|Home | About | Journals | Submit | Contact Us | Français|
Several studies have reported that men with lower urinary tract symptoms (LUTS) are more likely to experience erectile dysfunction (ED). All but one of these studies was cross-sectional, limiting inferences about whether LUTS precipitate ED.
The association between LUTS and incident ED was examined prospectively in the Health Professionals Follow-up Study. LUTS were assessed biennially by the American Urological Association Symptom Index, which captures symptoms of frequency, urgency and force of urinary stream. Severe LUTS was defined as a symptom score of ≥ 20 points and no LUTS was defined as a score of ≤ 7 points, both among men not treated for LUTS. In 2000, the men were asked to rate their erectile function for several time periods. ED was defined as “poor” or “very poor” function or use of ED medications and no ED was defined as “very good” or “good” function and no use of ED medications. We estimated risk ratios using Poisson regression adjusting for age and other potentially confounding factors.
We observed 3,953 incident ED cases among 17,086 men. Men with severe LUTS in 1994 or earlier had a statistically significant 40% higher risk of ED subsequently than men without LUTS. The risk of ED increased with increasing LUTS severity (p trend < 0.0001). The positive association between LUTS and ED was stronger among younger than older men (p interaction = 0.03).
This study provides evidence that men with LUTS are more likely to develop ED subsequently.
Lower urinary tract symptoms (LUTS), often secondary to benign prostatic hyperplasia (BPH), and erectile dysfunction (ED) are highly prevalent in older men. The prevalence of LUTS and ED in older men may be as high as 31% 2 and 52% 1, respectively. Both ED and LUTS are associated with decreased quality of life. 3,4
Many studies have found a positive association between LUTS and ED 2,5–16; all but one 16 were cross-sectional. Because the cross-sectional design cannot address the temporal relationship between LUTS and ED it remains unclear whether this association is causal. The one prospective study 16 supports a positive association between LUTS and incidence of ED, although the sample size was relatively small (n=1,683).
To further address whether LUTS is prospectively associated with incident ED, we conducted an investigation in the Health Professionals Follow-up Study (HPFS), a large study of US men.
The HPFS is a prospective cohort of 51,529 US male health professionals aged 40–75 years in 1986. Follow-up questionnaires are mailed every 2 years to update exposures and outcomes.
Men were excluded if they died before 2000 (13.7%), did not return the 2000 questionnaire on which ED was assessed (9.6%), or filled out only the short version, which did not ask about ED (9.2%). Men were excluded if they had cancer before 1986 (4.1%), had prostate, bladder, or testicular cancer before 2000 (4.5%), or responded in 2000 but did not provide ED information (4.5%).
LUTS was assessed in 1992, 1994, and 1998. Surgery history, including transurethral resection of the prostate (TURP), was collected on all questionnaires. Use of medications to treat BPH was collected in 1998. Men were asked to indicate frequency (0,10, 25, 50, 75, or almost 100% of the time) of the following: 1) sensation of incomplete bladder emptying (incomplete emptying), 2) having to urinate again after less than 2 hours (frequency), 3) stopping and starting several times during urination (intermittency), 4) difficulty postponing urination (urgency), 5) weak urinary stream (weak stream), and 6) having to push or strain to begin urination (hesitancy). 17 Scores ranging from 0 to 5 were assigned, with a score of 0 corresponding to “0% of the time” and a score of 5 corresponding to “almost 100% of the time”. Men were also asked how many times they typically had to get up at night to urinate (0, 1, 2, 3, 4, 5, or 6 or more times) during the past month (nocturia). A score ranging from 0 to 5 was assigned with “5 times” or “6 or more times” corresponding to a score of 5. These scores were summed to create a score ranging from 0 to 35, which we categorized into 4 groups: no or low (0–7), low-moderate (8–14), high-moderate (15–19), and severe (20–35) symptoms.
In 2000, men were asked, “Please rate your ability (without treatment) to have and maintain an erection good enough for intercourse for the following time periods: Before 1986, 1986–1989, 1990–1994, 1995 or later, in the last 3 months”. Men could rate their ability as “very good”, “good”, “fair”, “poor”, or “very poor”. Two cohorts were constructed: Cohort 1) Men with “very good” or “good” erectile function in every period prior to 1995, and Cohort 2) Men with “very good” or “good” erectile function in every period prior to the past three months (Figure 1). For the first cohort, incident ED was defined as “poor” or “very poor” erectile function in either “1995 or later” or “in the last 3 months” or treatment for ED “in the last 3 months” (irrespective of reported erectile function), and no ED was defined as “very good” or “good” erectile function and no treatment for ED “in the last 3 months”. For the second cohort, incident ED was defined as “poor” or “very poor” erectile function or treatment for ED (irrespective of erectile function) “in the last 3 months” and no ED was defined as “very good” or “good” erectile function and no treatment for ED “in the last 3 months”. To improve the specificity of the outcome definitions, men who reported “fair” function in “1995 or later” and “ in the last 3 months” (Cohort 1) or “in the last 3 months” (Cohort 2) were excluded from the analysis.
We estimated the risk ratio (RR) of ED and 95% confidence intervals (CI) using Poisson regression in SAS v 9.1 (SAS Institute, Cary, NC). Models included indicator variables for LUTS severity groups along with terms for TURP and BPH medications use. For cohort 1, LUTS severity was defined based on the 1992 and 1994 questionnaires and TURP was defined based on surgery in 1994 or earlier (Figure 1). For cohort 2, LUTS was defined based on the 1992, 1994, and 1998 questionnaires, medications to treat BPH was defined based on the 1998 questionnaire, and TURP was defined based on surgery in 1998 and earlier (Figure 1). Multivariable models were adjusted for: age, body mass index (BMI), lifetime cumulative cigarette smoking, physical activity, race, fruit and vegetable intake, alcohol intake, diabetes, hypertension, and cardiovascular disease. Trend tests were modeled using the median score for each symptom category. Sensitivity analyses were conducted restricting to healthy men (i.e., excluded 1,963 who ever had diabetes, cancer, myocardial infarction, stroke, coronary artery bypass, angioplasty, or angina). We conducted analyses stratified by age (<65, ≥ 65 years); statistical interaction was assessed using the likelihood ratio test.
In 1992 (the first time in which we assessed LUTS), men with higher AUA symptom scores were older. After age-standardization, men with higher scores were less physically active, were more likely to smoke, and were more likely to have diabetes, hypertension, and cardiovascular disease (Table 1). Similarly, men with ED that developed in 1995 or later were older, had higher BMI, were less physically active, and ate fewer fruits and vegetables than men without ED. Further, men with ED were more likely to smoke and have diabetes, hypertension, and cardiovascular disease.
In the analysis of incident ED that developed in 1995 or later (Figure 1, cohort 1), we included 3,953 cases among 17,086 men who reported good or very good erectile function before 1995. Men with severe LUTS in 1994 or earlier had a statistically significant 40% higher risk of ED than men without LUTS (Table 2). Risk of ED increased with increasing LUTS severity (p trend < 0.0001). In the analysis of incident ED that developed in the last three months prior to the return of the 2000 questionnaire (Figure 1, cohort 2), we included 953 cases among 14,426 men. Men with severe LUTS in 1998 or earlier had a 59% higher risk of ED than men without LUTS (Table 2). Again, risk of incident ED increased with increasing LUTS severity (p trend = 0.0002). None of the covariates was an important confounder of the LUTS-ED association. When 0 LUTS was the reference instead of 0–7, the dose-response relation was steeper for both ED in 1995 or later (1–7: RR=1.47 95% CI 1.29–1.67, 8–14: RR=1.70 95% CI 1.48–1.96, 15–19: RR=1.95 95% CI 1.65–2.31, 20–35: RR= 2.00 95% CI 1.64–2.45; p trend<0.0001) and in the last three months (data not shown). The association was also increasing across quintiles of the score for ED in 1995 or later (Q2: RR=1.16 (95% CI 1.07–1.26) Q3: RR=1.40 (95% CI 1.28–1.53), Q4: RR=1.40 (95% CI 1.28–1.53), Q5: RR=1.51 (95% CI 1.39–1.64) p trend<0.0001) and in the last three months (data not shown). The association between LUTS and incident ED, both in 1995 or later and in the last three months, was similar in magnitude when the men were categorized by their most recent AUA score prior to ED assessment and when duration of BPH was used (data not shown).
No differences in risk of ED were observed for obstructive (incomplete emptying, hesitancy, weak stream, and intermittency) versus irritative (frequency, urgency, and nocturia) symptoms (data not shown). Men who had undergone TURP or took medications to treat BPH had a higher risk of ED compared with men with no or low LUTS (Table 2). The association between LUTS and ED was stronger in younger than older men (Table 3).
The results were unchanged after restricting to healthy men (data not shown). After further restricting to men without hypertension the results were unchanged. When we restricted the analysis to healthy men who were normal weight (BMI <25), active (≥17.5 MET-hr/wk), ate the recommended 5 servings of fruits and vegetables daily, and never smoked, the results were similar (data not shown).
Results from this prospective study provide support for a positive association between LUTS and incident ED. The increasing risk of ED with increasing AUA symptom score was observed with both standard categories and quintiles of the score. The positive association was stronger when men without any of the six symptoms were the referent. We also found that the association between LUTS and ED may be stronger in younger men.
Because most studies to date have been cross-sectional, we also performed a cross-sectional analysis (versus 0–7, 8–14: RR=1.28, 15–19: RR=1.36, 20–35: RR=1.44) and these findings were comparable to our prospective results. The inferences from our cross-sectional results and those previously published are the same. Cross-sectional studies have reported ORs for the LUTS-ED association ranging from 1.39 to 9.9. 2,5,8,9,12,13,15 Only one study 18 reported no association between LUTS and ED, but it had methodologic limitations.
Our results are similar to those from the one published prospective study (comparing LUTS ≥12 to 0 OR=3.1, 95%CI 1.5–6.4). 16 However, our substantially larger study allowed us to categorize LUTS into finer categories and conduct stratified analyses, as well as multiple sensitivity analyses.
Although our results are consistent with the existing literature, it is difficult to compare the magnitude of our results with those from previous studies. First, LUTS was not consistently categorized among studies. In the other prospective study, men with a score ≥ 12 were compared with men with a score of 0 16; other studies selected different cutpoints. Second, the previous studies reported odds ratios as estimates of the risk ratio, whereas we report risk ratios. Because LUTS and ED are common among older men, for positive associations, the odds ratio will overestimate the risk ratio. Therefore, our directly calculated risk ratios are somewhat smaller than previous estimates for this association.
Several possible noncausal and causal explanations for the apparent association between LUTS and ED exist.
Our results suggest that the association between LUTS and ED is not due to residual confounding by age or solely the effect of treatment: we controlled for age as a continuous variable and created separate categories of exposure for men who had received treatment for LUTS and still observed in untreated men an association between LUTS and ED. We observed an increased risk of ED among men who had undergone surgery or were taking medication for BPH compared with men with no or low LUTS, although their ED risk was lower than for men with severe LUTS. It is not clear whether this finding is due to BPH treatment or reflects that men who received treatment had worse LUTS. Alternatively, these findings may suggest that treatment for severe LUTS moderates ED risk in men with severe LUTS.
Residual confounding by risk factors for LUTS and ED, particularly co-morbidities, is unlikely because the associations were unchanged when restricted analyses were conducted. Nevertheless, it is possible that our results could be partly explained by confounding by unrecognized, shared risk factors for LUTS and ED. Our results cannot distinguish among the other proposed explanations for the association between LUTS and ED.
The association between LUTS and ED was more pronounced among younger than older men, although the absolute increase in risk was comparable (~15% increase comparing severe to no/low symptoms). In younger men, a greater proportion of ED may be caused by the same pathophysiologic mechanisms that cause LUTS in older men, but in older men a greater proportion of ED may be caused by other factors, such as co-morbidities. Alternatively, the difference by age may reflect a difference in the accuracy of reporting ED by age because of differential expectations by age about the quality of erectile function.
The prospective design and large sample size are strengths of our study. We assessed both LUTS and ED by mailed questionnaires, which may have minimized embarrassment and increased the accuracy of report. Further, because LUTS status was collected prior to occurrence and assessment of ED, the extent of any inaccuracy in the report of LUTS is unlikely to be different by ED status, and any such bias would tend to attenuate the association. We asked the men in 2000 when in the past they first experienced ED, and thus, there is the potential for inaccurate recollection of ED. However, the association between LUTS and incident ED was comparable for both remotely recollected ED (back to 1995) and for very recently recollected ED (in the past 3 months), suggesting that inaccurate remote recall is not greatly influencing the observed association.
This study provides evidence that men who have LUTS are more likely to develop ED subsequently. Work is needed to determine whether targeting the pathologies underlying LUTS through dietary and lifestyle modifications or therapies would prevent ED.
The National Institutes of Health provided funding for HPFS general follow-up (grants # CA55075, HL35464). Pfizer, Inc., provided funding for the examination of causes of sexual dysfunction. Ms. Mondul was supported by the National Institutes of Health National Research Service Award T32 CA009314. The sponsor, Pfizer Inc, represented by Dr. Glasser, was involved in the review of the manuscript. The content of this manuscript is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.