International Journal of Impotence Research (2007) 19, 303–309 & 2007 Nature Publishing Group All rights reserved 0955-9930/07 $30.00 www.nature.com/ijir

ORIGINAL ARTICLE Penile size and somatometric parameters among Iranian normal adult men

D Mehraban1, M Salehi2 and F Zayeri3

1Department of Urology, School of Medical Sciences, Tehran University of Medical Sciences, Tehran, Iran; 2Department of Biostatistics and Epidemiology, School of Public Health, Tehran University of Medical Sciences, Tehran, Iran and 3Department of Social Sciences, Faculty of Medicine, Shahed University, Tehran, Iran

This study aimed to determine the penile size and its correlation with somatometric parameters in physically normal Iranian adult men. To do this, a random sample of 1500 normal men aged between 20 and 40 years underwent tape measurements of penile dimensions in the stretched state under the same condition. The mean total penile length was 11.5871.45 cm, the mean glanular length was 3.0470.33 cm and the mean girth was 8.6671.01 cm. Multivariate regression analysis showed that penile dimensions are significantly correlated with age (P ¼ 0.018), height (Po0.001) and index finger length (Po0.001). This analysis provided no evidence for significant effect of waist/ hip ratio and weight on penile dimensions. As the penile augmentation is still in its experimental stage and its indications have not yet been clearly established, providing standardized data on penile dimensions seems to be necessary to make convenient decisions in the counseling and/or treatment of people with short penis concerns. International Journal of Impotence Research (2007) 19, 303–309. doi:10.1038/sj.ijir.3901532; published online 7 December 2006

Keywords: penile size; somatometric parameters; adult men; regression analysis; generalized estimating equations

Introduction ‘short penis’ is increasing.3 Most commonly these patients are otherwise physically normal males, Measurement of penile length is necessary in many who seek validation of their sexual function and situations. Some of these include diagnosis of fertility potential, true confirmation of genital organ micropenis (including pituitary and hypothalamic normality and adequacy. Furthermore, the defini- hormone deficiencies), anatomic malformations of tion of normal penile size is currently of consider- the genitalia, defining the situations requiring able importance in the diagnosis and treatment of penile-lengthening procedures, evaluating the re- sexual dysfunction and in the light of the claims of sults of medical or surgical intervention of the penis, penile augmentation in the lay press.4,5 and experimental studies.1 A review of previously published urologic surveys In the everyday clinical practice, urologists in different populations reveals an absence of quite often have to deal with young patients who universally acceptable parameters and standardized present with complaints of inadequate size of criteria on the proper size of the external genital their external genitalia.2 Concerns over penile size organs. This poses major difficulty in the counseling and a desire for a longer penis are common in the and/or treatment of young adult men with worries male population and the number of male patients of sexual inadequacy.2 seeking andrological consultation for the problem of In our country, Iran, standardized and proper criteria on the size of external genitalia is not available; thus, we decided to conduct this study to determine the penile length profile of the Correspondence: Dr D Mehraban, Department of Urology, Faculty of Medicine, Tehran University of Medical physically normal adult Iranian males aged between Sciences, Tehran 1235647, Iran. 20 and 40 years. In the present study, we also aimed E-mail: [email protected] to evaluate the dependence between penile dimen- Received 27 June 2006; revised 14 August 2006; accepted sions and some somatometric/anthropometric para- 29 September 2006; published online 7 December 2006 meters in this population. The obtained results Penile size among Iranian adult men D Mehraban et al 304 would aim to provide a framework for urologists in of 10 volunteers were measured by them. Then, the counseling and/or treatment of adult men with a random effect model was used to assure the the described concerns. interexaminer reliability according to the method presented by Fleiss.6 The estimated reliability coefficient was R ¼ 0.953, which indicated a high Methods interexaminer reliability among these caregivers. In addition, we used the method discussed by Fleiss6 to evaluate the intraclass reliability. To do this, a Sampling technique and study sample caregiver was randomly chosen and the penile sizes In this survey, 1500 Iranian healthy men aged of five volunteers were measured by four times by between 20 and 40 years old were randomly him. The obtained intraclass correlation coefficient recruited. The random selection was performed was R ¼ 0.991, which indicated a high intraexami- using a table of random numbers. According to ner reliability in this study. some reports, over 10 millions people live in Tehran During the study, the study clinics were visited on a and suburbs now. This population is composed of regular basis and quality of examination was checked different ethnicities of Iran distributed in its 22 again by the investigators. This was believed to be municipality districts. Thus, a large sample of 1500 crucial for the concordance of the measurements. adult men in this city may be considered as a representive and reasonable sample for the same population in Iran. Certain clinics in 22 districts of Tehran were Measurements approached and set as the examination venue for Tape measurement of penis, glans and penile interested healthy volunteers of 20–40 years of age circumference was made in centimeters with one if they met the inclusion/exclusion criteria. To select decimal point. To measure the total penile length, the study sample, a multistage sampling technique we placed the butt of a rigid tape measure tool on was used. In the first stage, one clinic was randomly the pubic skin over the dorsum of the fully starched selected in each municipality district of Tehran using penis held by the other hand and stretched once. a simple random sampling method. In the second The distance between the pubic skin and the stage, in each selected clinic at least 25 healthy men external urethral meatus was recorded. The glans were recruited. To choose the study sample in each length was measured from the corona to the external clinic, we chose at least 25 random numbers between urethral meatus level. The penile circumference or 1 and 100 (proportional to sample size in each clinic) girth was measured at the midshaft. The height and and then according to the selected numbers the study weight were measured as routine. samples were selected among the consecutive refer- The waist circumference was measured at the rals to each clinics. The study objectives were exactly umbilical level and the hips circumference was explained to each selected man and then he was measured at the level of the anterior–superior iliac recruited if he signed the study consent. As only few spines. The length of the extended index finger was numbers of these persons (less than 10 men) did not measured from the metacarpophalangeal joint to the accept to participate in our survey, the selected adult tip of the distal phalanx, excluding the nail tip. men may be considered as a good random sample for The examination room had to be quiet, comfort- the healthy adult men population. ing, and with the temperature between 20 and 251C At the time of measurement, all the selected subjects to alleviate any anxiety and to permit relaxation of appeared physically normal, with normal sexual the external genitalia. function (erection and ejaculation), with no history of infertility, cryptorchidism, hypospadias or any surgery other than circumcision, or any inflamma- Statistical analysis tory/malignant condition of the penis or the testicles. The data analysis was performed using the Statis- tical Analysis System, version 9.1 (SAS 9.1). Descriptive statistics, Pearson’s correlation coeffi- Training and reliability analysis cient, univariate and multivariate regression models In each clinic, only a trained caregiver measured the were utilized for data analysis. In the multivariate penile size. The principal investigator invited the modeling process, the generalized estimating equa- caregivers and held training sessions for them by tions (GEE) methodology was used to account for the examining volunteers in front of them and then correlation between the response data. observing them to do the measurement themselves. For reliability assurance, the measurements by the caregivers were rechecked by the investigators in Results some volunteers before the conduction of the study. To assess the interexaminer reliability, five care- As stated earlier, a random sample of 1500 Iranian givers were randomly selected and the penile sizes physically normal adult males were assessed in this

International Journal of Impotence Research Penile size among Iranian adult men D Mehraban et al 305 study. Table 1 shows the characteristics of the study correlation between age and total penile length sample. In addition, Table 2 shows the descriptive (r ¼À0.006, P ¼ 0.808). In addition, we found no statistics for the penile dimensions among this significant relation between age and glanular length sample. (r ¼ 0.033, P ¼ 0.204). Moreover, weight, height and Using the Pearson’s coefficient of correlation, index finger length showed statistically positive and we found a significant and positive relationship rather strong correlation with all penile dimensions between these penile measurements. The correla- (all the P-values were below 0.001). There was also tion between total penile length with glanular length an inversely significant association between waist/ and girth were 0.472 (Po0.001) and 0.183 hip (W/H) ratio and glanular length (r ¼À0.197, (Po0.001), respectively. In addition, glanular length Po0.001), but we found no evidence for the and girth were statistically correlated with a Pear- relationship between W/H ratio and total penile son’s correlation coefficient of 0.315 (Po0.001). length (r ¼À0.046, P ¼ 0.074) as well as W/H ratio We also determined the correlation between these and granular length (r ¼ 0.018, P ¼ 0.479). penile measurements and the described somato- We also assessed the association between the metric parameters. Table 3 shows these findings. explanatory variables (age, weight, height, W/H ratio According to the results of Table 3, age was and index finger) and the penile dimensions using significantly and positively correlated with girth univariate and multivariate regression models. (r ¼ 0.197, Po0.001), but there was no significant Table 4 shows the obtained results. The univariate regression analysis showed that index finger length is an independent and direct Table 1 Descriptive statistics for the general characteristics of predictive variable for total penile length (Po0.001), the Iranian male glanular length (Po0.001) and girth (Po0.001). Characteristics Min Max Median Mean s.d. According to these results, the subject’s height was another independent and positive predictive vari- Age (years) 20 40 29.0 29.61 5.50 able for total penile length (Po0.001), glanular Weight (kg) 54.2 115.6 78.0 78.67 8.41 length (Po0.001) and girth (P ¼ 0.023). This analysis Height (cm) 148.1 196.5 174.0 174.30 6.31 revealed that age has a positive association with Waist (cm) 70.7 123.3 89.7 88.90 7.03 girth (Po0.001). Additionally, the subject’s weight Hip (cm) 41.6 120.8 98.0 97.63 6.87 W/H ratio 0.78 1.85 0.91 0.91 0.05 showed a significant and indirect relation with total Index finger length (cm) 6.8 12.0 9.0 8.97 0.89 penile length (P ¼ 0.047). Finally, the significant estimates for W/H ratio provide a strong evidence

Table 2 Descriptive statistics for penile measurements of Iranian adults

Penile parameter Min Max Median Mean s.d. 5th percentile 95th percentile

Total penile length (cm) 7.5 19.0 11.50 11.58 1.42 9.50 14.00 Glanular length (cm) 2.0 4.5 3.00 3.04 .33 2.50 3.50 Girth (cm) 4.4 13.5 9.00 8.66 1.01 7.10 10.00

Table 3 Pearson’s correlation coefficient (r) for assessing the relationship between penile measurements and somatometric parameters

Penile measurement Somatometric parameter

Age Weight Height Inx.a W/H ratio

Total penile length r À0.006 0.179 0.307 0.231 À0.046 P 0.808 o0.001** o0.001** o0.001** 0.074

Glanular length r 0.033 0.268 0.229 0.353 0.018 P 0.204 o0.001** o0.001** o0.001** 0.479

Girth r 0.197 0.208 0.180 0.459 À0.197 P o0.001** o0.001** o0.001** o0.001** o0.001**

**Significant at a ¼ 0.01. a Index finger length.

International Journal of Impotence Research Penile size among Iranian adult men D Mehraban et al 306 Table 4 Univariate and multivariate regression analysis for penile measurements and somatometric parameters

Response variable(s) Explanatory variables

Age Weight Height Inx.a Waist/hip

Total penile length Est.b À0.001 À0.010 0.052 0.189 À0.512 s.e.c 0.006 0.005 0.007 0.041 0.756 P 0.835 0.047* o0.001** o0.001** 0.499

Glanular length Est. 0.002 0.003 0.004 0.087 0.164 s.e. 0.001 0.002 0.001 0.009 0.172 P 0.189 0.085 o0.001** o0.001** 0.340 Girth Est. 0.019 0.003 0.008 0.425 À1.535 s.e. 0.004 0.005 0.003 0.028 0.510 P o0.001** 0.477 0.023* o0.001** 0.003**

Total penile length, glanular length, girth Est. 0.007 0.001 0.019 0.234 À0.289 s.e. 0.003 0.002 0.003 0.019 0.322 P 0.018* 0.771 o0.001** o0.001** 0.370

*Significant at a ¼ 0.05. **Significant at a ¼ 0.01. a Index finger length. b Est, estimate of model parameter. c s.e., standard error of the estimate.

for the inverse effect of this predictive variable on his sexual organs. If a man perceives his penis as girth (P ¼ 0.003). inadequate, whether real or imagined, negative In Table 4, we also considered a multivariate feelings invade his interaction with his sexual regression model for assessing the relationship partner and even his professional associates.7 between penile dimensions (simultaneously) and Recently, the request for penile augmentation is somatometric parameters. According to the reported growing in different countries. However, it is results, penile dimensions (total penile length, claimed that penile augmentation is still in its glanular length and girth) were positively associated experimental stage and its indications have not yet with predictive variables age (P ¼ 0.018), height been clearly established and many complications (Po0.001) and index finger length (Po0.001). This have been reported.8 Thus, providing standardized multivariate analysis provided no evidence for the data on penile dimensions seems to be necessary to relationship between penile dimensions and expla- make convenient decisions in the counseling and/or natory variables W/H ratio (P ¼ 0.370) and weight treatment of people with short penis concerns. (P ¼ 0.771). Figure 1 provides these guidelines in Iranian 20– Finally, total penile length was considered to 40 years old males. In another study, it has been construct a nomogram using 2.5th, 50th and 97.5th suggested that only men with a stretched or erect percentiles (Figure 1). In different ages, values length of less than 7.5 cm should be considered below the 2.5th percentile and more than 97.5th candidates for penile lengthening.4 Here, it should percentile may be considered as abnormal penile be noted that although all the efforts were directed length. towards including all eligible healthy young adult men, however, some men with self-conceived small genitalia size may have shunned entering the study and this might have affected the results. In addition, Discussion there are apparent ethnicity variations among different ethnics such as Fars, Kurd, Azeri and so In recent years, the definition of normal penile size on in our country, but we found no documented has become of increasing interest in performing survey about these variations in the related scien- correct diagnostic assessments and therapeutic tific websites. However, we are going to conduct a choices in patients with concerns regarding penile similar study in different Iranian ethnic groups in dimensions.2 A man’s self-esteem can be affected by the future. his genitalia profile, mainly penile size, in as much In the present study, we only measured the penile as his self worth may be judged by the ‘adequacy’ of dimensions in the flaccid stretched state. Owing to

International Journal of Impotence Research Penile size among Iranian adult men D Mehraban et al 307 16.0 2.5th percentile Median 97.5th percentile

15.0

14.0

13.0

12.0

11.0 Total penile length (cm)

10.0

9.0

8.0 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 Age (year) Figure 1 Nomogram for total penile length in Iranian healthy adults.

Table 5 Penile length reported by various investigators (1899 to present)

Reference Sample size (n) Age range (years) Length (cm)

Flaccid Stretched Erect

Loeb11 50 17–53 9.51 NAa NA Schoenfeld and Beebe10 54 20–25 NA 13.02 NA Kinsey et al.12 2770 20–59 9.7 NA 15.5 Bondil et al.13 905 17–91 10.7 16.74 NA da Ros et al.14 150 NA NA NA 14.5 Richters et al.15 156 NA NA NA 15.99 Wessels et al.4 80 21–82 8.85 12.45 12.89 Smith et al.16 184 NA NA NA 15.71 Schneider et al.17 Young 111 18–19 8.60 NA 14.48 Older 32 40–68 9.22 NA 14.18 Spyropoulos et al.2 52 19–38 NA 12.18b NA Sengezer et al.1 200 20–22 6.80 8.98 12.73 Ponchietti et al.18 3300 17–19 9 12.5 NA Ajmani et al.19 320 17–23 8.16 NA NA Son et al.8 123 20–25 6.9 9.6 NA Awwad et al.20 271 Adult men 9.3 13.5 NA Present study 1500 20–40 NA 11.58 NA a Not available. b Flaccid–Stretched state. ethical concerns and religious beliefs in our country, flaccid and erect lengths (R2 ¼ 65.5 and 65.3%, we did not measure penile dimensions in erect state. respectively). In addition, the study of Wessells However, previous studies showed that stretched et al.4 resulted in R2 ¼ 0.793 for the predictability of penile length measurement offers a good alternative penile erect length by stretched length. to erect and flaccid penile length measure- We have compiled the results of different surveys ments.1,4,9,10 For instance, regression equations from different ethnicities published so far in Table 5. in the study of Sengezer et al.1 showed that penile As might be seen, because of differences in stretched length is a good predictor for penile measurement technique, age and the health status

International Journal of Impotence Research Penile size among Iranian adult men D Mehraban et al 308 of the study samples, one could not relate the cluded that index finger length is the best predictive apparent difference in the measurements to the variable among the described parameters. Other ethnicity factor only. However, our results are close studies suggested that finger length may correlate to the Greek and US studies.2,4 In addition, compar- with function or disorders of the male reproductive ing the results of the present study with other Asian system.22,23 This is based on the HOXA and HOXD surveys shows that penile size in the flaccid genes’ common embryological control of finger stretched state in Iranian adult men (11.58 cm) is development and differentiation of the genital between recorded sizes for the Jordanian20 (13.50) bud.24 Although the measurements of finger lengths and Korean8 (9.6 cm) normal adult men. do not have the power to predict the testicular In the previous decades, numerous urologic function of adult men, its correlation to penile studies have focused on penile measurements in dimensions have been reported.2 different populations, but few studies are available about the relationship between penile measure- ments and different explanatory variables such as somatometric parameters. As penile length and Conclusion circumference correlate with anthropometric para- meters such as weight, height, W/H ratio and index Data from 1500 young adult male provide a range finger length, we considered these parameters as that could be used as a standard by andrologists and bodily measures, that display a wide extent of other medical practitioners pending a larger nation- normal variability along the general population. A wide study. The results demonstrate that like other study on 52 Greek males aged 19–38 years revealed studies, penile and glanular lengths and girth are that the glanular and total penile lengths are somatometric characteristics and as such are related insignificantly inversely correlated to age, weight, to other anthropometric measures. BMI and W/H ratio and positively to height and correlated positively (statistically significant) with index finger length.2 In addition, a study on 3300 young Italian males aged 17–19 years showed that Acknowledgments penile measurements (length and circumference at We express our special thanks to the volunteers for the midshaft) are correlated to height, weight and their patience and the research deputy of Tehran BMI.18 In another interesting survey, the relation- University of Medical Sciences for the financial ship between shoe size and penile length was supports. studied by Shah and Christopher.21 They concluded that the supposed association of penile length and shoe size has no scientific basis. In our study, univariate regression analysis showed significant References and positive relationship between total penile length and subject’s height and index finger length 1 Sengezer M, Ozturk S, Deveci M. Accurate method for and negative relationship between total penile determining functional penile length in Turkish young men. length and weight. We did not attempt to include Ann Plast Surg 2002; 48: 381–385. 2 Spyropoulos E, Borousas D, Mavrikos S, Dellis A, Bourounis the pubic fat pad in our measurements. This may M, Athanasiadis S. Size of external genital organs and explain why there is a negative effect of weight on somatometric parameters among physically normal men total penile length. It is well known that weight gain younger than 40 years old. Urology 2002; 60: 485–489. increases the amount of the supra-pubic fat pad, 3 Mondaini N, Ponchietti R, Gontero P, Muir GH, Natali A, which can cause a relative decrease in penile length. Caldarera E et al. Penile length is normal in most men seeking penile lengthening procedures. Int J Impot Res 2002; 14: In our study, another univariate regression analysis 283–286. revealed statistically significant and positive asso- 4 Wessels H, Lue TF, McAninch JW. Penile length in the flaccid ciation between glanular length and subject’s height and erect stage: guidelines for penile augmentation. J Urol and index finger length. Moreover, similar analysis 1996; 156: 995–997. 5 Austoni E, Juarneri A, Gatti G. Penile elongation and showed that age, height and index finger length had thickening – a myth? – Is there a cosmetic or medical significant and direct effect on girth. Finally, W/H indication. Andrologia 1999; 31: 45–51. ratio was the only significant and inverse effective 6 Fleiss JL. The Design and Analysis of Clinical Experiments. factor on girth. John Wiley & Sons: New York, 1986, pp 1–32. We could not find a similar study with the 7 Alter GJ. Augmentation phalloplasty. Urol Clin North Am 1995; 22: 887–902. multivariate analysis of penile measurements and 8 Son H, Lee H, Huh JS, Kim SW, Paick JS. Studies on self- anthropometric parameters in the published arti- esteem of penile size in young Korean military men. Asian J cles. In this study, the multivariate regression Androl 2003; 5: 185–189. analysis revealed a direct effect of age, height and 9 Son H. Normal penile size and self esteem about penile size of the third decade men in Korea. Korean J Urol 1999; 40: index finger length on penile dimensions. 1037–1042. Regarding the results of both univariate and 10 Schonfeld WA, Beebe GW. Normal growth and variation in the multivariate regression analyses, it can be con- male genitalia from birth to maturity. J Urol 1942; 48: 759–777.

International Journal of Impotence Research Penile size among Iranian adult men D Mehraban et al 309 11 Loeb H. Harnro hrencapacitat und Tripperspritzen. Muench 18 Ponchietti R, Mondaini N, Bonafe M, Di Loro F, Biscioni S, Med Wochenschr 1899; 46: 1016–1019. Masieri L. Penile length and circumference: a study on 3,300 12 Kinsey AC, Pomeroy WB, Martin CE. Sexual Behavior in the young Italian males. Eur Urol 2001; 39: 183–186. Human Male. WB Saunders: Philadelphia, 1948. 19 Ajmani ML, Jain SP, Saxena SK. Anthropometric study of male 13 Bondil P, Costa P, Daures JP, Louis JF, Navratil H. external genitalia of 320 healthy Nigerian adults. Anthropol Clinical study of the longitudinal deformation of the Anz 1985; 43: 179–186. flaccid penis and of its variations with aging. Eur Urol 1992; 20 Awwad Z, Abu-Hijleh M, Basri S, Shegam N, Murshidi M, 21: 284–286. Ajlouni K. Penile measurements in normal adult Jordanians 14 da Ros C, Teloken C, Sogari P, Barcelos M, Silva F, Souto C and in patients with erectile dysfunction. Int J Impot Res 2005; et al. Caucasian penis: what is the normal size. J Urol 1994; 17: 191–195. 151: 323A. 21 Shah J, Christopher N. Can shoe size predict penile length? 15 Richters J, Gerofi J, Donovan B. Are condoms the right size(s): BJU Int 2002; 90: 586–587. a method for self measurement of the erect penis. Venereology 22 Bang AK, Carlsen E, Holm M, Petersen JH, Skakkebaek NE, 1995; 8: 77–81. Jorgensen N. A study of finger lengths, semen quality and sex 16 Smith AM, Jolley D, Hocking J, Benton K, Gerofi J. Does penis hormones in 360 young men from the general Danish size influence condom slippage and breakage? Int J STD AIDS population. Hum Reprod 2005; 20: 3109–3113. 1998; 9: 444–447. 23 Koehler N, Simmons LW, Rhodes G. How well does second-to- 17 Schneider T, Sperling H, Lummen G, Syllwasschy J, fourth-digit ratio in hands correlate with other indications Rubben H. Dose penile size in younger men cause of masculinity in males? Proc Biol Sci 2004; 271(Suppl 5): problems in condom use? A prospective measurement of 296–298. penile dimensions in 111 young and 32 older men. Urology 24 Kondo T, Zakany J, Innis J, Duboule D. Of fingers, toes and 2001; 57: 314–318. penises. Nature 1997; 390: 29.

International Journal of Impotence Research