Introduction

Since the 1940s, physicians have tried to treat premature ejaculation by prescribing local anaesthetics,¹ sympatholytic drugs,² dopamine antagonists³ and monoamine oxidase inhibitors.⁴ However, the compliance for these drugs was low due to disturbing, and sometimes quite serious, side effects.⁵

In 1973, Eaton⁶ published the first report on the efficacy of clomipramine, a tricyclic antidepressant with a rather low side effect profile in low dosage, to treat premature ejaculation. However, in general, sexologists were not very keen on using clomipramine for the treatment of premature ejaculation in the following two decades. A new era was heralded in the early 1990s. In 1994, Waldinger et al⁷ published the first study on the efficacy of paroxetine, a selective serotonine reuptake inhibitor (SSRI), for the treatment of premature ejaculation. Until 2003, another 35 studies on modern antidepressant treatment for the indication of premature ejaculation were published. Although most of these studies claimed efficacy of antidepressants in the clinical management of premature ejaculation, methodological shortcomings confounded many of their outcomes.

Objectives and methodology

With regard to premature ejaculation, a solid design for a psychopharmacological study warrants clearly defined clinical end points, an operational definition of premature ejaculation, an objective accurate instrument to measure ejaculation time, a representative sample of study participants, a prospective randomized double-blind placebo-controlled design, and adequate baseline measurements before any intervention.

The objective of this review is to critically discuss and compare differences in methodological approaches on previously reported clinical trials on men with premature ejaculation. The importance of reproducible quantification of the ejaculation time will be emphasized and study results will be systematically reviewed. Finally, based on these systematic clinical studies, a manual for evidence-based clinical research on drug treatment of premature ejaculation will be recommended.

Clinical end points

In sexological literature, one never succeeded in clearly defining premature ejaculation. Premature ejaculation has been studied in terms of qualitative measures of clinical outcome, for example, female partner satisfaction or dissatisfaction in the male over his voluntary control. Quantitative clinical end points are the duration of ejaculatory latency, or the number of thrusts prior to ejaculation. In 1994, Waldinger et al⁷ introduced and defined the intravaginal ejaculation latency time (IELT) as an objective measure of outcome for the ejaculation time. The IELT is defined as the time from the start of vaginal intromission to the start of intravaginal ejaculation.⁷ An advantage of the definition of the IELT is its clear starting and end point, which is particularly important for research purposes. Clearly, an ejaculation before intromission (ejaculatio ante portas) has an IELT rating of 0. Each ejaculation after vaginal intromission has a rating in seconds or minutes.

Operational definition of premature ejaculation

Regarding ejaculatory time and number of thrusts, authors used self-chosen, subjective, cutoff points of the length of time prior to ejaculation. These cutoff points varied widely, from 1 to 7 min after vaginal penetration. Subjective cutoff points have also been proposed for the number of thrusts as a criterion for premature ejaculation: ejaculation within 8–15 thrusts.

In 1998, Waldinger et al ⁸ conducted a stopwatch study in 110 consecutively enrolled men with lifelong premature ejaculation in order to obtain an empirically operationalized definition for premature ejaculation (Figure 1). In this study, the female partners of men with a complaint of premature ejaculation handled a stopwatch to measure the IELT at home during each intercourse for a period of 4 weeks. The study demonstrated that 80% of men ejaculated within 30 s, 10% ejaculated between 30 and 60 s, and 10% ejaculated between 1 and 2 min independent of age and duration of relationship.⁸ Based on these results, Waldinger et al empirically defined lifelong premature ejaculation as an IELT that takes place within 1 min after vaginal penetration in more than 90% of intercourses, independent of age and duration of relationship. It is of importance to emphasize that this study in Dutch men was the first stopwatch study conducted in a Western European Caucasian population. Other patient populations need to be investigated with the same stopwatch method for any cultural differences and associated behaviour that is linked to complaints of premature ejaculation.⁵

Figure 1.

IELT measured with a stopwatch in 110 men with lifelong premature ejaculation. In all, 90% of men ejaculate within 1 min after vaginal penetration, with 80% actually ejaculating within 30 s. (With permission of copyright: Waldinger et al. Int J Psychiatry Clin Pract 1998; 2: 287–293.)

Full figure and legend (86K)

Correlations between stopwatch method (golden standard), open questions and questionnaires

Generally, people cannot estimate time accurately. The most accurate method to measure time is the use of a stopwatch. In a large prospective nonpharmacological stopwatch study by Waldinger et al,⁸ 110 males with a complaint of rapid ejaculation and their female partners, separately, assessed man's IELT both at an open question, a questionnaire and later by using a stopwatch at home at every intercourse during 4 weeks. The study showed moderate to low correlations between open questions and a questionnaire with stopwatch assessment (r=0.56, P<0.001; r=0.59, P<0.001, respectively). These correlations were even lower when the female partners were interviewed. Finally, there was a low correlation between the intergender estimate of ejaculation time; men and their female partners differed highly in the judgement of man's IELT.⁸ These results are important to bear in mind for interpretation of drug treatment studies as it shows that the outcome values of the ejaculation time assessment is dependent on both gender (eg male, female) and method (eg open question, questionnaire, stopwatch) with which the assessment has taken place.

Evaluation of sexual psychopharmacological studies

In the next paragraph, the importance of stopwatch measurement, baseline IELT, and assessment of the IELT at each intercourse will be demonstrated by citing large differences in the results of methodologically differently performed studies.

Stopwatch and active drug response

In a double-blind study, Kim et al⁹ investigated the IELT of men with premature ejaculation by comparing 40 mg/daily fluoxetine to placebo. Waldinger et al¹⁰ also used a double-blind study to investigate the IELT by comparing 20 mg/daily fluoxetine to placebo. The first authors used a questionnaire while the latter authors used the stopwatch method. The results of these two studies were strikingly different. The questionnaire study⁹ showed a baseline IELT of 4641 s. At week 4, the placebo group had an IELT of 2.273.78 min (2.9-fold increase), while after fluoxetine the IELT did not increase significantly from placebo (2.302.08 min) (3.0-fold increase).

In the stopwatch-based study,¹⁰ the baseline IELT was 1813 s. At week 6, the placebo group had an IELT of 2925 s (1.6-fold increase), while after fluoxetine the IELT significantly increased 3.514.18 min (11.7-fold increase). In other words, in the stopwatch study 20 mg of fluoxetine exerted a three-double higher fold increase than the double dosage (eg 40 mg fluoxetine) in the questionnaire study. Moreover, the use of a stopwatch did not lead to the measurement of a clinically relevant placebo response in contrast to the questionnaire study. The comparison of both studies illustrates the scientific merit of the stopwatch in the accuracy of ejaculation time assessment at active drug treatment.

Stopwatch and low placebo response

As shortly mentioned before, the use of a stopwatch also diminishes the risk of a placebo response, as will be further illustrated by the following example. Waldinger et al¹⁰ and McMahon¹¹ reported two different placebo-controlled studies investigating the effects of 50 mg sertraline per day. In both studies, a stopwatch was used to measure IELT. The study designs were double blind¹⁰ and single blind,¹¹ respectively. In the first study, Waldinger et al ¹⁰ found an IELT of 1813 s at baseline, and at week 6 they found an IELT of 2925 s in the placebo group (1.6-fold increase), and an IELT of 1.951.45 min in the sertraline group (6.5-fold increase). In the second study of MacMahon,¹¹ similar results were reported, for example, an IELT of 18 s at base-line, and at week 4 an IELT of 30 s in the placebo group (1.6-fold increase). After sertraline the IELT raised to 3.4 min (11.3-fold increase). Regardless of differences in the study design and number of weeks, the stopwatch method showed remarkably similar (low) changes in the placebo groups.

In contrast to the use of a stopwatch, a retrospectively given subjective assessment of the ejaculation time by the patient increases the risk of a placebo response, as will be illustrated by the following example.

Biri et al¹² published a double-blind placebo-controlled study investigating the effects of 50 mg sertraline per day. His patients measured their ejaculation latency by using subjective reports. Latency to ejaculation before treatment was 43.5320.2 and 40.9312.6 s in the placebo and sertraline group, respectively. After 4 weeks, the latency to ejaculation increased to 114.493.7 s (2.6-fold increase) in the placebo group and 325.4261.7 s (7.9-fold increase) in the sertraline group. It is emphasized that in contrast to the stopwatch measurement in the studies of Waldinger et al ¹⁰ and McMahon,¹¹ the subjective reports of the ejaculation time in the study of Biri et al,¹² led to a statistically significant increase of the ejaculation time in the placebo group.

Baseline IELT as a marker for SSRI-induced fold increase of IELT and absolute IELT

It was questioned whether baseline IELT could predict the outcome of the SSRI response in men with premature ejaculation. In order to answer this question, Waldinger et al¹⁰ used paroxetine, an SSRI that is well known for its ejaculation-delaying effects. In a double-blind placebo-controlled stopwatch study, men with premature ejaculation were divided into two groups; group I with an IELT of less than 1 min and group II with an IELT between 1 and 3 min.¹⁰

After a 1-month baseline period in which the IELTs were assessed at home by stopwatch, both groups were randomized in a 20 mg/day paroxetine and placebo group. The baseline IELT in group I was 14 s. After 6 weeks, the IELT in the placebo and paroxetine group were 14 s (zero-fold increase) and 92 s (6.5-fold increase), respectively. In group II, the baseline was 83 s. After 6 weeks, the IELT in the placebo group and paroxetine group were 85 s (zero-fold increase) and 602 s (7.2-fold increase), respectively.

In summary, this stopwatch study demonstrated no placebo response after 6 weeks. Furthermore, after the daily use of 20 mg paroxetine, a similar fold increase in both groups (6.5- and 7.2-fold increase, respectively) were noted. It should be stressed however, that although the fold increase in both groups was similar, the absolute IELT found after 6 weeks was 92 s in group I, whereas it was 602 s in group II. In other words, the same dose of paroxetine elicited a much higher ejaculation time in men who had between 1 and 3 min IELTs at baseline, compared to individuals with less than 1 min at baseline.

Based on these data, it appears to be very important to include homogeneous baseline IELT groups in clinical trials on premature ejaculation. In other words, we can predict that study populations with higher IELTs at baseline will end up with longer ejaculation times on serotonergic antidepressants.

Baseline IELT as a marker for clinical relevancy of the placebo response

The baseline IELT value is not only important to predict the outcome of an active drug, but also for the placebo response. A nice example can be found by comparing two stopwatch studies of Waldinger et al ¹⁰ and Althof et al.¹³ As has been mentioned before, in the study of Waldinger et al,¹⁰ men with premature ejaculation had a baseline IELT of 1813 s, which raised to 2925 s (1.6-fold increase) after placebo. Men in the study of Althof et al ¹³ had a baseline IELT of 81 s, which raised to 116 s (1.4-fold increase) at week 4 after placebo. Thus, it appeared that placebo treatment led to about 1.4–1.6-fold IELT increase. This example illustrates an important issue for clinical research. In men with a baseline IELT of about 20 s, a 1.5-fold increase is clinically not very relevant (about 10 s delay). However, in men with a baseline IELT of 90 s, a 1.5-fold increase will be experienced to be clinically relevant (about 45 s delay). This example therefore also illustrates the relevance of an exact measurement of baseline IELTs to be able to predict the effects of placebo in clinical studies. Exact measurements are only possible with the use of a stopwatch and not with a retrospectively used questionnaire or a subjective guess.

Consequences of a less firm definition of premature ejaculation on placebo and active drug effects

From the aforementioned results of drug treatment studies, it can be predicted that a shift of the definition of premature ejaculation from an IELT of less than 1 min to less or equal to 2 min, may give rise to two phenomena.

1. In men with a baseline IELT between 1 and 2 min, the magnitude of placebo effects will be higher compared to men with an IELT of less than 1 min.
2. In a study in which men with an IELT of less than 1 min are under-represented, there is a risk that, in spite of randomization, more men with an IELT >1 min. receive drug A than drug B. This risks a false conclusion that drug A is more effective in delaying ejaculation than drug B.

Consequences of the use of a questionnaire on placebo effects

The use of a questionnaire will include an absolute risk that the baseline IELTs and follow-up IELTs during active drug treatment are not adequately assessed. The use of a questionnaire may lead to:

1. a higher placebo response and
2. a higher variability of the effects of active drugs under investigation.

Conclusion

In contrast to the prospective use of a stopwatch at each intercourse, retrospective estimation of the ejaculation time by spontaneous answer or use of a questionnaire may lead to the measurement of higher placebo and inadequate active drug treatment values. Studies using subjective estimation and questionnaire assessments of the IELT may therefore lead to a higher variability of clinical outcome measure. The absolute ejaculation time measured in clinical trials is dependent on the individual baseline IELT of each individual man and the type of drug used in the trial. Moreover, the IELT assessment is influenced by gender (male, female) and method (open question, questionnaire, stopwatch) with which the IELT assessment is performed.

In order to avoid inadequate assessments, it is recommended to include homogeneous populations in randomized clinical studies in terms of IELT values at baseline. As long as large-scale stopwatch studies have not demonstrated otherwise, rapid (premature) ejaculation should be defined as an IELT of less than 1 min, independent of the amount of distress the short IELT performs for a man or his partner. A broadening of the definition implies a risk for higher placebo responses in drug treatment trials.

Hence, sufficient evidence is provided to support the use of a methodology consisting of prospective stopwatch assessment at each intercourse at baseline and during the drug trial, the use of the IELT as an outcome measure, a randomized, double-blind placebo-controlled design and a definition of premature ejaculation as an ejaculation that occurs within 1 min after vaginal penetration in more than 90% of intercourses. Finally, it should be emphasized that evidence-based data on the impact of methodology on treatment outcome can only be provided by a meta-analysis of the methodology of all drug treatment studies on premature ejaculation.

References

1. Damrau F. Premature ejaculation: use of ethylaminobenzoate to prolong coitus. J Urol 1963; 89: 936–939. | PubMed |
2. Shilon M, Paz GF, Hommonai ZT. The use of phenoxybenzamine treatment in premature ejaculation. Fertil Steril 1984; 42: 651–659.
3. Mellgren A. Treatment of ejaculatio praecox with thioridazine. Psychother Psychosom 1967; 15: 454–460.
4. Bennett D. Treatment of ejaculatio praecox with monoamine oxidase inhibitors [letter]. Lancet 1961; 2: 1309.
5. Waldinger MD. The neurobiological approach to premature ejaculation [review article]. J Urol 2002; 168: 2359–2367. | Article | PubMed | ISI |
6. Eaton H. Clomipramine in the treatment of premature ejaculation. J Int Med Res 1973; 1: 432–434.
7. Waldinger MD, Hengeveld MW, Zwinderman AH. Paroxetine treatment of premature ejaculation: a double-blind, randomized, placebo-controlled study. Am J Psychiatry 1994; 151: 1377–1379. | PubMed | ISI | ChemPort |
8. Waldinger MD, Hengeveld MW, Zwinderman AH, Olivier B. An empirical operationalization study of DSM-IV diagnostic criteria for premature ejaculation. Int J Psychiatry Clin Pract 1998; 2: 287–293.
9. Kim SC, Seo KK. Efficacy and safety of fluoxetine, sertraline and clomipramine in patients with premature ejaculation: a double-blind, placebo controlled study. J Urol 1998; 159: 425–427. | PubMed | ISI | ChemPort |
10. Waldinger MD, Hengeveld MW, Zwinderman AH, Olivier B. Effect of SSRI antidepressants on ejaculation: a double-blind, randomized, placebo-controlled study with fluoxetine, fluvoxamine, paroxetine and sertraline. J Clin Psychopharmacol 1998; 18: 274–281. | Article | PubMed | ISI | ChemPort |
11. McMahon CG. Treatment of premature ejaculation with sertraline hydrochloride: a single-blind placebo controlled crossover study. J Urol 1998; 159: 1935–1938. | Article | PubMed | ISI | ChemPort |
12. Biri H et al. Sertraline in the treatment of premature ejaculation: a double-blind placebo controlled study. Int Urol Nephrol 1998; 30: 611–615. | PubMed |
13. Althof SE et al. A double-blind crossover trial of clomipramine for rapid ejaculation in 15 couples. J Clin Psychiatry 1995; 56: 402–407. | PubMed | ISI | ChemPort |