The validated Spanish version by Aluja and Farré (cited by Blázquez et al., 2008) of the Golombok Rust Inventory of Sexual Satisfaction (GRISS; Rust & Golombok, 1986) was used. It assesses sexual function of heterosexual couples who have an active sexuality. There is one version for men and one for women, each with 28 items, which are answered on a Likert scale of 5 points (from Never to Always). Nine different sub-scales are included: Infrequency, No communication, Dissatisfaction, Non-sensuality, Avoidance, Impotence (Erectile Dysfunction), Premature Ejaculation, Anorgasmia and Vaginismus. This study only analyzed data from the Erectile Dysfunction scale. Cronbach's alpha of the questionnaire for each of the two groups was calculated, resulting in a score of .73 for the substance user group and .74 for non-users.

Sexual Opinion Survey (SOS; Fisher, Byrne, White, & Kelley, 1988) was employed to assess the erotophobia-erotophilia. These constructs as defined by Fisher et al. (1988) are a learned disposition to respond to sexual stimuli along a continuum that extends from negative (erotophobia) to positive (erotophilia). Thus, people who score high on erotophobia tend to have more negative attitudes to sexual stimuli, and those who score highly in erotophilia will show more positive attitudes to sexual stimuli. The Spanish validation by Carpintero and Fuertes (1994) consists of 21 items answered on a Likert scale with 7 response options (1 = strongly agree; 7 = strongly disagree) was used. The authors reported a Cronbach's alpha of .86. In this study, a coefficient of .75 was obtained in the consumer group and .87 in the control.

State-Trait Anxiety Questionnaire (STAI; Spielberger, Gorsuch, & Lushene, 1982) is designed to measure anxiety. It has two self-rating scales to measure two independent concepts of anxiety: state and trait. Both scales have 20 items rated on a Likert scale of 0-3 points. The Spanish adaptation has good internal consistency between .90 and .93 for anxiety/state and between .84 and .87 for anxiety/trait. In this study was obtained from the anxiety scale a value of .89 in the substance user group and .90 in the non-user group, and the trait anxiety scale, .88 to .84 for users and non-users.

Non-probability sampling by conglomerates was performed, consisting of selecting treatment centers for drug addicts distributed throughout the country and asking them to participate in the study. To ensure that the sample was sufficiently large, provinces from all of the Autonomous Communities (at least one province in each) were selected. Participants were divided into two groups: the substance users group of people who were undergoing treatment for drug addiction and group of non-users of people who belonged to the same cluster; they were workers or treatment center workers. Once the questionnaires were answered, different treatment centers forwarded to the study authors for correction and subsequent statistical analysis.

In Table 2, the descriptive data for the two samples are presented. It is observed that the substance user group obtained higher scores on the Erectile Dysfunction GRISS scale, these scores indicate greater severity. When the raw score is transformed, using the criteria of the authors of the questionnaire (Rust & Golombok, 1986) it shows that the percentage of dysfunctional participants is higher in the group of male consumers of substances. Furthermore, those in the substance user group obtained higher anxiety scores: both state anxiety and trait anxiety, and lower scores on the SOS. These results indicate that male substance users are more erotophobic than nonusers.

The mean abstinence time of consumers was 8.66 months, with a standard deviation of 8.25. The minimum time was one month and maximum two years. The mode was 1 month. To check if withdrawal time is related to the presence of ED, the substance user group was split into three groups depending on withdrawal time at the moment of evaluation: 1 to 4 months, 5 to 12 months and over 12 months without substance use. The average score on the ED scale decreased as abstinence time increased. To check whether there were statistically significant differences in scores based on the time of withdrawal the Kruskal-Wallis test was performed. The result indicates that the differences are not statistically significant (χ2(2) = 2.30, p = .316). In addition, a bivariate correlation analysis was performed, analyzing the relationship between the withdrawal time variable and ED scale scores. Because the data did not meet the requirements for the use of parametric tests, Spearman Rho correlation was carried out. The results indicate no significant correlation between the time of withdrawal and the score on the scale of (Rho = -.04, p = .155).

To analyze the relationship between ED and the type of addictive substance consumed, a Kruskal-Wallis test was used to assess whether there were statistically significant differences between the different substances. The result indicated that there were no significant differences (p = .095) between the different drugs taken separately. For this reason it was decided to group people according to the type of substance. Were divided participants into groups by the effects of the substances on the central nervous system: stimulant, depressant or psychedelic. The classification was carried out as follows, people who had used cocaine or stimulant psychotropic drugs were included in the group of stimulants, persons who had used heroin, alcohol or psychotropic depressants were included in the group of depressants and finally those who had used cannabis were included in the group of psychedelic substances. Participants who had used mixtures of substances that did not pertain to the same group (cocaine and alcohol, cocaine and heroin, or three or more substances) and nonusers were left out of the analysis. Mean scores on the scale of Erectile Dysfunction depending on the substance used were: stimulants (M = 2.73, SD = 2.39), depressants (M = 3.25, SD = 2.41) and psychedelic (M = 3.27, SD = 2.31). The results of the nonparametric Kruskal-Wallis test are statistically significant (χ2(2) = 7.85, p = .020). Dunn's test showed that statistically significant differences were found between stimulants and depressants, indicating that people who have used depressants are more likely to have ED than those who consumed stimulants.

To test whether age negatively influences erectile response, the relationship between age and response to the scale of erectile dysfunction was analyzed using Spearman's Rho test. A significant result was yielded, thus taking all the data together (Rho = .14, p = .000), how the data using only consumer group (Rho = .15, p = .000). When the two groups are compared taking into account the age variable, it is observed that the differences are between 30-39 and 40-49 years of age (Table 3).

The main focus of this study was to analyze the difficulties in erectile response that may occur in men who have been regular users of addictive substances. In view of the results obtained, it can be concluded that people who have had experiences of consumption have more erectile difficulties than nonusers. Several studies found similar results. Vallejo-Medina et al. (2009) reported that the erection scale scores obtained by people with a history of substance abuse were lower than those of non-users. Jiann (2009) indicated that, in a study of 701 men in drug treatment centers in Taiwan, 36.4% had ED; 10.4% of which being severe cases. Jiann (2009) concluded that both snuff and alcohol and illegal drugs have adverse effects on erectile function. Therefore, the results obtained in this study are consistent with the few studies on ED in drug users.

As for abstinence, if the substance can cause ED in men, cessation of use should cause said dysfunction to disappear and improve sexual response over time. However, the results obtained indicate that abstinence does not improve erectile response in men who have been substance users. In the literature, we find similar results, indicating that sexual response does not improve with the cessation of substance use (Del Río, Fernandez, & Cabello, 2013). Vallejo-Medina and Sierra (2013) found that substance users did not significantly improve in the sexual area after a year of abstinence. Jiann (2008) obtained similar results, claiming that the neurological damage caused by addictive substances are long lasting. This is a very important aspect to consider therapeutically with people consuming of addictive substances, as against common sense, abstinence does not improve sexual response. Therefore, work of a specific depth regarding sexual dysfunction in drug treatment centers is necessary. One that includes evaluating and puts within reach therapeutic strategies that may mitigate these effects. Currently, in most of these centers, little sexological intervention is performed. This could be due to the little training of the people who work in these facilities in this respect, the limited time that one can spend on issues that are outside of substance consumption, or the resistance of patients to talking about these issues. Although, over time, the substance is no longer present in the body, the behaviors learned over years of consumption are maintained, probably being the reason behind that the sexual response does not improve with abstinence.

Regarding the erotophilia-erotophobia dimension, people who have had a history of use obtained lower mean scores on the SOS, so they are more erotophobic. Previous research shows that sexual attitudes are positively related to sexual satisfaction (Sánchez-Fuentes, Santos-Iglesias, & Sierra, 2014). It seems logical that men with the highest average score on the ED scale have greater resistance to having sexual encounters with potential mates, fear and anxiety that can cause them to fail in sexual encounters. It is also assumed that people who feel more secure in their relationships (suffer less ED) may tend to look more actively and be more open to new experiences (more erotophilic) than people who are feel more insecure. Similar results were found when erotophobia among women with a history of use and women with no history of use was analyzed (Del Río, Cabello, Cabello, & Lopez, 2012) as well in people with no history of substance use (Sierra et al., 2014).

As for anxiety, men who are substance users have greater anxiety on average than men who were not users. The relationship between substance use and anxiety is known (Cardoso-Moreno & Tomás-Aragonés, 2013; Charriau et al., 2013; Córdova & García, 2011). Men who are substance users scored statistically significantly higher on both state anxiety and trait anxiety than nonusers. Anxiety is an important factor in the development of different sexual dysfunctions including erectile dysfunction (Heiman & LoPiccolo, 2006; Lykins, Janssen, Newhouse, Heiman, & Rafaeli, 2012; Mosack et al., 2011).

The results also show that the kind of substance (stimulant, depressant or psychedelic) influences erectile response differently. Thus, men who have consumed depressants have a statistically significantly higher average score than men who have used stimulants. To date, there are few studies comparing the effects on sexual function depending on the kind of substance. Jiann's (2008) study, in which the influence of different substances is discussed in ED is noteworthy. This author concludes that different substances affect the erectile function of men, but does not indicate whether there is a substance that influences more than another, because, as indicated, we must be cautious in interpreting the results, since these are affected by different factors (psychological, physiological, environmental, cultural), not only by the substance consumed. The possible effects of substances used to adulterate the drugs of abuse should be added to the different factors that Jiann (2008) points out. Because of this, the conclusions from these data need to be taken cautiously.

Finally, it is found that variable of age influences erectile response. The result of this hypothesis is not surprising, since many studies examining age and sexual dysfunction point to a direct correlation between the two variables, i.e., the older the age, the higher the prevalence of sexual dysfunction (Bechara, Romano, & De Bonis, 2013; Sierra et al., 2014). Among the reasons it is argued that there are higher prevalences of other diseases (hypertension, diabetes, vascular disorders, musculoskeletal disorders, etc.), decreased libido, lack of vitamin B, the involvement of the vascular endothelium and decreased testosterone. Also, if one starts with the assumption that substance use affects sexual response, the longer that they were consumed (the older individuals may have a greater history of use) this should lead to greater impairment of the sexual response than with a short period of consumption. Although this research can not rule out this hypothesis, since it does not include data relating to the length of use to unravel which of the two variables have a greater influence.

Some of the limitations of this study should be noted. Such as the fact that data collection was carried out by different people. Information was collected by professionals in various treatment centers and although they had been previously given precise instructions on how to collect data, it could have been a variable which influenced the results. Another limitation has to do with the substances consumed by people in treatment for drug abuse. The information provided by those who seek treatment for drug addiction is not without bias, mainly for two reasons. The first is the fact that the person reporting on a substance that thinks is the main source of the problem (e.g. cocaine), without reporting, intentionally or not, the use of other substances (alcohol, cannabis) that may also influence their sexual behavior. Second, we must take into account that substances normally consumed are often adulterated with other active ingredients that the consumer has no information about, and these can also influence their behavior or have long-term effects. However this cannot be known because these are illicit substances. Therefore, it must be considered that the results are subject solely to the information the user provides regarding their history of consumption, with the imprecision that this could involve. Finally, it should also be considered a limitation that a higher percentage of substance users no partner than nonusers, as it could affect differentially on the answers provided, being a protective factor against different pathologies (Ross & Mirowsky, 2013; Zheng & Thomas, 2013).