FOCUS ON: Pediatric Imaging
Original Research
Retrospective Review of Diagnosis and Treatment in Children Presenting to the Pediatric Department With Acute Scrotum
Audio Available | 
Share
OBJECTIVE. Testicular torsion is a common acute condition in boys requiring prompt accurate management. The objective of this article was to evaluate ultrasound accuracy, findings, and clinical predictors in testicular torsion in boys presenting to the Stollery pediatric emergency department with acute scrotal pain.
METHODS. Retrospective review of surgical and emergency department ultrasound records for boys from 1 month to 17 years old presenting with acute scrotal pain from 2008 to 2011 was performed. Clinical symptoms, ultrasound and surgical findings, and diagnoses were recorded. Surgical results and follow-up were used as the reference standard.
RESULTS. Of 342 patients who presented to the emergency department with acute scrotum, 35 had testicular torsion. Of 266 ultrasound examinations performed, 29 boys had torsion confirmed by surgery. The false-positive rate for ultrasound was 2.6%, and there were no false-negative findings. Mean times from presentation at the emergency department to ultrasound and surgery were 209.4 and 309.4 minutes, respectively. Of the torsed testicles, 69% were salvageable. Sensitivity, specificity, and diagnostic accuracy of ultrasound for testicular torsion were 100%, 97.9%, and 98.1%, respectively. Sonographic heterogeneity was seen in 80% of nonviable testes at surgery and 58% of patients with viable testes (p = 0.41). Sudden-onset scrotal pain (88%), abnormal position (86%), and absent cremasteric reflex (91%) were most prevalent in torsion patients.
CONCLUSION. Color Doppler ultrasound is accurate and sensitive for diagnosis of torsion in the setting of acute scrotum. Despite heterogeneity on preoperative ultrasound, many testes were considered to be salvageable at surgery. The salvage rate of torsed testes was high.
Keywords: acute scrotum, Doppler ultrasound, pediatrics, testicular torsion
Testicular torsion is a common acute condition that affects approximately one in every 125 boys under the age of 20 years in the United States each year [1], with an estimated occur-rence of 25–35% of acute pediatric scrotal disease [2]. It is defined as the rotation of the longitudinal axis of the spermatic cord, resulting in the absence of testicular blood flow. However, the true incidence of torsion in patients presenting with acute scrotal pain (acute scrotum) is still unclear, and incidence rates in the literature have varied significantly from 9% to 72% [3–12]. Testicular torsion should be diagnosed and managed early to prevent testicular necrosis and loss. It has been suggested that if the torsion is successfully repaired within 6 hours of symptom onset, there is an 80–100% salvage rate, whereas the salvage rate drops to 20% after 12 hours [1]. Experimental studies have shown that testicular hemorrhagic infarction begins within 2 hours of onset of torsion, irreversible damage can occur after 6 hours, and complete infarction within 24 hours [13, 14].
The diagnosis of testicular torsion currently involves a thorough clinical history, physical examination, and color Doppler ultrasound. However, there have been previously documented cases of missed testicular torsions on color Doppler ultrasound and studies of the presence of testicular torsion with a corresponding negative ultrasound result [15, 16]. The use of particular signs and symptoms associated with testicular torsion has also been studied [3, 17], but there is no current standard set of clinical criteria for diagnostic purposes of torsion. Thus, the aim of our study was to review the accuracy and utility of ultrasound and clinical predictors for the pediatric acute scrotum and to review the care of children presenting to our hospital emergency department with acute scrotum with respect to accuracy of diagnosis and management as well as to determine any potential factors that may yield incorrect diagnosis or management of these children.
| Materials and Methods |   |
|---|
With University of Alberta health research ethics board approval, a retrospective study of all patients between the ages of 1 month and 17 years who presented from July 1, 2008, to July 1, 2011, to the Stollery Children's Hospital (SCH) at the University of Alberta with acute scrotum. The data collected were from three different databases (ultrasound, emergency department, and urology department) to ensure that all patients were included. The ultrasound data were taken from the PACS system with inclusion of all male children within the specified age range with a scrotal ultrasound. Patients with scrotal ultrasound with chronic symptoms were then manually excluded. A chart review of all children who presented to the emergency department and to the pediatric urologists at SCH with acute scrotal complaint was used to ensure that our dataset was inclusive. Patients who were not within the appropriate age range or did not have acute complaints were excluded. The collection of patients from the three departments constituted the patients in the study. All other hospitals within the Northern Alberta region transfer acutely ill patients with suspected testicular torsion to SCH pediatric urology. Thus, all torsion patients within the region would have been included within our database. A centralized electronic medical record system was used to ensure that all hospital visits, imaging studies, and patients follow-up were included.
For each patient, the following information was obtained: demographics, physical examination findings, duration since onset of symptoms, imaging techniques and findings from ultrasound with or without Doppler imaging, ultrasound diagnosis, specialty training of the radiologist involved, surgical intervention, surgical findings and surgical diagnosis, and follow-up results. A pediatric radiologist (blinded to surgical and clinical outcome) reviewed each scrotal ultrasound to document the technique and specific findings (e.g., heterogeneity) if they were not previously specified in the original report. Surgical results and follow-up were used as the reference standard. However, if surgical or follow-up diagnosis was not available (e.g., nontorsion patients), ultrasound was used as the standard of diagnosis. Torsion-detorsion patients were defined as patients who were described as likely torsion and detorsion on ultrasound. False-positive findings were defined as those in patients diagnosed as torsed on ultra-sound and found to be nontorsed on surgical exploration or clinical follow-up.
Results were expressed as means and SD or as median values and range. Differences between the groups were analyzed using the chi-square test and Student two-tailed t test. Values of p less than 0.05 were considered statistically significant.
| Results | |
|---|
This 3-year retrospective study included 342 children who presented to SCH with acute scrotum. Of these boys, 35 (10.2%) had testicular torsion confirmed by surgical diagnosis; seven had possible torsion-detorsion; and 302 (88%) were classified as nontorsion, consisting of diagnoses including torsion of appendix testes (n = 3), epididymoorchitis (n = 135), cellulitis, hematoceole, varicoceole, and hydroceole. Seven patients (2.0%) who presented with possible torsiondetorsion were diagnosed by ultrasound. Of the 35 patients with testicular torsion, 29 were diagnosed by ultrasound and confirmed with surgery, whereas six were sent straight to surgery without imaging. Of the seven patients who went straight to surgery without imaging, one did not have testicular torsion; the surgical diagnosis was torsion of the appendix testes. The torsed testicles were salvageable in 24 patients (69%), and none of these testicles showed atrophy in subsequent follow-up; 11 patients had nonsalvageable testicles (Fig. 1). The mean time from presentation at the emergency department to ultrasound was 209.4 minutes (n = 29) and from presentation at the emergency department to surgical exploration was 309.4 minutes (n = 28). This time was inclusive from initial presentation of the patient to the emergency department to the service described and included the waiting time for the patient to be seen by the physician. Mean time for testicular torsion patients from ultrasound to surgery was 161.9 minutes (n = 26) and from symptom onset to surgery was 1216 minutes (n = 27).
 View larger version (142K) | Fig. 1 —Photograph shows surgical findings of 12-year-old boy with testicular torsion who presented very late (more than 12 hours) after acute onset of scrotal pain. |
The average age of the boys with testicular torsion was 12.2 years. The average age of the salvageable testes group was 12.1 years, whereas it was 9.96 years in the nonsalvageable group (p = 0.11). The mean times for torsion patients from symptom onset and from the emergency department to ultrasound and surgery are shown in Table 1. The only significant time difference between the salvageable and nonsalvageable testes group was the time from presentation to emergency department to surgery. However, within the testicular torsion patients, two patients (with low clinical suspicion presentations) were kept waiting in the emergency department overnight for ultrasound the next morning. When these patients are excluded, there is no longer any significant difference between the measured times (Table 1).
A total of 266 patients (77.7%) underwent scrotal ultrasound. The sensitivity and specificity of ultrasound for testicular torsion were 100% and 97.9%, respectively. The diagnostic accuracy of ultrasound compared with surgery was 98.1% for testicular torsion. Positive predictive value was 85.3%, and negative predictive value (NPV) was 100%. Doppler ultrasound resulted in false-positive findings for five patients (1.9%) and no false-negative findings. Of these five patients, four studies were interpreted by adult radiologists and one by a pediatric radiologist. Eight (3.0%) ultrasound reports were deemed inconclusive. None of these eight patients had testicular torsion at surgical exploration or clinical follow-up. All eight inconclusive ultrasound studies were interpreted by radiologists who were not specialized in pediatric imaging (p = 0.0225).
Ultrasound results confirmed that absent blood flow in the testes is characteristic of torsion patients. Patients with testicular torsion were found to have a significantly higher incidence of decreased and heterogeneous testicular echogenicity on ultrasound (p < 0.0001 for both) (Table 2). Similarly, a significantly higher incidence of decreased flow in the epididymis and enlarged scrotal size was found in testicular torsion patients versus nontorsion patients (p < 0.0001 for both) (Table 2). There was no significant difference in ultrasound findings between the salvageable versus the nonsalvageable groups (Table 3). Heterogeneous echotexture was seen both in salvageable and nonsalvageable testes (p = 0.414). Figures 2 and 3 show examples of testicular torsion patients who had heterogeneous and homogeneous echogenicity within the testes on ultrasound.
 View larger version (158K) | Fig. 2A —Testicular torsion with nonsalvageable testicles. A, Doppler ultrasound images of four different testicular torsion patients with nonsalvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (61K) | Fig. 2B —Testicular torsion with nonsalvageable testicles. B, Doppler ultrasound images of four different testicular torsion patients with nonsalvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (57K) | Fig. 2C —Testicular torsion with nonsalvageable testicles. C, Doppler ultrasound images of four different testicular torsion patients with nonsalvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (53K) | Fig. 2D —Testicular torsion with nonsalvageable testicles. D, Doppler ultrasound images of four different testicular torsion patients with nonsalvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (37K) | Fig. 3A —Testicular torsion with salvageable testicles. A, Doppler ultrasound images of four different testicular torsion patients with salvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (148K) | Fig. 3B —Testicular torsion with salvageable testicles. B, Doppler ultrasound images of four different testicular torsion patients with salvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (121K) | Fig. 3C —Testicular torsion with salvageable testicles. C, Doppler ultrasound images of four different testicular torsion patients with salvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
 View larger version (116K) | Fig. 3D —Testicular torsion with salvageable testicles. D, Doppler ultrasound images of four different testicular torsion patients with salvageable testicles that are heterogeneous (A and B) and homogeneous (C and D). |
Clinical predictors of testicular torsion patients showed that sudden-onset scrotal pain (88%), nausea (68%), vomiting (67%), abnormal position or lie (86%), and abnormal ipsi-lateral cremasteric reflex (91%) were all significant clinical predictors of testicular torsion in patients presenting to the emergency department with acute scrotum (p < 0.0001) (Table 4).
| Discussion | |
|---|
Testicular torsion is a frequent condition and although the exact prevalence in patients presenting with acute scrotum has been estimated in the literature to vary from 9% to 72% [3–12], recent literature has shown a prevalence varying from 12% to 25% [18, 19]. The most recent study by Boettcher et al. [19] in 2012 involved 328 boys, and an incidence rate of testicular torsion of 13.8% was shown. Our study showed a 10.2% prevalence rate of testicular torsion among all children who presented to our emergency department with acute scrotum. Our incidence rate was similar and well within the range of the recently published testicular torsion prevalence rates [18, 19]. Torsion does occur in the preadolescent population as seen in our study (average age 12.2 years), and a younger age is associated with testicular loss, contrary to the previously published literature that testicular torsion is most common during adolescence [12, 20].
At our institution, ultrasound is commonly used in the workup of acute scrotal pain. A study by Ciftci et al. [21] previously showed that, at their institution, 41.3% of patients who presented with acute torsion also had undergone ultrasound [21]. Our study showed a significantly higher ultrasound usage, at 77.6% of the diagnosis of torsion in boys presenting with acute scrotum. Studies in the literature have reported sensitivity rates varying from 78.6% to 100% and specificity rates varying from 76.9% to 100% [22–26]. On the basis of sensitivity, specificity, and accuracy of 100%, 97.9%, and 98.1%, respectively, at our institution, color Doppler ultra-sound is a good examination to determine the absence or presence of testicular torsion. A 100% NPV indicates that a negative acute scrotum examination using color Doppler ultrasound can confidently rule out a testicular torsion at time of the examination. Although color Doppler ultrasound has been shown in this study to be useful for determining the absence and presence of testicular torsion, it should be noted that eight reports, all by adult radiologists, were deemed inconclusive for a final diagnosis. Thus, it is important to emphasize the utility of defining a clear final diagnosis in the dictation report to enable the rest of the team to efficiently and effectively approach management.
It is important to note that seven patients were sent straight to surgery with suspicion of testicular torsion without ultrasound imaging. Of these, one patient did not have torsion. Thus, with a 100% NPV on ultrasound and the case of a false-positive finding in a patient using uro-logic clinical suspicion, imaging is useful in preventing unnecessary surgical exploration.
Absent flow in the testes was confirmed as a defining characteristic on ultrasound for testicular torsion. In their retrospective study, Yang et al. [12] showed that swollen testes, abnormal testicular texture, and decreased or absent testicular blood flow were common findings associated with testicular torsion on color Doppler ultrasound. Our study confirmed similar findings of decreased and heterogeneous echogenicity and enlarged scrotal size on ultrasound. Nussbaum et al. [24] showed that in pediatric patients with testicular torsion, an enlarged hypovascular and echogenic epididymis was common. Similarly, a significantly higher incidence of decreased flow in the epididymis in testicular torsion patients was reported in our study. In 2008, Kaye et al. [27] reported 55 boys with scrotal exploration; 37 were heterogeneous on ultrasound, and all the heterogeneous testes were found to be nonsalvageable on exploration. Chmelnik et al. [28] reported 16 boys with testicular torsion, of whom those with focal sonographic heterogeneity had a 100% orchiectomy or atrophy rate. However, in our study, we showed that the finding of heterogeneous testes on ultrasound does not always indicate atrophy in patients with testicular torsion because 58% of salvageable testes and 80% of nonsalvageable testes had heterogenic echo-genicity (p = 0.4137).
The literature shows testicular salvage rates varying from 50% to 94.7%, depending on the interval between onset of pain and surgical intervention [18, 19, 29–32]. When detorsed within 6 hours of the onset of pain, there is a reported 90–100% salvage rate, and this rate decreases to 20–50% after 12 hours and 0–10% when greater than 24 hours [30–32]. Our study showed a 69% salvage rate, well within the range documented in the literature; however, it is important to note that our findings of necrosis were based on surgical and not actual histopathologic findings. Despite some very long times to presentation, many testicles did not show long-term atrophy. On average, boys with torsed testes that were viable on exploration reported pain for an average of 18.7 hours before surgery, whereas boys with torsed nonsalvageable testes reported an average of 21.7 hours of pain (p = 0.360). These times were not significantly different, indicating that the length of time may not be a significant factor for determining ability to salvage the torsed testicle. However two patients with presentations that were not suspicious for acute torsion were kept waiting in the emergency department overnight for ultrasound to be performed the next morning and diagnosed with nonsalvageable testes at surgery. This practice was likely related to a combination of atypical clinical presentation and limited after-hours ultrasound access for nonemergency indications. Since our study concluded, on the basis of these results, our institution has decreased the threshold for performing ultrasound in boys with atypical presentations in the setting of acute scrotum. Thus, it is important to note that based on our results, definitive investigations, such as color Doppler ultrasound, and treatment should never be delayed because of presumed testicular nonsalvageability. Improved education within the community may also decrease the time to emergency presentation.
The most reliable clinical findings from our study that are consistent with torsion included acute onset, nausea, abnormal lie, and loss of ipsilateral cremasteric reflex. These clinical predictors are in agreement with previous reports [3, 17]. In 523 emergency department visits, with 17 patients having testicular torsion, Beni-Israel et al. [3] found the clinical symptoms of pain duration of less than 24 hours, nausea and vomiting, abnormal cremasteric reflex, abdominal pain, and high position of the testes associated with increased likelihood of torsion. Similarly, in eight boys with testicular torsion of 79 boys with acute scrotum, Srinivasan et al. [17] showed that absent ipsilateral cremasteric reflex and scrotal changes were statistically predictive of torsion.
There are several limitations to our study. First, although our study had 342 boys with acute scrotum, we only had a small number of boys (n = 35) with testicular torsion. With an increased number of torsion patients, our data may have more statistical power and be better for extrapolation to the pediatric population. There was also no reference standard for a nontorsion diagnosis. For example, the torsion patients were confirmed with either follow-up clinical results or surgical exploration; however, the diagnosis for the nontorsion patients was based on the history, physical examination, ultrasound findings, and clinical follow-up. The clinical follow-up was thought to be relatively reliable on the basis of the centralized nature of urologic referral patterns in the health care region and the completeness of the electronic medical records. It should be recognized that patients with false-positive results may have been correctly diagnosed because surgical findings were used for the reference standard and patients may have torsed and detorsed by time of exploration. Our study was retrospective. Thus, the data available for history, clinical examination, and imaging findings were limited in some patients. Of note, the duration of symptom onset was based on an estimated time from the patient's memory instead of an accurately measured time period. To overcome the missing ultrasound data limitation, we had one blinded pediatric radiolo-gist review and provide imaging findings for the ultrasound images of every patient in our study to ensure that the imaging data collection was complete. Finally, because only one radiologist reviewed the ultrasound images, the definition of heterogeneous echotexture used in our study may be different from that used in previously published studies because there may be some interobserver variability. Thus, it would be ideal to provide a standardized definition for “heterogeneous echotexture” for future studies to allow interstudy comparisons.
We conclude from our study that color Doppler ultrasound is an accurate tool for detection of testicular torsion in patients presenting to the emergency department with acute scrotum, and, correspondingly, our institution has a heavy reliance on the tool for diagnosis. A negative ultrasound examination is highly predictive of the absence of torsion at the time of imaging. Although there is some variability regarding what findings on color Doppler ultra-sound are associated with testicular torsion, absent testicular flow, heterogeneous echotexture, enlarged scrotum, and decreased epididymal flow were shown to be good imaging indicators in our study. Although heterogeneity of the testis may be seen in torsion, we showed that it is not a good predictor for necrosis, which is a contradiction to previous proposals in the literature. It is important to recognize that investigation and treatment should not be delayed on the presumption that a testicle is no longer salvageable when a patient presents longer than 6 hours after symptom onset. Finally, there is a need for education of nonpediatric radiologists to decrease the number of inconclusive reports, which could result in delay in diagnosis and management or inappropriate surgery. At our institution, all pediatric ultrasound examinations are now performed by pediatric radiologists who have greater expertise and have a decreased number of indeterminate diagnoses.
WEB
This is a Web exclusive article.
| References | |
|---|
