Reduced risk of prostate cancer in a cohort of Lithuanian diabetes patients

Background: The aim of this study was to determine whether type 2 diabetes mellitus and treatment with metformin is associated with prostate cancer risk. Methods: Male patients with diagnosis of type 2 diabetes during the period of 2000 – 2016 were identified in the National Health Insurance Fund database. The prostate cancer cases were identified by pooling these records from the national Cancer Registry. The calculation of prostate cancer stardardized incidence ratios (SIRs) was composed as a ratio of observed number of cancer case in people with diagnosis of diabetes to the expected number of cancer cases in the underlying general population. Results: Overall, 64,000 males diagnosed with diabetes in Lithuania between 2000 and 2016 were included in the final cohort. 2,754 prostate cancers were observed versus 3,111.26 expected within the period of observation entailing an SIR of 0.89 (95% CI: 0.85– 0.92). Significantly lower risk of prostate cancer was found in diabetes patients in all age groups, there were no differences in prostate cancer risk by time since diagnosis of diabetes. Significantly lower risk of prostate cancer also was found in both metformin users and never-users’ groups, with higher risk reduction in metformin users (SIR 0.71, 95% CI: 0.68–0.75) than in diabetes patients never-users (SIR 0.88, 95% CI: 0.80–0.96). Conclusions: In this large population-based study we found a significantly decreased risk of prostate cancer among men with diabetes and metformin users. However, the risk of prostate cancer showed no clear trend with diabetes duration and cumulative metformin dose.

Conclusions: In this large population-based study we found a significantly decreased risk of prostate cancer among men with diabetes and metformin users. However, the risk of prostate cancer showed no clear trend with diabetes duration and cumulative metformin dose.

Background
Prostate cancer is a leading cancer in the male patient population across the Western world [1]. Approximately 450.000 men in Europe are diagnosed with prostate cancer and 107000 died from this disease in 2018 [2]. It is thought that such epidemiology in Western countries in part might be associated with changing life style, environment and development of various chronic diseases. During the past decade a huge amount of interest was devoted to the metabolic syndrome and especially type 2 diabetes mellitus (T2DM) and their associations with prostate cancer development [3] [4]. However, results from published studies that analyse T2DM association with prostate cancer are conflicting.
Some have found that T2DM reduces prostate cancer incidence risk by 11% to 44% [5] [6,7] Pooled analysis of 45 studies showed that patients with T2DM were associated with an estimated reduction of 14% in the risk of developing prostate cancer as compared with those without diabetes [8]. Conversely, some other studies lack significant association [9][10][11].
Metformin is the most widely used in the treatment of T2DM and reduces insulin resistance and diabetes-related morbidiaty and mortality. Population-based studies show that metformin treatment is associated with a dose-dependent reduction in cancer risk [12]. Data on the influence of metformin on prostate cancer risk, incidence, growth, and aggressiveness and survival rate are still conflicting and controversial. Some epidemiological studies indicate that metformin decreases prostate cancer incidence [13,14], while other authors do not report any beneficial effect [15,16]. A recently published meta-analysis by Soranna et al. found no association between metformin use and prostate cancer risk in T2DM patients [17]. Therefore, it is still difficult to characterize the exact role of metformin in prostate cancer chemoprevention.
The aim of this analysis was to evaluate whether T2DM and metformin use could be associated with decreased prostate cancer risk in Lithuanian male patients with T2DM.

Dataset
The National Health Insurance Fund (NHIF) database was created in 1999, seeking to reimburse healthcare institutions for the healthcare services provided from the NHIF. This NHIF database is utilized to accumulate, administrate, analyse and report data of all health care services provided across the country. The database encompasses demographic data and all records of patients entries to outpatient clinics, admissions to emergency rooms and hospitals as well as all prescriptions reimbursed by the government.

Study design and population
This study was approved by the Vilnius Regional Bioethics Committee, No. 158200-17-913-423. A retrospective cohort design was used to examine the relationship between diabetes and prostate cancer risk. The cohort was composed of patients (informed consent of study subject is not required for a record-linkage studies) identified in the NHIF database with  Only patients with the diagnosis of T2DM aged 40 or more years were included. We excluded 9 records because of incorrect date of diagnosis or death, 1938 records with prostate cancer diagnosis before diabetes diagnosis and 482 with prostate cancer diagnosis less than 1 year after diabetes diagnosis. We finally had 64000 males with a previous diabetes diagnosis in our analysis.

Statistical methods
The person-time of observation was computed from the date of the first recorded diabetes mellitus diagnosis in the NHIF database until diagnosis of prostate cancer, emigration or end of the observation period (31 December 2016), whichever came first.
We calculated standardized incidence ratios (SIRs) for prostate cancers as a ratio of observed number of cancer cases in people with diabetes diagnosis to the expected number of cancer cases in the underlying general population. Expected numbers were calculated as multiplication of the exact person-years under observation in the cohort by calendar year-and 5-year-age-groups-specific national incidence rates. 95% confidence intervals for the SIRs were estimated assuming number of observed cases follows Poisson distribution.
We computed SIRs by age of diabetes diagnosis, duration of follow-up until occurrence of prostate cancer and use of metformin. Patients with a first diabetes diagnosis in 2000 also included prevalent cases, therefore they were excluded from analysis by duration of follow-up. Patients who had never been prescribed metformin after entry were defined as never-users; ever-users were defined as those who had been prescribed metformin more than 6 times and first prescription more than 1 year before prostate cancer diagnosis. To evaluate dose-response relationship we calculated cumulative dose from the NHIF database.
All statistical analyses were carried out using STATA 11 statistical software (StataCorp.

Discussion
In this analysis involving 64000 men with more than 1 years' duration of T2DM we found an 11% reduction in the risk of developing prostate cancer compared to the general population. Our findings were in accordance with the results of a meta-analysis by Bansal et al. which covered forty five studies involving approximately 8 million participants with 132 thousand prostate cancer cases detected and showed that T2DM lowers the risk of prostate by 14% ( 95% CI 0.8-0.9,2 p<0.01) [8].
Hypothesized mechanisms for decreased prostate cancer risk among diabetics include decreased levels of hormones and other cancer-related growth factors among diabetics and protective effects of diabetes medications [18]. Inverse association between T2DM and prostate cancer can be explained by the changes in insulin levels in men with diabetes. Initially many of them are hyperinsulinemic, but as the disease progresses, the level of insulin may decline and thus have growth inhibitory effect on cancer cell development [19]. Prolonged hyperglycaemia, hyperinsulinemia and insulin resistance have been reported to cause destruction of the Leydig cells, causing lower testosterone levels [20,21] with longer duration of diabetes; risk of prostate cancer was slightly increased during the first 3 years after diagnosis of diabetes, but was reduced among men diagnosed 4 or more years before [24]. Another study showed that prostate cancer risk was not reduced in the first year after diabetes diagnosis, was lower for men diagnosed for 1-6 years, and was even lower for men who had been diagnosed for more than 6 years after diabetes diagnosis [25]. In our study first year of follow-up was excluded from the analysis.
It was previously published that younger patients have lower risk of prostate cancer. Two studies both revealed that patients diagnosed with diabetes mellitus before 30 might have a relatively lower risk of PCa than those diagnosed with diabetes after age 30 [25, 26].
However, these studies included all types of diabetes mellitus and very young patients, therefore knowing that prostate cancer is predominantly a disease of the elderly, it is possible that studying a younger cohort and with shorter follow up these results could be inaccurate. Our T2DM cohort data demonstrated that prostate cancer risk was reduced similarly through all age groups starting from 40 years of age.
Metformin as the most common medication used in the management of T2DM, has also been suggested to decrease overall risk of prostate cancer [7]. We found a slightly lower risk of prostate cancer in patients who have used metformin compared to never users, 29% vs 12% respectively. Data from a Taiwanese study suggest that the longer T2DM is controlled and the higher the cumulative metformin dose, the more pronounced metformin protective effect against prostate cancer can be seen [13]. Recent data coming from Finland confirm that patients with T2DM and a normo glycemic status have a lower risk for developing prostate cancer, supporting the risk lowering effect of this antidiabetic drug [27]. However, in a study from Sweden men with T2DM on metformin had no decrease risk compared to men with T2DM not on anti-diabetic drugs [16]. Therefore, more clinical prospective randomized trials are needed to consider it for everyday use as a prostate cancer chemopreventive drug.
The relationship between diabetes and metformin use and prostate cancer may be suspected to be causal due to evidence of decreasing prostate cancer risk with increasing diabetes duration and longer duration of metformin use, or higher cumulative dose. Our study failed to show dose-response relationship between duration of diabetes and cumulative dose of metformin.
The power of this analysis lays in its population-based nature, allowing comparisons within the population from which the cases were identified and omits selection bias of control group. However, the strength of the inferences of registry-based studies is limited by inability to control for the individual level confounders other than age. T2DM is a chronic disease and patients, probably after being diagnosed, can be thought to adjust their dietary habits, alcohol consumption or increase physical activity. However, it is not possible to control and measure these changes and account that they are crucial in reducing prostate cancer risk in T2DM patients. Other drawbacks of this study are provided further. The patients with T2DM are under increased surveillance and the possibility to discover prostate cancer might be higher, patients have more frequent visits to the doctors and are more likely to undergo additional medical examinations including PSA testing. If such bias exists it would imply that in our study the reduction of prostate cancer risk among T2DM is under-estimated. On the other hand, men with diabetes have been found to have lower PSA levels compared to men without diabetes [28], while lower PSA levels could also contribute to reduced detection of prostate cancer in men with diabetes and they had significantly more advanced tumours with higher PSA levels [29].
Another limitation of this study lack of information on grading and staging of prostate cancer, therefore we were not able to analyse risk by aggressiveness of prostate cancer.

Conclusions
In this large population-based study we found significantly decreased risk of prostate cancer among men with T2DM. The risk of prostate cancer seemed not be influenced by patients age and increasing time since T2DM diagnosis. Significantly lower risk also was found in both metformin users and never-users groups, with lover risk of prostate cancer in ever-users, however, there was no clear trend in the risk according cumulative dose. The data from the national database were requested in the context of the study which was aprooved by the Vilnius Regional Bioethics Committee, No. 158200-17-913-423. A retrospective cohort design was used to examine the relationship between diabetes and prostate cancer risk. The cohort was composed of patients (informed consent of study subject is not required for a record-linkage studies) identified in the NHIF database with the diagnosis of T2DM.

Not applicable
Availability of data and material -All data relevant to the study are included in the article or uploaded as supplementary information.

Competing interests
The authors declare that they have no competing interests.