Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Oct 16.
Published in final edited form as: Pharmacoepidemiol Drug Saf. 2019 Nov 12;29(1):48–56. doi: 10.1002/pds.4909

The role of 5α-reductase inhibitors in gastro-oesophageal cancer risk: a nested case-control study

John Busby 1, Reema Karasneh 1, Peter Murchie 2, Úna McMenamin 1, Shahinaz M Gadalla 3, M Constanza Camargo 3, Lisa Iversen 2, Amanda J Lee 4, Andrew D Spence 1, Chris R Cardwell 1
PMCID: PMC8520491  NIHMSID: NIHMS1742304  PMID: 31713940

Abstract

Purpose:

The strong male predominance of gastro-oesophageal cancer suggests sex hormones play an important role. 5α-reductase (5AR) inhibitors have antiandrogen effects, and have been shown to decrease cancer cell proliferation and metastasis. We conducted the first epidemiologic investigation into the association between 5AR inhibitor use and gastro-oesophageal cancer risk.

Methods:

We conducted a nested case-control study within the Scottish Primary Care Clinical Information Unit Research database. Male cases diagnosed with oesophageal or gastric cancer between 1999 and 2011 were matched to up to five male controls based on birth year, diagnosis year and general practice. We used electronic prescribing records to ascertain medication use. We used conditional logistic regression to calculate odds ratios (ORs) for the association between 5AR inhibitor use and cancer risk, after adjusting for comorbidities and aspirin, statin or proton pump inhibitor use.

Results:

The study included 2,003 gastro-oesophageal cancer cases and 9,650 controls. There was some evidence of reduced gastro-oesophageal cancer risk among 5AR inhibitor users (adjusted OR=0.75; 95% CI: 0.56, 1.02), particularly for finasteride (adjusted OR=0.68; 95% CI 0.50, 0.94). These decreases were more marked among those who received at least 3 years of 5AR inhibitors (adjusted OR= 0.54; 95% CI: 0.27, 1.05; p-value=0.071) or finasteride (adjusted OR=0.49; 95% CI 0.24, 0.99; p-value=0.046).

Conclusions:

We found evidence of reduced gastro-oesophageal cancer risk among users of 5AR inhibitors, particularly finasteride. However, larger epidemiological studies are required before randomised controlled trials are considered.

Keywords: Gastric cancer, Oesophageal cancer, Androgen, 5α-reductase inhibitor

Introduction

Oesophageal and gastric cancer are among the most common cancers in the world, with around 456,000 and 952,000 new cases diagnosed annually.(1) Prognosis is extremely poor, even in high-income countries such as the UK, where nearly 55% of patients die within one year of diagnosis.(2) The strong male predominance of gastro-oesophageal cancer(3–5) has prompted much interest in the role of sex hormones.(6, 7) This hormonal hypothesis is supported by findings of reduced gastric cancer incidence among hormone replacement therapy users, and women with a longer time between menarche and menopause.(8)

5α-reductase (5AR) inhibitors (including finasteride and dutasteride) are widely used(9) medications for the treatment of benign prostatic hyperplasia(10), the most common urologic disease among elderly males causing symptoms including nocturia urinary urgency.(11) Once prescribed, they are used indefinitely to reduce prostate size by inhibiting the androgen receptor thus preventing the conversion of testosterone to the more biologically active dihydrotestosterone.(12, 13) 5AR inhibitors have been shown to increase the ratio of oestrogen to testosterone(14), and have been linked with some hormonal side effects such as gynaecomastia.(15) Preclinical evidence that androgen receptors are present within oesophageal (16–18) and gastric cancer tumours (19–21), and that their inhibition reduces cell proliferation and migration(22, 23), suggests that medications with hormonal effects could influence gastro-oesophageal cancer risk. Surprisingly, clinical studies have yet to assess the impact of 5AR inhibitor use on gastro-oesophageal cancer.

Consequently, in a population-based primary care cohort from Scotland, we conducted the first study to investigate whether 5AR inhibitor use was associated with a reduced risk of gastro-oesophageal cancer.

Methods

Data

We conducted a nested case-control study within the Primary Care Clinical Information Unit Research (PCCIUR) database.(24) Between 1993 and 2011, the PCCIUR collected computerised medical records from approximately 15% of the Scottish general practice population, and includes details on patient demographics (e.g. age, gender), primary care encounters, clinical diagnoses and prescriptions. Access to the PCCIUR data was approved by the Research Applications and Data Management Team, University of Aberdeen. Ethical approval for this study was supplied by the Queen’s University Belfast, School of Medicine Ethics Committee (reference number: 15.43).

Cases and controls

Our primary outcome was gastro-oesophageal cancer since classifying tumours arising close to the oesophagogastric junction is difficult.(25, 26) Cases and controls were restricted to males as 5AR inhibitors are not used in females. Cases were defined as patients with a primary oesophageal (Read code: B10..) or gastric (Read code: B11..) cancer diagnosis between 1st January 1999 and 30th April 2011. Up to five controls were randomly selected for each case matched on year of birth and general practice (equivalent to US family practice centres). Controls were chosen to ensure they were registered with the general practice during the year of the cases cancer diagnosis to prevent confounding due to secular time effects or changes in coding practices. The index date was defined as the diagnosis date of the case in each matched group.

Definition of exposure

We identified 5AR inhibitor use from electronic prescription records. We used the British National Formulary to compile a list of proprietary and generic names for 5AR inhibitor medications (Appendix 1). The start of the exposure assessment window was the latest of 1st January 1996 (as prescriptions before this were less likely to be generated electronically) or the date of general practice registration. The exposure assessment window ended one year before cancer diagnosis to prevent reverse causation. Additionally, the exposure assessment window was truncated to ensure it was identical across the matched groups.(27) Cases and controls with an earlier cancer diagnosis (other than non-melanoma skin cancer), and those with less than a three year exposure assessment window, were excluded from the study. We defined patients as users if they had at least one prescription during the exposure assessment window.

To enable the testing of dose‐response relationships, we extracted data on the medication prescribed, number of packs/tablets and medication strength, and calculated defined daily doses (DDDs). The DDD system is a validated measure of drug consumption maintained by the World Health Organization.(28) A single DDD is the average maintenance dose per day of a drug used for its main indication in adults (e.g. BPH for finasteride). We calculated the total number of DDDs received during the exposure window and categorized patients into those receiving 0 (never users), between 1 and 365 DDDs (less than 1 years usage), between 366 and 1095 DDDs (between 1 and 3 years usage), and more than 1095 (more than 3 years usage).

Confounding factors

We identified eleven comorbidities (myocardial infarction, heart failure, peripheral vascular disease, cerebrovascular disease, connective tissue disease, dementia, chronic obstructive pulmonary disease, rheumatoid arthritis, diabetes, renal disease, liver disease) using Read codes recorded by the GP during the exposure assessment window.(29) This list was based on the comorbidities comprising the Charlson comorbidity index after removing rare diseases within the UK (e.g. AIDS), cancer (as those with previous cancer are excluded from the study), and those which may lie on the causal pathway between 5ARI use and gastro-oesophageal cancer (e.g. peptic ulcer). Read codes are a thesaurus of clinical terms broadly based on ICD codes and have been widely used within UK general practice since 1985.(30)

Use of aspirin, statins, and proton pump inhibitors within the exposure assessment window was identified from prescription records (Appendix 1), as associations with oesophageal and/or gastric cancer have been identified previously.(31–34) Lifestyle data including obesity (body mass index, BMI>30), smoking status (never, ex, current) and alcohol use (none, low [e.g. moderate or light drinker], high [e.g. above recommended limits, chronic alcoholism]) were also recorded in the PCCIUR data using Read codes. Lifestyle factors were assessed using the record closest to the date of cases cancer diagnosis within the exposure assessment window.

Statistical analysis

We calculated descriptive statistics, and compared the demographics and clinical characteristics of the cases and controls. We used conditional logistic regression to estimate odds ratios (ORs) with 95% confidence intervals (CIs) for the association between 5AR inhibitor use and gastro-oesophageal cancer risk. The matched design accounted for the effect of age, general practice and year of diagnosis with additional adjustments made for statin use (yes/no), aspirin use (yes/no), proton pump inhibitor use (yes/no), and the presence of each comorbidity (yes/no) using regression. We conducted separate site-specific analysis for oesophageal and gastric cancer, and used an interaction test to formally test for differences across sites. Similarly, we conducted separate gastro-oesophageal analyses for finasteride and dutasteride, as they use different mechanisms to inhibit testosterone metabolism.(35)

Sensitivity and subgroup analyses

We investigated the impact of additionally adjusting for lifestyle factors (smoking, alcohol consumption and obesity) using complete case and multiple imputation with chained equations methods. The imputation used ordered logit models with age and deprivation (Scottish Index of Multiple Deprivation quintile based on the postcode of the GP practice), separately for cases and controls. Briefly, this is a simulation-based approach for handling missing data which leads to valid statistical inferences.(36) Sensitivity analyses were also conducted investigating the impact of excluding prescriptions in the two years prior to the index date (as opposed to one in the main analysis), and defining medication users as patients with at least three 5AR inhibitor prescriptions (as opposed to one in the main analysis).

Finally, to investigate confounding, we conducted separate negative control analyses(37) for liver (Read code: B15..) and pancreatic (Read code: B17..) cancer, as there is little evidence of androgen receptor expression in these sites(38, 39), but they should be subject to many of the same potential biases as gastro-oesophageal cancer. Therefore, if confounding was driving our gastro-oesophageal cancer results we would expect to see similar associations for liver and pancreatic cancer. Conversely, findings of meaningful associations for gastro-oesophageal cancer, which are not replicated among the negative controls, would support a causal interpretation. Analyses were conducted using Stata version 13.(40)

Results

Descriptive statistics

We identified 2,003 cases of gastro-oesophageal cancer (1,364 oesophageal and 639 gastric cancer) (Table 1). An average of 4.8 controls were matched for each case, with a median exposure assessment window of 5.4 years (min 3.0, max 15.1). A larger proportion of cases than controls had a history of COPD (12.3 vs. 8.6%), and were more likely to be current or ex-smokers (70.8 vs. 61.0%), drink high levels of alcohol (9.7 vs. 7.3%) and have a BMI under thirty (85.7 vs. 80.1%).

Table 1:

Characteristics of controls and cases with gastro-oesophageal cancer

Cases (%) Controls (%)
Count 2,003 9,650
Median Exposure Years (IQR) 5.4 (4.0,7.6) 5.4 (4.0,7.6)
Year of Diagnosis
 1999–2003 670 (33.4) 3,238 (33.6)
 2004–2007 921 (46.0) 4,416 (45.8)
 2008–2011 412 (20.6) 1,996 (20.7)
Mean Age (SD) 68.2 (11.0) 67.7 (10.8)
 <39 18 (0.9) 89 (0.9)
 40–59 421 (21.0) 2,093 (21.7)
 60–79 1,264 (63.1) 6,193 (64.2)
 80+ 300 (15.0) 1,275 (13.2)
Smoking status
 Non-smoker 498 (29.2) 2,871 (39.0)
 Ex-smoker 649 (38.0) 2,655 (36.1)
 Current smoker 561 (32.8) 1,835 (24.9)
 Missing 295 2,289
Alcohol Consumption
 No 259 (17.6) 1,123 (17.4)
 Low 1,071 (72.7) 4,857 (75.3)
 High 143 (9.7) 474 (7.3)
 Missing 530 3,196
Obesity
 Not Obese 1,716 (85.7) 7,725 (80.1)
 Obese 287 (14.3) 1,925 (19.9)
Deprivation Quintile
 1 (Least Deprived) 235 (11.9) 1,113 (11.7)
 2 358 (18.1) 1,713 (18.0)
 3 431 (21.8) 2,092 (22.0)
 4 476 (24.1) 2,319 (24.4)
 5 (Most Deprived) 473 (24.0) 2,271 (23.9)
 Missing 30 142
Common Comorbidities a
 Connective Tissue Disease 853 (42.6) 4,145 (43.0)
 Diabetes 229 (11.4) 1,019 (10.6)
 MI 216 (10.8) 936 (9.7)
 COPD 246 (12.3) 830 (8.6)
 CVD 199 (9.9) 785 (8.1)
Other Medication Use
 Aspirin 611 (30.5) 2,802 (29.0)
 Statin 507 (25.3) 2,316 (24.0)
 Proton pump inhibitors 521 (26.0) 1,910 (19.8)
Open in a new tab
a

For brevity, only the 5 most common comorbidities are listed. The full analysis included myocardial infarction (MI), heart failure, peripheral vascular disease, cerebrovascular disease (CVD), connective tissue disease, dementia, chronic obstructive pulmonary disease (COPD), rheumatoid arthritis, diabetes, renal disease and liver disease

Association between 5AR inhibitor use and gastro-oesophageal cancer

Overall, 2.7% of cases and 3.2% of controls used 5AR inhibitors. Finasteride (2.9% of study population) was more commonly used than dutasteride (0.3%). There was some evidence of a reduction in gastro-oesophageal cancer risk among 5AR inhibitor users (ORadj= 0.75; 95% CI: 0.56, 1.02; Table 2), particularly those using finasteride (ORadj= 0.68; 95% CI: 0.50, 0.94). In contrast, there was no significant association between dutasteride use and gastro-oesophageal cancer risk (ORadj= 1.53; 95% CI 0.67, 3.50). There were more marked associations among patients who received at least 3 years of 5AR inhibitors (adjusted OR= 0.54; 95% CI: 0.27, 1.05; p-value=0.071) or finasteride (adjusted OR=0.49; 95% CI 0.24, 0.99; p-value=0.046). There was no evidence of a difference in associations between gastric (ORadj= 0.68; 95% CI: 0.39, 1.17) or oesophageal (ORadj= 0.78; 95% CI: 0.55, 1.13) cancer (p-value for interaction=0.633).

Table 2:

5AR inhibitor use and upper gastro-oesophageal cancer riskb

Medication Cases n,(%) Controls n,(%) Unadjusted OR (95% CI)c Adjusted OR (95% CI)d P-value
Gastro-oesophageal
Any 5AR inhibitor
 Never use 1,949 (97.3) 9,343 (96.8) Ref Ref
 Ever use 54 (2.7) 307 (3.2) 0.80 (0.59,1.08) 0.75 (0.56,1.02) 0.065
 <365 DDDs 22 (1.1%) 127 (1.3%) 0.77 (0.49,1.22) 0.74 (0.47,1.17) 0.201
 366 – 1095 DDDs 22 (1.1%) 100 (1.0%) 1.03 (0.64,1.65) 0.94 (0.58,1.53) 0.813
 >1095 DDDs 10 (0.5%) 80 (0.8%) 0.57 (0.29,1.10) 0.54 (0.27,1.05) 0.071
Finasteride
 Never use 1,956 (97.7) 9,357 (97.0) Ref Ref
 Ever use 47 (2.3) 293 (3.0) 0.73 (0.53,1.00) 0.68 (0.50,0.94) 0.020
 <365 DDDs 18 (0.9%) 119 (1.2%) 0.69 (0.42,1.13) 0.65 (0.39,1.07) 0.090
 366 – 1095 DDDs 20 (1.0%) 95 (1.0%) 0.98 (0.60,1.61) 0.90 (0.54,1.48) 0.667
 >1095 DDDs 9 (0.4%) 79 (0.8%) 0.51 (0.25,1.03) 0.49 (0.24,0.99) 0.046
Dutasteride
 Never use 1,995 (99.6) 9,627 (99.8) Ref Ref
 Ever use 8 (0.4) 23 (0.2) 1.53 (0.67,3.48) 1.53 (0.67,3.50) 0.312
 <365 DDDs 6 (0.3%) 18 (0.2%) 1.41 (0.55,3.61) 1.45 (0.56,3.73) 0.439
 366 – 1095 DDDs 1 (0.0%) 4 (0.0%) 1.21 (0.12,12.06) 1.16 (0.12,11.65) 0.899
 >1095 DDDs 1 (0.0%) 1 (0.0%) 5.00 (0.31,79.94) 4.03 (0.25,65.80) 0.328
Oesophageal
Any 5AR inhibitor
 Never use 1,326 (97.2) 6,366 (96.9) Ref Ref
 Ever use 38 (2.8) 202 (3.1) 0.85 (0.59,1.21) 0.78 (0.55,1.13) 0.191
 <365 DDDs 14 (1.0%) 86 (1.3%) 0.71 (0.40,1.26) 0.67 (0.38,1.20) 0.178
 366 – 1095 DDDs 15 (1.1%) 64 (1.0%) 1.09 (0.61,1.95) 1.00 (0.55,1.80) 0.990
 >1095 DDDs 9 (0.7%) 52 (0.8%) 0.79 (0.38,1.63) 0.72 (0.35,1.51) 0.387
Gastric
Any 5AR inhibitor
 Never use 623 (97.5) 2,977 (96.6) Ref Ref
 Ever use 16 (2.5) 105 (3.4) 0.71 (0.42,1.21) 0.68 (0.39,1.17) 0.162
 <365 DDDs 8 (1.3%) 41 (1.3%) 0.91 (0.43,1.95) 0.86 (0.40,1.86) 0.701
 366 – 1095 DDDs 7 (1.1%) 36 (1.2%) 0.93 (0.41,2.11) 0.87 (0.38,2.00) 0.745
 >1095 DDDs 1 (0.2%) 28 (0.9%) 0.16 (0.02,1.18) 0.16 (0.02,1.20) 0.075
Open in a new tab
b

OR: odds ratio; CI: confidence interval

c

Conditioned on year of birth, general practice and year of diagnosis

d

Additionally adjusted for use of aspirin, statins and proton pump inhibitors, and the presence of myocardial infarction, heart failure, peripheral vascular disease, cerebrovascular disease, connective tissue disease, dementia, chronic obstructive pulmonary disease, rheumatoid arthritis, diabetes, renal disease and liver disease.

Sensitivity and subgroup analysis

In general, our results were little altered in sensitivity analyses (Table 3). Similar findings were observed when excluding prescriptions in the two years prior to diagnosis, or when the exposure definition of ‘ever use’ was based upon three or more prescriptions. Additionally adjusting for lifestyle factors (smoking, alcohol consumption, and obesity) using multiple imputation methods resulted in similar findings to the main analysis. Our estimates were slightly attenuated when adjusting for lifestyle factors using complete-case methods, although our conclusions were unaltered. Finally, in negative control analyses, there was no evidence of a protective association between 5AR inhibitor use and liver (ORadj= 1.03; 95% CI: 0.50, 2.11) or pancreatic (ORadj= 0.91; 95% CI: 0.50, 1.69) cancer risk.

Table 3:

Sensitivity and negative control analyses for 5AR inhibitor usee

Cases n,(%) Controls n,(%) Unadjusted OR (95% CI)f Adjusted OR (95% CI)g
Gastro-oesophageal
2- year exposure lag
 Never use 1,964 (98.1) 9,404 (97.5) Ref Ref
 Ever use 39 (1.9) 246 (2.5) 0.73 (0.52,1.04) 0.68 (0.48,0.97)
Ever use ≥3 prescriptions
 Never use 1,955 (97.6) 9,384 (97.2) Ref Ref
 Ever use 48 (2.4) 266 (2.8) 0.82 (0.59,1.12) 0.77 (0.56,1.07)
MI lifestyle adjusted h
 Never use 1,949 (97.3) 9,343 (96.8) Ref Ref
 Ever use 54 (2.7) 307 (3.2) 0.80 (0.59,1.08) 0.78 (0.58,1.06)
Lifestyle complete case i
 Never use 1,396 (96.8) 5,927 (96.2) Ref Ref
 Ever use 46 (3.2) 234 (3.8) 0.85 (0.61,1.19) 0.86 (0.61,1.21)
Liver
Any 5AR inhibitor
 Never use 289 (96.0) 1,410 (96.5) Ref Ref
 Ever use 12 (4.0) 51 (3.5) 1.17 (0.60,2.26) 1.03 (0.50,2.11)
Pancreatic
Any 5AR inhibitor
 Never use 396 (96.4) 1,914 (96.4) Ref Ref
 Ever use 15 (3.6) 72 (3.6) 0.97 (0.54,1.74) 0.91 (0.50,1.69)
Open in a new tab
e

OR: odds ratio; CI: confidence interval

f

Conditioned on year of birth, general practice and year of diagnosis

g

Additionally adjusted for use of aspirin, statins and proton pump inhibitors, and the presence of myocardial infarction, heart failure, peripheral vascular disease, cerebrovascular disease, connective tissue disease, dementia, chronic obstructive pulmonary disease, rheumatoid arthritis, diabetes, renal disease and liver disease.

h

Additionally adjusted for smoking, alcohol consumption and obesity using multiple imputation with chained equations with age, gender and deprivation used in the imputation, separately for cases and controls, using chained ordered logit models

i

Additionally adjusted for smoking, alcohol consumption and obesity

Discussion

Summary of main findings

In this novel population-based study of gastro-oesophageal cancer cases and controls, there was evidence that users of 5AR inhibitors, and particularly finasteride, had lower cancer risk than non-users. These associations were most marked in those receiving at least 3 years of 5AR inhibitors, and were not replicated in negative control analyses for liver or pancreatic cancer.

Strengths and limitations

Our study is the first to investigate the effect of 5AR inhibitor use on gastro-oesophageal cancer risk. It is based on high-quality and nationally representative data.(41) We used prescribing information collected as part of routine clinical care, in many cases, several years before the onset of oesophageal or gastric cancer, which accurately reflects GP prescribing practices and negates the risk of recall bias. Although a weakness of this approach is that we do not know if patients used their medications, the main conclusions were similar when restricting our analysis to those who received multiple repeat 5AR inhibitor prescriptions (>12), where non-compliance is likely to be less of a concern.

Our study is observational and hence open to confounding. Although we have controlled for several of the key determinants of cancer risk through the matched design and analysis (e.g. age, comorbidities), some other risk factors, including ethnicity and nutrition, were not available. However, the lack of a protective association for pancreatic and liver cancer, which should share many of the same potential biases as gastro-oesophageal cancer, suggests that residual confounding is not solely driving our results. As almost all patients received 5ARI inhibitors for BPH we were unable to adjust for this factor in our analysis. Although this could have introduced confounding by indication in our analysis, we think this is unlikely given the lack of any evidence of an association between BPH and gastro-oesophageal cancer risk. The statistical power of our analysis is limited by our sample size and the relatively low use of 5AR inhibitors within our study population.

Our analysis is based on cancer diagnoses recorded in GP electronic records. Although a recent Clinical Practice Research Datalink study found that over 95% of gastro-oesophageal cancers are captured within GP records(42), a higher proportion of oesophageal cancers were recorded the PCCIUR data than in the Scottish Cancer Registry, which could suggest some misclassification of cancer site in our study.(43) This potential issue would not affect our primary analysis which combined oesophageal and gastric cancers. Finally, histological data were not available to allow a separate analysis of squamous cell carcinoma and adenocarcinoma, the two most common forms of gastro-oesophageal cancer. One preclinical study has demonstrated a higher concentration of androgen receptors in adenocarcinoma compared to squamous cell carcinoma, suggesting that 5AR inhibitors may have stronger effects in adenocarcinoma tumours.(44) Our exposure data is based on prescription data so we can not be sure that patients consumed their medications. We would expect the magnitude of this potential misclassification to be small and to result in slightly attenuated estimates (as it is unlikely to act non-differentially between study arms).

Comparisons with other research

To our knowledge, this is the first study to investigate the association between 5AR inhibitors and gastro-oesophageal cancer risk. Several studies have previously demonstrated lower prostate cancer risk among 5AR inhibitor users.(45–49) For example, one randomised controlled trial reported that patients treated with finasteride had a 30% (95% CI: 24, 35) lower risk of prostate cancer than those assigned to a placebo.(47) Epidemiological studies have also reported a reduced risk of bladder cancer, but not breast cancer, among 5AR inhibitor users.(50–52)

Implications for clinicians, policymakers and researchers

Our results suggest that the use of 5AR inhibitors may be associated with reduced gastro-oesophageal cancer risk. These results are compatible with preclinical evidence from oesophageal and gastric cancer cell lines demonstrating that androgen receptor inhibition leads to decreased proliferation and metastasis (22, 23), and are consistent with ecological findings of lower gastro-oesophageal cancer incidence among females(3) and prostate cancer survivors (who would have typically been treated with anti-androgen therapy).(53) Our finding of a weaker association for dutasteride is difficult to explain as it is a stronger androgen receptor inhibitor, and has a longer half-life than finasteride.(35) However, it is possible that our inconsistent findings reflect the very low use of dutasteride within our study population (<0.3%), which hindered our statistical power. Differences in gastro-oesophageal cancer risk by 5AR inhibitor type should be explored further in larger studies and/or in settings where dutasteride is more commonly prescribed.

The role of 5AR inhibitors as a chemopreventive agent should be explored further, particularly as they are inexpensive (finasteride costs £1.73 [$2.24] per 28-tablet pack)(54), have no major safety concerns, and are well tolerated by patients.(55, 56) These medications may have greatest utility among patients at high risk of gastro-oesophageal cancer, such as those with Barrett’s oesophagus or advanced gastric premalignant lesions.(57, 58) In this paper we have demonstrated that the association with gastro-oesophageal cancer adheres to several of Hill’s criteria for causation including biological plausibility, experimental evidence, strength of association, temporality, biological gradient, and specificity.(59) However, our findings should be replicated in larger epidemiological studies, with detailed information on tumour morphology and/or androgen receptor status, before randomised controlled trials can be considered.

Conclusions

In this novel population-based study, we found a 25% lower gastro-oesophageal cancer risk among 5AR inhibitor users. This association should be replicated in larger epidemiological studies before randomised controlled trials are considered.

Key points:

  • 5α-reductase (5AR) inhibitors have antiandrogen effects, and have been shown to decrease cancer cell proliferation and metastasis.

  • We conducted a nested case-control study within the Scottish Primary Care Clinical Information Unit Research database to assess the association between 5AR inhibitor use and gastro-oesophageal cancer.

  • There was some evidence of reduced gastro-oesophageal cancer risk among 5AR inhibitor users, particularly for finasteride.

  • Larger epidemiological studies are required before randomised controlled trials are considered.

Appendices

Appendix 1:

List of generic and proprietary drug names for each exposure and confounder medication

Substance Name Drug Name
5AR Inhibitor Avodart, Dutasteride, Finasteride, Proscar
Aspirin Asasantin, Aspirin, Caprin, Co-codaprin, Micropirin, Migramax, Nu-Seals
Dutasteride Dutasteride, Avodart
Finasteride Finasteride, Proscar
Proton pump inhibitor Esomeprazole, Heliclear, Helimet, Lansoprazole, Losec, Nexium, Omeprazole, Pantoprazole, Pariet, Protium, Rabeprazole Sodium, Zoton
Statin Atorvastatin, Cholib, Crestor, Dorisin, Fluvastatin, Inegy, Lescol, Lipitor, Lipostat, Luvinsta, Pinmactil, Pravastatin, Rosuvastatin, Simvador, Simvastatin, Stefluvin, Zocor
Open in a new tab

Footnotes

Disclosures

Conflict of Interest: he authors declare that they have no conflict of interest

References

  • 1.GLOBOCAN. Estimated Cancer Incidence, Mortaility and Prevelence Worldwide in 2012 2012. [ [Google Scholar]
  • 2.Office for National Statistics. Cancer survival in England: Patients diagnosed between 2010 and 2014 and followed up to 2015 2016 [Available from: http://www.ons.gov.uk/peoplepopulationandcommunity/healthandsocialcare/conditionsanddiseases/bulletins/cancersurvivalinenglandadultsdiagnosed/2010and2014andfollowedupto2015#5-year-survival.
  • 3.Coupland VH, Allum W, Blazeby JM, Mendall MA, Hardwick RH, Linklater KM, et al. Incidence and survival of oesophageal and gastric cancer in England between 1998 and 2007, a population-based study. BMC Cancer. 2012;12:11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Sipponen P, Correa P. Delayed rise in incidence of gastric cancer in females results in unique sex ratio (M/F) pattern: etiologic hypothesis. Gastric cancer : official journal of the International Gastric Cancer Association and the Japanese Gastric Cancer Association. 2002;5(4):213–9. [DOI] [PubMed] [Google Scholar]
  • 5.Rutegard M, Shore R, Lu Y, Lagergren P, Lindblad M. Sex differences in the incidence of gastrointestinal adenocarcinoma in Sweden 1970–2006. European journal of cancer (Oxford, England : 1990). 2010;46(6):1093–100. [DOI] [PubMed] [Google Scholar]
  • 6.Tian Y, Wan H, Lin Y, Xie X, Li Z, Tan G. Androgen receptor may be responsible for gender disparity in gastric cancer. Medical Hypotheses. 2013;80(5):672–4. [DOI] [PubMed] [Google Scholar]
  • 7.Sukocheva OA, Li B, Due SL, Hussey DJ, Watson DI. Androgens and esophageal cancer: What do we know? World Journal of Gastroenterology : WJG. 2015;21(20):6146–56. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Camargo MC, Goto Y, Zabaleta J, Morgan DR, Correa P, Rabkin CS. Sex Hormones, Hormonal Interventions and Gastric Cancer Risk: A Meta-Analysis. Cancer epidemiology, biomarkers & prevention : a publication of the American Association for Cancer Research, cosponsored by the American Society of Preventive Oncology. 2012;21(1):20–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Powell-Smith A, Goldacre B. OpenPrescribing.net 2016 [cited 2016 26/09/2016].
  • 10.Roehrborn CG. Benign Prostatic Hyperplasia: An Overview. Reviews in Urology. 2005;7(Suppl 9):S3–S14. [PMC free article] [PubMed] [Google Scholar]
  • 11.Robert G, Descazeaud A, Allory Y, Vacherot F, de la Taille A. Should We Investigate Prostatic Inflammation for the Management of Benign Prostatic Hyperplasia? European Urology Supplements. 2009;8(13):879–86. [Google Scholar]
  • 12.McConnell JD, Roehrborn CG, Bautista OM, Andriole GLJ, Dixon CM, Kusek JW, et al. The Long-Term Effect of Doxazosin, Finasteride, and Combination Therapy on the Clinical Progression of Benign Prostatic Hyperplasia. New England Journal of Medicine. 2003;349(25):2387–98. [DOI] [PubMed] [Google Scholar]
  • 13.McConnell JD, Bruskewitz R, Walsh P, Andriole G, Lieber M, Holtgrewe HL, et al. The Effect of Finasteride on the Risk of Acute Urinary Retention and the Need for Surgical Treatment among Men with Benign Prostatic Hyperplasia. New England Journal of Medicine. 1998;338(9):557–63. [DOI] [PubMed] [Google Scholar]
  • 14.Yao S, Till C, Kristal AR, Goodman PJ, Hsing AW, Tangen CM, et al. Serum estrogen levels and prostate cancer risk in the prostate cancer prevention trial: a nested case–control study. Cancer Causes & Control. 2011;22(8):1121–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Ramot Y, Czarnowicki T, Zlotogorski A. Finasteride induced Gynecomastia: Case report and Review of the Literature. International Journal of Trichology. 2009;1(1):27–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Awan AK, Iftikhar SY, Morris TM, Clarke PA, Grabowska AM, Waraich N, et al. Androgen receptors may act in a paracrine manner to regulate oesophageal adenocarcinoma growth. European journal of surgical oncology : the journal of the European Society of Surgical Oncology and the British Association of Surgical Oncology. 2007;33(5):561–8. [DOI] [PubMed] [Google Scholar]
  • 17.Nordenstedt H, Younes M, El-Serag HB. Expression of Androgen Receptors in Barrett Esophagus. Journal of clinical gastroenterology. 2012;46(3):251–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Yamashita Y, Hirai T, Mukaida H, Kawano K, Toge T, Niimoto M, et al. Detection of androgen receptors in human esophageal cancer. The Japanese journal of surgery. 1989;19(2):195–202. [DOI] [PubMed] [Google Scholar]
  • 19.Kominea A, Konstantinopoulos PA, Kapranos N, Vandoros G, Gkermpesi M, Andricopoulos P, et al. Androgen receptor (AR) expression is an independent unfavorable prognostic factor in gastric cancer. Journal of cancer research and clinical oncology. 2004;130(5):253–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Matsui M, Kojima O, Kawakami S, Uehara Y, Takahashi T. The prognosis of patients with gastric cancer possessing sex hormone receptors. Surgery today. 1992;22(5):421–5. [DOI] [PubMed] [Google Scholar]
  • 21.Wu CW, Chi CW, Chang TJ, Lui WY, P’Eng F K. Sex hormone receptors in gastric cancer. Cancer. 1990;65(6):1396–400. [DOI] [PubMed] [Google Scholar]
  • 22.Zhang B-g, Du T, Zang M-d, Chang Q, Fan Z-y, Li J-f, et al. Androgen receptor promotes gastric cancer cell migration and invasion via AKT-phosphorylation dependent upregulation of matrix metalloproteinase 9. Oncotarget. 2014;5(21):10584–95. [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]
  • 23.Zhang Y, Pan T, Zhong X, Cheng C. Androgen receptor promotes esophageal cancer cell migration and proliferation via matrix metalloproteinase 2. Tumour biology : the journal of the International Society for Oncodevelopmental Biology and Medicine. 2015;36(8):5859–64. [DOI] [PubMed] [Google Scholar]
  • 24.The Institute of Applied Health Sciences. PCCIU Background 2016 [cited 2016 02/11/16]. Available from: http://www.abdn.ac.uk/iahs/research/primary-care/pcciur/background.php.
  • 25.Chandrasoma P, Wickramasinghe K, Ma Y, DeMeester T. Adenocarcinomas of the distal esophagus and “gastric cardia” are predominantly esophageal carcinomas. The American journal of surgical pathology. 2007;31(4):569–75. [DOI] [PubMed] [Google Scholar]
  • 26.Rüdiger Siewert J, Feith M, Werner M, Stein HJ. Adenocarcinoma of the Esophagogastric Junction: Results of Surgical Therapy Based on Anatomical/Topographic Classification in 1,002 Consecutive Patients. Annals of Surgery. 2000;232(3):353–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Suissa S, Dell’aniello S, Vahey S, Renoux C. Time-window bias in case-control studies: statins and lung cancer. Epidemiology (Cambridge, Mass). 2011;22(2):228–31. [DOI] [PubMed] [Google Scholar]
  • 28.World Health Organization. The Anatomical Therapeutic Chemical Classification System With Defined Daily Doses 2016. [Available from: http://www.who.int/classifications/atcddd/en/.
  • 29.Khan NF, Perera R, Harper S, Rose PW. Adaptation and validation of the Charlson Index for Read/OXMIS coded databases. BMC Family Practice. 2010;11:1–. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Khan NF, Perera R, Harper S, Rose PW. Adaptation and validation of the Charlson Index for Read/OXMIS coded databases. BMC Family Practice. 2010;11(1):1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Hippisley-Cox J, Coupland C. Unintended effects of statins in men and women in England and Wales: population based cohort study using the QResearch database. Br Med J. 2010;340:12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Rothwell PM, Fowkes FGR, Belch JFF, Ogawa H, Warlow CP, Meade TW. Effect of daily aspirin on long-term risk of death due to cancer: analysis of individual patient data from randomised trials. Lancet. 2011;377(9759):31–41. [DOI] [PubMed] [Google Scholar]
  • 33.Singh S, Garg SK, Singh PP, Iyer PG, El-Serag HB. Acid-suppressive medications and risk of oesophageal adenocarcinoma in patients with Barrett’s oesophagus: a systematic review and meta-analysis. Gut. 2014;63(8):1229–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Ahn JS, Eom C-S, Jeon CY, Park SM. Acid suppressive drugs and gastric cancer: A meta-analysis of observational studies. World Journal of Gastroenterology : WJG. 2013;19(16):2560–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Nickel JC. Comparison of Clinical Trials With Finasteride and Dutasteride. Reviews in Urology. 2004;6(Suppl 9):S31–S9. [PMC free article] [PubMed] [Google Scholar]
  • 36.White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Statistics in medicine. 2011;30(4):377–99. [DOI] [PubMed] [Google Scholar]
  • 37.Lipsitch M, Tchetgen ET, Cohen T. Negative Controls: A Tool for Detecting Confounding and Bias in Observational Studies. Epidemiology (Cambridge, Mass). 2010;21(3):383–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Munoz J, Wheler JJ, Kurzrock R. Androgen receptors beyond prostate cancer: an old marker as a new target. Oncotarget. 2015;6(2):592–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Targarona EM, Pons MD, Gonzalez G, Boix L, Marco V, Marco C. Is exocrine pancreatic cancer a hormone-dependent tumor? A study of the existence of sex hormone receptors in normal and neoplastic pancreas. Hepato-gastroenterology. 1991;38(2):165–9. [PubMed] [Google Scholar]
  • 40.StataCorp. Stata Statistical Software: Release 13. College Station, TX: StataCorp LP; 2013. [Google Scholar]
  • 41.Elder R, Kirkpatrick M, Ramsay W, Macleod M, Guthrie B, Sutton M, et al. Measuring quality in primary medical services using data from SPICE. 2007. [Google Scholar]
  • 42.Boggon R, van Staa TP, Chapman M, Gallagher AM, Hammad TA, Richards MA. Cancer recording and mortality in the General Practice Research Database and linked cancer registries. Pharmacoepidemiology and drug safety. 2013;22(2):168–75. [DOI] [PubMed] [Google Scholar]
  • 43.Information Services Division (Scotland). Cancer Statistics [cited 2016 04/11/2016]. Available from: http://www.isdscotland.org/Health-Topics/Cancer/Cancer-Statistics/.
  • 44.Tihan T, Harmon JW, Wan X, Younes Z, Nass P, Duncan KL, et al. Evidence of androgen receptor expression in squamous and adenocarcinoma of the esophagus. Anticancer research. 2001;21(4b):3107–14. [PubMed] [Google Scholar]
  • 45.Roehrborn CG, Andriole GL, Wilson TH, Castro R, Rittmaster RS. Effect of dutasteride on prostate biopsy rates and the diagnosis of prostate cancer in men with lower urinary tract symptoms and enlarged prostates in the Combination of Avodart and Tamsulosin trial. European urology. 2011;59(2):244–9. [DOI] [PubMed] [Google Scholar]
  • 46.Andriole GL, Bostwick DG, Brawley OW, Gomella LG, Marberger M, Montorsi F, et al. Effect of Dutasteride on the Risk of Prostate Cancer. New England Journal of Medicine. 2010;362(13):1192–202. [DOI] [PubMed] [Google Scholar]
  • 47.Thompson IMJ, Goodman PJ, Tangen CM, Parnes HL, Minasian LM, Godley PA, et al. Long-Term Survival of Participants in the Prostate Cancer Prevention Trial. New England Journal of Medicine. 2013;369(7):603–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Murtola TJ, Tammela TL, Maattanen L, Ala-Opas M, Stenman UH, Auvinen A. Prostate cancer incidence among finasteride and alpha-blocker users in the Finnish Prostate Cancer Screening Trial. British journal of cancer. 2009;101(5):843–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Robinson D, Garmo H, Bill-Axelson A, Mucci L, Holmberg L, Stattin P. Use of 5α-reductase inhibitors for lower urinary tract symptoms and risk of prostate cancer in Swedish men: nationwide, population based case-control study. BMJ : British Medical Journal. 2013;346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50.Morales EE, Grill S, Svatek RS, Kaushik D, Thompson IM Jr., Ankerst DP, et al. Finasteride Reduces Risk of Bladder Cancer in a Large Prospective Screening Study. European urology. 2016;69(3):407–10. [DOI] [PubMed] [Google Scholar]
  • 51.Duijnhoven RG, Straus S, Souverein PC, de Boer A, Bosch J, Hoes AW, et al. Long-term use of 5 alpha-reductase inhibitors and the risk of male breast cancer. Cancer Causes & Control. 2014;25(11):1577–82. [DOI] [PubMed] [Google Scholar]
  • 52.Robinson D, Garmo H, Holmberg L, Stattin P. 5-alpha reductase inhibitors, benign prostatic hyperplasia, and risk of male breast cancer. Cancer Causes & Control. 2015;26(9):1289–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Cooper SC, Croft S, Day R, Thomson CS, Trudgill NJ. Patients with prostate cancer are less likely to develop oesophageal adenocarcinoma: could androgens have a role in the aetiology of oesophageal adenocarcinoma? Cancer causes & control : CCC. 2009;20(8):1363–8. [DOI] [PubMed] [Google Scholar]
  • 54.Joint Formulary Committee. Losartan. In: Joint Formulary Committee. British National Formulary.[BNF online] 2016. [cited 2017 13/04/17]. Available from: https://www.medicinescomplete.com/mc/bnflegacy/current/PHP1219-losartan-potassium.htm. [Google Scholar]
  • 55.Yuan JQ, Mao C, Wong SYS, Yang ZY, Fu XH, Dai XY, et al. Comparative Effectiveness and Safety of Monodrug Therapies for Lower Urinary Tract Symptoms Associated With Benign Prostatic Hyperplasia A Network Meta-analysis. Medicine (Baltimore). 2015;94(27):7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56.Naslund MJ, Miner M. A review of the clinical efficacy and safety of 5 alpha-reductase inhibitors for the enlarged prostate. Clin Ther. 2007;29(1):17–25. [DOI] [PubMed] [Google Scholar]
  • 57.de Jonge PJ, van Blankenstein M, Looman CW, Casparie MK, Meijer GA, Kuipers EJ. Risk of malignant progression in patients with Barrett’s oesophagus: a Dutch nationwide cohort study. Gut. 2010;59(8):1030–6. [DOI] [PubMed] [Google Scholar]
  • 58.Solaymani-Dodaran M, Logan RFA, West J, Card T, Coupland C. Risk of oesophageal cancer in Barrett’s oesophagus and gastro-oesophageal reflux. Gut. 2004;53(8):1070–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59.Hill AB. The Environment and Disease: Association or Causation? Proceedings of the Royal Society of Medicine. 1965;58(5):295–300. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES