Patterns of Occult Hip Fractures and Mimics Revealed by MRI

Prevalence Prevalence and Patterns of Occult Hip Fractures and Mimics Revealed Revealed by MRI Mayumi Oka1,2 Johnny U. V. Monu1

OBJECTIVE. The diagnosis of hip fractures can be difficult on radiography alone. MRI is frequently used to confirm or deny the presence of a minimally displaced hip fracture. This study evaluates the patterns of injury seen on MRI that ar e difficult to diagnose on radiography. MATERIALS AND METHODS. MRIs of 73 patients who were examined for possible hip fractures and whose radiographic findings were negative or equivocal for hip fracture were reviewed. Seventy-six studies were performed in 73 patients who were between 24 and 102 years old. MRIs were evaluated for the pr esence and location of bone or soft-tissue sof t-tissue injury. Muscle injuries were categorized on the basis of location and type of injury. RESULTS. Forty-six percent (35/76) of the studies showed subtle fractures. Seventeen fractures were in the proximal femur and 18 in the innominate bone. Soft-tissue abnormalities were common, found in 65% of the studies. Twenty percent of the MRI findings were considered normal because there was no apparent finding on the images to explain the patients’ symptoms. CONCLUSION. Soft-tissue abnormalities are commonly seen alone or in association with subtle fractures on MRI in the evaluation of patients with a clinical suspicion of hip fracture. MRI is recommended for all symptomatic patients whose radiographic findings are negative for hip fracture.

H

Received April 5, 2002; accepted after revision August 19, 2003. Presented at the 2001 annual meeting of the American Roentgen Ray Society, Seattle, WA. 1

Department of Radiology, University of Rochester School of Medicine and Dentistry, University of Rochester Medical Center, 601 Elmwood Ave., Box 648, Rochester, NY 14642. Address correspondence to J. U. V. Monu.

2

Present address: Department of Radiology, Johns Hopkins Hospital, 600 N Wolfe St., Baltimore, MD 21287.

AJR 2004;182:283–288

0361–803X/04/1822–283 © American Roentgen Ray Society

AJR:182, February 2004

ip injury is a growing medical problem, mostly because of an increase in the elderly population and in high-velocity motor vehicle trauma affecting young people [1]. Prompt and early diagnosis is important in minimally displaced fractures of the femoral neck because delayed diagnosis and treatment may result in significant displacement of the fracture fragment [2]. Significant displacement will alter the treatment and require more extensive surgery, such as hemiarthroplasty instead of stabilization with internal fixation. Diagnosis of minimally displaced hip fractures on radiographs can be challenging, especially in elderly patients with osteoporosis [3–7]. In these patients, in the face of reasonable clinical suspicion for fracture, MRI is recommended for further evaluation when conventional radiographic findings are negative or equivocal [7–10]. A subset of patients exist who have clinical signs and symptoms of hip fracture but do not show a fracture on MRI. Various softtissue injuries have been found on MRI of  some of these patients [11, 12].

Evaluating the incidence and type of bone and muscle injuries in cases of clinical suspicion of hip fracture will help clarify injury patterns that clinically mimic femoral neck fractures. Materials and Methods Patients with a clinical suspicion of hip fracture who underwent MRI of the hip because radiographic findings were reported to be negative were retrospectively identified from our radiology database. Seventy-six studies from 73 patients—28 men and 45 women—satisfied the selection criteria and form the population of this study. One patient was evaluated for bilateral hip fractures, and one patient with systemic lupus erythematosus who was undergoing steroid therapy was evaluated on three occasions. The patients were 24–102 years old (average, 67 years). All images were reviewed in conference by both authors, one a fellowship-trained and experienced musculoskeletal radiologist and the other a trainee radiologist. Except for three patients who were evaluated with only frontal radiography of the pelvis, all patients underwent radiography of the pelvis in the anteroposterior projection and of the hip in anteroposterior and cross-table lateral projec-

283

Oka and Monu tions. The radiographs were reviewed for adequacy of the image quality and the presence of fractures. A study was deemed adequate when the acetabulum, femoral head, and femoral neck were visualized on two images. No patients were excluded because of inadequate radiographs. Sixty-four of the 76 MRIs of the hip were obtained within 48 hr after the radiographs. However, the interval between radiographs and MRI studies ranged from 5 days to 3 weeks in the other 12 cases. All MRI studies were performed on a 1.5-T unit (Signa, General Electric Medical Systems, Milwaukee, WI) using a commercially available pelvic phased array surface coil as a receiver. Imaging protocols varied but generally included T1-weighted spin-echo images (TR/TE range, 600/14–18; slice thickness, 5 mm; interslice gap, 0.5–1.0 mm; matrix, 256 × 192) in the axial and oblique coronal planes; T2-weighted fast spin-echo images (TR range/TE range, 3,200–5,000/96–102; slice thickness, 5 mm; interslice gap, 0.5–1.0 mm; matrix, 256 ×192) with fat saturation; and occasional STIR images (TR/TE, 4,800/60; inversion time, 150 msec; slice thickness, 5 mm; interslice gap, 0.5–1.0 mm; matrix, 256 × 192) in the coronal or axial plane. Images in the sagittal plane using dual-echo spin-echo pulse sequences were infrequently obtained. Field of view ranged between 16 and 22 cm to cover the index hip joint. The MRIs were evaluated for the presence and site of bone or soft-tissue injury. The bone injuries were characterized as bone contusion or fracture. A fracture was diagnosed when a linear lowsignal focus was surrounded by an intermediatesignal area on T1-weighted images and the linear low-signal focus was surrounded by high signal on

T2-weighted images [9, 10, 12]. Ill-defined areas of altered signal intensity on T1-weighted images and high signal intensity on T2-weighted images were interpreted as bone marrow edema. Soft-tissue injuries were further characterized as muscle edema, hematoma, partial muscle tear, or complete muscle tear using criteria that have been previously defined [13–19]. Acute muscle injuries have associated edema and hemorrhage that cause prolongation of the T1 and T2 relaxation times of the injured tissue [14]. Muscle in juries were diagnosed when areas of increased signal intensity were seen on T2-weighted or STIR images. Complete tears showed total disruption of the muscle, muscle retraction, blood, and edema in the tear. A partial tear is seen as a focal interruption of continuity of  the muscle without complete transection of the muscle or tendon. Muscle edema was seen as a nongeographic area or an infiltrative pattern of low signal on T1-weighted and high signal on T2-weighted images that could not be distinguished from early interstitial hemorrhage. Any confined area of altered signal within or outside the muscle fibers was identified as a hematoma, which is a form of muscle injury [13, 14]. In a setting of trauma, it may be difficult to differentiate an isolated hematoma from a hematoma arising from a muscle injury; a hematoma was therefore considered to be muscle injury in this study. A hematoma may show varying signal intensities on both T1- and T2-weighted images, depending on the evolving state of hemoglobin [20]. We further analyzed our results in terms of  muscle function and stratified the muscles around the hip joint into groups as follows: thigh adductor, thigh abductor, external rotator, internal rota-

TABLE 1

Fractures and Locations (n = 35)

Fracture Location

No.

Femoral fractures Neck

8

Intertrochanteric

5

Greater trochanteric

4

Total

17

Pelvic fractures Pubic rami alone

13

Acetabulum alone

2

Pubic rami and acetabulum

3

Total

18

Grand total

35

tor, flexor, and extensor. This stratification was based on their primary function or function when the hip was in a neutral position [21]. The individual muscles and the functional muscle groups are shown in Appendix 1.

Results

Our retrospective review showed that no fractures were missed on radiographs. Thirty-five (46%) of 76 studies were found to have fractures on MRI (Table 1). The fractures were four fractures of the greater trochanter (Fig. 1), five intertrochanteric fractures, and eight femoral neck fractures.

A

B

Fig. 1.—77-year-old woman with fracture of greater trochanter after fall. Fracture was difficult to visualize on conventional radiographs. A, Frontal radiograph of left hip appears to show normal findings. B, T1-weighted coronal image of left hip shows li near low-signal focus (arrowheads ), indicating fracture.

284

AJR:182, February 2004

MRI of Hip Fractures

A

B

Fig. 2.—67-year-old woman who presented with persistent right hip pain after fall 3 weeks earlier. A, Frontal radiograph of right hip shows unremarkable findings. B, Coronal T2-weighted image shows abnormally high signal in anterior column of hip and in area of superior pubic ramus, consistent with trabecular fracture. Linear abnormally high signal is seen in adductors and obturator externus and is compatible with presence of interstitial edema or hemorrhage.

In the innominate bones, there were five acetabular fractures and 13 fractures of the obturator ring. Two patients had bone marrow edema without focal cortical disruption (Fig. 2).

In the group with fractures on MRI, 24 (69%) had associated muscle injury. The obturator externus muscle, a major external rotator, was abnormal in 15 studies (43%). Injury to

A

the other external rotators of the hip, which include the gluteus maximus, piriformis, obturatorius internus, gemelli, and quadratus femoris, was seen in 17 studies (49%). Injury

B

Fig. 3.—73-year-old man who experienced hip pain after fall. He had no fractures, but MRI showed muscle injury. A, T1-weighted coronal image shows abnormal signal of interstitial muscle hemorrhage (arrows ) in obturator externus muscle near its attachment to greater trochanter and in musculotendinous junction. B, T2-weighted coronal image shows abnormal signal in obturator externus muscle near its attachment to greater trochanter and in musculotendinous junction. These findings are compatible with interstitial hemorrhage (arrows ).

AJR:182, February 2004

285

Oka and Monu

TABLE 2

Muscle Injury, by Individual Muscles ( n = 76)

Muscle

No Fracture ( n = 41)

Fracture ( n = 35)

Total (n = 76)

Iliopsoas

5

5

10

Obturator internus

3

7

10

Obturator externus

13

15

28

Pectineus

3

5

8

Adductor longus

6

7

13

Adductor brevis

7

10

17

Adductor magnus

7

5

12

Gluteus minimus

7

8

15

Gluteus medius

7

6

13 9

Gluteus maximus

6

3

Tensor fasciae latae

1

1

Rectus femoris

2

2

Vastus lateralis

2

2

Piriformis

2

2

to the thigh adductor was seen in 22 studies (63%). In this group of patients with adductor muscle injuries, five (23%) were associated with femoral fractures and 17 (77%) were related to fractures of the innominate bone. On the basis of MRI findings, 41 patients (56%) did not have a fracture. Twenty-six (63%) of the patients in this group had muscle injuries. The obturator externus muscle was the most frequently injured muscle, with abnormality seen in 13 (32%) of the 41 patients (Fig. 3). Injury to the other external rotators of the hip was seen in 14 studies (34%). Twelve studies (29%) showed injuries in the abductor group (Tables 2 and 3 and Appendix 1). Trauma was the indication for the study in 42 patients (58%), including 27 (77%) of the 35 studies showing fractures and 15 ( 37%) of  the 41 studies showing no fracture. The remaining 31 patients (43%) had pain but no history of recent trauma, and the study was requested to consider unexpected fractures. Among these patients were eight fractures (26%) and 13 muscle injuries (42%).

Overall, muscle injuries were seen in 50 studies (66%). Muscle edema, hemorrhage, and partial tear represented most of the muscle injuries and were seen in 49 studies. One patient had an iliopsoas muscle abscess (Fig. 4). Mean patient age was 70 years in the fracture group and 63 years in the nonfracture group. The distribution of abnormalities detected in individual muscles and in various functional muscle groups is shown in Appendix 1 and Table 2. Injury to the gluteus medius and gluteus minimus, which are strong internal rotators of  the hip and thigh abductors, was seen with equal frequency in patients with and without fractures. Injury to the gluteus maximus was more frequently seen in patients without fracture. The external rotators are the most commonly injured muscles in both groups. Fifteen patients (21%) (13 women and two men) had no fractures or muscle injuries on MRI. The mean age in this group was 57.3 years (range, 33–81 years).

Discussion TABLE 3

Muscle Injury, by Muscle Group (n = 76)

Muscle Group Extensors

No Fracture (n = 41)

Fracture (n = 35)

11

8

Flexors

5

5

Adductors

7

12

Abductors

11

9

Internal rotators

13

14

External rotators

17

19

286

The term “hip fracture” implies a proximal femoral fracture, usually involving the femoral neck. Proximal femoral fractures involving the femoral neck or intertrochanteric regions are devastating injuries to the hip that need to be promptly diagnosed. However, injuries that involve the innominate bone or adjacent muscles may present with similar signs and symptoms. In this series, 24% of the studies showed pelvic fractures that were occult on radiography, with fractures of the obturator ring (21%) being the most common.

Generally, patients who sustain obturator ring fractures or muscle injury around the hip joint do not require surgery; however, they can still have prolonged immobilization, difficulty with ambulation as a result of pain, and subsequent impaired hip function if the fracture is not properly treated. Physical therapy and early mobilization are essential for these patients to recover as much function as possible. Muscle injuries were present in 65% (49/  76) of studies regardless of the presence or absence of fracture. Muscle injury may be caused by a direct blow (contusion) or an indirect injury (strain due to abnormal stretching) and may result in edema, hemorrhage, muscle tear, and hematoma. Edema and hemorrhage present as high signal intensity interspersed in muscles on T2-weighted images. In this series, the adductor muscle group was frequently injured in association with an obturator ring fracture. Muscle tears tend to occur at or near the musculotendinous junction and are seen as focal high signal intensity on T2-weighted images. Moderate strain and partial tears are difficult to distinguish from hemorrhage when the latter are located near the junction. MRI may not always differentiate between a muscle strain or tear that is associated with a fracture and hemorrhage into the muscles from a fracture. MRI findings reflect only the relative presence of edema, deoxyhemoglobin, and methemoglobin (Fig. 5). The frequency and distribution of muscle injuries were similar in the group of patients with fractures (69%) and in the group with no bone injuries (63%). Some of the observed clinical signs and symptoms are caused by associated muscle injuries [11]. The 46% occult fracture rate in this study is similar to a fracture rate of 54% in the study by Bogost et al. [11]. The differences in our figures may be partly explained by the fact that Bogost et al. studied only patients with a clear history of trauma. Because a clear history of  trauma may not be readily available in elderly patients, we included in our study population patients with a clinically suspected hip fracture with no history of trauma. Our position is supported by the subsequent observation that nearly 23% (35 patients) of the fractures were in our group of patients who did not recall or give a clear history of trauma. In addition to obturator ring fractures, Bogost et al. [11] reported a small number of  sacral fractures. We found no sacral fractures in our study population. The sacrum was not

AJR:182, February 2004

MRI of Hip Fractures

A

B

Fig. 4.—81-year-old woman with abscess in iliopsoas muscle who presented with left hip pain, no history of trauma, and clinical suspicion of occult fracture. A, T1-weighted axial image shows rounded low-signal focus (arrow ) adjacent to left iliacus muscle. B, T2-weighted image shows central area of high signal surrounded by concentric focus of low- and high-signal zones, consistent with abscess ( solid arrow ). Abnormally high signal in iliacus muscle (open arrows ) is result of muscle inflammation.

routinely included in our images because our imaging protocol called for a smaller field of  view. The incidence of muscle injury in our study population was similar to that of Bogost et al., which was 61% (40% in the fracture group and 21% in the no-fracture group). The elderly patient is generally fragile and does not require a major traumatic event to sustain a fracture or significant muscle in jury. A history of trauma may not be readily available or recalled. Furthermore, hip fractures in the elderly may present with atypical symptoms, such as several weeks of pain or gait instability. Occasionally in this population, some nontraumatic causes of hip pain such as an iliopsoas abscess (Fig. 4) and even exacerbation of hip arthritis may present acutely and mimic a hip fracture. Frequently, these elderly patients have coexisting morbidity that confuses the clinical picture and confounds accurate diagnosis. Because morbidity and mortality after hip injury in the elderly have a significant socioeconomic impact, prompt and accurate diagnosis is imperative [1]. MRI not only facilitates the diagnosis of  nondisplaced fractures not seen on radiographs [8, 10, 22, 23] but also provides information that may be beneficial for the appropriate treatment of other injuries. Limited MRI protocols with T1-weighted coronal images were previously reported to be sufficient to detect proximal femoral frac-

AJR:182, February 2004

tures [9, 24]. However, we emphasize the usefulness of T2-weighted sequences in the diagnosis of injuries other than hip fractures. T2-weighted images facilitate recognition of  soft-tissue injuries. In addition, areas of bone

marrow edema cannot be distinguished from sclerosis using T1-weighted images alone. The main limitation of this study is that it was retrospective. The study population was too small for statistical confirmation. The protocols

Fig. 5.—74-year-old woman with iliopsoas avulsion injury after fall 2 days before MRI examination. T2-weighted axial fast spin-echo fat-suppressed image shows abnormally high signal ( arrows ) in and surrounding iliopsoas muscle, which is consistent with partial tear of muscle and surrounding hemorrhage. Note also some abnormal signal around gluteus minimus muscle ( asterisks ).

287

Oka and Monu for MRI of the hip varied during the study period. However, our entire patient population had both T1-weighted and T2-weighted sequences as part of their study. Although multiple radiologists interpreted the conventional radiographs, our retrospective review showed that no fractures were missed on radiographs. In conclusion, soft-tissue injuries commonly accompany hip fractures but may exist in isolation and may mimic fractures in that they can produce similar symptoms. Soft-tissue injuries alone can be a cause of morbidity and merit additional attention. MRI is useful in the diagnosis of soft-tissue injuries, especially in the setting of clinically suspected hip fractures. MRI is recommended for all symptomatic patients whose conventional radiographs do not reveal a hip fracture. References 1. Rudman N, McIlmail D. Emergency department evaluation and treatment of hip and thigh injuries.  Emerg Med Clin North Am 2000;18:29–66 2. Lu-Yao GL, Keller RB, Littenberg B, Wennberg JE. Outcomes after displaced fractures of the femoral neck: a meta-analysis of one hundred and six published reports.   J Bone Joint Surg Am 1994;76:15–25 3. Schultz E, Miller TT, Boruchov SD, Schmell EB, Toledano B. Incomplete intertrochanteric fractures: imaging features and clinical management.

 Radiology 1999;211:237–240 4. Ganel A, Engel J, Oster Z, et al. Bone scanning in assessment of fractures of the scaphoid.   J Hand  Surg 1979;4:540–543 5. Matin P. Bone scintigraphy in the diagnosis and management of traumatic injury. Semin Nucl Med  1983;13:104–122 6. Grainger C, Garcia J, Howart NR, May M, Rosier P. Role of MRI in the diagnosis of insufficiency fractures of the sacrum and acetabular roof. Skeletal Radiol 1997;26:517–524 7. Haramati N, Staron RB, Barax C, Feldman F. Magnetic resonance imaging of occult fractures of the proximal femur. Skeletal Radiol 1994;23:19–22 8. Rubin SJ, Marquardt JD, Gottlieb RH, Meyers SP, Totterman SMS, O’Mara RE. Magnetic resonance imaging: a cost-effective alternative to bone scintigraphy in the evaluation of patients with suspected hip fractures. Skeletal Radiol 1998;27:199–204 9. Quinn SF, McCarthy JL. Prospective evaluation of patients with suspected hip fracture and indeterminate radiographs: use of T1-weighted MR images. Radiology 1993;187:469–471 10. Deutsch AL, Mink JH, Waxman AD. Occult fractures of the proximal femur: MR imaging.  Radiology 1989;170:113–6 11. Bogost GA, Lizerbram EK, Crues JV III. MR imaging in evaluation of suspected hip fracture: frequency of unsuspected bone and soft-tissue injury. Radiology 1995;197:263–267 12. May DA, Purins JL, Smith DK. MR imaging of  occult traumatic fractures and muscular injuries of the hip and pelvis in elderly patients.  AJR 1996;166:1075–1078

13. Kneeland JB. MR imaging of sports injuries of the hip.  Magn Reson Imaging Clin N Am 1999;7:105–115 14. El-Khoury GY, Brandser EA, Kathol MH, Tearse DS, Callaghan JJ. Imaging of muscle injuries. Skeletal Radiol 1996;25:3–11 15. De Smet AA. MR imaging of acute and remote muscle injuries. In: De Smet AA, ed.  Musculoskeletal MRI: normal anatomy and key pathology. (ARRS categorical course syllabus) Leesburg, VA: American Roentgen Ray Society, 2001:97–103 16. Deutsch AL, Mink JH. Magnetic resonance imaging of musculoskeletal injuries.  Radiol Clin North  Am 1989;27:983–1002 17. Palmer WE, Kuong SJ, Elmadbouh HM. MR imaging of myotendinous strain. AJR 1999;173:703–709 18. De Smet AA. Magnetic resonance findings in skeletal muscle tears. Skeletal Radiol 1993;22:479–484 19. Arrington ED, Miller MD. Skeletal muscle injuries. Orthop Clin North Am 1995;26:411–422 20. Bush CH. The magnetic resonance imaging of  musculoskeletal hemorrhage. Skeletal Radiol 2000;29:1–9 21. Warwick R, Williams PL. Gray’s anatomy, 35th ed. Edinburgh, Scotland: Longman Group, 1973:559–571 22. Pandey R, McNally E, Ali A, Bulstrode C. The role of MRI in the diagnosis of occult hip fractures. Injury 1998;29:61–63 23. Rizzo PF, Gould ES, Lyden JP, Asnis SE. Diagnosis of occult fractures about the hip: magnetic resonance imaging compared with bone scanning.  J   Bone Joint Surg Am 1994;75:395–401 24. Mlinek EJ, Clark KC, Walker CW. Limited magnetic resonance imaging in the diagnosis of occult hip fractures. Am J Emerg Med 1998;16:390–392

APPENDIX 1. Functional Muscle Groups About the Hip Extensors

Primary: Gluteus maximus, adductor magnus Other: Gluteus medius, gluteus minimus, biceps femoris, semitendinosus, semimembranosus

Flexors

Primary: Iliopsoas, pectineus, sartorius, tensor fasciae latae Other: Adductors, rectus femoris

Adductors

Primary: Adductor longus, adductor brevis, adductor magnus, gracilis Other: Obturator externus, pectineus, gluteus maximus

Abductors

Primary: Gluteus medius, gluteus minimus, tensor fasciae latae, sartorius Other: Piriformis, quadratus femoris

Internal Rotators

Primary: Gluteus medius, gluteus minimus, tensor fasciae latae, pectineus Other: Semitendinosus, semimembranosus

External Rotators

Primary: Gluteus maximus, piriformis, obturator internus, obturator externus, superior gemellus, inferior gemellus, quadratus femoris

288

AJR:182, February 2004