The relationship between obstructive sleep apnoea and erectile dysfunction: An underdiagnosed link? A prospective cross‐sectional study

Abstract This cross‐sectional study aimed to investigate the prevalence and clinical characteristics of erectile dysfunction in patients with obstructive sleep apnoea. We enrolled 133 male patients with suspected obstructive sleep apnoea. Ear, nose and throat evaluation, laboratory tests, body mass index, Epworth sleepiness scale, 5‐international index of erectile function, overnight ambulatory polygraphy and drug‐induced sleep endoscopy patterns were assessed. Eighty patients reported obstructive sleep apnoea. 60% (n = 48) reported erectile dysfunction. Statistically significant correlations were found between 5‐International Index of Erectile Function and age, hypertension, diabetes, Epworth sleepiness scale, apnoea‐hypopnea index score, O2 saturation‐nadir, and oxygen desaturation index. Age, diabetes and O2 saturation‐nadir were independent predictors of erectile function. Epworth sleepiness scale, apnoea‐hypopnea index score, O2 saturation‐nadir, oxygen desaturation index and albumin were higher compared to patients without erectile dysfunction. No statistically significant differences were reported for drug‐induced sleep endoscopy patterns and erectile dysfunction. Patients with obstructive sleep apnoea were at significant risk of having erectile dysfunction. Males with obstructive sleep apnoea should be investigated for erectile dysfunction.

in the general population ranges from 9% to 38%, with over 100 million people affected worldwide (Benjafield et al., 2018;Zhang et al., 2019). OSA, which results in snoring, episodes of hypoxia, disturbed sleep and daytime somnolence, has become a serious health problem with negative effects on the quality of life ( _ Irer et al., 2018;Silva et al., 2016;Stepnowsky et al., 2019). Its risk factors include increasing age, obesity, and male gender (Gabbay & Lavie, 2012;Shazia Jehan et al., 2017;Nigro et al., 2018). Moreover, OSA often coexists with other systemic comorbidities, such as diabetes, hypertension, cardiovascular diseases, and sexual dysfunctions (Bonsignore et al., 2019;Pinto et al., 2016;Zheng et al., 2020). Erectile dysfunction, i.e the inability to achieve or maintain a rigid penile erection suitable for sexual intercourse, affects 52% of men between 40 and 70 years and similarly represents an increasing health concern due to its effects on quality of life and its increasing prevalence, which is estimated to exceed 320 million men worldwide by 2025 (Kessler et al., 2019;Yafi et al., 2016). Among the different causes of ED, such as surgery of the lower urinary tract, prostatic surgery, neurological diseases and metabolic/hormonal disorders, several studies and meta-analyses confirmed the increased prevalence of ED in patients with OSA (Chen et al., 2015;Crocetto et al., 2021;Kellesarian et al., 2018;Liu et al., 2015;Manfredi et al., 2021;Romero-Otero et al., 2021;Sperlongano et al., 2014). In addition, the treatment with continuous positive airway pressure (CPAP) can improve erectile function (Li et al., 2010;Pascual et al., 2018). Nevertheless, connections between sleep disorders and sexual problems are largely understudied and OSA syndrome is one of the lesserstudied risk factors for ED (Kalejaiye et al., 2017). Despite the mechanisms that underlie ED remain unclear, sexual problems are common among men with sleep disorders (Hoyos et al., 2015). So far, the exact pathogenesis linking OSA and ED is not well established albeit it has been postulated the role of shared comorbidities such as diabetes, hypertension and metabolic syndrome in inflammation and vascular impairment present in both diseases (Bouloukaki et al., 2014;Hoyos et al., 2015;Kellesarian et al., 2018). Other possible theories focused on reduced nocturnal erections in the rapid eye movement (REM) periods due to sleep fragmentation, sympathetic hyperactivity after each OSA episode and modifications in the hormonal status of OSA patients (Andersen et al., 2010). Several validated questionnaires as the Berlin questionnaire (BQ), STOP-Bang questionnaire and the Epworth Sleepiness Scale (ESS), allow nonspecialist clinicians to identify patients at high risk of having OSA (Amra et al., 2018;Pereira et al., 2013). The gold standard for the diagnosis of OSA is polysomnography (PSG) which requires the patient to sleep overnight in a sleep laboratory under observation.
As result, it is costly, time-consuming, often inaccessible and uncomfortable, requiring, in addition, expert technicians (Greg orio et al., 2011). For these reasons, most sleep study centres use

| Polygraphy and drug-induced sleep endoscopy procedure
All subjects underwent overnight ambulatory RP (Weinmann SOM-NOlab 2, Hamburg, Germany) according to the American Academy of Sleep Medicine Guidelines (Berry et al., 2012

| RESULTS
According to polygraphic parameters, 80 male patients were diagnosed with OSA, met inclusion criteria and were enrolled in the study  Table 1.
We found a statistically significant correlation between IIEF-5 and age, hypertension, diabetes, ESS, AHI, ODI and SaO2-nadir, reporting, in particular, an inverse correlation between OSA parameters and IIEF-5 ( Figure 2

| DISCUSSION
OSA is a frequent medical condition, associated with sleep fragmentation, episodes of hypoxia and daytime somnolence (Kalejaiye et al., 2017;Zheng et al., 2020). Several cross-sectional studies reported a prevalence of ED in patients with OSA ranging from 41% to 80%; interestingly, the OSA treatment with CPAP improves not only the ED but also the possible deficits in reproductive hormones (Hoyos et al., 2015;Zheng et al., 2020). Likewise, in our sample size, we found that 60% of OSA patients were diagnosed with ED with a mean IIEF-5 value of 18.1 ± 5.6 (SD). Among possible causative fac- In our study population, we found normal values of testosterone in the overall cohort and, similarly, both in patients with and without ED. This could be however related to the relatively young age of patients involved in our study. Hypoxia, reduced sleep time and sleep fragmentation can lead to the reduction of testosterone levels (Wittert, 2014). Moreover, as obesity is common among patients with OSA, BMI could be one of the most important determinants of testosterone levels in men with OSA Shamim et al., 2015). This could be secondary to the increased expression of aromatase in adipose tissue and the reduction of sex hormone-binding globulin (Colleluori et al., 2020). The relatively low BMI (27 ± 3.3) in our cohort of patients could probably explain the normal levels of testosterone and the lack of correlation between OSAS and hormonal status reported. It is however to be stated that the threshold of testosterone required to maintain an erection is relatively low and therefore, the impact of BMI and the expression of adipose-tissue aromatase in erectile dysfunction could be significant only in men with severe cases of hypogonadism (Isidori et al., 2014). All enrolled subjects in our study complained of symptoms such as snoring, sleepiness, morning dry mouth and tiredness. The diagnosis of OSA was confirmed by overnight ambulatory RP. Our study population presented the commonest  (8.1 ± 4.6;Johns, 1991). These data stressed the importance to perform at least an overnight RP to define a clear diagnosis of OSA (Laratta et al., 2017).
In our cohort of 80 OSA patients, we found 26.3% of smokers, 41% of hypertensive, and 15% of type 2 diabetic subjects. So far, the relationship between OSA and smoking in the literature is inconclusive (Bielicki et al., 2019). Albeit some studies suggested that the AHI increases with the increase in the smoking rate, this relation was not further confirmed (Hsu et al., 2019). We found 26.3% of smokers in OSA patients, consistent with data reported in the literature which shows that 22% of newly diagnosed OSA patients were current/ former smokers (Shao et al., 2020).
Regarding the relation between systemic hypertension and OSA, approximately half of our OSA patients suffered from systemic hypertension, in line with recent literature (Goldberger et al., 2008). In addition, we found 15% of patients suffering DM2 and overweight. The relationship between OSA, DM2 and obesity is probably multi-directional. In particular, obesity and increased visceral fat are, in turn, perpetuating factors for DM2 in patients with OSA, causing leptin and insulin resistance (Berger & Polotsky, 2018;Jehan et al., 2018). In addition, OSA, in association with DM2, may affect glycemic control increasing the risk of DM2 complications (Khaire et al., 2020).
Among ED patients, we found a higher mean age and a higher prevalence of diabetes, confirming their role as risk factors for ED. The ESS score was higher than in non-ED patients, although still below the cut-off of 10. In addition, we found a significant correlation between ESS and RP parameters (AHI, ODI and SatO 2 -nadir). In our study, we found a slightly higher level of albumin in OSA patients with ED. This data is in contrast with different studies reported in the literature that correlates hypoalbuminemia and ED in patients with chronic hepatitis and cirrhosis (Hunter et al., 2014). However, the causes of ED in patients with liver disease are not clear and the related disruption of vascular, hormonal and neurological integrity may result in an increased and independent sexual dysfunction risk (Kim et al., 2015). In addition, the altered ratio between free testosterone and albumin-bound testosterone could explain ED in patients with hypoalbuminemia (Demir & Barlas, 2021 these data pointed out that the treatment and prevention of diabetes is essential to prevent ED, especially in patients with OSA, where diabetes itself represents an important comorbidity and a predictor of the ineffectiveness of phosphodiesterase-5 inhibitors (PDE5-I) therapy (Garrido-Abad et al., 2021). In addition, the improvement of oxygenation with the treatment of OSA using CPAP or following oral appliances and/or surgery could, as well, improve ED.
We are conscious of several limitations which affect our study.
First, the study design is prone to potential sampling bias; second, the RP and DISE evaluations, although made in the same clinic, were performed by different operators, without assessing the concordance among evaluations; third, the IIEF-5 and the ESS assessment remain two subjective questionnaires that could have a wide intervariability; fourth, our sample size, although consistent with other studies reported in the literature, is still limited to evaluate the influence of DISE patterns in ED. Due to those limitations, the results obtained have to be, therefore, cautiously evaluated.

| CONCLUSIONS
This study confirmed that OSA is a risk factor for ED. In particular, our data demonstrated that OSA parameters correlate with ED and age, while SatO2-nadir and diabetes are independent predictors of ED. Our research pointed out that men presenting to the Ear, nose and throat (ENT) clinic with OSA are at significant risk of having ED. Therefore, males with OSA should be investigated for ED. ACKNOWLEDGMENT Open Access Funding provided by Universita degli Studi di Napoli

AUTHOR CONTRIBUTIONS
Federico II within the CRUI-CARE Agreement.

CONFLICT OF INTEREST
The authors declare no conflict of interest.