 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 2004 The Journal of Bone and Joint Surgery, Inc. What's New in Orthopaedic Trauma
1 Department of Orthopaedic Surgery, Regions Hospital, 640 Jackson Street, St.
Paul, MN 55101. E-mail address:
peter.a.cole{at}healthpartners.com.
Specialty Update has been developed in collaboration with the Council of
Musculoskeletal Specialty Societies (COMSS) of the American Academy of
Orthopaedic Surgeons.
Since its inception in 1990, the term "evidence-based medicine" has evolved to include evidence-based nursing, physiotherapy, occupational therapy, podiatry, critical care, and surgery. Surgical subspecialties, such as orthopaedics, have led the way in a paradigm shift favoring evidence-based medicine over traditional approaches1. The most sophisticated practice of evidence-based medicine requires, in turn, a clear delineation of relevant clinical questions, a thorough search of the literature relating to these questions, a critical appraisal of available evidence and its applicability to the clinical situation, and a balanced application of the conclusions to the clinical problem. The balanced application of the evidence (i.e., clinical decision-making) is the central point of practicing evidence-based medicine and involves, according to the principles of evidence-based medicine, integration of our clinical expertise and judgment with patient and societal values and the best available research evidence (Fig. 1)2.
A common misunderstanding and ultimate misapplication of evidence-based medicine is the assumption that evidence-based medicine equals randomized trial evidence. Relying solely on research evidence from randomized trials is problematic for several reasons: (1) fewer than 14% of published scientific articles in The Journal of Bone and Joint Surgery are randomized trials3, (2) observational study designs are overlooked as potentially valuable sources of information to guide practice4, and (3) patients' preferences, clinical circumstances, and surgeon expertise are undervalued in the decision-making process. The use of the literature in orthopaedics requires a fundamental understanding of the hierarchy of evidence, from meta-analyses of high-quality randomized trials showing definitive results that are directly applicable to individual patients to a reliance on physiologic rationale or previous experience with a small number of similar patients. The hallmark of the surgeon who practices evidence-based medicine is that, for particular clinical decisions, he or she knows the strength of the evidence and therefore the degree of uncertainty. Ultimately, we must weigh the current "best" evidence with our previous experience, our skills, the preferences of our patients, and our clinical situation (i.e., our hospital setting and resources). However, understanding the association between study design and level of evidence remains important. As of January of 2003, The Journal of Bone and Joint Surgery has published the level of evidence associated with each scientific article to provide readers with a gauge of the validity of the study results3. In the hierarchy of evidence, classifications are based on the validity of each study design. Thus, designs that limit bias to the greatest extent are at the top of the pyramid and those that are inherently biased are at the bottom (Fig. 2).
According to the traditional paradigm, clinicians evaluate and solve clinical problems by reflecting on their own clinical experience, by assessing the underlying biology and pathophysiology of the problem, or by consulting a textbook or local expert. For many traditional practitioners, reading the Introduction and Discussion sections of a research article is adequate for gaining relevant information, and observations from day-to-day clinical experience are a valid means of building and maintaining knowledge about patient prognosis, the value of diagnostic tests, and the efficacy of treatment. Because this paradigm of "eminence-based practice" places high value on traditional scientific authority (i.e., opinion leaders in the field) and adherence to standard approaches, traditional medical training and common sense provide an adequate basis for evaluating new tests and treatments, and content expertise and clinical experience are sufficient to generate guidelines for clinical practice5. Evidence-based practice posits that while pathophysiology and clinical experience are necessary, they alone are insufficient guides for practice. The evidence-based-medicine approach includes several additional steps. These steps include using experience to identify important knowledge gaps and information needs, formulating answerable questions, identifying potentially relevant research, assessing the validity of evidence and results, developing clinical policies that align research evidence and clinical circumstances, and applying research evidence to individual patients with their particular experiences, expectations, and values. Practicing evidence-based medicine is not easy. Not all surgeons have time to search the literature for the best available external clinical evidence from systematic research. Surgeons then look to preappraised resources to guide their clinical practice. In response to the need for evidence summaries, The Journal publishes the Evidence-Based Orthopaedics, User's Guide to the Orthopaedic Literature, and Specialty Update reviews to provide clinicians with the tools that they need to critically appraise the methodological quality of individual studies as well as systematic reviews of the current evidence. Indeed, we reported that a rigorous systematic review can receive more than twice the number of citations compared with other traditional (i.e., nonsystematic, narrative) reviews (mean, 13.8 compared with 6.0, p = 0.008)6. In this Specialty Update on Orthopaedic Trauma, we used approaches to systematically identify most articles. Such a systematic review is defined by several criteria: (1) a specific health-care question, (2) a comprehensive, transparent, and reliable search strategy, (3) an assessment of study validity, (4) the inclusion of relevant outcome measures, and (5) an evaluation of study heterogeneity (i.e., differences in the treatment effect across different studies)7. We conducted both database and manual searches of all clinical studies with a focus on orthopaedic trauma that were published from May 2003 through May 2004. Specifically, we searched the Cochrane Database, the McMaster University database, five orthopaedic journals, one general trauma journal, and four high-impact medical journals (Table I). Of the 8315 papers that were reviewed, we identified 156 potentially eligible studies. The complete abstracts of these 156 studies were reviewed in duplicate for final inclusion. After the review of the abstracts, forty-six studies were included in the present review. Consensus on the final studies was achieved. In the present review, we summarize the salient findings of eleven level-I studies, seven level-II studies, three level-III studies, twenty-three level-IV studies, one level-V study, and one reliability study across subspecialties of orthopaedic trauma. Of these, thirty-three studies pertained to therapy, nine involved prognosis, three evaluated diagnostic tests, and one was a reliability study.
The current evidence in studies on orthopaedic trauma that were published from May 2003 through May 2004 has been summarized according to the type of study, the level of evidence, and the sample size (Table II). When managing patients, surgeons often are confronted with dilemmas about a treatment (therapy), the outcome of a disease process (prognosis), or the utility of a diagnostic test (diagnosis)1. Evidence summaries are most helpful to surgeons and clinicians when they provide answers to relevant questions in clinical practice, such as "Does one treatment offer a significant advantage over another one?" (a question about therapy), "How confident can I be about the diagnosis if this test is positive?" (a question about a diagnosis), or "Is my patient at risk for a poor outcome following surgery?" (a question about prognosis). Table II provides such a summary in context with the hierarchy of evidence. Readers also should note that level-I studies do not always represent a definitive result suggesting that no further research in the area is required. In most circumstances, further research is necessary to confirm or refute the findings of a level-I study, to improve the generalizability to other patient populations, and to further explore findings for the subgroups in such studies.
In addition to the systematic review, we included eleven papers that were initially presented at the 2003 Annual Meeting of the Orthopaedic Trauma Association in Salt Lake City, Utah. These papers were subsequently designated as OTA Highlight Papers at the 2003 Annual Meeting of the American Academy of Orthopaedic Surgeons in San Francisco, California.
From Psychological Impact to Postoperative Complications Orthopaedic trauma surgeons have historically focused most acutely on the discipline of fixing fractures. More recently, there has been a broadening of focus, in both investigation and practice, from the narrow technical aspects of surgery and treatment to the realms of biology and function. This widerangled emphasis was underscored by a couple of articles that explored the emotional and psychological consequences of orthopaedic injury. Investigators from the Lower Extremity Assessment Project (LEAP) studied the psychological distress associated with a severe lower-limb injury8. Of the 569 patients who had been admitted to one of eight level-I trauma centers with a unilateral lower limb injury that either resulted in traumatic amputation or was at risk for therapeutic amputation, 545 (96%) completed at least one Brief Symptom Inventory (BSI) within two years and 385 (68%) completed four. Forty-eight percent of the patients screened positive for a likely psychological disorder at three months after the injury, and this percentage remained high (42%) at twenty-four months. Two years after the injury, almost one-fifth of the patients reported severe phobic anxiety and/or depression. Factors that were associated with a likely psychological disorder included poorer physical function, younger age, nonwhite race, poverty, a likely drinking problem, neuroticism, a poor sense of self-efficacy, and limited social support. Relatively few patients reported receiving any mental health services after the injury (12% at three months and 22% at twenty-four months). In another study, performed by Holbrook and Hoyt as an extension of the Trauma Recovery Project, gender differences emerged in the psychological manifestations of trauma9. The Trauma Recovery Project is a prospective, epidemiologic study from the San Diego Regionalized Trauma System that is designed to examine certain outcomes after major trauma, such as quality of life, as well as psychologic sequelae of trauma, such as depression and early symptoms of acute stress reaction. In the study by Holbrook and Hoyt, which included 313 women and 735 men, women were significantly more likely to have poorer quality-of-life outcomes than men at both twelve months (odds ratio = 2.2, p < 0.001) and eighteen months (odds ratio = 2.0, p < 0.001). Quality of well-being scores at the six, twelve, and eighteen-month follow-up time-points were markedly and significantly lower for women than for men, independent of injury severity, serious and moderate injury status, lower extremity injury, intentional or unintentional injury type, and blunt or penetrating injury. Women also were significantly more likely to have early combined depression and symptoms of acute stress reaction at the time of discharge (odds ratio = 1.7, p < 0.01) and to have continuous depression throughout the eighteen-month follow-up period (odds ratio = 2.3, p < 0.001). Moving on to studies on the biological manifestations of treatment, Pape et al.10 investigated the impact of intramedullary nailing as opposed to temporary external fixation for the treatment of femoral fractures in severely injured patients. This theme of "damage-control" orthopaedic surgery recently has been advocated for the treatment of femoral shaft fractures in severely injured patients because surgical procedures have been found to represent a second-hit phenomenon regarding the operative burden. In this prospective, randomized, multicenter study, Pape et al. attempted to determine the operative burden by evaluating the perioperative concentrations of proinflammatory cytokines (interleukin-1, interleukin-6, and interleukin-8) in serial samples of central venous blood. A significant and sustained inflammatory response was measured after intramedullary nailing performed within twenty-four hours but not after initial external fixation or after secondary conversion to an intramedullary implant. These findings suggest that damage-control orthopaedic surgery appears to minimize the additional physiologic impact induced by acute intramedullary fixation of the femur. Part of this work was presented and distinguished as one of the OTA Highlight Papers at the 2003 meeting of the AAOS that was held in San Francisco11. Measuring injury severity is a difficult and inexact science, particularly as it relates to the relative burden of fractures. It has long been thought that fractures have been undervalued by the Injury Severity Score (ISS) with regard to their influence on patient sickness. Balogh et al.12 compared the scoring efficacy of the Injury Severity Score with that of the New Injury Severity Score (NISS) for predicting an extended length of stay in the hospital and admission to the intensive care unit in order to determine the effect of multiple orthopaedic injuries. In that prospective study of 3100 patients with multiple orthopaedic injuries, 7.5% of the patients had a higher NISS than ISS. These patients spent more days in the hospital (twenty-two compared with eight; p < 0.001), spent more days in the intensive-care unit (3.4 compared with 0.1; p < 0.001), and had a higher mortality rate (8% compared with 1.2%; p < 0.001) compared with patients with identical NISS and ISS scores but without multiple orthopaedic injuries. The NISS was found to be more predictive of longer length of stay (ten days or more) (p < 0.0001) and admission to the intensive care unit (p < 0.0001). The authors emphasized that the recognition of relative injury severity in this high-risk group is not possible with use of the traditional ISS and that the NISS is easier to calculate and has better predictive power. Complications of injury and treatments rendered after fracture surgery have probably been underreported in the literature for a number of reasons. During the past year, postoperative infection, heterotopic ossification, and deep venous thrombosis were three such complications that gained substantial attention. The Multidisciplinary Alliance against Device-Related Infections is an organization that was established in 2002 to organize groups of experts for the purpose of developing guidelines for treatment. In a Current Concepts article that was published in the New England Journal of Medicine this past year, a member of this group reported on the treatment of infections associated with surgical implants13. The author estimated that 2,000,000 fracture fixation devices (including internal and external fixation devices) are implanted annually in the United States and cited a postoperative infection rate of 5% (2.5 times that associated with joint prostheses). Certainly, a relatively greater proportion of these infections occur after open fractures. Gosselin et al.14, in a Cochrane Database Systematic Review, attempted to quantify the evidence on the effectiveness of antibiotics in the initial treatment of open fractures of the limbs. From a review of randomized or quasi-randomized controlled trials involving patients of any age with open fractures of the limbs, 913 participants in seven studies were identified and analyzed. The use of antibiotics had a protective effect against early infection when compared with no antibiotics or placebo (relative risk = 0.41, 95% confidence interval = 0.27 to 0.63; absolute risk reduction = 0.08, 95% confidence interval = 0.04 to 0.12; number needed to treat = 13, 95% confidence interval = 8 to 25). Two studies that were recognized as 2003 OTA Highlight Papers underscored the importance of the investigation of infection, particularly in patients with open fractures. In the first study, which was another extension of the LEAP study, Pollak et al.15 analyzed the data on infections for 315 patients with severe open extremity fractures. The overall infection rate was 27% within the first three months after injury. Interestingly, there were no time-related differences in the infection rate when the authors analyzed the period between the injury and the first débridement, the period between admission and the first débridement, and the period between the first débridement and definitive wound coverage. However, a significant difference was observed when the authors analyzed the period between the injury and admission to a level-I trauma center. The rate of infection was 22.2% among patients who were admitted within six hours after the injury, compared with 38.9% among patients who were admitted more than six hours after the injury (p < 0.01). These findings indicate that certain qualities of the interventions received at these centers have a positive influence on the infection rates in patients with severe open fractures of the extremities. Such findings are certainly relevant because once a bone infection has been established, it is difficult to eradicate. The second OTA Highlight Paper elucidated why this may be the case16. In a basic-science project designed to analyze antibiotic resistance in patients with staphylococcal osteomyelitis, a group of investigators from Wake Forest Medical Center used a human osteoblast culture model to study the penetration of bacteriocidal and bacteriostatic antibiotics in osteoblasts infected with Staphylococcus aureus. At twelve hours after infection (and intracellular penetration), the bacteriocidal effects of antibiotics were negated by some cellular defense mechanism. These results help to explain the development of antibiotic resistance and the chronic and unpredictable nature of osteomyelitis. Much like osteomyelitis, heterotopic ossification is a difficult complication to treat after musculoskeletal trauma and thus has also spawned substantial research on its prevention. Karunakar et al.17 performed a prospective, randomized, double-blind, placebo-controlled trial on the efficacy of indomethacin as prophylaxis against heterotopic ossification after operative treatment of acetabular fractures. One hundred and seven of 125 patients who had undergone surgical treatment of an acetabular fracture through a posterior approach were available for long-term follow-up. Sixty-one patients were randomized to receive a placebo and forty-six were randomized to receive 75 mg of slow-release indomethacin, in a blinded fashion, once per day for six weeks. There was no significant difference between the indomethacin group and the placebo group with regard to the overall rate of Brooker class-III and IV heterotopic ossification (17% compared with 21%). Unfortunately, the rate of compliance in the indomethacin group was poor, with 40% of the patients having no detectable serum level of the indomethacin at two weeks. In addition, the indomethacin group had a higher dropout rate because of treatment complications. Further complicating the decision of whether or not to provide prophylaxis against heterotopic ossification is the fact that nonsteroidal anti-inflammatory drugs have been implicated in the retardation of fracture-healing, although strong evidence in support of this notion has been scant in the clinical literature. A recent study demonstrated indirect evidence in support of this clinical concern18. Two hundred and eighty-two high-risk patients who had had open reduction and internal fixation of an acetabular fracture were randomized to receive either radiation therapy or indomethacin in a protocol designed to assess the relative efficacy of these treatments. One hundred and twelve patients had sustained at least one concomitant fracture of a long bone. Thirty-six patients had no prophylaxis, thirty-eight received focal radiation, and thirty-eight received indomethacin. Overall, sixteen nonunions developed in fifteen patients. When patients who had received indomethacin were compared with those who had not, a significant difference was noted with regard to the rate of long-bone nonunion (26% compared with 7%, p = 0.004). One of the most inconclusive areas of clinical investigation is the diagnosis and treatment of deep venous thrombosis after skeletal trauma. Stannard et al.19, in another 2003 OTA Highlight Study, helped to sort out the appropriate roles of ultrasound and magnetic resonance venography in the diagnosis of deep venous thrombosis after skeletal injury. In a study of 322 patients who had been enrolled in two deep venous thrombosis prophylaxis protocols, they found a high false-negative rate (77%) in association with the use of ultrasound for the detection of pelvic deep venous thromboses and a high false-negative rate (40%) in association with the use of magnetic resonance venography for the detection of lower extremity deep venous thromboses. On the contrary, magnetic resonance venography was highly sensitive for the detection of pelvic deep venous thromboses (with thirteen of thirteen such lesions being detected), and ultrasound was sensitive for the detection of lower-limb deep venous thromboses (with twenty of twenty-three such lesions being detected). A relative weakness of the study was that the gold standard was assumed to be the two diagnostic tests used in the study; however, the study was clearly instructive regarding the relative roles of each diagnostic test for different anatomic locations.
The most appropriate workup for spinal injury in trauma patients has evolved with the technology used in the emergency department. The helical truncal computed tomographic scanning that is performed to evaluate visceral trauma also yields diagnostic information about the spine. In the study by Hauser et al.20, helical truncal computed tomographic scans were compared with dedicated radiographs of the thoracolumbar spine for the evaluation of 222 consecutive patients who required clinical screening for thoracolumbar injury. The accuracy of computed tomographic scanning of the chest, abdomen, and pelvis for the detection of thoracolumbar fractures was 99% (95% confidence interval, 96% to 100%). The accuracy of radiographs of the thoracolumbar spine was 87% (95% confidence interval, 82% to 92%). The sensitivity, specificity, and positive and negative predictive values were better for computed tomographic scanning of the chest, abdomen, and pelvis than for radiographs of the thoracolumbar spine. No fractures were missed with computed tomographic scanning, whereas radiographs were associated with a misclassification rate of 12.6%. Computed tomographic scanning also yielded far faster spinal "clearance" times (fifty-seven compared with 295 minutes). The authors recommended that computed tomographic scanning of the chest, abdomen, and pelvis should replace plain radiographs for the evaluation of high-risk trauma patients who require thoracolumbar spine screening. Moving on to the cervical spine, great efforts have been made to narrow the indications for a radiographic workup of the alert trauma patient. The Canadian C-Spine (cervical-spine) Rule (CCR) and the National Emergency X-Radiography Utilization Study (NEXUS) Low-Risk Criteria (NLC) are simple algorithms that are used to guide the use of cervical spine radiography for the evaluation of such patients. Stiell et al.21, in a prospective cohort study that was performed at nine Canadian emergency departments, compared the CCR with the NLC for the evaluation of alert trauma patients who were in stable condition. Among 8283 patients, 169 (2.0%) had a clinically important cervical spine injury. The CCR was more sensitive (99.4% compared with 90.7%, p < 0.001) and more specific (45.1% compared with 36.8%, p < 0.001) than the NLC for the detection of these injuries, and its use would have resulted in lower rates of radiography (55.9% compared with 66.6%, p < 0.001). The CCR would have missed one patient with an important injury, and the NLC would have missed sixteen such patients. Thus, the authors concluded that the Canadian C-Spine Rule is a superior tool for the diagnosis of cervical spine injury and that its use would result in reduced rates of radiography in alert patients with trauma. The CCR simply combines the exclusion of certain risk factors (an age of sixty-five years or more, a dangerous mechanism, paresthesias in the extremities, a simple rear-end motor-vehicle collision, a sitting position in the emergency department, an ability to walk at any time, a delayed onset of neck pain, or an absence of midline cervical spine tenderness) with the patient's ability to rotate the neck left and right by 45° to determine the need for radiographs. Moving on to the treatment of spine fractures, Wood et al.22 compared the results of operative treatment (anterior or posterior arthrodesis) with those of nonoperative treatment (application of a cast or orthosis) in a prospective, randomized trial of forty-seven consecutive patients with thoracolumbar burst fractures without neurological deficit. After an average duration of follow-up of forty-four months (minimum, twenty-four months), there was no difference between the groups with regard to return to work or pain scores; however, patients who had been treated nonoperatively reported less disability, and the results of the SF-36 and Oswestry questionnaires revealed certain trends that favored nonoperative treatment. Furthermore, the rate of complications in the operative treatment group (sixteen of twenty-five) was greater than that in the nonoperative treatment group (two of twenty-three). On the low-energy end of the spine-fracture spectrum, Rhyne et al. evaluated the results of minimally invasive balloon reduction followed by polymethylmethacrylate fixation (kyphoplasty) in a study of fifty-two patients with eighty-two thoracolumbar osteoporotic vertebral compression fractures23. All patients had had a failure of nonoperative treatment, and all had documented changes on magnetic resonance imaging scans. After an average duration of follow-up of thirty-seven weeks (range, four to ninety-nine weeks), improved heights of 4.6 and 3.9 mm were measured in the anterior and middle columns, respectively. In addition, the Cobb angle increased by 14% (p < 0.05), the visual analog pain scale score improved by 7 points (p < 0.05), and the Roland-Morris Disability Survey score improved by 11 points (p < 0.05). There were no medical or procedural complications, other than a cement leakage rate of 9.8%.
Validated functional outcome instruments have been widely used for certain skeletal injuries and infrequently used for others. For example, the modified Merle d'Aubigné clinical hip score has been the outcome instrument that typically has been used for the evaluation of acetabular fractures since its principal utilizations by Letournel and Judet in their classic publications. Moed et al.24, in a study of 150 patients with an acetabular fracture, compared the modified Merle d'Aubigné clinical hip score with the Musculoskeletal Function Assessment score after a minimum duration of follow-up of two years. The mean modified Merle d'Aubigné score was 16.8 points (range, 9 to 18 points), and the mean Musculoskeletal Function Assessment score was 24.9 points (range, 0 to 79 points). The Merle d'Aubigné score data were asymmetric, demonstrating a ceiling effect. This finding brought into question the usefulness of this score as a method for evaluating the outcome of treatment of acetabular fractures. The results also provided initial reference values for the evaluation of acetabular fractures with the Musculoskeletal Function Assessment score. The principles of percutaneous pelvic ring treatment underscore the importance of a well-reduced sacrum in the presence of a sacral fracture, both for the safe placement of iliosacral screws and for the maintenance of reduction. Two studies chosen this year underscored the importance of this principle, adding basic-science and clinical ammunition to support the argument. In a study from the University of Texas Southwestern Medical Center, iliosacral screw failure was recorded along with pelvic injury patterns in sixty-two patients25. Residual postoperative displacement and late displacement of the posterior part of the pelvis were measured. The main outcome measure was failure, defined as at least 1 cm of combined vertical displacement of the posterior part of the pelvis compared with the immediate postoperative position. Thirty-two patients had a dislocation or fracture-dislocation of the sacroiliac joint, and thirty had a vertical fracture of the sacrum. Fixation failed within the first three weeks after surgery in four patients, all of whom had a vertical sacral fracture. In two other patients with a vertical sacral fracture, there was evidence that the fracture had only barely been held. The vertical sacral fracture pattern was significantly associated with failure (p = 0.04); the excess risk of failure compared with sacroiliac joint injury was 13% (95% confidence interval, 1% to 25%). There was no significant association between failure and anterior fixation method, iliosacral screw arrangement or length, or any demographic or injury variable. These clinical results were supported by the results of a cadaveric study by Beebe et al.26, one of the 2003 OTA Highlight Papers. In this study, three cranially malreduced pelvic fractures that were displaced by 1 cm and three anatomically reduced sacral fractures (both of which were associated with disruption of the pubic symphysis) were compared with regard to the contact area between fracture fragments and biomechanical stiffness (load-displacement curves). A significant increase in stiffness of 19% was found in association with the anatomically reduced pelves, and a significant association was found between the cross-sectional area and the final stiffness of internally fixed fractures.
Hip fractures present an enormous and expanding burden to the medical system and society at large, so it is appropriate that this type of injury is represented heaviest in this annual review. The fact that this fracture presents an ever-expanding burden is confirmed by an article from the British Medical Journal, in which Roberts and Goldacre evaluated trends in mortality after admission to a hospital for a femoral neck fracture27. From 1968 to 1998, 32,590 patients who were sixty-five years old or more were admitted, and, amazingly, there was no change (decrease) in mortality rates after 1984. A further twist on this theme was elucidated by Richmond et al.28 in a study of 836 patients who had been diagnosed with a hip fracture. The mortality risk was highest within the first three months after the fracture, with standardized mortality ratios approaching that of the control population by two years. The data also demonstrated that a hip fracture was not associated with significant excess mortality among patients who were more than eighty-five years old. Among younger patients, however, those with an ASA classification of 3 or 4 had significantly greater mortality that persisted for as long as two years after a hip fracture. In a prospective cohort study that was published in the Journal of the American Medical Association, 1206 patients with an age of more than fifty years who had been admitted to four New York City hospitals for the treatment of a hip fracture were enrolled in a trial designed to examine the association between the timing of surgical repair and outcome29. Earlier surgery (performed less than twenty-four hours after the injury) was not associated with improved mortality (hazard ratio, 0.75; 95% confidence interval, 0.52 to 1.08) or improved locomotion (difference, –0.04 point; 95% confidence interval, –0.49 to 0.39), but it was associated with fewer days of severe and very severe pain (difference, –0.22 day; 95% confidence interval, –0.41 to –0.03) and a shorter length of stay by 1.94 days (p < 0.001). Postoperative pain and length of stay did not differ between the groups. Early surgery also was associated with fewer major complications (odds ratio, 0.26; 95% confidence interval, 0.07 to 0.95). Parker and Handoll30, in a Cochrane Database Systematic Review, compared cephalocondylic intramedullary nails with extramedullary implants for the surgical treatment of extracapsular hip fractures in adults. Eighteen randomized and quasi-randomized trials comparing the Gamma nail with the sliding hip screw were included, with data available for 2575 patients. The Gamma nail was associated with an increased risk of operative as well as later fractures of the femur and, of course, an increased rate of reoperation. There were no major differences between the implants with regard to the rates of wound infection, mortality, or medical complications. Data were inadequate for other outcomes. The authors rightfully concluded that additional studies are required to determine if different types of intramedullary nails produce similar results or if intramedullary nails have advantages for selected fracture types. Sentiment seems to be increasing for the replacement of the fractured proximal part of the femur in light of relatively high complication rates that have been reported in association with fixation of these fractures in the elderly. Bhandari et al.31 examined this very issue in a meta-analysis involving nine trials that included a total of 1162 patients. The authors found a trend toward an increase in the relative risk of death in the first four months after arthroplasty compared with the risk in the first four months after internal fixation (relative risk, 1.27); however, at one year, the relative risk of death was 1.04. Fourteen trials that included a total of 1901 patients provided data on revision surgery. The relative risk of revision surgery after arthroplasty compared with the risk after internal fixation was 0.23 (p = 0.0003), although arthroplasty was associated with a significantly increased risk of infection (relative risk, 1.81; p = 0.009). Pain relief and the attainment of overall good function were similar in the two groups. Only larger trials will resolve remaining critical questions regarding the optimal treatment, such as relative successful outcomes in both treatment groups or technique-dependent variables. Certainly, strategies must evolve to prevent hip fractures in the first place; however, a study from Amsterdam did not seem to show that the use of hip protectors was effective toward this end32. In that study, 561 high-risk patients with an age of seventy years or more were enrolled in a randomized, controlled trial. The mean duration of follow-up was 69.6 weeks. Eighteen hip fractures occurred in the hip-protector group, compared with twenty in the control group. Four of the eighteen hip fractures in the intervention group occurred while the patient was wearing a hip protector. There was no significant difference between the intervention and control groups with regard to the time to a first hip fracture. The rate of compliance as detected during unannounced visits was found to be 62% after one month and 37% after twelve months. However, even the comparison of compliant participants did not show a significant difference between the groups with regard to the risk of hip fracture (hazard ratio, 0.77; 95% confidence interval, 0.25 to 2.38). The failure of treatment of proximal femoral fractures presents substantial clinical challenges. Bartonicek et al.33 evaluated the results of valgus intertrochanteric osteotomy for the treatment of varus nonunion and malunion of trochanteric fractures. In that study, fifteen patients (age range, twenty-nine to eighty-four years) with varus malunion (eleven patients) or varus nonunion (four patients) were treated with a valgus intertrochanteric osteotomy and fixation with a 120° double-angled blade plate. The indication for surgery included a varus malunion with limb-shortening of >2 cm that was associated with limp, abductor muscle insufficiency, hip pain, and back pain. After a mean duration of follow-up of 5.5 years (range, two to ten years), fourteen patients had healing without complications; twelve patients had healing within four months. All patients recovered at least 90° of flexion, and the mean Harris hip score improved from 73 points (range, 61 to 83 points) preoperatively to 92 points (range, 76 to 98 points) postoperatively. Osteoarthritis or avascular necrosis of the femoral head did not develop in any patient. Haidukewych and Berry34, in a study of sixty elderly patients, demonstrated that successful salvage treatment with hip arthroplasty is possible after the failed treatment of an intertrochanteric hip fracture. Thirty-two patients had a total hip arthroplasty with a cemented cup (twenty-four patients) or an uncemented cup (eight patients), twenty-seven had a bipolar hemiarthroplasty, and one had a unipolar hemiarthroplasty. A calcar-replacement design, extended-neck stem, or long-stem implant was used in fifty-one of the sixty hips. Forty-four patients were followed for a mean of five years. Thirty-nine patients had no or mild pain, and five had moderate or severe pain. Pain was commonly located in the region of the greater trochanter. Forty patients were able to walk, twenty-six with one-arm support or less. Five reoperations were performed, and two dislocations were recorded. Survivorship analysis with revision of the implant for any reason as the end point revealed a survival rate of 100% at seven years and of 87.5% (95% confidence interval, 67.3% to 100%) at ten years. Certainly, an extension of the natural history of total hip arthroplasty is the periprosthetic femoral fracture. Although allograft supplementation of plate fixation has been advocated for the treatment of such fractures, Ricci and Borrelli35, in an OTA Highlight Study that was presented at the 2003 annual meeting of the AAOS, reported on the treatment of twenty patients with use of a biological plating technique without allograft supplementation. Of the seventeen patients who were available for follow-up, all healed with satisfactory alignment and thirteen returned to baseline walking status. Although there has been a substantial evolution in the treatment of distal femoral fractures over the last thirty years, no substantial series has documented the long-term results after open reduction and internal fixation of such fractures. Rademakers et al.36 reported the long-term functional and radiographic results for thirty-two patients with monocondylar or bicondylar distal femoral fractures. After one year of follow-up, forty (60%) of the original sixty-seven fractures had healed and another twenty-two (33%) demonstrated a fixation callus. One patient had a nonunion. The mean range of motion of the knee was 111° (range, 10° to 145°), with the exclusion of two knees that required arthrodesis. After a mean duration of follow-up of fourteen years (range, five to twenty-five years), the mean range of motion of the knee was 118° (range, 10° to 145°). The Neer score showed good to excellent results in 84% of the patients, and the Hospital for Special Surgery knee score showed good to excellent results in 75% of the patients. The Ahlback score showed moderate to severe development of secondary osteoarthritis in 36% of the patients, although 72% of these patients still had a good to excellent functional result. Seven fractures (10%) were associated with local complications in the form of a deep wound infection. This study represents yet another valuable long-term follow-up report that is of substantial historical value and that provides a clinical perspective that seems to suggest remarkable forgiveness in the natural history of knee trauma.
Running against this theme, a study from the Hospital for Special Surgery identified risk factors for poorer outcomes following the operative treatment of displaced tibial plateau fractures in elderly patients37. In that study, thirty-nine displaced tibial plateau fractures (unicondylar and bicondylar) in patients older than fifty-five years of age were treated operatively. At an average of 2.5 years after surgery, external fixation and increasing age were associated with significantly poorer results according to the criteria of Rasmussen and the Short Musculoskeletal Functional Assessment (SMFA) examinations (p < 0.0001). However, preexisting degenerative joint disease and the Schatzker fracture class were not predictive of poorer outcomes according to these same measures.
Moving on to higher-energy knee trauma, Mills and
Tejwani38
investigated whether the Injury Severity Score (ISS) is predictive of the
occurrence of heterotopic ossification after knee dislocation. Thirty-six knee
dislocations were analyzed in order to develop a classification system for
heterotopic ossification. This was the first report to discuss the prevalence
of heterotopic ossification after trauma, and it also was the first attempt to
develop a classification system for the severity of heterotopic ossification.
The authors also discussed the prevalence of heterotopic ossification after
trauma and developed a classification system for its severity. The types of
heterotopic ossification ranged from none (type 0) to ankylosing (type IV).
Twenty-nine patients with type-0 to III heterotopic ossification (Group A)
recovered an average range of motion of 126° at an average of fourteen
months. Six patients had ankylosing type-IV heterotopic ossification (Group
B). The ISS ranged from 9 to 26 points in Group A, compared with 26 to 50
points in Group B (p < 0.001). Regarding the formation of type-IV
heterotopic ossification, the sensitivity of an ISS of In another recent study, Bhandari et al.39 sought to identify predictors of reoperation following operative treatment of fractures of the tibial shaft. In a retrospective observational study of 200 patients with tibial shaft fractures, twenty possible prognostic variables were assessed to predict the likelihood of reoperation. Reoperation was defined as any surgical procedure performed less than one year after the initial index operation for the injury. Only an open fracture wound (relative risk, 4.32; 95% confidence interval, 1.76 to 11.26), lack of cortical continuity between the fracture ends following fixation (relative risk, 8.33; 95% confidence interval, 3.03 to 25.0), and the presence of a transverse fracture (relative risk, 20.0; 95% confidence interval, 4.34 to 142.86) correlated with a higher incidence of reoperation.
Moving on to high-energy ankle trauma, Pollak et al.40 performed a retrospective cohort analysis of pilon fractures that had been treated at the Harborview Medical Center in Seattle and the Shock Trauma Medical Center in Baltimore between 1994 and 1995. Seventy-eight percent of 103 eligible patients were evaluated at a mean of 3.2 years after surgery to determine factors influencing outcome. Treatment with external fixation (with or without limited internal fixation) was significantly related to poorer results. All five amputations in the cohort occurred in patients who had been treated with external fixation, although it should be pointed out that this group had a greater prevalence of open fractures and OTA type-C injuries than did the group managed with open reduction and internal fixation. Additionally, two or more comorbidities, being married, having an annual personal income of less than $25,000, and not having attained a high-school diploma also were associated with poorer outcomes. Relatively poorer outcomes for this injury were further supported by the findings that 35% of the patients reported substantial ankle stiffness, 29% reported persistent swelling, and 33% reported ongoing pain. Furthermore, 43% of the employed patients did not return to work. In patients with lower-energy ankle fractures, soft-tissue injury is a frequent determinant of the treatment rendered. Nielson et al.41, in a prospective study of seventy-three patients with ankle fractures, reported that thirty patients had a complete tear of the interosseous membrane on magnetic resonance imaging. In seven patients, the interosseous membrane tear was proximal to the fibular fracture as detected with magnetic resonance imaging. Five of these seven patients had a Weber type-B fracture, and two had a Weber type-C fracture. Conversely, three patients with a Weber type-C fracture had a tear that remained distal to the level of the fibular fracture, leading the authors to conclude that one cannot consistently estimate the integrity of the interosseous membrane and subsequent need for transsyndesmotic fixation solely on the basis of the level of the fibular fracture. These findings underscore the need for an intraoperative syndesmotic stress test to determine the competence of the interosseous membrane. Tornetta et al.42, in another 2003 OTA Highlight Study, presented interesting data on the high frequency of syndesmotic injuries in patients with Weber type-B ankle fractures. In that study, 291 adults with ankle fractures were evaluated intraoperatively with regard to syndesmotic competence. The criteria for the diagnosis of syndesmotic instability were 2 mm of subluxation of the talus, syndesmotic widening, or posterior subluxation. Syndesmotic instability was found in association with 36% of fixed ankle fractures (including 30% of bimalleolar type-IV supination-external rotation fractures and 40% of type-IV supination-external rotation fractures associated with a deltoid ligament injury). Ninety-four percent of the bimalleolar fractures with syndesmotic incompetence occurred in association with an anterior collicular medial malleolar fracture. Egol et al.43 used a computerized driving simulator to help to develop guidelines for the timing of safe driving after an ankle fracture. Eleven healthy volunteers (Group I) were tested once to establish normal mean values, and thirty-one volunteers with a fracture of the right ankle (Group II) were tested at six, nine, and twelve weeks following operative repair. The subjects were tested in a series of driving scenarios (city, suburban, and highway). The increase in the total braking time at six weeks meant an increase in the distance traveled by the automobile before braking of 22 ft (6.7 m) at 60 mph (96.6 km/hr), and the increase at nine weeks meant an increase of 8 ft (2.4 m). The authors concluded that, by nine weeks, the total braking time of patients who have undergone fixation of a displaced right ankle fracture returns to the normal, baseline value. It is surprising that so few studies have documented surgical and functional outcomes after talar fractures; this year, however, we present two valuable contributions. In the first study, Vallier et al.44 identified fifty-seven talar body fractures that had been treated operatively. Twenty-three patients had a concomitant talar neck fracture, and eleven had an open fracture. Thirty-eight patients were evaluated after an average duration of follow-up of thirty-three months. Early complications occurred in eight patients. Ten of the twenty-six patients who had a complete set of radiographs had development of osteonecrosis of the talar body. Five of these ten patients experienced collapse of the talar dome at a mean of 10.2 months after surgery. All patients with a history of both an open fracture and osteonecrosis experienced collapse. Fractures of both the talar body and neck led to the development of advanced arthritis more frequently than did fractures of the talar body only (p = 0.04). All patients with open fractures had end-stage posttraumatic arthritis (p = 0.053). In another study, Sanders et al.45 examined seventy patients with displaced talar neck fractures. They found that the incidence of secondary reconstructive surgery increased from 24% ± 5% at one year after the injury to 48% ± 10% by ten years after the injury. The functional outcome was highly variable and was most dependent on the development of complications. Of the forty-four patients who did not require reconstructive surgery, 70% returned to employment and rated their overall function as 63% of the preinjury level. The incidence of secondary reconstructive surgery following talar neck fractures increased over time, and surgery was most commonly performed to treat subtalar arthritis or misalignment. A new generation of studies on the operative treatment of calcaneal fractures is indicating that catastrophic complications are not likely and that the historically high rates of wound infection and dehiscence are avoidable. Benirschke and Kramer46, in a study of 341 closed and thirty-nine open calcaneal fractures that were treated with open reduction through an extensile lateral approach, reported that 1.8% of patients with closed fractures and 7.7% of those with open fractures experienced serious infections that required intervention beyond oral antibiotics. All incisions and fractures eventually healed. Berry et al.47, in a study of thirty open calcaneal fractures in twenty-nine patients, reported no deep infections and no late amputations. Twenty-one of the twenty-eight patients returned for follow-up at an average of forty-nine months. The average SF-36 scores were within one standard deviation of published Canadian norms. Worse function was observed in patients with plantar wounds and severely comminuted fractures. To continue with the theme of safe treatment of calcaneal fractures, Aldridge et al.48, in a retrospective review of nineteen operatively treated open calcaneal fractures, reported that two patients had development of chronic, draining calcaneal osteomyelitis, for which one patient underwent a below-the-knee amputation. The rate of complications in this series was 11%, although it should be noted that five patients required free tissue transfer for wound coverage. It has been stated that calcaneal fracture surgery is among the most painful of orthopaedic procedures. Cooper et al.49 reported the results of a randomized, prospective trial involving thirty patients having open repair of a calcaneal fracture. The patients were divided into three groups of ten. Group 1 received patient-controlled morphine analgesia alone, whereas Groups 2 and 3 received patient-controlled morphine analgesia and a "one-shot" bupivacaine sciatic nerve blockade either preoperatively (Group 2) or postoperatively (Group 3). Sciatic nerve blockade conferred significant pain relief compared with morphine alone as measured with use of a number of variables. Postoperative sciatic nerve blockade provided the longest possible postoperative block duration, whereas preoperative blockade conferred no other advantage compared with postoperative blockade. Another long-term series from the University of Iowa revealed that acceptable results are possible in association with the nonoperative treatment of lower extremity fractures50. In this 2003 OTA Highlight Study, Allmacher et al. evaluated twenty-four closed calcaneal fractures that had been treated without surgery between 1970 and 1985. The patients were assessed at two time-points, once at an average of 12.8 years and again at an average of twenty-two years, to determine sequential outcomes associated with this injury. The authors found a deterioration of outcome between the first and the second decade after the injury. The rate of excellent and good results decreased from 63% to 47%. Subtalar arthrosis on computed tomographic scans correlated closely with deterioration in outcome.
McKee et al.51 challenged the widely held belief that nonoperative treatment of clavicular fractures is invariably associated with a good functional outcome. In a study that won the Bovill Award at the Nineteenth Annual Meeting of the OTA in 2003, these investigators evaluated a series of twenty-five patients with united fractures of the clavicle and documented that strength-testing showed significantly decreased endurance. The mean DASH (Disabilities of the Arm, Shoulder and Hand) score was 25.1 points, which also indicated significant functional impairment. Patients with clavicular shortening of >2 cm were relatively more affected. McKee et al.52 also brought it to our attention that surgical treatment may be an option for patients with malunited clavicular fractures. The authors reported on fifteen patients who underwent corrective osteotomy for the treatment of a symptomatic malunion. All patients had a significant improvement in function. Clavicular shortening improved from 2.9 cm preoperatively to 0.4 cm postoperatively. The mean DASH score improved from 32 points preoperatively to 12 points at a mean of twenty months postoperatively. Other investigators have shed more light on problems associated with fractures of the clavicle. Robinson et al.53 performed a prospective cohort study of 868 patients with fractures that were treated nonoperatively to identify the factors that were associated with clavicular nonunion. The overall rate of nonunion was 6.2% at twenty-four weeks. The nonunion rate was higher in patients with distal-fifth clavicular fractures (11.5%). The factors that were associated with nonunion of a diaphyseal fracture were advanced age, female gender, fracture displacement, and comminution. For fractures of the distal part of the clavicle, only age and displacement were predictive of nonunion. Moving slightly distally, Gaebler et al.54 presented the results of a five-year prospective study on the epidemiology and outcome of 507 minimally displaced proximal humeral fractures that were treated nonoperatively. Adequate outcome information was available for 376 fractures. Patients were managed with immobilization in a sling for two weeks before range of motion was instituted. There was gradual improvement in functional status between six weeks and one year. Eighty-eight percent of the patients had a good or excellent outcome at one year. The only two factors that predicted a good outcome were a younger age and a longer duration of physical therapy. Galatz et al.55, in a study of thirteen patients who had been referred to two university centers over a ten-year period, documented the functional outcome of open reduction and internal fixation of nonunions of the surgical neck of the humerus. The results were excellent in eleven patients, demonstrating that fixation of surgical neck nonunions is associated with predictable and encouraging results if patients are carefully selected.
Park et al.56 evaluated the results for twenty-seven patients in whom two-part and three-part proximal humeral fractures had been treated with heavy, braided, nonabsorbable, rotator cuff-incorporating sutures. Seventy-eight percent of the patients had an excellent outcome. The average forward elevation was 155°. Radiographically, 86% of the patients had anatomic alignment of the proximal part of the humerus. This minimalist approach seems to provide interesting results and likely requires the authors' conservative approach to rehabilitation as described by Neer and Hughes. Distal humeral fractures in elderly patients can present a complex challenge to the orthopaedic traumatologist. Failure of fixation is always a concern in patients with osteoporotic bone. The option of performing an acute total elbow arthroplasty was studied by Frankle et al.57, who retrospectively compared open reduction and internal fixation with total elbow arthroplasty. Eight of twelve patients who had been treated with open reduction and internal fixation and twelve of twelve patients who had been treated with a total elbow arthroplasty had an excellent or good result after a minimum duration of follow-up of two years. Three patients in the open reduction and internal fixation group had loss of fixation requiring revision to total elbow arthroplasty. None of the arthroplasties required revision. Kamineni and Morrey58 reviewed forty-nine acute fractures that had been treated with total elbow arthroplasty. The average arc of motion was from 24° of extension to 131° of flexion. The average Mayo elbow performance score was 93 of a possible 100 points. Sixty-five percent of the patients had no complications, and only ten additional surgical procedures (including five revisions) were performed. Total elbow arthroplasty, performed acutely, seems to be a reasonable alternative to open reduction and internal fixation in elderly patients with osteoporotic bone.
Fractures in children generally receive a paucity of attention in the context of the overall orthopaedic trauma literature, perhaps as indicated by the current update. One of the most controversial areas is the treatment of pediatric femoral fractures, perhaps because the complications of any chosen treatment are either commonplace or disastrous. Flynn et al.59 prospectively studied eighty-three children between the ages of six and sixteen years who were treated with titanium elastic nails or with traction followed by spica casting. The choice of treatment was left up to the staff surgeon. Thirty-five children (thirty-five fractures) with a mean age of 8.7 years were treated with traction and application of a spica cast, and forty-eight children (forty-nine fractures) with a mean age of 10.2 years were treated with titanium elastic nails. Three patients who had been treated with traction and a spica cast had an unsatisfactory result, compared with none of the patients who had been treated with titanium elastic nails. Twelve patients (34%) who had been treated with traction and a cast had a complication, compared with ten patients (21%) who had been treated with titanium elastic nails. Furthermore, patients who had been treated with titanium elastic nails had a shorter hospitalization, walked with support sooner, walked independently sooner, and returned to school sooner. All of these differences were significant (p < 0.0001). There was no difference in total hospital charges between the two groups. Finally, with regard to the treatment of articular fractures in children, Cutler et al.60 reported that computed tomographic scans aided in the identification of the ideal position for the screw (perpendicular to and at the midpoint of the fracture) in a cohort of sixty-two patients with distal tibial physeal fractures. Specifically, there was significant improvement (p < 0.0001) in the accuracy of the point of insertion (mean, 12 mm) and the direction of the screw (mean, 31°) on the preoperative plan when computed tomographic scans were used rather than plain radiographs.
This orthopaedic trauma update represents somewhat of a genesis. This work is a first-time attempt to use the tools of a scientific systematic review to choose publications most relevant to orthopaedic trauma and then to summarize them in a new framework of evidence-based medicine. The methodology, we believe, provides a new set of eyeglasses that should aid the reader in seeing a more relevant context for contemporary clinical investigation, indeed one that helps to filter solutions for the diagnosis, workup, and treatment of patients. Because of the newfound relevance of this emerging framework, we spent a fair amount of discussion on the history and definition of evidence-based medicine as well as the process of systematic review before launching into the selection and summary of forty-six publications and eleven papers from the Nineteenth Annual Meeting of the Orthopaedic Trauma Association, which were further distinguished as OTA Highlight Papers at the 2003 AAOS meeting in San Francisco. Although in Table II we chose to classify each of the Level I, II, and III studies, we wish to underscore the importance of Figure 1. Specifically, Figure 1 helps to emphasize that scientific evidence is only one of three factors that should influence clinical decision-making, along with patient preferences and clinical circumstances. All three influences interface with the importance of a surgeon's own clinical experience in choosing a method of diagnosis, workup, or therapy. Level-I and II evidence, although most unbiased, is not possible or feasible for all desired hypotheses, and it simply never exists in the natural evolution of new ideas and treatments from the outset. Therefore, it is paramount that we consider that all levels of evidence are necessary to build the pyramid of understanding, lest we rebuke the value of more biased evidence that may in fact be all that we have for certain periods of the natural history of understanding of a clinical question.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Introduction
General Topics in Fracture...
26 points was 100%,
the specificity was 97%, and the positive predictive value was 86%. Patients
in Group B had a greater prevalence of documented closed head injury (p <
0.025). Neither the timing of surgical ligament reconstruction nor the number
of ligaments reconstructed had a significant influence on the degree of
heterotopic ossification. | home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |