In a 1987 editorial in the Journal of Bone and Joint Surgery , Dr Sigvard Hansen, of Harborview Medical Center in Seattle, noted the profound physical, mental, social, and financial implications of futile attempts at limb salvage in the setting of severe lower extremity trauma. He acknowledged the evolution of microvascular and external fixation techniques, which made “heroic” limb salvage procedures possible. Although for some patients the outcomes represent true progress in medicine, other patients had to go through years of repeated surgeries, infections, and bone grafting only to end up with a compromised amputation on a delayed basis. Many of these patients ended up demoralized, divorced, destitute, and drug addicted.

Dr Hansen called for the development of objective guidelines to influence clinicians’ decisions about “amputation versus salvage” and concluded his article with the following: “Perhaps the best source would be a multicenter study done by members of the Orthopaedic Trauma Association. The development of such guidelines would help to provide an answer to this problem and thus allow both patients and their doctors to avoid prolonged, costly, and fruitless salvage procedures when such a course is not indicated.”

Subsequently, Johansen and colleagues established the Mangled Extremity Severity Score (MESS) criteria in an effort to provide such a set of guidelines. This system, first published in 1990, attempted to stratify 4 variables: skeletal/soft tissue injury, limb ischemia, shock, and patient age. An analysis of retrospective and prospective study groups led to a recommendation that a MESS value greater than or equal to 7 was an indication for amputation. In the prospective portion of the analysis, this proved to be 100% accurate in predicting amputation. The weakness of this study is the self-fulfilling nature inherent in studying this algorithm on a prospective basis.

Even with such guidance, orthopedic surgeons continued to struggle with the decision of whether to amputate or salvage the severely injured lower extremity. Commonly held beliefs, including that an insensate plantar foot was an indication for amputation, were based on little or no evidence.

Given this obvious gap in the literature of the emerging subspecialty of orthopedic traumatology, a prospective longitudinal study was undertaken at 8 level I trauma centers nationally. Ultimately headed by Ellen MacKenzie, PhD, a professor of Health Policy and Management at Johns Hopkins University, and Dr Michael Bosse, an orthopedic traumatologist at Carolinas Medical Center, the Lower Extremity Assessment Project (LEAP) study set out to answer many of the questions surrounding the decision to amputate or salvage. A National Institutes of Health–funded, multicenter, prospective observational study, the LEAP study represented a milestone in orthopedic trauma research, and perhaps in orthopedics.

The inclusion criteria for the LEAP study included the following:

• 1.
  

  Traumatic amputations below the distal femur

  
  
• 2.
  

  Gustilo grade IIIA open tibia fractures with hospital stays greater than 4 days, 2 or more limb procedures, and a high degree of nerve, muscle, or bone injury

  
  
• 3.
  

  Gustilo grade IIIB and IIIC open tibia fractures

  
  
• 4.
  

  Dysvascular injuries below the distal femur

  
  
• 5.
  

  Major soft tissue injuries below the distal femur, excluding the foot

  
  
• 6.
  

  Grade III open pilon fractures

  
  
• 7.
  

  Grade IIIB open ankle fractures

  
  
• 8.
  

  Severe open hindfoot and midfoot injuries with degloving and nerve injury.

The LEAP study attempted to account for all variables in patients sustaining these injuries. Surgeons caring for patients participating in the study would be permitted to carry out patient management as they saw fit. There was no set treatment algorithm to the study. The LEAP study attempted to define the characteristics of the individuals who sustained these injuries, the characteristics of their environment, the variables of the physical aspects of their injury, the secondary medical and mental conditions that arose from their injury and treatment, their ultimate functional status, and their general health.

Besides age, gender, and comorbidities, the values, beliefs, and psychological profiles of patients’ were also assessed. The characteristics of their environment included their physical surroundings, their educational and economic backgrounds, and their social and work surroundings. Assessing all of these aspects entailed the use of many assessment tools never previously used in orthopedic research.

In terms of evidence-based medicine, the LEAP study and the resulting publications are somewhat different from many evidence-based trials. There is no control of the therapeutic intervention. However, the data was prospectively collected, and the investigators have accumulated 7-year follow-up on most of the patients. With the 7-year follow-up data, examiners have attempted to record, as objectively as possible, all the potential variables in what is a complicated clinical situation. It seems that the existing lower extremity trauma scoring systems are not predictive of the outcome of salvage or amputation, but unfortunately the LEAP studies have not yielded a better or more accurate method for predicting optimal treatment.

In the final analysis, it has been learned that many of the patients sustaining this high degree of extremity trauma have a great number of social, economic, and personality disadvantages, even before their injury. Functional outcomes and quality of life outcomes seem more related to many of these preexisting factors than to interventions provided by the health care system, regardless of whether they underwent amputation or salvage. This fact may be distressing news to the orthopedic traumatologist, as this could mean that all of the prevarication and anxiety surrounding the treatment of these injuries have little long-term effect on the outcome. The alternative view would promote keeping these variables in mind and trying to maximize the support for those patients most in need at the time of severe lower extremity trauma.

Given the nature of the study, the levels of evidence are all level I or level II. There are no data in the orthopedic trauma literature to date to compare with the LEAP study, in terms of the breadth of the data recorded or the number of patients enrolled. From the standpoint of evidence-based medicine, this is the best information available.

Based on the inclusion criteria, 601 patients were enrolled for more than a period of 44 months at 8 centers. Initial patient characteristics data showed that 77% were male, 72% were white, and 71% were between the ages of 20 and 45 years. Of the patients, 70% had graduated from high school (compared with 86% nationally) and 25% lived below the federal poverty line (16% nationally), 38% had no health insurance (20% nationally), and there were twice as many heavy drinkers in this sample as in the population at large. On personality inventories, these patients were noted to be slightly more neurotic, extroverted, and less open to new experiences than the remainder of the population. None of these characteristics appeared to have an influence on the likelihood of the affected limb that was going to be amputated or salvaged.

A 2002 LEAP publication in the New England Journal of Medicine compared the outcome of surgeries between those patients who underwent reconstruction and those who underwent primary amputation. The major outcome variable examined was the Sickness Impact Profile (SIP). The SIP is a measure of self-reported health status, which relies on 136 statements of limitation in each of 12 subcategories. A low score on the scale represents a lower sickness impact and a score of greater than 10 in any area represents severe disability.

Two years after injury, there was no significant difference in the SIP scores between the group that underwent amputation and the group that underwent reconstruction (12.6 vs 11.8, P = .53). Even after adjusting for characteristics of the patients and of their injuries, functional outcomes were similar between the 2 groups. Self-efficacy is one of the characteristics assessed with SIP scores, and describes the confidence a subject has in his or her ability to perform specific tasks or activities. A person with low self-efficacy may disengage from the coping process because they expect to fail. Self-efficacy and social support turned out to be highly predictive of outcome in the group that underwent reconstruction and the group that underwent amputation. Other predictors of a lower score on the SIP, included rehospitalization for a major complication, lower educational level, nonwhite race, poverty, lack of private health insurance, poor social support network, low self-efficacy, smoking, and the involvement of the patient in disability or compensation litigation. Patients who underwent reconstruction were more likely to have a secondary hospitalization for major complication than those who underwent amputation (47.5% vs 33.9%, P = .002). Two years after their injury, only 53% of those who underwent amputation had returned to work and 49% of those who underwent reconstruction had returned to work. Unfortunately, many of the predictors of poor SIP scores coincide with patient characteristics that are overrepresented in the severe lower extremity trauma population (lower educational level, poverty, and lack of health insurance).

The cohort of 161 patients who had undergone amputation above the ankle within 3 months of injury were further examined with the SIP, in an effort to determine the influence of amputation level on functional outcome. The patients treated with amputations above the knee showed no significant difference in their SIP scores from those treated with amputations below the knee. Patients with through-knee amputations had worse regression-adjusted SIP scores than either above or below the knee amputees. Patients with amputations below their knee had the fastest walking speeds. The study also failed to find any link between outcomes and the technological sophistication of the prosthetic device used by the amputees. Considering there are approximately 3500 traumatic major lower limb amputations per year in the United States, these findings offer some guidance to physicians; however, the level and the severity of the trauma is still the most likely determining factor for the level of amputation.

Another subset of the LEAP studies that attempted to evaluate the decision-making processes in amputation versus reconstruction was published in 2001. This study included 556 of the LEAP study patients and evaluated them by 5 different injury severity scoring systems. The MESS, the Limb Salvage Index, the Predictive Salvage Index, the Nerve Injury, Ischemia, Soft Tissue Injury, Skeletal Injury, Shock, and Age of Patient Score, and the Hannover Fracture Scale-97 for ischemic and nonischemic limbs were assessed. Immediate amputation and amputation within a 6-month period were assessed. There were 63 immediate amputations and 86 delayed amputations.

The LEAP study did not support the usefulness of any of the examined lower extremity injury severity indices for determining limbs that require amputation and those likely to be successfully salvaged. Overall, these scores lacked sensitivity, but were in some cases specific. Scores at the time of injury were not useful in identifying patients that would eventually require amputation, but they might have some use in predicting which limbs could be successfully salvaged. This study tried to address the initial question of whether there are objective criteria that could be used for indicating amputation. Unfortunately, no established system was predictive, and no evidence-based alternative has been proposed.

Before the LEAP study, a widely held indication for amputation in case of lower extremity trauma was the absence of plantar sensation at the initial presentation. In 2005, an article by Bosse and colleagues attempted to determine the long-term outcomes after treatment of those patients who present with the absence of plantar sensation. The study included 26 insensate plantar feet that were amputated, 29 insensate feet that were salvaged, and 29 matched controls from among the larger cohort of sensate limbs that were salvaged. At 2-year follow-up, normal plantar sensation was present in an equal proportion of those who initially had an insensate foot that was salvaged and in those who had a sensate foot that was salvaged (approximately 55%). Only 1 patient of the 29 of the insensate salvaged group had completely absent plantar sensation at 2-year follow-up, the others having at least some level of improvement. There were no significant differences in the SIP between any of the 3 groups. In summary, absent plantar sensation at the time of presentation did not prove to be an indication for amputation, a predictor of functional outcome, or even a predictor of eventual plantar sensation.

A separate study examined the 7 lower extremity injury severity indices, as they might be related to functional outcomes, 2 years after salvage. Median SIP scores were 15.2 at 6 months and 6.0 at 24 months. None of the examined scoring systems were predictive of ultimate functional outcome or improvement in functional outcome between 6 and 24 months.

A subsequent grouping of the LEAP studies evaluated treatment variables that may have affected outcome. A 2007 article by Webb and colleagues examined a subcohort of 156 patients to describe surgeon-controlled variables that may have affected union, complication, and functional outcome in severe open diaphyseal tibia fractures. It appeared that the timing of wound debridement (<6 hours vs 6–24 hours), the timing of soft tissue coverage (more or less than 3 days after injury), and the timing of bone grafting (more or less than 3 months after injury) did not seem to have any influence on infection rates, union rates, or functional outcome. Those fixed with intramedullary fixation had slightly less severe injuries than those fixed with external fixation, and understandably had lower SIP scores and fewer complications. The entire cohort did show deterioration in SIP scores between 2-year and 7-year follow-up.

The effect of smoking on fracture healing and complications was examined in a cohort of 268 patients with open tibia fracture. A multivariate regression analysis examined those who had never smoked, those who had quit smoking, and those who were current smokers regarding their ability to heal the fracture within 24 months. The study also evaluated their time to union, the presence or absence of infection, and the presence or absence of osteomyelitis. Current smokers were 37% less likely to achieve union than nonsmokers, and previous smokers were 32% less likely to achieve union than nonsmokers. Current smokers were 2.2 times more likely to develop an infection and 3.7 times more likely to develop osteomyelitis than nonsmokers. Those who quit smoking were at no greater risk of infection overall, but were at 2.8 times greater risk for developing osteomyelitis. The effect of smoking on open tibia fractures has been examined previously, but the LEAP study certainly represents the largest cohort of prospectively collected data in which multivariate analysis could be used to isolate the effect of smoking.

The optimal type of flap coverage has been a source of dispute almost since the advent of microvascular free tissue transfer techniques. Pollak and colleagues examined a cohort of 190 patients in the LEAP database who required flap coverage, and who had at least 6 months of follow-up. End points included short-term complications, such as wound infection, necrosis, and loss of the flap. There were 87 limbs treated with a rotational flap and 107 limbs treated with a free tissue transfer. In terms of selection bias, the group treated with free tissue transfer represented more severe injuries to the limb and the group treated with rotational flap had a significantly higher injury severity score for overall trauma. There were no significant differences with respect to overall complication rates. After controlling for other variables, a single difference in short-term complications was identified specifically in those patients with the most severe osseous injuries. Limbs that featured an Orthopaedic Trauma Association/AO type C bony injury that was treated with a rotational flap were 4.3 times more likely to have a wound complication requiring reoperation than those treated with a free flap. It would seem that the severity of the osseous injury might be predictive of injury to the surrounding soft tissues, which may diminish the success rate of rotating local tissue.

A separate study reported the overall complication rate for the groups with amputation and the groups with reconstruction at 2-year follow-up. Of 149 limbs that underwent amputation during the initial hospitalization, the revision amputation rate was 5.4%. The complication rate at 3 months was 24.8% and one-third of these were wound infections. Out of 371 limb reconstructions, 3.9% required late amputation, 37.7% reported a complication by 6 months, one-quarter of which were wound infections, 23.7% had nonunion, and 7.7% had osteomyelitis. These numbers overall may best be used in counseling patients at the time of initial hospitalization. Although a few patients in such circumstances may be capable of participating in a truly informed decision, reconstruction does come with a higher rate of complication and rehospitalization than primary amputation, and 4% of reconstruction patients will end up with a late amputation.

A study by O’Toole and colleagues attempted to examine the variables specific to patient satisfaction. None of the patient demographics, treatment characteristics, or injury characteristics was found to correlate with patient satisfaction. Five key outcome measures seemed to account for more than one-third of the overall variation in patient satisfaction: return to work, depression, the physical functioning component of their SIP, their self-selected walking speed, and their pain intensity. It seems that patient satisfaction is determined most significantly by function, pain, and presence or absence of depression.

An area of trauma care that traditionally receives little attention is psychological distress, and the LEAP study attempted to determine the rate of this particular comorbidity in severe lower limb injury. Forty-eight percent of patients tested positive a likely psychological disorder 3 months after injury, and this number only diminished to 42% at the 2-year mark. Almost 20% reported severe phobic anxiety or depression, and all examined subscales of psychological distress tested higher than normative values. Unfortunately, only 12% and 22% of patients reported receiving any mental health services 3 months and 24 months after injury, respectively. This study clearly identified an area of trauma care with a room for improvement.

Another subset of injuries examined was knee dislocations with vascular injury, but there were only 18 patients in the LEAP cohort that met these criteria. Four of the 18 limbs were amputated (22%) and a prolonged warm ischemia time was the factor most highly associated with amputation. SIP scores for successful reconstruction were 12 at 1 year and 7 at 2 years, compared with a 2-year SIP score of 16 for those who underwent amputation. Patients whose limbs were salvaged did much better, but these were presumably the less severe injuries.

One of the factors identified early on in the decision to reconstruct or amputate the severely traumatized lower extremity was the monetary cost associated with each treatment option. Limb salvage incurs the costs of an increased rate of subsequent hospitalization and subsequent operation for infection, soft tissue coverage, and union, whereas amputation bears the lifetime costs of prosthetics manufacture and repair. The LEAP study addressed this issue in a 2007 article in the Journal of Bone and Joint Surgery . The cost calculations included initial hospitalization, subsequent hospitalization related to the injured limb, inpatient rehabilitation, outpatient doctor visits, outpatient physical and occupational therapy, and the purchase and maintenance of prosthetic devices. Lifetime costs were projected based on expected life years. When the prosthetic costs at 2 years were included, salvage averaged $81,316 and amputation averaged $91,106. Projected lifetime costs, however, were 3 times higher for the amputation group ($509,275 vs $163,282). It may be difficult to include monetary costs as a factor in the decision algorithm for salvage versus amputation, but these are data that treating surgeons and potential patients may want to be aware of to fully appreciate the implications of treatment decisions.

Further analysis of the LEAP data yielded 3 publications looking at 7-year follow-up. Examining long-term work disability, 58% of 423 patients followed up to 7 years had returned to work at the 7-year mark (47% of amputees and 62% of reconstructions, the difference not being statistically significant). However, even those patients who returned to work were judged to be, on average, limited in their ability to perform their job 20% to 25% of the time. Factors significantly associated with a higher rate of return to work included lower age, white race, higher education level, nonsmoking, average to high self-efficacy, preinjury job tenure, and absence of litigation in the case. The assessment of pain and physical function at just 3 months post injury was a significant predictor of ultimate return to work at 7 years.

In an attempt to examine the long-term persistence of disability at 7 years, telephone interviews of almost 400 patients were conducted and correlated with SIP scores. At 7 years, half of the patients had a SIP score greater than or equal to 10 points which, according to the SIP index, indicates severe disability. One-third of all patients had a score typical of the general population for their age and gender. When adjusting for other factors, poor physical SIP subscores were significantly associated with both those limbs that had severe soft tissue injury without fracture in the reconstruction group and those who underwent through-knee disarticulation in the amputation group. There were no significant differences in the psychosocial outcome scores between the group that underwent amputation and the group that underwent salvage. Familiar patient characteristics were significantly associated with prolonged disability, and these included increasing age, female gender, nonwhite race, lower education level, poverty, current or previous smoking, low self-efficacy score, poor self-reported preinjury health status, and involvement with litigation over disability. With the exception of age, predictors of poor outcome at 24 months were the same at 7 years.

The prevalence of chronic pain at 7 years was also reported. Only 23% of the LEAP study population was pain-free at 84 months, compared with 42.3% of the general population. In terms of the severity of their chronic pain on a graded scale, scores reported for the LEAP patients were similar to the primary care migraine headache population and the chronic back pain population. Significant early predictors of chronic pain were less than high school level education, less than college education, low self-efficacy, and high levels of alcohol consumption. Those patients who continued to be treated with narcotic medication 3 months post discharge had lower levels of chronic pain at the 84-month mark.

Some aspects of the LEAP study also examined the role of physical therapy in recovery. Patients with amputation and reconstruction used comparable amounts of physical therapy services. The percentage of patients with a perceived need for physical therapy services but receiving no therapy increased over the course of 2-year follow-up, reaching a rate of 68% at 2 years. Risk factors for not receiving therapy included lack of private health insurance, increased pain, lower level of education, lower level of fitness at the time of injury, being a smoker, and presence of a severe muscle injury. In a subsequent study, patients whose need for physical therapy was not met (as assessed by a physical therapist) were statistically less likely to improve in all selected domains of physical impairment and functional limitation when compared with those patients whose physical therapy needs were met. These 2 studies taken in aggregate seem to show the benefit of physical therapy in severe lower extremity trauma, and suggest that orthopedic surgeons may need to be more aggressive in assisting patients to receive physical therapy.

In conclusion, the LEAP study offers a wide variety of preinjury, injury, treatment, and outcome variables to examine lower extremity injuries. Although treatment was in no way randomized, these articles collectively give enhanced insight into the factors that drive measurable outcomes. Surgeons treating these patients are now more capable of properly counseling the patient and understanding prognostic factors, but perhaps are no better prepared to alter the outcome.

Dr Hansen’s quest for a score that will predict which limbs should be salvaged and which should be amputated remains unfulfilled. Ironically, a patient’s degree of “self-efficacy” (ie, how well they believe that they can handle change and maximize their future potential) may be the single greatest determining factor studied, and something completely out of the surgeon’s control. In the realm of evidence-based medicine, the LEAP studies provided a wealth of data, but still failed to completely determine the treatment at the onset of severe lower extremity trauma.