On the relationship between androgen-deprivation therapy for prostate cancer and risk of infection by SARS-CoV-2

Abstract

information for conducting a meta-analysis. Our statement 'no statistically significant associations were reported for prostate and pancreatic cancers', which refers to cancer risk, is supported by summary estimates provided in the text and in Figure 3 (RR per 1 score 0.99, 95% CI 0.97e1.02; I 2 ¼ 0%, P het 0.43; six studies and 0.86, 95% CI 0.62e1.18; I 2 ¼ 90%, P het <0.001; two studies, respectively). Apart from lacking a doseeresponse relationship, the results for pancreatic cancer were only reported in two studies and were highly heterogeneous. Unfortunately, a meta-analysis was not possible for mortality data, because not enough studies assessed the relationship between meeting the 2007 World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) recommendations and survival in patients with pancreatic or prostate cancer (Table 2B). As already stated in our discussion, upcoming studies using the recently updated WCRF/AICR score are warranted to draw firmer conclusions regarding mortality or cancer survival. 2 Concerning the comment regarding the use of relative risks (RRs), we used this term to uniformly refer to risk estimates [i.e. hazard ratio (HR), risk ratio, odds ratio (OR)] in tables and figures. In particular, all cohort studies provided results in HRs and only caseecontrol studies used ORs. Thus, a primary meta-analysis synthesizing findings within the same study design does not have an issue of combining different measures of association. In addition, results for HR and OR were pooled to produce a summary estimate in the meta-analysis. This was done under the assumption that when events are rare and absolute cancer risks are small, these measures will be approximately equal in the studies. 3 Finally, we assessed between-study heterogeneity and small-study effects using standard methods (Cochran's Q, I 2 , Egger's regression asymmetry test and funnel plot), and we appreciate the efforts of Jayaraj et al. 1 to complement our analysis with additional statistics that make our findings more transparent.

The pivotal role of TMPRSS2 in coronavirus disease 2019 and prostate cancer

“…Another recent study in 19 high-volume medical oncology departments in Italy examined the frequency of COVID-19 in prostate cancer patients undergoing ADT, either alone or in combination with another agent [ 23 ]. A concomitant confirmed diagnosis of SARS-CoV-2 infection was found in 1.8% (36 out of 1949 total patients) of the examined population.…”

“…ADT did not seem to have a protective effect considering the higher lethality of SARS-CoV-2 in patients <70 years (25% in prostate cancer patients compared with 13% in infected Italian males). The findings of these two analyses on prostate cancer patients seem conflicting [ 22 , 23 ], supporting the need for further retrospective or prospective evidence on the role of antiandrogen agents in patients with COVID-19.…”

Clinical significance of sex hormones in COVID-19

“…Eleven of these 36 (30.6%) patients died from COVID-19 [21]. This lethality rate was similar and not lower than that of Italian men with similar age infected with COVID-19 [21].…”

“…Nevertheless, the increased mortality reported by Caffo, et al. [21] may be partly due to the fact that their patients had advanced metastatic prostate cancer.…”

“…Meanwhile, a third study from Italy did not support the latter concept. Caffo, et al [21] reported that 36 of 1949 (1.8%) with prostate cancer had COVID-19.…”