Education degree predicts cardiovascular outcomes in men suffering from erectile dysfunction

The level of education has been recognized as a cardiovascular risk factor; nevertheless, it is often neglected in cardiovascular risk prediction.

forthcoming major adverse cardiovascular events. Therefore, education level should be considered as a costless but valuable information in the assessment of cardiovascular risk in patients with erectile dysfunction.

K E Y W O R D S
cardiovascular risk, erectile dysfunction, level of education, metabolic syndrome INTRODUCTION Cardiovascular (CV) diseases (CVD) represent the highest cause of death in the world. Since 2000, ischemic heart disease has been responsible for up to 16% of the world's total deaths, becoming the biggest killer (WHO 2020 https://www.who.int/news-room/factsheets/detail/the-top-10-causes-of-death). Hence, it is pivotal to individuate precocious CV predictors for correcting the modifiable ones by ameliorating lifestyle or by using specific medications, thus preventing fatal and non-fatal CV events.
It is well known that erectile dysfunction (ED) is a marker of arterial damage and an early and independent predictor for major adverse CV events (MACE). 1 Therefore, a medical examination for sexual dysfunction represents an unmissable chance for defining a patient's CV risk status. ED shares several risk factors with CV diseases, 2 most of which are included in common risk algorithms for CV risk stratification (e.g., Framingham risk score, SCORE). Despite their practicality and simplicity, these engines are not accurate for all populations as they could be deemed reliable only if applied in contexts comparable to those in which they were developed. 3 Hence, in men suffering from ED, known to be at higher CV risk, it becomes fundamental focusing attention also on unconventional risk factors not included in current risk equations. In the last years, our research group focused the attention on this topic: several hormones (low testosterone and prolactin), intrapsychic (depression) and couple conditions (partner's decreased libido and cheating on the partner), as well as parenthood, 4 emerged as markers of an increased CV risk. Among the risk factors tested in our population of ED men, a low level of education emerged to play a relevant role in the assessment of CV risk, so much that, in low educated subjects with ED, the predictive efficacy of the Progetto Cuore risk engine, developed for the Italian population, is undermined. 5 The level of education attained is one of the most widely explored socioeconomic factors in the context of CV risk assessment, providing the most consistent results. 6 A lower level of education represents a surrogate marker of lower socioeconomic conditions as well as individuals' capability to catch and apply health advice provided in both healthcare and non-professional settings. Lower education levels have been associated, independently of other sociodemographic factors, with a higher lifetime risk of CV events, a higher CV morbidity and mortality. Despite this evidence, patients' level of education is not routinely collected in clinical settings, and it is not generally considered for CV risk assessment. Moreover, the CV risk profile according to the level of education attained in high-risk subjects, such as ED men, has not been thoroughly assessed. Therefore, the aim of the present study is to evaluate the association between the level of education and clinical, biochemical, and instrumental parameters known as related with higher CV risk in men seeking medical care for ED. Moreover, taking advantage of information on the occurrence of MACE retrospectively obtained in a subset of the same population, the present study assesses the role of a lower level of education as a predictor of incident MACE.

Cross-sectional study
The data of a consecutive series of 3733 men attending for the first time an outpatient andrology clinic for ED were collected in the period between 2002 and 2015. to the Italian education system, as follows: none/primary education (that begins at age 6 and lasts 5 years); lower secondary education (that begins at age 11 and lasts 3 years); upper secondary education (that begins at age 14 and lasts 5 years); and higher education (e.g., university and post-university degree). Men were interviewed before the beginning of any specific diagnostic procedure or treatment using the Structured Interview on Erectile Dysfunction (SIEDY) 7,8 and ANDROTEST. 9 SIEDY is a 13-item validated structured interview investigating several characteristics of ED with answers reported as a Likert scale. 7,10 The questions grouped into three scales allow identifying and quantifying the three pathogenic components concurring with ED (i.e., organic, relational, and psychic). 7 ANDROTEST is a 12item validated structured interview for the screening of testosterone deficiency in subjects with ED, with answers reported as a Likert scale. 9 Information on severity of ED was obtained with the following question (SIEDY, question 1A): "Is the erection sufficient for penetration?" Answers were categorized as a dummy variable being "erection sufficient for penetration in more than 50% of cases" corresponding to "no/mild ED," whereas "erection sufficient for penetration in less than 50% of the cases" corresponds to "moderate/severe ED." Libido was assessed by question #14 of SIEDY7 and the replies were then Total score (∑MHQ) 28.6 ± 15.5 33.6 ± 12.6 Note: Data are expressed as mean ± standard deviation when normally distributed, median (quartiles) when not normally distributed, and as percentages when categorical.
categorized as a dummy variable being "unmodified or moderately reduced desire" correspondent to "No/mildly impaired sexual desire," whereas "remarkably reduced or absent desire" correspondent to "Moderately/severely impaired sexual desire." Frequency of masturbation was investigated by question #7 of ANDROTEST 9 and categorized in a dummy variable with autoeroticism in the last 3 months more frequent than or equal to three times for week was defined as "normal," whereas autoeroticism less frequent than three times for week was  The American Heart Association and the National Heart, Lung, and Blood Institute (AHA/NHLBI) criteria for metabolic syndrome (MetS) definition were used. 13 Penile color Doppler ultrasound (PCDU) was performed according to international standard procedures 14 using 10 µg of alprostadil for inducing erection. Based on the medications used, the Chronic Diseases Score (CDS) was calculated and used as a summary of chronic illnesses. 15 The assessment of CV risk was evaluated using the Progetto Cuore risk engine developed in the Italian population to estimate the 10-year risk of occurrence of a first MACE, considering sex, age, and the presence of several conventional CV risk factors. 16

Longitudinal study
A longitudinal study was also performed for 956 out of the 3733 individuals who were included in the cross-sectional cohorts (Table 1).
MACE occurred in this cohort were retrospectively collected for a mean of 3.9 ± 2.4 years. The identification of CV events through the International Classification of Diseases codes was previously described. 5

STATISTICAL ANALYSIS
Mean ± standard deviation and median (quartiles) were used as measures of central tendency for parameters with normal or non-normal distribution, respectively, unless otherwise specified.
Analysis of covariance (ANCOVA) and binary logistic regression were applied for multivariate analysis for continuous and binary dependent  Age, smoking habit, and alcohol intake showed a significant association with the level of education and, therefore, were introduced as confounders in the following analyses.

Cross-sectional study
When compared to the category with the highest degree of education, patients belonging to the other groups showed, independently of age, smoking habit, and alcohol intake, a higher body mass index (BMI) ( Figure 1A) and WC ( Figure 1B). After adjusting for the aforementioned confounders, a lower level of education was associated with higher values of glycated hemoglobin ( Figure 1C) and blood glucose ( Figure 1D).
In contrast, no significant difference was found when considering blood lipid levels (data not shown).
As compared with subjects with university and post-university degrees, patients in the lowest categories of education showed a higher number of comorbidities, as expressed by the stepwise increase in CDS (Figure 2A). In particular, they presented a higher prevalence of diabetes, MetS, and CVD ( Figure 2B).
Among sexual parameters, after the adjustment for the aforemen- Interestingly, at variance with other outcomes explored, subjects who completed primary education scored similar to patients in the highest education group in most of the sub-categories, reaching even significantly lower scores for phobic anxiety symptoms ( Figure 5B-F).
Accordingly, patients with a primary education degree reported a history of psychopathology less frequently than those in the highest education group ( Figure 5A).

F I G U R E 3
Association between education level and sexual function parameters. Odds ratio (OR) and 95% confidence interval (CI) of erectile dysfunction, reduced masturbation frequency, and decreased sexual desire among different educational level groups (upper secondary: red symbol, lower secondary: orange symbol, and primary: green symbol, panel A); means and 95% CI of flaccid peak systolic velocity among different education level groups (panel B). Data are adjusted for age, smoking habit and alcohol intake, waist circumference, and chronic diseases score. *p < 0.05 versus higher education (university and post-university degree)

F I G U R E 4
Association between education level and hormone levels. Means and 95% confidence interval of plasmatic luteinizing hormone (LH, panel A) and total testosterone (panel B) among different educational level groups. Data are adjusted for age, smoking habit and alcohol intake, waist circumference, and Chronic Diseases Score. *p < 0.05, ***p < 0.0001 versus higher education (university and post-university degree) The comparison of the estimated 10-year CV risk derived from the Progetto CUORE algorithm showed that patients with lower or upper secondary education had a similar risk to those with university or post-university degrees, whereas the risk was significantly higher in patients with primary education, even after adjusting for confounders ( Figure 6A).

Longitudinal study
Among  Higher MHQ scores denote more severe symptoms. Data are adjusted for age, smoking habit and alcohol intake, waist circumference, and Chronic Diseases Score. *p < 0.05 versus higher education (university and post-university degree)

DISCUSSION
In the present study, we demonstrated that in people suffering from ED, a lower level of education is associated with a worse CV risk profile.
This result translates into an increased incidence of MACE associated with a lower education level.
It is well known that low socioeconomic status (SES) is associated with an increased risk of CVD, and level of education has widely been used as an indicator for socioeconomic position. 6 A lower educational level is associated with a higher prevalence of traditional CV risk factors. Some studies showed an inverse correlation between blood pressure and level of education. 17 Diabetes seems to affect more often low-educated and low-income people, and this is probably secondary to obesity and physical inactivity. 18  In line with the adverse CV risk profile and reduced flaccid PSV, 32 patients with the lowest education level had a higher estimated CV risk. When actual incident MACE has been considered, patients with primary education confirmed a higher risk as compared with more educated ones. Conversely, patients with lower secondary education, whose predicted risk was similar to subjects with higher education, had significantly higher occurrence of actual MACE. Not only the CV risk in comparison with higher educated patients was higher but also the absolute risk within the lower educated groups was underestimated. Indeed, when comparing the predicted versus the observed CV risk among subjects with primary or lower secondary education, an average of 20%-25% experienced a MACE at 8 years of follow-up, whereas according to the Progetto Cuore algorithm, less than a half of these would be expected. Although CV risk assessment using few major CV risk factors included in clinical prediction models is a simple and efficient approach in many settings, there is still a gap between the estimated and the observed CV events (the so called "residual risk").
This can be explained by factors other than those considered in the standard-risk algorithms, with a specific and independent role in determining an adverse CV outcome. Despite its role as a CV risk factor, the level of education (as any other socioeconomic factor) is not included This, in turn, may reduce health disparities among the most vulnerable sections of society.
The possible underestimation of the CV risk when considering only few major CV risk factors is particularly important for men with ED who represent a high-risk population independently of traditional risk factors. In these patients, a more precise CV risk assessment using other secondary risk factors may be useful. We previously reported that in men with ED and two unconventional CV risk factors such as low education and reported partner's hypoactive sexual desire, the observed number of events is significantly higher than the number estimated using a common CV risk algorithm (i.e., Progetto CUORE). 5 The risk of underestimating CV risk is particularly true for relatively healthy and younger individuals. In a recent study, we demonstrated that in men with ED, family history for cardio-metabolic diseases is associated with an adverse cardio-metabolic profile, and it is a predictor for incident MACE, particularly in low-risk individuals (younger and without major CV risk factors). 36 Hence, incorporating a simple information such as the level of education in the CV assessment in men with ED may better define CV risk profile resulting in more precise treatment decisions, especially in people without clearly determined risk factors.
Several limitations should be recognized. These results are derived from men consulting an andrology clinic for sexual dysfunctions; therefore, they could not be directly applied to subject without ED or with ED consulting non-specialist healthcare providers or not seeking medical care. Moreover, information about the level of education was obtained through participants' reports; however, this self-reported information has limited risk of recall bias. Finally, the identification of non-fatal MACE through registers may lead to misclassification. However, for the last two issues, the large sample size is likely to dilute possible mistakes.

CONCLUSION
In men suffering from erectile dysfunction, a lower education is associated with a more severe erectile dysfunction of atherogenic pathogenesis and with a worse cardio-metabolic profile. Moreover, it is a predictor of forthcoming major adverse cardiovascular events with a risk that overcomes that predicted by the Progetto Cuore risk scoring system.
There is a significant relationship between socioeconomic factors and cardiovascular health. Among them, the level of education is a costless but valuable information that should be always part of the cardiovascular risk assessment in men with erectile dysfunction. A low level of education, along with other secondary cardiovascular determinants, can help in the identification of men who may benefit from stricter risk factor control and whose risk would otherwise be underestimated.

ACKNOWLEDGMENTS
Open Access Funding provided by Universita degli Studi di Firenze within the CRUI-CARE Agreement.