A systematic review of the literature exploring the interplay between prostate cancer and type two diabetes mellitus

Danielle Crawley¹, Florence Chamberlain², Hans Garmo¹, Sarah Rudman², Björn Zethelius^3,4, Lars Holmberg^1,5, Jan Adolfsson⁶, Par Stattin⁵, Paul Carroll⁷ and Mieke Van Hemelrijck¹

¹Translational Oncology and Urology Research Group, King’s College London, London SE1 9RT, UK

²Department of Medical Oncology, Guy’s and St Thomas’ NHS Foundation Trust, London SE1 9RT, UK

³Department of Public Health and Geriatrics, Uppsala University, Uppsala 751 05, Sweden

⁴Medical Products Agency, Uppsala 751 03, Sweden

⁵Department of Surgical Sciences, Uppsala University, Uppsala 751 05, Sweden

⁶Department of Clinical Science, Intervention and Technology, Karolinska Institutet, Stockholm 171 77, Sweden

⁷Department of Diabetes and Endocrinology, Guy’s and St Thomas’ NHS Foundation Trust, London, SE1 9RT UK

Correspondence to: Danielle Crawley. Email: Danielle.crawley@kcl.ac.uk

Abstract

Prostate cancer (PCa) and type two diabetes mellitus (T2DM) are both increasing prevalent conditions and often occur concurrently. However, the relationship between the two is more complex than just two prevalent conditions co-existing. This review systematically explores the literature around the interplay between the two conditions. It covers the impact of pre-existing T2DM on PCa incidence, grade and stage, as well as exploring the impact of T2DM on PCa outcomes and mortality and the interaction between T2DM and PCa treatments.

Keywords: type two diabetes, prostate cancer, review

Copyright: © the authors; licensee ecancermedicalscience. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Published: 25/01/2018; Received: 23/10/2017

Introduction

Prostate cancer (PCa) remains the commonest cancer in men, affecting one in eight men in the UK [1]. There are also 3.2 million people in the UK who have been diagnosed with type 2 diabetes mellitus (T2DM) and it is estimated that this will rise to 5 million by 2025 [2]. PCa and T2DM thus often occur together in the same individual

T2DM increases the risk of cancer-specific death from several solid malignancies, including colorectal and breast cancer [3], but conflicting evidence exists in the case of PCa. The impact of pre-existing T2DM has also been studied in regard to grade and stage of PCa at presentation, with conflicting results [4]. Moreover, there is emerging evidence that the presence of T2DM and other metabolic abnormalities (dyslipidaemia, hypertension, obesity) is associated with a more rapid progression of PCa [5, 6]. This relationship is further complicated by the fact that standard treatment for advanced PCa, androgen deprivation therapy (ADT), has been suggested to increase incidence of T2DM [7],as well as worsen glycaemic control in those with pre-existing T2DM.

To provide a background on the complex association between PCa and T2DM, this review is set out to explore the following areas:

1. Impact of pre-existing T2DM on PCa incidence

2. Impact of pre-existing T2DM on PCa grade and stage

3. Impact of pre-existing T2DM on PCa outcomes and mortality

4. Interplay between T2DM and PCa treatments

Impact of pre-existing T2DM on PCa incidence

T2DM that increases the risk of some solid malignancies is now well established, although not all the literature is robust [8]. Some studies suggest that people with T2DM are as much as twice as likely to die from cancer than those without [3]. However, the opposite is seen in PCa, with an inverse association reported in several published meta-analyses [8].

Bonovas et al. published the first meta-analysis examining T2DM and risk of PCa in 2004. They included 14 studies and concluded that T2DM confers a statistically significant 9% decrease in relative risk of developing PCa [9]. This was followed in 2006 by a meta-analysis conducted by Kasper et al. which included 19 studies and reported an inverse relationship of a similar magnitude, relative risk (RR): 0.84 (95%CI: 0.76–0.93) [10]. Following this, Bansal et al. published an updated meta-analysis including 45 studies, involving 8.1 million participants and 132,331 PCa cases, which also reports an inverse association with an RR of 0.86 (95% CI 0.80–0.92) [11]. More recently, Gang et al. published a further updated meta-analysis reviewing the literature up to and including April 2012. This meta-analysis included 56 studies and also reported an inverse association, RR: 0.88 (95%CI: 0.82–0.93) [12]. Here, we perform a further systematic review of the literature up until June 2017.

Evidence acquisition

The systematic review was performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [13], with search terms, inclusion and exclusion criteria all defined a priori.

Search strategy

A computerised literature search of Pubmed to identify full text and abstracts published was performed. The search was done with and without MESH terms (diabetes, diabetes mellitus, PCa, prostate neoplasm, incidence, risk). All references of the selected articles were checked, including hand searches.

Study eligibility

The final articles were chosen based on the following set of inclusion criteria:

• examined association of T2DM with PCa incidence/risk

• case control or cohort study

• English Language

• not included in the prior published meta-analyses described above

Articles were excluded if they:

• examined association of T2DM with PCa mortality

• examined association of T2DM treatments (i.e. drugs) and PCa incidence

• were a review article or meta-analysis

Initially, titles were reviewed to assess whether they met inclusion criteria. If, after assessing the abstract, there was any doubt regarding whether it met the relevant criteria, it was kept for more thorough, subsequent assessment. The list of potential articles was further shortened by performing detailed evaluations of the methods and results of each remaining paper. Figure 1 provides more detailed information regarding the exclusion process. The strength of each study was assessed using the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria [14] and is shown in Table 1.

Data collection

The following details were recorded for each study: author, year of publication, country where study was undertaken, study design, number of patients, population/setting, outcome reported and variables adjusted for in the analysis.

Figure 1. Preferred reporting items for systematic reviews and meta-analysis (PRISMA) flow diagram of article identification, screening, eligibility and inclusion for systematic review on impact of T2DM on PCa incidence.

Table 1. Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist for observational studies included in the systematic review on impact of T2DM on PCa incidence.

Evidence synthesis

The literature search identified a total of 896 studies, of which 44 were deemed initially relevant. Using the above inclusion and exclusion criteria, 36 were excluded (Figure 1). The reasons for exclusion were included in previously published meta-analysis (n = 22), only T2DM patients included (n = 6), full article not available (n = 3), outcome not PCa incidence (n = 2), meta-analysis (n = 1), duplicate data (n = 1), genetic variant of T2DM examined only (n = 1) (Figure 1). A total of eight studies were included in the systematic review (Table 2).

Of these eight studies, five were cohort studies [15–20] and three case-control studies [21, 22]. Three studies were from European populations, two from the USA, two from Israel and one from Australia.

The studies combined included 2,716,302 subjects. Six reported an inverse association between T2DM and PCa incidence [15–18, 20, 22], one reported no association [21] and in one a positive association was reported [19]. Of the six studies reporting an inverse association, they all reported similar measures of association in the magnitude of a 20% reduction in risk of PCa in those exposed to T2DM as compared to those without T2DM. The one paper which reports no statistically significant association is a case-control study in which cases of T2DM were patients enrolled in the Freemantle Diabetes Study [21]. It is the smallest of the studies, with 1,289 cases and 5,156 controls – which may account for the non- statistical significance of their findings (HR: 0.83, 95%CI: 0.60–1.14) – though the direction and magnitude of the association reported is in line with the other studies. The study which reported a positive association is a retrospective review on 3,162 consecutive men who underwent a prostate biopsy due to either an elevated PSA and/or an abnormal DRE [19]. This design is different from the other studies included here, which were largely based on the general population, not on a selected population attending for a prostate biopsy. This heterogeneity in the design may account for the findings of a 26% increased odds of a positive biopsy in patients with T2DM, compared to those without (OR: 1.26 95%CI: 1.01–1.55).

Table 2. Characteristics of the eight studies included in the systematic review on impact of T2DM on PCa incidence.

Discussion

The biological mechanism underlying the inverse association between T2DM and PCa risk is not elucidated. First, several metabolic alterations occur in people with T2DM which may protect from PCa. The Insulin-IGF-1 theory of carcinogenesis suggests that prolonged hyperinsulinaemia results in reduced insulin binding proteins and therefore increased free IGF-1, which results in cellular changes that can lead to carcinogenesis via increased mitosis and decreased apoptosis [23]. There is both laboratory and epidemiological evidence supporting that raised insulin levels are associated with increased PCa risk [24, 25]. Patients with T2DM, though initially may have raised insulin levels, over time develop hypoinsulinaemia. Hence, patients with T2DM who have lower levels of insulin over time would be protected in terms of PCa risk [26]. Several studies have reported a strengthening of the inverse association between T2DM and PCa risks with the duration of T2DM [27–29], which serves to strengthen this hypothesis. Prolonged hypoinsulinaemia may also result in a reduction of leptin, a hormone involved in energy homeostasis [30], raised levels of which have been associated with PCa risk [31, 32]. However, there are no published studies specifically examining this relationship between insulin and leptin levels in T2DM and risk of PCa.

Another metabolic change which occurs in T2DM is a reduction in testosterone levels which has been shown both in animal models and in vitro [33, 34]. PCa is testosterone driven [35]; therefore, a decrease in testosterone is expected to be associated with a decreased risk.

Genetic factors as well as metabolic changes have also been postulated to be involved in the protective effect which T2DM appears to have on PCa risk. The TCF2 gene confers a predisposition to T2DM and has also been shown to have a potential protective effect in PCa. Similarly, other studies have identified different variants in the JAZF1 gene, one associated with T2DM and another associated with PCa [36].

Some cross-sectional studies have shown that men with T2DM have lower PSA levels, compared to those without T2DM [37] and the rate of change over time is also lower [38]. This could result in less screen detected PCa and at least in part account for the difference in the risk seen. This is supported both by studies in which enrolled participants undergo prostate biopsy, which report increased risk of positive biopsies in T2DM [19], and by those which show higher grade PCa detected in those with T2DM [4].

Finally, treatments used for T2DM, including metformin, could be potential confounders in the association between T2DM and PCa risks, which is discussed later in this review.

Conclusion

The updated systematic review of the literature examining the association between T2DM and PCa risks presented here concurs with the previously published findings of several meta-analyses, indicating that T2DM has a protective effect on PCa risk. However, the underlying biological mechanisms are yet to be elucidated.

Impact of pre-existing T2DM on PCa grade and stage

PCa severity can be described in terms of its grade and its clinical or TNM stage. Some studies have examined whether pre-existing T2DM is associated with a particular grade or stage of disease. There is considerable overlap with the literature presented above examining T2DM and PCa incidence and that examining particular stage and grade of PCa. In the systematic review of T2DM and PCa incidence above which includes studies published after 2012, of the eight new studies identified, only two presented subgroup analysis on stage and/or grade [15, 18] (Table 3). These are discussed below but another systematic review was not deemed informative, as one including literature up until 2013 has previously been published [4].

Existing literature

In 2013, Xu et al. undertook a meta-analysis of all studies examining the association between T2DM and PCa risks including subgroup analysis by different grade and stage [4]. They included nine studies: five examining stage only, two grade only and two which explored both. They reported findings of an inverse association between T2DM and PCa for both low- and high-grade PCa, defined as Gleason 2–6 and Gleason 7–10 (RR: 0.74, 95%CI: 0.64–0.86 and 0.78, 95%CI: 0.67–0.90). They reported an RR of a similar direction and magnitude for localised and advanced disease (RR: 0.72, 95%CI: 0.68–0.76 and 0.85, 95%CI: 0.75–0.97). This meta-analysis included all studies published up until October 2012.

Table 3. Overview of the eight papers included in the systematic review on T2DM and PCa incidence, by subgroup analysis including PCa stage and grade.

Updated literature review

Two new studies were identified in the systematic review described above which considered stage and grade of PCa. A nested case-control study by Fall et al. included 44,352 men with PCa in Prostate Cancer data Base Sweden (PCBaSe) Sweden, which is based on the National Prostate Cancer register of Sweden (NPCR) [39]. They showed an inverse association between T2DM and risk of PCa across all risk groups, low, intermediate and high risk/metastatic (OR: 0.71, 95%CI: 0.64–0.80; 0.76, 95%CI: 0.69–0.84; 0.86, 95%CI: 0.80–0.93, respectively). Although they showed a slightly less clear risk pattern for those with high risk and metastatic disease, no significant difference between T2DM and risk category of PCa emerged from this study.

Tsilidis et al. included 139,131 men from the European Prospective Investigation into Cancer and Nutrition (EPIC) prospective cohort, 4,531 of whom went on to develop PCa. They reported no statistical evidence for an inverse association between T2DM and PCa risks. There was no evidence that the association differed by stage (p-heterogeneity, 0.19) or grade (p-heterogeneity, 0.48) of the disease, although the numbers were small in some subgroups and the study may have been under powered to detect differences.

Conclusion

The meta-analysis by Xu et al. [4] concluded that the protective effect of T2DM on PCa risk is seen across different disease grades and stages; however, the number of studies available was too small to confirm this finding. The two new studies identified in the systematic review of T2DM and PCa presented above, which considered grade or stage [15, 18], reported similar findings of no significant difference between different stages and grades of PCa. However, sample sizes remain small and no definitive conclusion can be drawn on the impact of T2DM on risk of PCa of different grades and stages. Larger studies are required to address this question in detail.

Impact of pre-existing T2DM on prostate cancer outcomes and mortality

Introduction

In 2008, a systematic review and meta-analysis examining all-cause mortality in cancer patients reported worse outcomes in cancer patients who had pre-existing diabetes mellitus (DM) [40]. However, the magnitude of the association varied widely between different types of malignancy. This led to a call for research focusing on individual cancer types. Following this, a body of literature has emerged and three published meta-analyses examined the association between pre-existing DM and PCa specific and all-cause mortality [41–43]. The most recent of which was published in 2016. Much of this literature does not distinguish between type 1 and type 2 diabetes; hence in this section refers to DM, encompassing both types, unless otherwise specified.

Existing literature

In 2010, Synder et al. performed a systematic review of the literature (seven papers), but was only able to include four studies in a meta-analysis. They could only investigate all-cause mortality and reported a pooled HR of 1.57 (95%CI: 1.12–2.20) for DM versus no DM. They concluded that more rigorous research was necessary before firm conclusions could be drawn.

Subsequently, a further meta-analysis by Cai et al. [42] was published in 2015, which included 11 cohort studies and looked at outcomes of all-cause mortality, PCa-specific mortality and non-PCa mortality. It reported that DM was positively associated with all three outcomes, with pooled HR of 1.50 (95%CI: 1.25–1.79), 1.26 (95%CI: 1.20–1.33), 1.83 (95%CI: 1.33–2.52), respectively. They concluded that DM was associated with an adverse prognosis in PCa and that clinicians treating patients with both conditions should pay more attention to the dual diagnosis and even consider more aggressive treatment strategies.

The final and most recent meta-analysis by Lee et al. [43] included 17 cohort studies which included 274,677 men. The studies were mainly from the USA (eight) and Europe (six), with two from Taiwan and one from Korea. They reported a 29% increase in PCa-specific mortality (95%CI 1.37–2.96), alongside a 37% increase in all-cause mortality (95%CI: 1.29–1.45). They performed a subgroup analysis of three cohort studies which considered T2DM separately from type 1 DM. This analysis showed a twofold increase in all-cause mortality for those with T2DM as compared to those without DM (95%CI:1.37–2.96), and could not exclude a positive association with PCa-specific mortality (RR:1.17, 95%CI: 0.96–1.42).