Downloaded from www.ajronline.org by 46.3.193.109 on 01/20/18 from IP address 46.3.193.109. Copyright RRS. For personal use only; all rights reserved Nancy M. Major 1 Steven T. Crawford 1,2 Received July 3, 2001; accepted after revision ugust 21, 2001. 1 Department of Radiology, Duke University Medical Center, Erwin Rd., ox 3808, Durham, NC 27710. ddress correspondence to N. M. Major. 2 Present address: Mallinckrodt Institute of Radiology, 510 S. Kingshighway lvd., St. Louis, MO 63110. JR 2002;178:413 418 0361 803X/02/1782 413 merican Roentgen Ray Society Elbow Effusions in Trauma in dults and Children: Is There an Occult Fracture? OJECTIVE. The purpose of this study was to evaluate whether a detectable abnormality was present on MR imaging without a visible fracture on conventional radiography in the setting of trauma. recent retrospective study based on the presence or absence of periosteal reaction on follow-up radiographs concluded that fractures were not always present. The discrepancies in the literature over the usefulness of joint effusions as an indicator of fracture caused us to evaluate whether fractures were present more often than identified by conventional radiography. To do this, we used MR imaging. MTERILS ND METHODS. Thirteen consecutive patients (age range, 4 80 years; seven children and six adults), whose posttrauma elbow radiographs showed an effusion but no fracture, underwent screening MR imaging. RESULTS. ll patients showed bone marrow edema. Four of the seven children had fractures on screening MR imaging, and all adults had some identifiable fractures. CONCLUSION. Preliminary data using screening MR imaging suggests that an occult fracture usually is present in the setting of effusion without radiographically visualized fracture. E lbow effusions are seen on lateral radiographs by the presence of the fat pad sign. This was first described by Norell in 1954 [1]. These fat pads can be elevated by the presence of fluid in the joint caused by an acute fracture. We wondered how often an occult fracture was present. For years, residents were taught that elbow effusions without a visible fracture are synonymous with an occult fracture in the setting of trauma. Indeed, multiple articles have indicated a strong relationship between an occult fracture and a joint effusion [2 5]. In 1984, Swischuk et al. [6] disputed this association, and more recently, the relationship between an occult fracture and a joint effusion has come into question in a retrospective study based on the presence or absence of periosteal reaction on follow-up radiographs [7]. Here, the authors found a prevalence of only 15% and 17%, respectively. recent prospective study, again based on follow-up radiographs, showed that 76% of patients had evidence of a fracture [8]. With this wide variation of data, the usefulness of the posterior fat pad sign has come into question. The purpose of our study is to evaluate the presence of an occult fracture using screening MR imaging as the standard for fracture detection in both the pediatric and adult populations. Materials and Methods Thirteen consecutive patients (seven children and six adults) were imaged prospectively over a 5-month interval. The age range was 4 80 years. The patients were selected consecutively and included in the study when posttraumatic elbow radiographs from the emergency department showed an effusion but no visible fracture. The patients underwent screening MR imaging. The radiographs obtained in the emergency department consisted of frontal and lateral views and a radial head view of the elbow obtained in the standard fashion. The emergency department radiology resident, a senior radiology resident, and a musculoskeletal radiologist reviewed the radiographs. patient was included in the study when an effusion was present but no fracture was identified by any of the three radiologists. The mean time between the injury and the MR imaging was 3 days (range, 0 6 days). MR imaging was performed using a 1.5-T Signa scanner (General Electric Medical Systems, Milwaukee, WI), and a flexible coil (send-receive) T2-weighted fast spin-echo image with fat suppression (TR/effective TE, 3500/65; slice thickness, 4 mm; interslice gap, 0.4 mm; excitations, 2; matrix, 256 192) was obtained in the coronal plane (parallel to the epicondyles). The field of view included the distal humerus to the proximal radius and ulna. In one patient, an additional T1- weighted coronal image (TR/TE, 600/13) was obtained inadvertently. Routine elbow MR imaging was not used because the purpose of the study was to as- JR:178, February 2002 413
Major and Crawford Downloaded from www.ajronline.org by 46.3.193.109 on 01/20/18 from IP address 46.3.193.109. Copyright RRS. For personal use only; all rights reserved sess the presence of an injury. We felt this could be addressed quickly with a marrow-sensitive sequence. The musculoskeletal radiologists and the senior resident reviewed the MR images in consensus. fracture was defined by the presence of focal linear low signal intensity surrounded by high signal intensity or disruption of the cortex. contusion was defined as high signal intensity without a visible fracture line. Results ll patients were noted to have a joint effusion on MR imaging. Of the pediatric patients, two showed Salter-Harris Type III fractures of the epicondyle; one, Salter-Harris Type I fracture of the medial epicondyle (Fig. 1); one, radial head fracture (Fig. 2); and three, bone contusions one of the lateral epicondyle, one of the radial head, and one in the supracondylar area (Table 1). Therefore, 57% of the pediatric patients studied had a fracture that was only seen as an effusion on the initial radiographs. ut all (100%) had a detectable injury as evidenced by bone marrow edema. Fig. 1. 5-year-old girl with fracture of medial epicondyle., nteroposterior and lateral radiograph shows no fracture., Lateral radiograph shows evidence of joint effusion. C, Coronal fast spin-echo fat-suppressed T2-weighted MR image (TR/effective TE, 3500/65) shows high signal intensity in trochlea compatible with contusion (large arrow). Patient also has physeal injury that corresponds to Salter-Harris Type I fracture as shown by focus of high signal intensity seen along physis (small arrows). C Fig. 2. 10-year-old girl with radial head fracture., nteroposterior radiograph shows no evidence of fracture., Coronal fast spin-echo fat-suppressed T2-weighted MR image (TR/effective TE, 3500/65) shows radial head fracture as evidenced by cortical disruption (arrow) and surrounding bone marrow edema. (Fig. 2 continues on next page) 414 JR:178, February 2002
Elbow Effusions in Trauma Fig. 2. (continued) 10-year-old girl with radial head fracture. C, Consecutive posterior MR image shows linear high signal intensity located along fracture line (arrows). Downloaded from www.ajronline.org by 46.3.193.109 on 01/20/18 from IP address 46.3.193.109. Copyright RRS. For personal use only; all rights reserved Three of the six adult patients had osteochondral fractures, two in the capitellum (Fig. 3), and the third in the radial head (Fig. 4). Two of these patients showed a loose body in the joint. The remaining adult patients showed a radial head fracture (Fig. 5) and a nondisplaced medial epicondyle avulsion and a coronoid process fracture with considerable extensor tendon complex injury from lateral epicondyle that required additional therapy (Fig. 6) (Table 1). ll of the adult patients had radiographically occult fractures as seen by MR imaging. Ten (77%) of the 13 patients had fractures that were present on the MR image that were Fig. 3. 80-year-old man with osteochondral fracture in capitellum., nteroposterior radiograph shows no evidence of fracture., Coronal fast spin-echo fat-suppressed T2-weighted MR image (TR/ effective TE, 3500/65) shows focal cartilage defect in capitellum with underlying marrow high signal intensity, consistent with osteochondral fracture (arrow ). Defect should not be confused with pseudodefect of capitellum because abnormality is too anterior, and high signal intensity present around lesion would not be seen. Note remainder of images determined high signal intensity in radial shaft to be residual red marrow. not seen on the initial radiographs. Thirteen of the 13 patients showed bone marrow edema indicative of an injury. Discussion For many years, the fat pad sign has been used as an indication of an occult fracture. The posterior fat pad sign has been reported as a reliable sign of an elbow joint effusion [1 3, 5]. However, the presumed incidence of a fracture has varied widely in both the radiologic and orthopedic literature. Numerous studies have evaluated for fractures retrospectively, using conventional radiographic follow-up to determine the presence of a fracture [6 11]. Our goal was to determine the presence of an injury versus a fracture of the elbow using the gold standard of a marrow-sensitive sequence (fast spin-echo with fat suppression) after identifying a joint effusion on conventional radiography without a visible fracture noted by any of the three radiologists. We chose fast spin-echo with fat suppression as the sequence because of the increased conspicuity of bone marrow edema seen with this sequence. We found that four pediatric patients and six adult patients had fractures seen on the MR images that were not seen on C JR:178, February 2002 415
Major and Crawford Downloaded from www.ajronline.org by 46.3.193.109 on 01/20/18 from IP address 46.3.193.109. Copyright RRS. For personal use only; all rights reserved the initial radiographs. Ten (77%) of the 13 patients had fractures that were seen on the MR imaging follow-up, and all patients had bone marrow edema. In other words, all patients had an abnormal finding to explain the presence of a joint effusion. In retrospect, the results of the MR imaging follow-up confirmed that none of the fractures were identified on the radiographs. One explanation for the higher incidence of fracture in our study compared with other reports is that the sensitivity of MR imaging is much greater than that of conventional radiography. The discrepancy between the initial studies [2 4] that stated a high incidence of occult fracture and the more recent studies that dispute this claim [2, 7, 9, 10] has been previously discussed [8] and is likely due to improvements in radiographic technique, which allows better detection of the fractures. Only a single layer of periosteum is within the joint capsule that can produce periosteal reaction at the edge of the joint. Therefore, periosteal reaction can be difficult to identify on conventional radiographs, making conventional radiography less sensitive than MR imaging for an accurate diagnosis. The current treatment for occult fractures about the elbow is posterior splinting. Two of C Fig. 4. 34-year-old man with osteochondral fracture in radial head., nteroposterior radiograph shows no evidence of fracture., T1-weighted coronal spin-echo MR image (TR/TE, 600/13) shows focal cartilage defect in radial head with surrounding marrow T1 shortening consistent with osteochondral fracture (arrow). C, Coronal fast spin-echo fat-suppressed T2-weighted MR image (TR/ effective TE, 3500/65) shows loose body in joint space from previously described osteochondral defect (arrow). one marrow edema is present in radial head. the patients (15%) in our study, both adults, required surgery instead of posterior splinting [12]. One patient required repair of the extensor tendon complex and another required removal of a loose body after an osteochondral fracture. Neither of these abnormalities was suspected on conventional radiography. The findings at surgery were consistent with the findings on MR imaging. The two patients with osteochondral lesions had no history of prior elbow trauma. The osteochondral lesions were felt to be acute, given the presentation of trauma and the findings of edema and a loose body. The remainder of the pa- 416 JR:178, February 2002
Elbow Effusions in Trauma Downloaded from www.ajronline.org by 46.3.193.109 on 01/20/18 from IP address 46.3.193.109. Copyright RRS. For personal use only; all rights reserved tients did well with conservative treatment (posterior splinting). Previous reports have stated that even if a fracture was present on follow-up radiographs, the clinical treatment was not altered [12]. Griffith et al. [13] revealed a variety of bone and soft-tissue injuries in the pediatric population with MR imaging. Despite the information provided by the study, no change in treatment or clinical outcome occurred. We have the same conclusion in our pediatric population, although our study included a smaller number of patients. Shortcomings of our study include the small number of patients. The design of this study was to investigate whether an abnormality exists when an elbow joint effusion is present in the setting of trauma and to refute the idea by Donnelly et al. [7]. Our study was not designed to separate the pediatric population from the adult population but rather to see what abnormalities might exist within each. The screening protocol chosen was to enable detection of fractures and bone marrow edema that would indicate injury. n additional shortcoming was the use of a three-view conventional radiographic series to evaluate the elbow. lthough this is routine at our institution, four views are often obtained at other institutions. Perhaps a fracture would be evident on the additional view. Fig. 5. 26-year-old man with radial head fracture., nteroposterior radiograph shows no evidence of fracture., Coronal fast spin-echo fat-suppressed T2-weighted MR image (TR/ effective TE, 3500/65) reveals radial head fracture shown by linear low signal intensity involving medial cortex of radial head (arrows) surrounded by bone marrow edema. High signal intensity in proximal ulna represents partial volume averaging with joint fluid. In reviewing our data, children can have both supracondylar injuries (which are usually felt to be the likely site of the abnormality) and radial head injuries. Similarly, adults can sustain injuries to any of the bony structures about the elbow joint (radial head is most often suspected). lthough most patients will be treated with a posterior splint TLE 1 Note. ll patients had joint effusion. and will go on to full recovery, the presence of an osteochondral defect and possible loose body or ligamentous injury may be the cause of continued pain or accelerated degenerative changes if unrecognized. In conclusion, our investigation using MR imaging, rather than previous studies using findings on follow-up radiography, suggests MR Imaging Findings for 13 Pediatric and dult Patients Patient ge (yr) MR Imaging Findings 1 4 Supracondylar contusion 2 5 Salter-Harris type III fracture of medial epicondyle 3 8 Salter-Harris type III fracture of medial epicondyle with tremendous soft-tissue swelling and brachialis muscle strain 4 10 Radial head fractures 5 10 Lateral epicondylar contusion 6 11 Salter-Harris type I medial epicondyle fracture 7 15 Radial head contusion 8 26 Radial head fracture 9 34 Radial head osteochondral fracture loose body 10 37 Osteochondral fracture at lateral aspect of capitellum 11 39 Medial epicondylar avulsion fracture 12 65 Coronoid process fracture with complete disruption of extensor tendons 13 80 Capitellar osteochondral fracture with loose body JR:178, February 2002 417
Major and Crawford Downloaded from www.ajronline.org by 46.3.193.109 on 01/20/18 from IP address 46.3.193.109. Copyright RRS. For personal use only; all rights reserved that an occult fracture usually is present in the setting of effusion without radiographically visualized fracture. Indeed, all of the imaged patients had intraarticular injury, which confirms the traditional teaching that occult fractures or injuries are present. cknowledgment We thank Mirjana Cudic for her assistance with preparation of this manuscript. References 1. Norell H-G. Roentgenologic visualization of the extracapsular fat. ct Radiol 1954;42:205 210 Fig. 6. 65-year-old man with nondisplaced medial epicondyle avulsion and coronoid process fracture with extensor tendon complex injury from lateral epicondyle., nteroposterior radiograph shows no evidence of fracture., Coronal fast spin-echo fat-suppressed T2-weighted MR image (TR/effective TE, 3500/65) shows high signal at olecranon process (small arrow). In addition, bone marrow edema is seen in lateral epicondyle (large arrow). Common tendon of wrist-hand extensor-supinator complex has been avulsed from its origin (curved arrow). 2. ledsoe RC, Isenstark JL. Displacement of aft pads in disease and injury of the elbow: a new radiographic sign. Radiology 1959;73:717 724 3. Kohn M. Soft-tissue alterations in elbow trauma. JR 1959;82:867 874 4. ohrer SP. The fat pad sign following elbow trauma: its usefulness and reliability in suspecting invisible fractures. Clin Radiol 1970;21:90 94 5. Rogers LF. Fractures and dislocations of the elbow. Semin Roentgenol 1978;13:97 107 6. Swischuk LE, Hayden CK, Kupfer MC. Significance of intraarticular fluid without visible fracture in children. JR 1984;142:1261 1262 7. Donnelly LF, Klostermeier TT, Klosterman L. Traumatic elbow effusions in pediatric patients: are occult fractures the rule? JR 1998;171:243 245 8. Skaggs DL, Mirzayan R. The posterior fat pad sign in association with occult fracture of the elbow in children. J one Joint Surg m 1999;81: 1429 1433 9. Quinton DN, Finlay D, utterworth R. The elbow fat pad sign: brief report. J one Joint Surg r 1987;69:844 845 10. Morewood DJ. Incidence of unsuspected fractures in traumatic effusions of the elbow joint. r Med J (Clin Res Ed) 1987;295:109 110 11. de eaux C, eattie T, Gilbert F. Elbow fat pad sign: implications for clinical management. J R Coll Surg Edinb 1992;37:205 206 12. Henriksen M, Gehrchen PM, Jorgensen M, Gerner-Smidt H. Treatment of traumatic effusion in the elbow joint: a prospective, randomized study of 62 consecutive patients. Injury 1995;26:475 478 13. Griffith JF, Roebuck DJ, Cheng JCY, et al. cute elbow trauma in children: spectrum of injury revealed by MR imaging not apparent on radiographs. JR 2001;176:53 60 418 JR:178, February 2002