Impacts of androgen deprivation therapy on the risks and outcomes of SARS-CoV-2 infection in patients with prostate cancer

Abstract

Studies have investigated the effects of androgen deprivation therapy (ADT) use on the incidence and clinical outcomes of coronavirus disease 2019 (COVID-19); however, the results have been inconsistent. We searched the PubMed, Medline, Cochrane, Scopus, and Web of Science databases from inception to March 2022; 13 studies covering 84 003 prostate cancer (PCa) patients with or without ADT met the eligibility criteria and were included in the meta-analysis. We calculated the pooled risk ratios (RRs) with 95% confidence intervals (CIs) to explore the association between ADT use and the infection risk of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) and severity of COVID-19. After synthesizing the evidence, the pooled RR in the SARS-CoV-2 positive group was equal to 1.17, and the SARS-CoV-2 positive risk in PCa patients using ADT was not significantly different from that in those not using ADT (P = 0.544). Moreover, no significant results concerning the beneficial effect of ADT on the rate of intensive care unit admission (RR = 1.04, P = 0.872) or death risk (RR = 1.23, P = 0.53) were found. However, PCa patients with a history of ADT use had a markedly higher COVID-19 hospitalization rate (RR = 1.31, P = 0.015) than those with no history of ADT use. These findings indicate that ADT use by PCa patients is associated with a high risk of hospitalization during infection with SARS-CoV-2. A large number of high quality studies are needed to confirm these results.

INTRODUCTION

Infection with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes coronavirus disease 2019 (COVID-19), which results in damage to the pancreas, kidney, heart, brain, and especially the lungs.1 The World Health Organization (WHO) has declared that COVID-19 is a serious public health emergency of increasing international concern.2 By March 2022, SARS-CoV-2 was circulating in more than 150 countries and had affected over 464 million people, resulting in more than 5 000 000 deaths.3 Although several COVID-19 vaccines are currently available, vaccines have not increased in popularity worldwide due to substantial variation in acceptance rates among countries; additionally, some vaccines are less effective against the Delta and Omicron variants.4,5 No valid interventions have been officially identified to protect against the high morbidity and mortality rates of COVID-19 to date.6 Therefore, risk factors for SARS-CoV-2 infection and COVID-19 severity as well as the mechanism underlying disease invasion and progression have attracted extensive attention.

Acute lung injury and respiratory distress syndrome are common and serious complications in COVID-19 patients.7 These systemic effects are chiefly caused by the extensive distribution of angiotensin-converting enzyme 2 (ACE2).1 ACE2 is highly expressed in alveolar type II cells, renal tubular epithelial cells, Sertoli cells, and Leydig cells in the testis.8 SARS-CoV-2 uses ACE2 as a functional receptor to enter host cells.9 Hoffmann et al.10 reported that the entry of SARS-CoV-2 into host cells is dependent on the binding of the SARS spike (S) protein to ACE2 on cells and on SARS S protein priming by transmembrane protease serine 2 (TMPRSS2). The gene encoding TMPRSS2, an androgen-regulated protease from the family of type 2 transmembrane serine proteases,11 is characterized by several androgen receptor elements (AREs) upstream of the first intron and the transcription start site.12 The TMPRSS2 protease is principally expressed in the prostate, with relatively low expression levels in other organs, including the liver, testis, kidney, and pancreas; it is involved in multiple signaling pathways that influence the occurrence and development of tumors.13,14 In particular, TMPRSS2 plays a pivotal role in infection with influenza virus and coronavirus.12 The pathogenesis of viral spread can be summarized as a disturbance of antibody-mediated neutralization or the SARS S protein priming for cell–cell fusion/virus–cell membrane fusion via SARS S protein cleavage,12,15,16 as observed in SARS-CoV-2.17 Therefore, TMPRSS2 may be a potential target for developing new agents against COVID-19. According to in vitro evidence, inhibitors of TMPRSS2, such as camostat mesylate, are likely effective at reducing viral infections.10 Notably, the sex difference in androgen levels might be responsible for the increased susceptibility of men to COVID-19 compared with women, and androgen-driven immune modulation may be involved in SARS-CoV-2 infection. A likely mechanism is that the androgen/androgen receptor signaling pathway regulates TMPRSS2 expression and thereby mediates SARS-CoV-2 invasion.18

Moreover, the androgen-regulated protease TMPRSS2 exhibits markedly higher expression in primary and metastatic prostate tumors and plays a role in tumor differentiation, invasion, and metastasis.19,20 The growth and progression of prostate cancer (PCa), the second most common cancer in males worldwide,21 depend on androgen stimulation; inhibitors can suppress the development of PCa.22 Androgen deprivation therapy (ADT) is one of the basic methods of treating PCa. Additionally, studies have shown that androgens positively regulate the expression of TMPRSS2 not only in prostatic cells13,20 but also in lung-derived cell lines.23 Overall, exploration of methods of downregulating TMPRSS2 to target SARS-CoV-2 and thereby prevent infection and severity of COVID-19 via androgen suppression has clinical importance. Karimi et al.24 and Sari Motlagh et al.25 have performed systematic reviews and meta-analyses, but their conclusions might change substantially given the rapid publication of relevant studies. Accordingly, we have updated prior systematic reviews and meta-analyses by evaluating whether ADT for PCa exerts protective or pathogenic effects on SARS-CoV-2 infection and COVID-19 progression. We believe that this study provides an important contribution to understanding the risk and prognosis factors of COVID-19 as well as investigating potential methods for COVID-19 prevention or treatment.

MATERIALS AND METHODS

Search strategy

The protocol of our meta-analysis adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The corresponding checklist is shown in Supplementary Table 1. In addition, this systematic review has been registered in the International Prospective Register of Systematic Reviews (PROSPERO; registration ID: CRD42021260612). Further details of our meta-analysis can be found on the official website (https://www.crd.york.ac.uk/prospero). We systematically searched for publications from database inception to March 2022 in the PubMed, Medline, Cochrane, Scopus, and Web of Science databases without restrictions. The retrieval terms were as follows: “prostatic neoplasms”, “prostate neoplasms”, “cancer of prostate”, “PCa”, “coronavirus disease 2019”, “coronavirus disease-19”, “COVID-19”, “novel coronavirus 2019”, “severe acute respiratory syndrome coronavirus 2”, “SARS coronavirus 2”, “SARS-CoV-2”, “androgen deprivation therapy”, “ADT”, “endocrine therapy”, and “androgen inhibitor”. Potentially relevant records were subjected to a full-text review. Moreover, the reference lists of the selected articles were manually searched to identify other potentially relevant studies. This iterative process was continued until no additional articles could be identified.

Supplementary Table 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist

Section/topic	Number	Checklist item	Reported on page number
Title
Title	1	Identify the report as a systematic review, meta-analysis, or both	1
Abstract
Structured summary	2	Provide a structured summary including, as applicable: Background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number	1
Introduction
Rationale	3	Describe the rationale for the review in the context of what is already known	1-2
Objectives	4	Provide an explicit statement of questions being addressed with reference to PICOS design	2
Methods
Protocol and registration	5	Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number	2
Eligibility criteria	6	Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale	2
Information sources	7	Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched	2
Search	8	Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated	2
Study selection	9	State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis)	2
Data collection process	10	Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators	2
Data items	11	List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made	2
Risk of bias in individual studies	12	Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis	2
Summary measures	13	State the principal summary measures (e.g., risk ratio, difference in means)	2
Synthesis of results	14	Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I2) for each meta-analysis	2
Risk of bias across studies	15	Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies)	2
Additional analyses	16	Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified	2
Results
Study selection	17	Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram	2; Figure 1
Study characteristics	18	For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations	2-3; Table 1; Supplementary Tables 2 and 3
Risk of bias within studies	19	Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12)	4
Results of individual studies	20	For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot	3-4; Figures 2–5
Synthesis of results	21	Present results of each meta-analysis done, including confidence intervals and measures of consistency	3-4; Figures 2–5
Risk of bias across studies	22	Present results of any assessment of risk of bias across studies (see item 15)	4; Supplementary Figure 1 (251.2KB, tif) ; Supplementary Table 4
Additional analysis	23	Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see item 16])	3-4; Figures 2–5
Discussion
Summary of evidence	24	Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policy makers)	5-6
Limitations	25	Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias)	7
Conclusions	26	Provide a general interpretation of the results in the context of other evidence, and implications for future research	7
Funding
Funding	27	Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review	7

PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses, PICOS: participants, interventions, comparisons, outcomes, and study

Selection criteria

The inclusion criteria were as follows: (1) population: PCa patients; (2) intervention: any type of ADT treatment; (3) comparison: no ADT treatment; (4) outcome: SARS-CoV-2 positive risk, the rates of hospitalization and intensive care unit (ICU) admission, and death risk; and (5) design: randomized controlled trials/observational studies that provided adequate data for the meta-analysis.

Data extraction

Two researchers (YBH and WLL) independently screened the abstracts and full texts of the included studies, extracted the data, and assessed the risk of bias. The data extracted from the included studies were the author, country, study design, time span, sample size, age, ADT treatment (i.e., type, frequency, duration, and dosage), duration of therapy on COVID-19, chemotherapy, castration resistance, comorbidities, reported indicators, and risk ratios (RRs). Any divergence was resolved through discussion and consensus. Further disagreement during this process was resolved by consensus through involving the third and fourth authors (MS and XCL).

Quality assessment

The most recent version of the Cochrane tool (ROBINS-I) was used to assess the risk of bias in our meta-analyses.26 The confounding bias, selection bias, intervention bias, bias due to missing data or outcome measurements, and reporting bias were all taken into account. In addition, the quality of evidence was assessed via the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) approach.27,28 Five limitations were assessed, including the risk of bias, inconsistency, indirectness, imprecision, and publication bias. Two researchers (YBH and WLL) individually evaluated studies based on the above domains. Then, using GRADEpro software (version 3.6; McMaster University, Hamilton, Ontario, Canada), we classified the overall quality of evidence as “very low”, “low”, “moderate”, or “high”. A higher classification corresponds to more reliable evidence.

Statistical analyses

In this study, PCa cases were classified as those with or without ADT. Forest plots were created to illustrate the incidence and severity of COVID-19 in ADT and non-ADT groups, as described by RRs and 95% confidence intervals (CIs). Heterogeneity was assessed by the I² statistic and categorized as low, medium, or high; the lower limits of I² in each category were 25%, 50%, and 75%, respectively. We applied a random-effects model if there was significant heterogeneity; otherwise, we used a fixed-effects model. The heterogeneity in each study was evaluated using the leave-one-out sensitivity analysis method. Potential publication bias was evaluated by Begg’s and Egger’s tests. All analyses were conducted using Stata 13.0 statistical software (Stata Corporation, College Station, TX, USA). P < 0.05 was considered statistically significant.

RESULTS

Description of study selection and characteristics

In total, 714 potentially relevant studies were identified using the search terms described above. After excluding duplicated studies, 168 publications remained. Subsequently, we excluded irrelevant studies by screening titles and abstracts (n = 96); afterward, 72 studies remained and their full texts were reviewed. Then, 59 publications were excluded for the following reasons: (1) review articles or comments without data (n = 20); (2) animal or molecular biology studies (n = 4); (3) insufficient data reported to calculate RR and 95% Cl (n = 21); and (4) no relevant outcomes (n = 14). Finally, 13 articles met our eligibility criteria.11,29,30,31,32,33,34,35,36,37,38,39,40 The characteristics of these studies are described in detail in Table 1, and a flowchart of study selection following the PRISMA guidelines is provided in Figure 1.

Table 1.

Main characteristics of eligible studies included in the meta-analysis concerning impact of androgen deprivation therapy on risk of incidence and severity of COVID-19 in PCa patients

Study	Country	Study type	Recruitment period	Total cases (n)	Median or range of age (year)	Chemotherapy cases (n)	Castration resistance cases (n)	Reported indicators	Adjusted RR (95% CI)
Kwon et al.29 2021	USA	Cohort study; retrospective	February 2020–December 2020	5211	NR	NR	NR	SARS-CoV-2 positive	1.18 (0.70–1.99)
								Death	0.56 (0.07–4.88)
Duarte et al.30 2021	Brazil	Cohort study; retrospective	January 2020–April 2021	199	75	NR	NR	Death	0.63 (0.29–1.37)
Gedeborg et al.31 2022	Sweden	Case-control study; retrospective	December 2019–December 2020	24 174	51–97	NR	NR	Death	5.05 (4.18–6.10)
Welén et al.32 2022	Sweden	Cohort study; retrospective	July 2020–November 2020	5824	50–81	NR	NR	Hospitalization	1.24 (0.97–1.59)
								ICU admission	0.87 (0.51–1.48)
								Death	0.89 (0.54–1.46)
Kazan et al.33 2021	Turkey	Cohort study; retrospective	August 2020–June 2021	365	NR	NR	NR	SARS-CoV-2 positive	0.68 (0.34–1.36)
								Hospitalization	3.36 (0.61–18.58)
Jiménez-Alcaide et al.34 2021	Spain	Cohort study; retrospective	March 2020–May 2020	1349	64.2	1	1	Death	0.67 (0.26–1.74)
Patel et al.352021	USA	Cohort study; retrospective	March 2020–May 2020	465	61–81	NR	NR	Hospitalization	1.07 (0.61–1.87)
								Death	1.28 (0.79–2.08)
Schmidt et al.362021	USA	Cohort study; retrospective	May 2020–February 2021	1228	73	266	NR	Death	0.77 (0.42–1.42)
Tucker et al.37 2021	USA	Cohort study; retrospective	December 2019–July 2020	589	NR	NR	NR	Death	1.41 (0.99–2.01)
Koskinen et al.38 2020	Finland	Cohort study; retrospective	March 2020–May 2020	352	51–96	NR	NR	SARS-CoV-2 positive	0.88 (0.32–2.44)
Klein et al.39 2021	USA	Cohort study; prospective	March 2020–June 2020	1779	38–101	NR	NR	SARS-CoV-2 positive	0.90 (0.54–1.61)
								Hospitalization	2.00 (0.57–7.32)
								ICU admission	1.30 (0.38–4.09)
								Death	2.90 (0.86–9.81)
Montopoli et al.11 2020	Italy	Cohort study; retrospective	December 2019–April 2020	42 434	NR	NR	NR	SARS-CoV-2 positive	4.05 (1.55–10.59)
								Hospitalization	3.93 (1.31–11.77)
								ICU admission	4.40 (0.76–25.50)
Caffo et al.40 2020	Italy	Cohort study; retrospective	February 2020–June 2020	34	75	9	34	Death	1.17 (0.17–8.09)

No studies reported the duration of therapy on COVID-19. COVID-19: coronavirus disease 2019; RR: risk ratio; NR: not reported; SARS-CoV-2: severe acute respiratory syndrome coronavirus 2; ICU: intensive care unit; CI: confidence interval

Figure 1.

The PRISMA flowchart of the systematic review and meta-analysis. RR: risk ratio; CI: confidence interval; ADT: androgen deprivation therapy; SARS-CoV-2: severe acute respiratory syndrome coronavirus 2; ICU: intensive care unit; PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

The studies included in our meta-analysis encompassed a total of 84 003 PCa patients. Regarding region, 5 studies were from the USA29,35–37,39 (including 9272 PCa patients), 2 studies were from Sweden31,32 (29 998 PCa patients), 2 studies were from Italy11,40 (42 468 PCa patients), 1 study was from Brazil30 (199 PCa patients), 1 study was from Turkey33 (365 PCa patients), 1 study was from Spain34 (1349 PCa patients), and 1 study was from Finland38 (352 PCa patients). All the studies were retrospective cohort studies, except one prospective cohort study conducted by Klein et al.39 and one case–control study conducted by Gedeborg et al.31 In total, 276 PCa patients underwent chemotherapy, and 35 PCa patients progressed to castration resistance. Additionally, 7 studies reported ADT administration details (Supplementary Table 2).29,30,31,32,35,36,38 Of these, most (n = 4) reported that ADT involved a gonadotrophic-releasing hormone (GnRH) analog/antagonist, surgical castration, and anti-androgen/anti-androgen receptor drugs.31,32,35,38

Supplementary Table 2.

Characteristics of androgen deprivation therapy usage in eligible studies

First author (year)	ADT types					ADT selection	Duration of ADT use	Assessed variables
	①	②	③	④	⑤
	BAT	GnRH analogue	GnRH antagonist	Surgical castration	Anti-androgen/anti-AR
	Kwon et al.29 2021	NR	√	√	NR				NR	②+③	≤6 months	SARS-CoV-2 positive Death
Duarte et al.30 2021	NR	√	NR	√	NR	②+④	NR	Death
Gedeborg et al.31 2022	NR	√	√	√	√	②+③+④+⑤	NR	Death
Welén et al.32 2022	NR	√	√	√	√	②+③+④+⑤	NR	Hospitalization ICU admission Death
Kazan et al.33 2021	NR	NR	NR	NR	NR	NR	NR	SARS-CoV-2 positive Hospitalization
Jiménez-Alcaide et al.34 2021	NR	NR	NR	NR	NR	NR	NR	Death
Patel et al.35 2021	NR	√	√	√	√	②+③+④+⑤	≤6 months	Hospitalization Death
Schmidt et al.36 2021	NR	√	√	√	NR	②+③+④	≤3 months	Death
Tucker et al.37 2021	NR	NR	NR	NR	NR	NR	NR	Death
Koskinen et al.38 2020	NR	√	√	√	√	②+③+④+⑤	NR	SARS-CoV-2 positive
Klein et al.39 2021	NR	NR	NR	NR	NR	NR	NR	SARS-CoV-2 positive Hospitalization ICU admission Death
Montopoli et al.11 2020	NR	NR	NR	NR	NR	NR	NR	SARS-CoV-2 positive Hospitalization ICU admission
Caffo et al.40 2020	NR	NR	NR	NR	NR	NR	50 months	Death

ADT: androgen deprivation therapy; BAT: bipolar androgen therapy; GnRH: gonadotropin-releasing hormone; AR: androgen receptor; NR: not reported; SARS-CoV-2: severe acute respiratory syndrome coronavirus 2; ICU: intensive care unit

Ten studies reported comorbidities (Supplementary Table 3),29,30,31,32,33,34,36,38,39,40 and six of these enrolled ADT and non-ADT populations with various comorbidities.30,33,34,36,38,39 For example, in the study of Duarte et al.,30 patients receiving ADT were more likely to have kidney disease (RR = 0.37, 95% CI: 0.17–0.81; P = 0.010), and in the study of Klein et al.,39 patients receiving ADT were more likely to have a respiratory disease (RR = 0.65, 95% CI: 0.44–0.94; P = 0.020), a history of smoking (RR = 1.15, 95% CI: 1.05–1.25; P = 0.002), or immunosuppressive diseases (RR = 1.25, 95% CI: 1.04–1.49; P = 0.010).

Supplementary Table 3.

Stratified analysis between androgen deprivation therapy and comorbidities in eligible studies

First author (year)	Comorbidities
	Diabetes mellitus	Kidney disease	Respiratory disease	Obesity	Smoking	Immune suppressive disease	Hypertension	Cardiovascular disease	Neurological disease
Kwon et al.29 2021	763	658	321	340	NR	NR	NR	577	NR
Duarte et al.30 2021	ADT (+): 37 ADT (−): 12 RR: 0.85 95% CI: 0.49–1.48 P=0.570	ADT (+): 12 ADT (−): 9 RR: 0.37 95% CI: 0.17–0.81 P=0.010	ADT (+): 17 ADT (−): 6 RR: 0.78 95% CI: 0.33–1.86 P=0.580	NR	NR	NR	NR	ADT (+): 60 ADT (−): 21 RR: 0.79 95% CI: 0.55–1.13 P=0.200	ADT (+): 8 ADT (−): 3 RR: 0.74 95% CI: 0.20–2.65 P=0.640
Gedeborg et al.31 2022	2458	NR	1320	NR	NR	NR	NR	2184	NR
Welén et al.32 2022	8	NR	3	NR	NR	NR	13	6	NR
Kazan et al.33 2021	ADT (+): 38 ADT (−): 60 RR: 1.04 95% CI: 0.74–1.47 P=0.820	NR	ADT (+): 16 ADT (−): 13 RR: 2.02 95% CI: 1.00–4.08 P=0.050	NR	ADT (+): 90 ADT (−): 145 RR: 1.02 95% CI: 0.87–1.19 P=0.790	NR	ADT (+): 70 ADT (−): 107 RR: 1.08 95% CI: 0.87–1.33 P=0.500	ADT (+): 31 ADT (−): 39 RR: 1.31 95% CI: 0.86–1.99 P=0.210	NR
Jiménez-Alcaide et al.34 2021	NR	NR	ADT (+): 0 ADT (−): 13 RR: 0.16 95% CI: 0.01–2.47 P=0.19	ADT (+): 1 ADT (−): 9 RR: 0.51 95% CI: 0.07–3.59 P=0.49	ADT (+): 3 ADT (−): 11 RR: 1.24 95% CI: 0.41–3.71 P=0.70	NR	ADT (+): 6 ADT (−): 41 RR: 0.55 95% CI: 0.29–1.07 P=0.08	ADT (+): 0 ADT (−): 9 RR: 0.22 95% CI: 0.01–3.58 P=0.29	NR
Patel et al.35 2021	NR	NR	NR	NR	NR	NR	NR	NR	NR
Schmidt et al.36 2021	ADT (+): 40 ADT (−): 80 RR: 0.91 95% CI: 0.66–1.27 P=0.580	ADT (+): 24 ADT (−): 54 RR: 0.81 95% CI: 0.52–1.26 P=0.350	ADT (+): 21 ADT (−): 46 RR: 0.83 95% CI: 0.51–1.35 P=0.450	NR	ADT (+): 86 ADT (−): 151 RR: 1.04 95% CI: 0.86–1.25 P=0.700	NR	NR	ADT (+): 70 ADT (−): 114 RR: 1.12 95% CI: 0.89–1.41 P=0.340	NR
Tucker et al.37 2021	NR	NR	NR	NR	NR	NR	NR	NR	NR
Koskinen et al.38 2020	ADT (+): 16 ADT (−): 17 RR: 1.53 95% CI: 0.80–2.93 P=0.200	NR	ADT (+): 8 ADT (−): 12 RR: 1.08 95% CI: 0.46–2.58 P=0.850	NR	ADT (+): 18 ADT (−): 17 RR: 1.72 95% CI: 0.92–3.22 P=0.090	NR	ADT (+): 30 ADT (−): 47 RR: 1.04 95% CI: 0.69–1.56 P=0.850	ADT (+): 51 ADT (−): 71 RR: 1.17 95% CI: 0.88–1.56 P=0.290	NR
Klein et al.39 2021	ADT (+): 107 ADT (−): 466 RR: 1.11 95% CI: 0.94–1.32 P=0.210	NR	ADT (+): 28 ADT (−): 210 RR: 0.65 95% CI: 0.44–0.94 P=0.020	NR	ADT (+): 207 ADT (−): 875 RR: 1.15 95% CI: 1.05–1.25 P=0.002	ADT (+): 104 ADT (−): 405 RR: 1.25 95% CI: 1.04–1.49 P=0.010	ADT (+): 252 ADT (−): 1194 RR: 1.02 95% CI: 0.97–1.08 P=0.410	ADT (+): 186 ADT (−): 925 RR: 0.98 95% CI: 0.88–1.08 P=0.620	NR
Montopoli et al.11 2020	NR	NR	NR	NR	NR	NR	NR	NR	NR
Caffo et al.40 2020	4	2	3	NR	NR	NR	20	12	NR

ADT: androgen deprivation therapy; +: use; −: no use; RR: risk ratio; CI: confidence interval; NR: not reported

Association between ADT use and the risk of SARS-CoV-2 positive in PCa patients

To evaluate the impact of ADT use to treat PCa on the risk of SARS-CoV-2 positive, we included five studies11,29,33,38,39 on 50 141 PCa patients. The results of the heterogeneity test showed moderate heterogeneity (I² = 60.1%; P = 0.040); thus, a random-effects model was applied. The individual heterogeneity of each study is shown in Figure 2a. There was no significant difference in SARS-CoV-2 positive between the ADT and non-ADT groups (P = 0.544), indicating that ADT use did not protect the patients against infection with SARS-CoV-2 (RR = 1.17, 95% CI: 0.71–1.92; Figure 2b).

Figure 2.

The correlation between androgen deprivation therapy and SARS-CoV-2 positive risk was illustrated by (a) leave-one-out sensitivity analysis and (b) a forest plot. The best models were applied to estimate each meta-analysis. For the forest plot, the line perpendicular to the X-axis indicates an RR equal to 1 (baseline). The RR values and 95% CIs of individual studies are represented by each enclosed circle and horizontal line parallel to the X-axis, respectively. Overlap between the baseline and a diamond indicates no statistical significance. CI: confidence interval; RR: risk ratio; SARS-CoV-2: severe acute respiratory syndrome coronavirus 2.

Association between ADT use and the rate of COVID-19 hospitalization in PCa patients

A total of 50 867 PCa patients from 5 studies11,32,33,35,39 concerning the relationship between ADT and the rate of COVID-19 hospitalization were included in the meta-analysis. As shown in Figure 3a, we detected no significant individual heterogeneity. The overall heterogeneity of these studies (I² = 34.7%; P = 0.190) suggested that a fixed-effects model could be used for analyses. According to the combined outcomes, there was a significant association between ADT use and a high rate of COVID-19 hospitalization in PCa patients (RR = 1.31, 95% CI: 1.05–1.62; P = 0.015; Figure 3b).

Figure 3.

The correlation between androgen deprivation therapy and the rate of COVID-19 hospitalization was analyzed by (a) leave-one-out sensitivity analysis and (b) forest plot. CI: confidence interval; RR: risk ratio; COVID-19: coronavirus disease 2019.

Association between ADT use and the rate of ICU admission for COVID-19 in PCa patients

Additionally, we collected data from 50 037 patients from three eligible studies11,32,39 and conducted a meta-analysis to explore whether the use of ADT correlated with the rate of emergency medical care necessitated for COVID-19. The heterogeneity test and leave-one-out sensitivity analysis did not reveal significant overall or individual heterogeneity (I² = 36.6%; P = 0.206; Figure 4a). Our meta-analysis applied a fixed-effects model, revealing that ADT use had no significant effect on ICU admission resulting from COVID-19 (95% CI: 0.65–1.66; P = 0.872), with the pooled RR of 1.04 (Figure 4b).

Figure 4.

The correlation between androgen deprivation therapy and the rate of COVID-19 ICU admission was shown in (a) leave-one-out sensitivity analysis and (b) forest plot. CI: confidence interval; RR: risk ratio; COVID-19: coronavirus disease 2019; ICU: intensive care unit.

Association between ADT use and the death risk of COVID-19 in PCa patients

Death is the most unfavorable outcome when assessing the severity of COVID-19; the death risk due to COVID-19 was explored in 10 studies29,30,31,32,34,35,36,37,39,40 including 40 852 PCa patients. Among these studies, we found high heterogeneity (I² = 92.7%; P < 0.001); therefore, we utilized a random-effects model. The outcomes of the leave-one-out sensitivity analysis are shown in Figure 5a. No significant difference (RR = 1.23, 95% CI: 0.64–2.38; P = 0.53) in the death risk was observed between the ADT and non-ADT groups (Figure 5b).

Figure 5.

The correlation between androgen deprivation therapy and the death risk of COVID-19 was illustrated by (a) leave-one-out sensitivity analysis and (b) forest plot. CI: confidence interval; RR: risk ratio; COVID-19: coronavirus disease 2019.

Publication bias and quality assessment

Begg’s and Egger’s tests were performed in the four groups of studies, and none exhibited significant publication bias (all P > 0.05 for the SARS-CoV-2 positive, COVID-19 hospitalization, ICU admission, and death groups). The ROBINS-I checklists of individual studies included in these four meta-analyses (the above groups) are shown in Supplementary Figure 1 ^{(251.2KB, tif)} . We determined that our meta-analyses had a low-to-moderate risk of bias. Moreover, details on the quality of evidence for each meta-analysis according to the GRADE approach are summarized in Supplementary Table 4. In the meta-analysis of the risk of SARS-CoV-2 positive, a low quality of evidence was found due to high overall heterogeneity. In the meta-analysis of the risk of COVID-19 hospitalization, the range of values was excessively large; therefore, the evidence of outcomes related to the rate of hospitalization was determined to be of low quality. For the meta-analysis of ICU admissions due to COVID-19, the range of true values in each pooled study was excessively large and variable, which severely impacted the GRADE score and resulted in a very low quality of evidence. Similarly, in the meta-analysis exploring the relationship between ADT use and the death risk, imprecision and inconsistency resulted in very low quality of evidence.

Supplementary Table 4.

The overall quality of evidence in pooled findings from eligible studies^a

Outcomes	Number of studies	Design	Certain quality assessment					Summary of findings
			Risk of bias	Inconsistency	Indirectness	Imprecision	Other considerations	Relative effect, RR (95% CI)	Quality
Risk of SARS-CoV-2 positive	5	Observational studies	Not serious	Seriousb	Not serious	Not serious	None	1.17 (0.71–1.92)	Low
Rate of hospitalization	5	Observational studies	Not serious	Not serious	Not serious	Seriousc	None	1.31 (1.05–1.62)	Low
Rate of ICU admission	3	Observational studies	Not serious	Not serious	Not serious	Very seriousc	None	1.04 (0.65–1.66)	Very low
Death risk	10	Observational studies	Not serious	Seriousb	Not serious	Seriousc	None	1.23 (0.64–2.38)	Very low

^aThe quality assessment was based on GRADE approach; ^bHigh levels of heterogeneity in pooled findings; ^cThe range of true values in each pooled studies was too large in varying degrees. GRADE: Grading of Recommendations assessment, Development and Evaluation, SARS-CoV-2: severe acute respiratory syndrome coronavirus 2; ICU: intensive care unit; CI: confidence interval; RR: risk ratio

DISCUSSION

The sheer magnitude of the COVID-19 outbreak and pandemic has triggered an unprecedented global health emergency. The potential role of testosterone and other androgens in COVID-19 occurrence and development has been highly controversial. Worse clinical outcomes and higher mortality of COVID-19 have been reported among the male population in many countries,41 and this discrepancy may be largely due to hormonal differences. The expression of TMPRSS2 is mediated by testosterone, and TMPRSS2 plays a vital role in proteolysis of the SARS S protein and binding of SARS-CoV-2 to ACE2, possibly explaining the increased severity of COVID-19 in males.42 Additionally, Kissick et al.43 indicated that high levels of circulating testosterone had immunosuppressive effects, as it inhibited B-cell growth and the differentiation of CD4⁺ T cells, thereby impairing the immune response and reducing the production of immunoglobulin. The effects could increase the probability of SARS-CoV-2 invasion.

However, the level of circulating preoperative testosterone may also predict biochemical recurrence (BCR) in patients who received radical prostatectomy (RP);44 thus, beneficial effects of testosterone on COVID-19 have been reported by an increasing number of studies. High testosterone levels are associated with a shorter duration of SARS-CoV-2 viral positive, while low testosterone levels can lead to poor prognosis, worse clinical COVID-19 phenotype, and higher mortality rate in SARS-CoV-2 positive male patients.45,46,47 Rastrelli et al.48 indicated that lower baseline levels of testosterone predicted higher likelihood of poor prognosis and mortality rate by analyzing male patients with COVID-19 pneumonia. Montaño et al.49 found that sufficient levels of testosterone exerted a protective effect on the respiratory system, improving pulmonary ventilation and respiratory muscle contraction; in contrast, low levels of testosterone during hospitalization may be a risk factor for clinical disease severity and inflammation.50,51,52 These findings can be explained by the inhibitory effect of testosterone on proinflammatory cytokine production. In acute multiorgan damage induced by COVID-19, crucial proinflammatory cytokines, including interleukin (IL)-6, IL-1β, and tumor necrosis factor-α (TNF-α), rapidly lead to a severe inflammatory cytokine storm.53 Testosterone exerts a protective effect by inhibiting these cytokines.54 Papadopoulos et al.55 and Sun et al.56 indicated that a high level of IL-6 could lead to increased COVID-19 severity but testosterone inhibited IL-6 synthesis and released and downregulated IL-6 receptor expression. In addition, testosterone inhibits the nuclear factor kappa-B (NF-κB) inflammatory signaling pathway, thus reducing lung inflammation and fibrosis.57 In sum, the possible effects of testosterone on COVID-19 are very complicated and span a wide spectrum.

The relationship between androgen levels and PCa has been fully elucidated.58 ADT, which reduces androgen levels, is the mainstay of therapy in advanced or metastatic PCa patients.59 However, routine ADT therapies, such as GnRH analogs, GnRH antagonists, and surgical castration, cannot permanently prevent the progression of castration-resistant PCa. Therefore, androgen receptor signaling blocking agents, such as enzalutamide and abiraterone, have been developed and applied in the treatment of metastatic castration-resistant PCa.60 Enzalutamide has antagonism against the overexpression of androgen receptor, and abiraterone downregulates the expression of cytochrome P450, family 17, subfamily A, polypeptide 1 (CYP17A1), which is an androgen synthesis enzyme.61,62 Both of these agents inhibit androgen receptor signaling, interfering with the malignant biological behavior of metastatic castration-resistant PCa. Inspired by the findings above, we investigated whether ADT can regulate the androgen receptor pathway, mediate TMPRSS2 transcription, and impede SARS-CoV-2 from infecting host cells. Montopoli et al.11 found that PCa patients who received ADT had a lower risk of SARS-CoV-2 positive and ICU admission than patients who did not receive ADT, indicating that ADT may protect against SARS-CoV-2 infection in PCa patients. However, a subsequent study conducted by Kwon et al.,29 including 5211 PCa patients with COVID-19, observed the opposite pattern, with COVID-19 more prevalent in PCa patients who received ADT. In addition, several studies did not find that ADT was associated with reductions in the mortality rates of PCa patients who were SARS-CoV-2 positive.39,40 Given these contradictory results, a systematic review is needed to verify the findings and update the prior hypothesis. We established inclusion and exclusion criteria and conducted a comprehensive literature search, extracting data from 13 eligible studies to conduct four meta-analyses. Significant differences in the hospitalization rates (RR = 1.31, P = 0.015) were identified between PCa patients who received ADT and those who did not. Our findings indicated that ADT use by PCa patients was associated with a high risk of COVID-19-related hospitalization.

There are several possible explanations for this finding. First, ADT resistance develops rapidly, either to routine therapy or enzalutamide/abiraterone, resulting in further progression of metastatic castration-resistant PCa; this could influence the rate of hospitalization of PCa patients with COVID-19. Additionally, reactivation of the androgen receptor signaling pathway is one of the main factors underlying metastatic castration-resistant PCa.63 Mediated by resistance to enzalutamide/abiraterone, overexpression and mutations of androgen receptors or high levels of androgen receptor splice variants could also cause SARS-CoV-2 invasion.64 Second, according to stratification analyses based on various comorbidities, the use of ADT was significantly associated with high incidence rates of comorbidities, such as a history of smoking and immunosuppressive diseases, in two studies.30,39 These comorbidities would also yield increased risks of hospitalization. Moreover, the patient population included in our study was mainly comprised of elderly males. A reduction in serum testosterone levels and related risky health habits or basic diseases, such as smoking and hypertension, are prevalent in populations of older male patients. These factors also cause higher rates of hospitalization for COVID-19. Third, as we discussed previously, androgen is a double-edged sword regarding the rapid progression of COVID-19. Testosterone exerts diverse roles on different components of the immune system: low levels of testosterone may have protective effects in the early stages of infection but lead to worse outcomes in advanced COVID-19 cases.46 Therefore, we hypothesized that in the early stage, COVID-19 patients exhibit a relatively weak inflammatory response. As ADT inhibits the proinflammatory effects of testosterone, it may downregulate the levels of TMPRSS2 and prevent SARS-CoV-2 from replicating in host cells. In advanced stages of COVID-19, SARS-CoV-2 has undergone massive viral replication in the host. A hyperactive inflammatory response is produced, causing a series of pneumonia-like symptoms, such as fever, coughing, sneezing, and leukocytosis. Patients are more likely to come to the hospital for treatment and hospitalization in this stage. Finally, in COVID-19 patients exhibiting acute multiorgan damage, a severe inflammatory cytokine storm is rapidly activated. At this point, control of the inflammatory response is more important than preventing viral replication.

Some studies have indicated that prostate inflammation is related to the severity of PCa. From macroscopic and clinical perspectives, inflammatory-related factors, such as the neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and eosinophil-to-lymphocyte ratio (ELR), are significantly correlated with increases in the Gleason score in PCa patients who underwent active surveillance and RP.65 Moreover, a single-center study compared histological specimens from 260 patients with or without active surveillance and prostate biopsy after RP; the outcomes revealed that systemic inflammatory markers in the prostate, consisting of the NLR, PLR, and monocyte-to-lymphocyte ratio (MLR), may be associated with a higher risk of adverse pathology.66 In addition, higher levels of testosterone predict shorter durations of SARS-CoV-2 viral positive,47 and ADT reduces testosterone secretion, possibly mitigating the damaging effects of inflammation due to COVID-19. From a microscopic, molecular biology perspective, abnormal levels of testosterone influence reactive oxygen species (ROS) production, cellular senescence, autophagy, and mitophagy, as well as vitamin D levels. Higher levels of ROS may activate nucleotide-like receptors, especially NOD-like receptor protein 3 (NLRP3), a member of the nucleotide-like receptor family, and induce pyroptosis. Steroidogenic acute regulatory protein (StAR) could act as a barrier to pyroptosis by inducing testosterone production. The crosstalk between diverse SARS-CoV-2-related immune cells in the microenvironment and inflammatory cytokines is listed in Supplementary Figure 2 ^{(128.3KB, tif)} .67,68,69,70,71,72,73 Furthermore, testosterone inhibits the hypersecretion of adipokines, including TNF-α and IL-6, in senescent adipocytes. Testosterone-related immunosenescence may alter the ratio of CD4⁺/CD8⁺ T cells; interfere with the distributions of immature T cells, memory T cells, the Th1 response (IL-2 and interferon [IFN]-γ), and the Th2 response (IL-4); and modify the production of proinflammatory cytokines, including IL-2, IL-8, IL-17, monocyte chemoattractant protein 1 (MCP1), TNF-α, and chemokine (C-C motif) ligand 3 (CCL3). Since studies have reported a significant positive association between testosterone levels and 25-hydroxyvitamin D levels, shorted as 25(OH) D levels in elderly patients with COVID-19, testosterone is also implicated in modulating 25(OH) D levels.74,75 After 25(OH) D combines with the vitamin D receptor, it can influence the autophagic activity of macrophages, increase antimicrobial products, and reduce the expression of proinflammatory cytokines, including TNF-α, IL-8, IL-13, and IFN-γ.74,75,76,77,78 Additionally, male patients with long-term low testosterone levels may have reduced muscle mass and strength, which could lead to progressively reduced lung capacity and excessive reliance on ventilators.79,80,81 This may be the primary explanation for the association between lower testosterone levels after long-term treatment with ADT and increases in the hospitalization rate of COVID-19 patients. Moreover, ADT has severe side effects, including anaphylactoid reactions, shock, diabetes, cerebral hemorrhage, and cardiac diseases, which may indirectly lead to the hospitalization of COVID-19 patients.82

We acknowledge that shortcomings and limitations in the present study may have affected our findings. First, since individuals have different rates of exposure to sources of infection during a pandemic, the actual rate of SARS-CoV-2 infection is challenging to determine. Second, although all authors participated in the systematic literature review, development of inclusion and exclusion criteria, and crosschecking of the included studies, selection bias with regard to the inclusion and exclusion of certain studies was unavoidable. Third, in terms of the inclusion criteria, studies did not control for baseline information, such as PCa stage, medical history, and the age of patients. Additionally, SARS-CoV-2 positive has a wide range of clinical manifestations, from asymptomatic cases to complicated cases. The main limitation of the statistical power of our meta-analyses was a shortage of relevant studies that reported the data of interest. Moreover, further consideration of the eligibility of studies included in our meta-analyses is needed to confirm our results. Statistical significance might be achieved with the inclusion of more studies, and longitudinal studies are needed. Additionally, as research on this topic is new and sparse, large-scale, multicenter clinical prospective and randomized controlled trial (RCT) studies have not yet to be conducted. All the included studies were retrospective and observational, directly resulting in low or very low evidence quality according to the GRADE criteria. Therefore, more cogent results might be obtained by systematic reviews following extensive study in this field and the inclusion of more high-quality studies.

In addition, we believe that unstudied confounding factors, which cannot be fully ruled out, are likely to have interfered with our findings. For example, the administration route, drug type, dosage, frequency, and treatment course of ADT were not identical among the included studies, and different effects of ADT would directly impact the progression of COVID-19. Additionally, the risk of cardiovascular adverse events varies across ADT regimens. For example, Albertsen et al.83 integrated 6 phase-III prospective randomized trials that included 2328 male patients; they found that compared with a GnRH antagonist, GnRH agonists doubled the risk of cardiovascular adverse events in men with a history of cardiovascular disease. More importantly, surgical castration significantly increases the risks of cardiovascular thrombotic events, coronary heart disease, acute myocardial infarction, and ischemic stroke compared to GnRH agonists in PCa patients.84,85 An innovative therapeutic strategy, bipolar androgen therapy (BAT), rapidly fluctuates between supraphysiological and near-castrate serum testosterone levels,86 which could restore androgen receptor sensitivity in castration-resistant PCa patients after progression on enzalutamide/abiraterone.87 Recently, Xiong et al.88 reported that BAT caused low-grade cardiovascular adverse events in castration-resistant PCa patients. Therefore, we attempted to conduct a subgroup analysis to assess the relationship between cardiovascular events and ADT regimens, as well as their potential effects on hospitalization rates for COVID-19. However, the limited number of studies with available information on the ADT type and our selection criteria restricted this subgroup analysis; therefore, large, prospective trials are needed to further confirm its value in the future. Additionally, uses of diverse supportive care interventions, such as antiviral drugs and corticosteroids, merit examination. Fatigue is one of the most common and debilitating PCa-related symptoms as well as acute COVID-19 sequelae. Fatigue and/or cognitive impairment appear in a substantial proportion of COVID-19 patients,89,90 which may lead to poor prognosis. Ferro et al.91 found that the chronic, high-dose use of corticosteroids was significantly associated with a high level of fatigue in PCa patients undergoing systemic treatment. However, complete information on the above variables could not be extracted from eligible studies. These issues will be the focus of our future study.

CONCLUSIONS

As COVID-19 continues to spread worldwide, vast swaths of populations have been infected. Therefore, strong evidence regarding the validity and safety of ADT in COVID-19 is urgently needed. According to our systematic review and updated meta-analyses, we concluded that in PCa patients, ADT use is associated with a higher rate of hospitalization for COVID-19. Our meta-analyses provide valuable contributions to the development of prevention and treatment methods for COVID-19, especially for PCa patients. Future studies with large and standardized samples, as well as those that analyze immune parameters, pathological examinations, and other variables, are needed to confirm our conclusions. Additionally, further exploration of new techniques could provide insights for repurposing and exploiting PCa and COVID-19 drugs.

AUTHOR CONTRIBUTIONS

BF and XCL conceptualized and designed this work and undertook project leadership. BF and XCL funded this work. YBH, WLL, MS, XD, YTW, LXZ, ZHX, and ZFY provided significant inputs on this manuscript. YBH, WLL, MS, and XCL collected and checked the published data; and YBH and BF performed the analysis. YBH, WLL, BF, and XCL wrote, reviewed, edited, and revised the final manuscript. All authors read and approved the final manuscript.

COMPETING INTERESTS

All authors declared no competing interests.

Supplementary Figure 1

The review authors’ judgments based on ROBINS-I checklists about risk of bias for each meta-analyses, including (a) SARS-CoV-2 positive risk, (b) rate of COVID-19 hospitalization, (c) rate of COVID-19 ICU admission, and (d) death risk of COVID-19. The different colors represent various of evaluations for each methodological quality items. SARS-CoV-2: severe acute respiratory syndrome coronavirus 2; COVID-19: coronavirus disease 2019; ICU: intensive care unit.

Supplementary Figure 2

Summary of the connections between diverse SARS-CoV-2–related immune cells in the microenvironment and inflammatory cytokines. Authorized illustrations came from the Smart Medical Art (https://smart.servier.com). AEC: airway epithelial cell; IL: interleukin; TNF-α: tumor necrosis factor-α; TGF-β: transforming growth factor-β; IFN-γ: interferon-γ; SARS-CoV-2: severe acute respiratory syndrome coronavirus 2.