Psychobiological correlates of smoking in patients with erectile dysfunction

Abstract

Although it is clear that cigarette abuse is closely linked to sexual dysfunction, it is still unclear which are the psychobiological correlates of smoking among individuals with sexual dysfunction. The aim of the present study is the assessment of the organic, psychogenic and relational correlates of erectile dysfunction (ED) in outpatients with different smoking habits. We studied the psychobiological correlates of smoking behaviour in a consecutive series of 1150 male patients, seeking medical care for ED. All patients were investigated using a Structured Interview (SIEDY^©), which explores the organic, relational and intra-psychic components of ED, and a self-administered questionnaire for general psychopathology (MHQ). In addition, several biochemical and instrumental parameters were studied, to clarify the biological components underlying ED. Current smokers (CS) showed a higher activation of the hypothalamus–pituitary–testis axis (higher LH, testosterone and right testicular volume) and lower levels of both prolactin and TSH. Hormonal changes were reverted after smoking cessation. CS showed a higher degree of somatized anxiety and were more often unsatisfied of their occupational and domestic lifestyle. Smoking, as part of a risky behaviour, was significantly associated with abuse of alcohol and cannabis. Both CS and past smokers (PS) showed an impairment of subjective and objective (dynamic peak systolic velocity at penile duplex ultrasound) erectile parameters. This might be due to a direct atherogenic effect of smoking, a cigarette-induced alteration of lipid profile (higher triglyceride and lower HDL cholesterol in CS than in non-smokers or PS), or due to a higher use of medications potentially interfering with sexual function. This is the first comprehensive evaluation of the biological and intrapsychic correlates to the smoking habit. Our report demonstrates that smoking has a strong negative impact on male sexual life, even if it is associated at an apparently more sexual-favourable hormonal milieu.

Introduction

The deleterious effect of cigarette smoking on erectile function has been demonstrated in the majority of cross-sectional epidemiological studies performed worldwide,^{1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17} including Italy,^{18, 19, 20} and confirmed in longitudinal surveys on selected populations.^{21, 22, 23, 24}

Although it is clear that cigarette abuse is closely linked to sexual dysfunction, it is still unclear which are the psychobiological correlates of smoking among individuals with sexual dysfunction. We recently developed and validated the Structured Interview for Erectile DYsfunction (SIEDY^©), which quantifies organic, marital and intrapsychic contribution to the determinism of erectile dysfunction (ED).²⁵ The aim of the present study is the assessment of the organic, psychogenic and relational correlates of ED in outpatients with different smoking habits.

Patients and methods

A consecutive series of 1150 patients attending for the first time the outpatient clinic for sexual dysfunction was studied. Patients with mental retardation, or coming from foreign countries, were excluded. Patients with frank hypogonadism (ie testosterone <7 nmol/l), severe hyperprolactinaemia (ie PRL >700 mU/l) and hypo- or hyperthyroidism (ie TSH <0.001 or >10 mU/l, respectively) were also excluded from the analysis. Characteristics of the sample are summarized in Tables 1 and 2. The site at which the study was performed is a university outpatient clinic specifically devoted to the diagnosis and treatment of male sexual dysfunction. Patients are referred to the clinic by their family doctors, or, occasionally, by other specialists.

Table 1 Sociodemographic characteristics of NS, PS and CS

Table 2 Biochemical characteristics of NS, PS and CS

Patients were interviewed prior to the beginning of any treatment, and before any specific diagnostic procedures, using the SIEDY^© Structured Interview.²⁵ This brief, multidimensional instrument designed for the pathogenetic assessment of ED provides scores for organic, marital and psychopathological factors affecting sexual function, which can be simultaneously present in the same patient. Self-reported smoking history was obtained from a specific question. Men were defined as current smokers (CS) if their history of smoking had lasted for at least 1 y and past smokers (PS) if they had smoked at least 1 y in their life and were not CS, and nonsmokers (NS) if they had never smoked or smoked less than 1 y. Patients were asked to specify any current pharmacological treatment; drugs listed in Broderick et al²⁶ were considered capable of interfering with sexual function. Hypoactive sexual desire (HSD) was assessed with item #14 of SIEDY^© Structured Interview.²⁵ Sexual intercourse frequency was investigated with a standard question: ‘how frequently did you have sexual intercourses in the last three months?’. Severity of ED was categorized on a four-point scale using question 1A of SIEDY^© appendix A (0, 1, 2 and 3 if full erection in less than 25, 25–49, 50–74 and over 74% of occasions, respectively), as previously described.²⁵

Patients were asked to complete the Middlesex Hospital Questionnaire, modified (MHQ),²⁷ a brief self-reported questionnaire for the screening of mental disorders, which provides scores for free-floating anxiety (MHQ-A), phobic anxiety (MHQ-F), obsessive–compulsive traits and symptoms (MHQ-O), depressive symptoms (MHQ-D), somatization (MHQ-S), and hysterical traits and symptoms (MHQ-H).

All patients underwent a complete physical examination, with measurement of blood pressure (mean of three measurements 5 min apart, in sitting position, with a standard sphygmomanometer), height and weight. Right testis volume was measured with a Prader orchidometer. Left testis volume was omitted, because its measurement is often biased by a concomitant varicocele.²⁸ Blood samples were drawn in the morning, after overnight fast, for determination of blood glucose (glucose oxidase method; Aeroset Abbott, Rome, Italy), total and HDL cholesterol, and triglyceride (automated enzymatic colorimetric method; Aeroset Abbott, Rome, Italy), testosterone (T), prolactin (PRL), LH, TSH and PSA (electrochemiluminescent method; Modular Roche, Milan, Italy). Penile duplex ultrasound examination was performed before and after PGE₁ intracavernous injection (10 μg) in all patients. The following parameters were considered: (a) in basal condition: basal peak systolic velocity and acceleration; (b) after PGE₁ intracavernous injection (10 μg): peak systolic velocity (Vpmax), end diastolic velocity and resistance index.²⁹

Data were expressed as mean±s.d. when normally distributed, and as median [quartiles] for parameters with non-normal distribution, unless otherwise specified. Differences among more than two groups were assessed with one-way ANOVA or Kruskal–Wallis test, whenever appropriate. Unpaired two-sided Student's t-tests were used for comparison of means of normally distributed parameters; when distribution could be normalized through logarithmic transformation, the same test was applied to logarithmically transformed data. In all other cases, Mann–Whitney U-test was used for comparisons between groups. Correlations were assessed using the Spearman's or Pearson's method whenever appropriate. Stepwise multiple linear regression was applied for multivariate analysis, categorizing current smoking and previous smoking as dummy variables. In order to verify more accurately the association of smoking with testosterone levels, separate linear regression models were applied to current, former and never smokers, in order to identify predictors of testosterone levels in each group. Those parameters which resulted to be significantly associated with testosterone levels in at least one of the three groups were used as potential predictors of testosterone in the whole sample, together with present smoking status (yes/no); in a separate analysis, lifetime smoking (current or previous smoking vs never smoking) was used as a dummy variable in the place of present smoking status. Chi-square test was used for comparison of categorical parameters. All statistical analysis was performed on SPSS for Windows 11.0.

Results

Sociodemographic parameters

Among the 1150 patients studied, 353 (30.7%) were NS, 382 (33.2%) PS and 415 (36.1%) CS. The sociodemographic characteristics of the sample are summarized in Table 1. PS were older (P<0.0001) when compared to NS and CS, while no difference was observed between NS and CS patients. PS had more often a stable relationship and a lower educational level when compared to CS and NS, but these differences were not confirmed after adjustment for age. CS were more often employed (as explored by question #1 of SIEDY^© Structured Interview) and showed a significantly higher prevalence of alcohol and cannabis abuse when compared to the rest of the sample; all these differences were confirmed after adjustment for age.

Metabolic and hormonal parameters

PS showed a higher BMI when compared with CS and NS, while no difference was found between CS and NS (27.5±4.5, 25.8±3.8 and 25.8±3.5 kg/m² for PS, NS and CS, respectively; P<0.0001 for PS vs NS and CS). The difference was confirmed even after adjustment for age (data not shown). CS showed significantly (P<0.05) lower HDL cholesterol and higher triglyceride levels when compared to the rest of the sample (Table 2). The difference between CS and PS in HDL and triglyceride levels was confirmed after adjustment for BMI (HDL: B=−4.92 ±1.22 mg/dl; P<0.0001; triglyceride: B=35.32± 9.88 mg/dl; P<0.0001; see also Table 2). No difference among groups was observed for glycaemia and total cholesterol levels (Table 2).

Logarithmically transformed LH levels were significantly (P<0.05) higher in PS and CS patients when compared to the rest of the sample, but the difference between PS and NS was not confirmed after adjustment for age (Table 2).

CS patients showed significantly (P<0.005) higher testosterone levels when compared to the rest of the sample (Table 2 and Figure 1); the difference between CS and PS was confirmed even after adjustment for age (B=1.6±0.5 nmol/l; P<0.0001) or BMI (B=1.3±0.5 nmol/l; P<0.01). Among CS, no differences in testosterone levels were observed between patients smoking less or more than 20 cigarettes (17.8±6.0 and 18.0±6.3 nmol/l; P=NS). When applying stepwise linear regression models separately to current, former and never smokers, using testosterone as dependent variable, and age, BMI, logarithmically transformed LH, HDL and triglyceride levels as putative factors, testosterone levels were significantly associated with BMI (B=−0.32±0.11, −0.34±0.09 and −0.48±0.13 nmol/l for NS, PS and CS, respectively; all P<0.01) in all age groups, while logarithmically transformed LH levels were significantly correlated with testosterone only among CS and PS (4.13±1.65 and 6.18±1.82 nmol/l, respectively; both P<0.05), triglyceride levels among NS and PS (adj r=−0.351 and −0.170, respectively; both P<0.05) and age in CS (adj r=−0.141; P<0.05). After adjustment for BMI, LH, age and triglyceride, current smoking was significantly associated with higher testosterone levels (B=1.17±0.52 nmol/l; P<0.05), while previous smoking did not show any significant association with androgen milieu (data not shown).

Figure 1

Logarithmically transformed testosterone levels in NS, PS and CS; P<0.01 for CS vs PS and NS.

Right testicular volume was significantly (P<0.05) higher in CS than in the rest of the sample (20.9±4.1 vs 20.3±4.4 ml; P<0.05), even after adjustment for age (data not shown). No difference among groups was observed in the prevalence of HSD or intercourse frequency (data not shown). CS and PS showed significantly lower (P<0.05) logarithmically transformed PRL and TSH levels when compared to the rest of the sample; at multiple regression analysis, differences between PS and NS were not confirmed after adjustment for age. No difference among groups was observed for logarithmically transformed PSA levels (Table 2).

Penile vascular parameters

Current and past smokers showed a significantly (P<0.01) lower Vpmax when compared to the rest of the sample (Figure 2). The difference between PS and NS was confirmed after adjustment for age (B=–5.4±2.1 cm/s; P<0.05). Moreover, a significant inverse correlation between Vpmax and duration of smoking was observed (r=−0.261; P<0.0001).

Figure 2

Logarithmically transformed PGE-1 stimulated peak systolic (Vpmax) in NS, PS and CS; P<0.01 for CS and PS vs NS.

A higher duration of smoking was observed in patients reporting previous cardiovascular diseases (CVD) or hypertension when compared to the rest of the sample (16[0–31] vs 31[18–40] y and 16[0–32] vs 26[0–36] y for patients without and with CVD and hypertension, respectively P<0.0001). At multiple regression analysis, both age and duration of smoking were significantly associated with CVD (adj r=0.224 and 0.102, respectively; P<0.0001), while no correlation between hypertension and smoking was observed after adjustment for age (data not shown). We did not find any difference in duration of smoking between patients with or without diabetes mellitus (data not shown).

Current and past smokers showed a higher prevalence of use of drugs potentially interfering with erectile function when compared with NS (11.5, 23.9 and 16.6% for NS, PS and CS, respectively; P<0.05); the difference between NS and PS was confirmed after adjustment for age (data not shown).

Intrapsychic parameters

Current and past smokers showed significantly higher (P<0.05) MHQ-S score when compared to the rest of the sample (Figure 3), but the difference between PS and NS was not confirmed after adjustment for age (data not shown). No differences were observed among groups for MHQ-A, MHQ-F, MHQ-D, MHQ-O and MHQ-H scores (data not shown). CS reported a significantly (P<0.05) higher prevalence of conflicts at home and lower prevalence of satisfaction at work (as explored by question #11 and #2 of SIEDY^©) when compared to the rest of the sample; these differences were confirmed after adjustment for age (data not shown).

Figure 3

Logarithmically transformed MHQ-S scores (somatization symptoms) in NS, PS and CS; P<0.05 for CS and PS vs NS.

Erection and SIEDY^© parameters

The proportion of subjects with full erection in over 75% of occasions was 11.9, 6.1 and 7.6% in NS, PS and CS, respectively (P<0.05 for NS vs PS and CS); the difference between NS and PS was confirmed after adjustment for age (data not shown). Among CS or PS, we observed a significant (r=0.258; P<0.0001) correlation between SIEDY^© scale 1 scores and duration of smoking. No differences were observed among groups for SIEDY^© scale 2 (data not shown). SIEDY^© scale 3 scores were higher in CS when compared to the rest of the sample (3[1–4], 3[1–4] and 3[2–5] for NS, PS and CS, respectively; P<0.05 for CS vs the rest of the sample), but the difference between CS and PS was not confirmed after adjustment for age (data not shown).

Discussion

In this study, we report for the first time the simultaneous evaluation of biological, relational and intrapsychic correlates of the smoking habit in 1150 patients, seeking medical care for sexual dysfunction.

Among biological factors, we found that CS have the highest degree of activation of the hypothalamus–pituitary–testis (HPT) axis, as revealed by the highest concentration of circulating LH and testosterone and by the largest volume of the right testis. The association between smoking and HPT axis activation retains significance after correcting for possible confounding factors as confirmed at multivariate analysis. In one hand, this might imply that nicotine, or any other tobacco-derived chemical (pyiridine, pyrazines, nitrosamines, phenols, indoles, hydrocarbons, tar, quinolines), chronically stimulates LHRH-secreting cells in the hypothalamus, which, in turn, activate pituitary and, downstream, the testicular production of T. This is in keeping with recent results on increased release of LH after acute exposure to high nicotine cigarette smoking in nicotine-dependent subjects³⁰ and with the LHRH-secreting activity of nicotine in LHRH-secreting neurons GT1–7.³¹ A positive effect of tobacco exposure on LH and T was previously described in a prospective study on smoking in a large series of infertile men.³² This finding is also in agreement with results from several cross-sectional studies, showing a positive association between smoking and T plasma levels;^{33, 34, 35, 36, 37, 38, 39, 40, 41} men who quit smoking have, during an 11-y follow-up, an increased risk of hypogonadism. Accordingly, in the present and Trummer et al³² study, quitting smoking abrogated the tobacco-induced stimulation of T secretion. Interestingly, in our, as well as in other, reports,^{32, 40} the negative effect of smoking cessation on HPT was independent of tobacco-withdrawal-induced weight gain, which per se could substantially contribute to the aging-related T decline.^{41, 42, 43} In fact, T decline after smoking cessation retains significance even after correction for confounding factors, as age and BMI. However, BMI plays an independent and crucial role in the control of testosterone plasma levels; in fact in our study BMI was the only parameter significantly correlated to each smoking group even after adjustment for confounding factors. This is in accordance with previous reports showing that BMI is the best predictor of testosterone plasma level decline with age.^{41, 43} On the other hand, it is possible that the smoking habit acts as a potential confounder in cross-sectional studies on T, because smoking, as a part of risky behaviour, is initiated more often in sensational seekers, which are characterized by higher androgen levels.^{44, 45, 46}

Cigarette smoking is indeed a socially accepted addictive disorder⁴⁷ often present in sensational seekers.⁴⁸ This study confirms that cigarette smokers display health risk behaviour; in fact, CS reported a greater abuse of recreational substances, as cannabis and alcohol, than the rest of the sample, even after adjustment for age (Table 1). This increased appetitive and health risk behaviour that, in our sample, characterizes the lifestyle of ED smokers may be either the cause or the consequence of the referred higher degree of domestic conflicts and/or job dissatisfaction and might underlie the highest score in Scale 3, which explores problems in the intrapsychic domain.²⁵

Recent studies suggested that nicotine dependence is significantly associated with anxious and depressive symptoms and disorders, and that the prevalence of mental disorders in nicotine-dependent smokers is higher than in nonsmokers.^{49, 50, 51} Data obtained in our study, showing no significant statistical differences between CS, NS and PS on MHQ-D scores, do not confirm such an association between smoking and depressive symptoms. Moreover, no significant differences between the three groups were observed on the other MHQ subscales, with the exception of the somatization one, whose scores were significantly higher in CS than in NS. This result suggests that, unlike NS, smokers often suffer from physical symptoms due to high levels of anxiety; moreover, the higher prevalence of conflicts at home and the lower prevalence of satisfaction at work detected in CS subjects could suggest the use of smoking as a tool to cope with different stressors, as previously observed.^{52, 53} It should also be considered that a selection bias could have determined an underestimation of psychological disturbances in CS. In fact, the sample is composed of patients referring to a clinic for sexual dysfunction; the greater prevalence of vascular (ie, organic) abnormalities in CS and PS could be associated with an under-representation of psychogenic and relational components of the pathogenesis of ED in comparison with the rest of the sample.

The association of smoking with higher androgen levels could be the consequence of the effect of smoking on the hypothalamo–pituitary–gonadal axis; alternatively, an increased androgen milieu could facilitate health risk behaviours, including smoking.^{42, 43, 44, 45, 46} Cross-sectional surveys, such as this one, do not allow any conclusion on the direction of causal relationships. However, it is notable that PS show LH, testosterone and testicular volume similar to those of never smokers, suggesting that cigarette smoking stimulates HPT axis activity, and that the effect is lost with smoking cessation. Conversely, it is still possible that successful smoking quitters are atypical of the broader population of smokers and, possibly, less androgenized than those who still pursued the hazard habit. It should be considered that present data do not provide any information on the pathophysiological mechanisms underlying increased testosterone levels in smokers, which remain largely speculative. However, the difference in androgen milieu between smokers and NS is relevant and should be considered in clinical practice in the characterization of patients with ED.

We also found that CS have significantly lower PRL concentration than the rest of the sample, after adjustment for age, a finding in agreement with previous studies in cohorts of male³² and female⁵⁴ subfertile subjects of similar size. In addition, we found that CS have lower TSH levels than NS or PS, after adjustment for age. This has been previously observed only in large population studies.^{55, 56, 57} The lack of measure of thyroid hormones and auto-antibodies does not allow us to draw conclusions on this particular finding. However, it is interesting to note that nicotine administration or exposure to cigarette smoke induces, in experimental animals, a decrease in PRL and TSH, due to activation of hypothalamic nicotinic receptors, which in turn leads to an increased activity of tubero-infundibular dopamine neurons (for a review, see Fuxe et al⁵⁸).

The smoking-related advantage in androgen milieu and PRL levels is not of great help in improving sexual activity, because indeed CS do not have a higher sexual desire and intercourse frequency than the rest of the sample. In contrast, they have more difficulties in achieving an erection, most probably because of a smoking-related damage to the vascular system that, in our sample, is persistent even in PS. The mechanisms underlying smoke-induced vascular damage are complex, and only partly understood. It has been suggested that smoking increases oxidative stress, therefore initiating and perpetuating an endothelial dysfunction (for reviews, see Ambrose and Barua⁵⁹ and Landmesser et al⁶⁰). In addition, the current smoking habit has a reversible but still deleterious effect also on lipid metabolism. In fact, according to several previous reports,^{61, 62, 63, 64} CS have a higher triglyceride and a lower HDL circulating level than NS or PS. These lipid changes may contribute to the development of the atherogenic process underlying vascular damages in penile arteries. It is interesting to note that, after smoking cessation, both triglyceride and HDL modifications were no more apparent, further suggesting the usefulness of quit smoking for sexual health.

Both CS and PS have a lower peak systolic velocity at penile colour duplex ultrasound, which was strongly correlated with duration of the habit. This is in partial contrast with a recent report showing a lack of difference in peak systolic velocity between smokers and NS ED patients.⁶⁵ However, that study has been conducted in a cohort of ED patients at least 10-fold smaller than the present one. According to colour duplex ultrasound results, our current and past-smoking patients referred more often a history of CVD and reported a higher use of medicaments potentially interfering with erectile function and, overall, higher scores in SIEDY^© scale 1, which explores the organic component of ED.²⁵ It is noteworthy that several medications used for the treatment of CVD (eg, beta blockers) are associated with ED.

In conclusion, we report for the first time that CS, albeit showing a favour sex-oriented hormonal profile (high activation of HPT and low PRL), display health risk behaviour and are actually more dissatisfied of their private and job lifestyle, resulting in a higher degree of somatized chronic anxiety. As a final result, they have a worse erectile activity, because of pronounced signs and symptoms of vasculogenic ED. Smoking cessation, although associated with weight gain and decline in T, reverses the negative metabolic changes associated to cigarette exposure. Therefore, as indicated in warning labels on cigarette packages, smoking is deleterious for several psychobiological aspects of (male) sexual life and specific community programmes for smoking cessation should be encouraged, in view of a better sexual health.