Prevalence & predictors of sexual functioning & sex hormone profiles among men with opioid dependence: A community-based, cross-sectional study

Study design

This study is part of a larger longitudinal cohort study with a three-month follow up assessment; however, the present paper focuses only on the baseline findings from the original study. This community clinic at NDDTC, AIIMS specialises in treatment of opioid dependence, offering methadone and buprenorphine-based opioid agonist treatment (OAT). It serves a catchment area of 5-7 km radius, catering to ⁓1,000 patients per month. New patients are registered twice a week, while follow up visits are available on six working days.

Study population and sampling

The study population included patients diagnosed with opioid dependence (ICD-11: 6C43.2, WHO 2022) and seeking treatment at the community drug treatment clinic. All newly registered patients meeting the selection criteria were invited to participate in the study; those providing written informed consent were enrolled in the study. The original sample size was calculated for the longitudinal cohort design, which required 72 participants completing follow-up for adequate paired analysis. Anticipating 50 per cent dropout rate, we set a baseline recruitment target of 144 participants, and 143 were ultimately enrolled. As the present analysis reports only baseline cross-sectional data, we additionally conducted a post-hoc power and precision check using G*Power (Exact test family, correlation: bivariate normal model, two-tailed). Given the clinic’s fixed registration days and structured intake process, every eligible patient was approached consecutively, ensuring strict consecutive sampling.

Selection criteria

Eligible participants were men, within the age group of 18-50 yr, sexually active in the past three months (defined as at least one sexual intercourse with a partner), using heroin as their primary opioid for at least 25 days in the past month, newly registered at the clinic for receiving OAT, not on treatment for opioid dependence in the past month, and willing to provide blood and urine samples at baseline and follow-up. Exclusion criteria included severe mental illness affecting their ability to participate, harmful use or dependence on other psychoactive substances (except nicotine) in the past three months, presence of chronic medical or endocrinological conditions, and use of any medications affecting sexual functioning or sex hormone levels in the past three months (such as antidepressants, antipsychotics, hormonal replacements, and anabolic steroids).

Data collection

All participants were interviewed in a private setting by the first author, in a single session. Data were collected using a semi-structured proforma covering socio-demographic, substance use, clinical, and sexual history (sexual orientation, age at first intercourse, partners’ gender/type, number of partners, sexual practices, sexually transmitted infections, intercourse under intoxication, substances used to enhance performance, premature ejaculation history, privacy, and partner’s substance use). The standardised instruments included:

• Alcohol, smoking, and substance involvement screening test (ASSIST 3.0): It was used for assessing substances use other than opioids²⁵.Those with scores of ≥11 for alcohol and ≥4 for other psychoactive substances (denoting harmful or dependent use) were excluded from the study.

• Leeds dependence questionnaire (LDQ): It was used to measure opioid dependence severity over the last week. A score >20 indicates severe dependence²⁶.

• The International index of erectile function (IIEF-15): It was used to assess sexual functioning, covering five domains (erectile function, orgasmic function, sexual desire, intercourse satisfaction, and overall satisfaction). Each item is scored on a Likert scale, with higher scores indicating better sexual function. The IIEF-15 had acceptable internal consistency (α > 0.70) and test-retest reliability (r > 0.70), except for the orgasmic function scale²⁷.

• Fagerström test for nicotine dependence (FTND/FTND-ST): These were used to assess nicotine dependence severity for smoking and smokeless tobacco respectively^28,29.

• Clinical opiate withdrawal scale (COWS): It was used for opioid withdrawal severity³⁰.

• Depression, anxiety, and stress scale (DASS-SF): It was used to evaluate depression, anxiety, and stress symptoms. It gives a total raw score, along with three subdomain raw scores for depression, anxiety, and stress³¹.

• HIV risk taking behaviour scale (HRBS): It is an 11-item, clinician-rated tool measuring risky behaviours among injection drug users (IDU) related to HIV transmission³².

Sample collection

Blood and urine samples were collected from each participant at both time points using aseptic precautions. The samples were transported, processed, analysed, and stored at -80°C for further testing. Blood samples (10 mL) were drawn, prior to initiation of opioid agonist therapy, between 8:00 - 11:00 AM for haematological, biochemical, and hormonal assays. Most had last used heroin in the early morning prior to attending the clinic and were either intoxicated or about to enter early withdrawal at the time of sampling. Hormonal assays were conducted using electrochemiluminescence immunoassay (ECLIA) for serum-based hormones. GnRH was analysed separately using enzyme-linked immunosorbent assay (ELISA) in plasma.

Data management and statistical analysis

Data were managed using REDCap (research electronic data capture)^33,34, and statistical analysis was performed using licensed IBM-SPSS statistics for Windows, Version 29.0.2.0 (IBM Corp, Armonk, NY). Continuous data were summarised using mean (standard deviation) or median (interquartile range) based on the data distribution, while frequencies (percentages) were used for categorical variables. The normality of continuous data was assessed using Shapiro-Wilk test. Given that most continuous variables were not normally distributed, non-parametric tests were used for analysis. Spearman’s rank correlation was used to assess the relationship between the sexual functioning scores and sex hormone levels. A two-sided P value of <0.05 was considered statistically significant for all analyses.

The false discovery rate (FDR) correction using the Benjamini-Hochberg method was applied to adjust for multiple comparisons, ensuring control over Type I errors (adjusted P<0.00125).To examine whether hormonal variables independently predicted sexual functioning after accounting for other covariates, hierarchical multiple linear regression was conducted with IIEF total score as the dependent variable. Predictors were entered in four blocks: (i) demographics (age, education, occupation, marital status), (ii) opioid use variables (duration of use, daily expenditure, LDQ score), (iii) psychosocial/confounder variables (DASS total, HIV risk-taking, FTND), and (iv) hormones (total testosterone, prolactin, GnRH, DHEAS). In addition to the full-sample model, a sensitivity analysis restricted to married participants was conducted to include intimate partner violence (IPV) as an additional psychosocial predictor. Regression assumptions (linearity, homoscedasticity, collinearity) were checked and adequately met. Because residual normality was not satisfied, robust confidence intervals were estimated using bias-corrected accelerated (BCa) bootstrapping with 1000 samples.

Socio-demographic characteristics

The median (IQR) age of the study participants was 26 (24-30) years. The socio-demographic of the participants details are provided in table I.

Opioid and other substance use history

The median age of onset for heroin use was 19 years (IQR 17-24). All the participants reported using heroin by inhalational route (smoking or chasing) in their lifetime. About 19.6 per cent (n=28) reported injecting heroin at some point in life, and 7 per cent (n=10) had injected heroin in the past month. Participants reported near-daily heroin use, with an average of 29.5 days (SD = 1.2) of use in the past month. The median LDQ score was 27 (IQR 26-28), indicating severe OD.

About 7.7 per cent (n=11) reported lifetime dependence on alcohol, 1.4 per cent (n=2) on cannabis, and 88.8 per cent (n=127) on tobacco. About 93 per cent (n=133) participants were classified as moderate-risk and 6.3 per cent (=9) as high-risk for tobacco use according to ASSIST scores. According to ASSIST scores, all participants (100%, n=143) were classified as high-risk for opioid use.

All, except one, 142 participants reported current tobacco smoking. Based on the Fagerström test for nicotine dependence (FTND), 31 (21.8%) had low dependence, 24 (16.9%) had low-to-moderate dependence, 65 (45.8%) had moderate dependence, and 22 (15.5%) had high dependence. Among the 97 participants who reported smokeless tobacco use, 18 (18.6%) had low dependence, 25 (25.8%) had low-to-moderate dependence, 20 (20.6%) had moderate dependence, and 34 (35.1%) had high dependence. DASS-21 scores indicated that none of the participants had depression, anxiety, or stress. No participant was receiving psychotropics or hormonal agents at baseline, consistent with the exclusion criteria.

Marital history

About 55.2 per cent (n=79) were married [average frequency of marriage once (IQR 1-1)] and had one child (IQR 0-2). Among married participants, 87.3 per cent (n=69) reported being ‘very happy’ in their marriage, 3.8 per cent (n=3) as ‘happy,’ and 8.9 per cent (n=7) as ‘very unhappy.’ Among married participants, 81 per cent (n=64) reported perpetrating intimate partner violence (IPV), with verbal abuse being most common (79.7%, n=63), followed by physical abuse (50.6%, n=40), and sexual abuse (19.0%, n=15). Conversely, 39.2 per cent (n=31) of participants reported experiencing IPV from their partners, with verbal abuse being the most common, reported by all (39.2%, n=31), followed by physical abuse (6.3%, n=5). None reported experiencing sexual abuse from their partners.

Sexual history

Majority of the participants (93%, n=133) identified themselves as heterosexual in orientation and 7 per cent (n=10) as bisexual. The median age at first sexual intercourse was 17 yr (IQR 14-18), and the median time since last sexual intercourse was 7 days (IQR 3-30). All participants reported having female sex partners over their lifetime; 6.3 per cent (n=9) reported having male partners and 7.7 per cent (n=11) having transgender partners. In the last three months, all the participants had female sex partners, and 1.4 per cent (n = 2) had transgenders as sex partners.

About 55.2 per cent (n=79) engaged in sexual activity with their wives, 50.3 per cent (n=72) with girlfriend, 46.2 per cent (n=66) with FSWs, and 11.2 per cent (n=16) with men in their lifetime. In the past three months, 53.8 per cent (n=77) engaged in sexual activity with their wives, 28 per cent (n=40) with girlfriend, 18.9 per cent (n=27) with casual partners (females other than wife and girlfriend), and 17.5 per cent (n=25) with FSWs. Only 1.4 per cent (n=2) reported having sex with males in the past three months. Among the 66 participants who reported lifetime sexual intercourse with FSWs, 9.1 per cent (n=6) had unprotected intercourse. The majority of participants (93%, n=133) reported having had sexual intercourse under the influence of drugs at some point. Opioids were the most commonly involved substance during these encounters (97%, n=129), followed by alcohol (2.3%, n=3) and cannabis (0.8%, n=1). One-fourth of the participants (25.2%, n=36) reported using any over-the-counter medications, while more than half (55.9%, n=80) specifically used heroin for enhancing their sexual performance. Additionally, 6.3 per cent (n=9) of participants reported that their sexual partners use any psychoactive substances, with opioids (4.2%, n=6) and alcohol (2.1%, n=3) being the most common. Notably, all participants in this group (6.3%, n=9) had engaged in sexual intercourse while their partner was under the influence of drugs.

Sexual functioning and sex hormones

Erectile dysfunction (ED) was reported by 93.7 per cent (n=134) participants, while 95.1 per cent (n=136) had difficulty in achieving orgasm. About 94.4 per cent (n=135) reported decreased libido, while 99.3 per cent (n=142) expressed dissatisfaction in sexual intercourse. About 88.8 per cent (n=127) reported overall dissatisfaction with their sexual experiences. Premature ejaculation (defined as intravaginal ejaculatory latency time (IELT) of less than 2 min) was reported by 72 per cent (n=103) participants. Severity stratification of various domain of sexual dysfunction is shown in figure. Levels of sex hormones are shown in table II.

Close

Figure.

Severity of sexual dysfunction as per IIEF-15. The image presents the classification of male sexual dysfunction across five domains using validated scoring criteria. Erectile dysfunction (total score 1-30) is categorised into normal (25-30), mild (19-24), mild to moderate (13-18), moderate (7-12), and severe (1-6). Orgasmic dysfunction (total score 1-10) is classified as normal (9-10), mild (7-8), mild to moderate (5-6), moderate (3-4), and severe (1-2). Reduced sexual desire (total score 2-10) follows a similar classification with normal (9-10), mild (7-8), mild to moderate (5-6), moderate (3-4), and severe (2). Intercourse dissatisfaction (total score 0-15) is categorised as normal (13-15), mild (10-12), mild to moderate (7-9), moderate (4-6), and severe (0-3). Lastly, overall dissatisfaction (total score 2-10) is classified into normal (9-10), mild (7-8), mild to moderate (5-6), moderate (3-4), and severe (2).

Export to PPT

Correlations between sexual functioning and hormonal levels

Spearman’s rank correlation analysis revealed no significant relationship between most sexual functioning domains and sex hormone levels. However, LH showed a weak negative correlation with erectile function (ρ_s=-0.237, P=0.034), but this association did not remain statistically significant after FDR correction for multiple comparisons (adjusted P value <0.00125). No other sexual functioning domains showed significant correlations with hormone levels. (Supplementary Table I).

Hierarchical regression analyses

Hierarchical regression results are shown in supplementary tables II-V. In the full-sample model (N=143), demographic factors explained 13 per cent variance in IIEF scores (R²=0.133), with marital status (β=0.31, P=0.002) and occupational status (β=0.18, P=0.031) as significant predictors. Adding opioid variables modestly improved the model (ΔR²=0.045, P=0.07), with education emerging as significant (β=0.16, P=0.048) and LDQ showing a borderline negative association. Inclusion of psychosocial variables further improved fit (ΔR²=0.057, P=0.024), with HIV risk-taking emerging as a consistent positive predictor (β=0.25, P=0.003). In the final block, hormones (total testosterone, prolactin, GnRH, DHEAS) did not improve the model (ΔR²=0.015, P=0.63) and none were significant predictors. The final model explained 17 per cent of variance (Adj R²=0.17). Bootstrapped CIs confirmed these findings, with no change in significance patterns (Supplementary Table IV).

A sensitivity analysis restricted to married participants (N=78) allowed inclusion of IPV. In this model, IPV emerged as a significant negative predictor of IIEF scores (β=-0.29, P=0.016), while other psychosocial and hormonal variables were not significant. Bootstrapped regression confirmed this association (P=0.019; Supplementary Table V). Thus, across both models, demographic and psychosocial factors (marital status, occupation, HIV risk-taking, IPV) were more strongly associated with sexual functioning than hormonal measures.