Impact of Duloxetine on Male Fertility: A Randomized Controlled Clinical Trial

Abstract

This study assessed the impact of duloxetine (serotonin and norepinephrine reuptake inhibitor) on semen parameters, sperm DNA fragmentation, and serum hormones. We performed a double-blind, placebo-controlled, randomized clinical trial of duloxetine 60mg or placebo daily for 6 weeks (5 weeks full-dose and 1 week taper). The primary outcome was the proportion of men with abnormal DNA fragmentation during and after duloxetine administration. Secondary outcomes were changes in semen parameters and hormones on treatment (2 and 6 weeks) and after discontinuation (8 and 10 weeks). Sixty-eight healthy males aged 18–65 were included. Duloxetine was not associated with an increase in the proportion of participants with abnormal sperm DNA fragmentation terminal deoxynucleotidyl transferase dUTP nick-end labeling scores (>25%) on treatment (p=0.09) or after treatment (p=0.56), nor did median sperm DNA fragmentation increase on treatment. Compared with placebo, there were no changes in bulk semen parameters during treatment. Limited changes in hormonal values were detected. This first published human study of serotonin and norepinephrine reuptake inhibitors on male fertility revealed no clinically meaningful effects on sperm DNA fragmentation, semen parameters or serum hormones. Duloxetine, and possibly other serotonin and norepinephrine reuptake inhibitors, may be considered for men desiring fertility who require anti-depressant treatment.

INTRODUCTION

Despite the documented negative impact of some anti-depressants on male fertility, an estimated 5.5% of young, fertile men between the ages of 18 and 39 from 2015–2018 received an anti-depressant medication prescription(Brody DJ, 2020; Tanrikut, Feldman, Altemus, Paduch, & Schlegel, 2010; Tanrikut & Schlegel, 2007). To date, the majority of the studies investigating male infertility have focused on one specific class of anti-depressants, selective serotonin reuptake inhibitors (SSRIs)(Beeder & Samplaski, 2020). However, there are numerous other classes of anti-depressants that also have a broad range of indications, including treatment of psychiatric conditions, premature ejaculation, and neuropathic pain(Gebhardt, Heinzel-Gutenbrunner, & Konig, 2016; Higgins, Nash, & Lynch, 2010; Schatzberg, 2000). One commonly prescribed class is the serotonin and norepinephrine reuptake inhibitors (SNRIs), which include venlafaxine and duloxetine(Cascade, Kalali, & Thase, 2007; Ferguson, 2001).

The putative negative impact of SSRIs on male fertility include substantial decreases in sperm quality and testosterone production. Our prior study demonstrated nearly 50% of healthy men will have abnormal sperm DNA fragmentation, reduction in testosterone levels by approximately 200 ng/dL and impacts on sexual function (worsened erectile and ejaculatory function)(Tanrikut et al., 2010). Other studies have also described variable adverse effects on semen parameters(Beeder & Samplaski, 2020; Koyuncu et al., 2011; Safarinejad, 2008; Tanrikut & Schlegel, 2007). The exact physiologic mechanism of the effect of SSRI on male fertility has not been directly evaluated, but the timing of observed effects of SSRIs suggests that sperm DNA damage occurs as a consequence of alterations in sperm transport(Tanrikut & Schlegel, 2007). The adverse effects on sperm DNA and semen parameters during SSRI treatment are similar to changes seen in men with known defects of sperm transport (e.g., partial vasal obstruction or spinal cord injury)(Iammarrone, Balet, Lower, Gillott, & Grudzinskas, 2003). Another possible mechanism may be a direct spermicidal effect through the binding of drug to sulfhydryl groups on mature spermatozoa, although drug effects wash out faster than would be expected from this mechanism(de Lamirande & Gagnon, 1998; Kumar et al., 2006; Tanrikut et al., 2010; Wolf & Kuhn, 1992). The only other class of anti-depressants studied in human male fertility is tricyclic anti-depressants (TCAs), and these studies were conducted more than forty years ago(Levin, Amsterdam, Winokur, & Wein, 1981; Padron & Nodarse, 1980). Very limited evidence exists for the impact of SNRIs on male infertility aside from in vitro and animal studies, which have shown no substantial impact on semen parameters(Beeder & Samplaski, 2020).

This study aimed to evaluate the impact of an SNRI, specifically duloxetine, on various parameters used to evaluate male fertility potential including semen parameters, sperm DNA fragmentation, and serum hormone levels.

METHODS

Participants

Healthy male volunteers, aged 18–65 years, were screened by research personnel and excluded if they had varicoceles, abnormal semen analysis, ongoing attempts to initiate pregnancy, current sexual dysfunction (moderate or severe on the International Index of Erectile Function (IIEF) survey), history of seizure disorder, history of previous chemotherapy or radiation, current psychiatric disorder, history of bipolar disorder, or family history (including cousins and grandparents) of bipolar disorder, major depressive disorder, or suicide. Subjects also were excluded if they used any psychotropic agents (prescription or herbal), anticonvulsants, or sleeping pills more than once per week, hormonal medications (daily or intermittent, including glucocorticoid pills, inhalers or creams in the previous three months), medications that impact sexual function, or prescription or non-prescription medications that enhance sexual function. Furthermore, those who had an inability to follow instructions and complete questionnaires in English, consumed tobacco or illicit drugs, or who consumed more than two ounces of alcohol daily were excluded. All included participants provided written informed consent before enrollment.

Study Design

A double-blinded, placebo-controlled, randomized clinical trial was conducted. Enrolled participants were randomized 1:1 into two groups, receiving either placebo or active study drug using block randomization. Block size was 5 and a seed was provided to replicate the randomization sequence. Participants were enrolled by members of the research team. The randomization scheme was programmed by the study statistician, and patient assignment was completed randomly by the investigational pharmacy based on the randomization algorithm. The active study drug group received 60mg of duloxetine once daily for 5 weeks, followed by 1 week of 30mg of duloxetine daily (to reflect standard therapeutic drug tapering) for a total of 6 weeks on duloxetine. The placebo group was given an identical appearing placebo for 6 weeks.(Frampton & Plosker, 2007)

Participant semen and serum samples were collected at a total of five timepoints: baseline, week 2 (on treatment), week 6 (at completion of treatment), week 8 (off treatment) and week 10 (study completion). At each timepoint, participants provided a semen sample via masturbation following standard collection protocols. Participant serum hormone levels, including luteinizing hormone (LH), follicle-stimulating hormone (FSH), testosterone (T), and estradiol (E), were drawn between 8:00AM and 10:00AM. All semen samples were analyzed for total volume, sperm concentration, motility and morphology and also assessed for sperm DNA integrity utilizing the TUNEL (terminal deoxynucleotidyl transferase dUTP nick-end labeling) assay(Tanrikut et al., 2010).

Participants were asked to complete an IIEF survey prior to medication initiation (baseline), and at weeks 6 and 10, as well as the Men’s Sexual Health Questionnaire (MSHQ) at baseline, weeks 2, 6, 8, and 10.

The study was approved monitored by the Institutional Review Board at Weill Cornell Medicine and registered with ClinicalTrials.gov (NCT03038867).

Laboratory Testing

All laboratory and research personnel were blinded to randomization group.

Semen Analysis:

Samples were collected into sterile containers and were kept at 37°C for 30 minutes to allow for liquefaction. Samples were analyzed by a single trained andrology laboratory technician at 400x magnification and semen analysis values were reported in accordance with the World Health Organization Laboratory Manual for the Examination and Processing of Human Semen, 5^th edition(Organization, 2010).

Hormone Evaluation:

Peripheral blood was obtained from each participant via venipuncture, with serum separated and stored at −20°C. Samples were run in duplicate using commercially available enzyme immunoassay kits for testosterone (IB79106, Immuno-Biological Laboratories, Minneapolis, MN, USA) and estradiol (IB79331, Immuno-Biological Laboratories, Minneapolis, MN, USA) and enzyme-linked immunosorbent assay kits for follicle stimulating hormone (FSH) (11-FSHHU-E01, ALPCO, Salem, NH, USA), LH (11-LUTHU-E01, ALPCO, Salem, NH, USA) and prolactin (25-PROHU-E01, ALPCO, Salem, NH, USA).

Sperm DNA Fragmentation:

For each participant sample, four glass slides with air-dried semen smears were utilized. The In-Situ Cell Death Detection Kit with Fluorescein isothiocyanate (FITC; Roche Diagnostics GmbH, Mannheim, Germany) was used. Slides were fixed with 4% paraformaldehyde (1mL) in phosphate-buffered saline (PBS) and incubated for 1 hour at room temperature in darkness. Slides were then washed with ice-cold PBS and permeabilized with Triton-X in 0.1% sodium citrate for 5 minutes. Slides were washed with PBS and incubated with a mixture of TUNEL enzyme solution. Parafilm M strips (Alcan Packaging, Darien, CT, USA) were applied to each slide and they were incubated in dark, moist chambers for 1 hour at 37° Celsius. Following incubation, the Parafilm M strips (Alcan Packaging, Darien, CT, USA) were removed from the slides and were then washed with PBS. Vectashield (Vector Laboratories, Burlingame, CA, USA) with DAPI was applied for nuclear counterstaining and a coverslip added before incubation overnight. Negative and positive controls were tested with each batch. Slide analysis was completed using a fluorescent microscope at 400x magnification. DAPI-positive and FITC-positive cells were counted from the same field. At least 100 DAPI positive cells were counted for one single tally. The number of FITC-positive cells detected was divided by DAPI-positive cells to produce the percentage of TUNEL-positive cells in the sample. At least four separate fields were analyzed for each sample and a mean of the slide evaluations taken.(Tanrikut et al., 2010)

Statistical Analysis

An a priori power calculation was completed assuming 5% of control group subjects would have abnormal sperm DNA fragmentation, and 30% of subjects would have abnormal sperm DNA fragmentation in the treatment group. With a power of 80% and two-sided a of 0.05, a total of 35 enrolled participants were required for each group (total of 70 participants).

Descriptive statistics for demographic and clinical factors of interest were analyzed for the placebo and treatment groups, separately. The Wilcoxon rank-sum test and Fisher’s exact test were used to compare 1) mean TUNEL values and 2) proportions of subjects with TUNEL values >25%, respectively, between the treatment and placebo groups, at each time point. The paired t-test and McNemar’s chi-square test were used to compare change in mean TUNEL value and change in proportion of abnormal results, respectively, between groups at each time point. The secondary outcomes of interest were compared between the two groups using similar statistical tests as noted above. For secondary and exploratory predictors, no correction for multiplicity of predictors was performed given the small sample size. All p-values were two-sided with statistical significance evaluated at the 0.05 alpha level. Ninety-five percent confidence intervals (95% CI) were calculated to assess the precision of the obtained estimates. Multiple imputation methods were used to impute values for missing-at-random data. All analyses were performed in R Version 4.0.2 (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

A total of 77 men were screened for inclusion and 68 subjects enrolled (Figure 1) between 2016–2018. The study concluded following targeted accrual, and all missing datapoints occurred secondary to lost to follow-up. Mean age of the participants was similar between groups (placebo 40.9 years ± 11.7 vs. duloxetine 41.5 years ± 12.4, p = 0.85). There were no differences in ethnicity, number of children, and sexual activity or function between placebo and duloxetine treatment groups (Table 1). No baseline differences were observed in baseline IIEF and MSHQ responses (Table 1).

Figure 1 –

Consort Diagram for Clinical Trial Enrollment

Table 1 –

Baseline demographic data

	Placebo (N=34)	Duloxetine (N=34)	p-value
Age (years, mean, SD)	40.9 (11.7)	41.5 (12.4)	0.85
Race/Ethnicity (n, %)			0.58
Asian	1 (2.9%)	2 (5.9%)
Black	13 (38.2%)	18 (52.9%)
Hispanic	3 (8.8%)	4 (11.8%)
Mixed	3 (8.8%)	1 (2.9%)
Native American	1 (2.9%)	0 (0.0%)
White	13 (38.2%)	9 (26.5%)
Children			0.81
No	21 (61.8%)	19 (57.6%)
Yes	13 (38.2%)	15 (44.1%)
Sexually Active			1.00
No	2 (5.9%)	1 (2.9%)
Yes	32 (94.1%)	33 (97.1%)
Sexual Dysfunction
No	33 (97.1%)	33 (97.1%)	1.00
Yes	1 (2.9%)	1 (2.9%)
IIEF Scores (median, IQR)
ED Function	29.0 (24.2–30.0)	29.0 (23.5–30.0)	0.94
Orgasmic Function	10.0 (7.0–10.0)	9.5 (6.3–10.0)	0.79
Sexual Desire	9.0 (8.0–9.0)	9.0 (8.0–9.8)	0.89
Intercourse Satisfaction	11.5 (9.0–13.0)	13.0 (8.5–14.0)	0.56
Overall Satisfaction	8.0 (6.0–10.0)	8.5 (5.3–10.0)	0.77
MSHQ (median, IQR)
Erection	15.0 (12.0–15.0)	14.0 (12.3–15.0)	0.54
ED Bother	5.0 (4.0–5.0)	5.0 (5.0–5.0)	0.19
Ejaculation	32.0 (28.0–33.0)	32.0 (30.0–33.0)	0.79
Bother	5.0 (5.0–5.0)	5.0 (5.0–5.0)	0.10
Satisfaction	25.0 (22.0–30.0)	26.0 (23.0–30.0)	0.49
Additional	28.0 (25.3–29.0)	27.5 (26.0–30.0)	0.91

SD: standard deviation; IQR: interquartile range

DNA integrity

When comparing the number of subjects with abnormal TUNEL scores (TUNEL positive >25%), there were no differences between duloxetine and placebo groups at baseline, on treatment or off treatment (0 vs. 0, p = 1.00; 5 vs. 1, p = 0.09; and 2 vs. 1, p = 0.56; respectively) (Table 2). Similarly, there were no differences in TUNEL scores between the placebo and treatment groups at baseline, on treatment or off treatment (Week 0: 6.8% (IQR 3.7–9.3) vs. 6.2% (IQR 4.5–9.0), p = 0.97; Week 2: 6.1 (IQR 3.4–11.7) vs. 7.7% (IQR 3.8–9.5), p = 0.60; Week 6: 6.4% (IQR 2.7–10.6) vs. 6.2% (IQR 4.2–14.7), p = 0.40; Week 8: 5.0% (IQR 3.8–11.3) vs. 5.3% (IQR 3.0–8.4), p=0.40; Week 10: 5.2% (IQR 3.7–12.7) vs. 9.5% (IQR 3.2–9.7), p = 0.19) (Table 2). No differences were seen when each time point was compared to baseline (Table S1).

Table 2 –

Counts of Significantly Elevated TUNEL (>25%) and Mean TUNEL (%) at All Time Points

	Placebo	Duloxetine	p-value
TUNEL > 25% (n)
Baseline	0/34	0/34	1.00
On Treatment (Week 2 and 6 combined))	5/68	1/68	0.09
Week 2	2/34	1/34
Week 6	3/34	0/34
Off Treatment (Week 8 and 10 combined)	2/68	1/68	0.56
Week 8	0/34	1/34
Week 10	2/34	0/34
TUNEL % (median, IQR)
Week 0	6.8 (3.7–9.3)	6.2 (4.5–9.0)	0.97
Week 2	6.1 (3.4–11.7)	7.7 (3.8–9.5)	0.60
Week 6	6.4 (2.7–10.6)	6.2 (4.2–14.7)	0.40
Week 8	5.0 (3.8–11.3)	5.3 (3.0–8.4)	0.40
Week 10	5.2 (3.7–12.7)	9.5 (3.2–9.7)	0.19

IQR: interquartile range

Hormone effects

There were no clinically significant differences in the serum concentrations of T in both groups at any time points of the study (Figure 2). A more substantial transient decrease of T from baseline was noted during treatment at week 2 (−2.2 ng/dL for placebo versus −28.1 ng/dL for duloxetine, p = 0.02) (Table S1). Serum E was higher for those on treatment at week 8 only (57.8 ± 30.0 pg/ml for duloxetine versus 43.9 ± 24.3 pg/ml for placebo, p=0.04). Serum LH, FSH, and PRL were similar between groups across weeks (Table S2). When comparing serum hormones to baseline (Table S1), serum E was higher during duloxetine treatment at 8 weeks (−9.47 pg/ml for placebo versus 4.60 pg/ml for duloxetine, p=0.05). No other differences were observed for T, E, LH, FSH or PRL at any other time points.

Figure 2 –

Impact of Duloxetine on Serum Testosterone, Sperm Concentration and Sperm Motility at All Time Points For All Collected Data

Semen effects

There were no differences in semen parameters (sperm concentration and sperm motility) between the groups across each time point (Figure 2).

IIEF and MSHQ Questionnaires

There were no differences in either IIEF or MSHQ scores compared to baseline levels at all times points of duloxetine treatment (Table S3 and Table S4), except for greater ejaculatory bother on the IIEF for participants on treatment at Week 2 (p=0.03), MSHQ bother scores at week 2 (p<0.01) and MSHQ erection scores at week 10 (p=0.04).

DISCUSSION

Previous studies examining the potential adverse impact of anti-depressants on male infertility have focused on SSRIs(Beeder & Samplaski, 2020). We present the first known published study of SNRI effects on parameters of human male fertility potential, specifically examining the effect of duloxetine on sperm DNA integrity, semen parameters, and serum hormone levels. Our results demonstrate no clinically relevant differences during duloxetine treatment, compared to placebo at all observed evaluation times for sperm DNA integrity, semen quality, and most serum hormone levels (E, LH, FSH, PRL). Although statistically significant differences in serum T levels compared to baseline were observed at week 2 as well as E levels at week 8, these variations did not appear to be clinically important.

Studies evaluating SSRIs have demonstrated a negative impact of this class of medications on male fertility. Sperm DNA integrity changes are hypothesized to occur secondary to serotonin effects resulting in altered post-testicular sperm transport and subsequently, DNA damage(Tanrikut et al., 2010). Previous studies report negative impacts to semen parameters and have hypothesized that these effects may be a consequence of altered central hormone levels such as prolactin(Safarinejad, 2008). However, current evidence indicates that changes in serum prolactin levels are not the underlying cause for the adverse effects of SSRIs on male fertility(Nargund, 2015; Tanrikut et al., 2010). In our study, we observed no change in central hormone (LH, FSH, and PRL) production. We did, however, observe a greater statistically significant decrease in serum T for men taking duloxetine compared to those taking placebo, at week 2 (during full dose duloxetine) in our study. However, this decrease in serum T levels between the groups, as well as within groups compared to baseline levels, are likely of limited clinical significance. The lack of a persistent change in testosterone at 6 weeks (still during treatment) further suggests that duloxetine does not have a substantial adverse effect on testosterone production.

No differences were observed in sperm DNA quality, as indicated by mean TUNEL scores, between treatment and placebo groups. The proportion of men with abnormal DNA fragmentation (as well as mean DNA fragmentation levels) were not significantly different on duloxetine versus placebo.

Anti-depressant SSRIs have been used off-label for the treatment for ejaculatory dysfunction, namely premature ejaculation(Waldinger, Hengeveld, Zwinderman, & Olivier, 1998). They have also been suggested to cause erectile dysfunction(Higgins et al., 2010). Therefore it was not unexpected that there could be some differences in ejaculatory bother and erection scores for some men on treatment compared to placebo after initiation of medication, as noted on week 2 (during) treatment (Table S3 and Table S4).

The only prior published study that we can identify examining SNRI effects on male fertility potential was performed in mice. In this study, mice received venlafaxine with or without vitamin C and were compared to controls(Bandegi et al., 2018). Mice treated with SNRI had better sperm morphology, non-progressive motility and viability, but no difference in sperm concentration(Bandegi et al., 2018). Interestingly, the authors reported these differences in contrast to studies with SSRIs, citing that differences in medication effect may have been due to drug dosage and the animal model itself(Bandegi et al., 2018). This rodent model is limited in its application to humans, since sperm quality is assessed in the epididymis, rather than evaluating ejaculated sperm quality. The difference in evaluation is important, as SSRIs may exert their effect of sperm transport after leaving the epididymis. Overall, these author’s findings, are in line with the results of our study that demonstrated no significant negative impact of the studied SNRI on sperm quality.

SSRI and SNRI medications are effective for the treatment of various psychological and pain-related disorders. There are some reports of minor differences in efficacy for treatment of depression and anxiety, however, chronic pain symptoms may be better treated with use of SNRIs(Machado & Einarson, 2010; Stahl, Grady, Moret, & Briley, 2005). Additionally both SSRIs and SNRIs are relatively well-tolerated, although some adverse effects associated with SNRIs have been attributed to alterations in norepinephrine levels in subjects using SNRIs(Santarsieri & Schwartz, 2015). Similarly, the difference in effect of SSRI compared to SNRI on clinical parameters of male infertility may be due to an opposing mechanistic effect of norepinephrine, but this remains to be elucidated. Thus, whether someone is prescribed SSRI or SNRI for treatment of their psychologic or pain disorder may be equivocal, however if the patient is a young man interested in preserving fertility or in the process of actively trying to conceive a child, consideration of SNRI rather than SSRI use may be beneficial. Any subtle effects of hormonal changes associated with anti-depressants on male fertility should be further investigated. In addition, other SSRIs beyond paroxetine should be studied to determine whether the adverse effect on sperm quality and fertility potential is a medication class effect.

The main limitation of our study is that a subset of subjects was lost to follow-up at different timepoints, limiting the power of analysis at each subsequent time point studied. We did, however, utilize multiple imputation methods in an attempt to overcome these missing data. We did not have a biomarker or method for monitoring adherence to taking the pills as instructed, thus, we cannot confirm participant adherence to placebo or medication administration. Furthermore, the men in our study were administered duloxetine for a short time period, which may be less generalizable to those taking anti-depressant medications for long periods of time. However, we hypothesized that a six week duloxetine exposure to maturing spermatozoa in the testis and epididymis should be adequate to affect DNA fragmentation, should such effects occur. Of note, shorter duration of SSRI use had a substantive effect on sperm DNA fragmentation. Despite these limitations, our study has several strengths. First, our study design was randomized with double-blinding and controlled with a placebo group, and had multiple time points for outcome evaluation. Additionally, to our knowledge, this is the first published human study examining the effects of duloxetine, an increasingly prescribed SNRI anti-depressant medication, on male fertility.

CONCLUSION

This is the first known human study to evaluate potential adverse effects of an SNRI (duloxetine) on male fertility. We found no clinically relevant negative impacts on semen parameters or sperm DNA integrity. For men taking SSRIs who are interested in fertility and have abnormal sperm DNA integrity, change of medication to an SNRI may be considered to maintain fertility potential. In addition, only subtle changes in serum testosterone were observed with duloxetine in this study, unlike the profound changes in serum T seen previously on paroxetine.

Supplementary Material

tS1-S4

Table S1 – Median TUNEL and Hormone Comparisons At Each Week versus Baseline

Table S2 – Mean Hormone Comparisons for FSH, LH, Estrogen and Prolactin

Table S3 – Median IIEF Scores at Baseline, Weeks 6 and 10

Table S4 – Median MSHQ Scores at Baseline, Weeks, 2, 6, 8 and 10

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available on clinicaltrials.gov (NCT03038867) at https://clinicaltrials.gov/ct2/show/NCT03038867.