Does spironolactone reduce the risk of urinary tract cancers in hypertensive patients?

Hypertension and cancers are among the most common chronic medical conditions in middle-aged and elderly populations [1]. Hypertension has been believed to be a cancer risk factor, although the relationship still remains unclear. Cancers of the lung, prostate and genitourinary systems are common in the same middle-aged and elderly individuals who also have hypertension and in whom substantial proportions receive drug treatment [2]. As a result of the large numbers involved, the potential associated risk between antihypertension drugs and cancers is of large interest to the medical community and the general population. Occasionally, findings that a commonly used drug or more often class of similar drugs are associated with increased risk of the common cancers would lead to intense interest by the public. Indeed, reports that some hypertensive individuals whose BP had been well controlled had stopped taking their medications out of concern for the risk of cancer have been made anecdotally in the daily press. Therefore, questions may be asked what evidence is there for the association between the risk (or benefit) of cancer and the antihypertensive treatment and what is the likelihood that as both treatment for hypertension and cancers are common in the same middle-aged and elderly population, the association of risk may be due to chance? In the article by Chuang et al.[3] in this issue of the Journal of Hypertension, the authors reported that spironolactone was associated with a decreased risk of the urinary tract cancers (UTCs) – kidney, bladder and prostate cancer – whereas excess risk was observed with AT1 blockers (ARBs) and a neutral effect with angiotensin-converting enzyme (ACE) inhibitors [3]. The data for the study came from a national health insurance programme between 2005 and 2011, in which 32 167 UTCs were registered, and appropriate events tracking and data-linkage procedures were carried out. A retrospective case–control approach was adopted, with each eligible patient having both hypertension and UTC matched by sex, age (within 5 years), baseline comorbidities and year of UTC diagnosis. Methodology used in the analyses was generally of high quality and appropriate. Small but plausible reduced risks of UTC were observed with spironolactone [odds ratios (ORs) 0.91, 95% confidence interval (CI) 0.87–0.96] and excess risk with ARB (OR 1.22, 95% CI 1.18–1.26), whereas the effect of ACE inhibitors was neutral (OR 1.00, 95% CI 0.96–1.04). How would such findings be consistent with current knowledge? A suitable analogy may be made with the controversy regarding ARB, in which excess risk of cancers with ARB was reported in some studies [4,5] and in meta-analyses that selectively included some, but not all reports [6]. Others, including meta-analyses of outcomes reported in clinical trials [7–9] and data-linkage analysis of large databases [10], however, did not find such excess risk. The medical community, on balancing the benefits from well tolerated and effective control of hypertension using the ARB versus the small and uncertain potential risk of cancer, concluded that there was more to be gained in the continued use of ARBs in hypertension. As with ARBs and other antihypertension drugs, a large adequately powered long-term clinical trial addressing the benefit or risk of spironolactone on the cancer risks is desirable but very unlikely be carried out. Data addressing this question would have to depend on meta-analyses or retrospective database linkages. In the current study, the latter approach was followed. The authors have adopted a retrospective case–control study matching for sex, age and other baseline comorbidities to provide some comparability of background risk between the patients with cancer and controls. This is an appropriate approach but still does not get around the weaknesses inherent in the lack of randomization in a clinical trial. Although the case–control approach allows a degree of comparability, known confounders that could be adjusted and unknown confounders that could not be adjusted still leaves open a large possibility of the play of chance. This is an important consideration given that the magnitude of benefit was modest, confounders that are unknown or unadjusted could change the modest benefit to no benefit or harm. This could be the case with spironolactone in the study. Investigators who report on such findings of risk or benefit routinely invoke postulated pharmacological mechanisms of action of the agent to explain their observed findings. In the case of spironolactone, it was pointed out that spironolactone could concomitantly bind to sex hormone receptors and has a predominant affinity for androgen receptors. It was suggested that this androgen pathway could play a critical role in many hormone-related tumours, including those of prostrate, bladder, lung, breast and kidneys [11]. It was further suggested that androgen receptors positively influence renal cell cancer. This accounts for the high incidence rates in men. Targeting androgen receptors could inhibit cancer cell migration and invasion by modulating hypoxia-inducible factor 2 alpha/vascular endothelial growth factor (HIF2a/VEGF) signals [12]. The presence of androgen receptors in various genitourinary neoplasms shown in histological immunochemistry studies may explain the effects of androgen receptors in these tissues and may explain the reduced risk of prostate cancer in men and bladder cancer in women in this study [13]. Recent new postulated mechanisms on the antitumour effect of spironolactone include suggestions that this compound could contribute to the repair of some severe DNA damage and thus reduce the risk of tumour formation [14]. However, it is not known whether these postulated mechanisms are the basis for the anticancer effect of spironolactone and whether single isolated mechanisms could influence the overall development of cancer in the complex living organism in which many other systems interact in tandem. Possible mechanisms of action have similarly been used to explain the benefits and harm of other renin–angiotensin system (RAS) modulators. In the case of ARBs and ACE inhibitors, it was postulated that these antagonists of angiotensin II type I receptors exert protective effects by preventing tumour progression, vascularization and metastasis, whereas antagonism of the angiotensin II type 2 may be involved in the development of cancer. These mechanisms were used to explain their influence on cancer development, either protective or otherwise, but it is clear that by themselves, they cannot tell us which individual would or would not develop cancer. Thus, whether or not a postulated mechanism will bring about the desired effect probably does not just depend on one mechanism but on other multiple factors, some internal and others external and environmental, that would lead to the development and manifestation of cancers and other noncommunicable chronic diseases. It remains unknown whether the postulated mechanisms of benefit for spironolactone would be much more ‘positive’ than those for ACE inhibitors and ARBs. The next step at establishing whether or not the use of spironolactone for treatment of hypertension, heart failure and liver disease would also be associated with a reduction in risk of cancers, particularly UTC, would be to look at data from clinical trials. Ideally, these should be clinical trials of spironolactone on development of cancer, especially UTC. The trials should be properly designed, with adequate statistical power to detect a clinically useful difference (if any exists) and therefore must have adequate sample sizes. Cancer outcomes should be clearly defined, data systematically collected and events adjudicated. Such trials are unlikely to be possible or feasible. The fallback position would be to perform a meta-analysis to examine outcomes from the numerous trials of spironolactone designed to treat the other chronic conditions. As has been found in previous meta-analyses of ARB trials, how the meta-analysis is performed can result in findings that are inconclusive and confusing. This is not surprising given that definitions and ascertainment of cancer outcomes differ from one trial to another, the prevalence and incidence of cancer in the patients being studied in the trials may differ and the cancer incidence is generally low such that the numbers of events observed may not be adequate for any conclusive analysis of outcomes. In situations such as these, the risk of type I errors can be high. One additional common pitfall is the deliberate or inadvertent selective inclusion of trials, such as including data from trials that report high incidence of cancers with a particular treatment while omitting others in which cancer incidence is low or events are not reported. Findings from such analyses could be misleading. One may ask whether spironolactone is different from the other RAS-modulating agents and whether as hypertension and cancers are common in the middle-aged and elderly individuals, the association of hypertension or its treatment with cancers are coincidental. However, the androgen receptor blocking properties of spironolactone may be unique in this class of agents and may be involved in cancer suppression. Indeed, antiandrogenic therapy has been used in the treatment of prostate cancers. However, whether or not spironolactone can actually reduce UTC, as suggested in this study, needs further confirmatory studies and preferably randomized controlled clinical trials. As it stands, there is inadequate evidence to show that the finding as reported here is not due to a play of chance. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.