Abstract
Background/Aim: The clinical hazard of prostate cancer development after five-alpha reductase inhibitors (5ARI) treatment among benign prostate hyperplasia (BPH) patients is still controversial. The aim of this study was to evaluate the epidemiological features of BPH patients treated in a single institute to identify risk factors associated with prostate cancer development. Patients and Methods: We retrospectively analyzed patients who were diagnosed with BPH and received alpha blockers (AB) only or 5ARI between January 2007 and December 2012 and followed up until death or December 2020. The primary study outcome was prostate cancer and high-grade prostate cancer. Results: Of the 5,122 included patients, 14.9% (762/5,122) received 5ARI during their BPH treatment. The median age, initial prostate specific antigen (PSA) levels and the PSA change were significantly higher in the 5ARI group compared to those of the AB group. The prostate cancer diagnosis rate was higher in the 5ARI group, and the percentage of high-grade prostate cancer was not different between the two groups. In total, 1,715 (33.5%) patients were recorded dead, and the median follow-up period was longer in the 5ARI group. In Cox regression analysis, only age and initial PSA levels were significantly associated with prostate cancer. Late PSA was the only independent factor associated with high-grade prostate cancer development. Conclusion: Our real-world data revealed that age, initial PSA, and late PSA levels were associated with prostate cancer and high-grade prostate cancer diagnosis among BPH patients. Furthermore, 5ARI use had no effect on prostate cancer patient survival. However, PSA assessment during follow-up is still required in our institutional practice to avoid delayed diagnosis.
- Alpha blocker
- benign prostate hyperplasia
- five-alpha reductase inhibitor
- prostate cancer
- prostate specific antigen
Five-alpha reductase inhibitors (5ARI) are common first line prescriptions for benign prostate hyperplasia (BPH). 5ARI provides persistent and reliable long-term efficacy in reducing prostate gland size compared to alpha blockers (AB). In the Medical Therapy of Prostatic Symptoms (MTOPS) and Combination of Avodart and Tamsulosin (COMBAT) trials, 5ARI reduced urinary retention rate, surgery rate, and 4-year long-term voiding symptoms improvement (1, 2). In addition, through modulation of serum dihydrotestosterone levels, 5ARI treatment results in decreased levels of the prostate specific antigen (PSA) levels masking the presence of prostate cancer. Although the Prostate Cancer Prevention Trial (PCPT) and the Reduction by Dutasteride of Prostate Cancer Events (REDUCE) trial both found a decrease in the incidence of prostate cancer grade 7 or less and overall prostate cancer, a higher rate of high-grade prostate cancer was detected in the study group (3, 4). Both clinical trials raised the question whether 5ARI treatment is associated with high grade prostate cancer development. Analysis of several real-world databases also examined the effect of 5ARI treatment on prostate cancer development and the results were controversial. In 2014, Preston et al. reported that 5ARI was not associated with lethal, high-grade, or grade 7 prostate cancer (5). Van Rompay et al. used a 20-year cohort database and found that 5ARI treatment was associated with lower prostate cancer diagnosis compared with AB treatment and no treatment (6). However, Sarkar et al. analyzed the veteran affair system database in the United States and showed that 5ARI treatment may delay diagnosis of prostate cancer and jeopardize cancer-specific survival (7). Since the clinical trials and real-world database results varied, we conducted a clinical database study with first-hand patient information to identify the association of 5ARI use with prostate cancer.
Patients and Methods
Database. The database was constructed in a web-based format including all clinical information of patients who received medical care in a single medical institute since January 2000. It contained diagnosis, English-based charts, and lab examinations in outpatient, hospitalizations, emergency department admissions and clinical trials with de-identification management.
Patient selection and treatments. Between January 2007 and December 2012, patients who were newly diagnosed with BPH and prescribed with AB or 5ARI were included. The BPH diagnosis was coded with ICD-9-CM 600. AB and 5ARI were coded using anatomical therapeutic chemical classification (ATC) codes including Doxazosin (C02CA04), Alfuzosin (G04CA01), Tamsulosin (G04CA02), Terazosin (G04CA03), Silodosin (G04CA04), Alfuzosin plus Finasteride (G04CA51), Tamsulosin plus Dutasteride (G04CA52), Dutasteride (G04CB02), and Finasteride (G04CB01). The exclusion criteria included: previously diagnosed cancers with ICD-9-CM 140-239 and prostate cancer diagnosis within one month after inclusion. All included patients were followed up till December 31, 2020, last hospital visit or death.
Study assessment. This study was conducted through retrospective clinical data claiming and analysis process with an external validation and was approved by the institute review board of Taichung Veterans General Hospital (number CE21170A). The study end-point was prostate cancer diagnosis or patient death. Patients who only received AB were classified into the AB group. In contrast, the 5ARI group included patients who received any 5ARI medications during the follow-up period. Data collection included age, body mass index (BMI) in the nearest record while inclusion, serum PSA levels at initial inclusion and the latest measurement during the study period, PSA change, comorbidities during study periods (hypertension, ICD9 401-405, ICD10 I10-I16; diabetes mellitus, ICD9 250, ICD 10 E08-E13; ischemic heart disease ICD9 410-414, ICD10 I20-I25; cerebrovascular disease ICD9 430-438, ICD I60-I69; hyperlipidemia ICD9 272, ICD10 E78; chronic obstructive pulmonary disease ICD9 490-496, ICD10 J40-J47; peripheral vascular disease ICD9 440-449, ICD10 I70-I79; chronic kidney disease ICD9 585, ICD10 N18; sleep disorder ICD9 327, ICD10 G47; gout ICD9 274, ICD10, M10), duration of medical therapy with 5ARI, incidence of prostate cancer, incidence of high-grade prostate cancer, patient death, and duration of follow-up. Instead of doubling the PSA in the 5ARI group as other clinical trials, we used primary PSA as a study variable for both groups. PSA change was defined using latest PSA minus initial PSA. High grade prostate cancer was defined as Gleason score sum 8-10. All the continuous variables are displayed as mean and categorical variables as percentages. The study flow chart is listed in Figure 1.
Flow diagram showing the process of benign prostate hyperplasia patient sampling and participation. BPH: Benign prostate hyperplasia; 5ARI: five-alpha reductase inhibitor.
Statistical analysis. The Mann-Whitney U-test was applied for the continuous variables, such as age, PSA levels, and follow-up duration. The Chi-square test and Fisher’s t-test were used for categorical variables, such as comorbidities, as well as the use of 5ARI. Univariate and multivariate Cox hazard proportional regression was used to estimate the hazard ratio (HR) and 95% confidence interval (CI) for association between variables and the development of prostate cancer or high-grade prostate cancer. All statistical analyses were performed using SAS software version 9.2 (SAS Institute, Inc., Cary, NC, USA).
Results
There were 5,246 patients in our database diagnosed with BPH and treated with AB or 5ARI. After exclusion, a total of 5,122 patients were included in the analysis. Amongst them, 4,360 patients received AB as the major treatment without 5ARI and the other 762 patients received 5ARI with or without AB (Figure 1). The overall prostate cancer development rate was 1.8% (91/5,122, Table I). The median age of all included patients was 66 years and the 5ARI group had a significantly elder age than the AB group (69 vs. 66, p<0.001). Median BMI was 24.6 kg/m2 in total and the median initial PSA levels were higher in the 5ARI group (2.8 vs. 1.4, p<0.001). Although the latest PSA levels revealed no difference between the two groups, the PSA change showed a significant decrease in the 5ARI group (−0.1 vs. 0, p<0.001). There were also significantly higher percentages of hypertension, chronic kidney disease, and sleep disorder in the 5ARI group. Both groups were similar regarding diabetes mellitus, ischemic heart disease, cerebrovascular disease, hyperlipidemia, chronic obstructive pulmonary disease, peripheral vascular disease, and gout. The rate of prostate cancer was higher in the 5ARI group (3.5% vs. 1.5%, p<0.001) and the high-grade prostate cancer rate had no difference between the two groups. In total, 1,715 patients died during follow-up and the median follow-up period was longer in the 5ARI group (7.5 vs. 3.7 years, p<0.001).
Baseline characteristics and demographics of benign prostate hyperplasia (BPH) patients followed up for prostate cancer.
In Cox regression analysis, age, initial PSA levels, final PSA levels, and the use of 5ARI were significant in univariate analysis. However, only age and initial PSA levels showed a statistically significant association with prostate cancer (HR=1.02, 95%CI=1.00-1.04, p=0.043, and HR=1.02, 95%CI=1.01-1.02, p<0.001 respectively, Table II).
Univariate and multivariate analysis for prostate cancer development.
Among 91 patients who developed prostate cancer, high-grade prostate cancer was recorded in 13 (14.29%, Table III). There were no differences between the 5ARI and AB group in the rate of high-grade prostate cancer development or other variables. The rate of high-grade prostate cancer accounted for 18.5% in the 5ARI group and 12.5% in the AB group. However, in multivariate analysis, final PSA level was the only independent factor associated with high-grade prostate cancer development (HR=1.04, 95%CI=1.01-1.07, p=0.022, Table IV).
Characteristics of benign prostate hyperplasia (BPH) patients who developed prostate cancer.
Risk analysis for high grade prostate cancer.
Discussion
There is an ongoing debate regarding the use of 5ARI and the risk of high-grade prostate cancer. Hsieh et al. found upregulation of androgen receptors after finasteride treatment in an androgen-dependent prostate cancer cell line and clinical patient tissues (8). Through an in vitro model, 5ARI sensitized prostate cancer cells and contributed to cell growth. Chhipa et al. treated different prostate cancer cell lines with dutasteride and finasteride and found that 5ARI induced more cell death in low grade cancer (9). These findings suggested the potential role of 5ARI in the chemoprevention of low-grade prostate cancer and acceleration of high-grade prostate cancer. However, in real-world practice, the diagnosis rate of prostate cancer is influenced by many factors, such as human race, patients’ preference, and clinical practice policy. For examples, some elderly people would hesitate to have tests for prostate cancer while the lower urinary tract symptoms are regarded as parts of the aging process. In United States and Taiwan, PSA screening is not approved by the government. In our study, 5ARI use was not associated with subsequent prostate cancer diagnosis; instead, age and baseline PSA levels were independent factors associated with prostate cancer development. This finding reflected several healthcare features in Taiwan. First, the overall prostate cancer diagnosis rate was 1.8%, which was much lower than the 9.3% in the Canadian regional database study by Van Rompay et al. (6). One of the reasons is the low prostate cancer incidence in Taiwan. Although the national incidence of prostate cancer raised rapidly from 35.47 cases per 100,000 in 2006 to 60.79 per 100,000 in 2019 in Taiwan, this was still much lower than the 113.57 cases per 100,000 in Canada (10, 11). Second, unlike other large regional or national database cohorts, we used an institutional database to obtain patient clinical information, especially the PSA levels (6, 12-16). Third, in Taiwan, 5ARI prescription is not allowed unless prostate cancer is diagnosed. Based on the strict patient selection criteria, most potential prostate cancer patients were excluded before or in the early phase of the study. Therefore, the overall diagnostic rate of prostate cancer decreased after inclusion in this study. Fourth, our results also showed that 5ARI was widely used in elder men to avoid surgical intervention for prostate obstruction. Fifth, the median follow-up duration in our study was 4.1 years compared with 8 years in the series of Wallerstedt et al. They found that the longer the duration of 5ARI treatment, the less the risk of prostate cancer Gleason grade 7 or less (12). In our multi-variate analysis, 5ARI treatment duration did not show any association with prostate cancer or high-grade prostate cancer risk. Sixth, the median PSA levels at prostate cancer diagnosis were 15.7 ng/ml in the 5ARI group compared to 5.4 ng/ml in the AB group (p=0.005, Table III). It revealed that initial PSA levels are still a useful primary tool to identify potential prostate cancer patients and is different from using PSA density as in the series of Inoue et al. (17).
There are several limitations to our study. First, the cohort database contained only institutional information and the real-world results regarding prostate cancer diagnosis or 5ARI treatment outside our hospital were missed. Second, the prostate cancer diagnosis rate was too low. It may result from the fundamentally low prostate cancer incidence in the Taiwanese population, or the low biopsy rate in clinical practice. Third, following the low prostate cancer diagnosis rate, the high-grade prostate cancer rate may also be biased. Fourth, the median follow-up duration was 4.1 years, which was not long enough to observe prostate cancer development.
In conclusion, our study revealed that 5ARI use was not associated with prostate cancer and high-grade prostate cancer development. Age and PSA levels were associated with increased risk of prostate cancer diagnosis. Moreover, twofold higher PSA levels were observed in the diagnosed prostate cancer patients and the use of PSA as a follow-up tool was found to be important in our institute.
Acknowledgements
This work was supported by the Ministry of Science and Technology, Taiwan, R.O.C. Grant number: MOST 109-2314-B-075A-007-MY3.
Footnotes
Authors’ Contributions
Jian-Ri Li, Li-Wen Chang, Sheng-Chun Hung, Shian-Shiang Wang, Chuan-Shu Chen, Cheng-Kuang Yang, and Kun-Yuan Chiu contributed to study design. Li-Wen Chang, and Sheng-Chun Hung contributed to data collection. Statistical analysis was performed by Kevin Lu and Chiann-Yi Hsu. The manuscript was written by Li-Wen Chang, Jian-Ri Li and revised by Kevin Lu and Chen-Li Cheng.
Conflicts of Interest
The Authors have declared no conflicts of interest in relation to this study.
- Received September 13, 2022.
- Revision received September 22, 2022.
- Accepted September 26, 2022.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
