It’s no secret that SSRIs and SNRIs cause sexual dysfunction (SD) in a large proportion of patients who take them. The exact size of that proportion, though, is difficult to pinpoint with any reliability. Most of the large clinical trials of antidepressants ascertain side effects through spontaneous reports, and research subjects are understandably embarrassed about coming out and complaining about sexual problems. This methodology led to early estimates of SD incidence in the range of 2%–16%, numbers considered preposterously low to the modern clinical eye. By contrast, in a prospective study of 1,022 outpatients, all of whom were asked specifically about sexual functioning, the authors estimated that SSRIs and venlafaxine caused rates of SD ranging from 58% to 73% (Montejo AL, J Clin Psych 2001;62 Suppl 3:10–21).
There are now several specific SD scales used in clinical trials, such as the ASEX (Arizona Sexual Experiences Scale) and the CSFQ (Changes in Sexual Functioning Questionnaire). If you plan on reading any of these studies, be forewarned that these scales are somewhat confusing. For the ASEX, higher scores mean more SD, or worse sex, while for the CSFQ, higher scores mean less SD, or better sex. To keep the findings clear, I came up with the following mnemonic: ASEX sounds like “asexual,” and more asexuality implies less sex, so a high ASEX is a “bad” sex score; for the CSFQ, the first word is “change” and psychiatrists usually see change as good, so a high CSPQ is a “good” sex score.
Beyond such quirks in research methods, there is also the issue of pharmaceutical industry bias. Almost all studies of antidepressant-induced SD are funded by companies that stand to gain market share if their drug is shown to be more sexually benign than those of their competitors. Therefore, we have to scrutinize the studies’ methods to ensure they were not manipulated to achieve a victory for the featured product.
I’ll take a look at the most recent SD data on the newer antidepressants in a second. But first, a recent metaanalysis, funded by the U.S. Agency for Healthcare Research and Quality, did a good job reviewing the literature on SD associated with the 13 older “second-generation antidepressants,” which included all six SSRIs (fluoxetine, sertraline, paroxetine, citalopram, escitalopram, fluvoxamine), three SNRIs (duloxetine, venlafaxine, and desvenlafaxine), and a smattering of others (bupropion, mirtazapine, nefazodone, and trazodone). The researchers identified 63 studies, many of them randomized controlled trials directly comparing two or more active drugs. They also included five large observational studies, all of which had the advantage of using SD rating scales to assess side effects (Carvalho J et al, J Sex Med 2011;8(9):2501–2511).
The results, by and large, were in line with our clinical experience. Bupropion produced less SD than SSRIs, paroxetine was the most sexually deadening of all SSRIs, and escitalopram was worse than mirtazapine and nefazodone (which isn’t surprising) but also somewhat worse than fluoxetine (which is surprising, given escitalopram’s reputation as a “clean” SSRI) (Reichenpfader U et al, Drug Saf 2014;37(1):19–31. doi: 10.1007/s40264–013–0129–4).
Let’s move on to the antidepressants approved since 2010, namely levomilnacipran (Fetzima), vilazodone (Viibryd), and vortioxetine (Trintellix).
When vilazodone was first approved by the FDA in 2011, there were rumors about a lower potential for sexual side effects, partly because its mechanism of action combines SSRI qualities with 5-HT1A partial agonism, and 5-HT1A stimulation is thought to enhance sexual functioning. The company provided evidence in the form of clinical trials using the ASEX to measure sexual functioning in patients assigned to drug or placebo. Unfortunately, they did not separate out patients who began the study with normal sexual functioning—meaning that they did not report the important treatment- emergent sexual dysfunction (TESD).
In that study, the average ASEX score ranged from 16.5 in men to 21.2 in women before being randomized to either vilazodone or placebo. ASEX scores in the range of 19–20 are typically used as thresholds for defining SD. After 8 weeks of treatment, patients on vilazodone saw an improvement of 0.4 points, while patients on placebo improved by 1.3 points (Rickels K et al, J Clin Psychiatry 2009;70(3):326–333). Because vilazodone did not worsen preexisting SD, one might be tempted to conclude that the drug is clean of sexual side effects. But that’s not necessarily true, because giving antidepressants to patients with preexisting SD tends to be an inaccurate way of assessing sexual side effects. A good example of this is a study of paroxetine, which has the most severe sexual side effects of any SSRI. In this study, patients started with ASEX scores of 20 and showed improved sexual functioning on Paxil after 27 weeks (Baldwin DS et al, Int Clin Psychopharmacol 2006;21(3):159–169). The fact that paroxetine could be found to improve sexual functioning casts doubt on the validity of the vilazodone data using similar methods.
Recently, another clinical trial of vilazodone was published; this one appropriately separated out data on patients with normal baseline sexual functioning. Patients with major depression were randomly assigned to vilazodone 20 mg (n = 288), vilazodone 40 mg (n = 284), citalopram 40 mg (n = 280), or placebo (n = 281) (Clayton A et al, Int Clin Psychopharmacol 2015;30:216– 223). About 40%–45% of patients came into the study with normal sexual functioning, according to the CSFQ. The table below shows what percentage of men and women developed TESD after 8 weeks of treatment.
The authors did not do significance testing because the sample size wasn’t large enough to make these statistics meaningful. Nonetheless, the raw numbers are informative, and hint that for women, vilazodone and citalopram may be equally likely to cause SD, while for men, there may be a small advantage for vilazodone—though we would need a study with a larger sample size to know for sure. The bottom line is that vilazodone and citalopram (and by extension, most SSRIs) are likely comparable in SD liability.
As with vilazodone, there is a tenuous neurotransmitter case to be made that vortioxetine might be cleaner sexually than SSRIs, because among its many effects on various serotonin receptors, it is an agonist at 5-HT1A. A pooled analysis of seven placebo-controlled trials found that higher doses of vortioxetine caused progressively more TESD per the table below.
In terms of the statistics, this was a “noninferiority trial,” a diabolically confusing study design that has come into vogue over the last several years. In a standard clinical trial, a statistical test is done to see whether a new treatment is superior or equivalent to another treatment. In a non-inferiority trial, by contrast, statistical tests are run to show that a drug is not “unacceptably worse” than placebo. Vortioxetine passed this low bar for the two lowest doses—5 mg and 10 mg. We can therefore say with confidence that low-dose vortioxetine is not significantly inferior to placebo with regard to sexual side effects.
But what about those higher vortioxetine doses of 15 mg or 20 mg, doses that are more likely to be needed for a robust antidepressant effect? These doses cause higher rates of SD than placebo numerically (42.9% and 46.1% vs. 32% for placebo)—but these differences were not statistically significant. The 48.2% rate of SD caused by duloxetine, on the other hand, was significantly higher than placebo, and was also higher than the lowish rates of SD caused by 5 mg and 10 mg of vortioxetine.
According to the authors, interpretations are limited by the fact that these data were taken from different studies, none of which were designed specifically for making statistical comparisons of sexual side effects. Nor were any statistical adjustments made for multiple comparisons, which increases the probability that any differences, or lack thereof, may have been due to chance. Nonetheless, given that the rates of SD went up steadily with increasingly higher vortioxetine doses, it’s reasonable to conclude that vortioxetine does, in fact, cause SD to some degree—though possibly a somewhat smaller degree than duloxetine.
In another study, researchers tested whether switching patients who had SSRI-induced SD to vortioxetine would improve their sex lives more than switching them to another SSRI, in this case escitalopram (Jacobsen PL et al, J Sex Med 2015;12(10):2036–2048). This is an important clinical issue, since we often must decide which antidepressant to switch patients to when the first one is causing SD. The best answer to this quandary is generally going to be bupropion or mirtazapine, but some patients may not do as well on those agents.
A total of 447 patients were enrolled and randomly assigned to double-blind treatment with either vortioxetine or escitalopram, and dosing was identical for both drugs, starting at 10 mg/ day for one week and increasing to 20 mg/day thereafter. Sexual functioning was assessed with the CSFQ. After 8 weeks, patients assigned to vortioxetine improved by 8.8 points on the scale, significantly more than the 6.6-point improvement with escitalopram (the scale ranges from 14 to 70). In terms of the percent of patients whose sexual functioning significantly improved, vortioxetine also beat escitalopram on this outcome numerically (75% vs. 66.2%) but not quite statistically (p = 0.057); nonetheless, that’s a low enough p value for many to consider the difference real. In practical terms, the number needed to treat is 12. This means that you would have to switch 12 patients to vortioxetine to finally arrive at a patient whose SD would not have improved on escitalopram. Given vortioxetine’s brand-name expense and high rate of nausea, its slight potential sexual side effect advantage over escitalopram is probably not enough to convince us to prescribe it.
The only sexual side effect information we have about levomilnacipran is from the package insert, which lists erectile dysfunction (ED) as one of the “common” side effects of Fetzima. ED is reported as occurring in 6% of men as opposed to 1% of men on placebo. Ejaculation problems occurred in 5% of men on Fetzima and fewer than 1% of men on placebo. Fewer than 2% of female patients on Fetzima reported SD (http://www.allergan.com/assets/pdf/fetzima_pi#page=1). As you might have surmised, these numbers were based on spontaneous reports, so it is highly likely they significantly underestimate the true sexual side effect burden.
TCPR VERDICT: All three of these newer antidepressants cause significant SD. Vilazodone might be slightly less likely to cause SD in men than citalopram. Vortioxetine may have a slight advantage over duloxetine for new-onset SD, and over escitalopram for decreasing preexisting TESD. The specifics of levomilnacipran are anybody’s guess, but it probably will not have side effect advantages over venlafaxine or duloxetine, the other two SNRIs. If you are looking for a sexually neutral antidepressant, your number one choice is still bupropion.