HAM-D, Hamilton Rating Scale for Depression; SRI, serotonin reuptake inhibitor.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Nurnberg HG, Hensley PL, Gelenberg AJ, Fava M, Lauriello J, Paine S. Treatment of Antidepressant-Associated Sexual Dysfunction With Sildenafil: A Randomized Controlled Trial. JAMA. 2003;289(1):56–64. doi:10.1001/jama.289.1.56
Sexual dysfunction is a common adverse effect of antidepressants that frequently results in treatment noncompliance.
To assess the efficacy of sildenafil citrate in men with sexual dysfunction associated with the use of selective and nonselective serotonin reuptake inhibitor (SRI) antidepressants.
Design, Setting, and Patients
Prospective, parallel-group, randomized, double-blind, placebo-controlled trial conducted between November 1, 2000, and January 1, 2001, at 3 US university medical centers among 90 male outpatients (mean [SD] age, 45  years) with major depression in remission and sexual dysfunction associated with SRI antidepressant treatment.
Patients were randomly assigned to take sildenafil (n = 45) or placebo (n = 45) at a flexible dose starting at 50 mg and adjustable to 100 mg before sexual activity for 6 weeks.
Main Outcome Measures
The primary outcome measure was score on the Clinical Global Impression-Sexual Function (CGI-SF); secondary measures were scores on the International Index of Erectile Function, Arizona Sexual Experience Scale, Massachusetts General Hospital-Sexual Functioning Questionnaire, and Hamilton Rating Scale for Depression (HAM-D).
Among the 90 randomized patients, 93% (83/89) of patients treated per protocol took at least 1 dose of study drug and 85% (76/89) completed week 6 end-point assessments with last observation carried forward analyses. At a CGI-SF score of 2 or lower, 54.5% (24/44) of sildenafil compared with 4.4% (2/45) of placebo patients were much or very much improved (P<.001). Erectile function, arousal, ejaculation, orgasm, and overall satisfaction domain measures improved significantly in sildenafil compared with placebo patients. Mean depression scores remained consistent with remission (HAM-D score ≤10) in both groups for the study duration.
In our study, sildenafil effectively improved erectile function and other aspects of sexual function in men with sexual dysfunction associated with the use of SRI antidepressants. These improvements may allow patients to maintain adherence with effective antidepressant treatment.
Depression is a public health problem that adversely affects patients, their partners and families, the workplace, and general health care. In the United States, depressive disorders cause substantial disability, with an annual economic burden estimated at $65 billion.1 Approximately 18 million Americans are diagnosed with major depressive disorder (MDD) annually; 12 million to 18 million are treated with antidepressants, primarily for depression and anxiety disorders.2
Even when MDD is recognized and treated, up to 70% of primary care patients taking antidepressants are poorly compliant, giving adverse effects as the principal reason of treatment discontinuation during both short-term and long-term antidepressant treatment.3,4 A significant proportion (60%-75%) of these adverse event–related treatment discontinuations are treatment emergent sexual dysfunction, weight gain, and sleep disturbance. Discontinuation may in turn lead to increased relapse, recurrence, morbidity, or mortality rates associated with depression.2,5,6 The importance of the continuation of treatment is reflected by increased relapse rates (25%-70%) of MDD during 2 to 12 months following premature discontinuation of antidepressants.7
Selective serotonin reuptake inhibitors (SSRIs) replaced tricyclic antidepressants as first-line agents by demonstrating comparable efficacy, simpler titration, improved tolerability, and greater safety in an event of overdose.1,2 As SSRIs became widely used, a growing recognition of sexual adverse effects attributed to increased serotonin tempered the initial enthusiasm. Recognizing that further improvements in antidepressant treatment remained, pharmacological refinements were introduced that, in addition to blocking serotonin reuptake, block serotonin receptor function (modulators), provide dual norepinephrine-serotonin reuptake inhibition, or are without serotonin reuptake inhibition.
However, most antidepressants, with or without serotonin reuptake inhibition, can adversely affect sexual function.8-10 Selective and nonselective serotonin reuptake inhibitor (SRI) agents represent almost 90% of prescribed antidepressants, with 25 million individuals, and 95 million prescriptions per year.1 With reports that antidepressant-associated sexual dysfunction (AASD) occurs in 30% to 70% of patients taking SRIs10 and that almost 90% of patients developing AASD prematurely discontinue use, 7 to 16 million patients may potentially develop sexual adverse effects, leading to nonadherence and treatment failure.2
The sexual response consists of 4 stages: desire (libido), arousal (subjective or physiologic), orgasm, and resolution. Associations between depression, medication, and sexual dysfunction are prevalent and multidirectional.11 The Massachusetts Male Aging Study demonstrated increasing rates of erectile dysfunction (ED) in men older than 40 years.12 Among men with ED, 55% to 95% have symptoms of depression and 50% to 90% of men with depression have ED.5,13-16 Effective treatment of MDD not only can improve depression-associated ED and sexual dysfunction but also can cause AASD to emerge.9,17-20 Generally AASD occurs early in treatment and can improve, exacerbate, or persist, but rarely remits spontaneously.6,20 Decreased libido, ED, and delayed orgasm are common complaints.21,22 Numerous strategies have been proposed for treating AASD, but only limited efficacy has been reported in placebo-controlled trials.6,22-25
Sildenafil citrate is a selective and competitive inhibitor of phosphodiesterase type 5, the predominant catabolic isozyme for cyclic guanosine monophosphate in the corpus cavernosum.26,27 Inhibition of cyclic guanosine monophosphate catabolism enhances the relaxation of cavernosal smooth muscle to enable a response to sexual stimulation.28,29 Randomized clinical trials demonstrated sildenafil efficacy as an oral treatment in men with ED of diverse etiologies,30 diabetes,31,32 treated or untreated hypertension,33-35 multiple sclerosis,36 spinal cord injury,37-39 postprostatectomy,40,41 stable angina,42 and antiandrogen chemotherapy.43 In men with ED and depressive symptoms, sildenafil-treated patients had significant improvement in ED.44 In a retrospective analysis from 10 phase 2/3 double-blind, placebo-controlled trials, sildenafil improved ED in men taking concomitant SSRIs.23 These reports and open trials suggest further investigation of treatment with selective phosphodiesterase type 5 inhibitor for SRI-AASD.21,45
This prospective, randomized, double-blind, placebo-controlled investigation specifically examined sildenafil treatment for patients with MDD in remission, who were taking selective and nonselective serotonin reuptake inhibitors, and who were experiencing AASD. The following specific objectives were addressed: (1) to compare the efficacy of sildenafil with placebo in patients with remitted MDD experiencing SRI-AASD; (2) to determine whether effects of sildenafil treatment are limited to erectile function or extend to other domains of sexual function (ie, desire, orgasm, satisfaction); (3) to examine whether sildenafil treatment of SRI-AASD is associated with changes in depression; and (4) to compare adverse events (AEs) occurring with sildenafil and placebo treatment.
This was a prospective, parallel-group, randomized, double-blind, placebo-controlled study. A flexible dose of sildenafil was implemented. It was a 6-week trial conducted from November 1, 2000, to January 1, 2001, simultaneously at 3 US outpatient medical centers: University of New Mexico, Albuquerque; University of Arizona, Tucson; and Massachusetts General Hospital, Boston. Each center's institutional review board approved the protocol, and all patients provided written informed consent before study enrollment.
Male patients were eligible if they (1) were between ages 18 and 55 years, (2) had a diagnosis of MDD in remission, (3) were taking an antidepressant with a selective or nonselective serotonin reuptake inhibition mechanism for at least 12 weeks, (4) were on a stable dose for at least 6 weeks, and (5) were experiencing AASD for at least 4 weeks. Patients were required to be in good health, to have regular (minimum once weekly) sexual activity for the study duration, and to have had satisfactory sexual function before the onset of depression or antidepressant treatment. Patients eligible for study inclusion were those who substantiated that any prior sexual dysfunction was limited to previous episodes of depression or antidepressant treatment that remitted on clinical improvement and discontinuation of medication. Although these criteria cannot establish an infallible etiologic connection between antidepressant treatment and sexual dysfunction, the heuristic applies to clinical practice.
Major depressive disorder in remission was diagnosed according to Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria.46 Patients had to score 10 or less on the Hamilton Rating Scale for Depression (HAM-D).47 The HAM-D consists of 17 items clinician-rated on a 5-point scale from 0 (absent) to 4 (severe). The remission threshold score of 10 or less, rather than 7 or less, was selected to adjust for AASD adverse effects inflating the score.48 Similarly, patients had to score 10 or less on the Hamilton Rating Scale for Anxiety, indicating absence of significant symptomatic anxiety.49
Serotonin reuptake inhibitor–AASD is defined by DSM-IV criteria for substance-induced sexual dysfunction,46 which includes specific items for impaired desire, arousal (ED), orgasm, and sexual pain. Patients had to have substantial impaired sexual function defined by at least 1 of the following criteria that caused significant distress: (1) ED as defined by persistent or recurrent inability to attain or maintain an adequate erection until completion of sexual activity, (2) inability to have an orgasm, or (3) ejaculatory delay of at least 10 minutes for masturbation or intercourse.
Patients were excluded for any of the following: anatomical penile deformities, primary or prior diagnosis of a sexual disorder other than AASD, spinal cord injury, uncontrolled psychiatric disorder, diabetes mellitus, proliferative retinopathy, current/recent alcohol or substance abuse, significant hematologic, renal, or hepatic abnormalities, stroke or myocardial infarction within the last 6 months, cardiac failure, unstable cardiac condition or arrhythmia, current or anticipated use of nitrates, major relationship changes, blood pressure lower than 90/50 or higher than 170/100 mm Hg, known hyperprolactinemia, retinitis pigmentosa, investigational drug use within 3 months, or current use of drugs or therapies for AASD.
Patients were enrolled between November 1, 2000, and January 1, 2001, and recruited from outpatient settings, newspaper advertisements, postings, and referrals. All patients were assessed for eligibility at screening (N = 117) (Figure 1), and all consenting patients (n = 90) completed the Sexual Health Inventory for Men, a 5-item questionnaire derived from the International Index of Erectile Function (IIEF),50 to establish their ability for self-assessment of sexual dysfunction.51 All patients received a physical examination, including blood pressure, 12-lead electrocardiogram, and standard biochemistry and hematological laboratory tests.
Using SPSS version 10 (SPSS Inc, Chicago, Ill), an unrestricted, computer-generated randomization schedule was developed and given to the independent pharmacy. The randomization resulted in 45 patients assigned to active drug and 45 to placebo. The only restriction to this randomization was that the groups be of equal size. The largest difference in numbers assigned to the 2 groups at any point in the trial was 4 (excludes completions). Additionally, there were no statistically significant differences between assigned groups at baseline in demographic characteristics (Table 1). At baseline, eligible patients were allocated the next available number that randomly assigned them to receive sildenafil (50 mg) or matching placebo, with instruction to take 1 tablet approximately 1 hour before anticipated sexual activity but not more than once daily. Medications were sealed in sequentially numbered identical containers according to allocation sequence, and all study personnel and participants were blinded to treatment assignment for the duration of the study. Patients were instructed to make at least 2 attempts at sexual activity per week. Based on the investigator's judgment of efficacy and tolerability, the dose of study drug could be adjusted from 1 to 2 tablets (Sildenafil and placebo were provided by Pfizer Inc, New York, NY).
Drug accountability, concomitant antidepressant treatment, vital signs, and self-rated and physician-rated symptom reports were assessed at each visit. Throughout the study, the investigators monitored and collected any spontaneous reports of AEs and assessed the severity of the events and their relationship to the study drug.
Efficacy was assessed using 4 validated measurements. The primary outcome measure was the Clinical Global Impression Scale adapted for Sexual Function (CGI-SF).52 The secondary outcome measures were the IIEF,50 Arizona Sexual Experience Scale (ASEX),53 and Massachusetts General Hospital-Sexual Functioning Questionnaire (MGH-SFQ).54
The CGI-SF is a clinician-rated severity scale for measuring improvement in sexual function with anchored scores from 1 (normal) to 7 (among the most extreme sexual dysfunction).52 The IIEF is a self-rated 15-item measurement for sexual function in 5 functional domains: erection, orgasm, desire, intercourse satisfaction, and overall satisfaction.50 Questions are anchored on a 6-point scale with 1 corresponding to "almost never/never" and 5 corresponding to "almost always/always." A score of 0 corresponds to "no attempt at sexual intercourse." The minimum possible total score is 5, and the maximum total is 75. For purposes of this study, ASEX (rated by patient) and MGH-SFQ (rated by clinician) are 5-item anchored scales quantifying sexual drive, arousal (subjective excitement), penile erection, ability to reach orgasm, and satisfaction. Scores for each item range from 1 to 6, and scale totals range from 5 to 30 (higher scores indicating greater sexual dysfunction).53,54
The IIEF was administered at baseline and week 6, and the ASEX, CGI-SF, and MGH-SFQ were administered at baseline and weeks 2, 4, and 6. The 17-item HAM-D was administered at baseline and weeks 2, 4, and 6 to monitor depression severity. The Hamilton Rating Scale for Anxiety was administered only at baseline. The clinical assessment of each patient and medical record was used to confirm DSM-IV–defined MDD remission and to determine the presence of any exclusionary diagnoses.
It might appear that too many measures were being taken with various sexual function rating tests. However, the standard IIEF is specifically weighted toward erectile function in 6 of the 15 questions and is not validated in populations with psychiatric disorders. The ASEX and MGH scales weight sexual dysfunction domains equally and have established use in patients with psychiatric disorders. Use of all 3 in our population provides a measure of concurrent validity between scales. With different wording, anchors, and administration, patients did not appear bored or complain of being burdened by the assessments.
Baseline demographics, safety, and tolerability evaluations were compared using descriptive statistics by χ2 and Fisher exact tests (when cell sizes were <5). Independent t tests compared patient characteristics and CGI-SF scores between the study groups at end point. χ2 analyses were used to evaluate group differences in categorical measures. Per protocol and the last observation carried forward analyses were performed on all variables and included data from all protocol-treated patients. These patients received and took at least 1 dose of study medication and had at least 1 efficacy assessment, regardless of protocol deviations or whether they completed the study. The change in sexual functioning measured by CGI-SF, IIEF, ASEX, and MGH questionnaires from baseline to each patient's own end point were the dependent measures of efficacy. A repeated measures analysis of variance was used to determine any study group (placebo-treated vs sildenafil-treated group) differences in the change from baseline to end point for the measures of efficacy and depression severity (time × group interaction). In addition, exact nonparametric methods were applied to the efficacy measures to substantiate results that rely on distributional assumptions. Findings also were confirmed with analysis of covariance and Wilcoxon rank sum tests for primary analyses. Each sexual function rating was measured for Cohen D effect size.
All statistical tests were 2-sided, and all hypotheses were evaluated at the 5% significance level. To correct for the use of multiple similar comparisons used to measure sexual dysfunction, a Bonferroni adjustment correction was applied, and only P values <.03 were considered significant for the sexual dysfunction scales. The F test of the overall hypothesis test was first conducted before multiple comparisons analyses.
Sample size calculations were based on detecting a difference in full response rates at 6 weeks, assuming a response rate of 50% for sildenafil and 15% for placebo. Thus, a sample size of 72 evaluable patients (36 per group) was expected to detect a significant difference with 90% power for a type I error rate of α = .05 between sildenafil and placebo (2-sided). Assuming 20% attrition, 90 patients were entered. The sample size determination assumed no interactions of treatment with site or antidepressant.
The primary analysis was according to assignment at randomization. In addition to determination of this narrow measure of efficacy based on all randomized patients and imputing the worst rank scores for early exclusions because of protocol violations before and without taking the study drug, there was a general efficacy analysis for all protocol-treated patients and all patients who completed the study.
Adjusted means (SDs) were determined and reported. Where applicable, 95% confidence intervals (CIs) are provided. Analyses were performed with SAS version 8.01 (SAS Institute Inc, Cary, NC).
Ninety men who consented to participate were screened and randomized to sildenafil (n = 45) or placebo (n = 45). Mean (SD) age was 44.9 (7.9) years. Mean (SD) duration of antidepressant treatment with comparable distribution of prescribed antidepressants was 27 (39.4) months. There were no statistically significant differences between treatment groups in baseline demographics (Table 1).
Systematic review of all protocol deviations in patient enrollment was undertaken before unblinding. Six patients who did not receive protocol treatment and who were not evaluable for a postrandomization assessment were withdrawn between baseline and week 2 without having taken any study drug for the following independent external events: 2 patients had antidepressant dose changes by their primary physician; 1 patient spontaneously, against advice, discontinued antidepressant treatment with improved sexual function; 1 patient reported increased depression with an HAM-D score greater than 10 following job layoff; 1 patient received a new antihypertension medication prescription (potential to affect sexual function) from his primary physician; and 1 patient took no study drug and exerted no sexual effort following his partner's serious accident. These 6 non–protocol-treated patients were included in the main analysis by imputing worst rank scores. In addition, 1 sildenafil-assigned patient taking maprotiline, which has the weakest SRI mechanism, represented a protocol violation and was excluded from the main analysis following completion of the study.
That provided 83 of 89 (93.3%) randomized patients (n = 42 for sildenafil and n = 41 for placebo) who were treated per protocol with a minimum of 1 dose of study drug taken for inclusion in last observation carried forward analysis. Seventy-six patients (85.3%; n = 41 for sildenafil and n = 35 for placebo) completed all baseline to week 6 end point assessments; data for 5 placebo patients were discontinued for lack of efficacy and were lost to follow-up (Figure 1). Two patients were discontinued prematurely because of AEs: 1 in the placebo group after presenting to the emergency department with complaints of acute panic and chest pain (recurrent panic attack was diagnosed) and 1 in the sildenafil group with concurrent borderline personality disorder for a significant suicidal gesture during a domestic dispute. Patients who did not receive protocol treatment (n = 6) were discontinued for ineligibility or those who withdrew (n = 8) were not associated with significantly different baseline demographic characteristics compared with protocol-treated patients.
The difference between assigned groups was not significant at baseline (Table 1). Patients reported the following extensive sexual dysfunction symptoms (mean precentage): 86.7% erectile dysfunction, 87.8% subjective arousal difficulties (sense of sexual excitement), 64.4% libido problems, 69.9% ejaculatory delay, 21.1% absent orgasm, and 5.5% other ejaculation and/or orgasm difficulties (ie, premature, retrograde, anejaculatory, lack of pleasure, pain). The mean (SD) number of baseline symptoms by domains was 3.6 (0.9). Patient self-reported mean (SD) number of sexual attempts was 8.5 (7.9) per month; 2.8 (3.7) (32.9%) were considered successful.
CGI-SF. End point mean CGI-SF scores significantly improved (ie, they were lower) among sildenafil-treated (2.3 [1.3], 95% CI, 1.9-2.8) compared with placebo-treated patients (3.9 [0.8], 95% CI, 3.6-4.2; P<.001; Table 2). A categorical improvement of "much/very much improved" (CGI-SF score ≤2) was reported by 54.5% of sildenafil-assigned (24/44; 95% CI, 37.9%-68.3%) vs 4.4% (2/45; 95% CI, 0.5%-15.1%) placebo-assigned patients; 57.1% (24/42; 95% CI, 41.0%-72.3%) per protocol sildenafil-treated vs 4.9% (2/41; 95% CI, 0.6%-16.5%) per protocol placebo-treated patients; and 58.5% (24/41; 95% CI, 39.9%-70.9%) sildenafil-assigned vs 5.7% (2/35; 95% CI, 0.7%-19.2%) placebo-assigned patients who completed the study. In sensitivity analyses, the number of patients excluded did not affect the outcome.
At end point (Table 2), total IIEF scores improved from baseline for patients receiving sildenafil (21.2 [19.6], 95% CI, 21.1-30.4) compared with those receiving placebo (0.9 [9.4], 95% CI, 0.0-6.9), reflecting 54% and 2.3% improvements from baseline to end point (P<.001), respectively. Patients receiving sildenafil showed significant improvements from baseline to end point on IIEF domains of erectile function (P = .004), orgasm function (P = .007), intercourse satisfaction (P<.001), and overall satisfaction (P = .02) compared with patients receiving placebo.
Mean scores for ASEX and MGH-SFQ improved significantly (P<.001 for both) between baseline and end point for sildenafil-treated compared with placebo-treated patients on all 5 ASEX and 4 of 5 MGH-SFQ items (Table 2).
Scores for HAM-D at baseline and end point were 10 or less for sildenafil-treated and placebo-treated patients, indicating persisting remission in depression (Table 2). Baseline HAM-D means (SDs) were 5.1 (3.0) for sildenafil and 4.8 (2.9) for placebo and 3.5 (2.9) and 5.5 (3.9) at end point, respectively. Although the differences achieve statistical significance (P = .001), this can be a halo effect of improved sexual dysfunction reflected in a lower HAM-D item score for sexual function.
Among the 83 study patients who received treatment according to protocol, 4 (4.8%; n = 3 for placebo and n = 1 for sildenafil) developed intermittent HAM-D scores between 10 and 15. These were self-limited, transient, symptomatic changes clearly distinguished from MDD relapse and not considered clinically meaningful to warrant intervention. No MDD recurrences occurred in any study patients.
At 6-week end point, 78.1% of sildenafil-treated (32/41) and 85.7% of placebo-treated (30/35) patients were taking 100 mg of sildenafil or equivalent placebo dose. Patients in the sildenafil group were taking the study drug 5.3 (2.2) times per 2-week interval, while patients in the placebo group were taking medication 4.5 (2.1) times (P = .19), supporting that lack of efficacy in placebo patients was not because of lack of attempts.
Sildenafil was well tolerated (Table 3). The most common AE was headache, reported by 40.5% of sildenafil-treated and 10% of placebo-treated patients (P = .002). Less frequent AEs were flushing, reported by 16.7% of sildenafil-treated and 2.4% of placebo-treated patients (P = .05), dyspepsia (7.1% vs 0.0%), nasal congestion (11.9% vs 2.4%), and transient visual disturbances (11.9% vs 4.9%). No serious AEs related to study drug were reported. The 2 AEs that resulted in study termination were not considered treatment related.
Our sildenafil study is the first prospective, well-controlled trial to demonstrate significant efficacy and effectiveness of an antidote for treating SRI-AASD. Taken before sexual activity, sildenafil enabled 55% of all sildenafil-assigned or 59% of completed protocol-treated patients to report "much/very much improved" sexual function, including erectile function, ejaculation, orgasm, and satisfaction. By contrast, 4% of all placebo-assigned or 6% of completed protocol-treated patients reported such improvements. The maintained remission of depression supports the importance of maintaining medication adherence by managing treatment-emergent adverse effects to improve treatment outcomes. In everyday practice, less than 30% of patients complete a guidelines-recommended course of antidepressant treatment, discontinuing primarily for adverse effects, particularly sexual dysfunction.55
The assay sensitivity in this study is comparable or exceeds other randomized sildenafil trials.23,44,56 For example, men with residual antidepressant-associated ED following effective depression treatment, regardless of whether or not patients continued antidepressant treatment,56 and men with ED taking concomitant SSRIs while receiving sildenafil23 showed significant improvements in erectile function, ejaculation and/or orgasm, and satisfaction compared with those receiving placebo. Although the US Food and Drug Administration's approved indication is specified for ED, open-label studies21,45 and numerous case reports57-59 are consistent with empirical data and suggest that sildenafil provides significant improvements in SRI-AASD not limited to ED (ie, desire, delayed ejaculation/orgasm, satisfaction) that might directly or indirectly improve with sildenafil treatment.
These sildenafil treatment study data for SRI-AASD can be compared with data for other treatment approaches, such as antidote, avoidance, augmentation, switching, and adaptation (ie, waiting, dose reduction, drug holiday). Only a small number of double-blind, placebo-controlled trials have investigated these approaches.22,24,60,61 Antidotes have not demonstrated significant improvement for AASD above the 30% to 40% found with placebo.62-64 However, even those trials limit conclusions because of methodological shortcomings of dose variability, underpowered sample sizes, or post hoc findings.62-64 For example, 1 placebo-controlled study65 reporting efficacy of buspirone hydrochloride for AASD used higher doses than another randomized study reporting no significant effect.63 However, in the former, the efficacy reported was posthoc to failed buspirone augmentation of MDD resistant to treatment with SSRI.65 In other approaches to AASD, mixed receptor or nonserotonergic agents (ie, bupropion hydrochloride, nefazodone hydrochloride, mirtazapine) have only selective effects as add-on treatments, being more effective as initial agents to avoid treatment-emergent sexual dysfunction.66-68 Overall, current approaches have not established evidence-based clinical data comparable with sildenafil effectiveness in this study, which leaves an excess of medication being given to patients with AASD.61
The significant efficacy and effect sizes for treating SRI-AASD in this study indicate further investigation is needed to determine whether sildenafil can become a first-line treatment for this highly prevalent clinical problem. The establishment of effective evidence-based treatment options to lower the high rates of premature discontinuation of medication because of adverse effects, like AASD, must require randomized clinical trials in those specific populations.
This study has limitations to generalizability. Cause and effect between antidepressant and sexual dysfunction is not definitive because other etiologic contributions—such as age, diabetes, cardiovascular disease, lifestyle risk factors (eg, smoking, alcohol, obesity, stress), coincident conditions, MDD trait markers, subsyndromal conditions, halo effects, or drug interactions—cannot be ruled out.
Regarding the various instrument measures, parsimony prefers a criterion standard to evaluate ED and sexual dysfunction. While IIEF had not been previously tested extensively in psychiatric patients with AASD, it was highly correlated in this study with ASEX and MGH-SFQ.50 It may seem that the assignment of integers to ordinal categories of an outcome measure is arbitrary, but moderate differences among various scoring systems seldom produce marked changes in conclusions. As in this case, it has become common practice to use analysis of variance models for such sample sizes when the phenomenon in question has an underlying continuous scale.69,70 The lower effect for sexual desire supports earlier reports that sildenafil does not directly enhance libido.30,71 Another consideration is a selection bias for patients highly motivated to recover preexisting sexual function attenuating libido effects.
There were also inadvertent errors in performing the study. The 6 randomized patients who were removed from the study before the first visit and never underwent protocol treatment because of protocol deviations, and the 1 patient receiving maprotiline, who was removed following completion of the study, might have been better identified to avoid required assignment with worst outcome in last observation carried forward data analysis that underestimates the results. Patients taking nefazodone, imipramine hydrochloride, or venlafaxine hydrochloride are appropriate because these agents have a serotonin reuptake inhibition mechanism and associated sexual dysfunction adverse effects.24,61 Significant efficacy and effect size (1.14) persists if worst-case scenarios are assigned to these confounds or whether the data are analyzed parametrically or nonparametrically.
These findings are restricted to the effects of sildenafil on the specific group of men who fulfilled protocol criteria and cannot be generalized to women or other subgroups unless future randomized controlled trials are conducted in those populations.
Corresponding Author and Reprints: H. George Nurnberg, MD, Department of Psychiatry, University of New Mexico School of Medicine, 2400 Tucker NE, Albuquerque, NM 87131-5288 (e-mail: firstname.lastname@example.org).
Financial Disclosures: Dr Nurnberg has received research support from Bristol-Myers Squibb Co, Eli Lilly & Co, Lilly-Icos, Pfizer Inc, Wyeth, Novartis, Roche Pharmaceuticals, Pharmacia, Abbott Laboratories, Johnson & Johnson, and Bayer. He is a paid consultant for Wyeth, Pfizer Inc, Eli Lilly & Co, GlaxoSmithKline, and Abbott Laboratories and on the speakers bureau for Pfizer, Wyeth, Eli Lilly & Co, GlaxoSmithKline, Abbott Laboratories, and Bayer.
Dr Hensley has received research and grant support from Eli Lilly & Co, Forest Pharmaceuticals, Novartis, Pfizer Inc, and Roche Pharmaceuticals. She is a paid consultant for Forest Pharmaceuticals and Pfizer Inc and is on the speakers bureau for Forest Pharmaceuticals, Janssen Pharmaceuticals, Eli Lilly & Co, Pfizer Inc, and Wyeth.
Dr Gelenberg has received research support from Bristol-Myers Squibb Co, Cyberonics, Hoechst Marion Roussel, Eli Lilly & Co, Pfizer Inc, and Wyeth. He owns stock in Johnson & Johnson and Pfizer Inc; is a paid consultant for AstraZeneca, Best Practice, Bristol-Myers Squibb Co, GlaxoSmithKline, Eli Lilly & Co, Novartis, Vela Pharmaceuticals, and Wyeth; and is on the speakers bureau for Bristol-Myers Squibb and Wyeth.
Dr Fava has received research support and honoraria from Abbott Laboratories, Bristol-Myers Squibb Co, Eli Lilly & Co, Forest Pharmaceuticals, GlaxoSmithKline, Janssen Pharmaceuticals, Knoll Pharmaceuticals, Litchwer Pharma GmbH, Lorex Pharmaceuticals, Lundbeck, Novartis, Organon Inc, Parke-Davis, Pharmacia, Upjohn Co, Pharmavite, Pfizer Inc, Roche Pharmaceuticals, Sanofi/Synthelabo Pharmaceuticals, GlaxoSmithKline, Solvay Pharmaceuticals, Somerset Pharmaceuticals, and Wyeth.
Author Contributions: Dr Nurnberg, as the principal author of this study and Ms Paine, as the principal statistician for the analyses, had full access to all the data in this study and take full responsibility for the integrity of the data and the accuracy of the data analyses. All coauthors also had access to the data. Study concept and design: Nurnberg, Hensley, Gelenberg, Fava, Lauriello.
Acquisition of data: Nurnberg, Hensley, Gelenberg, Fava, Lauriello.
Analysis and interpretation of data: Nurnberg, Hensley, Gelenberg, Fava, Paine.
Drafting of the manuscript: Nurnberg, Hensley, Gelenberg, Fava, Lauriello, Paine.
Critical revision of the manuscript for important intellectual content: Nurnberg, Hensley, Gelenberg, Fava, Lauriello, Paine.
Statistical expertise: Nurnberg, Paine.
Obtained funding: Nurnberg.
Administrative, technical, or material support: Nurnberg, Hensley, Gelenberg, Fava.
Study supervision: Nurnberg, Hensley, Gelenberg, Fava, Paine.
Funding/Support: This study was supported by an independent investigator-initiated grant from Pfizer Inc, New York, NY.