Effective treatment and prevention of attempted suicide, anxiety, and aggressiveness with fluoxetine, despite proven use of androgenic anabolic steroids

The treatment of a man who attempted suicide after experiencing symptoms of anxiety and aggressiveness associated with the use of androgenic-anabolic steroids (AAS) is described. This report includes 30 days of inpatient treatment and a 6-month follow-up. Regular use of fluoxetine apparently prevented the onset of anxiety, depression, aggressiveness, and suicide ideation, even with the concurrent use of AAS. The urinary concentration of androgens, metabolites of AAS, and fluoxetine were monitored through analysis of urinary samples by the Brazilian Laboratory of Doping Control. Our results are congruent with previous findings describing the risk of suicide prompted by AAS use as well as the efficacy of fluoxetine in the treatment of mood disorders associated with the use of anabolic steroids.

. [1][2][3][4][5][6] Likewise, suicide has been reported as one the most frequent causes of unnatural death of anabolic steroid users. 7 Although the management of several AAS-related disorders has been described, 8 reports of treatment and the prevention of psychiatric symptoms concomitant to the detection of AAS and other drugs are scarce. To the best of our knowledge, a single case series 9 has reported the successful treatment of AAS-related withdrawal and depression in humans with fluoxetine. Similar responses were found in a controlled experiment using a mouse model. 10 We present the case of a 24-year-old male who attempted suicide following the use of AAS. The individual was also affected by depressive symptoms during a brief AAS withdrawal. These symptoms were successfully treated with fluoxetine. After discharge, the patient relapsed back into the use of AAS but did not present symptoms of anxiety, aggressiveness, or suicide ideation, if a regular intake of fluoxetine was maintained. The use of AAS and fluoxetine by the patient was monitored by analysis of urinary samples to minimize confounding factors related to the self-reported use of substances, such as memory and disclosure bias. 11,12 Naturalistic observation of AAS use, such as this study, is additionally challenged by the use of substances of unknown composition from illicit markets, 13 therefore highlighting the importance of complementing self-reported data with the analysis of biological samples. 14

| MATERIALS AND METHODS
The patient's clinical records were reviewed and described by obtaining prior informed consent from the patient. A summary of interventions is described in the "Timeline" image presented in  The following substances and   periods of detectable use are informed by the manufacturer: cocaine,   3 days; cannabis, 1 to 13 days; amphetamine, 9 days; methamphetamine, 3 to 5 days; benzodiazepines, 3 to 7 days; opioids, 2 to 3 days; and 3,4-methylenedioxymethamphetamine (MDMA), 2 to 3 days.
During inpatient treatment, urine samples were collected for 6 days. The patient provided four further samples during outpatient follow-up. Samples were analyzed for concentrations of fluoxetine, endogenous androgens, and the xenobiotic metabolites of AAS. The results and reference ranges are shown in Figure 1. Absolute values were normalized to a 0 to 1 scale so that parameters with distinct units and dispersions could be visualized in a single graph. The lower value detected of each androgen and metabolite corresponds to 0 in the graph, and the higher value corresponds to 1. A complete list of urinary concentrations of androgens, metabolites and fluoxetine can be foundat Supporting Information.
The urine sample preparation was based on enzymatic hydrolysis and liquid-liquid extraction of AAS. Aliquots of 2 mL urine samples were spiked with 40 μL of internal standard ( Table 1).
During the initial testing procedure (screening), the steroid profile was estimated by single point calibration. Calibration samples, ie, quality control of endogenous steroids (CQENDO) were prepared by spiking synthetic urine with endogenous steroids standards in their free  Table 2.
The pH was adjusted with 750 μL of 0.8 M phosphate buffer, and 50 μL of β-glucuronidase obtained from Escherichia coli (E. coli) was added. It was incubated at 50 C for 1 hour, and then 500 μL of aqueous buffer solution containing K 2 CO 3 /KHCO 3 20% (w/w) and 4 mL of methyl tert-butyl ether (TBME) was added. The mixture was stirred for 5 min and centrifuged at 3000 rpm for 5 min. After phase separation, the organic phase was evaporated to dryness under nitrogen and  and argonium (99.999% purity) gases were purchased from White Martins (São Paulo, Brazil). All reference materials were purchased from the National Measurement Institute (Australia).
The GC electron impact source mass spectrometry (GC-EI-MS/MS) conditions were: oven temperature program 140 to 230 C at 40 C/min, 230 to 280 C at 3 C/min, 280 to 300 C and held at 300 C for 3 min. The transfer line was set to 300 C and the ion source was set to 320 C. Electron ionization was performed using electron energy of 70 eV.
The collision energies were optimized according to the software Auto SRM. The dwell time was set to reach ten points across the peak for the narrowest peak. was set to 1e6 and maximum ion time (IT) was set to 100 ms. The sheath gas flow rate and the auxiliary gas flow rate were set to 60 and 20 respectively, the spray voltage was 3.90 kV, the capillary temperature was 380 C, the S-lens radio frequency (RF) level 80, the auxiliary gas heater temperature was 380 C.
Fluoxetine could be detected in t R = 6.15 min with the exact mass equal to m/z 310.14120. As internal standard 7-propilteofiline was used. It is detected at tR = 3.75 min and the monitored transition is m/z 372 ! m/z 210. This internal standard in 50 ng/mL was used to infer the concentration of fluoxetine in patient urine samples.
The data were evaluated using Thermo Fisher Scientific   Figure 1 for details of pharmacological treatment and the use of substances informed by the patient).
On admission, a urine toxicological analysis was performed for screening of recent use of drugs, as described in the Methods section.
This test showed solely the recent use of benzodiazepines, a result compatible with the drugs recently used by the patient. As the urine toxicological kit did not include screening for anabolic agents, these substances were tracked by applying the doping control analytical protocols described previously. 1352 ng/mL to 1500 ng/mL. As shown in Figure 1, urinary analysis revealed a tendency for regularization of androgen concentrations during inpatient treatment. For the duration of this report, the patient did not present health complications or symptoms, apart from those hereby described, that could have interfered in the study.
Mr Y was discharged from inpatient treatment 30 days after admission, with an alleged resolution to quit the use of AAS. The patient refused to attend meetings with a psychologist after discharge, but agreed to regular consultations with the first author, Dr Amaral. Despite continuous treatment and medical advice, the patient relapsed back into steroid use, as confirmed by urinary analysis. On the fourth follow-up consultation (D128), the use of AAS had raised all androgen concentrations above the 95% upper reference limit. 15,16 Mr Y informed us that he had voluntarily halted the use of fluoxetine, 14 days before this consultation because he believed the antidepressant was no longer necessary. However, the patient reported severe symptoms of anxiety, leading to our decision to initiate again treatment with fluoxetine, re-introducing the antidepres- are necessary to understand their interaction. Studies with animals exposed to AAS observed an effect of fluoxetine in reducing aggressiveness 22 and anxiety-like symptoms during AAS withdrawal. 10 Similar effects were reported when using the 5HT1A receptor agonist 8-OH-DPAT. 23 To the best of our knowledge, there are no reports of pharmaceutical interactions between fluoxetine and AAS. Besides, no association is observed between the treatment with fluoxetine and testosterone levels. 24 Regarding recommendations for treatment, it is important to highlight that AAS can trigger or aggravate the symptoms of mania and hypomania. 24 Since fluoxetine has the potential to induce mania in some patients, 25 this drug must be carefully monitored when prescribed to AAS users, namely in outpatient settings.
Underlying mechanisms of action of AAS on the serotonergic system in animals have been described by several authors, 22,23,26-28 and serotonin signaling is apparently involved in the anxiogenic and depressive effects observed during AAS withdrawal. 29 An increased risk of suicide in AAS users 7 seems to be related to the depressive and aggressive symptoms experienced by those patients. 30,31 This may be related to the effects of testosterone on the serotonergic pathways related to suicidality. 32 AAS also seem to decrease serotonergic activity in the basal forebrain and dorsal striatum. They also decrease 5HT1A receptors on the anterior hypothalamus and downregulate 5HT1B receptors in the globus pallidus and in the CA1 area of the hippocampus. 22 Therefore they affect the regions involved in the control of aggression. Another hypothesis is that AAS-related anxiety, aggressiveness, and depression are related to decreased synthesis of serotonin and increased cerebral levels of neuroactive kynurenines. 33 The limitations of this study include the lack of access to the pre-

| CONCLUSIONS
The onset of anxiety, aggressiveness, and a suicide attempt were