|Home | About | Journals | Submit | Contact Us | Français|
The diabetic Charcot foot syndrome is a serious and potentially limb-threatening lower-extremity complication of diabetes. First described in 1883, this enigmatic condition continues to challenge even the most experienced practitioners. Now considered an inflammatory syndrome, the diabetic Charcot foot is characterized by varying degrees of bone and joint disorganization secondary to underlying neuropathy, trauma, and perturbations of bone metabolism. An international task force of experts was convened by the American Diabetes Association and the American Podiatric Medical Association in January 2011 to summarize available evidence on the pathophysiology, natural history, presentations, and treatment recommendations for this entity.
The Charcot foot in diabetes poses many clinical challenges in its diagnosis and management. Despite the time that has passed since the first publication on pedal osteoarthropathy in 1883, we have much to learn about the pathophysiology, and little evidence exists on treatments of this disorder. The international task force was convened in January 2011 at the Salpêtrière Hospital in Paris, France, to review the literature and report on the definition, pathogenesis, diagnosis, and treatment of the diabetic Charcot foot. Recommendations in this report are solely the opinions of the authors and do not represent the official positions of the American Diabetes Association or the American Podiatric Medical Association.
Charcot neuropathic osteoarthropathy (CN), commonly referred to as the Charcot foot, is a condition affecting the bones, joints, and soft tissues of the foot and ankle, characterized by inflammation in the earliest phase. The Charcot foot has been documented to occur as a consequence of various peripheral neuropathies; however, diabetic neuropathy has become the most common etiology. The interaction of several component factors (diabetes, sensory-motor neuropathy, autonomic neuropathy, trauma, and metabolic abnormalities of bone) results in an acute localized inflammatory condition that may lead to varying degrees and patterns of bone destruction, subluxation, dislocation, and deformity. The hallmark deformity associated with this condition is midfoot collapse, described as a “rocker-bottom” foot (Fig. 1), although the condition appears in other joints and with other presentations. Pain or discomfort may be a feature of this disorder at the active (acute) stage, but the level of pain may be significantly diminished when compared with individuals with normal sensation and equivalent degrees of injury.
The set of signs and symptoms that occur together with CN qualifies this condition as a syndrome, the “Charcot foot syndrome.”
There is no singular cause for the development of the Charcot foot, but there are factors that predispose to its development, as well as a number of likely precipitating events. The current belief is that once the disease is triggered in a susceptible individual, it is mediated through a process of uncontrolled inflammation in the foot. This inflammation leads to osteolysis and is indirectly responsible for the progressive fracture and dislocation that characterizes its presentation (1). The evidence to support this hypothesis is largely circumstantial. A neurally mediated vascular reflex leading to increased peripheral blood flow and active bone resorption has been proposed as an etiological factor in the development of bone and joint destruction in neuropathic patients. However, the relationship between increased blood flow to bone and active bone resorption has not been conclusively defined.
When a bone is fractured, the release of proinflammatory cytokines including tumor necrosis factor-α and interleukin-1β leads to increased expression of the polypeptide receptor activator of nuclear factor-κB ligand (RANKL) from any of a number of local cell types. RANKL triggers the synthesis of the nuclear transcription factor nuclear factor-κβ (NF-κβ), and this in turn stimulates the maturation of osteoclasts from osteoclast precursor cells. At the same time, NF-κβ stimulates the production of the glycopeptide osteoprotegerin (OPG) from osteoblasts. This “decoy receptor” acts as an effective antagonist of RANKL (2). The fracture will also be associated with pain, and this leads to splinting of the bone, and the rise in proinflammatory cytokines is usually relatively short-lived. In the person who develops an acute Charcot foot, however, the loss of pain sensation allows for uninterrupted ambulation, with repetitive trauma. It has been suggested that this results in continual production of proinflammatory cytokines, RANKL, NF-κβ, and osteoclasts, which in turn leads to continuing local osteolysis (1). This has subsequently been shown by an increase in proinflammatory phenotypes of monocytes in those with active Charcot foot when compared with diabetic control subjects (3).
Osteoclasts generated in vitro in the presence of macrophage colony-stimulating factor and RANKL from patients with active CN have been shown to be more aggressive and exhibit an increase in their resorptive activity compared with control subjects. However, these changes are only partially inhibited by OPG, indicating that other cytokines may also be important (4).
Neuropathy is a universal feature of the affected limb. Although it has been suggested that people with a Charcot foot may have particular patterns of sensory loss reflecting involvement of different fibers (5,6), this is not generally accepted. Nevertheless, three groups have shown that people who have had an acute Charcot foot exhibit retention of vasodilatory reflexes in contrast to diabetic individuals with distal symmetrical neuropathy without CN (7–9).
Despite these observations, it should be noted that the syndrome might also occur in patients with a spectrum of unrelated diseases complicated by nerve damage. These include distal neuropathies caused by toxins (ethanol, drug related) and infection (leprosy), as well as diseases of the spinal cord and nerve roots (tabes dorsalis, trauma, syringomyelia) and a number of other conditions (Parkinson's disease, HIV, sarcoidosis, rheumatoid disease, and psoriasis). Although the neuroarthropathy is typically more proximal in those with disease of the spinal cord, the presentation may be otherwise indistinguishable.
Loss of protective sensation will increase the likelihood of trauma to the foot, while motor neuropathy could result in altered structure of the foot (with exaggeration of the plantar arch and clawing) and changed gait with resultant abnormal loading.
Finally, it is possible that peptides normally secreted from nerve terminals are also important in the underlying pathophysiology. Of these, calcitonin gene–related peptide (CGRP) is a likely candidate because it is known to antagonize the synthesis of RANKL. Hence, any reduction of CGRP through nerve damage will result in an increase in RANKL expression. It is of particular interest that CGRP has been reported to be necessary for the maintenance of the normal integrity of joint capsules, and it follows that any reduction in CGRP release by nerve terminals could facilitate joint dislocation (10).
Because it is not possible to identify those most likely to develop the Charcot syndrome, it is impossible to determine with any degree of confidence whether preexisting osteopenia is a significant predisposing factor. One group has, however, reported an apparent reduction in bone mineral density (BMD) of the femoral neck in the contralateral (unaffected) limb at the time of presentation. The researchers also reported an association between BMD and the relative prevalence of fracture and of dislocation in the affected foot (11).
Diabetes may be associated with osteopenia, but the available evidence suggests that reduction of BMD is a feature of type 1 diabetes more so than type 2 diabetes (12). Any reduction of BMD in type 1 diabetes may relate to loss of islet peptides such as insulin and amylin (IAPP), both of which act as growth factors for bone. Despite this, the fracture risk in type 2 diabetes may be no less than in type 1 diabetes (12), and this would explain the fact that the presentation does not appear to differ between the two types of the disease. Any associated deficiency of vitamin D—with or without renal failure and secondary hyperparathyroidism—would increase the possibility of reduced BMD in diabetes. The use of thiazolidinediones could theoretically increase the likelihood of an acute Charcot foot through an effect on bone density, but this has not yet been reported. The use of corticosteroids as immunosuppressants in people with diabetes who have had a renal and/or pancreatic transplant (13) may explain the apparent high incidence of the Charcot foot in this group.
Although the Charcot syndrome may occur in a variety of conditions, diabetes is ostensibly the most common worldwide. Diabetes may predispose to its occurrence through a number of mechanisms. Apart from the presence of neuropathy and possible osteopenia, these include the effects of advanced glycation end products, reactive oxygen species, and oxidized lipids, which may all enhance the expression of RANKL in diabetes (10). The effect of local inflammation on this pathway would similarly compound the expression of RANKL. Furthermore, a single study has reported an apparent association between two OPG-related polymorphisms in people with a history of an acute Charcot foot in diabetes (14).
Many patients recall that the onset of the condition was precipitated by trauma that is often minor in nature (15). Other cases may be triggered by different causes of local inflammation, including previous ulceration, infection, or recent foot surgery. In this respect the occurrence of an acute Charcot foot as a complication of osteomyelitis is increasingly recognized in people with diabetes. Very occasionally, the onset of an acute Charcot foot may follow successful revascularization.
The initial manifestations of the Charcot foot are frequently mild in nature, but can become much more pronounced with unperceived repetitive trauma. Diagnostic clinical findings include components of neurological, vascular, musculoskeletal, and radiographic abnormalities. There have been no reported cases of CN developing in the absence of neuropathy. Accordingly, peripheral sensory neuropathy associated with reduced sensation of pain is the essential predisposing condition that permits the development of the arthropathy (16–19). Because of the very presence of insensitivity, a personal history concerning antecedent trauma is often unreliable (18,20,21). Typical clinical findings include a markedly swollen, warm, and often erythematous foot with only mild to modest pain or discomfort (16,18–20,22). Acute local inflammation is often the earliest sign of underlying bone and joint injury (23). This initial clinical picture resembles cellulitis, deep vein thrombosis, or acute gout and can be misdiagnosed as such. There is most often a temperature differential between the two feet of several degrees (20,24). The affected population typically has well preserved or even exaggerated arterial blood flow in the foot. Pedal pulses are characteristically bounding unless obscured by concurrent edema. Patients with chronic deformities, however, can develop subsequent limb-threatening ischemia. Musculoskeletal deformity can be very slight or grossly evident most often due to the chronicity of the problem and the anatomical site of involvement (16,17,19,25). The classic rocker-bottom foot, with or without plantar ulceration, represents a severe chronic deformity typical for this condition (16,26,27). Radiographic and other imaging modalities can detect subtle changes consistent with active CN.
Radiographs are the primary initial imaging method for evaluation of the foot in diabetic patients. Easily available and inexpensive, they provide information on bone structure, alignment, and mineralization. X-rays may be normal or show subtle fractures and dislocations or later show more overt fractures and subluxations. In later stages, the calcaneal inclination angle is reduced and the talo-first metatarsal angle is broken (Fig. 2). Medial calcification of the arteries is present in most Charcot feet and is a frequent secondary finding on radiographs (25). However, radiographic changes of CN are typically delayed and have low sensitivity (28).
Magnetic resonance imaging (MRI) allows detection of subtle changes in the early stages of active CN when X-rays could still be normal. MRI primarily images protons in fat and water and can depict anatomy and pathology in both soft tissue and bone in great detail. Because of its unique capability of differentiating tissues with high detail, MRI has a high sensitivity and specificity for osteomyelitis and has become the test of choice for evaluation of the complicated foot in diabetic patients (29). Although not required for diagnosis when X-rays are diagnostic for Charcot bone and joint changes, MRI is very useful in making the diagnosis at its earliest onset before such changes become evident on plain films.
Nuclear medicine includes a number of exams based on the use of radioisotopic tracers. Three-phase bone scans, based on technetium-99m (99mTc), are highly sensitive for active bone pathology. However, diminished circulation can result in false-negative exams and, perhaps more importantly, uptake is not specific for osteoarthropathy. Labeled white blood cell scanning (using 111In or 99mTc) provides improved specificity for infection in the setting of neuropathic bone changes (30), but it can be difficult to differentiate soft tissue from bone. Therefore, this exam can be combined with a three-phase bone scan or sulfur colloid marrow exam when superimposed osteomyelitis is suspected (31). More recently, positron emission tomography scanning has been recognized as having potential for diagnosis of infection and differentiating the Charcot foot from osteomyelitis (32,33). However, this remains investigational at this time.
Evaluation of bone density may be useful in those with diabetes to assess onset of CN as well as fracture risk. BMD can be assessed using dual-energy X-ray absorptiometry or calcaneal ultrasound. BMD has been related to the pathological pattern of CN, whereby joint dislocation is more prevalent in those with normal mineralization versus fracture in those with diminished BMD (11).
Experts agree that radiographs are important as the first exam in virtually all settings (33,34). However, a negative result obviously should not offer any confidence regarding lack of disease. In a patient with low clinical suspicion of osteomyelitis and no sign of CN on radiographs, either three-phase bone scan or noncontrast MRI is very effective at excluding osseous disease. If the patient has an ulceration with a high likelihood of deep infection, MRI is the best diagnostic modality. Nonetheless, one test may not be adequate for full evaluation. In this setting where MRI diagnosis is indeterminate, a subsequent labeled white blood cell scan can provide more specificity and should be correlated with clinical findings. The decision of nuclear imaging versus MRI is largely based on personal preference, availability, and local experience. In general, if metal is present in the foot, nuclear medicine exams are preferred, whereas diffuse or regional ischemia makes MRI the preferred exam.
The medical treatment of CN is aimed at offloading the foot, treating bone disease, and preventing further foot fractures (34). Because of the various etiologies of increased local bone resorption and/or secondary osteoporosis in patients with CN and limited randomized placebo-controlled trials in this area, treatment guidelines are largely based on professional opinion rather than the highest level of clinical evidence.
Offloading at the acute active stage of the Charcot foot is the most important management strategy and could arrest the progression to deformity. Ideally, the foot should be immobilized in an irremovable total contact cast (TCC), which is initially replaced at 3 days, then checked each week. Edema reduction is often remarkable in the first few weeks of treatment. The cast should be changed frequently to avoid “pistoning” as the edema subsides. If possible, the patient should use crutches or wheelchair and should be encouraged to avoid weight bearing on the affected side. The casting is continued until the swelling has resolved and the temperature of the affected foot is within 2°C of the contralateral foot (35). An alternative device for offloading the acute active stage of CN is a prefabricated removable walking cast or “instant TCC” technique, which transforms a removable cast walker to one that is less easily removed (36,37). It is important to take into consideration that TCC may actually have unfavorable consequences on the non-Charcot limb and induce unnatural stress patterns causing ulcerations and even fractures. Furthermore, patients with CN have increased instability and risk for falling and fracture as a result of multiple comorbidities including loss of proprioception and postural hypotension. Nonetheless, it should be noted that total immobility has disadvantages in itself with a loss of muscle tone, reduction in bone density, and loss of body fitness.
Duration and aggressiveness of offloading (nonweight bearing vs. weight bearing, nonremovable vs. removable device) are guided by clinical assessment of healing of CN based on edema, erythema, and skin temperature changes (34,35). Evidence of healing on X-rays or MRI strengthens the clinical decision to transition the patient into footwear. To prevent recurrence or ulceration on subsequent deformities, various devices are recommended after an acute or active episode has resolved, including prescriptive shoes, boots, or other weight-bearing braces. Frequent monitoring is required.
Treatment by antiresorptive drugs has been proposed because bone turnover in patients with active CN is excessive. However, there is little evidence to support their use, but both oral and intravenous bisphosphonates (38) have been studied in the treatment of CN in small randomized, double-blind, controlled trials (39,40) or in retrospective controlled studies (41). Some patients cannot tolerate oral bisphosphonates but may benefit from intravenous therapy using pamidronate or zoledronic acid (42). Intranasal calcitonin is another antiresorptive agent that has been studied in CN. This treatment was associated with a significantly greater reduction in cross-linked carboxy-terminal telopeptide of type I collagen and bone-specific alkaline phosphatase than standard treatment in the control group that received only calcium supplementation and offloading. Calcitonin has a safer profile in renal failure when compared with bisphosphonate therapy (43–46). However, a single dose of intravenous bisphosphonate generally does not require renal adjustment. There is no conclusive evidence for using bisphosphonates in active Charcot foot, and our understanding is evolving as more trials are currently underway.
There is limited evidence for the use of external bone stimulation in CN. Ultrasonic bone stimulation was reported for the treatment of CN of the ankle and for the healing of fresh fractures. Direct current electrical bone growth stimulators have been used specifically in CN patients undergoing arthrodesis and clinically tested to promote healing of fractures in the acute phase of CN in small case series. Although these findings are promising, there have been no subsequent studies to validate this method, and its use has been supported only as an adjunct therapy during the postsurgical period (45–49).
Surgical treatment of Charcot arthropathy of the foot and ankle is based primarily on expert opinion. Surgery has generally been advised for resecting infected bone (osteomyelitis), removing bony prominences that could not be accommodated with therapeutic footwear or custom orthoses, or correcting deformities that could not be successfully accommodated with therapeutic footwear, custom ankle-foot orthoses, or a Charcot Restraint Orthotic Walker (50). This clinical approach is based on expert opinion and small, uncontrolled retrospective case series. There has been an increasing trend in the literature to advise earlier surgical correction of deformity and arthrodesis, based on the assumption that surgical stabilization would lead to an improved patient-perceived quality of life (51).
Several investigators have suggested that Achilles tendon lengthening combined with total contact casting has the potential to decrease the deforming forces at the midfoot and decrease the morbidity associated with CN (52–57). Exostectomy offers the potential to reduce pressure caused by bony prominences. This treatment is often combined with accommodative bracing and appears to obtain more favorable results in patients without associated ulcers (58–60).
Surgery has generally been avoided during the active inflammatory stage because of the perceived risk of wound infection or mechanical failure of fixation. Two recent case series would suggest potentially favorable outcomes with early correction of deformity combined with arthrodesis (61,62). Most case series have focused on reconstruction of the deformity by reduction and arthrodesis using standard methods of internal fixation. Because of the poor bone quality, expert opinion has advised an extended period of nonweight bearing after surgery to account for the poor bone healing and inherent weakness of the underlying osseous structures. Early surgical series showed improvement in restoring a plantigrade foot and preventing recurrence of ulceration, although nonunion, failure, and loss of initial correction were common (63–67). The concept of an internal fixation “superconstruct” that extends internal fixation beyond the zone of fusion has evolved to address these issues (68). The combination of poor bone quality and a tenuous soft tissue envelope in a relatively immune-impaired population has led many surgeons to use a modification of the external fixation method of Ilizarov to correct deformity with a limited risk for surgical-associated morbidity (69–74).
Given the common failures of nonsurgical management of CN of the ankle, the task force members agree that surgical management could be considered a primary treatment. Surgical correction of deformity at the level of the ankle is likely more necessary due to the poor tolerance of deformity in the coronal plane (i.e., ankle varus or valgus) and the resultant prominence of the malleoli and their vulnerability to pressure-induced ulceration. Several small uncontrolled series have recommended augmented internal fixation followed by prolonged periods of immobilization and nonweight bearing in neuropathic patients who sustain acute ankle fractures (75–77). Acute ankle fractures in patients with complicated diabetes are associated with significantly higher rates of noninfectious complications and need for surgical revision when compared with diabetic patients without other organ system comorbidities (78). Numerous techniques have been reported without comparative effectiveness (79–82). All of the surgical studies are retrospective in nature without a control group and are based on a limited number of patients. While a strong, stable construct is required, inconclusive data exist to recommend one form of fixation over another (i.e., internal, external, or combined) in the surgical reconstruction of the foot and ankle in patients who are not infected.
The Charcot foot syndrome is a complex complication of diabetes and neuropathy. Its destructive effects on the foot and ankle begin with a cycle of uncontrolled inflammation. The classic rocker-bottom foot deformity is a late stage of the syndrome and can be avoided by early recognition and management. Offloading is the most important initial treatment recommendation. Surgery can be helpful in early stages involving acute fractures of the foot or ankle or in later stages when offloading is ineffective. An algorithm summarizing the approach to the Charcot foot can be seen in Fig. 3.
The American Diabetes Association/American Podiatric Medical Association Task Force meeting was supported by unrestricted educational grants from Small Bone Innovations, sanofi-aventis, and Integra LifeSciences. No other potential conflicts of interest relevant to this article were reported.
The authors thank Cécile Swiatek, chief librarian, from the J.-M. Charcot library at the Hospital Pitié-Salpêtrière for her assistance with the site preparation for the task force and references in this article.