Abstract
Introduction: Regular cocaine use has been associated with hormonal dysfunction including hypogonadism, which can lead to fatigue, reduced stamina, sexual dysfunction, and impaired quality of life. However, cocaine’s endocrine effects are largely under-reported in the scientific addiction literature and, in many cases, are not addressed within treatment services. The low profile of these adverse effects might be attributable to a lack of awareness and linkage with cocaine use, such that they are recognized only when an acute/emergency problem arises. Methods: We assessed endocrine diurnal function (adrenocorticotrophic hormone [ACTH], cortisol, and testosterone) in 26 healthy and 27 cocaine-dependent men and examined changes in hormone levels in response to a single 40 mg dose of the noradrenaline re-uptake inhibitor atomoxetine in a double-blind, placebo-controlled experimental medicine study. Results: When compared with healthy controls, diurnal and atomoxetine-induced changes in ACTH and cortisol showed greater variability in cocaine-dependent men. Interestingly, despite an exaggerated rise in ACTH following atomoxetine, an attenuated cortisol response was observed, and one-third of cocaine-dependent men had subnormal testosterone levels. Conclusion: Our findings point to a potential disconnection between the pituitary and adrenal responses in cocaine-dependent men, a higher rate of hypogonadism, and a pressing need for more research into the endocrine effects of cocaine and their clinical implications.
Introduction
Cocaine is the most popular illicit stimulant drug in Europe, especially amongst men, who are six times more likely than women to seek treatment for chronic cocaine use [1]. This gender discrepancy has been attributed to the endocrine actions of cocaine underlying its reinforcing effects [2]. Most work on the effects of cocaine on gonadal and adrenal function has, however, been conducted in women, with a particular focus on changes to contraceptive treatment and usage patterns at different stages in the menstrual cycle [3]. The endocrine effects of cocaine have also been widely studied in the context of stress-related relapse in drug-abstinent individuals [4], but the potential clinical sequelae of endocrine dysfunction in cocaine-dependent men has received relatively little attention. Converging lines of evidence from animals and humans suggest that cocaine acutely activates the hypothalamic-pituitary-adrenal (HPA) axis by stimulating endogenous hypothalamic corticotropin-releasing factor [5], possibly modulated by monoaminergic neurotransmitter systems [6]. Whilst low doses of cocaine have been shown to acutely increase both adrenocorticotrophic hormone (ACTH) and cortisol secretion, basal levels of cortisol and ACTH seem to remain stable within their respective reference ranges during chronic cocaine exposure [5, 7]. Blunting of the circadian rhythm of ACTH/cortisol secretion has also been noted and putatively linked to enhanced glucocorticoid-mediated inhibition of the HPA axis [8]. There are relatively few experimental studies on the effects of cocaine on the hypothalamic-pituitary-gonadal (HPG) axis in humans. These studies have shown that cocaine acutely increases luteinizing hormone, and decreases prolactin, whilst leaving testosterone levels unchanged [5, 9, 10]. Testosterone levels, however, decline following chronic or binge use of cocaine [11‒13]. This is in keeping with studies that measured adrenal and/or sex-steroid hormone levels in men with long-term cocaine dependence (frequently with co-morbid opiate dependence), which observed significant reductions in testosterone [14, 15]. The mechanism by which cocaine interacts with the HPG axis is largely elusive, but it has been speculated that cocaine changes the secretion, metabolism, and clearance rate of steroids [16]. Given that gonadal steroids are not only critical in regulating reproductive function but also interact with the central nervous system [16, 17], we sought to investigate the effects of the noradrenaline re-uptake inhibitor atomoxetine – a potential pharmacotherapy for cocaine dependence [18] – on HPA and HPG function in cocaine-dependent men. We hypothesized that atomoxetine-induced stimulation of the HPA axis might ameliorate HPA dysfunction in men with an addiction to cocaine.
Methods
Fifty-three male volunteers (20–60 years) were recruited from the local community. Twenty-seven were actively using cocaine, as verified by a cocaine-positive urine screen, and satisfied the DSM-IV-TR criteria for cocaine dependence [19]; 15 also had co-morbid opiate dependence (see Table 1). The remaining 26 individuals were healthy, without a history of drug/alcohol dependence. Individuals with HIV-infection, steroid medication (e.g., corticosteroids) and/or abuse of anabolic steroids were excluded from the study. All participants were medically screened before they were administered a single oral dose of 40 mg atomoxetine in a randomized, double-blind, placebo-controlled, crossover experimental medicine design. Blood samples for measurements of serum testosterone, serum cortisol (4.7 mL serum gel tube, room temperature), and plasma ACTH (2.6 mL EDTA tube on ice, plasma separated within 30 min) were drawn at a fixed time in the morning (either at 9:30 a.m. or at 10:30 a.m.), 30 min before dosing and repeated 5 h later. Morning sampling has been recommended for hormones with diurnal variation such as testosterone and cortisol [20]. Moreover, sampling at fixed time points on separate testing sessions has been used in previous studies to verify temporal stability [17]. All samples were processed in a CPA-accredited laboratory using the Siemens Immulite platform. The crossover design also allowed us to compare circadian pre- and post-treatment changes in cocaine-dependent men with those seen in healthy male controls, which takes individual variations in circadian rhythm into account without the need for additional adjustments. The study was approved by the National Research Ethics Committee and all participants provided informed consent in writing. Data reported here are part of larger study on the cognitive effects of atomoxetine, which have been published elsewhere [21‒26].
Demographics, clinical variables, and pooled baseline levels of steroid hormones of both testing days in cocaine-dependent men
No. . | Age, years . | Body mass index . | Average amount (stones) . | Last cocaine use (days ago) . | Regular cocaine use, years . | Co-morbid opiate dependence . | Methadone, Buprenorphine,* g . | Serum cortisol, nmol/L . | Plasma ACTH, ng/L . | Serum testosterone, nmol/L . |
---|---|---|---|---|---|---|---|---|---|---|
1 | 47 | 21.9 | 2 | 0.5 | 24 | - | 422.5 | 28.0 | Na | |
2 | 55 | 25.9 | 2 | 1 | 29 | - | 334.0 | 26.0 | Na | |
3 | 45 | 26.6 | 4 | 1 | 12 | - | 399.0 | 23.5 | 15.5 | |
4 | 32 | 18.7 | 3.5 | 1 | 12 | Yes | 40 | 359.5 | 29.5 | 16.7 |
5 | 50 | 20.3 | 2 | 1 | 22 | Yes | 4* | 281.0 | 14.5 | 21.0 |
6 | 51 | 22.9 | 2 | 1 | 28 | - | 553.5 | 31.0 | 13.1 | |
7 | 36 | 17.9 | 2 | 1 | 16 | Yes | 4* | 401.5 | 17.5 | 10.5 |
8 | 42 | 18.9 | 5 | 2 | 12 | - | 240.5 | 20.0 | 10.2 | |
9 | 32 | 21.6 | 2 | 1 | 11 | Yes | 2* | 242.5 | 25.5 | 14.3 |
10 | 27 | 19.1 | 2 | 2 | 11 | Yes | 16* | 152.5 | 15.0 | 10.1 |
11 | 28 | 20.3 | 1 | 1 | 12 | Yes | 60 | 192.5 | 11.5 | 8.6 |
12 | 38 | 21.4 | 2–3 | 1 | 20 | Yes | 8* | 347.5 | 9.0 | 13.9 |
13 | 40 | 21.0 | 2–3 | 1 | 16 | Yes | 30 | 421.5 | 19.0 | 4.1 |
14 | 36 | 22.2 | 1/8 oz+ | 2 | 11 | - | 389.5 | 24.0 | 17.3 | |
15 | 33 | 24.4 | 1 | 1 | 16 | Yes | 60 | 241.5 | 18.5 | 2.5 |
16 | 48 | 26.6 | 2–3 | 2 | 14 | - | 327.5 | 14.0 | 22.7 | |
17 | 36 | 18.2 | 2 | 1 | 15 | - | 290.0 | 10.5 | 11.0 | |
18 | 50 | 25.8 | 1 | 1 | 22 | - | 415.0 | 31.0 | 14.4 | |
19 | 41 | 21.0 | 2–3 | 1 | 17 | Yes | 8* | 456.5 | 21.0 | 6.9 |
20 | 44 | 22.9 | 1 | 1 | 19 | Yes | 12* | 323.0 | 17.5 | 20.0 |
21 | 37 | 18.2 | 2 | 1 | 23 | Yes | 40 | 282.0 | 10.5 | 7.6 |
22 | 39 | 25.5 | 2–3 | 0.5 | 14 | Yes | 284.0 | 11.0 | 14.2 | |
23 | 41 | 30.8 | 3 | 0.5 | 6 | - | 365.5 | 15.0 | 4.4 | |
24 | 42 | 22.3 | 4 | 0.5 | 20 | Yes | 165.5 | 8.5 | 2.6 | |
25 | 46 | 26.8 | 2 | 1 | 26 | - | 287.0 | 18.0 | 7.9 | |
26 | 53 | 24.4 | 3 | 0.5 | 22 | Yes | 414.0 | 58.0 | 5.0 | |
27 | 37 | 27.2 | 5 | 1 | 18 | - | 228.5 | 25.5 | 5.5 |
No. . | Age, years . | Body mass index . | Average amount (stones) . | Last cocaine use (days ago) . | Regular cocaine use, years . | Co-morbid opiate dependence . | Methadone, Buprenorphine,* g . | Serum cortisol, nmol/L . | Plasma ACTH, ng/L . | Serum testosterone, nmol/L . |
---|---|---|---|---|---|---|---|---|---|---|
1 | 47 | 21.9 | 2 | 0.5 | 24 | - | 422.5 | 28.0 | Na | |
2 | 55 | 25.9 | 2 | 1 | 29 | - | 334.0 | 26.0 | Na | |
3 | 45 | 26.6 | 4 | 1 | 12 | - | 399.0 | 23.5 | 15.5 | |
4 | 32 | 18.7 | 3.5 | 1 | 12 | Yes | 40 | 359.5 | 29.5 | 16.7 |
5 | 50 | 20.3 | 2 | 1 | 22 | Yes | 4* | 281.0 | 14.5 | 21.0 |
6 | 51 | 22.9 | 2 | 1 | 28 | - | 553.5 | 31.0 | 13.1 | |
7 | 36 | 17.9 | 2 | 1 | 16 | Yes | 4* | 401.5 | 17.5 | 10.5 |
8 | 42 | 18.9 | 5 | 2 | 12 | - | 240.5 | 20.0 | 10.2 | |
9 | 32 | 21.6 | 2 | 1 | 11 | Yes | 2* | 242.5 | 25.5 | 14.3 |
10 | 27 | 19.1 | 2 | 2 | 11 | Yes | 16* | 152.5 | 15.0 | 10.1 |
11 | 28 | 20.3 | 1 | 1 | 12 | Yes | 60 | 192.5 | 11.5 | 8.6 |
12 | 38 | 21.4 | 2–3 | 1 | 20 | Yes | 8* | 347.5 | 9.0 | 13.9 |
13 | 40 | 21.0 | 2–3 | 1 | 16 | Yes | 30 | 421.5 | 19.0 | 4.1 |
14 | 36 | 22.2 | 1/8 oz+ | 2 | 11 | - | 389.5 | 24.0 | 17.3 | |
15 | 33 | 24.4 | 1 | 1 | 16 | Yes | 60 | 241.5 | 18.5 | 2.5 |
16 | 48 | 26.6 | 2–3 | 2 | 14 | - | 327.5 | 14.0 | 22.7 | |
17 | 36 | 18.2 | 2 | 1 | 15 | - | 290.0 | 10.5 | 11.0 | |
18 | 50 | 25.8 | 1 | 1 | 22 | - | 415.0 | 31.0 | 14.4 | |
19 | 41 | 21.0 | 2–3 | 1 | 17 | Yes | 8* | 456.5 | 21.0 | 6.9 |
20 | 44 | 22.9 | 1 | 1 | 19 | Yes | 12* | 323.0 | 17.5 | 20.0 |
21 | 37 | 18.2 | 2 | 1 | 23 | Yes | 40 | 282.0 | 10.5 | 7.6 |
22 | 39 | 25.5 | 2–3 | 0.5 | 14 | Yes | 284.0 | 11.0 | 14.2 | |
23 | 41 | 30.8 | 3 | 0.5 | 6 | - | 365.5 | 15.0 | 4.4 | |
24 | 42 | 22.3 | 4 | 0.5 | 20 | Yes | 165.5 | 8.5 | 2.6 | |
25 | 46 | 26.8 | 2 | 1 | 26 | - | 287.0 | 18.0 | 7.9 | |
26 | 53 | 24.4 | 3 | 0.5 | 22 | Yes | 414.0 | 58.0 | 5.0 | |
27 | 37 | 27.2 | 5 | 1 | 18 | - | 228.5 | 25.5 | 5.5 |
Reference ranges for early morning measurements provided by the laboratory: cortisol 280–650 nmol/L, ACTH 5–50 ng/L, testosterone 8–29 nmol/L.
The majority of patients used cocaine in its free-based form, known as crack-cocaine; only one patient consumed powdered cocaine (indicated by +).
Twelve of 15 patients who also met the DSM-IV criteria for opiate dependence were maintained on methadone or buprenorphine (the latter were indicated by *).
For statistical analysis, data were square-root transformed, as appropriate, but untransformed values are shown in the online supplementary Figure and Table (for all online suppl. material, see https://doi.org/10.1159/000535584). Separate repeated-measures analysis of co-variance models were used to examine group differences in hormone levels, main effects of atomoxetine compared to placebo, and interaction effects between group and atomoxetine treatment. Age and body mass index were included as covariates to control for group differences. For the calculation of coefficients of variation and the group comparisons, MedCalc software (www.medcalc.org) was used. All statistical tests were two-tailed with significance levels at 0.05.
Results
Participants were in generally good physical health, matched for sex (all male) but differed with respect to age (t51 = 2.2, p = 0.032), body mass index (t51 = 3.7, p < 0.001), and smoking status (Fisher’s exact p < 0.001). Cocaine-dependent participants were on average 4 years younger (Mcocaine = 41.0 ± 7.5 SD, Mcontrol = 45.4 ± 7.0 SD), had a lower body mass index (Mcocaine = 22.7 ± 3.4 SD, Mcontrol = 26.0 ± 3.1 SD), and were all smokers compared with their healthy peers (4% smokers). The groups did not differ with respect to alcohol consumption (t32.1 = −0.2, p = 0.828), as measured by the Alcohol Use Disorder Test [27].
Baseline
Mean hormone levels did not differ between the groups (Fig. 1a–c) or between cocaine-dependent patients with and without co-morbid opiate dependence (online suppl. Fig. S1a–c), but the variability was notably greater in cocaine-dependent men – irrespective of co-morbid opiate use. Testosterone levels in all control participants were within the reference range; but some cocaine-dependent men showed markedly low serum testosterone levels, suggesting hypogonadism (one-third [n = 9] had serum testosterone <8 nmol/L; cut-off score provided by the laboratory) (see Fig. 1a; Table 1; online suppl. Table S3). Although significantly more cocaine-dependent men showed abnormally low testosterone levels compared with healthy male control participants (Fisher’s exact p = 0.018), this abnormality was concealed when comparing mean levels at baseline (pooled F1,43 = 0.37, p = 0.549), but coefficient variability was significantly increased in the cocaine group (F24,24 = 3.81, p = 0.002). Similarly, mean levels of cortisol (pooled F1,49 = 0.001, p = 0.989) and ACTH (pooled F1,49 = 0.8, p = 0.365) were not significantly different from controls at baseline, but cocaine-dependent men again showed significantly greater variability in ACTH (F26,25 = 2.22, p = 0.050) but not in cortisol (F26,25 = 1.23, p = 0.613) (Fig. 1a, b). The aforementioned findings were internally consistent at both pre-dosing sessions (i.e., pre-placebo and pre-atomoxetine) (see also online suppl. Table S1). Interestingly, the expected circadian change in cortisol on placebo was more marked in patients than in controls, as reflected by a trend of a time-by-group interaction (F1,49 = 3.20, p = 0.080) (Fig. 1a).
Changes in steroidal hormone levels following randomized administration of a single dose of 40 mg of atomoxetine and placebo using a crossover design in 26 healthy male control volunteers (blue) and 27 cocaine-dependent men (red). a Testosterone levels in controls were all within the normal range and remained stable over time; by contrast, a third of cocaine-dependent patients presented with abnormally low testosterone levels at baseline. This abnormality was concealed when comparing mean levels at baseline, which did not significantly differ from controls (pooled F1,43 = 1.2, p = 0.298). b ACTH levels in the cocaine group showed greater overall variability, although mean baseline levels were not significantly different from controls (pooled F1,49 = 0.8, p = 0.365). There was a significant atomoxetine-induced increase in ACTH in both groups, which was of a greater magnitude in the cocaine group (F1,34 = 5.4, p = 0.026). c Cortisol levels in the cocaine group showed greater overall variability at baseline, although mean levels for the cohort as a whole fell within the normal range on both sessions and were not significantly different from controls (pooled F1,49 = 0.001, p = 0.989). The diurnal changes differed significantly between the groups such that in the cocaine group, the decline in cortisol levels tended to be steeper on placebo (F1,49 = 3.2, p = 0.080), whereas the atomoxetine-induced increase tended to be blunted (F1,49 = 3.8, p = 0.058).
Changes in steroidal hormone levels following randomized administration of a single dose of 40 mg of atomoxetine and placebo using a crossover design in 26 healthy male control volunteers (blue) and 27 cocaine-dependent men (red). a Testosterone levels in controls were all within the normal range and remained stable over time; by contrast, a third of cocaine-dependent patients presented with abnormally low testosterone levels at baseline. This abnormality was concealed when comparing mean levels at baseline, which did not significantly differ from controls (pooled F1,43 = 1.2, p = 0.298). b ACTH levels in the cocaine group showed greater overall variability, although mean baseline levels were not significantly different from controls (pooled F1,49 = 0.8, p = 0.365). There was a significant atomoxetine-induced increase in ACTH in both groups, which was of a greater magnitude in the cocaine group (F1,34 = 5.4, p = 0.026). c Cortisol levels in the cocaine group showed greater overall variability at baseline, although mean levels for the cohort as a whole fell within the normal range on both sessions and were not significantly different from controls (pooled F1,49 = 0.001, p = 0.989). The diurnal changes differed significantly between the groups such that in the cocaine group, the decline in cortisol levels tended to be steeper on placebo (F1,49 = 3.2, p = 0.080), whereas the atomoxetine-induced increase tended to be blunted (F1,49 = 3.8, p = 0.058).
Atomoxetine Treatment
Atomoxetine administration increased levels of ACTH (F1,47 = 4.17, p = 0.047) and cortisol (F1,48 = 6.88, p = 0.012) in both groups; the variability of responses did not significantly differ between the groups (Fig. 1a, b). However, the atomoxetine-induced increase in ACTH was of a significantly greater magnitude in the cocaine group, as reflected by a significant time-by-drug-by-group interaction (F1,47 = 5.04, p = 0.030). Subsequent comparisons between cocaine-dependent men both with and without co-morbid opiate dependence did not reveal significant differences (online suppl. Table S2).
Discussion
This present study reveals evidence of endocrine dysfunction involving the HPA and/or HPG axes in a substantial number of cocaine-dependent men (Fig. 1; Table 1). The observed response profiles point to a potential disconnection between pituitary and adrenal responses that warrants further investigation. Importantly, co-morbid opiate use did not alone explain the increased prevalence of subnormal levels of testosterone in our cocaine-dependent patients as low testosterone levels were also observed in non-opiate using patients. Although our findings support previous studies, which suggest that the endocrine effects of cocaine and opiates are distinct and not additive [17], further research is warranted to investigate the distinct effects of both substances. In contrast to the well-known influence of opiates on the male endocrine system [28, 29], relatively little is known about endocrine changes associated with cocaine dependence. In fact, there are only sporadic reports in the literature of cocaine affecting gonadal and adrenal function in regular users [30, 31]. One reason for this lack of awareness might be that most studies solely report the mean changes in hormone levels across cohorts, which fail to identify the subgroups of patients with endocrine dysfunction. We did not formally diagnose hypogonadism but noticed that one-third of our patient sample repeatedly showed subnormal testosterone levels (Fig. 1a), which was not reflected by significant differences at group level. Given that all our participants were generally healthy, i.e., medical conditions that could potentially account for these dysfunctions (e.g., diabetes mellitus, steroid abuse) were not present, further research is warranted to elucidate the nature and clinical implications of this heterogeneity in cocaine dependence.
Almost all our patients were enrolled with drug treatment services at the time of the study, further supporting the notion of limited awareness of subnormal levels of testosterone in cocaine-dependent men, which increases the risk of such cases being overlooked in routine clinical practice, and potentially predisposing to medical complications (e.g., metabolic disturbance, reduced bone density, cardiovascular risk) and reduced quality of life (e.g., fatigue, reduced stamina, dysphoria, sexual dysfunction). In the case of hypoadrenalism, this could be a cause of major morbidity and may even be life-threatening. Greater medical attention to individuals with cocaine dependence could also benefit drug addiction treatment, given that current strategies largely rely on psychosocial approaches, which are only modestly effective in the short term [32], increasing the need for more effective treatments [33]. A discussion about potential endocrine treatment in a subgroup of cocaine-dependent patients is interesting, also in light of preliminary preclinical work suggesting that testosterone might enhance dopamine turnover [34], thereby providing additional support for further studies addressing this issue.
Clearly, the interactions between drugs of abuse and gonadal and adrenal function are complex. However, both affective and cognitive dysfunction have been shown to improve in response, for example, to androgen replacement therapy in other contexts [35]. Imbalances between the HPA and HPG axes (as reflected by peripheral levels of cortisol and testosterone) have also been shown to mediate facial affect processing [36‒38], social behaviour [39], and decision-making [40]. Similar observations have also been reported in cocaine-dependent patients [15]. Further work is thus warranted to understand whether hormone replacement therapy is indicated for a subgroup of individuals with cocaine dependence. We recommend that all patients with a history of chronic cocaine use should be considered for formal endocrine assessment.
Acknowledgments
The authors thank all participants for their contribution to this study and the staff at the NIHR Clinical Research Facility at Addenbrooke’s for their dedicated support. The authors also thank Claire Whitelock, Ilse Lee, and Miriam Pollard for their assistance with data collection and quality control, Sanja Abbott for support with data management, and Tsen Vei Lim for graphical assistance.
Statement of Ethics
The study protocol was reviewed and approved by the National Research Ethics Committee (12/EE/0519; PI: KDE). All participants provided written informed consent prior to study enrolment.
Conflict of Interest Statement
Karen D. Ersche is the recipient of an Alexander von Humboldt Fellowship for senior researchers (GBR 1202805 HFST-E) and receives editorial honoraria from Karger Publishers. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Funding Sources
The study was funded by the Medical Research Council (MR/J012084/1) and supported by the NIHR Cambridge Biomedical Research Centre (NIHR203312) and the NIHR Applied Research Collaboration East of England. The views expressed are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care. The study was jointly sponsored by Cambridge University Hospitals NHS Trust and the University of Cambridge.
Author Contributions
Karen D. Ersche is responsible for the concept and the design of the study, data acquisition, and data integrity. Karen D. Ersche, Jan Stochl, and Annette B. Brühl had full access to the data and take responsibility for the accuracy of the data analysis. Karen D. Ersche, Jan Stochl, Annette B. Brühl, and Mark Gurnell contributed equally to the interpretation of the data and the writing up.
Data Availability Statement
All data collected and analysed for this study are included in this article. Further enquiries may be directed to the corresponding authors.