Factors associated with erectile dysfunction in men with lower urinary tract symptoms (LUTS): a Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN) study

Abstract

Background

Erectile function (EF) and its associated covariates have not been extensively studied in a large covariate of patients seeking treatment for lower urinary tract symptoms (LUTS). Our objective is to determine the relationship between urinary symptoms and comorbidities with erectile dysfunction (ED) in 447 treatment-seeking men with LUTS.

Methods

Data from the Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN) observational cohort study were analyzed using multivariable logistic regression models to quantify the relationship between LUTS and ED. Models also included anxiety, depression, obesity, cardiovascular disease (CVD), diabetes, and medication use.

Results

Men reporting sexual activity were younger (median age 61 vs. 69 years) and had lower prevalence of diabetes and CVD (13% and 15%, respectively) compared with men reporting no sexual activity (24% for diabetes and 27% for CVD). Among sexually-active participants, higher odds of ED were associated with diabetes [odds ratio (OR) =2.4; 95% confidence interval (CI): 1.1–5.0], age (per 5 years, OR =1.4; 95% CI: 1.2–1.6), urinary incontinence (UI) (OR =1.2; 95% CI: 1.0–1.4), and anxiety (per 10 T-score units, OR =1.4; 95% CI: 1.0–2.1). Low erectile confidence was related to older age in non-sexually-active men (per 5 years, OR =1.5; 95% CI: 1.2–1.8). In sexually-active men, anxiety (per 10 T-score units, OR =1.6; 95% CI: 1.1–2.4), age (OR =1.3; 95% CI: 1.2–1.5), PDE5-inhibitor use (OR =2.1, 95% CI: 1.0–4.3), and diabetes (OR =2.2; 95% CI: 1.1–4.7) were also associated with low erectile confidence.

Conclusions

The association of UI with ED highlights the importance of screening men with UI for ED. Modifiable health variables, such as anxiety and diabetes, were related to ED and erectile confidence in treatment-seeking men with LUTS.

Highlight box.

Key findings

• Urinary incontinence (UI) was associated with moderate to severe erectile dysfunction (ED).

• Anxiety was associated with higher odds of ED and lower erectile confidence in a large cohort of men with urinary symptoms, even after statistically controlling for several covariates.

• Age and diabetes mellitus were associated with ED in a large group of sexually active men seeking care for lower urinary tract symptoms (LUTS).

What is known and what is new?

• Previous studies have demonstrated an age-independent association between LUTS and ED. In a treatment-seeking group of men with LUTS, UI was significantly associated with ED, whereas other LUTS were not found to be similarly associated.

• Anxiety has been previously linked to ED. This study further confirmed that anxiety was significantly associated with higher odds of ED and lower erectile confidence in a cohort of men seeking treatment for LUTS.

What is the implication and what should we change now?

• Men being evaluated for UI should also be screened for ED. Additional studies are needed to clarify the mechanisms underlying the relationship between UI and ED in men.

• Anxiety should be routinely assessed when evaluating patients with ED.

Introduction

Erectile dysfunction (ED) is a distressing condition resulting from pathologies that interfere with achieving or maintaining an erection sufficient for sexual intercourse (1). Estimates of ED prevalence reach 6% in the 40–49-year age range and increase to upwards of 44% in the 70–79-year age range (2). Organic and psychogenic factors have been associated with ED, but they have not been extensively analyzed together in a large, multivariable study in men seeking treatment for lower urinary tract symptoms (LUTS) (3).

LUTS are associated with ED. In a systematic review, De Nunzio et al. found an age-independent association between LUTS and ED, although the causal direction was unclear (4). Associations have also been observed between cardiovascular disease (CVD) and ED, with ED shown to have a strong association with future cardiovascular events and severity of coronary artery disease (5,6). One such study showed that ED has a sensitivity of 34–57% as a risk factor for future myocardial infarctions (7). Additionally, metabolic disorders, such as diabetes and obesity, have been shown to be positively associated with ED (8,9). Psychiatric conditions are also associated with ED, with one study having found a 39% risk of developing ED in men with depression, and another having found that the prevalence of ED in men with anxiety was 20% (10,11).

Certain medications also affect erectile function (EF). Most notably, phosphodiesterase-5 (PDE5) inhibitors improve EF (12). Regarding psychotropics, systematic reviews have showed that serotonin-norepinephrine reuptake inhibitors (SNRIs) were associated with significantly higher odds for ED, whereas there was no significant association between ED and selective serotonin reuptake inhibitors (SSRIs); both drugs were associated, however, with higher odds of ejaculatory dysfunction and decreased libido (13). Studies investigating benzodiazepines and anxiolytics show varying results, with some suggesting that these medications cause ED, whereas others maintain that they increase sexual function through disinhibition (14).

The Symptoms of Lower Urinary Tract Dysfunction Research Network (LURN) Observational Cohort Study presents a unique opportunity to elucidate the patient factors that are associated with ED in a population of men seeking treatment for LUTS. In addition, it provides the opportunity to examine anxiety and depression as they relate to ED. This study aimed to describe the relationships between a comprehensive set of organic and psychogenic factors previously identified in the literature and the presence and severity of ED in a large cohort of patients with LUTS. Additionally, it aims to quantify the correlation between severity of LUTS and ED. We present this article in accordance with the STROBE reporting checklist (available at https://tau.amegroups.com/article/view/10.21037/tau-2025-86/rc).

Methods

Study design and population

Between January 2015 and January 2017, men seeking care for urinary symptoms at six tertiary care clinics were recruited as part of the LURN Observational Cohort Study (15). Patients over age 18 years who reported at least one urinary symptom according to the LUTS Tool were recruited to participate in the study (16). Patients with neurological conditions and a history of prostate or bladder cancer were excluded. Patient-reported outcome data, as well as demographic and clinical data, were collected from each patient at baseline, 3 months, and 12 months. Additional information regarding inclusion criteria has been previously reported (15,17).

All relevant ethical guidelines were followed. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Measures

EF was determined using the six erectile-function items of the International Index of Erectile Function (IIEF-6) (18), which has a high degree of validity as a clinical tool to diagnose ED (19). The American Urological Association Symptom Index (AUA-SI) (20) was used to quantify LUTS. Because urinary incontinence (UI) is not assessed by the AUA-SI, it was assessed using UI-specific questions on the LUTS Tool, which were combined using a Euclidean-distance scoring algorithm used in other LURN research (21). Patient-Reported Outcomes Measurement Information System (PROMIS) questionnaires were used to quantify depression and anxiety (22). Patient self-reported medical histories were used to determine CVD and diabetes status (present or absent). Patient current medication lists collected at enrollment were evaluated to assess use of 5-alpha reductase inhibitor (5ARI), PDE5 inhibitors, and psychotropic medications known to be associated with ED (Table S1).

Statistical analysis

Demographic and clinical characteristics were described using median and interquartile range (IQR) for continuous variables and frequencies and percentages for categorical variables, stratified by self-reported sexual activity based on responses to the IIEF items. The distribution of IIEF scores was explored as a continuous measure, using thresholds of no ED (score ≥26), minimal ED (18≤ score <26), moderate ED (11≤ score <18), and severe ED (score <11). Correlations between the IIEF and urinary symptoms were assessed using Pearson correlations. Responses to the erectile confidence item (a 5-choice rating scale from very low to very high) were explored by self-reported sexual activity. The distribution of IIEF scores was explored to determine the most appropriate model. Due to the skewed nature of responses (even among those reporting sexual activity), logistic regression, using a threshold of IIEF ≤18 representing moderate-to-severe ED, was the selected approach. Multivariable logistic regression models were fitted to assess associations with depression and anxiety, among sexually-active participants, adjusted for medication use at enrollment, LUTS, age, body mass index (BMI), history of CVD, and diabetes. Separate models were fitted for depression and anxiety, given the high correlation between these measures (i.e., the potential for large multicollinearity between anxiety and depression). Additional multivariable logistic models were fitted to assess associations between probability of low erectile confidence and PROMIS depression and anxiety adjusted for the same covariates. Models were fitted separately for those who did and did not report sexual activity at baseline. Analyses were conducted using SAS 9.4 (SAS Institute, Cary, NC, USA).

Results

Sample characteristics

Among 477 male participants, 330 (69%) reported being sexually-active at enrollment. Those reporting sexual activity were younger (median age 61 vs. 69 years) and had lower prevalence of diabetes (13% vs. 24%) and CVD (15% vs. 27%, Table 1). Race and ethnicity were similar between groups, with 4% reporting Asian race, 10% reporting Black race, 4% reporting Hispanic ethnicity, and 81% reporting White race. At enrollment, 15% of the sexually-active group reported PDE5 inhibitor use compared with 12% in the non-sexually-active group. Use of medications to treat depression and anxiety was relatively low at enrollment (<5%) except for SSRIs, reported in 11% of the non-sexually-active group and 7% of the sexually-active group. About 10% reported taking medication for overactive bladder; 16% and 8% reported 5ARI and alpha blocker use, respectively.

Table 1. Male participant characteristics (N=477).

Characteristics	Not sexually active participants (N=147)	Sexually active participants (N=330)	P value
Age (years)	69 [61–75]	61 [51–68]	<0.001
Age category (years)
<45	9 [6]	51 [15]	<0.001
45–54	10 [7]	44 [13]
55–64	31 [21]	111 [34]
≥65	97 [66]	124 [38]
Race†
Asian	4 [3]	12 [4]	0.56
Black	13 [9]	37 [11]
Other¶/multi-racial	10 [7]	14 [4]
White	118 [81]	264 [81]
Ethnicity† (Hispanic/Latino)	5 [3]	17 [5]	0.48
BMI‡ (kg/m2)	28 [25–33]	29 [26–32]	0.40
BMI‡ category (kg/m2)
BMI <25	34 [23]	67 [20]	0.54
25≤ BMI <30	53 [36]	136 [41]
30≤ BMI	60 [41]	127 [38]
Current smoker†	13 [9]	28 [9]	>0.99
Diabetes‡	35 [24]	43 [13]	0.005
History of cardiovascular disease‡	40 [27]	50 [15]	0.002
Medication use^
PDE5-inhibitor	17 [12]	49 [15]	0.39
Anti-anxiety medications	5 [3]	14 [4]	0.80
SSRIs	16 [11]	23 [7]	0.15
SNRIs	5 [3]	5 [2]	0.30
TCAs	2 [1]	6 [2]	>0.99
NDRIs	6 [4]	13 [4]	>0.99
Other anti-depressants	11 [7]	8 [2]	0.02
Anti-psychotics	9 [6]	4 [1]	0.004
Testosterone	2 [1]	8 [2]	0.73
OAB medication	17 [12]	33 [10]	0.63
5-alpha-reductase inhibitor	32 [22]	42 [13]	0.01
Alpha blocker	14 [10]	23 [7]	0.035
Erectile dysfunction (IIEF-6 score)§	2 [1–3]	23 [12–29]	<0.001
Erectile dysfunction category§
No erectile dysfunction	0 [0]	143 [43]
Minimal erectile dysfunction	0 [0]	66 [20]
Moderate erectile dysfunction	0 [0]	48 [15]
Severe erectile dysfunction	133 [100]	73 [22]

Values are presented as median [interquartile range] or count [frequency among non-missing observations]. ^†, missing 1–2%; ^‡, missing <1%, ^§, missing 3%; ^¶, other race includes American Indian, Alaska Native, Native Hawaiian, other Pacific Islander; ^, Medication use at enrollment, groups are not mutually exclusive. BMI, body mass index; IIEF, International Index of Erectile Function; NDRI, norepinephrine and dopamine reuptake inhibitor; OAB, overactive bladder; PDE5, phosphodiesterase-5; SNRI, serotonin and norepinephrine reuptake inhibitor; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant.

Correlations between IIEF, LUTS, and PROMIS anxiety and depression

Correlations between the IIEF-6 and urinary symptoms were small-to-negligible, ranging from −0.15 for the LUTS Tool UI scale to 0.03 for the AUA-SI voiding subscale. Correlations between the IIEF and PROMIS depression and anxiety scores were similarly small (r=−0.11 and −0.04, respectively). PROMIS depression and PROMIS anxiety were highly correlated (r=0.78; Figure 1).

Figure 1.

Correlation between LUTS (AUA total score and sub-domains, LUTS Tool UI scale), PROMIS anxiety and depression and IIEF. AUA, American Urological Association; IIEF, International Index of Erectile Function; LUTS, lower urinary tract symptoms; PROMIS, Patient-Reported Outcomes Measurement Information System; QOL, quality of life; UI, urinary incontinence.

ED and erectile confidence in sexually-active patients

Among those sexually-active (n=330), 63% reported at least minimal ED, defined as a score of IIEF ≤18, with 22% reporting severe ED (IIEF <11). Median IIEF-6 score in this group was 23 (IQR, 12–29, Table 1). Thirty-two percent of sexually-active participants reported low to very low erectile confidence, whereas 49% reported high or very high erectile confidence (Figure 2).

Figure 2.

Erection confidence at baseline by sexual activity.

After adjusting for urologic symptoms, age, BMI, history of CVD and diabetes, and medication use (including PDE5 inhibitors, 5ARIs, and medications to treat anxiety), each 10-unit increase in PROMIS anxiety score (i.e., more severe anxiety) was associated with a 40% increase in the odds of moderate-to-severe ED [odds ratio (OR) =1.4; 95% confidence interval (CI): 1.0–2.1; Figure 3]. After similar adjustment, the effect size was similar for depression, although the CI crossed 1.0; each 10-unit increase in PROMIS depression score was associated with a 30% increase in odds of moderate-to-severe ED (OR =1.3; 95% CI: 0.9–1.9). In both models, more severe UI was also associated with higher odds of moderate-to-severe ED (OR =1.2 per 10-unit increase in LUTS Tool UI score; 95% CI: 1.0–1.4), but no association between AUA-SI score and odds of more severe ED was detected. In this sample, older age and history of diabetes were also associated with higher odds of moderate or severe ED.

Figure 3.

Baseline multivariable erectile dysfunction logistic regression models, among those sexually-active (n=330). *, condition-specific medication use is use of anxiolytics in the anxiety model and SSRI use in the depression model. Logistic models model the probability of moderate or severe erectile dysfunction versus mild to no erectile dysfunction. AUA-SI, American Urological Association Symptom Index; BMI, body mass index; CI, confidence interval; LUTS UI, lower urinary tract symptoms urinary incontinence; OR, odds ratio; PDE5, phosphodiesterase-5; PROMIS, Patient-Reported Outcomes Measurement Information System; SSRI, selective serotonin reuptake inhibitor.

In the erectile confidence model, each 10-unit increase in PROMIS anxiety was associated with 60% higher odds of reporting low or very low erectile confidence (OR =1.6; 95% CI: 1.1–2.4) after adjustment for the same covariates as the ED model. For both the anxiety and depression models, older age, PDE5 inhibitor use, and history of diabetes were associated with higher odds of low or very low erectile confidence (Figure 4A).

Figure 4.

Baseline multivariable erectile confidence logistic regression models among participants sexually-active at baseline (A) and participants who were not sexually-active at baseline (B). *, condition-specific medication use is use of anxiolytics in the anxiety model and SSRI use in the depression model. Logistic models model the probability of having low to very low confidence versus moderate to very high confidence to get and keep an erection. AUA-SI, American Urological Association Symptom Index; BMI, body mass index; CI, confidence interval; LUTS UI, lower urinary tract symptoms urinary incontinence; OR, odds ratio; PDE5, phosphodiesterase-5; PROMIS, Patient-Reported Outcomes Measurement Information System; SSRI, selective serotonin reuptake inhibitor.

Erectile confidence in non-sexually-active patients

Among those not reporting sexual activity, 65% reported low or very low erectile confidence (Figure 2). In this subgroup, older age was associated with odds of low or very low erectile confidence in both models (OR =1.5; 95% CI: 1.2–1.8). In the anxiety model, more severe UI was also associated with higher odds of low or very low erectile confidence (OR =1.3 per 10-unit increase in LUTS Tool UI score; 95% CI: 1.0–1.7; Figure 4B).

Discussion

This study evaluated associations between ED and urologic and psychological health adjusted for known covariates in a sample of men with LUTS. A majority of sexually-active patients (57%) had some degree of ED. The classification of ED in non-sexually-active participants was limited since five of six questions in the IIEF-6 relate to sexual activity, with the sixth question being erectile confidence. UI was associated with the presence of moderate-to-severe ED but not with other LUTS, including symptoms listed on the AUA-SI, which does not query for UI. Although there was a significant association between UI and moderate-to-severe ED, this association was relatively small in magnitude (OR =1.2), indicating that although UI contributes to sexual dysfunction, it may not be the primary or most impactful factor. Previous research suggests that men with overactive bladder experience higher rates of ED (23), but a causal relationship has not yet been established. Tsikis et al. found higher odds of UI in men with ED following robotic prostatectomy, suggesting a potential shared anatomic pathway, although this study excluded patients with a history of prostate cancer (24). Additionally, psychogenic components may contribute to ED, where increased urinary leakage may negatively impact sexual confidence and desire (25). This pattern mirrors findings in the female population, where UI is associated with reduced sexual desire, comfort, and satisfaction (26). Taken together, these results highlight the importance of evaluating men with UI for ED and other sexual dysfunctions. This study adds to this body of literature and underscores the need for further research to better understand the complex relationship between UI and male sexual dysfunction.

The rate of ED was higher in this sample as compared with other studies analyzing prevalence of ED at similar age ranges (2,23,27). This may be explained by two factors: (I) the older age of our sample, which is skewed toward the >55-year age range, given that EF is known to decline with age (3); and (II) all participants were seeking care for LUTS and thus may be a less healthy cohort within the general population. As expected, there was lower erectile confidence in patients who were not sexually-active compared with those who reported sexual activity (Figure 2). Correlations between IIEF-6 scores and urological symptom severity were negligible, in contrast with studies of the general population showing a positive association between LUTS and ED (4,28). These different results may be due to our study being focused on treatment-seeking participants with bothersome LUTS. Additionally, correlations between IIEF-6 and anxiety and depression scores were small in this sample overall, but associations between ED with depression and anxiety were present when focusing only on sexually-active men. This finding highlights the need to assess sexual activity status in men when understanding correlates and comorbidities of sexual function. This finding also adds to previous studies of psychogenic contributions to EF (3,10,11).

In the multivariable logistic regression for ED and low erectile confidence, among sexually-active patients, higher anxiety was associated with a greater probability of moderate-to-severe ED and low erectile confidence in those reporting sexual activity. Additionally, older age and presence of diabetes were associated with increased odds of ED and lower erectile confidence, which is consistent with past studies establishing these as risk factors (8). Interestingly, in multivariable models of sexually-active patients, increased odds of ED and lower erectile confidence were not associated with BMI and CVD, despite strong evidence establishing them as independent risk factors (Figures 3,4A) (5-7,9). No significant associations were detected between condition-specific medications and the odds of ED, consistent with other studies that have reported no link between ED and the use of SSRIs or anxiolytics (13). This supports the idea that SSRI-related sexual dysfunction may be more closely associated with ejaculatory issues and reduced libido rather than ED. However, the CIs for these variables were very wide (Figure 3), limiting the interpretability of these estimates. PDE5 inhibitor use was associated with low erectile confidence in sexually-active patients, which may be due to indication bias, given the observational nature of the study (i.e., patients taking the medication because of low confidence) (12).

In patients who were not sexually-active, condition-specific medication use, PDE5 inhibitor use, and CVD covariates were not included in this model due to low subsample prevalence of these factors. Among the non-sexually-active patients, age had the largest relationship with erectile confidence (OR =1.5 in both the anxiety and depression models; Figure 4B). A surprising pattern of results in Figure 4, although not significant, was that 5ARIs were positively associated low erectile confidence in non-sexually-active men, but inversely associated with low erectile confidence in sexually-active men. Perhaps men who were more sexually confident were more likely to try 5ARIs as part of their treatment, but the medication still negatively affected erections.

Limitations of this study include a majority white sample (81%), with ages skewed toward >55 years, which limits generalizability to other racial groups and younger patients. The sample was treatment-seeking and therefore probably more health-conscious than the general population. Additionally, depression and anxiety are modifiable risk factors with severity that may have varied depending on treatment goals and success, as well as situational factors (e.g., life stressors). There were also too few patients taking other psychogenic medications, which excluded these drugs from analysis. SSRIs were exclusively used in depression analysis but are also part of pharmacotherapy for anxiety. In addition, CVD was defined broadly and likely encapsulates a wide range of severity of disease, which may be why no relationship was detected. Diabetes was not subdivided into Type I and II, which may have differing effects on EF. Furthermore, the IIEF-6 only measures EF and excludes libido and ejaculatory dysfunction, which limits assessment of overall sexual function. There was also limited information on the reasons why PDE5 inhibitors were prescribed (i.e., for LUTS or ED or both). In addition, there was limited dosage information, preventing analysis of dose-specific effects on EF. It is possible that the effect of LUTS on EF was partially confounded by medication status, which is why we adopted a statistically conservative approach by including medication class as covariates.

Future directions include further analysis of erectile and overall sexual function in the setting of a larger, more representative sample of men with LUTS, with improved characterization of CVD, diabetes, and reason for psychogenic medication use.

Conclusions

Among sexually-active men with LUTS, UI measured by the LUTS Tool was associated with the presence of moderate-to-severe ED, whereas LUTS, as measured by the AUA-SI (which does not query UI), was not. This finding highlights the importance of assessing sexual dysfunction in men seeking evaluation for UI. Older age and the presence of diabetes were associated with increased odds of ED; in addition to older age and the presence of diabetes, anxiety was associated with low erectile confidence in sexually-active men with LUTS. Further studies with a larger, more diverse patient population will help characterize sexual function in men with LUTS, in order to refine management strategies for men with LUTS and sexual dysfunction.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.