Testicular Torsion and Risk Factors for Orchiectomy

ARTICLE Testicular Torsion and Risk Factors for Orchiectomy

Jonathan M. Mansbach, MD; Peter Forbes, MA; Craig Peters, MD

Objective: To determine risk factors for testicular loss benign and malignant testicular tumors is 1.2 cases per due to testicular torsion. 100 000. Of the estimated 2248 males diagnosed nationally in 1998 with testicular torsion, 762 (34%) had an Design and Participants: Medical records of pa- orchiectomy. In the final multivariate model estimating tients aged 1 to 25 years with a principal diagnosis of tes- the probability of orchiectomy, only age was significant. ticular torsion were extracted from the 1998 Nation- wide Inpatient Sample. Population-based rates of testicular Conclusions: For males aged 1 to 25 years, testicular torsion and orchiectomy were determined. Logistic re- torsion is more common than testicular tumors, and in- gression was used to create a predictive model for orchi- creasing age is the sole identifiable risk factor for orchiectomy. For comparison, medical records of patients aged ectomy. We suggest that health care professionals edu- 1 to 25 years with a principal diagnosis of testicular neo- cate prepubertal male patients about testicular torsion plasm were extracted. and the necessity of seeking timely care to reduce the risk of orchiectomy and of possible subsequent reduced Results: The sample comprised 436 participants. The fertility. estimated incidence of testicular torsion for males aged 1 to 25 years in the United States is 4.5 cases per 100 000 male subjects per year, and the estimated incidence of Arch Pediatr Adolesc Med. 2005;159:1167-1171

ESTICULAR TORSION IS A nography is a helpful tool to differentiate urologic emergency. There torsion from other causes of an acutely is approximately a 4- to painful scrotum.5 8-hour window from the Further delays in diagnosis or treat- onset of torsion symptoms ment may be caused by individual pa- until surgical intervention is required to tient factors. Appendicitis is a similar acute T 1 save the affected testis. Delays in care may surgical condition in which delayed diag- necessitate orchiectomy, which has been nosis or delayed surgery may lead to mor- associated with reduced fertility.2 One bidity, specifically rupture of the appen- study3 found that 57% of patients had low dix. In this condition, differences in the sperm counts a median of 5 years after uni- individual patient factors, including in- lateral testicular loss from torsion. surance status and race, have been asso- Providing the necessary medical and ciated with the risk of rupture of the ap- surgical services for a patient with tes- pendix.6,7 The objectives of this study were ticular torsion requires 3 steps, namely, to determine by using a national data- timely presentation, rapid diagnosis, and base if there are identifiable risk factors for curative intervention. The surgical proce- testicular loss due to testicular torsion and dure performed would be orchiectomy to place the epidemiology of testicular tor- for those patients with a nonviable tes- sion in the context of the more com- ticle and septopexy for those with viable monly discussed condition of testicular testes. One potential barrier to providing neoplasm. ideal care is delayed presentation. In fact, male subjects may be hesitant to seek METHODS medical attention for conditions involv- Author Affiliations: Divisions ing their genitals, even for torsion.4 For- of Adolescent/Young Adult tunately, once a patient presents to a The 1998 Nationwide Inpatient Sample (NIS) Medicine (Dr Mansbach) and is part of the Healthcare Cost and Utilization Urology (Dr Peters) and health care professional, the diagnosis of Project, sponsored by the Agency for Health- Clinical Research Program testicular torsion can usually be made care Research and Quality.8 The NIS is a data- (Mr Forbes), Children’s from his history and physical examina- base of hospital inpatient stays, containing data Hospital Boston, Harvard tion. However, when the clinical diagno- from approximately 7 million hospital stays. Medical School, Boston, Mass. sis is uncertain, color Doppler ultraso- The 1998 NIS contains all discharge data from

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, DEC 2005 WWW.ARCHPEDIATRICS.COM 1167

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 tered sample design used by the NIS, the SAS Proc Sur- veyMeans was used to compute variances for totals and confi- 400 dence intervals for torsion, orchiectomy, and testicular neoplasm. Cancer (n = 120) Logistic regressions accounted for clustering and stratifica- Torsion (n = 436) tion using SAS Proc SurveyLogistic. 300 Patients with missing covariate levels (Ͻ4% of the total) were included in logistic regression models by assigning them to the majority covariate level. The missing data from the race vari- 200 able were analyzed 3 ways, namely, with the missing data treated as a separate covariate level, with the missing data removed from the analysis, and with the missing data assigned to the major- 100

Cumulative No. of Patients ity covariate level (white race). Race was not significantly associated with orchiectomy no matter how the missing values were assigned. As a result, in the final analyses, missing covar- 0 iates were assigned to the majority covariate level. 1510 15 20 25 The ICD-9-CM code 608.2 includes torsion of the testicle, Age, y appendix testis, and appendix epididymis. It is possible, therefore, that some patients, most likely the younger ones who had Figure 1. Cumulative number of patients with testicular torsion and with an orchiopexy, did not have torsion of the testicle but rather testicular cancer. torsion of the appendix testis or appendix epididymis. If patients with torsion of the appendix testis or appendix epididy- 984 hospitals located in 22 states, approximating a 20% sample mis had an orchiopexy, these misclassified patients would er- of US community hospitals. The overall goal for the NIS is to roneously lower the percentage of younger males requiring create a sample of hospitals that is generalizable to hospitals orchiectomies. Therefore, a sensitivity analysis was per- in the entire United States. The NIS data set includes a weight formed to determine the potential effect of misclassification on variable for each observation so that a weighted analysis can our conclusions. Under the assumption that some of the young- produce national estimates, with confidence intervals, of total est nonorchiectomy patients could be misclassified (ie, had an cases and rates of procedures of interest. The Children’s Hos- orchiopexy but not torsion of the testicle), random samples of pital Boston Institutional Review Board approved this study. 10%, 33%, and 50% of the nonorchiectomy patients in the Medical records of patients aged 1 to 25 years with the prin- youngest age category were removed, and the relationship be- cipal diagnosis of testicular torsion (International Classifica- tween orchiectomy and age category (youngest vs oldest males) tion of Diseases, Ninth Revision, Clinical Modification [ICD- was reassessed. Although the correct proportion of the nonor- 9-CM] code 608.2) in 1998 were extracted. The ICD-9-CM code chiectomy patients to be removed from the youngest age cat- 608.2 includes torsion of the testicle and torsion of the appen- egory is not known, we anticipate that much fewer than half dix testis and appendix epididymis. We compared the group of these patients, all of whom had a technically unnecessary of patients who received neither orchiectomy nor orchiopexy surgical procedure, had torsion of the appendix epididymis or (n=65) with the group who had a procedure (n=436); the mean appendix testis. ages for the groups were 11.9 vs 15.0 years (PϽ.001). We ex- cluded the 65 patients who did not have a surgical procedure, as these youth most likely had torsion of the appendix testis. RESULTS All 436 patients who had a surgical procedure were included in the analysis. Age-specific census estimates of the US male population were The final sample comprised 436 eligible participants from used to compute rates of testicular torsion and orchiectomy per 231 hospitals in 22 states. The number of torsion cases 100 000 male subjects. The July 1998 US male population aged nationally, estimated using a weighted analysis of the data, 1 to 25 years was extrapolated from census estimates for this was 2248 cases (95% confidence interval [CI], 1950- population at July 1, 2000, and July 1, 2001. For comparison, 2547 cases). Using census results, we estimated the total medical records of patients aged 1 to 25 years with a principal male population aged 1 to 25 years in 1998 to be 50.25 diagnosis of testicular neoplasm (ICD-9-CM codes 186.9 and million. Using these 2 results, the incidence of testicu- 222.0) were also extracted. lar torsion is 4.5 cases (95% CI, 3.9-5.1 cases) per 100 000 Statistical analyses were performed using SAS software ver- male subjects per year in the United States. Within the sion 9 (SAS Institute Inc, Cary, NC). In univariate analyses, t tests and Wilcoxon signed rank tests were used for continu- age range considered, testicular torsion is most com- ous variables, and Fisher exact tests were used for categorical mon in males aged 10 to 19 years, with an incidence of variables. 8.6 cases (95% CI, 7.1-10.1 cases) per 100 000 male sub- Analyses were conducted to determine the effect of differ- jects per year. For comparison, the database contained ent patient- and hospital-specific variables on the risk for or- 120 patients with benign or malignant testicular tu- chiectomy. The patient-specific variables included age (treated mors. Using a weighted analysis, the number of such cases categorically and continuously), race (white, black, or other), nationally was estimated to be 621 cases (95% CI, 300- insurance status (Medicaid, self-pay, or private insurance), and 912 cases). Using the population estimate, the esti- annual household income (low income vs other, based on the mated incidence of benign and malignant testicular tu- median income of the patient’s ZIP code of residence). Hospital- mors is 1.2 cases (95% CI, 0.7-1.8 cases) per 100 000 male specific variables included census region (4 levels) and hospital location (urban or rural, in which urban is within a census subjects per year, and the disorder is most common in statistical metropolitan area). males aged 16 to 24 years, with an incidence of 2.8 cases Logistic regression was used to model orchiectomy rates con- (95% CI, 1.4-4.2 cases) per 100 000 male subjects per year. trolling for the effects of hospital and patient covariates and to Figure 1 shows that 86% of testicular torsion cases oc- compute adjusted odds ratios. Because of the stratified and clus- cur in males older than 10 years (median age, 15 years),

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, DEC 2005 WWW.ARCHPEDIATRICS.COM 1168

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 Table 1. Comparison of Males Aged 1 to 25 Years With Table 2. Multivariate Predictors of Orchiectomy for Males Testicular Torsion Having vs Not Having an Orchiectomy* Aged 1 to 25 Years With Testicular Torsion*

Orchiectomy No Orchiectomy Odds Ratio Variable (n = 149) (n = 287) Variable (95% Confidence Interval) P Value Age, y Age† 1.08 (1.03-1.13) .003 1-9 8 (5.4) 34 (11.8) Race 10-17 89 (59.7) 179 (62.4) White Reference 18-25 52 (34.9) 74 (25.8) Black 1.22 (0.74-2.02) .43 Race Other 1.00 (0.54-1.86) .99 White 47 (31.5) 97 (33.8) Insurance status Black 38 (25.5) 68 (23.7) Medicaid 1.42 (0.84-2.39) .19 Other 34 (22.8) 72 (25.1) Self-pay 1.16 (0.66-2.01) .61 Missing 30 (20.1) 50 (17.4) Private Reference Insurance status Annual household income, $ Medicaid 34 (22.8) 58 (20.2) Ն25 000 Reference Self-pay 27 (18.1) 42 (14.6) Ͻ25 000 0.56 (0.23-1.36) .20 Private 86 (57.7) 187 (65.2) Census region Missing 2 (1.3) 0 Northeast 0.73 (0.36-1.45) .37 Annual household income, $ Midwest 1.16 (0.51-2.64) .72 Ն25 000 133 (89.3) 244 (85.0) South 1.03 (0.50-2.12) .94 Ͻ25 000 11 (7.4) 31 (10.8) West Reference Missing 5 (3.4) 12 (4.2) Hospital location Census region Urban Reference Northeast 41 (27.5) 102 (35.5) Rural 0.75 (0.39-1.41) .37 Midwest 27 (18.1) 43 (15.0) South 51 (34.2) 91 (31.7) *Odds ratios greater than 1 indicate increased likelihood of having an West 30 (20.1) 51 (17.8) orchiectomy. For categorical variables, the levels of the categories are Hospital location compared with each category’s reference level. Urban 133 (89.3) 253 (88.2) †The odds ratio indicates the increase in the odds of having an Rural 14 (9.4) 34 (11.8) orchiectomy per each 1-year increase in age. Missing 2 (1.3) 0

*Data are given as number (percentage). P = .03 for age, Fisher exact test. Other tests did not reach statistical significance. Missing value rows were not the raw data with an overlay of the modeled relation- included in univariate tests for statistical association between individual risk ship between age and the proportion of patients with tes- factors and orchiectomy. ticular torsion having an orchiectomy. To show the raw data at a finer level of detail, the figure displays more age categories than were used in the analysis. while 82% of testicular cancer cases occur in males older In the weighted analysis of the full data set, we ob- served orchiectomy rates among the youngest and old- than 15 years (median age, 20 years). Ͻ Of the 436 sample participants, 149 (34%) had an or- est males of 18% and 41%, respectively (P .001). We chiectomy. Other characteristics of the sample are given conducted a sensitivity analysis to evaluate the potential effect of misclassification on this result. With a ran- in Table 1. Using the weighted data, of the estimated 2248 males diagnosed nationally as having testicular tor- dom 10% of the males who did not undergo orchiec- sion in 1998, an estimated 762 (34%) had an orchiectomy from the youngest age category removed, the estimated orchiectomy rates for the youngest and oldest tomy. When age was grouped for descriptive purposes, Ͻ the estimated percentage undergoing orchiectomy in- age categories were 20% and 41%, respectively (P .001). With a random 33% removed, the rates were 25% and creased as age increased, namely, 19% among those aged Ͻ 1 to 9 years, 33% among those aged 10 to 17 years, and 41%, respectively (P .001). With a random 50% re- 41% among those aged 18 to 25 years. In the weighted moved, the rates were 31% and 41%, respectively (P=.02). analysis, the rates were nearly the same, namely, 18% for Therefore, the effect of age remained strong even with those aged 1 to 9 years, 33% for those aged 10 to 17 years, moderate to large misclassification of the youngest non- and 41% for those aged 18 to 25 years. When the cutoff orchiectomy subjects. for the low annual household income group was changed from $25 000 to $35 000, the results were comparable. COMMENT The final multivariate model estimating the probability of having an orchiectomy included adjustment for race, Testicular torsion is a treatable urologic emergency, with insurance status, annual household income, census re- an incidence that is 3.75 times greater than the inci- gion, and hospital location (Table 2). In the final model, dence of testicular tumors. The frequency of testicular only age was significant. For a 1-year increase in age, the torsion increases before puberty and the sole identifi- adjusted odds of having an orchiectomy increased by 1.08 able risk factor for orchiectomy due to testicular torsion (95% CI, 1.03-1.13), or an increase of 8% in the odds per is increasing age. We identified no differences in orchi- year. For each 10-year increase in age, the odds of hav- ectomy rates for other patient factors such as race, in- ing an orchiectomy doubles (1.0810=2.2). Figure 2 shows surance status, or hospital location.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, DEC 2005 WWW.ARCHPEDIATRICS.COM 1169

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 time to presentation. Because the time to presentation 0.6 was similar between the 2 groups, the group aged 21 to 34 years was presumed to have a greater degree of sper- 0.5 matic cord twisting, causing the increased orchiectomy rate. Sessions et al16 reported greater median degrees of 0.4 torsion in patients requiring orchiectomy (median, 540°) compared with patients with salvaged testicles (median 0.3 360°), but the range in both groups was identical and large (range, 180°-1080°). The national data in our study are 0.2 consistent with the findings of the more recent study15 and identify a parallel between increasing age and in- Proportion Having Orchiectomy 0.1 creased orchiectomy rates. Misdiagnosed testicular torsion resulting in loss of a 0.0 1510 15 20 25 testicle has legal implications, but the actual scope of mal- Age, y practice claims related to torsion is difficult to determine.17 Most concerning for patients, and the basis of some Figure 2. Estimated proportion of patients by age with testicular torsion malpractice claims, is the potential association between having an orchiectomy. Error bars indicate 95% confidence intervals. unilateral testicular loss due to testicular torsion and reduced fertility and sperm counts.2,3 Although the true pathogenesis of reduced fertility remains unproven, ex- For a male with testicular torsion to save his testicle, perimental data indicate that the contralateral intact tes- he must recognize the symptoms of torsion, access health tis may be harmed by antisperm antibodies18 or by in- care, and have a timely surgical procedure. Unfortu- creased germinal epithelial apoptosis.19 Other reports have nately, male subjects may be hesitant to seek medical at- suggested the presence of a congenital testicular dyspla- tention for conditions involving their genitals, even for sia as the basis for reduced fertility in torsion.20,21 The torsion.4,9 One study9 found that 85% of male respon- conclusions about the role of congenital pathologic con- dents did not think that it was necessary to seek atten- ditions, however, were based on the lack of relationship tion for testicular swelling and that 36% did not think between the histological changes in the contralateral tes- that it was necessary to seek attention for testicular swell- tis and the duration of torsion or the interval since tor- ing and pain. Despite this knowledge gap, we are not aware sion.20,21 Because this observation might indicate that re- of a formal recommendation or an adolescent health duced fertility in torsion is not preventable, preservation guideline advising health care providers to discuss the of as much testicular tissue as possible would seem to signs and symptoms of testicular torsion with their male be a priority. patients. One limitation of this study is that these data are from Because the frequency of testicular torsion begins to 1998, but there have been no important changes in the increase when males are 10 years old, more than 5 years management of patients with testicular torsion during this earlier than testicular tumors, it makes sense for health period.22,23 The database was not designed to address the care professionals to educate prepubertal male patients possible explanations for the linear association between about testicular disorders. Moreover, older adolescents increasing age and the rate of orchiectomies. For ex- have a documented knowledge deficit about testicular ample, there is no variable in the data set that permits pathologic conditions10 and have fewer physician visits adjustments for the duration of the torsion before pre- than preadolescents.11 To educate preadolescents about sentation or for the degree of torsion. However, the lack testicular disorders during health maintenance visits, of an explanation does not alter the message that males health care professionals must not only believe that it is need to seek care quickly when they have scrotal pain. an important topic but also be comfortable discussing The study is also limited by the fact that the ICD-9-CM problems involving the testicle.12 The message to pa- code 608.2 includes more diagnoses than torsion of the tients should be that scrotal pain, especially severe pain, testicle, but our sensitivity analysis suggests that the requires immediate evaluation. In fact, any male in the major finding would remain significant even if half of peripubertal age group or older with scrotal pain should the group aged 1 to 9 years did not have torsion of the be presumed to have torsion until proven otherwise. testicle. Although testicular salvage has been reported with pro- longed torsion symptoms,13 delayed presentation and age affect orchiectomy rates. In a 10-year retrospective study14 CONCLUSIONS of 30 patients, males younger than 18 years had delayed presentation and had more orchiectomies than those 18 These data suggest that testicular torsion is more com- years and older. In contrast, a more recent 9-year retro- mon than testicular neoplasm. Increasing age is a pre- spective study15 of 44 patients demonstrated that males dictor of orchiectomy due to testicular torsion but not aged 21 to 34 years had more orchiectomies than those race, insurance status, or hospital location. We suggest aged 8 to 20 years. Although these data by Cummings et that health care providers begin discussing testicular dis- al15 confirm that delays in presentation increase orchi- orders, specifically testicular torsion, with prepubertal ectomy rates, there was no statistically significant age- patients at the time of routine testicular examination. related difference between the 2 age groups in the mean Moreover, national organizations may want to consider

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, DEC 2005 WWW.ARCHPEDIATRICS.COM 1170

©2005 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 revising their male health guidelines to include earlier 7. Guagliardo MF, Teach SJ, Huang ZJ, Chamberlain JM, Joseph JG. Racial and education about testicular health. Males should under- ethnic disparities in pediatric appendicitis rupture rate. Acad Emerg Med. 2003; 10:1218-1227. stand that acute or intermittent scrotal pain needs im- 8. Healthcare Cost and Utilization Project. 1998 Nationwide Inpatient Sample. Rock- mediate medical evaluation and that delays may affect ville, Md: Agency for Healthcare Research and Quality; 1998. their fertility. 9. Nasrallah P, Nair G, Congeni J, Bennett CL, McMahon D. Testicular health aware- ness in pubertal males. J Urol. 2000;164:1115-1117. Accepted for Publication: June 6, 2005. 10. Goldenring JM, Purtell E. Knowledge of testicular cancer risk and need for self- examination in college students: a call for equal time for men in teaching of early Correspondence: Jonathan M. Mansbach, MD, Divi- cancer detection techniques. Pediatrics. 1984;74:1093-1096. sion of Adolescent/Young Adult Medicine, Children’s Hos- 11. Marcell AV, Klein JD, Fischer I, Allan MJ, Kokotailo PK. Male adolescent use of pital Boston, Harvard Medical School, 300 Longwood Ave, health care services: where are the boys? J Adolesc Health. 2002;30:35-43. Boston, MA 02115 (Jonathan.Mansbach@childrens 12. Brenner JS, Hergenroeder AC, Kozinetz CA, Kelder SH. Teaching testicular self- .harvard.edu). examination: education and practices in pediatric residents. Pediatrics [serial Funding/Support: This study was supported by a grant online]. 2003;111:e239-e244. 13. Jones DJ, Macreadie D, Morgans BT. Testicular torsion in the armed services: from the Novack Family Foundation, Boston (Dr Mans- twelve year review of 179 cases. Br J Surg. 1986;73:624-626. bach). 14. Barada JH, Weingarten JL, Cromie WJ. Testicular salvage and age-related delay Previous Presentation: This study was presented in part in the presentation of testicular torsion. J Urol. 1989;142:746-748. at the 2002 Annual Meeting of the Society for Adoles- 15. Cummings JM, Boullier JA, Sekhon D, Bose K. Adult testicular torsion. J Urol. cent Medicine; March 24, 2004; St Louis, Mo. 2002;167:2109-2110. 16. Sessions AE, Rabinowitz R, Hulbert WC, Goldstein MM, Mevorach RA. Testicu- lar torsion: direction, degree, duration and disinformation. J Urol. 2003;169: REFERENCES 663-665. 17. Matteson JR, Stock JA, Hanna MK, Arnold TV, Nagler HM. Medicolegal aspects 1. Schneck FX, Bellinger MF. Abnormalities of the testes and scrotum and their sur- of testicular torsion. Urology. 2001;57:783-787. gical management. In: Walsh P, Retik A, Vaughan E, Wein A, eds. Campbell’s 18. Williamson RC, Thomas WE. Sympathetic orchidopathia. Ann R Coll Surg Engl. Urology. 8th ed. Philadelphia, Pa: WB Saunders Co; 2002:2379. 1984;66:264-266. 2. Schutte B, Becker H, Vydra G. Exocrine and endocrine testicular function follow- 19. Hadziselimovic F, Geneto R, Emmons LR. Increased apoptosis in the contralat- ing unilateral torsion: a retrospective clinical study of 36 patients [in German]. eral testes of patients with testicular torsion as a factor for infertility. J Urol. 1998; Urologe A. 1986;25:142-146. 160:1158-1160. 3. Ferreira U, Netto Junior NR, Esteves SC, Rivero MA, Schirren C. Comparative 20. Dominguez C, Martinez Verduch M, Estornell F, Garcia F, Hernandez M, Garcia- study of the fertility potential of men with only one testis. Scand J Urol Nephrol. Ibarra F. Histological study in contralateral testis of prepubertal children follow- 1991;25:255-259. ing unilateral testicular torsion. Eur Urol. 1994;26:160-163. 4. Rampaul MS, Hosking SW. Testicular torsion: most delay occurs outside hospital. 21. Hagen P, Buchholz MM, Eigenmann J, Bandhauer K. Testicular dysplasia caus- Ann R Coll Surg Engl. 1998;80:169-172. ing disturbance of spermiogenesis in patients with unilateral torsion of the testis. 5. Baker LA, Sigman D, Mathews RI, Benson J, Docimo SG. An analysis of clinical Urol Int. 1992;49:154-157. outcomes using color Doppler testicular ultrasound for testicular torsion. Pediatrics. 22. Kass EJ, Stone KT, Cacciarelli AA, Mitchell B. Do all children with an acute scro- 2000;105:604-607. tum require exploration? J Urol. 1993;150:667-669. 6. Braveman P, Schaaf VM, Egerter S, Bennett T, Schecter W. Insurance-related 23. Herbener TE. Ultrasound in the assessment of the acute scrotum. J Clin Ultrasound. differences in the risk of ruptured appendix. N Engl J Med. 1994;331:444-449. 1996;24:405-421.

Announcement

Submissions. The Editors welcome contributions to Picture of the Month. Submissions should describe common problems presenting uncommonly, rather than total zebras. Cases should be of interest to practicing pediatricians, highlighting problems that they are likely to at least occasionally encounter in the office or hospital setting. High-quality clinical images (in either 35-mm slide or electronic format) along with parent or patient permission to use these images must accompany the submission. The entire discussion should comprise no more than 750 words. Articles and photographs accepted for publication will bear the contributor’s name. There is no charge for reproduction and printing of color illustra- tions. For details regarding electronic submission, please see: http://archpedi.ama-assn.org.

(REPRINTED) ARCH PEDIATR ADOLESC MED/ VOL 159, DEC 2005 WWW.ARCHPEDIATRICS.COM 1171