COVID-19 infection may reduce serum testosterone levels and increase the risk of erectile dysfunction: A two-sample Mendelian randomization study

Abstract

Purpose

Coronavirus disease 2019 (COVID-19) infection may affect serum hormones levels and male sexual function. This study aims to provide evidence for the causal relationship between COVID-19 infection, serum testosterone levels and the risk of erectile dysfunction (ED) using a two-sample Mendelian randomization (MR) approach.

Materials and Methods

Summary-level data for serum testosterone levels (199,569 samples and 12,321,875 single nucleotide polymorphisms [SNPs]) were obtained from Rebecca's study, while data for ED (6,175 cases and 217,630 controls) were sourced from Bovijn's study. Genetic variations linked to COVID-19 were used as instrumental variables (IVs) in meta-analyses of genome-wide association studies (GWASs) involving 6,406 cases and 902,088 controls from the COVID-19 Host Genetics Initiative. The inverse-variance weighted (IVW) method was primarily employed to evaluate the potential associations between COVID-19 infection, serum testosterone levels, and the risk of ED. The weighted mode, weighted-median and simple-median method were employed to evaluate the sensitivity. Heterogeneity and pleiotropic outlier were assessed using Cochran's Q test and MR-Egger regression.

Results

The MR analysis demonstrated that COVID-19 infection was associated with reduced serum testosterone levels (odds ratio [OR] 0.966, 95% confidence interval [CI] 0.938–0.993, p=0.016) and an increased risk of ED (OR 1.205, 95% CI 1.063–1.367, p=0.004) when using IVW methods. Sensitivity analyses utilizing various IV sets and MR approach remained consistent.

Conclusions

COVID-19 infection is associated with a decrease in serum testosterone levels and an increased risk of ED. Male patients recovering from COVID-19 need to pay special attention to their sex hormone levels and sexual health.

Graphical Abstract

INTRODUCTION

The emergence of coronavirus disease 2019 (COVID-19), caused by the novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) virus, has initiated a global health crisis that extends beyond the respiratory distress typically associated with the infection. As the pandemic progresses, an increasing number of studies highlight a range of long-term complications caused by the virus in different organ systems, including the cardiovascular, neurological and male reproductive system [1]. Research indicates that COVID-19 infection may impact serum hormones levels and male sexual function [2]. Numerous studies indicate COVID-19 infection decreases serum testosterone levels in patients [3]. This phenomenon may be attributed to the damage inflicted by COVID-19 infection on testicular tissue, potentially impairing testosterone production [4]. Consistent with these findings, the association between COVID-19 and sexual dysfunction, specifically erectile dysfunction (ED), has been empirically supported by various case reports and observational studies that demonstrate an increased prevalence of ED among individuals who have recovered from the virus [5,6]. Potential mechanisms underlying this trend include inflammatory responses triggered by psychological stress, vascular injury and infection, however, definitive conclusions about causation and the role of genetic factors in mediating this risk have yet to be established. Although COVID-19 has transitioned from a pandemic to an endemic disease, its impact on the long-term health of individuals, especially male reproductive health, remains of significant concern. This study aims to investigate the impact of COVID-19 infection on serum testosterone levels and the risk of ED to address the existing research gap. With the emergence of the concept of “post-viral syndrome”, the results of this study will offer new insights into understanding the long-term effects of COVID-19 on reproductive health, which is of great scientific and practical value.

Mendelian randomization (MR), an instrumental variable (IV) analysis approach, was employed to mitigate confounding and reverse causation biases often encountered in observational studies [7]. To achieve this, we utilized genetic variants associated with COVID-19 infection as IVs, which are considered to be randomly assigned and not subject to potential confounders. By leveraging the associations between genetic variants and COVID-19 infection, we estimated the causal effects of COVID-19 infection on serum testosterone levels and the risk of ED. Although several MR studies have explored the relationship between viral infection and physiological markers, research addressing COVID-19–specific impacts remain limited. Therefore, our study will contribute to the existing literature and provide a significant foundation for subsequent policy development and clinical practice.

In this study, by designing two-sample MR framework, we aim to provide evidence for the causal relationship between COVID-19 infection, serum testosterone levels, and the risk of ED.

MATERIALS AND METHODS

1. Study design and data source

Fig. 1 presents an overview of the flowchart illustrating the current study. Supplementary Table 1 summarizes the details of genome-wide association studies (GWAS) related to serum testosterone levels, ED, and COVID-19. All participants were recruited from Europe. For the genetic instruments of the COVID-19 as exposures, the IVs were obtained from the COVID-19 Host Genetics Initiative which includes 6,406 cases and 902,088 controls. For the outcomes, we obtained the summary-level data for serum testosterone levels (199,569 samples and 12,321,875 single nucleotide polymorphisms [SNPs]) from Rebecca's study and ED (6,175 cases and 217,630 controls) from Bovijn's study.

Fig. 1. An overall design of the present study. MR should satisfy three conditions to ensure the accuracy: (1) the genetic IVs must be closely related to exposures; (2) there should be no confounders that impact both the exposure and outcome; (3) IVs should solely influence the outcome through the exposure alone, rather than through any other pathways. MR, Mendelian randomization; IVs, instrumental variables; SNPs, single nucleotide polymorphisms; COVID-19, coronavirus disease 19; ED, erectile dysfunction.

2. Selection of genetic variants

To identify the potential IVs, we initially selected SNPs that were strongly correlated with COVID-19, meeting genome-wide significance threshold (p<5×10^-8). Subsequently, we pruned these SNPs to remove linkage disequilibrium (LD) using a threshold as r²<0.0001 and a distance=10,000 kb. Only 1 SNP was identified that reached the significance level of p<5×10^-8. Consequently, we expanded the significance threshold to p<5×10^-6 to select eligible IVs. Due to the decrease in significance threshold, we also computed F-values to evaluate whether the selected SNPs were closely related to the exposures. We excluded the SNPs with F-statistics less than 10, and the remaining SNPs were considered to have minimal bias caused by partial sample overlap. After these steps, a total of 18 SNPs were ultimately utilized for MR analysis in this study.

3. Statistical analyses

MR analysis must satisfy three critical conditions to ensure the accuracy: (1) the genetic IVs must be closely related to exposures; (2) there should be no confounders that impact both the exposure and outcome; (3) IVs should influence the outcome solely through the exposure, rather than through any other pathways.

We primarily used the inverse-variance weighted (IVW) approach to assess potential associations between COVID-19, serum testosterone levels, and ED, requiring SNPs to completely conform to the three fundamental principles of MR analysis. The IVW is a commonly used statistical method in MR analysis. This method reduces heterogeneity and enhances the accuracy of the overall estimate by weighting the effect of each genetic variant. This method is particularly useful for combining results from multiple genetic variants, as it allows weights to be assigned based on the precision of each variant estimate. Subsequently, we used the Cochran's Q test to evaluate the heterogeneity of the causal effect among the SNPs. The presence of invalid instruments could introduce bias in the IVW method. To ensure the robustness of our primary findings, we conducted a series of sensitivity analyses. First, the weighted-median, weighted mode and simple-median methods were used to evaluate the potential associations when the IVs contradicted to standard hypothesis. Next, to further assess the potential directional pleiotropy, we applied MR-Egger regression analysis, which can indicate pleiotropy corrected causal effects by examining the slope of the MR-Egger regression. The intercept of the MR-Egger regression model indicates the presence of horizontal pleiotropy or not. Furthermore, we utilized the MR pleiotropy residual sum and outlier (MR-PRESSO) test to identify and adjust for any outlier SNPs. MR-PRESSO is a tool for detecting and dealing with potential pleiotropy and outliers in MR analyses. Finally, we conducted a “leave-one-out” analysis, iteratively excluding one SNP at a time and re-performing the MR analysis on the remaining IVs using the IVW method.

All analyses were conducted using the “MendelianRandomization,” “MRPRESSO,” and “TwoSampleMR” packages in R software version 4.3.2.

4. Ethics approval and consent to participate

Since all of these data were openly accessible, patient consent had been obtained by corresponding studies, and ethical approval was not required for this study.

RESULTS

As shown in Supplementary Table 2, a total of 18 SNPs were identified as IVs for COVID-19, with corresponding F-statistics ranging from 2,278.3 to 12,324.1. All F-statistics exceeded 10, indicating that the IVs were unlikely to be influenced by weak instrumental bias.

1. Association of COVID-19 with serum testosterone levels

The MR analysis results of the correlation between COVID-19, serum testosterone levels, and ED are presented in Tables 1 and 2. Specifically, as shown in Fig. 2A and Table 1, COVID-19 infection was associated with reduced serum testosterone levels using IVW methods (odds ratio [OR] 0.966, 95% confidence interval [CI] 0.938–0.993, p=0.016). This finding indicates that, on average, men who have recovered from COVID-19 may experience a small but statistically significant decrease in serum testosterone levels compared to those who have not been infected. In the sensitivity analysis, the beta coefficients obtained from the simple mode, weighted mode, MR-Egger regression and weighted median approaches were consistent with those from the IVW method, although these methods exhibited lower statistical power and wider CIs (Table 1). Moreover, Cochran Q tests based on both IVW and MR-Egger methods revealed no substantial heterogeneity (p=0.281, p=0.605). No evidence of directional pleiotropy was found (p=0.191), and the MR-PRESSO test did not identify any outlier SNPs (number of outlier SNPs=0) (Table 1). In the leave-one-out sensitivity analysis, the overall association between COVID-19 infection and serum testosterone levels remained stable after sequentially excluding each SNP (Fig. 2D).

Table 1. Estimation of associations between COVID-19 and serum testosterone levels using MR analysis.

Exposure	Outcome	Method	No. of SNPs	Beta	OR (95% CI)	p for associations	p for pleiotropy	p for heterogeneity
COVID-19	Testosterone	IVW	18	-0.035	0.966 (0.938–0.993)	0.016		0.281
COVID-19	Testosterone	MR-Egger	18	-0.004	0.986 (0.943–1.067)	0.911	0.191	0.605
COVID-19	Testosterone	Weighted median	18	-0.018	0.982 (0.954–1.011)	0.218
COVID-19	Testosterone	Simple mode	18	-0.027	0.973 (0.928–1.020)	0.278
COVID-19	Testosterone	Weighted mode	18	-0.020	0.981 (0.943–1.020)	0.339

COVID-19, coronavirus disease 19; MR, Mendelian randomization; IVW, inverse variance weighted; SNPs, single-nucleotide polymorphisms; OR, odds ratio; CI, confidence interval.

Table 2. Estimation of associations between COVID-19 and the risk of ED using MR analysis.

Exposure	Outcome	Method	No. of SNPs	Beta	OR (95% CI)	p for associations	p for pleiotropy	p for heterogeneity
COVID-19	ED	IVW	18	0.187	1.205 (1.063–1.367)	0.004		0.819
COVID-19	ED	MR-Egger	18	0.223	1.249 (0.939–1.662)	0.146	0.787	0.771
COVID-19	ED	Weighted median	18	0.218	1.243 (1.041–1.484)	0.016
COVID-19	ED	Simple mode	18	0.055	1.057 (0.790–1.414)	0.713
COVID-19	ED	Weighted mode	18	0.122	1.130 (0.880–1.450)	0.352

COVID-19, coronavirus disease 19; ED, erectile dysfunction; MR, Mendelian randomization; IVW, inverse variance weighted; SNPs, single-nucleotide polymorphisms; OR, odds ratio; CI, confidence interval.

Fig. 2. Scatter plot (A) and Forest plot (B) represented the SNPs of COVID-19 associated serum testosterone levels using different MR methods. Funnel plot (C) showed no significant heterogeneity among the SNPs of serum testosterone levels. The Forest plot (D) of leave-one-out sensitivity analysis showed the impact of each SNP on the overall causal estimate to serum testosterone levels. SNPs, single nucleotide polymorphisms; COVID-19, coronavirus disease 19; MR, Mendelian randomization; IVW, inverse-variance weighted; SE, standard error.

2. Association of COVID-19 with ED

Additionally, we found that COVID-19 infection was associated with an increased risk of ED using IVW methods (OR 1.205, 95% CI 1.063–1.367, p=0.004) (Fig. 3A, Table 2). This result indicates that men who have recovered from COVID-19 may have a higher risk of developing ED compared to those who have not been infected. In sensitivity analysis, using various MR methods the association between COVID-19 infection and ED had no statistically significant effect in the simple mode, MR-Egger regression and weighted mode method (OR 1.057, 95% CI 0.790–1.414, p=0.713; OR 1.249, 95% CI 0.939–1.662, p=0.146; OR 1.130, 95% CI 0.880–1.450, p=0.352), but showed a modest statistically significance in the weighted median method (OR 1.243, 95% CI 1.041–1.484, p=0.016) (Table 2). There was no significant pleiotropy or heterogeneity in the results (p>0.05). No single SNP affecting the overall effect was identified by using the leave-one-out sensitivity analysis (Fig. 3D).

Fig. 3. Scatter plot (A) and Forest plot (B) represented the SNPs of COVID-19 associated risk of ED using different MR methods. Funnel plot (C) showed no significant heterogeneity among the SNPs of ED. The Forest plot (D) of leave-one-out sensitivity analysis showed the impact of each SNP on the overall causal estimate to ED. SNPs, single nucleotide polymorphisms; COVID-19, coronavirus disease 19; ED, erectile dysfunction; MR, Mendelian randomization; IVW, inverse-variance weighted; SE, standard error.

DISCUSSION

In this study, we systematically assessed the potential causal effects of COVID-19 infection on serum testosterone levels and the risk of ED using a two-sample MR approach. Our findings indicate that COVID-19 infection is associated with a decrease in serum testosterone levels and an increased risk of ED. The results of sensitivity analyses using various IV sets and MR methods remained robust.

Consistent with our findings, Ma et al. [8] found that COVID-19 hospitalization is associated with a reduction in total testosterone levels. The observed decrease in serum testosterone levels among COVID-19 patients may have multifactorial causes. First, the systemic inflammatory response triggered by the virus may lead to hypogonadism, which is characterized by low testosterone levels [9]. Cytokines, such as interleukin-6 (IL-6), which are elevated during severe infections, are known to negatively impact the hypothalamic-pituitary-gonadal axis, thereby reducing reduces testosterone production [10]. Additionally, angiotensin-converting enzyme 2 (ACE2) is a functional receptor for SARS-CoV-2, with high expression levels in target organs such as lungs and testicles [11]. SARS-CoV-2 binds to testicular ACE2, resulting in damage to Sertoli cells and Leydig cells, which in turn decreases testicular testosterone synthesis [12]. Furthermore, studies have shown that serum testosterone concentrations influence the severity of COVID-19. A clinical study involving 116 male patients with COVID-19 demonstrated that low serum testosterone levels were closely associated with disease severity [13]. Salonia et al. [14,15] followed 121 COVID-19 patients for 1 year and found that serum testosterone levels continued to increase over time. However, by the 7th month, more than 50% of the patients had low serum testosterone levels, and by the 12th month, 30% of patients still exhibited low serum testosterone levels [14,15].

It is known that testosterone plays a crucial role in male sexual function and its deficiency can lead to ED [16]. Testosterone or its combination with sildenafil is effective in treating ED in patients with low testosterone levels [17]. Therefore, our study establishes a link between COVID-19 infection and an increased risk of ED, which may be explained in part by reduced testosterone levels.

Our results suggest that COVID-19 infection may increase patients' risk of ED. Data from multiple surveys across different countries indicate that the proportion of ED attributed to COVID-19 infection ranges from 16% to 64% [18,19]. A multicenter study in Germany revealed that during the COVID-19 pandemic, there was an 11.26% increase in new ED diagnoses compared to non-pandemic periods [5]. Although the underlying biological mechanisms of ED resulting from COVID-19 infection have not been fully elucidated, several key viewpoints currently exist. Given that COVID-19 infection leads to widespread endothelial dysfunction, and ED is closely related to proper vascular flow and functional endothelium, this may contribute to new-onset ED [20]. Unal et al. [21] found that COVID-19 infection caused damage to cavernous smooth muscle, leading to ED as detected by cavernous electromyography on patients. Anxiety and depression resulting from COVID-19 infection are also one of the main contributors of ED; therefore, active therapeutic interventions for anxiety and depression in COVID-19–infected patients can effectively alleviate their ED [22]. Furthermore, studies have shown a higher prevalence of ED in COVID-19–infected patients with neurological manifestations such as taste and olfactory disturbances or with comorbidities such as hypertension, diabetes mellitus, chronic liver disease, and ischemic heart disease [23] . COVID-19 infection is known to increase the risk of ED. What is the impact of the COVID-19 vaccine on ED? Studies indicate that the COVID-19 vaccine does not affect male sexual function, including erection and overall sexual satisfaction [24]. On the contrary, vaccination has a positive effect on anxiety and sexual function [25]. Erectile function may deteriorate after COVID-19 infection; however, it tends to improve over time, especially from the first year of recovery onwards [26]. In addition, oral medications such as tadalafil have been shown to significantly improve ED resulting from COVID-19 infection [27].

Our findings suggest that men recovering from COVID-19 should be evaluated not only for immediate sequelae but also for hormonal and sexual health. The impact of long-term complications associated with post-acute sequelae of SARS-CoV-2 infection, or “long COVID”, could be significant if they include hormonal disturbances and sexual dysfunction. Therefore, clinically, a multidisciplinary approach to post COVID-19 care, including endocrinologists and urologists, may be necessary. Aggressive interventions should be implemented for male COVID-19 survivors with decreased testosterone levels or an increased risk of ED. This may include pharmacological interventions (e.g., testosterone replacement therapy), lifestyle modifications (e.g., dietary changes, exercise, and sleep habits), and psychological support [28].

Our study provides a comprehensive and systematic description of the potential causal effects of COVID-19 infection on serum testosterone levels and ED risk using MR methods. Compared to traditional observational studies, this study utilizes a large number of publicly available raw data from GWAS and minimizing the interference of reverse causality and confounding factors, thereby enabling reliable conclusions to be obtained. However, our study has several limitations. Firstly, all participants were of European ancestry, so whether this study can be generalized to other ethnicities is open to discussion. Future studies should explore these relationships in more diverse populations. Secondly, our expansion of the significance threshold to p<5×10^-6 for selecting eligible IVs may lead to false-positive variations and resulting biases. Although we excluded SNPs with F-statistics less than 10 and tried to minimize bias from partial sample overlap, a larger scale of MR analysis with more samples is required.

CONCLUSIONS

In conclusion, this study supports that COVID-19 infection is associated with a decrease in serum testosterone levels and an increased risk of ED. Male patients recovering from COVID-19 should pay more attention to their sex hormone levels and sexual health.

DATA AVAILABILITY STATEMENT

The UK Biobank is an open access resource, available at https://www.ukbiobank.ac.uk/researchers/. Data used in this project can be obtained from the UK Biobank by submitting a data request proposal.

SUPPLEMENTARY MATERIALS

Supplementary materials can be found via https://doi.org/10.4111/icu.20240384.

Supplementary Table 1

Details of the genome-wide association studies and datasets used in this study

Supplementary Table 2

Details of the instrumental variables for COVID-19, variance explained by the selected instruments, and F-statistics for each SNP in this study