As with individual studies themselves, several review papers on the epidemiology of PE have shown little consensus. For example, Irfan et al. [79] present prevalence ranges based on many studies, grouping data from Asia and Europe separately yet showing little difference across these two regions. Ranges in that review generally showed greater variance due to methodological differences than to Asian vs. European origin, with each group origin showing prevalence ranges between about 3% and 47%, an unhelpful range for deriving an overall credible prevalence. In an early review (2005), Jannini and Lenzi [80] suggested an overall rate of about 21%, noting possible higher rates in Asian and Latin American men, and potentially lower rates in men from the Middle East, presumably a collateral effect of religio-cultural factors. In a more recent review (2015), Saitz and Serefoglu [81] suggested a rate of only about 6–14% for combined lifelong and acquired PE but are mute about regional differences. In contrast, at about the same time (2106), McCabe et al. [3] reported a range of 6–30% in 2016, identifying Asia and Latin America as regions where prevalence appeared to be substantially higher. In a recent revisiting of the topic, Rowland et al. [22] argued that analyses that consider only studies having specific methodological improvements (e.g., using PROs, incorporating multiple criteria, avoiding self-identification) suggest a rate of combined lifelong and acquired PE of 6–10% for men with definite PE, and further, that the use of self-identification through single item responses, reliance on patient populations, or inclusion of men with “probable” PE tends to add another 10–15% to the prevalence rate. However, as with Saitz and Serefoglu [81], this study offers no assessment regarding world regional differences. As a group, these reviews do, nevertheless, underscore the difficulty of reaching a consensus range for prevalences rates in general, considering the methodological and rigor differences across studies, including the use of varying criteria, non-standardized measures, and sampling limitations. Thus, given the above disparities, such reviews further beg the question as to whether conclusions regarding different prevalence rates in different world regions could have any validity in a context where little consensus yet exists regarding overall prevalence rates.
4.1. Methodological Concerns
Given the typically high and often problematic prevalence rates generated by single-item self-report studies (e.g., 83%: [57]; or 65%: [55]), the question arises as to whether such values should ever be included in the calculation of overall prevalence rates. The lack of standardization of various PE criteria from PROs is also a problem; for example, not only does the use of ≥9 PEDT criterion double the rate compared to a criterion of ≥11 [35,38], but other PROs place different weightings on ejaculatory control vs. bother/distress. In fact, closer inspection of various anomalous prevalence rates often reveals specific methodological issues. For example, Tang & Khoo [25] reported a high prevalence rate (40.6%), but careful examination reveals that the sample size was small, the control measures were inadequate, and the subjects were drawn from the healthcare system. Shaeer and Shaeer [57] and Hanafy et al. [23] also present anomalous prevalence indices (e.g., 49.6%, 28%), yet examination of such studies reveals the lack fully replicable PE criteria or the focus solely on a patient population.
4.2. Drawing Conclusions About World Region and Cultural Differences
Given the above issues, most conclusions about differences in world region PE rates cannot be drawn with any degree of confidence. Although several early reports and reviews had suggested anomalous rates in men from Asia, the Middle East, and Latin America [1,2,3,32,80,82], consensus on such anomalies had been lacking [4,5,40]. Furthermore, the results of the current review offer little or no compelling evidence to support significant regional differences. For example, anomalous rates initially reported or assumed in Korean, Middle Eastern, and Chinese men have generally not been confirmed in subsequent studies. In fact, only data from Latin America suggest a higher prevalence rate in that population [8,47,48], although when the PE condition is restricted by an EL threshold of 1 or 2 min, the prevalence drops below 10%, thus falling in line with other world regions. At this point, the dearth of well-implemented prevalence studies from Latin America is notable, and more data are necessary to confidently infer higher PE prevalence rates in this world region.
Lacking from the literature are adequate data from sub-Saharan Africa. One study that assessed differences in PE diagnostic criteria across world regions found that sub-Saharan men not only placed greater emphasis on ejaculatory control than men from North America/Europe, but also reported greater bother/distress as well [8]. Such differences in baseline responding for men with and without PE regarding PE diagnostic criteria not only suggest possible differences in these populations, but also stress the dire need for well-designed prevalence studies from these world regions before drawing conclusions.
The current analyses would lead us to infer that religio-cultural variations are not particularly relevant to PE prevalence rates. However, we readily admit that we did not actually compare across groupings representing different cultural and religious traditions, as no data exist to make such comparisons. Rather, we used national and world regional identities as proxies for various cultural representations, and such identities/boundaries do not always overlap with homogenous cultural or religious identities. Thus, while we can conclude with some confidence that existing data do not support the idea that PE prevalence differs across national or world regions, we are less confident in concluding that religio-cultural differences play no role in the likelihood of PE occurrence (and therefore, prevalence) [80].
4.3. Choosing an Adequate Methodology for Estimating PE Prevalence Across World Regions
The major issue with making prevalence comparisons across countries and/or world regions is the lack of a consistent and consensus methodology. Our review has revealed a number of methodological strategies that might be helpful for future studies on this topic. First, in addition to probability studies, prevalence studies should focus on community-based samples rather than patient samples or specialized groups that may contribute significant bias. To this point, Appendix A Table A1 provides a sampling of studies based on patient populations, and depicts the high and/or variable rates that are often generated by such studies [63,83,84,85,86]. These studies were not included as part of our larger analyses. Second, descriptions of methodology need to clearly specify whether the sample included men with lifelong PE or acquired PE, or both.
Our review has further noted the potentially inflated PE rates in studies using single-item and/or self-assessment—rates that, in our view, provide justifiable reason for their exclusion from the pool of prevalence studies. Studies employing professionally based definitions (e.g., DSM [53], ISSM [54], or ICD [65]), on the other hand, appear to generate more moderate and consistent PE rates (e.g., 6–20%), as do studies using multiple items or PROs that tap more than one dimension of PE. As such these methods represent a major methodological improvement over single-item self-identification. Furthermore, both methods (using professional definitions or validated PROs) typically assess both ejaculatory control—the primary defining criteria for PE [9,12,14,87]—and bother/distress [78]. Because two PROs—the PEDT [21] and CHEES [67]—offer the distinct advantage of providing suggested cut-off scores for categorizing men into definite and probable PE, they provide less opportunity for error variance than other PROs (e.g., PEP [33] or IPE [66]) that lack standardized cut-off scores for determining PE status.
In addition to assessing the constructs of “ejaculatory control” and “bother/distress”, an adequate methodology should place a restriction on EL to eliminate respondents who might otherwise fall within the normal range of latencies. Studies that have done so, even when the restriction has been as high as 5 min, have generally reported substantially lower prevalence rates, in the range of 3–12% [16,17,34,39,42,47,48,50,88,89], although exceptions do occur [43].
4.4. Limitations
Our review included English language studies in indexed international publications, and thus the literature search was limited by these search parameters. Equally important, the vast majority of prevalence studies were based on community samples; few were based on probability samples. Sample sizes ranged greatly, from slightly over 200 to over 10,000, and our analysis did not weight studies based on this parameter. Third, our critical analysis focused on broad methodological issues related to variations in sample characteristics and procedures for PE categorization. We did not examine individual studies for other methodological issues, such as ones related to biases in age, education, specific recruitment procedures (other than elimination of studies based on patient populations or special pre-selected groupings), and so on. Accordingly, we focused on consistency across (and adequacy of) methodologies in broad groups of studies, deriving conclusions based on patterns of prevalence rates rather than on specific methodologies and rates unique to each individual study. Finally, we did not directly assess prevalence rates based on religio-cultural identities, but rather utilized national and world regional groupings as proxies for various cultural variations, the former not always fully overlapping with the latter.
Source link
Stella Tamas www.mdpi.com
