Diabetes and prostate cancer are both prevalent conditions in older men; however, little research has been conducted to evaluate the impact of pre-existing diabetes on prostate cancer treatment and outcomes. We conducted a systematic review to summarize the research to date on how diabetes affects prostate cancer mortality and morbidity. We also conducted a preliminary meta-analysis to quantify the effects of diabetes on long-term, overall mortality. While only four studies could be included in this meta-analysis, resulting in a pooled hazard ratio of 1.57 (95% CI: 1.12-2.20), perhaps of greater importance is the lack of evidence to conduct even a preliminary meta-analysis of the impact of diabetes on prostate cancer-specific mortality. Because diabetes is associated with worse mortality in general (aside from any cancer diagnosis), it is particularly important to investigate diabetes’ influence on prostate cancer-specific outcomes, and future research should focus on this area. Studies conducted to date indicate mixed results regarding the impact of diabetes on prostate cancer-specific and non-prostate cancer mortality. Additional research is critical to inform our understanding of whether the difference found in overall mortality is cancer-related, non-cancer related, or both. Diabetes was also associated with differences in treatment selection, complications, and recurrence.
What might explain an adverse biological interaction between diabetes mellitus and prostate cancer? As discussed in our previous paper,6
an environment of hyperinsulinemia and hyperglycemia might lead to increased tumor proliferation and metastasis. In vitro, insulin appears to be a growth factor for prostatic epithelia.20
Insulin-like growth factor-1 (IGF-1) and IGF binding protein-3 are also associated with prostate growth.21-22
Such findings have prompted the development of IGF receptor inhibitors, and prostate cancer has been cited as a tumor type that may benefit from treatment with such agents. Studies are underway to elucidate further the biology of diabetes mellitus, IGF, and prostate cancer.
There might also be biological interactions with regards to the effects of radiotherapy. Chan et al. found that younger men with diabetes who received radiotherapy had a shorter time to treatment failure.9
Herold et al. found that men with diabetes who underwent radiotherapy had higher rates of gastrointestinal and genitourinary late complications than men without diabetes.13
Potential biologic mechanisms for these worse outcomes among men with diabetes include alterations in IGF, its receptors, and their signaling pathways that may affect resistance to treatment.23
For example, heme oxygenase -1 and clusterin are up-regulated in diabetes mellitus and associated with worse treatment outcomes.24-25
There are at least two additional circumstances in which prostate cancer therapy can adversely interact with pre-existing diabetes. First, androgen-deprivation therapy initiation might aggravate insulin resistance and provoke deterioration in glycemia and/or dyslipidemia. Second, use of corticosteroids (e.g. in combination with docetaxel) can immediately aggravate glycemia and heighten diabetes-related susceptibility to bacterial infection. Further research should explore the biologic mechanisms for differences in treatment outcomes and evaluate whether differences in outcomes are related to selection bias or are related to the treatment itself.
This study benefits from rigorous methods, including a comprehensive, systematic review of the literature. The study team included experts in multiple disciplines, including cancer, diabetes, and epidemiology. All steps of the process involved two levels of review with disagreements addressed by consensus or a third review. Nevertheless, there are some limitations to our study that warrant mention. One challenge was the heterogeneity of studies in terms of the populations included and the outcomes assessed. For the four studies included in the meta-analysis, each used adjusted multivariate Cox proportional hazards models, and we allowed for between-study differences by using a random-effects model. However, two of the studies used population-based cohorts while two studies used cohorts undergoing radiation therapy, so these results should be interpreted with caution. Though we do not know the direction of effect in the two studies we were unable to include due to insufficient reporting, we are reassured by the consistent result of our sensitivity analysis. Variation in study design and other quality components across the 11 studies included in the systematic review may limit conclusions.
Perhaps the most important finding from this review is the lack of research investigating how diabetes affects prostate cancer prognosis. As noted above, only 11 studies addressed this issue. Six of those 11 studies estimated diabetes’ impact on overall, long-term mortality, but only four studies addressed the more important question of whether diabetes affects prostate cancer-specific mortality. Because both prostate cancer and diabetes are common conditions among older men, significantly more research is needed to investigate how diabetes affects prostate cancer management and prognosis and on the proper management of men with prostate cancer who have comorbid diabetes. Important research questions include whether men with diabetes should get different treatments than men without diabetes and what the appropriate treatments are. For example, are men with diabetes indeed more likely to receive radiotherapy? Are the worse outcomes and increased complication rates associated with radiotherapy in men with diabetes due to selection bias or a direct result of the disease? Would a modified radiotherapy regimen for men with diabetes improve the benefit to risk ratio? Can tighter control of diabetes improve survival for men with prostate cancer more generally? In all cases, additional research should use rigorous methods for diabetes ascertainment, sufficient follow-up time, and appropriate adjustment for confounders. The findings of this review and meta-analysis clearly support the need for further research in this area.