Abstract
Background:
Erectile dysfunction (ED) increases with age. Remarkably, the relationship between age and the risk of ED has only been described in crude categories, such as risk for men aged 50–59 yr, without taking comorbidities into account.
Objective:
To understand how the risk of patient-reported ED varies according to age and comorbidity status.
Design, setting, and participants:
This cross-sectional study included a cohort of 17 250 patients with prostate cancer who completed the International Index of Erectile Function erectile function domain (IIEF-EF) questionnaire before any prostate treatment.
Outcome measurements and statistical analysis:
We created a logistic regression model to predict the probability of ED using age and comorbidities such as cardiovascular disease, diabetes, and hypertension as predictors. We used age as a nonlinear term to allow a curvilinear relationship between age and ED.
Results and limitations:
The prevalence of patient-reported ED among men without any comorbidities increased from 10% to 79% from the age of 40 and 80 yr. The risk of ED increased sharply with comorbidity: the probability of ED for 50- and 75-yr-old individuals was 20% and 68% for healthy men, but 41% and 85% for those with hypertension, obesity, and diabetes. Men with several comorbidities have the same risk of ED as that of healthy men 15–25 yr older. Limitations include a healthier-than-average patient group and lack of information about some comorbidities and the severity of comorbidities.
Conclusions:
Our results allow us to better understand how the risk of ED changes with age and comorbidities. Further research should evaluate the impact of other risk factors not considered in the present study and should take risk factor severity into account.
Patient summary:
Our study shows how the probability of erectile dysfunction (ED) changes with increasing age, analyzed alone and when taking into account the presence of other risk factors for this condition (eg, diabetes, high blood pressure, and cardiovascular disease). Our results help in better understanding the probability of ED for men with and without comorbidities.
Keywords: Erectile dysfunction, Population estimates, Age, Comorbidity, Patient-reported outcomes
1. Introduction
Erectile dysfunction (ED) is defined as the persistent inability to obtain or maintain an erection that is sufficient to allow satisfactory sexual intercourse [1]. ED has important implications for the quality of life and psychosocial health of men and their partners [2,3].
Age is a well-known risk factor for ED, and several studies have reported on the prevalence of ED with advancing age [2,4–7]. Remarkably, all of these studies reported on ED prevalence using crude categories, such as 10-yr age intervals. This is problematic since there may be important variations within these groups. For instance, 50-yr-old men may not have the same ED risk as 59-yr-old men, but are treated the same in a paper reporting ED among men aged 50–59 yr. Moreover, all the studies used mixed populations of participants both with and without other risk factors for ED, such as obesity, hypertension, and diabetes [8]. Hence, the risks given reflect the respective combined distribution of comorbidities in the cohorts studied. For example, when the authors of the Cologne study [4] report an ED prevalence of 34% for men aged 60–69 yr, what they mean is that in any group of 100 men in their 60s with the same distribution of comorbidities as 60-yr-old men in the Cologne study, 34 will have ED. Similarly, the Massachusetts Male Aging Study [2] reported at least minor ED in 50% of men aged 50–54 yr, but again this is true only for groups of men with a similar level of comorbidity as in that study. In this context, our aim was to understand how the risk of patient-reported ED varies according to age, alone and together with the presence of other risk factors for this condition.
At Memorial Sloan Kettering Cancer Center (MSKCC, New York, NY, USA), sexual status surveys are used as part of routine care in the evaluation of patients with prostate cancer. This provides us with a large database of patients that can be used to describe the prevalence of ED by age with far more granularity on age and comorbidity status than prior studies, permitting us to take into account the effect of comorbidities separately. Because ED is a marker of endothelial dysfunction [9], our results provide insight into how this varies by age and comorbidities.
2. Patients and methods
2.1. Study population
In this cross-sectional study, after obtaining institutional review board approval (reference 17–629), we used a prospective database of 19 601 patients with a diagnosis of prostate cancer referred to MSKCC between 2007 and 2021 who completed ED questionnaires at their first visit. We excluded patients who had undergone previous prostate cancer therapy or surgery for ED. This resulted in a final population of 17 250 men.
2.2. Erectile function assessment
Erectile function was evaluated in our population using the patient-reported erectile function domain score of the International Index of Erectile Function (IIEF), which comprises questions 1–5 and 15 of the IIEF questionnaire (IIEF-EF) and is routinely used as a standalone instrument for this purpose [10]. We used the MSKCC modification of the IIEF-EF, which incorporates a wider range of sexual behaviors compared to the original IIEF-EF [10]. We defined patients affected by ED as those with an IIEF-EF score ≤24 [11].
2.3. Comorbidity assessment
Age, diabetes, dyslipidemia, hypertension, obesity, cardiovascular disease (CVD), lower urinary tract symptoms/benign prostatic hyperplasia (LUTS/BPH), and depression/anxiety were considered as ED risk factors for the present study. These are some of the ED risk factors reported by the American Urological Association [12] and European Association of Urology guidelines [8], and associations between these risk factors and ED have been extensively studied [13–18]. Metabolic syndrome [19], sleep apnea [20], and smoking [21] were not included because of inconsistent clinical recording of these risk factors. All risk factors were obtained from the medical history as documented by the physician who evaluated each patient. Dyslipidemia was defined as elevated total cholesterol, low-density lipoprotein cholesterol, or triglyceride levels, or low levels of high-density lipoprotein cholesterol. We defined patients as having CVD if they had a history of coronary artery disease, other heart diseases (eg, rhythm problems, cardiomyopathy, and heart valve diseases), or vascular disease (eg, peripheral vascular disease, stroke, transient ischemic attack, and carotid stenosis), but excluded hypertension and dyslipidemia, as these were considered separately. All risk factors except for age were considered as dichotomous variables (eg, diabetes: yes vs no).
2.4. Statistical models
To estimate the prevalence of ED due to the selected risk factors in our population, we first evaluated whether or not the association between age and ED risk was linear. We used restricted cubic splines for age with four knots placed at the 5th, 35th, 65th, and 95th percentiles, corresponding to 49.6, 60.2, 66.7, and 76.8 yr [22]. There was strong evidence of a nonlinear relationship between age and the logarithmic odds of ED (p < 0.001), and therefore the nonlinear terms were retained. To assess further how we should build our model, we evaluated wherever there was interaction between age and each of the other risk factors. We saw no evidence of interaction, with p ≥ 0.5 for all the risk factors except for diabetes (p = 0.2) and LUTS/BPH (p = 0.09). Interaction terms were therefore not considered. The predictors included in the final model were age as a nonlinear term and diabetes, dyslipidemia, hypertension, obesity, CVD, LUTS/BPH, and depression/anxiety as binary terms (yes vs no).
To investigate how ED varies according to age and the presence of other risk factors, we converted the coefficient for each risk factor from the multivariable logistic regression model to an integer score; Supplementary Tables 1 and 2 provide further details [23]. Then we plotted ED prevalence against age and the aggregate risk score obtained by summing the risk factor scores for each combination of possible comorbidities using two different methods. First, we used a line plot representing ED risk by age for men with different aggregate scores (0 = absence of risk factors; 10 = patients with, eg, LUTS/BPH and CVD; 20 = patients with, eg, LUTS/BPH, CVD, hypertension, and obesity; 30 = patients with, eg, depression/anxiety, diabetes, and CVD; and 45 = patients with all six risk factors). Second, we used a heatmap to represent ED prevalence by age and risk factors in a more granular way. As some urologists use an IIEF-EF score <26 to define ED [24], we repeated all of the analyses using this cutoff. Finally, we plotted the prevalence of moderate ED (IIEF-EF score <17) and severe ED (IIEF-EF score <11) using the same approach. All tests were two-sided with a significance level of 0.05. Statistical analyses were performed using R v4.0.2 statistical software (R Foundation for Statical Computing, Vienna, Austria).
3. Results
The cohort characteristics are summarized in Table 1. Approximately half of the study population was affected by ED (51%). All comorbidity risk factors were more frequent in the group of patients with ED, and approximately two-thirds of the men with ED had at least two concomitant risk factors.
Table 1 –
Population characteristics of 17 250 patients stratified by ED (IIEF-EF score ≤24)
| Characteristic | Non-ED group (n = 8416) | ED group (n = 8834) |
|---|---|---|
| Median age, yr (interquartile range) | 61 (55–66) | 66 (61–71) |
| Diabetes, n (%) | 604 (7.2) | 1445 (16) |
| Obesity, n (%) | 2002 (24) | 2323 (26) |
| Dyslipidemia, n (%) | 3591 (43) | 4505 (51) |
| Hypertension, n (%) | 3278 (39) | 4802 (54) |
| Cardiovascular disease, n (%) | 1216 (15) | 2167 (25) |
| Depression/anxiety, n (%) | 893 (11) | 1203 (14) |
| Lower urinary tract symptoms/benign prostatic hyperplasia | 2218 (26) | 3182 (36) |
| Number of risk factors per patient, n (%) | ||
| 0 factors | 1824 (22) | 1048 (12) |
| 1 factor | 2493 (30) | 1946 (22) |
| 2 factors | 2014 (24) | 2227 (25) |
| 3 factors | 1269 (15) | 1878 (21) |
| >3 factors | 816 (9.7) | 1735 (20) |
| Median IIEF-EF score (interquartile range) | 29 (28–30) | 12 (4–20) |
ED = erectile dysfunction; IIEF-EF = International Index of Erectile Function erectile function domain.
The multivariable logistic regression results are presented in Table 2. The intercept for our model was zero. All of the risk factors except for dyslipidemia were significantly associated with ED. Diabetes emerged as the strongest risk factor for ED (odds radio 2.01, 95% confidence interval 1.81–2.24). To ease interpretation, we converted the coefficients from the multivariable model to integer scores using the standard approach of dividing by the smallest coefficient [23]. We then multiplied by a factor of three and rounded the results to integer numbers (Supplementary Table 1).
Table 2 –
Multivariable logistic regression analysis for erectile dysfunction (International Index of Erectile Function erectile function score ≤24) a
| Characteristic | Odds ratio (95% CI) | p value |
|---|---|---|
| Age (per 10-yr increment) | 2.15 (1.84–2.53) | <0.001 |
| Diabetes | 2.01 (1.81–2.24) | <0.001 |
| Obesity | 1.14 (1.05–1.23) | <0.001 |
| Dyslipidemia | 0.97 (0.91–1.04) | 0.4 |
| Hypertension | 1.32 (1.24–1.42) | <0.001 |
| Cardiovascular disease | 1.22 (1.12–1.33) | <0.001 |
| Depression/anxiety | 1.48 (1.34–1.63) | <0.001 |
| Lower urinary tract symptoms/benign prostatic hyperplasia | 1.25 (1.16–1.34) | <0.001 |
CI = confidence interval.
Age was included as a nonlinear term (splits not reported). All risk factors except for age were included as dichotomous variables.
The prevalence of ED by age and aggregate risk factor scores is shown in Figures 1 and 2 for specific combinations of age and comorbidities. The data can also be serve as a visualization of endothelial function in aging males.
Fig. 1 –
Prevalence of erectile dysfunction (ED; International Index of Erectile Function erectile function domain score ≤24) by age for patients with different aggregate risk scores (0, 10, 20, 30, and 45). To obtain the aggregate score for a patient with specific risk factors, add the risk factor scores in the table.
CVD = cardiovascular disease; LUTS/BPH = lower urinary tract symptoms/benign prostatic hyperplasia.
Fig. 2 –
Prevalence of erectile dysfunction (ED; International Index of Erectile Function erectile function domain score ≤24) by age and aggregate risk factor score. Black curves denote ED risk (from 10% to 90%). To obtain the aggregate score for a patient with specific risk factors, add the risk factor scores in the table.
CVD = cardiovascular disease; LUTS/BPH = lower urinary tract symptoms/benign prostatic hyperplasia.
Aggregate risk scores calculated by summing the risk factor scores were used to plot the prevalence of ED for specific combinations of age and various comorbidities. ED prevalence ranged between 10% and 78% for men without comorbidities as age increased from 40 to 80 yr (Fig. 1). At a specific patient age, ED risk greatly increases with the presence of additional risk factors. However, a given relative increase in the odds of ED has differential effects on the absolute risk, depending on the baseline risk. Hence, the effect of comorbidities on the absolute risk of ED was higher for young patients than for older patients. For instance, for a 40-yr-old man the risk of ED increases from 10% to 44% (an increase of 34%) if all the risk factors are present, whereas for an 80-yr-old patient this increase is approximately 18% (from 78% to 96%). The prevalence of ED for 50-yr-old and 75 yr-old men was 20% and 68%, respectively, for healthy men, but 41% and 85%, respectively, for men with concomitant hypertension, obesity, and diabetes. Finally, the data show that healthy older men have the same risk of ED as younger patients who have several comorbidities. For instance, a 50-yr-old man with diabetes, obesity, and hypertension has the same risk of ED as a 65-yr-old man without comorbidities.
Supplementary Table 2 presents the formulae for calculating ED risk for a man with a particular age and risk factor combination. Supplementary Table 3 shows how to calculate risks for different combinations of age and comorbidities. Supplementary Tables 4 and 5 and Supplementary Figures 1 and 2 show results when using an IIEF-EF score <26 as the alternative definition of ED. The results for moderate and severe ED are given in Supplementary Figures 3 and 4, respectively. It is evident that the shapes of the curves are very similar, albeit with lower prevalence.
4. Discussion
We used a large database to estimate the prevalence of patient-reported ED for specific ages and combinations of comorbid conditions. A particular advantage of our approach is that we provide estimates for men at specific ages rather than for men in 5–10-yr age intervals. This contrasts with previous studies on age and ED [2,4–7]. By categorizing age in 5–10-yr ranges, authors overestimate ED risk for younger men and underestimate it for older men in each group. For example, the average risk of ED among healthy men aged 50–59 yr in our study was ~28%, but this varied over an approximately 1.75-fold range, from 20% risk at age 50 yr to 35% risk at age 59 yr.
A second advantage of our approach is that we provide estimates for specific combinations of comorbidities rather than for men with the average comorbidity. Prior studies that explored the relationship between age and ED [2,4,5,7] reported ED risk for mixed cohorts of patients both with and without comorbidities (eg, hypertension, cardiovascular disease, diabetes). Their results therefore correspond to the average ED prevalence in the general population and not to the ED risk due to age alone. Comparison of ED prevalence between studies is difficult owing to differences in study design, ED definitions, and the methods used to diagnose ED. However, the Massachusetts Male Aging Study [2] provides an opportunity to compare the methodological approaches. The authors of the Massachusetts study reported ED prevalence of 70% for 70-yr-old men, whereas we found that risk varies from 55% for healthy men to >80% for men with numerous comorbidities.
Our more granular approach allows a better understanding of how ED, and hence endothelial dysfunction in general, changes in the aging male. Previous analyses hypothesized that ED is a harbinger of CVD, suggesting that ED pathology may be the first manifestation of endothelial dysfunction [9,13,25,26].
We were able to find only one study that reported the prevalence of ED among men without concomitant diseases. Nicolosi et al [6] interviewed 2412 men across four countries (Brazil, Italy, Japan, and Malaysia) and reported results comparable to ours (eg, ED prevalence of 7.8% for healthy men aged 40–44 yr). However, their study had two main limitations. They reported probabilities for age groups comprising 5-yr increments. Furthermore, in contrast to our use of the validated IIEF-EF instrument, the authors assessed erectile function using the response to a single question (“How would you describe yourself?”) with four possible answers (“always, usually, sometimes or never able to get and keep an erection good enough for sexual intercourse”), which is not validated.
Numerous other studies have shown an association between ED and the comorbidities considered here [13–18]. Nevertheless, our study is the first to estimate the prevalence of ED in the presence of specific comorbidities. However, we evaluated only some of the risk factors associated with ED, and therefore further studies should assess the degree to which other risk factors such as cigarette smoking [21], metabolic syndrome [19], and obstructive sleep apnea [20] are independently associated with ED.
The population in our study presents a potential limitation. We evaluated patients with prostate cancer before they received definitive therapy. Cancer screening, as reported in a recent study [27], is associated with a lower hazard of all-cause mortality after accounting for other risk factors. The most convincing explanation for this effect is that nonadherence to protocol screening could be a marker for a general behavioral profile involving an unhealthy lifestyle and nonadherence to medical tests and treatments, and consequently a higher risk of mortality; moreover, patients who have short life expectancy because of other diseases should not be screened nor present for treatment of prostate cancer [28]. We previously showed that patients who undergo radical prostatectomy have life expectancy equivalent to that for men from the general population who are 3 yr younger [29]. Correcting for this effect would shift the curves in Figure 1 to the left: for instance, a 60-yr-old man with obesity and hypertension (aggregate score 10) would have ED risk of 50% rather than the 44% calculated with the original model. Supplementary Table 6 reports the risk of ED by age according to this correction. Nevertheless, further studies should confirm our findings in a population of patients without prostate cancer.
Another potential limitation of this study is that we included comorbidities in our model as binary variables (eg, hypertension yes vs no) obtained from the patient’s medical history and not defined using predefined cutoffs. Moreover, patients with uncontrolled comorbidities were not distinguished from those with comorbidities well controlled with specific therapies. Finally, we did not consider the risk factor duration. However, all of these pathologies have different severity levels. It is likely, for example, that the risk of ED is very different between men with severe and mild cases of hypertension, ands between men with controlled and uncontrolled hypertension and patients affected by hypertension for a different length of time. Similarly, dyslipidemia, which was not associated with ED in our analyses despite being a well-known risk factor [2,19], would probably emerge as a predictor if lipid status was used as a continuous variable instead of a categorical classification. There is room for future studies to further refine our approach by considering the severity of comorbidities and their treatments.
5. Conclusions
Our results allow a better understanding of how the risk of ED, and endothelial function in general, changes with age and comorbidities. Further research should evaluate the impact of other risk factors not considered in the present study and take risk factor severity into account.
Supplementary Material
Acknowledgments:
We thank Dr. David Kent for providing comments on an early version of this paper.
Funding/Support and role of the sponsor: This work was supported in part by the National Institutes of Health/National Cancer Institute via a Cancer Center Support Grant to Memorial Sloan Kettering Cancer Center (P30 CA008748), a SPORE grant in Prostate Cancer to H. Scher (P50-CA92629), and the Sidney Kimmel Center for Prostate and Urologic Cancers. The sponsors played no direct role in the study.
Financial disclosures: Francesco Pellegrino certifies that all conflicts of interest, including specific financial interests and relationships and affiliations relevant to the subject matter or materials discussed in the manuscript (eg, employment/affiliation, grants or funding, consultancies, honoraria, stock ownership or options, expert testimony, royalties, or patents filed, received, or pending), are the following: Andrew J. Vickers is named on a patent for a statistical method for detecting prostate cancer that has been commercialized by OPKO Health as the 4Kscore, and receives royalties from sales of the test and has stock options in OPKO Health.
References
- [1].NIH Consensus Conference. Impotence: NIH Consensus Development Panel on Impotence. JAMA 1993;270:83–90. [PubMed] [Google Scholar]
- [2].Feldman HA, Goldstein I, Hatzichristou DG, Krane RJ, McKinlay JB. Impotence and its medical and psychosocial correlates: results of the Massachusetts Male Aging Study. J Urol 1994;151:54–61. [DOI] [PubMed] [Google Scholar]
- [3].Fisher WA, Eardley I, McCabe M, Sand M. Erectile dysfunction (ED) is a shared sexual concern of couples I: couple conceptions of ED. J Sex Med 2009;6:2746–60. [DOI] [PubMed] [Google Scholar]
- [4].Braun M, Wassmer G, Klotz T, Reifenrath B, Mathers M, Engelmann U. Epidemiology of erectile dysfunction: results of the ‘Cologne Male Survey.’ Int J Impot Res 2000 12: 6 2001;12:305–11. [DOI] [PubMed] [Google Scholar]
- [5].Mulhall JP, Luo X, Zou KH, Stecher V, Galaznik A. Relationship between age and erectile dysfunction diagnosis or treatment using real-world observational data in the USA. Int J Clin Pract 2016;70:1012–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Nicolosi A, Glasser DB, Moreira ED, Villa M. Prevalence of erectile dysfunction and associated factors among men without concomitant diseases: a population study. Int J Impot Res 2003 15:4 2003;15:253–7. [DOI] [PubMed] [Google Scholar]
- [7].Corona G, Lee DM, Forti G, et al. Age-related changes in general and sexual health in middle-aged and older men: results from the European Male Ageing Study (EMAS). J Sex Med 2010;7:1362–80. [DOI] [PubMed] [Google Scholar]
- [8].Salonia A, Bettocchi C, Boeri L, et al. European Association of Urology guidelines on sexual and reproductive health—2021 update: male sexual dysfunction. Eur Urol 2021;80:333–57. [DOI] [PubMed] [Google Scholar]
- [9].Thompson IM, Tangen CM, Goodman PJ, Probstfield JL, Moinpour CM, Coltman CA. Erectile dysfunction and subsequent cardiovascular disease. JAMA 2005;294:2996–3002. [DOI] [PubMed] [Google Scholar]
- [10].Vickers AJ, Tin AL, Singh K, Dunn RL, Mulhall J. Updating the International Index of Erectile Function: evaluation of a large clinical data set. J Sex Med 2020;17:126–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Terrier JE, Mulhall JP, Nelson CJ. Exploring the optimal erectile function domain score cutoff that defines sexual satisfaction after radical prostatectomy. J Sex Med 2017;14:804–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Burnett AL, Nehra A, Breau RH, et al. Erectile dysfunction: AUA guideline. J Urol 2018;200:633–41. [DOI] [PubMed] [Google Scholar]
- [13].Gandaglia G, Briganti A, Jackson G, et al. A systematic review of the association between erectile dysfunction and cardiovascular disease. Eur Urol 2014;65:968–78. [DOI] [PubMed] [Google Scholar]
- [14].Binmoammar TA, Hassounah S, Alsaad S, Rawaf S, Majeed A. The impact of poor glycaemic control on the prevalence of erectile dysfunction in men with type 2 diabetes mellitus: a systematic review. JRSM Open 2016;7:2054270415622602. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Pizzol D, Smith L, Fontana L, et al. Associations between body mass index, waist circumference and erectile dysfunction: a systematic review and meta-analysis. Rev Endocr Metab Disord 2020;21:657–66. [DOI] [PubMed] [Google Scholar]
- [16].Liu Q, Zhang Y, Wang J, et al. Erectile dysfunction and depression: a systematic review and meta-analysis. J Sex Med 2018;15:1073–82. [DOI] [PubMed] [Google Scholar]
- [17].Seftel AD, de La Rosette J, Birt J, Porter V, Zarotsky V, Viktrup L. Coexisting lower urinary tract symptoms and erectile dysfunction: a systematic review of epidemiological data. Int J Clin Pract 2013;67:32–45. [DOI] [PubMed] [Google Scholar]
- [18].Li JZ, Maguire TA, Zou KH, Lee LJ, Donde SS, Taylor DG. Prevalence, comorbidities, and risk factors of erectile dysfunction: results from a prospective real-world study in the United Kingdom. Int J Clin Pract 2022;2022:5229702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Besiroglu H, Otunctemur A, Ozbek E. The relationship between metabolic syndrome, its components, and erectile dysfunction: a systematic review and a meta-analysis of observational studies. J Sex Med 2015;12:1309–18. [DOI] [PubMed] [Google Scholar]
- [20].Kellesarian SV, Malignaggi VR, Feng C, Javed F. Association between obstructive sleep apnea and erectile dysfunction: a systematic review and meta-analysis. Int J Impot Res 2018 30:3 2018;30:129–40. [DOI] [PubMed] [Google Scholar]
- [21].Cao S, Gan Y, Dong X, Liu J, Lu Z. Association of quantity and duration of smoking with erectile dysfunction: a dose–response meta-analysis. J Sex Med 2014;11:2376–84. [DOI] [PubMed] [Google Scholar]
- [22].Nieboer D, Vergouwe Y, Roobol MJ, et al. Nonlinear modeling was applied thoughtfully for risk prediction: the Prostate Biopsy Collaborative Group. J Clin Epidemiol 2015;68:426–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Steyerberg EW. Clinical prediction models: a practical approach to development, validation, and updating. ed. 1. New York, NY: Springer; 2009. [Google Scholar]
- [24].Cappelleri JC, Rosen RC, Smith MD, Mishra A, Osterloh IH. Diagnostic evaluation of the erectile function domain of the International Index of Erectile Function. Urology 1999;54:346–51. [DOI] [PubMed] [Google Scholar]
- [25].Uddin SMI, Mirbolouk M, Dardari Z, et al. Erectile dysfunction as an independent predictor of future cardiovascular events: the Multi-Ethnic Study of Atherosclerosis (MESA). Circulation 2018;138:540. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Ibrahim A, Ali M, Kiernan TJ, Stack AG. Erectile dysfunction and ischaemic heart disease. Eur Cardiol 2018;13:98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Schoenborn NL, Sheehan OC, Roth DL, et al. Association between receipt of cancer screening and all-cause mortality in older adults. JAMA Netw Open 2021;4:e2112062. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Pierre-Victor D, Pinsky PF. Association of nonadherence to cancer screening examinations with mortality from unrelated causes: a secondary analysis of the PLCO cancer screening trial. JAMA Intern Med 2019;179:196–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Kent M, Penson DF, Albertsen PC, et al. Successful external validation of a model to predict other cause mortality in localized prostate cancer. BMC Med 2016;14:25. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.