Self-Reported Premature Ejaculation Prevalence and Characteristics in Korean Young Males: Community-Based Data From an Internet Survey


  • Hwancheol Son,

    1. Department of Urology, Seoul National University Boramae Hospital, Seoul National University School of Medicine, Sindaebang 2-dong, Dongjak-gu, Seoul, Republic of Korea
    Search for more papers by this author
  • Sang Hoon Song,

    Corresponding author
    1. Department of Urology, Jeungpyung Health Center, Naeseong-ri, Jeungpyeong-eup, Jeungpyung-gun, Chungbuk, Republic of Korea
      Jeungpyung Health Center, Jeungpyung-gun, Republic of Korea 368-904 (e-mail:
    Search for more papers by this author
  • Soo Woong Kim,

    1. Department of Urology, Seoul National University School of Medicine, Yeongeoun-dong, Jongno-gu, Seoul, Republic of Korea
    Search for more papers by this author
  • Jae-Seung Paick

    1. Department of Urology, Seoul National University School of Medicine, Yeongeoun-dong, Jongno-gu, Seoul, Republic of Korea
    Search for more papers by this author

Jeungpyung Health Center, Jeungpyung-gun, Republic of Korea 368-904 (e-mail:


ABSTRACT: Premature ejaculation (PE) is suspected to be the most prevalent male sexual complaint, and the prevalence of PE is considerably high also in the younger generation. We investigated the PE prevalence based on the Diagnostic and Statistical Manual of Mental Disorders (4th ed text revision; DSM-IV-TR) definition and the risk factors of PE in Korean young men via Internet survey. Subjects (n = 3980) aged from 20 to 59, who performed sexual intercourse more than once a month during the past 6 months were asked to participate in this study. Participants were asked to complete a questionnaire that consisted of questions on general, medical, and sexual history related to ejaculation. A total of 600 subjects were included in this study. PE prevalence was found to be 18.3%. Prevalences were not significantly different across age groups, after excluding subjects with erectile dysfunction (ED). Educational level, marital status and duration, average income, sexual orientation, smoking, alcohol consumption, and circumcision status showed no difference in the PE and non-PE groups. Partners perceived satisfaction rates were 45.0% in the PE group and 63.9% in the non-PE group. Significant differences were found between the PE and non-PE groups in terms of ED, obesity, and depression prevalence. However, multiple logistic regression analysis revealed that the significant risk factors of PE were age and the frequency of conversations with partners about sexual intercourse. This Internet-based study is limited because participants probably represent a selected population of Internet users with non-representative educational and socioeconomic profiles. This study is the first to report the prevalence of both self-reported PE and PE on the basis of the DSM-IV-TR definition in the Korean population. This study demonstrates that PE in Korea is as prevalent as it is in European countries and the United States.

Premature ejaculation (PE) is believed to be the most common male sexual complaint and to affect about 20%–30% of the male population across all age groups (Laumann et al, 1999, 2005; Rowland et al, 2004; Porst et al, 2006). Researchers have published PE prevalence data in various countries. However, although characteristics of PE might differ culturally (Waldinger et al, 2005), fewer researches have reported its prevalence in Asian populations. Also, few large-scale community-based studies have been conducted in Asia; thus, normative data are lacking (Moreira et al, 2006; Ahn et al, 2007). Furthermore, the epidemiological studies performed surveyed only middle aged and elderly subjects (Moreira et al, 2006) or focused on patterns of sexual activity, help-seeking behavior, and erectile dysfunction (ED) rather than on the prevalence of PE and the characteristics of those with the condition (Ahn et al, 2007).

No universally accepted definition of PE had been established until quite recently; thus, clinical research studies employed a variety of definitions, physiological measurements, and psychometric instruments to evaluate PE (Montague et al, 2004). Recently, the International Society for Sexual Medicine (ISSM) offered an evidence-based definition of lifelong PE (McMahon et al, 2008, p 1602). The committee proposed that lifelong PE be defined as “a male sexual dysfunction characterized by ejaculation that always or nearly always occurs prior to or within about 1 minute of vaginal penetration and, the inability to delay ejaculation on all or nearly all vaginal penetrations, and negative personal consequences, such as, distress, bother, frustration, and/or the avoidance of sexual intimacy.” This proposed evidence-based definition was used because no universally accepted definition has been established.

The most commonly quoted definition is that of the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (4th ed text revision; DSM-IV-TR; Shabsigh and Rowland, 2007). DSM-IV-TR defines PE as the “persistent or recurrent onset of orgasm and ejaculation with minimal sexual stimulation before, on, or shortly after penetration and before the person wishes it, taking into account factors that affect the duration of the excitement phase, such as, age, novelty of the sexual partner or situation, and recent frequency of sexual activity” that “cause(s) marked distress or interpersonal difficulty and PE not exclusively due to the direct effects of a substance, for example, alcohol, opioids, and other drugs.” This definition, however, does not provide any ejaculation latency cutoff points and is criticized of its authority-based nature, many false-positive diagnoses, acceptance of unnecessary pathologization, and erroneous prescriptions of ejaculation-delaying drugs (Waldinger and Schweitzer, 2007). Still, despite the proposal of the ISSM ad hoc committee, considerable debate exists regarding normative intravaginal ejaculatory latency time (IELT) and its application to the definition of PE (Montague et al, 2004). Furthermore, a number of clinicians regard the DSM-IV-TR definition as valid and regard revision and the development of a new definition as unnecessary (Hellstrom, 2007; Shabsigh and Rowland, 2007). Therefore, we accepted that no standardized definition can be applied universally and adopted the most commonly used definition (DSM-IV-TR). As demonstrated by Porst et al. (2006), this definition is usefully operationalized with a couple of questions even in an Internet-based survey for classifying PE.

In the present study, we aimed to determine the community-based prevalence of PE and to identify the characteristics of PE in Korea with the DSM-IV-TR definition. We focused on younger males because PE is more prevalent in the young, which contrasts with ED (Laumann et al, 1999; Mercer et al, 2003; Rowland et al, 2007). In addition, the risk factors of PE and relationships between PE and other comorbid conditions were evaluated.

Materials and Methods

In June 2006, subjects were recruited from a population of more than 20 000 men enrolled with INR research (a population-based Internet research company based in Seoul). The research panel represents a national population-based sample of men who agreed to participate in Internet surveys; each subject was paid less than 1 dollar in cyber money for participating in this survey. Initially, e-mail was sent to 3980 men aged over 20 years, requesting that they visit the web portal and complete the study questionnaire. On the website, they were asked to participate in this survey if they were or had been involved in stable, sexually active relationship and had participated in sexual intercourse more than once per month during the past 6 months. We excluded those who did not meet this criterion because of the possibility of bias from the recollection of remote experiences that had happened more than a half year ago. The web page was designed such that a subject answered all questions on a page then moved to the next page. Only participants that completed all questions were included in the study. To exclude untrustworthy answers, we excluded answers with a response time of less than 20% of the average response time. This study protocol was reviewed and approved by the Institutional Review Board (IRB) at our hospital.

The survey contained 96 questions in Korean (not shown): questions 1–28 addressed general medical and sexual history and demographic information; questions 29–49 identified erectile function and included questions on the International Index of Erectile Function (IIEF) questionnaire; questions 50–76 assessed ejaculation and PE-related attitudes and behaviors; and questions 77–96 evaluated symptoms related to voiding and defecation along with International Prostate Symptom Score (IPSS). The questionnaire used in this study was revised 8 times on the basis of the findings of a pilot survey performed on 45 male volunteers recruited from among medical personnel and patients that visited our urology clinic to clarify questions related to comprehension.

The definition of PE was encompassed in 2 questions that reflected the DSM-IV-TR definition of PE. The first question was, “How often do you experience the onset of orgasm and ejaculation with minimal sexual stimulation before, on, or shortly after penetration and before you wish it when you have sexual intercourse with your partner?” The allowed responses were never, sometimes, usually, always. This question captured the DSM-IVTR criterion of ejaculating before the individual wishes. The second question was, “To what degree do you feel distressed with your ejaculatory function?” The allowed answers were none, a little, somewhat, very much. This question reflected distress or interpersonal difficulty. Men were classified as having PE if they answered “usually” or “always” to the first question and “somewhat” or “very much” to the second, and if they were not influenced by substances like alcohol or drugs, known to affect sexual function. Self-reported PE prevalence was also evaluated using the question, “Which category would you place yourself in?” The responses provided were normal, premature ejaculation, delayed ejaculation. Estimated IELT was obtained using the questionnaire but was not applied to the definition of PE. PE definition in this study did not classify lifelong or acquired PE.

Group demographic characteristics and medical and sexual histories were compared. Based on Asian criteria (Choo, 2002), body mass index (BMI) was classified as: normal, less than 23 kg/m2; overweight, 23–25 kg/m2; and obese, more than 25 kg/m2. IIEF EF domain scores were classified and analyzed as described Tokatli et al (2006); that is, a score of 25 or less was classified as ED, and a score greater than 26 was normal.

All statistical analyses were performed using SPSS version 13.0 (SPSS Inc, Chicago, Illinois). The chi-square test was used to compare categorical data, and the independent t test and 1-way analysis of variance were used to compare numerical data. A multiple logistic regression analysis model was used to identify independent risk factors of PE. All hypotheses were 2-sided and P < .05 was considered significant. Values are presented as x̄ ± SD.


Characteristics of the Subjects

During the 2-week study period, 2518 individuals opened the requesting e-mail and 1739 visited the web portal. Of these 883 were ineligible and 256 did not complete the questionnaire or were excluded because they exceeded the response time limit. Finally, 600 subjects were included, a response rate of 23.8%. Mean age was 35.5 ± 8.4 years, and the age distribution was as follows: 162 men were in their 20s, 267 in their 30s, 126 in their 40s, and 45 in their 50s. The average frequency of sexual intercourse was 5.4 ± 4.2 times per month. This frequency tended to decrease with age (ie, 5.8 ± 4.2 in the 20s, 5.6 ± 4.6 in the 30s, 4.6 ± 3.4 in the 40s, 4.8 ± 3.7 in the 50s). Of participants, 95.3% identified themselves as heterosexual, 2.8% as bisexual, and 1.8% as homosexual. Table 1 lists demographic characteristics, such as educational level, marital status, body mass index, smoking status, and alcohol consumption.

Table 1. . Subject demographics of PE and non-PE groups
 All MenPENon-PE
  1. Abbreviation: PE, premature ejaculation group defined by Diagnostic and Statistical Manual of Mental Disorders (4th ed, text revision).

n, %600 (100)110 (100)490 (100)
Age, yr (x̄ ± SD)35.5 ± 8.434.6 ± 7.335.7 ± 8.6
Age group, %   
Education level, %   
    High school11.810.012.2
Marital status, %   
    Separated or divorced0.30.90.2
Sexual orientation, %   
    Bisexual or homosexual4.75.54.5
Body mass index, %   
    Below 2337.540.036.9
    25 and above33.034.532.7
Smoking history, %   
    Stopped smoking10.810.910.8
    Currently smoking48.845.549.6
Drinking, %   

PE Prevalence

According to the definition of PE used in this study, 110 (18.3%) men were allocated to the PE group. PE prevalence by age, shown in Figure 1, was highest (22.8%) in the 30s and then decreased with age. However, after excluding those with ED, PE prevalence was not significantly different across the age groups. Educational level, marital status and duration, average income, sexual orientation, smoking, alcohol consumption, and circumcision status were not significantly different in the PE and non-PE groups (Table 1).

Figure 1.

. Premature ejaculation (PE) prevalence by age group.

One hundred and six PE subjects (96.4%) had not sought help from a health professional for PE. The other 4, who had sought medical help, had not experienced any disease other than PE. This low rate of help-seeking behavior is consistent with that found in a previous study among middle aged and elderly men in Korea (Moreira et al, 2006).

We evaluated partner satisfaction with the question, “How often are your partners satisfied with sexual intercourse?” Partners' satisfaction rates were 45.0% in the PE group and 63.9% in the non-PE group. However, men in the PE group tried to discuss sex with partners more than men in the non-PE group, and men in the PE group tended to regard themselves as being more sexually active than their partners (Table 2).

Table 2. . Partner relationship in PE and non-PE groups
  Conversation with Partner About Sexual IntercourseComparative Activeness in Sexual Relationship
  1. Abbreviation: PE, premature ejaculation group defined by Diagnostic and Statistical Manual of Mental Disorders (4th ed, text revision).

 % within row88.211.8100.063.625.510.9100.0
 % within column20.210.918.321.513.119.418.3
 % within row78.421.6100.052.037.810.2100.0
 % within column79.889.181.778.586.980.681.7
 % within row80.219.8100.054.235.510.4100.0
 % within column100.0100.0100.0100.0100.0100.0100.0

Average estimated IELT was 10.4 ± 8.4 minutes (median, 10 minutes; range, 0.5–50.5 minutes) for all subjects, 5.6 ± 3.4 minutes (median, 5 minutes; range, 1.0–20.0 minutes) in the PE group, and 11.3 ± 8.0 minutes (median, 10 minutes; range, 0.5–50.5 minutes) in the non-PE group. Figure 2 shows estimated IELTs in the PE and non-PE groups. Of men in the PE group, 77.3% reported an IELT of less than 5 minutes.

Figure 2.

. Distribution of estimated intravaginal ejaculatory latency time (IELT) reported for premature ejaculation (PE) and non-PE groups.

We also surveyed self-reported PE prevalence. One hundred and seventeen subjects (19.5%) felt that they belonged to the premature ejaculation group, and 8 subjects (1.3%) felt that they suffered from delayed ejaculation. This self-reported PE prevalence rate was remarkably similar to the prevalence rate obtained according to the DSM-IV-TR definition (18.3%).

Risk Factors of PE

The socioeconomic and health-related variables of men with and without PE were compared. The prevalences of comorbid conditions, such as hypertension, diabetes, renal disease, liver disease, tuberculosis, lumbago, prostate disease, and hemorrhoids, were not significantly different in the PE and non-PE groups. However, significant differences were found between the 2 groups in terms of ED, obesity, and depression. In particular, the prevalence of ED was higher in the PE group (49.6%) than in the non-PE group (39.8%), and that of obesity was also higher in the PE group. However, after excluding men with ED (250 men), obesity was not different in the 2 groups. When asked, “Do you feel depressed?” 37.6% in the PE group checked “yes,” whereas only 25.6% responded affirmatively in the non-PE group. Total IPSS scores and voiding and storage symptom scores were not found to be associated with PE. However, the prevalence of PE was greater in subjects with a lower quality of life (QoL) because of urinary symptoms. Multiple logistic regression analysis showed that only age and the frequency of discussing sex with partners were significantly associated with PE. Table 3 presents risk factors and their odds ratios.

Table 3. . Factors associated with premature ejaculation (PE)
 PE, Odds Ratio (confidence limits)a
  1. a Reference indicates the standard group for calculating odds ratio.

  2. b P < .01.

  3. c P < .05.

Age, yr 
    20–292.72 (0.82, 8.97)
    30–394.66 (1.50, 14.45)b
    40–496.49 (2.02, 20.80)b
    Current or previous smoker1.18 (0.76, 1.82)
Medical condition 
    Hypertension1.57 (0.68, 3.58)
    Diabetes0.72 (0.23, 2.18)
    Depression1.67 (0.51, 1.48)
    Prostate disease0.39 (0.02, 7.68)
    Yes1.38 (0.87, 2.17)
    Not done0.92 (0.57, 1.48)
Frequency of conversation with partner about sexual intercourse 
    Always/often2.25 (1.19, 4.26)c


This Internet-based, large-scale study is the first to report the prevalence, determined either by self-reporting or by the DSM-IV-TR definition, of PE in the Korean population. Even if those 2 definitions are commonly based on subjective measures, this result provides valuable information about the characteristics, such as the risk factors of PE, its relationship with comorbid conditions, and partner satisfaction, for Korean men that suffer from PE.

The prevalence of PE found in this study (18.3%) is similar to reported prevalences in Europe (Giuliano et al, 2008) and the United States (Porst et al, 2006). Previous studies have reported PE prevalences from 11% to 33% in the Korean population (Moreira et al, 2006; Ahn et al, 2007). However, in these studies, PE was self-reported, but even though definitions of PE varied depending on researchers' preferences, the prevalences of PE found were consistent with reports published in other countries. In the present study, almost identical prevalences were found using the DSM-IV-TR definition and self-reporting. However, the make-up of the PE group defined with the DSM-IV-TR criteria and the PE group defined by self-reporting were quite different, and the correlation between the 2 definitions was low (Pearson's R = .386, P < .001). Indeed, the percentage of subjects diagnosed to have PE in both of these groups was only 50% (Figure 3). Accordingly, the results of our analysis would probably have been different if it had been performed on the self-reported PE group.

Figure 3.

. Venn diagram demonstrates overlap of the 2 premature ejaculation (PE) definitions. Numbers in each circle are counts of subjects defined either by Diagnostic and Statistical Manual of Mental Disorders (4th ed text revision; DSM-IV-TR) or self report.

Although data on the prevalence of PE according to age is limited, there is a widespread belief that the prevalence of PE decreases with age (Althof, 2006). The present study demonstrates that PE prevalence peaks in the fourth decade and then decreases with age (Figure 1). This result appears to support the aforementioned belief that the prevalence of PE decreases with age. However, when we exclude ED subjects, the prevalence of PE remained constant across the age groups, which concurs with previous studies (Laumann et al, 1999; Porst et al, 2006; Carson et al, unpublished data). However, these studies did not follow men longitudinally to assess changes in ejaculatory function (Althof et al, 2006); thus, more study is needed to confirm the relationship between PE and age.

We documented IELT during the present study, but we did not apply IELT to define PE. IELT was first introduced by Waldinger et al (1994) in an attempt to systematize the definition of PE. IELT was defined as the time between vaginal intromission and intravaginal ejaculation (Waldinger et al, 1994). This parameter is either recorded at the time or documented later by recall. Self-reported estimates of IELT tend to be greater than those recorded with a stopwatch (Waldinger et al, 1998), and this was also suspected during the present study (the average IELT was 5.6 ± 3.4 minutes in the PE group). On the other hand, several authors have reported that estimated and stopwatch IELT values correlate reasonably well or are interchangeable when assigning PE status when estimated IELT is combined with patient-reported outcomes (Pryor et al, unpublished data; Rosen et al, 2007). In terms of its value as an outcome measure, IELT has the advantages of being relatively objective and reliable; furthermore, it is an important variable when assessing the efficacy of PE treatments (Rowland et al, 2001). However, it should be noted that the use of the IELT as a measure of ejaculatory performance has some limitations. IELT can be confounded by several variables, including the duration and content of foreplay, the sexual position, the depth, force, and frequency of penile thrusting, the period of time elapsed since the previous ejaculation, partner pelvic floor muscle tone, and the extent of vaginal lubrication (McMahon et al, 2008). Moreover, the use of IELT measures alone to define PE would not fully capture the subjective distressing aspects of the disorder, and overlaps of IELT distributions have been reported between PE and non-PE groups (Patrick et al, 2005). This overlap was well documented in the present study (Figure 2).

In a large observational study conducted in the United States, it was suggested that men with PE are as psychologically distressed and affected as men with ED (Patrick et al, 2005). They reported that men with PE reported lower levels of sexual functioning and less confidence in their relationships than non-PE controls. In an European study, a consistent result was seen in more men with PE reporting interpersonal difficulties than men without PE (11.2% compared with 0.4%; Giuliano et al, 2008). In the present study, we also found that men in the PE group reported less confidence concerning their relationships with partners, as evaluated using perceived partner satisfaction with sexual intercourse (45.0% compared with 63.9%) However, in contrast to the result of the above-mentioned study that men with PE in the United States avoided discussing sexual issues or problems with their partners compared with men without PE (62.4% compared with 43.3%; Rowland et al, 2004), we found that more men in our PE group discussed sex with their partners than men in the non-PE group (88.2% compared with 78.4%). Furthermore, they tended to regard themselves as being more sexually active than their partners (Table 2). This difference is interesting, because it probably reflects cultural differences, as reported previously (Waldinger et al, 2005). Nevertheless, more study is needed to evaluate this positive attitude shown by Korean men.

Previous studies demonstrated the high prevalence of PE among patients with medical comorbidities (eg, diabetes). PE was very common among non–insulin-dependent diabetic patients in Saudi communities (El-Sakka, 2003). However, the relationship between BMI and PE has been little studied, although some recently published evidence suggests an association between obesity and ED (Larsen et al, 2007). It is probable that in some, ED is caused by vascular diseases, which share the same risk factors as obesity. Furthermore, many patients with ED develop secondary PE, perhaps because of either the need for intense stimulation to attain and maintain an erection or the associated anxiety (Montague et al, 2004). In the present study, the prevalence of ED was high in both study groups, and after excluding men with ED (defined as an IIEF-EF domain score of <25), we found that the prevalence of obesity was not different in the 2 groups. This result concurs with the result of an earlier study, in which PE was found to be unrelated to overweight or obesity in a Danish population (ASCF principal investigators and associates, 1992).

The prevalence of ED was higher in the PE group than in the non-PE group in this study. This was consistent with the recent report that the increased severity of ED is possibly associated with the presence of PE (El-Sakka, 2008). In the aforementioned study, patients with PE and low desire were more likely to report severe than mild ED; 52.4% of the patients with severe ED had PE compared with 29.5% of the patients with mild ED. It is likely that some men with PE express their complaint as an erectile disorder since penile detumescence after ejaculation occurs rapidly, or erectile dysfunction might be superimposed on lifelong PE because of the efforts of these men to minimize sexual excitement or to general causes of erectile dysfunction (Waldinger, 2002). Some studies have shown a positive effect of phosphodiesterase type-5 inhibitor (sildenafil) in the treatment of PE with or without antidepressant agents (Seftel, 2003). On the other hand, in a prospective study, the efficacy of sildenafil was negatively affected by an increased duration and severity of ED; however, global efficacy and overall satisfaction were not attenuated by PE (El-Sakka, 2006). Such an association makes one wonder whether PE is associated with an increased neurobiological risk for erectile dysfunction, but evidence is not yet adequate (Waldinger, 2002).

Prevalence studies require substantial funding and effort to achieve satisfactory results. However, the recent Internet explosion has enabled population-based medical surveys. Undoubtedly, such Internet-based studies have limitations because they cannot adequately harvest information normally obtained during physician-patient interactions or during clinical examinations. Therefore, participants of these surveys could be misinformed in some cases. Furthermore, they introduce the probability of selection bias, because participants in this survey might represent a biased population of Internet users with educational and socioeconomic profiles removed from those of the general population. Furthermore, although the Internet is now popular among many age groups, the young are more likely to use it frequently and are more likely to register with Internet survey companies. Moreover, although some questions in the survey were validated questionnaires (eg, IIEF, IPSS), others were not, and this could cause suspicion that the questions were designed so that all of the verbal and visual components are working together to produce the desired end product. However, we tried to maintain the accessibility of the questions and to make them easily understood. We also revised the questionnaire several times on the basis of the findings of a pilot survey performed on 45 male volunteers and patients; however, the possibility remains for bias.

However, despite these pitfalls of Internet surveys, they do guarantee the anonymity of participants, which is critical in studies about personal sexual activities. On the other hand, face-to-face interviews could cause embarrassment or might be perceived beforehand to be so when sensitive personal issues are likely to questioned; thus, respondents might feel obliged to give socially acceptable answers (ASCF principal investigators and associates, 1992). Furthermore, Internet surveys allow investigators to gather substantial amounts of data quickly in a form that allows computerize analysis.

The low overall response rate (23.8%) achieved during the present study raises concerns about the possible introduction of bias; for example, the study might have appealed more to men that are comfortable addressing sexualmatters. The low response rate is possibly due to the excessive number of questions. Nevertheless, the prevalence of PE as determined by the present study is similar to those found in the Western studies, in which more than 80% of initially recruited subjects completed the study protocols (Patrick et al, 2005; Giuliano et al, 2008).


The present study demonstrates that PE in Korea is as prevalent as it is in European countries and the United States. During the course of this study, it became clear that knowledge of ejaculatory dysfunction and premature ejaculation among Korean men is inadequate. Accordingly, we suggest that extended studies using an evidence-based definition of PE and case control studies to identify the associated risk factors be undertaken among Korean men.