse survival (HR, 1.75 [1.10-2.78]) than controls, but that there was no survival difference in those with T2DM >2-years duration (HR, 1.30 [0.75-2.25]).8 However, these studies were also subject to limitations. Because the sample size was limited in the study by Olson et al, it is possible that a moderate effect may have been missed, and as noted previously, the study by Chu et al was designed primarily to evaluate postresection survival. Furthermore, both studies only evaluated a relatively short duration of preexisting T2DM (>3 years and >2 years, respectively).

Our study adds to this limited body of literature in several ways. First, the large sample size allowed for the detection of even modest but clinically important differences in survival. Second, a sufficient number of patients with preexisting T2DM of substantial durations allowed for a definitive examination of the effects of preexisting diabetes mellitus. Our study is the only one to date to identify long-standing T2DM as a poor prognostic indicator in PAC in a large population of both unresected and resected patients with PAC, with evaluation of T2DM durations from very recent disease onset (<90 days) to long-term (>5 years) disease.

Although the mechanism of this association remains unclear, hyperinsulinemia and insulin-resistance have been proposed as potential biologic mechanisms for the observed association between T2DM and increased pancreatic cancer risk.4 Hyperinsulinemia has been implicated as a potential contributor to increased mortality in other gastrointestinal cancers, including nonmetastatic colorectal cancer.22 Studies have demonstrated that insulin receptors are present on pancreatic cancer cells.23-25 In streptozotocin-treated mice, which is a model of hypoinsulinemic diabetes, the growth of pancreatic cancer was inhibited, and levels of plasma insulin and insulin-like growth factor-1 were significantly decreased.24 Feng et al found that insulin promoted the proliferation of pancreatic cancer cells in vitro and that both insulin and glucose contributed to the chemoresistance of pancreatic cancer cells to gemcitabine.26 Furthermore, treatment with metformin or rosiglitazone suppressed pancreatic cell proliferation, regardless of insulin and glucose concentrations.26 These findings lend insight into the potential mechanisms for the observed increased risk of pancreatic cancer after the diagnosis of diabetes and also suggest the potential for decreased survival rates from pancreatic cancer among subjects with preexisting T2DM.

Given the retrospective nature of the data, much of these data are dependent on physician entry into the computerized medical record. There is some reassurance provided by the fact that many of the main diagnoses, namely cohort-, exposure-, and outcome-defining variables, have been validated in previous studies, thus minimizing the possibility of misclassification bias.16-18 Although the use of medical record data may underestimate the prevalence of T2DM in incident PAC,5 the variable of T2DM exposure has previously been validated in THIN.18 However, due to potential left censoring, the observed durations of T2DM exposure may have been underestimated. All efforts were made to correctly classify subjects on the basis of all available data.

There were occurrences of missing data in this data set. One completely missing variable was that of race, because it is not recorded in THIN. This would be of concern if certain racial groups have worsened outcomes and if their medication usage and glycemic control are similarly different. However, this effect is likely minimized due to the national health care delivery system in the United Kingdom.

Staging information is also not available in the THIN database. However, because this study population reflects the general UK population, a vast majority of subjects will have presented in advanced stages of pancreatic cancer. Further, nearly all subjects with local disease should be removed by excluding those subjects who underwent pancreatic resection. Unfortunately, distribution of advanced stages and, more importantly, their effect on survival in the context of the diagnosis of T2DM is not available because of the nature of the database.

It is possible that the differences in survival are due to differences in time to diagnosis as it relates to the duration of T2DM. However, if such an effect were to exist, it would, if anything, bias results in the opposite direction, as was observed in this study, because subjects with long-standing diabetes would likely be diagnosed earlier due to closer follow-up. The complications of T2DM are not known to mask the effects, signs, and symptoms of advanced PAC.

In our study, we did not include the variables of therapies administered for treatment of PAC, except for history of pancreatic resection, which is the only potentially curative treatment for local disease.27 Furthermore, gemcitabine-based therapies are nearly universally used for the treatment of both metastatic and locally advanced PAC disease and are typically palliative in nature. Therefore, it is unlikely that differences in the administered therapies would significantly influence the results. Furthermore, the increased mortality due to >5 years exposure to T2DM is larger than the beneficial effect that one would see if all subjects with <90 days exposure to T2DM were given chemotherapy.

In summary, we observed a significant increase in overall mortality in patients with a long-standing duration of T2DM (>5 years). The implications of these results are magnified by the growing prevalence of diabetes mellitus, the incidence of which has been increasing over the past 2 decades28, 29 and is expected to nearly double over the next 25 years.30 Further studies are needed to investigate this relationship between long-standing preexisting T2DM and increased mortality in individuals with PAC. If hyperinsulinemia or insulin resistance is implicated as a biologic mechanism for this increased mortality, then it could represent an important therapeutic target to minimize potential chemoresistance and to improve patient outcomes.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES

We acknowledge The Clinical Translational Research Award (CTSA) and Kevin Haynes for assistance with data extraction.

FUNDING SOURCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES

Grant funding for this study provided by the NIH, grant T-32-DK007740-14 and by the NIH Clinical and Translational Science Award, grant UL1-RR024134.

Conflict of Interest Disclosures

The authors made no disclosures.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. Acknowledgements
  7. FUNDING SOURCES
  8. REFERENCES
  • 1
    American Cancer Society. Cancer Facts & Figures 2010. Atlanta, GA: American Cancer Society; 2010.
  • 2
    Huxley R, Ansary-Moghaddam A, Berrington de González A, Barzi F, Woodward M. Type-II diabetes and pancreatic cancer: A meta-analysis of 36 studies. Br J Cancer. 2005; 92: 2076-2083.
  • 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; 47: 1928-1937.
  • 4
    Wang F, Herrington M, Larsson J, Permert J. The relationship between diabetes and pancreatic cancer. Mol Cancer. 2003; 2: 4.
  • 5
    Pannala R, Leirness JB, Bamlet WR, Basu A, Petersen GM, Chari ST. Prevalence and clinical profile of pancreatic cancer-associated diabetes mellitus. Gastroenterology. 2008; 134: 981-987.
  • 6
    McWilliams RR, Matsumoto ME, Burch PA, et al. Obesity adversely affects survival in pancreatic cancer patients. Cancer. 2010; 116: 5054-5062.
  • 7
    Busaidy N, Yazbeck C, Shah P, et al. Survival of resectable pancreatic cancer patients with diabetes. J Clin Oncol. 2006; 24(18 suppl): abstract 4098.
  • 8
    Chu CK, Mazo AE, Goodman M, et al. Preoperative diabetes mellitus and long-term survival after resection of pancreatic adenocarcinoma. Ann Surg Oncol. 2010; 17: 502-513.
  • 9
    Wakasugi H, Funakoshi A, Iguchi H. Clinical observations of pancreatic diabetes caused by pancreatic carcinoma, and survival period. Int J Clin Oncol. 2001; 6: 50-54.
  • 10
    Olowokure O, Beg M, Ali S, et al. Is diabetes mellitus (DM) associated with worse outcomes in pancreatic cancer (PC)? J Clin Oncol. 2010; 28(15 suppl): abstract 4114.
  • 11
    Dandona M, Linehan D, Hawkins W, Strasberg S, Gao F, Wang-Gillam A. Influence of obesity and other risk factors on survival outcomes in patients undergoing pancreaticoduodenectomy for pancreatic cancer. Pancreas. 2011; 40: 931-937.
  • 12
    Ganti AK, Potti A, Koch M, et al. Predictive value of clinical features at initial presentation in pancreatic adenocarcinoma: a series of 308 cases. Med Oncol. 2002; 19: 233-237.
  • 13
    Olson SH, Chou JF, Ludwig E, et al. Allergies, obesity, other risk factors and survival from pancreatic cancer. Int J Cancer. 2010; 127: 2412-2419.
  • 14
    Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in the United Kingdom. BMJ. 1991; 302: 766-768.
  • 15
    Lewis JD, Schninnar R, Bilker WB, Wang X, Strom BL. Validation studies of the health improvement network (THIN) database for pharmacoepidemiology research. Pharmacoepidemiol Drug Saf. 2007; 16: 393-401.
  • 16
    Haynes K, Forde KA, Schinnar R, Wong P, Strom BL, Lewis JD. Cancer incidence in the health improvement network. Pharmacoepidemiol Drug Saf. 2009; 18: 730-736.
  • 17
    Hall GC. Validation of death and suicide recording on the THIN UK primary care database. Pharmacoepidemiol Drug Saf. 2009; 18: 120-131.
  • 18
    Mulnier HE, Seaman HE, Raleigh VS, Soedamah-Muthu SS, Colhoun HM, Lawrenson RA. Mortality in people with type 2 diabetes in the UK. Diabetic Med. 2006; 23: 516-521.
  • 19
    Lewis JD, Bilker WB, Weinstein RB, Strom BL. The relationship between time since registration and measured incidence rates in the general practice research database. Pharmacoepidemiol Drug Saf. 2005; 14: 443-451.
  • 20
    Charlson ME, Pompei P, Ales KA, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987; 40: 373-383.
  • 21
    Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol. 1989; 129: 125-137.
  • 22
    Wolpin BM, Meyerhardt JA, Chan AT, et al. Insulin, the insulin-like growth factor axis, and mortality in patients with nonmetastatic colorectal cancer. J Clin Oncol. 2009; 27: 176-185.
  • 23
    Fisher WE, Muscarella P, Boros LG, Schirmer WJ. Variable effect of streptozotocin-diabetes on the growth of hamster pancreatic cancer (H2T) in the syrian hamster and nude mouse. Surgery. 1998; 123: 315-320.
  • 24
    Fisher WE, Boros LG, O'Dorisio TM, O'Dorisio MS, Schirmer WJ. GI hormonal changes in diabetes influence pancreatic cancer growth. J Surg Res. 1995; 58: 754-758.
  • 25
    Fisher WE, Boros LG, Schirmer WJ. Insulin promotes pancreatic cancer: Evidence for endocrine influence on exocrine pancreatic tumors. J Surg Res. 1996; 63: 310-313.
  • 26
    Feng Y, Busaidy N, Yeung S. Differential impact of anti-diabetic treatment on pancreatic cancer cell growth in cell culture conditions mimicking different stages in the natural history of diabetes mellitus type 2. J Clin Oncol. 2008; 26(15 suppl): 4640.
  • 27
    Jimenez R, Fernández-del Castillo C. Pancreatic cancer. In: Sleisenger and Fordtran's Gastrointestinal and Liver Disease: Pathophysiology, Diagnosis, Management. 9th ed. Mark Feldman M, Friedman LS, Brandt LJ, eds. Philadelphia, PA: Saunders/Elsevier; 2010: 1017-1034.
  • 28
    Cowie CC, Rust KF, Byrd-Holt DD, et al. Prevalence of diabetes and impaired fasting glucose in adults in the U.S. population: national health and nutrition examination survey 1999-2002. Diabetes Care. 2006; 29: 1263-1268.
  • 29
    Cowie CC, Rust KF, Ford ES, et al. Full accounting of diabetes and pre-diabetes in the U.S. population in 1988-1994 and 2005-2006. Diabetes Care. 2009; 32: 287-294.
  • 30
    Huang ES, Basu A, O'Grady M, Capretta JC. Projecting the future diabetes population size and related costs for the U.S. Diabetes Care. 2009; 32: 2225-2229.