We are aware of only two previous studies evaluating the long-term outcomes of unstable pelvic fractures in skeletally immature patients or the specific factors that influence long-term clinical or functional outcomes1,2. Historically, pelvic fractures in children have been treated nonoperatively, but potential complications such as pelvic asymmetry, nonstructural scoliosis, back pain, lower-extremity length inequality, limp, and others have been recognized1,3-8. Additionally, an unstable pelvic fracture in a skeletally immature patient may not remodel to a symmetric configuration. Previous authors have suggested that nonoperative management of displaced pelvic fractures may result in pelvic asymmetry, leading to poor clinical results4,9. However, operative and nonoperative treatment may result in similarly good outcomes if an acceptable reduction is achieved and maintained1. In most studies, the results for children with and those without closed triradiate cartilages have been analyzed together1-10. Consequently, it is unclear how much asymmetry is tolerable following an unstable pelvic fracture in a patient with open triradiate cartilages.
The purpose of this study was to evaluate the long-term functional outcomes of unstable pelvic fractures in skeletally immature patients. We hypothesized that pelvic remodeling following a displaced fracture is minimal and that residual displacement of >1 cm results in inferior functional outcome scores.
Materials and Methods
This study was performed at the Denver Health Medical Center, a regional level-I trauma center, and the Denver Children's Hospital, the regional pediatric level-I trauma center, after approval was obtained from the institutional review board at each medical center. Eligible patients were identified through a computer search of the trauma registry for pelvic fractures in patients who were seventeen years of age or younger. The Tile classification system for pelvic ring disruption11 (see Appendix) was used to classify the fractures in this study. Patients with closed triradiate cartilages at the time of injury, a stable pelvic ring injury (Tile type A) such as an avulsion fracture, or an isolated acetabular fracture were excluded. The medical records, radiographs, and computed tomography scans of patients with open triradiate cartilages and an unstable or partially stable pelvic (Tile type-B or C) fracture seen between 1986 and 2000 were reviewed. A Modified Injury Severity Score (MISS)12 was calculated retrospectively from information in the medical record and trauma registry. A MISS of ≥25 points is associated with greater disability and morbidity in children with a traumatic injury2,12. Pelvic asymmetry was determined on the anteroposterior radiograph, with use of the method of Keshishyan et al. (Fig. 1)5, at the time of injury, at the time of fracture union, at the one-year follow-up evaluation, and at the final follow-up evaluation. With this method, pelvic asymmetry is determined by the difference in length (in centimeters) between two diagonal lines drawn from the border of the sacroiliac joint to the contralateral triradiate cartilage5. Asymmetry was correlated with a deformity index (the difference in the diagonal lengths divided by the sum of the diagonal lengths) to ensure that there was minimal error due to pelvic rotation on the anteroposterior radiograph. The reported normal anatomic variation is 4 mm of asymmetry5.
Patients who met the inclusion criteria underwent a standardized clinical and radiographic evaluation. Physical examination consisted of a detailed neurological examination, including assessment of deep tendon reflexes, sensory testing, manual strength assessment, evaluation for the Babinski sign, and clonus testing13,14. Additional, non-neurological tests included measurement of the range of motion of the hips, knees, ankles, and lumbar spine; evaluation of hip abductor strength (the Trendelenburg sign); the Adam forward-bend test for scoliosis; palpation of the pelvis; and observation of gait for limp13,14. Results were recorded for each patient. Anteroposterior, inlet, and outlet radiographs were assessed for all patients. Judet radiographs were made when a patient had an associated acetabular fracture, in order to assess the degree and effect of posttraumatic osteoarthritis on functional outcome. Reduction of the acetabular fracture was graded with the method described by Matta15, which involves assessment of the maximum displacement of the articular fracture fragments on postoperative anteroposterior and Judet radiographs. The largest displacement is recorded in millimeters. An anatomic reduction is 0 to 1 mm of displacement, an imperfect reduction is 2 to 3 mm of displacement, and a poor reduction is ≥4 mm of displacement or surgical secondary congruence (articular congruity with non-anatomic reduction of the joint surface to the adjacent pelvis).
The Short Musculoskeletal Function Assessment Questionnaire (SMFA) was used to evaluate functional status16. The SMFA is composed of multiple questions that are categorized into two main measures: bother and dysfunction. The lower the score on the SMFA for a given measure, the better the patient's perceived outcome for that measure. The bother and dysfunction measures are calculated on the basis of the patient's answers, and scores for the same patient can be compared over time or scores can be compared between patients. Scores of 0 indicate no bother or dysfunction, and scores of 100 indicate maximum bother and dysfunction. Previous investigators reported mean scores of 28.3 points for dysfunction and 31.8 points for bother in a follow-up study of adults who had sustained an acute fracture of the lower extremity16.
No formal treatment protocol for unstable pelvic fractures was utilized during the study period. However, the same guidelines for assessment and treatment were used at both institutions; they included complete primary and secondary survey examinations as described in the Advanced Trauma and Life Support (ATLS) system17,18, computed tomography scanning of the abdomen and pelvis, angiography of the pelvic arterial system with embolization for persistent hemodynamic instability, and abdominal exploration if indicated. Associated extremity fractures were splinted immediately and were fixed definitively when the patient was hemodynamically stable.
Treatment was based on the attending surgeon's preference. All unilateral pelvic ring disruptions were treated with attempted closed reduction with use of direct manual compression over the anterior superior iliac spine and longitudinal traction with countertraction. A bilateral long leg hip spica cast extending to below both knees was applied if the closed reduction was judged to be successful. The spica cast was applied in the operating room with the patient under general anesthesia, with a spica-cast table placed on a radiolucent operating table. The final position of the affected hip, which was based on the appearance on intraoperative fluoroscopy, was usually in neutral with 10° to 20° of flexion and slight internal rotation. If there was >5 mm of vertical or anteriorposterior displacement, as determined on the pelvic inlet and outlet radiographs, closed or open reduction was attempted. If the vertical or anterior-posterior displacement of >5 mm persisted, then external fixation and/or open reduction and internal fixation was performed. Unilateral pelvic fractures with an associated acetabular fracture that was displaced >2 mm were treated operatively with open reduction and internal fixation. Bilateral pelvic ring fractures were treated with open or closed methods. When pelvic external fixation was used, pins were placed in the iliac crests adjacent to the anterior superior iliac spines. In selected cases, percutaneous sacroiliac screws, as described by Routt et al.19, were placed with fluoroscopic guidance over a guidewire from the posterior aspect of the ilium into the body of the first sacral ala.
All patients were seen at one, three, six, and twelve months following discharge from the hospital and then, if possible, yearly thereafter.
Of 230 pelvic fractures treated during the study period, twenty-three patients (twenty-three fractures) met the inclusion criteria as they had a partially stable or an unstable fracture (Tile type B or C) and open triradiate cartilages. Three patients, all with a unilateral type-B fracture that was treated nonoperatively, were deported to Mexico, at a mean of 2.7 years (range, 1.9 to 3.3 years) following the injury. They were not available for evaluation and were therefore excluded from the study. Thus, twenty (87%) of the twenty-three patients were available for evaluation. Their mean age was 9.5 years (range, 2.7 to 12.8 years) at the time of injury and 16.0 years (range, 11.3 to 27.2 years) at the time of final follow-up. The mean duration of follow-up was 6.5 years (range, 1.6 to 14.8 years), and ten patients were followed for more than seven years. Four patients had a type-B pelvic fracture, and sixteen, including six with an associated acetabular fracture, had a type-C pelvic fracture. All six injuries with an acetabular fracture were type C3. All patients reported that they were in good health. The clinical and radiographic data and data on the associated injuries for the twenty patients in the study are listed in the Appendix.
Pelvic asymmetry, which was evaluated radiographically from the time of the injury through the time of the last examination, did not remodel in any of the twenty patients regardless of their radiographic age at the time of the injury. The mean pelvic asymmetry was 2.8 cm (range, 1.1 to 6.0 cm) at the time of injury, 1.5 cm (range, 0.5 to 5.2 cm) at one year after the injury, and 1.5 cm (range, 0.3 to 5.0 cm) at the time of union and the final follow-up. Eleven patients (55%) had ≤1.0 cm of pelvic asymmetry at the time of union and final follow-up, and three of those patients had 0.4 cm of asymmetry. The remaining nine fractures had ≥1.1 cm of asymmetry at the time of union and final follow-up. The mean asymmetry associated with the type-B fractures was 2.1 cm at the time of injury, 1.4 cm at the time of union, 1.2 cm at one year, and 1.4 cm at the time of the last follow-up. The respective values for the mean asymmetry associated with type-C fractures were 3.5 cm, 1.5 cm, 1.9 cm, and 1.5 cm. With the numbers available, no significant change (increase or decrease) in pelvic asymmetry was noted after union. Table I shows the mean asymmetry according to fracture type. There was a correlation between the measured asymmetry and the index deformity measurements on the anteroposterior pelvic radiographs for all fractures (r = 0.6 and p = 0.04).
Of the patients with asymmetry after treatment, two had been treated with closed reduction with manipulation and a spica cast, seven had had external fixation alone, six had had temporary external fixation with posterior internal fixation (Figs. 2-A, 2-B, and 2-C), and two had been treated with open reduction and internal fixation.
Fractures treated with closed reduction with manipulation and a spica cast were associated with less asymmetry or displacement, as measured on radiographs made at the time of injury and during the study period, than were fractures treated with surgical stabilization. Moreover, radiographs showed a 48% reduction in mean pelvic asymmetry (from 2.9 cm to 1.5 cm) after surgical stabilization compared with an 18% reduction (from 1.7 cm to 1.4 cm) after closed reduction, manipulation, and application of a spica cast (Table II).
Asymmetry was compared between two surgical treatment groups: those treated with external fixation only (seven patients) and those treated either with (1) external fixation as well as open reduction and internal fixation or percutaneous fixation of the sacroiliac joint or with (2) open reduction and internal fixation of the symphysis and open or percutaneous fixation of the sacroiliac joint (eleven patients) (Table III). The asymmetry immediately after the injury did not differ between the two groups (p > 0.05), but there was a significant difference between the groups with regard to asymmetry at the time of union and final follow-up (p < 0.05). With the numbers available, no significant difference was found when we compared specific forms of fixation with external fixation alone or when we compared specific forms of fixation on the basis of the Tile fracture type.
At the time of union, the mean asymmetry or displacement seen on the radiographs had decreased by 76% (from 2.9 cm to 0.7 cm), compared with that seen on the radiographs made at the time of injury, in the group treated with open or percutaneous fixation (with or without external fixation) and by 16% (from 3.2 cm to 2.7 cm) in the group treated with external fixation alone (Table III). There was no increase or decrease in asymmetry after union, at one year or at the time of final follow-up, regardless of treatment.
Results Based on Fracture Type (Tile Classification)
The four patients with a Tile type-B fracture had a mean MISS of 26.0 points; three of these patients required a blood transfusion, and two required pelvic arterial embolization. Two patients were treated with external fixation alone and two, with external fixation and posterior sacroiliac open reduction and internal fixation. The mean asymmetry values are shown according to the Tile fracture type in Table I. Of the four patients with a type-B fracture, the two treated with external fixation alone had had a mean pelvic asymmetry of 1.6 cm (1.1 and 2.1 cm) at the time of injury whereas the two treated with external fixation and posterior open reduction and internal fixation had had a mean asymmetry of 2.3 cm (1.4 and 3.1 cm). The mean pelvic asymmetry after union and at the time of final follow-up was 1.1 cm (0.9 and 1.3 cm) in the two patients treated with external fixation alone and 1.7 cm (0.9 and 2.4 cm) in the other two patients.
The sixteen patients with a Tile type-C fracture had a mean MISS of 28.7 points (range, 11 to 50 points). Twelve patients required blood transfusion (mean, 6.3 units; range, four to fourteen units), and four had pelvic arterial embolization. Fourteen patients were treated operatively. Five patients who were treated with external fixation alone had a mean pelvic asymmetry of 3.5 cm (range, 1.4 to 6.0 cm) at the time of the initial injury and 3.3 cm (range, 1.5 to 5.0 cm) at the time of union and final follow-up. Nine patients who were treated with open or percutaneous fixation, with or without external fixation, had a mean pelvic asymmetry of 3.9 cm (range, 2.5 to 5.8 cm) at the time of the initial injury and 0.6 cm (range, 0.3 to 1.2 cm) at the time of union and final follow-up. Two patients with a unilateral type-C fracture were treated nonoperatively, with closed reduction and application of a bilateral long leg hip spica, and they had a mean pelvic asymmetry of 1.7 cm (1.6 and 1.8 cm) at the time of the initial injury and 1.4 cm (0.9 and 1.8 cm) at the time of union and final follow-up.
Six of the sixteen patients with a Tile type-C fracture had an associated acetabular fracture. All were type-C3 fractures. Four of the six patients were treated with open reduction and internal fixation. Two of the six acetabular fractures were separations of the triradiate cartilage with <2 mm of displacement. One of the fractures associated with a separation of the triradiate cartilage was treated nonoperatively. The other was treated with anterior external fixation with percutaneous sacroiliac screw fixation, with the acetabular fracture treated nonoperatively. According to the classification system of Matta15, four of the six acetabular fractures had an anatomic reduction and the other two had a nearly anatomic reduction at the time of union. However, all of the reductions of the injuries to the triradiate cartilage appeared anatomic, without evidence of degenerative osteoarthritis, at the time of final follow-up. Premature closure of the triradiate cartilage was not observed in any patient.
The functional assessment was conducted with the SMFA questionnaire. The eleven patients with ≤1 cm of pelvic asymmetry at the time of final follow-up had a mean dysfunction index score of 6.9 points (range, 0 to 13.6 points) and a mean bother index score of 8.0 points (range, 0 to 15.6 points), compared with 36.5 points (range, 23.4 to 53.2 points) and 40.8 points (range, 28.4 to 58.8 points) for the nine patients with ≥1.1 cm of asymmetry (p < 0.05).
A MISS of ≥26 points has been reported to be associated with increased morbidity in children with a pelvic fracture2. The patients in this study who had a MISS of ≤25 points had mean dysfunction and bother index scores of 22 and 25 points, compared with a mean of 19 and 21 points, respectively, for the patients with a MISS of ≥26 points (p = 1).
There was a positive correlation (r = 0.6 and p = 0.04) between the displacement and asymmetry at the time of the injury and the SMFA score. There was a significant difference between the SMFA scores of the patients with ≥1.1 cm of asymmetry at the time of final follow-up and the SMFA scores of those with ≤1 cm of asymmetry at the time of final follow-up (p = 0.04). However, with the numbers available, there was no significant association between the SMFA score and the fracture type (B or C) or the MISS.
Results of Clinical Assessment
All twenty patients had a complete final radiographic and clinical assessment. Eleven had ≤1 cm of pelvic asymmetry and no lumbar pain, Trendelenburg sign, or lumbar scoliosis. However, two patients, each with 0.9 cm of pelvic asymmetry, had occasional sacroiliac joint pain and mild tenderness to palpation over the sacroiliac joint. Pelvic radiographs of these two patients showed in situ arthrodesis of the affected sacroiliac joint; one patient had been treated with percutaneous sacroiliac screw fixation, and the other had been treated nonoperatively.
All nine patients with ≥1.1 cm of pelvic asymmetry had three or more of the following: nonstructural scoliosis (nine patients), lumbar pain (eight patients), sacroiliac joint pain and tenderness (six patients), and the Trendelenburg sign and gait (four patients) on physical examination. There was a positive correlation between low-back symptoms (lumbar pain, sacroiliac joint pain and tenderness on examination, and nonstructural scoliosis) and pelvic asymmetry of ≥1.1 cm (r = 0.7 and p = 0.03).
No focal neurologic findings were identified in any patient. Five of the nine patients with a MISS of ≤25 points, compared with two of the eleven patients with a MISS of ≥26 points, had no symptoms or abnormalities on physical examination of the pelvis at the time of follow-up. This difference was not significant (p = 0.08). There was no correlation between the MISS and low-back or other physical symptoms at the time of follow-up (r = 0.1 and p = 0.7).
Four of the eighteen patients who were treated operatively had a complication. These included two pin-site infections, which resolved with more frequent cleaning and administration of oral antibiotics; one deep wound infection, treated with débridement and three days of intravenous antibiotics; and one case of osteomyelitis of the pubic symphysis that developed after open reduction and internal fixation and resolved with débridement and six weeks of intravenous antibiotics. There were no deaths.
Unstable pelvic ring disruptions are uncommon in skeletally immature patients1,3,9,11,12. Most pelvic fractures in children are the result of an automobile-pedestrian or motor-vehicle accident and, like adult pelvic ring disruptions, they are often associated with many injuries1,3,9,11,12. Additionally, there is a high potential for morbidity or mortality as a result of the pelvic ring disruption and/or the associated injuries3-5,9,10,20,21. The resultant hemodynamic instability often necessitates blood transfusion. However, the effect of associated injuries on the outcome in children is not completely known2,10.
Long-term morbidity has been correlated with the number of associated injuries (the MISS) at presentation and with the residual pelvic asymmetry1,3,9,11,12. Previous studies have included both unstable and stable pelvic fracture patterns as well as patients who were skeletally mature adolescents at the time of injury1-5,7-9. Functional outcomes were not evaluated with a validated outcome measure. In the present study, patients with a MISS of ≤25 points had slightly, but not significantly, better functional scores on the SMFA compared with patients with a MISS of ≥26 points. We found no correlation between the SMFA score and the MISS.
We measured musculoskeletal function with the SMFA, which had been previously validated for use in adults, eighteen to eighty-nine years of age, with musculoskeletal injuries16. The majority of our patients were adolescents or adults at the time of final follow-up; only one patient was less than twelve years of age. Clinical findings such as low-back and sacroiliac symptoms were uniformly more common in patients with, and correlated statistically with, asymmetry of ≥1.1 cm. Additionally, all patients who had worse scores on the bother and dysfunction indices had more low-back and sacroiliac pain.
Asymmetry of ≥1.1 cm was associated with functional disability as shown by higher scores for the dysfunction and bother indices of the SMFA. Patients with such asymmetry presented clinically with low-back pain, a Trendelenburg gait, sacroiliac pain and tenderness, and nonstructural scoliosis. Similar clinical findings have been reported previously1,9. Hargitai et al. found poor results, including pain, limb-shortening, and scoliosis, three to six years following nonoperative treatment of unstable pelvic fractures with skeletal traction9. They recommended posterior stabilization for displaced fractures. Schwarz et al. found that residual pelvic asymmetry correlated with low-back pain and recommended ensuring symmetric alignment by whatever means possible1. The best clinical and functional results in the present study were in patients with ≤1 cm of pelvic asymmetry.
In our study, the SMFA score and the findings of the clinical examination both correlated with the degree of pelvic asymmetry. We measured pelvic asymmetry with a previously described technique5, the advantage of which is that it takes into account multiplanar deformity; a maximum of 4 mm of asymmetry is considered normal. The disadvantage of the technique is that care must be taken to ensure that a true anteroposterior radiograph of the pelvis is made. The ratios utilized in the deformity index formula minimize error from obliquity.
On the basis of the findings in the small number of patients treated in this study, unstable fractures associated with pelvic asymmetry in children with open triradiate cartilages did not appear to remodel to a more symmetric alignment. In fact, asymmetry appeared to increase for up to one year after the injury in the nonoperatively treated patients and in many of those who had been treated operatively (Table II). Keshishyan et al. noted that, in three of the thirty-one children in their study, the deformity index decreased over the one-year period following the pelvic fracture5. In contrast, none of the twenty-three patients in our series had a decrease in either the asymmetry or the deformity index during the follow-up period. We did, however, note good agreement between the measured asymmetry and the deformity index (r = 0.6 and p = 0.04).
Long-term morbidity and functional problems may be minimized by restoring pelvic symmetry. Pelvic ring fractures (Tile type B or C) with more displacement or asymmetry at the time of injury had, on the average, a greater decrease in pelvic asymmetry when they were treated with surgical stabilization than when they underwent nonoperative stabilization. However, asymmetry of ≤1 cm was not always achieved with surgical stabilization, especially when external fixation was utilized as the sole treatment (the seven patients who were so treated had a mean of 2.7 cm of asymmetry at the time of follow-up). Moreover, radiographic evaluation demonstrated significantly less pelvic asymmetry in the patients treated with combined anterior and posterior forms of fixation than in those treated with anterior external fixation alone. These findings were consistent throughout the treatment and follow-up period. We are unable to make recommendations regarding a specific technique of operative treatment on the basis of our results. However, anterior and posterior fixation should be considered at the time of reduction to achieve a symmetric pelvis and limit asymmetry at the time of union to ≤1.0 cm.
Our study is limited by the relative rarity of unstable fractures in patients with open triradiate cartilages. Because of the small number of patients, we could not compare treatments or make specific recommendations regarding management of specific fracture patterns. However, a significant difference was found between the functional (SMFA) scores for patients with ≥1.1 cm of pelvic asymmetry and those for patients with ≤1 cm of pelvic asymmetry at the time of the final follow-up. Correlation of the findings of the physical examination, such as low-back symptoms, with asymmetry of ≥1.1 cm at the time of the final follow-up suggests that asymmetry may have a direct effect on functional (SMFA) outcome. Conversely, the magnitude of displacement or asymmetry at the time of injury may affect functional scores, as shown by the correlation of worse scores with greater asymmetry, and may confound the effect of ≥1.1 cm of asymmetry at the time of final follow-up on the SMFA scores at that time. Although patients with a MISS of ≤25 points had fewer physical symptoms, no correlation was found between the MISS and pelvic asymmetry at the time of injury or final follow-up or the SMFA score.
The pelvic ring did not remodel in any patient, regardless of the age at injury. Both closed and operative treatment appeared to reduce asymmetry, and the least asymmetry occurred in the patients who had had anterior and posterior stabilization. Authors of previous studies of pelvic fractures in children have drawn a variety of conclusions regarding optimal treatment. The cohorts in those studies were also small and included either isolated acetabular fractures or adolescents with closed triradiate cartilages1,3,9,11,12. Patients with closed triradiate cartilages at the time of injury were excluded from the present study. Pelvic fractures with an associated acetabular fracture were included, but an anatomic or nearly anatomic reduction of the acetabular fracture was achieved in each case. Clinical and radiographic follow-up revealed no evidence of degenerative osteoarthritis or dysfunction of the hip joint. Therefore, the functional outcome deficits that we noted most likely were the result of the pelvic injury and not the associated acetabular fracture.
A standardized protocol including early diagnosis, resuscitation, and aggressive and prompt treatment of associated injuries has been shown to be effective in reducing morbidity and mortality in adults with an unstable pelvic fracture and may also be effective in children with such injuries22. The initial asymmetry or displacement at the time of injury correlated with worse scores on the SMFA, and worse SMFA scores were also noted in patients with ≥1.1 cm of asymmetry at the time of final follow-up. Patients who had ≤1.0 cm of asymmetry had better findings and fewer symptoms on clinical examination. However, the effect of operative treatment on functional outcome is not clear, given the small number of patients. Larger, prospective studies will help to clarify this issue as well as the impact of initial displacement and residual asymmetry on functional outcome.
Tables showing the Tile classification of pelvic ring disruptions, the clinical and radiographic data for the patients, and the associated injuries are available with the electronic versions of this article, on our web site at jbjs.org (go to the article citation and click on “Supplementary Material”) and on our quarterly CD-ROM (call our subscription department, at 781-449-9780, to order the CD-ROM). ▪
NOTE: The authors thank Zachery Lahti, BA, for his help in the initial data collection.
The authors did not receive grants or outside funding in support of their research or preparation of this manuscript. They did not receive payments or other benefits or a commitment or agreement to provide such benefits from a commercial entity. No commercial entity paid or directed, or agreed to pay or direct, any benefits to any research fund, foundation, educational institution, or other charitable or nonprofit organization with which the authors are affiliated or associated.
Investigation performed at the Department of Orthopaedic Surgery, Denver Health Medical Center, and The Children's Hospital, Denver, Colorado