Obesity is consistently linked with prostate cancer (PCa) recurrence and mortality, though the mechanism is unknown. Impaired glucose regulation, which is common among obese individuals, has been hypothesized as a potential mechanism for PCa tumor growth. In this study, we explore the relationship between serum glucose at time of treatment and risk of PCa recurrence following initial therapy.
The study group comprised 1734 men treated with radical prostatectomy (RP) or radiation therapy (RT) for localized PCa between 2001–2010. Serum glucose levels closest to date of diagnosis were determined. PCa recurrence was determined based on PSA progression (nadir PSA+2 for RT; PSA⩾0.2 for RP) or secondary therapy. Multivariate Cox regression was performed to determine whether glucose level was associated with biochemical recurrence after adjusting for age, race, body mass index, comorbidity, diagnosis of diabetes, Gleason Sum, PSA, treatment and treatment year.
Recurrence was identified in 16% of men over a mean follow-up period of 41 months (range 1–121 months). Those with elevated glucose (⩾100 mg/dl) had a 50% increased risk of recurrence (HR 1.5, 95% CI: 1.1–2.0) compared with those with a normal glucose level (<100 mg/dl). This effect was seen in both those undergoing RP (HR 1.9, 95% CI: 1.0–3.6) and those treated with RT (HR 1.4, 95% CI: 1.0–2.0).
Glucose levels at the time of PCa diagnosis are an independent predictor of PCa recurrence for men undergoing treatment for localized disease.
Obesity appears to confer an increased risk of prostate cancer (PCa)-specific mortality.1, 2, 3, 4, 5, 6 Considering the epidemic of obesity in the United States, understanding the mechanisms between obesity and poor cancer prognosis holds great promise for improving patient outcomes. One potential link is impaired glucose regulation, which is a precursor to type 2 diabetes mellitus (DM), a commonly co-occurring condition with obesity.7, 8
There are several pathways by which alterations in glucose metabolism may lead to cancer progression. Glucose is required by cancer cells for growth,9 and cancer cells are known to use more glucose than non-cancerous cells. Hyperglycemia leads to elevated insulin levels, and the activation of the insulin/insulin-like growth factor pathway has been implicated in PCa growth through positive effects on cellular proliferation and anti-apoptosis.10 Hyperglycemia also leads to the production of advanced glycosylated end products11 with resultant oxidative stress that can lead to DNA damage. Finally, chronic inflammation is seen with DM and the metabolic syndrome and results in release of several cytokines that can promote tumor growth.12
Although a diagnosis of DM (self-reported or by diagnosis code) has been associated with poor PCa outcomes in some studies, this finding has not been consistent.13, 14, 15 These mixed findings may be due to imprecision in assessment of DM, which typically uses a diagnostic code and does not adequately describe the severity of DM within an individual. Limited work investigated hemoglobin A1C (HbA1C)14, 16 levels as more specific measures of severity of DM and glucose control. Whereas HbA1C levels are typically only obtained for patients with DM, glucose is a commonly collected test that is more widely available for analyses. Studies in breast and colon cancer have found that higher blood glucose levels are associated with a greater risk of disease progression,17, 18, 19 though such studies have not been performed in PCa. We hypothesize that impaired glucose regulation, represented by higher serum glucose levels, is associated with a higher risk of recurrence after primary PCa treatment. In this study, we explore the relationship between serum glucose at time of PCa diagnosis and risk of biochemical recurrence for men with localized PCa undergoing primary therapy.
Materials and Methods
The Northwest Veterans Integrated Services Network (VISN 20) electronic medical record was used to identify eligible men with PCa. VISN 20 encompasses Washington, Oregon, Alaska, Idaho, western Montana and northern California and includes 7 primary medical centers and 27 community-based outpatient clinics. VISN 20 uses a shared regional electronic medical record system and the data for this study were extracted from the VISN 20 regional data warehouse, a system of relational databases that pulls nearly 100% of the information contained in the electronic medical record (including clinic appointments, inpatient stays, pharmacy records, laboratory values, pathology reports, imaging tests and vital measures).20, 21 Institution review board approval was obtained.
We identified men diagnosed with localized prostate cancer during the years 2001–2010 who were treated with either radical prostatectomy (RP) or radiation therapy (RT) in the form of brachytherapy or external beam radiation therapy. We excluded men treated with active surveillance/watchful waiting and men with distant metastases at diagnosis. We also excluded men who received primary androgen-deprivation therapy (ADT) or long-term continuous ADT as part of their primary treatment as ADT has an impact on glucose regulation.22 In order to ensure that complete follow-up would be likely within the Veterans Affairs (VA) system, we included only those men who had an established relationship with a primary care provider in VISN20 prior to diagnosis and had a previously assessed PSA value measured within the network.
Data collection and coding
The primary exposure of interest was serum glucose closest to the date of diagnosis. DM was coded based on provider-recorded ICD9 diagnosis codes in the 12-month period prior to diagnosis. Recurrence following RP was defined as any PSA, 6 months or more after surgery, ⩾0.2 ng ml−1 (after achieving an undetectable PSA post-RP). For RT patients, PSA recurrence was defined by nadir PSA+2.0 ng ml−1. Initiation of salvage therapy was also considered as evidence of progression (salvage RT, salvage RP or salvage ADT). Salvage RT was defined as radiation received >1 year after primary treatment date. Several covariates were also identified from the EMR including demographic characteristics (age, race), clinical measures (body mass index (BMI), Charlson Comorbidity Index23 scores in the 12-month period prior to diagnosis) and tumor characteristics (year of diagnosis, Gleason Score, clinical stage (available for all subjects) and pathologic stage (available for subjects treated with RP)).
Descriptive statistics, including χ2 tests, were conducted to compare demographic and clinical characteristics by initial treatment and glucose level. Glucose level was analyzed using three different approaches. First, quartiles of glucose were determined and the lowest quartile served as the referent group. Second, the categories established by the American Diabetes Association (ADA) for normal (<100 mg/dl) impaired glucose (100–125 mg/dl) and diabetes (>125 mg/dl) were used with the normal category serving as the referent group. Finally, a binary analysis was performed using the normal ADA glucose level as the cutoff (<100 vs ⩾100 mg/dl). All three approaches are reported. Patients who may have moved or stopped using VA care were censored at the end of the calendar year of their last visit if they had no VA utilization in the subsequent calendar year. Multivariate Cox proportional hazards regression was performed adjusting for all covariates above (age, race, BMI, Charlson Comorbidity Index, year of diagnosis, treatment type, Gleason Score and tumor stage) included in the analysis. The proportional hazards assumption was assessed by examining the slope of the Schoenfeld residuals. Models including all patients and stratified by treatment were performed. Effect modification by treatment and by the National Comprehensive Cancer Network (NCCN) risk strata24 were evaluated with the likelihood ratio test comparing the full model (with the interaction term) with the reduced model. Two additional analyses were performed: first, excluding those with a clinical diagnosis of DM and/or a glucose level that would qualify for the diagnosis of DM by the ADA (>125 mg/dl), and second, excluding patient with high risk disease who did not receive concomitant ADT, as this may have an impact on the risk of disease recurrence. All statistical analyses were conducted using STATA software, version 12 (Stata, Inc., College Station, TX, USA).
A total of 1734 men were identified during the study period undergoing treatment with either RP (n=722) or RT (n=1012). Table 1 presents the demographic and pathologic features of these men by disease recurrence status. As expected, pathologic factors (PSA, Gleason and stage were highly associated with recurrence. Several differences between men undergoing RP or RT were identified (data not shown). Men treated with RT were older (15 vs 1% over age 75, P<0.001) had higher Charlson Comorbidity scores (P<0.001) and were more likely to have been diagnosed with DM prior to diagnosis (P<0.001). Men receiving RT had higher pretreatment PSA values compared with the RP group (35%⩾10 ng ml−1 vs 19%, P<0.001). The median follow-up time was 41 months (range 1–121 months). Recurrence was observed in 281 men (16%). Cumulative recurrence was similar following RT (15%) and surgery (17%). Recurrence events occurred in 7, 14 and 33% of low, intermediate and high risk cases (as defined by the NCCN),24 respectively.
Table 2 shows the distribution of demographic and pathologic features by different ADA glucose levels. A total of 514 (30%), 623 (36%) and 597 (34%) had glucose levels of <100, 100–125 and>125 mg/dl, respectively. Men in the highest glucose category were more commonly obese, with higher Charlson scores and having a diagnosis of diabetes mellitus (all P<0.001). Gleason Sum Scores of 7–10 were more common in the highest category of glucose (68%) compared with 57 and 59% for the lower two categories, respectively (P=0.001). Similarly, higher clinical stage (T2b–T3) was more common in those with glucose >125 mg/dl (23%) than those with lower glucose levels (16 and 17% respectively, P=0.04). Surgery was more common in those with the highest glucose levels.
The hazard ratios (HR) and 95% CI for risk of recurrence by quartile of glucose level are presented in Table 3 for the entire cohort. The lowest quartile (range 31–98 mg/dl) served as the referent group in the multivariate analysis adjusting for age, race, BMI, diagnosis of DM, treatment type and year of treatment, stage, Gleason and Charlson Score. In the overall cohort, we observed a modest trend in increased risk (range 35–57%) of recurrence for each of the glucose quartiles compared with the lowest quartile, though this was not significant. There was no evidence for effect modification by treatment or NCCN risk category (both likelihood ratio test P-value>0.6). In the RP group, all three higher quartiles had a more than two-fold increased risk of recurrence compared with the lowest quartile, with P-values ranging from 0.01–0.04, data not shown). In the RT group, there were non-significant differences in risk of recurrence with higher glucose levels (range 18–48%). The risk of recurrence by the three ADA glucose categories is presented in Table 4 along with dichotomous ADA categories. In the entire cohort, those with an abnormal glucose (>100 mg/dl) had a 46% increased risk of recurrence (HR: 1.46: 95% CI 1.09–1.95). This observed risk was present in both the RP group (HR 1.91, 95% CI 1.03–3.55) and the RT group (HR 1.42, 95% CI 1.00–2.01). Excluding those with either a clinical diagnosis of DM or a serum glucose that would qualify for the diagnosis of DM by the ADA (>125 mg/dl) did not change the results. In that analysis including both RP and XRT patients, the presence of a glucose level of 100–125 mg/dl was associated with a 44% increased risk of recurrence compared with those with a normal glucose level (HR 1.41, 95% CI: 1.01–1.97).
In this study, we found that men with elevated glucose levels at the time of PCa diagnosis have a greater risk of PCa recurrence following primary treatment compared with men with a normal glucose level. Considering the growing obesity epidemic and the consistent relationship between obesity and adverse PCa outcomes,7, 8 these findings are important in helping to identify modifiable risk factors for men undergoing treatment for newly diagnosed PCa.
Glucose and cancer outcomes have been evaluated in breast and colorectal cancer.17, 18, 19, 25 In a study of 202 women with breast cancer, there was a similar increased risk of recurrence in the second (HR 1.9, 95% CI: 1.1–3.2) and third (HR 1.8 (95% CI: 1.0–3.1) tertiles compared with the lowest tertile.17 In colorectal cancer, those with the highest quartile of glucose (HR 1.8, 95% CI 1.1–3.1)19 or post-load plasma glucose (HR 1.6, 95% CI 1.1–2.4)18 had higher risk of disease-specific mortality compared with those in the lowest quartile. These studies support the role of abnormal glucose regulation and cancer outcomes.
Small studies have evaluated serum biomarkers of glucose regulation and PCa. In a study of 89 men with DM treated with RP, a higher HbA1C pre-operatively was pre-operatively associated with the risk of extracapsular disease (pT3) and higher Gleason grade (⩾ 7).26 In another study of 247 men undergoing RP, HbA1C levels pre-operatively were associated with higher Gleason grade (P=0.001).16 Higher HbA1C levels were associated with a non-significant risk of recurrence: compared with those with HbA1C<6.3, the HR for men with HbA1C of 6.3–7.7 and⩾7.8 was 1.5 (95% CI 0.8–2.6) and 1.3 (0.7–2.4), respectively.16 Both of these studies are limited by their small size and only include men with DM. Studies have found associations between insulin levels and PCa-specific mortality.1, 27 In one study, a higher mean fasting insulin level at diagnosis was found in those who died of PCa.27 In a second study, levels of C-peptide (an insulin surrogate) were determined from PCa cases from the Physicians Health Study.1 Those in the highest C-peptide quartile had a more than two-fold increased risk of PCa-specific mortality (HR 2.4, 95% CI: 1.3–4.3) compared with those in the lowest C-peptide quartile. We do not have insulin or C-peptide levels available in our study and thus rely on serum glucose levels as an indicator of an individual’s glucose regulation.
We believe the present analysis is the first to specifically evaluate glucose levels and outcomes in men with PCa. We found that men with an elevated glucose at the time of PCa diagnosis had ∼50% increased risk of PCa recurrence even after adjusting for clinical and pathologic factors. Whether it is hyperglycemia or subsequent elevated levels of insulin/insulin-like growth factor which affects tumor recurrence cannot be determined from this study. Studies have not found a significant relationship between the diagnosis of DM (self-reported or diagnosis code)13, 14, 15 and PCa recurrence, but the diagnosis of DM may not reflect an individual’s current glycemic control. This could lead to misclassification as men with well-controlled DM may have glucose and insulin levels similar to men without DM and vice versa. Thus, using a serum biomarker is more sensitive for accurately determining the relationship between glucose regulation and disease outcomes. Our analysis is also unique in that we include both surgical and radiation patients. Although the magnitude of influence with hyperglycemia on recurrence is less in the XRT group, the direction of the effect is the same and there was no evidence for effect modification by treatment. Potential reasons for this attenuation include regression to the mean, differences in biochemical recurrence definitions, unmeasured confounding and the concomitant use of ADT in the radiation group alone.
There are limitations to our study. We only use a single glucose measurement that may not reflect an individual’s ongoing glucose regulation during their post-treatment time. A HbA1C measurement does represent one’s recent glucose regulation; however, such values were only available in those with DM, and similar to the pre-operative glucose level, will not necessarily reflect subsequent glucose regulation. Further, we cannot be certain that these glucose levels were drawn fasting; however, the median glucose for men without the diagnosis of DM was 107 (interquartile range 96–126). Data from the Center for Disease Control and Prevention Diabetes Fact Sheet 2011, state that 35% of adults aged >20 (and 50% of adults aged >65) had prediabetes based on fasting glucose levels.28 Thus our data, in which 72% of the patients are >60 years of age and 82% of patients are overweight or obese, are consistent with this distribution. Our population reflects veterans who are consistent users of VA care and may not be generalizable to other men. As the VA is an open system, some men may go to non-VA providers for follow-up care that may not get reported back to the VA. This appears to happen infrequently, a prior study of men with elevated PSA tests, 8% of men received subsequent care outside of the VA29; we specifically restricted our cohort to men who had an established primary care relationship prior to diagnosis to help reduce this potential limitation. Our definitions of biochemical recurrence may misclassify some men as not having evidence of recurrence and we cannot rule out type 1 error in this analysis. We also do not include medication usage for DM treatment that may have an impact on glucose levels and we cannot rule out additional sources of unmeasured confounding. Future study of medication usage for DM may be informative. Finally, recurrence after PCa is a heterogeneous event with many individuals not progressing to PCa-specific mortality. Although PCa-specific mortality is a more robust endpoint, we did not have enough events to study this relationship. Use of national VA data in the future may allow such an analysis.
In conclusion, men with an elevated glucose level at the time of treatment for localized PCa have a higher risk of recurrence than men with a normal glucose level. This analysis provides evidence that disease progression may be influenced by glucose, suggesting that glucose control could be a modifiable risk factor for PCa recurrence and progression.
Ma J, Li H, Giovannucci E, Mucci L, Qiu W, Nguyen PL et al. Prediagnostic body-mass index, plasma C-peptide concentration, and prostate cancer-specific mortality in men with prostate cancer: a long-term survival analysis. Lancet Oncol 2008; 9: 1039–1047.
Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ . Overweight, obesity, and mortality from cancer in a prospectively studied cohort of US adults. N Engl J Med 2003; 348: 1625–1638.
Andersson SO, Wolk A, Bergstrom R, Adami HO, Engholm G, Englund A et al. Body size and prostate cancer: a 20-year follow-up study among 135006 Swedish construction workers. J Natl Cancer Inst 1997; 89: 385–389.
Wright ME, Chang SC, Schatzkin A, Albanes D, Kipnis V, Mouw T et al. Prospective study of adiposity and weight change in relation to prostate cancer incidence and mortality. Cancer 2007; 109: 675–684.
Efstathiou JA, Bae K, Shipley WU, Hanks GE, Pilepich MV, Sandler HM et al. Obesity and mortality in men with locally advanced prostate cancer: analysis of RTOG 85-31. Cancer 2007; 110: 2691–2699.
Gong Z, Agalliu I, Lin DW, Stanford JL, Kristal AR . Obesity is associated with increased risks of prostate cancer metastasis and death after initial cancer diagnosis in middle-aged men. Cancer 2007; 109: 1192–1202.
What are the risk factors for prostate cancer? American Cancer Society, Inc. http://www.cancer.org/Cancer/ProstateCancer/DetailedGuide/prostate-cancer-risk-factors., Accessed 2012.
Chornokur G, Dalton K, Borysova ME, Kumar NB . Disparities at presentation, diagnosis, treatment, and survival in African American men, affected by prostate cancer. Prostate 2011; 71: 985–997.
Osborne CK, Bolan G, Monaco ME, Lippman ME . Hormone responsive human breast cancer in long-term tissue culture: effect of insulin. Proc Natl Acad Sci USA 1976; 73: 4536–4540.
Pollak M . Insulin and insulin-like growth factor signalling in neoplasia. Nat Rev Cancer 2008; 8: 915–928.
Shinohara M, Thornalley PJ, Giardino I, Beisswenger P, Thorpe SR, Onorato J et al. Overexpression of glyoxalase-I in bovine endothelial cells inhibits intracellular advanced glycation endproduct formation and prevents hyperglycemia-induced increases in macromolecular endocytosis. J Clin Invest 1998; 101: 1142–1147.
DeNardo DG, Andreu P, Coussens LM . Interactions between lymphocytes and myeloid cells regulate pro- versus anti-tumor immunity. Cancer Metastasis Rev 2010; 29: 309–316.
Chan JM, Latini DM, Cowan J, Duchane J, Carroll PR . History of diabetes, clinical features of prostate cancer, and prostate cancer recurrence-data from CaPSURE (United States). Cancer Causes Control 2005; 16: 789–797.
Jayachandran J, Aronson WJ, Terris MK et al. Diabetes and outcomes after radical prostatectomy: are results affected by obesity and race? Results from the shared equal-access regional cancer hospital database. Cancer Epidemiol Biomarkers Prev 2010; 19: 9–17.
Smith MR, Bae K, Efstathiou JA, Hanks GE, Pilepich MV, Sandler HM et al. Diabetes and mortality in men with locally advanced prostate cancer: RTOG 92-02. J Clin Oncol 2008; 26: 4333–4339.
Kim HS, Presti JC, Aronson WJ, Terris MK, Kane CJ, Amling CL et al. Glycemic control and prostate cancer progression: results from the SEARCH database. Prostate 2010; 70: 1540–1546.
Barba M, Sperati F, Stranges S, Carlomagno C, Nasti G, Iaffaioli V et al. Fasting glucose and treatment outcome in breast and colorectal cancer patients treated with targeted agents: results from a historic cohort. Ann Oncol 2011; 23: 1838–1845.
Colangelo LA, Gapstur SM, Gann PH, Dyer AR, Liu K . Colorectal cancer mortality and factors related to the insulin resistance syndrome. Cancer Epidemiol Biomarkers Prev 2002; 11: 385–391.
Trevisan M, Liu J, Muti P, Misciagna G, Menotti A, Fucci F . Markers of insulin resistance and colorectal cancer mortality. Cancer Epidemiol Biomarkers Prev 2001; 10: 937–941.
Maynard C, Chapko MK . Data resources in the Department of Veterans Affairs. Diabetes Care 2004; 27 (Suppl 2): B22–B26.
Zeliadt SB, Sekaran NK, Hu EY, Slatore CC, Au DH, Backhus L et al. Comparison of demographic characteristics, surgical resection patterns, and survival outcomes for veterans and nonveterans with non-small cell lung cancer in the Pacific Northwest. J Thorac Oncol 2011; 6: 1726–1732.
Basaria S, Muller DC, Carducci MA, Egan J, Dobs AS . Hyperglycemia and insulin resistance in men with prostate carcinoma who receive androgen-deprivation therapy. Cancer 2006; 106: 581–588.
Klabunde CN, Legler JM, Warren JL, Baldwin LM, Schrag D . A refined comorbidity measurement algorithm for claims-based studies of breast, prostate, colorectal, and lung cancer patients. Ann Epidemiol 2007; 17: 584–590.
National Comprehensive Cancer Network. Prostate Cancer (Version 3.2012). Accessed Date May 2012. URL http://www.nccn.org/professionals/physician_gls/pdf/prostate.pdf.
Flood A, Mai V, Pfeiffer R, Kahle L, Remaley AT, Lanza E et al. Elevated serum concentrations of insulin and glucose increase risk of recurrent colorectal adenomas. Gastroenterology 2007; 133: 1423–1429.
Hong SK, Lee ST, Kim SS, Min KE, Byun SS, Cho SY et al. Significance of preoperative HbA1c level in patients with diabetes mellitus and clinically localized prostate cancer. Prostate 2009; 69: 820–826.
Hammarsten J, Hogstedt B . Hyperinsulinaemia: a prospective risk factor for lethal clinical prostate cancer. Eur J Cancer 2005; 41: 2887–2895.
Centers for Disease Control and Prevention. National diabetes fact sheet: national estimates and general information on diabetes and prediabetes in the United States. US Department of Health and Human Services, Center for Disease Control and Prevention: Atlanta, GA, 2011.
Zeliadt SB, Hoffman RM, Etzioni R, Ginger VA, Lin DW . What happens after an elevated PSA test: the experience of 13 591 veterans. J Gen Intern Med 2010; 25: 1205–1210.
This study was supported by NIH grants: P50CA097186 from the National Cancer Institute; with additional support from the Fred Hutchinson Cancer Research Center. This material is the result of work supported by resources from the VA Puget Sound Health Care System, Seattle, Washington.
The authors declare no conflict of interest.
About this article
Cite this article
Wright, J., Plymate, S., Porter, M. et al. Hyperglycemia and prostate cancer recurrence in men treated for localized prostate cancer. Prostate Cancer Prostatic Dis 16, 204–208 (2013). https://doi.org/10.1038/pcan.2013.5
- biochemical recurrence
- veterans’ affairs
Journal of Molecular Medicine (2021)
The Adverse Effects of Androgen Deprivation Therapy in Prostate Cancer and the Benefits and Potential Anti-oncogenic Mechanisms of Progressive Resistance Training
Sports Medicine - Open (2020)
Pre-diagnostic carbohydrate intake and treatment failure after radical prostatectomy for early-stage prostate cancer
Cancer Causes & Control (2019)
Journal of Bioenergetics and Biomembranes (2018)
Metabolic Syndrome and Prostate Cancer: a Review of Complex Interplay Amongst Various Endocrine Factors in the Pathophysiology and Progression of Prostate Cancer
Hormones and Cancer (2016)