 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
The Journal of Bone and Joint Surgery (American). 2008;90:1570-1578. doi:10.2106/JBJS.G.01673 © 2008 The Journal of Bone and Joint Surgery, Inc.
The Management of Ankle Fractures in Patients with Diabetes
1 University of Pittsburgh Medical Center Comprehensive Foot and Ankle Center, Roesch-Taylor Building, Suite 7300, 2100 Jane Street, Pittsburgh, PA 15203. E-mail address: wukichdk{at}upmc.edu
Disclosure: The authors did not receive any outside funding or grants in support of their research for or preparation of this work. Neither they nor a member of their immediate families received 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, division, center, clinical practice, or other charitable or nonprofit organization with which the authors, or a member of their immediate families, are affiliated or associated.
Patients with diabetes mellitus have higher complication rates following both open and closed management of ankle fractures. Diabetic patients with neuropathy or vasculopathy have higher complication rates than both diabetic patients without these comorbidities and nondiabetic patients. Unstable ankle fractures in diabetic patients without neuropathy or vasculopathy are best treated with open reduction and internal fixation with use of standard techniques. Patients with neuropathy or vasculopathy are at increased risk for both soft-tissue and osseous complications, including delayed union and nonunion. Careful soft-tissue management as well as stable, rigid internal fixation are crucial to obtaining a good outcome. Prolonged non-weight-bearing and subsequently protected weight-bearing are recommended following both operative and nonoperative management of ankle fractures in patients with diabetes.
Ankle fractures are among the most common injuries treated by orthopaedic surgeons, with an estimated 260,000 occurring per year in the United States1,2. With an ever-aging population, the number of ankle fractures continues to rise3. While protocols for management of ankle fractures are generally well established, treatment of these injuries in patients with diabetes mellitus poses a challenge for orthopaedists. These patients have unique characteristics, including delayed fracture-healing, impaired wound-healing, vasculopathy, and neuropathy, that all must be taken into account when formulating a treatment plan. As the prevalence of diabetes mellitus has continued to increase, so too has the number of ankle fractures seen in this patient population. While much has been published on the impact of diabetes on the treatment of ankle fractures, the majority of studies have included a small number of patients and have been either retrospective reviews or case-control studies. In this review, we will report the impact of diabetes on the management of ankle fractures. Specifically, we will examine the epidemiology of diabetes as it relates to ankle fractures, the specific characteristics of diabetes that pose a problem in the management of ankle fractures (impaired wound-healing, delayed fracture-healing, and neuropathy), the information in the literature as it relates to outcomes and complications following the management of these fractures in diabetics, and current evidence regarding the optimal management of ankle fractures in patients with diabetes. A critical analysis of the existing evidence regarding the impact of diabetes on ankle fractures will lead to a better understanding of this problem and the ability to make better decisions for patient management.
According to statistics from the Centers for Disease Control and Prevention in 2005, 20.8 million people (7% of the population) in the United States had diabetes mellitus4. In the population over the age of sixty years, 10.3 million (20.2% of the population) were affected. Additionally, 1.5 million new cases of diabetes were diagnosed in 2005 alone4. The prevalence of diabetes increased by 61% between 1990 and 2001, and it is estimated that this number will further increase by 165% between 2000 and 2050, with the fastest rates of increase occurring in older and minority subpopulations5. Overall, the estimated cost incurred by the United States economy as a result of diabetes and its complications is more than 100 billion dollars annually6. Similar increases in the prevalence of diabetes and its associated costs have been reported throughout the world7,8. The authors of a recent study estimated the overall lifetime risk of diabetes developing in an individual born in the United States in the year 2000 to be 32.5% for males and 38.5% for females9. The highest rates were predicted to occur in the Hispanic population (a 45.4% risk for males and a 52.5% risk for females). The diagnosis of type-2 diabetes at the age of forty was predicted to be associated with a decreased life expectancy of 11.6 life-years for males and 14.3 life-years for females. Additionally, the number of quality-adjusted life-years was predicted to be decreased by 18.6 for males and 22.0 for females.
Diabetes mellitus constitutes a heterogeneous group of metabolic disorders that share the common manifestation of hyperglycemia. The diagnosis of diabetes is contingent on a fasting blood glucose level of  126 mg/dL (7 mmol/L) measured on two separate occasions, random glucose levels of 200 mg/dL (11 mmol/L) with symptoms (polyuria, polydipsia, or unexplained weight loss), or a positive glucose challenge of 200 mg/dL10. Broadly, diabetes can be considered to present in two forms (type 1 and type 2). Type-1 diabetes is caused by the autoimmune destruction of insulin-producing β-cells in the islets of Langerhans in the pancreas. This leads to an absolute decrease in the amount of circulating insulin. Circulating insulin is virtually absent, and pancreatic β-cells fail to respond to normal insulinogenic stimuli. Therefore, exogenous insulin must be administered in order to maintain adequate glycemic control. Type-2 diabetes (formerly known as adult-onset diabetes) is characterized by increased peripheral insulin resistance combined with a secretory defect in insulin by the pancreatic β-cells. Both components must be present for the clinical manifestation of diabetes. Overall, 90% of the cases of diabetes in the United States are type 2, and 90% of these patients are clinically obese. Regardless of the underlying cause of diabetes, the common manifestation is systemic hyperglycemia. In turn, this leads to the glycosylation of proteins and the increased formation of intracellular sorbitol and other polyols. The end product is tissue damage in a number of end organs. Patients all show some degree of immune dysfunction, peripheral neuropathy, nephropathy, retinopathy, and arthropathy.
Diabetic neuropathy profoundly impacts the management of ankle fractures. In the United States, 10% of diabetic patients have some degree of neuropathy at the time of the initial diagnosis of the diabetes, and up to 40% will be diagnosed with peripheral neuropathy within the first decade following that diagnosis1,11,12. In older patients, in whom ankle fractures are more common, neuropathy is even more prevalent. More than 50% of all diabetic patients over the age of sixty have some degree of peripheral neuropathy13. The subgroup of patients with a loss of protective sensation in the foot and ankle is at particular risk for complications following an ankle fracture. In general, peripheral neuropathy must be profound prior to the loss of protective sensation. Once protective sensation is lost, the risk of foot ulcerations increases sevenfold because of the increased vulnerability to unrecognized trauma14,15. While the so-called gold standard for diagnosing peripheral neuropathy remains nerve conduction studies, the most commonly used instrument clinically is the 10-g (5.07) nylon Semmes-Weinstein monofilament test16-19. This simple test can identify persons at an increased risk for foot ulceration with a sensitivity of up to 91% and a specificity of up to 86%20-22. Vibration testing with a 128-Hz tuning fork can also be used and may be an even more sensitive predictor of early neuropathy16. It is important to identify neuropathy in all diabetic patients and particularly in those who have sustained an ankle fracture. The severity of diabetic neuropathic complications is directly related to both the lack of control and the chronicity of abnormal glucose metabolism23. The importance of control is highlighted by the fact that a 1% reduction in the hemoglobin A1C level results in approximately a 25% to 30% reduction in the rate of complications10. Under normal glucose homeostasis, the entry of glucose into the cell is tightly regulated by insulin. However, in diabetic patients, insulin impairment leads to increased glucose levels in the bloodstream and subsequently to increased diffusion of glucose into the cells. The end product of this is the formation of advanced glycosylation end products, which are stable and irreversible. In red blood cells, these can be measured as the hemoglobin A1C level, a marker of long-term glucose control. Ultimately, the effects of hyperglycemia on peripheral nerves are twofold, direct neuronal injury and microvascular damage. On a molecular level, cellular homeostasis relies on the tightly regulated use and production of certain reactive oxygen species (nitric oxide, hydrogen peroxide, and superoxide) that play crucial roles in the normal functioning of the cell. These processes are very sensitive to glycemic control. In the presence of excess glucose, the tight regulation of reactive oxygen species is disrupted and excess reactive oxygen species are produced23,24. Within the nerve cells, the excess reactive oxygen species cause direct injury to both cellular proteins and membrane lipids. Additionally, toxic peroxidation products accumulate and bind to normal cellular nuclear material, leading to increased apoptosis, DNA damage, and decreased axonal transport. Hyperglycemia also ultimately results in a decreased production of neurotrophic factors that are responsible for the health and maintenance of normal nerve function25,26. On the vascular level, chronic excess production of superoxide species leads to a loss of normal nitric oxide function, resulting in vasoconstriction and nerve ischemia.
Charcot arthropathy is defined as a noninfectious, destructive process culminating in eventual dislocation and periarticular fracture in patients with peripheral neuropathy and the loss of protective sensation27-29. While diabetes is the most common cause of Charcot arthropathy today, it can potentially result from any clinical entity that renders a patient insensate to protective sensation. Although Charcot arthropathy is relatively uncommon, the potentially serious complications following a delayed diagnosis of this condition make early detection extremely important. As recent evidence has highlighted, a high index of suspicion and early intervention can potentially avoid some of the historically poor outcomes30. Charcot arthropathy was originally described in the 1860s by the neurologist Jean-Martin Charcot28. The hypertrophic, destructive neuroarthropathy that he initially described has since been shown to occur in diabetic patients with peripheral neuropathy. The prevalence of Charcot arthropathy is estimated to be approximately 0.3% among all diabetic patients, although it is thought that this may be an underestimate of its true burden31. The only symptom of diabetes that has been shown to be predictive of the potential for the development of Charcot arthropathy is the loss of protective sensation32,33. Two distinct theories have been postulated as potential causes of Charcot arthropathy: the neurotraumatic theory and the neurovascular theory27. In reality, the cause is likely multifactorial, and overlap between the theories is certainly feasible. The neurotraumatic theory postulates that an initial injury (either a single macrotraumatic event or repetitive microtrauma) activates the process. The loss of protective sensation leads to continued weight-bearing by the patient and the lack of protective unloading. Eventually, there is a loss of structural integrity, and finally a reparative process begins. The neurovascular theory predicates a state of hyperemia resulting from abnormal vasomotor control secondary to the underlying neuropathy. This leads to an increase in local blood flow and, secondarily, to an increased level of osteoclast stimulation. Eventually, this results in increased bone turnover and a relative local osteopenia, which leaves the osseous architecture susceptible to minor trauma. This theory is supported by the fact that local tissue samples from Charcot joints show an increased osteoclast-to-osteoblast ratio mediated by a local cytokine pathway34. Recently, overexpression of RANKL (receptor activator of nuclear factor kappa-B ligand) has been implicated as a mechanism for the development of Charcot arthropathy35. The classic clinical description of an early Charcot foot or ankle is one with painless erythema, warmth, and swelling of the involved joints. Many patients still have some degree of pain on initial presentation. Often, a patient with an acute Charcot foot is initially misdiagnosed with an infection secondary to the erythematous, inflamed appearance of the foot. However, these patients lack signs of systemic illness, such as fever or elevated white blood-cell counts. Additionally, the erythema will resolve rapidly with elevation of a Charcot foot but not with elevation of an infected foot. In the absence of an open, draining ulceration as a conduit for deep infection, infection is unlikely to be the cause of the acute inflammation. When observed prior to the development of osseous changes on radiographs, these clinical symptoms are often initially dismissed as a sprain, contusion, cellulitis, gout, or deep vein thrombosis. The initial classification system described by Eichenholtz was based on clinical and radiographic data of patients with Charcot arthropathy36. Stage I (development-fragmentation) refers to the initial, acute phase in which the cardinal signs of hyperemia, warmth, and swelling are present. Radiographically, a continuum of joint subluxation and dislocation, articular erosions, and osseous destruction is observed. Stage II (coalescence) is heralded by the resolution of the cardinal clinical signs and the transition from destruction to the reparative process. Radiographically, sclerosis is present and, as the bones begin to coalesce, their inherent stability increases and the deformity ceases to rapidly progress. Stage III (reconstruction-consolidation) refers to the hypertrophic, remodeling phase in which trabecular remodeling and the maturation of the osseous architecture into a stable, almost ankylosed state occur. The Eichenholtz classification ignores the initial phase of inflammatory swelling in which the radiographic appearance remains normal or nearly normal. It is patients with this stage of the disease who are often initially dismissed as having a bruise or strain, and unfortunately this is also the group that can most likely be successfully treated if diagnosed early. Shibata et al. were, to our knowledge, the first to describe Stage-0 neuroarthropathy in their series of patients with leprosy37. In 1995, Schon and Marks referred to Stage 0 as occurring in patients with neuropathy who sustain an acute fracture38. Chantelau et al. subsequently defined Stage-0 Charcot arthropathy as occurring in the six to twelve-month period in which the clinical signs of a Charcot foot develop without any abnormality seen on plain radiographs39,40. Magnetic resonance imaging at this stage will show a pattern of bone-marrow edema consistent with osseous stress injury. When diagnosed at this early stage, the Charcot foot can be managed with immobilization and non-weight-bearing and this can often successfully prevent the development of osseous destruction and deformity39.
Experimental and clinical studies have shown an association between diabetes and a delay in fracture-healing41. This may be due in part to effects on cellular proliferation, vascular ingrowth, mineralization of fracture callus, and/or remodeling41,42. We are aware of only a few small clinical studies in which the times to fracture-healing were investigated in diabetic populations, and none were prospective controlled studies. Loder evaluated all closed lower-extremity fractures in adult patients with diabetes treated at his institution between 1972 and 198243. Open fractures and hip fractures were excluded from the study. The time to fracture-healing was identified on the basis of radiographic parameters as well as clinical chart review (with healing indicated by a return to full weight-bearing). The times to healing in these diabetic patients were subsequently compared with the expected healing time for the same fracture type in the general population, as determined by a review of the available literature. Thirty-one fractures in twenty-eight patients who had been followed for an appropriate duration were identified. The average age of the patients was fifty-five years. Nineteen of the fractures were in female patients, and twelve were in male patients. Twenty-one fractures occurred in insulin-dependent diabetic patients, and ten occurred in patients who were taking oral hypoglycemic drugs or using diet to control the diabetes. The study included eight femoral fractures, eleven tibial fractures, and twelve ankle fractures. Overall, the average ratio of observed-to-expected healing times in this patient population was 1.63. The ratio was elevated for both insulin-dependent (1.57) and non-insulin-dependent (1.76) diabetic patients, for both male (1.77) and female (1.65) patients, for both patients older than the age of fifty years (1.43) and those younger than the age of fifty years (2.06), and for both patients with closed treatment of the fracture (1.42) and those with open treatment (1.86). More recently, Boddenberg reviewed both the available literature and his own series of diabetic patients with foot and ankle fractures44. Healing times, which were determined by both chart review and radiographic analysis, were noted to be slightly increased in this small group of eighty patients (average, 3.5 months compared with three months in nondiabetic patients). Boddenberg also found that healing of Charcot fractures was delayed an average of three months. Multiple basic-science studies have demonstrated a delay in healing in diabetic rat models. Macey et al. investigated fracture-healing in untreated diabetic rats, diabetic rats treated with exogenous insulin, and a control group of nondiabetic rats45. Two weeks after fracture, the diabetic rats were shown to have a decrease in callus size, tensile strength, stiffness, collagen content, and DNA content compared with the controls. Interestingly, treatment with exogenous insulin was able to restore the strength and stiffness of the fracture callus to control levels. Follak et al. investigated the influence of the diabetic metabolic state (well-controlled compared with uncontrolled diabetes) in a spontaneously diabetic rat model46. The rats were divided into a normoglycemic control group, a group with well-controlled diabetes (as determined by the degree of hyperglycemia and insulin requirements), and a group with poorly controlled diabetes. All studied parameters, including callus morphology and biomechanical findings, were worse in the poorly controlled group than in the well-controlled and normoglycemic groups. The diabetic group showed both a delay in cellular differentiation as well as a decrease in the peak failure load and stiffness of the fracture. Beam et al. investigated the effects of blood glucose control on fracture-healing in a control group, a spontaneously diabetic group, and a diabetic group treated with an insulin regimen47. Insulin administration was shown to restore the biomechanical properties, the degree of cellular proliferation, and the bone content of the callus to control levels. Again, the importance of close glycemic control was highlighted. More recent studies have been performed at the molecular and genetic level in order to elucidate clues to why fracture-healing is impaired in diabetics. Kayal et al. showed that, while initial cartilage formation was similar in diabetic and nondiabetic rat populations, with time there was a decrease in bone volume, callus size, and cartilage content that corresponded with an increase in the number of osteoclasts and an increase in cartilage resorption48. Other studies have shown a decrease in gene expression for the regulation of osteoblast differentiation and a decrease in local platelet-derived growth-factor levels leading to decreased cellular proliferation rates49. Additional studies have been performed to evaluate the potential use of adjuncts to enhance diabetic fracture-healing in the rat model. Gebauer et al. showed that, while low-intensity pulsed ultrasound did not have any effect on the degree of initial cellular proliferation in the fracture callus, it was able to significantly improve the biomechanical properties of the callus through enhanced matrix production (p = 0.0219 for torque and p = 0.0211 for stiffness)50. Gandhi et al. showed that local insulin delivery (without affecting systemic hyperglycemia) normalized both early (proliferative and chondrogenic) and late (mineralization and biomechanical strength) parameters of fracture-healing in rats51. On the basis of these findings, it was concluded that insulin itself plays a direct role in the healing process. Another study by Gandhi et al. showed that local administration of platelet-rich plasma normalized both early (cellular proliferation and chondrogenesis) and late (mechanical strength) parameters of fracture-healing in the diabetic rat model52.
Diabetic patients are at risk for both large-vessel arteriosclerosis and small-vessel angiopathy53. These vasculopathies lead to impaired oxygen delivery and local tissue ischemia. The relative hypoxia creates a poor environment for wound-healing. Collagen production and cross-linking are impaired, as is fibroblast function54. Because of potentially catastrophic wound-healing complications following the operative fixation of ankle fractures, a thorough evaluation of the preoperative vascular status is imperative. The initial evaluation of the vascular status of a diabetic patient with an ankle fracture should include palpation for the dorsalis pedis and posterior tibial pulses. In general, the presence of these pulses is considered a sign that distal blood flow is adequate for wound-healing. The absence of palpable lower-extremity pulses should trigger further evaluation of the patient's vascular status. Peripheral vascular disease can be measured noninvasively by calculating the ankle-brachial index, which is the ratio of the systolic blood pressure at the ankle to that of the brachial artery. In general, an ankle-brachial index of <0.90 suggests peripheral vascular disease and warrants referral to a vascular surgeon53. It should be noted that measurements of the ankle-brachial index may be falsely elevated in diabetic patients secondary to arterial calcinosis. An ankle-brachial index of >1.1 in a diabetic patient may suggest calcinosis, and additional testing should be pursued53. Transcutaneous oxygen measurements can also be extremely helpful in the initial evaluation of diabetic patients with an ankle fracture, particularly those with swelling and pain that preclude the acute measurement of the ankle-brachial index. A transcutaneous oxygen pressure (TcPO2) of >30 mm Hg is generally considered adequate for wound-healing53. A TcPO2 of <30 mm Hg is indicative of relative tissue hypoxia and should prompt a vascular surgery consultation for possible angiography or a revascularization procedure prior to surgical fixation of the ankle fracture. The importance of paying close attention to the vascular status of a diabetic patient with an ankle fracture cannot be overstated.
Many retrospective studies have demonstrated an increased risk of complications in diabetic patients who sustain an ankle fracture. Additionally, many authors have made recommendations for the treatment of these fractures on the basis of their own clinical experience. To our knowledge, however, there have been no randomized controlled trials comparing treatment modalities for displaced ankle fractures in diabetic patients. Costigan et al. reported on what we believe to be the largest clinical series of diabetic patients in whom the outcomes of ankle fractures were specifically investigated55. This was a retrospective case series (Level-IV evidence) of eighty-four patients with diabetes who had undergone open reduction and internal fixation of an unstable ankle fracture. The average age of the patients was 49.3 years, and the average duration of follow-up was 4.1 years. A total of fourteen complications developed in twelve (14%) of the eighty-four patients (both infection and Charcot changes developed in two patients). Infection (two superficial and eight deep) developed in ten (12%) of the eighty-four patients. Five of the eight deep infections required only one débridement and removal of hardware, two eventually resulted in a below-the-knee amputation, and one eventually resulted in Charcot changes requiring long-term bracing. Several patient variables, including sex, fracture pattern, whether the injury was open or closed, nephropathy, hypertension, vasculopathy, peripheral neuropathy, and diabetic control (insulin-dependent compared with non-insulin-dependent), were analyzed to determine their individual association with an increased risk of complications. Interestingly, only peripheral neuropathy and vasculopathy were associated with a significantly increased risk of complications (p < 0.0001 for both). Complications developed in ten of twelve patients with peripheral vascular disease (defined as an absence of pedal pulses) and in eleven of twelve patients with peripheral neuropathy. Blotter et al. performed a retrospective study (Level-III evidence) comparing the outcomes of operative fixation of ankle fractures between twenty-one diabetic patients and forty-six randomly selected controls matched for age, sex, and fracture severity56. A total of thirteen complications developed in nine (43%) of the twenty-one diabetic patients, whereas complications developed in seven (15%) of the forty-six patients in the control group. The complications in the diabetic group were also noted to be more severe, requiring seven additional procedures including two below-the-knee amputations (with a third patient refusing a recommended below-the-knee amputation); no additional procedures were needed in the control group. Overall, the risk of postoperative complications in the diabetic group was noted to be 2.76 times greater (95% confidence interval, 1.57 to 3.97) than the risk in the control group. Flynn et al. performed a retrospective study (Level-III evidence) comparing rates of infection associated with closed ankle fractures between seventy-three nondiabetic and twenty-five diabetic patients57. Patients with Charcot arthropathy were excluded from the study. In the diabetic group, infection was observed in four of six patients treated with closed reduction and a cast and in four of nineteen patients treated operatively. In the control group, infectious complications developed in none of five patients treated with closed management and in six of sixty-eight treated surgically. Peripheral vascular disease, diabetic neuropathy, and poor compliance with a diabetes treatment regimen were associated with an increased tendency toward infection. Jones et al. performed a retrospective study (Level-III evidence) comparing a group of forty-two diabetic patients with a closed ankle fracture with controls matched for age, sex, fracture type, and treatment modality (operative or nonoperative)58. Within the diabetic group, a secondary analysis was carried out between twenty-one patients who displayed diabetic comorbidities (nephropathy, retinopathy, neuropathy, vascular disease, a history of Charcot arthropathy, or a history of amputation) and twenty-one patients without any of those comorbidities. Patients with Charcot changes of the ankle at the time of the initial presentation were excluded from the study, whereas those with a history of Charcot neuropathy of the ipsilateral foot or contralateral foot or ankle were included. Complications developed in thirteen (31%) of the forty-two diabetic patients compared with seven (17%) of the forty-two patients in the control group; this difference did not reach significance (p = 0.133). Only the need for long-term bracing was noted to be significantly higher in the diabetic group (p = 0.025). In the subgroup analysis, diabetic patients without diabetic comorbidities had a complication rate comparable with that of the controls. However, in the group of diabetic patients with one or more major comorbidities, the complication rate (ten of twenty-one, 48%) was significantly higher than that of the controls (three of twenty-one, 14%) (p = 0.034). McCormack and Leith performed a retrospective study (Level-III evidence) comparing twenty-six patients who had an ankle fracture and diabetes mellitus with twenty-six nondiabetic controls59. A complication developed in eleven (42%) of the patients with diabetes compared with no patients in the control group. Within the diabetic group, wound problems developed in five of nineteen patients who had received operative intervention, and two of these wound problems eventually led to fulminate infection, amputation, and death. Five of seven diabetic patients who had been treated nonoperatively had a loss of reduction or malunion, although the authors stated that the malunions caused few clinical problems and all patients had a functional lower limb. Bibbo et al. performed a retrospective study of the operative management of ankle fractures in thirteen diabetic and forty-six nondiabetic patients (Level-III evidence)60. Complications developed in six (46%) of the thirteen diabetic patients compared with eight (17%) of the forty-six nondiabetic patients. Complications in the diabetic group included six superficial infections, three cases of Charcot neuroarthropathy, one delayed union, and one deep infection that resolved with intravenous administration of antibiotics. None of the patients in this small series required amputation or arthrodesis. In a study of Charcot neuroarthropathy (Level-IV evidence), Schon et al. presented the results of treatment of fifteen nondisplaced ankle fractures and thirteen displaced ankle fractures (exclusive of pilon-type fractures) in patients with neuropathy61. Of the fifteen nondisplaced fractures, seven were treated with either a cast or brace for three to nine months with non-weight-bearing for one to four months, five were treated with a cast or brace for two to four months with non-weight-bearing for zero to six months, and three that initially had undergone no treatment for six to twelve weeks were managed with delayed application of a cast or brace, which was worn for less than three months. All treatment regimens resulted in a stable, well-aligned ankle without the development of Charcot arthropathy in this group of patients with neuropathy and a nondisplaced ankle fracture. Of the thirteen displaced ankle fractures, four were initially treated with closed reduction and a cast or brace for three months. (Three of the four patients were noted to have been noncompliant with treatment recommendations in that they bore full weight.) Three of the four fractures eventually required arthrodesis, and the fourth required late open reduction and internal fixation at three months, leading to a good result. Nine patients with a displaced fracture were treated initially with open reduction and internal fixation. Two of the nine were managed postoperatively with immobilization for three months and non-weight-bearing for six weeks. The fracture progressed to an infected nonunion in one of these patients, and a Charcot valgus malunion requiring arthrodesis developed in the other. The remaining seven patients were treated postoperatively with immobilization for three to six months and non-weight-bearing for eight to twelve weeks. One patient required a free flap for wound closure. In another, a progressive malunion and osteonecrosis of the talus developed, eventually requiring an arthrodesis. The authors suggested that nondisplaced ankle fractures in patients with neuropathy usually heal when treated with a cast or brace and non-weight-bearing. However, displaced ankle fractures managed with closed means had high rates of progressive malunion or nonunion. The authors thus recommended open reduction and internal fixation of displaced fractures when possible. They also recommended long-term bracing for at least six months postoperatively. White et al. retrospectively reviewed the results in a series of fourteen open ankle fractures in thirteen patients with diabetes mellitus (Level-IV evidence)62. Wound complications developed in nine of the fourteen extremities, and ultimately five patients (six extremities) required a transtibial amputation for salvage. The open fracture healed uneventfully in only three of the fourteen extremities. An average of five operative procedures (range, two to nine) was performed in these patients. Lillmars and Meister performed a meta-analysis of five series of ankle fractures treated in diabetic patients and compared the results with those in nondiabetic controls63. A total of 127 diabetic patients underwent open reduction and internal fixation, and thirteen underwent closed management. Diabetic patients who had undergone open reduction and internal fixation had an overall complication rate of 30%, with an infection rate of 25%. The complication rate for those with nonoperative treatment of the fracture was ten of thirteen, with an infection rate of four of thirteen. Charcot arthropathy developed in six (7%) of eighty-three patients in whom Charcot changes were recorded as an outcome variable (all patients in this group were treated surgically). The overall amputation rate in diabetic patients treated either operatively or nonoperatively was 5% (seven of 139). The nondiabetic controls (223 patients treated operatively and five treated with closed means) had an overall complication rate of 7%, an infection rate of 6%, and an amputation rate of 0.4%. Given the high rates of nonunion and malunion following operative treatment of displaced ankle fractures in patients with neuropathy, several authors have advocated the use of supplemental fixation, consisting of transarticular fixation with Steinmann pins, multiple syndesmotic screws, or supplemental external fixation. Jani et al. performed a retrospective review (Level-IV evidence) of sixteen ankle fractures in fifteen patients with diabetes and neuropathy; the fractures were treated with supplemental calcaneal-talar-tibial fixation with Steinmann pins or screws in addition to standard techniques of open reduction and internal fixation (a fibular plate and fixation of the medial malleolus as necessary according to the AO technique)64. Postoperative management consisted of walking without weight-bearing with the limb in a short leg cast for twelve weeks, followed by removal of the Steinmann pins and partial weight-bearing for an additional twelve weeks. Overall, the major complication rate following treatment of these fractures was 25%. Two patients required amputation. No Charcot complications developed. A similar protocol with Steinmann pins crossing the subtalar and ankle joints was described earlier by Johnson65. In his Instructional Course Lecture, he recommended augmenting an unstable ankle fracture in a patient with Charcot arthropathy with one or two Steinmann pins placed across the ankle and subtalar joints to prevent the development of joint collapse and hardware failure. He also recommended cutting the pins off below the level of the plantar skin and removing them at six to eight weeks postoperatively. Perry et al. described their results in a retrospective case series (Level-IV evidence) in which six ankle nonunions in patients with neuropathy were salvaged with the use of a large-fragment 4.5-mm dynamic compression plate on the fibula with multiple 4.5-mm syndesmotic screws engaging two tibial cortices66. All patients were managed with immobilization and non-weight-bearing while walking for a minimum of three months postoperatively. The authors reported that an aligned and functional limb was achieved for all six patients, who were satisfied with the result. A similar protocol with multiple screws inserted from the fibula into the tibia was described by Schon and Marks38. Additional salvage operations following failed treatment of ankle fractures in patients with neuropathy include arthrodesis with the use of internal fixation or external fixation67-70. While a substantial number of these fractures do result in a fibrous union, the majority of patients are able to walk with the assistance of braces. In a recent large retrospective cohort study (Level-III evidence), the Nationwide Inpatient Sample (NIS) database was utilized to compare outcome measures in a group of diabetic patients who had undergone operative management of an ankle fracture with those in a nondiabetic control group71. The authors found a significant increase in in-hospital mortality, postoperative complications, length of hospital stay, and cost in the diabetic group (p < 0.001). The mortality rate was noted to be twenty-four (0.26%) of 9174 diabetic patients compared with 170 (0.11%) of 151,424 nondiabetic patients. Additionally, the inflation-adjusted cost was found to be $12,898 in the diabetic group compared with $10,794 in the nondiabetic group. In another recent prospective study (Level-I evidence), Egol et al. examined multiple factors and their effect on the short-term functional outcome of operative management of ankle fractures72. Diabetes was shown to portend a significant decrease in function at one year postoperatively. Overall, 154 (92%) of 167 patients without diabetes recovered >90% of function compared with only twenty-two (71%) of thirty-one patients with diabetes (p = 0.02).
Patients with diabetes mellitus have an increased rate of complications following both the operative and the nonoperative management of ankle fractures. The risk of complications is higher in patients with comorbidities (vasculopathy, neuropathy, or a history of Charcot arthropathy) than it is in either diabetic patients without comorbidities or nondiabetic patients. A complete neurological and vascular examination is imperative prior to pursuing operative treatment in the diabetic population. Patients with diminished or absent pulses warrant a vascular consultation before any operative intervention is undertaken. In order to achieve a successful outcome, the soft-tissue and vascular status of the limb must be optimized first. Some clinical recommendations73 can be made on the basis of this review (Table I).
1. Isolated stable, nondisplaced fractures of either the medial malleolus or the distal part of the fibula can be treated successfully with nonoperative management (Grade-B recommendation). Extreme vigilance and close follow-up with frequent radiographic and clinical examination are mandatory for a successful outcome. 2. Both operative and nonoperative management of unstable ankle fractures in diabetic patients have a high rate of complications. Operative management is more likely to result in a stable, functional lower extremity, and many unstable fractures that are initially treated with closed means will require surgical intervention in the future (either open reduction and internal fixation or arthrodesis) (Grade-B recommendation). 3. In diabetic patients without comorbidities, open reduction and internal fixation with use of standard orthopaedic fixation techniques can yield results comparable with those seen in patients without diabetes (Grade-B recommendation). 4. There is a trend toward the use of supplemental fixation, consisting of multiple syndesmotic screws, transarticular fixation, or supplemental external fixation devices, in patients with neuropathy and an ankle fracture. External fixation may be utilized as an off-loading device in patients who do not comply with weight-bearing restrictions. However, the data in the literature with regard to the efficacy of these supplemental fixation techniques are insufficient to make formal recommendations at this time (Grade-I recommendation). 5. Regardless of the treatment that is selected for an ankle fracture in a diabetic patient, the consensus is that a prolonged period of non-weight-bearing followed by protected weight-bearing is prudent (Grade-B recommendation).
Ankle fractures in diabetic patients with neuropathy remain a difficult problem, and more studies on the optimal treatment of these injuries are needed.
|
Top
Abstract
Introduction