Introduction

Diabetes mellitus (DM) affected approximately 382 million people worldwide in 2013 [1]. Diabetic complications including diabetic peripheral neuropathy (DPN) and peripheral arterial disease (PAD) remain prevalent in the USA and worldwide [2••, 3] and challenging to treat. Due to loss of protective sensation (LOPS) and impaired vascular supply, these can lead to serious foot complications including deformity, diabetic foot ulcerations (DFU), Charcot neuroarthropathy (CN), and infection [2••, 4, 5].

Among these, DFUs have a significant prevalence in the diabetic population. The lifetime risk of a diabetic patient to develop a foot ulcer is 34% [4]. More than 50% of DFUs become infected and 20% of DFU ultimately lead to amputation [4, 6]. In concert, CN remains a poorly understood and poorly recognized complication of DM. In a survey of faculty non-foot specialist providers, 70% had a self-described poor or complete lack of knowledge of this condition [7]. Up to 95% of CN cases are misdiagnosed [7, 8] and the delay in diagnosis leads to an up-to five-fold increase in complications including amputations and increased mortality risk [8,9,10,11].

Roughly 80% of the 100,000 amputations performed yearly in the USA are related to DM [12]. After years of decline, new evidence from the Center of Disease Control and Prevention (CDC) suggest that diabetes-related amputations have been rising for the past 4 years [13]. Mortality in patients who have major lower extremity amputation (LEA), below knee or above knee amputations, has a 5-year mortality above 60% [14]. Furthermore, mortality in patients with CN and neuropathic ulcer showed reduced life expectancy by 14 years [15].

In this review, we discuss mechanisms, clinical signs and diagnosis, challenges, and therapeutic strategies for both CN and DFU.

Diabetic Foot Ulcers

Mechanisms:

Diabetic foot ulcers are the result of multifactorial dysfunction. They are a frequent and severe complication of DM and the leading cause of LEA in the USA [16]. Peripheral diabetic wound complications cost $15 billion annually [17] and over 60% of the patients die within 5 years after major LEA [14].

Factors contributing to the development of a DFU include abnormal foot mechanics [18], peripheral neuropathy [2••], and PAD [19]. Asymptomatic neuropathy occurs in approximately 50% of patients with DM [2••]. Through the loss of distal sensory fiber loss, classically described as a ‘stocking-glove distribution’, diabetic patients often sustain repeated microtrauma. In combination with the aforementioned abnormal foot mechanics, this can lead to elevated pressures on the foot [19]. The shear stress on the foot along with focal pressure points leads to increased formation of DFUs (Fig. 1).

Fig. 1
figure 1

Shared mechanistic pathways of Charcot neuroarthropathy and diabetic foot ulcers

Full size image

Clinical Exam and Diagnosis of DFU:

History and Physical

The clinical diagnosis of a DFU begins with a complete assessment of the patient’s medical condition. A complete and thorough history and physical examination are necessary. Providers should incorporate an annual diabetic foot exam for all type 2 diabetics mellitus and all type 1 diabetics who have had the disease > 5 years [2••].

The annual diabetic foot exam consists of a history, assessment of pedal skin, vasculature, neurological testing, and musculoskeletal exam [20]. When evaluating the pedal skin, providers should note any areas of increased pressure with hyperkeratotic tissue and/or subdermal hemorrhaging. In addition, any foot deformities such as hammertoes, bunions, tailor’s bunions, high- or low-arched feet are to be noted and addressed.

If pedal pulses are weak or patient demonstrates intermittent claudication or rest pain, an urgent referral for non-invasive vascular studies is needed. Additionally, atrophied skin, pigment changes, reduced or absent pedal hair are suggestive of vascular disease. Non-invasive vascular studies help to establish a baseline and in a referral situation to a vascular surgeon should local ischemia or retarded healing be confirmed.

Neurological testing evaluates multiple nerve fibers in the distal extremity. We use both a 128-Hz tuning fork and Semmes Weinstein monofilament for small fiber assessment and temperature and sharp/dull testing for large fibers in the foot as recommended by the most recent statement by the American Diabetes Association [2••]. Deficits are noted, and education about risks associated with DPN is reviewed. It is estimated that up to 50% of all diabetic patients will develop peripheral neuropathy [2••]. If they have LOPS, more frequent visits are recommended.

Finally, a full assessment of the patient’s pedal joints is important. Range of motion should be supple and crepitant-free. When restriction or deformity is appreciated, excessive pressure may result peri-articular and lead to eventual skin changes, placing the patient at risk for ulcerative changes. A common example is seen with restricted first ray motion resulting in DFU to the plantar first metatarsal head [21].

Classification

Once the annual diabetic foot exam is performed, if a DFU is identified, a thorough assessment of the ulcer needs to be completed. We advise using a classification system to grade and assess ulcers consistently. Several common wound classification systems include the Wagner [22]; the University of Texas system (UTSA) [23]; and Wound, ischemia, and foot infection (Wifi) [24] classification systems.

The Wagner classification is an older, but ubiquitous system, originally used to depict the “dysvascular” foot [22]. It described depth of the wound, presence of osteomyelitis, and extent of gangrenous tissue (Table 1). However, it does not directly address ischemia and infection and is a major disadvantage.

Table 1 Diabetic foot ulcer classification systems
Full size table

The UTSA system takes into account the depth of the ulcer based on which layer of skin is penetrated and the presence of infection and/or ischemia (Table 1). The four depths (stages) are partial thickness, full thickness, tendon/capsule, and bone (0, 1, 2, 3, respectively). The grades include absence of infection and ischemia (A), presence of infection (B), presence of ischemia (C), and with both ischemia and infection (D). The depth of ulcer, infection status, and vascular status are combined to create a scale (i.e., 2B = infected full-thickness ulcer). Because it takes into account stage (depth), this is associated with increased risk of amputation and predictive of prolonged healing time [25].

Newer classification systems exist and further explore the importance of vasculature on wound healing. The Wifi system adds an additional decision-tree to the UTSA classification by including absolute toe pressures (TBIs). Absolute toe pressures are more reliable in diabetics to evaluate blood flow because ankle-brachial index can be obscured by vascular calcifications. Accounting for blood flow, wound depth, and presence of infection, the Wifi classification system predicts wound healing at 1 year in both crude and risk-adjusted analyses (Table 1) [26]. The Wifi system also demonstrated that increasing wound dimensions, presence of PAD, and longer time to initial diagnosis resulted in non-healing wounds [26].

Wound Evaluation:

A succinct description of the ulcerated skin relays vital information. It should include location, base consistency (granular, fibrotic, necrotic), depth, surface area (length x width) of the wound, outcome of probe-to-bone (PTB) testing, presence of undermining or tunneling, and a general assessment of the periwound skin (viable, non-viable, hyperkeratotic). Probe-to-bone testing should be performed with a sterile blunt metal instrument [27]. Pertinent assessment includes quantity of wound exudate, malodor, and signs of infection. However, the provider should be aware that diabetic foot infections (DFI) can be indolent because of reduced host response of cellular immunity [28, 29].

Diagnosis Challenges

Wound Healing:

Wound healing in DM is influenced by patient’s overall health status and molecular consequences of hyperglycemia. Many diabetics also have PAD [3] and PAD remains an independent risk factor for LEA. Because of decreased tissue perfusion distally, wound healing is impeded. Once the skin incurs a microscopic insult, it cannot appropriately heal. In fact, the healing process is complicated by the relative reduction in tissue perfusion and also as a result of the consequences of diabetic metabolism, such as hyperglycemia and formation of advanced glycation end products (AGEs). Hyperglycemia and AGEs can lead to impaired cell-mediated immunity and decreased phagocyte function [28, 29] further complicating healing.

Faithful Predictors of DFU Healing:

Determining the progress of (or lack thereof) DFU healing remains ambiguous. Rates of healing and amputation among patients with DFUs are variable and can be influenced by geographic location and local practice patterns [30, 31]. The most well established guideline to predict DFU healing is tracking the change in a wound’s surface area longitudinally. When a wound closes by more than 50% after 4 weeks of local wound care, it is a robust predictor of healing at 12 weeks [32]. This does not account for wound depth but it can guide additional management.

Another helpful determinant of healing is perfusion. Adequate pedal perfusion is indicated by TBIs > 45 mmHg and is associated with a higher chance of healing while ABIs < 0.5 or absolute ankle pressures < 50 mmHg places patient at higher risk of amputation [30].

Molecular approaches have shown that matrix metalloproteinases (MMP) and their tissue inhibitors (TIMP) play a major role in wound healing. The ratio of MMP-1/TIMP-1 strongly correlates to “good healing” with modest sensitivity and specificity [33]. However, this is heavily bench oriented and not feasible in a clinical setting.

The pathophysiologic understanding of the molecular and cellular changes underlying DFU healing and between- and within patient heterogeneity are largely unknown. The lack of robust prognostic biomarkers represents a major challenge.

DFU Recurrence

Diabetic foot ulcers heal with appropriate care and management. However, DFU recur and often times at high rates [34,35,36,37,38]. Commonly in published literature, patients with DFU are noted to be ‘in remission’ versus ‘cured’ when healing occurs [4].

Rates of DFU recurrence are noted to be 40% within 1 year of healing, 60% within 3 years of healing, and 65% within 5 years of healing [4]. Significant risk factors associated with recurrence are the presence of PAD [39], ulcer location on the plantar aspect of a toe [39], and presence of minor lesions (callus) [40].

The need to enroll patients who suffer, or have suffered, a DFU into surveillance programs, where regular monitoring of the patient’s feet can occur, is imperative. Our health care system requires all diabetic patients to have annual diabetic foot screenings. The aim of this best practice is to identify patients primarily who may be at greater risk and triage accordingly. This paradigm shift away from higher acuity care to an outpatient clinical setting is not a small task and requires a refocusing toward a multidisciplinary approach to succeed.

Multidisciplinary Care

The benefits of having a focused multidisciplinary approach to diabetic limb salvage are well established [41,42,43,44, 45•, 46,47,48]. When collaborative medicine is used to approach limb preservation, significant decreases in major LEA are realized [42, 45•, 49]. At our institution, the simple addition of podiatry services led to a significant decrease in major LEA while stabilizing the minor amputation counts [45•]. This becomes important because partial foot amputation (digit, ray, or transmetatarsal) are associated with decreased mortality compared with major LEA [50]. Financial benefits are also realized by better care organization. It has been shown that when podiatry services are excluded from a state Medicaid program, there is a 37.5% increase in DFI hospital admissions [51].

Therapeutic Approaches

The standard of care for DFU management includes eradication of infection, wound debridement, and offloading.

Infection Eradication

The risk of amputation in DFUs requires even subtle signs of infection be given the utmost attention. The Infectious Disease Society of America (IDSA) recommends classification of DFU based on infection severity and labels these as none, mild, moderate, or severe [52]. Risks factors for infection include: positive PTB test, DFU > 30 days, recurrent DFU, presence of PAD, previous amputation, LOPS, and kidney disease [52].

A DFU should be inspected and thoroughly debrided prior to assessing for infection. The IDSA system recommends that in DFU without clinical signs of infection (IDSA 1), antibiotic therapy is not warranted.

However, in a IDSA mild infection (IDSA 2), DFU are said to be superficial in depth with at least two signs of infection including: erythema 0.5–2 cm periwound, calor, local tenderness/pain, purulent discharge [52]. Patients remain hemodynamically stable with an IDSA mild infection. This stage warrants a wound culture with empiric antibiotic therapy against both Staphylococcus aureus and β-hemolytic Strepococcus. Narrow coverage once sensitivities are available [52]. Acting in this fashion may prevent progression to more severe infection, but this is not well studied in the literature.

In IDSA moderate and severe infection, erythema extends > 2 cm periwound with deeper tissue involvement [52]. The distinction between moderate and severe is systemic involvement [52]. When obtaining a wound culture, the culture should include the deepest tissue layer [52]. A discordance between superficial cultures and deep cultures, with the latter obtaining more gram-negative bacteria [53], is seen. If the appropriate culture is not obtained, providers may be undertreating DFI.

Wound Debridement

An important tenet of DFU ulcer care includes regular debridement of the wound. There are many ways to debride a wound: mechanical (surgical/sharp), enzymatic, larval, and ultrasonic/hydrosurgery. The benefits of debridement to the DFU are many; it stimulates healing, removes non-viable soft tissue, allows for full inspection of the wound, enhances contact between topical dressing preparations, and can be performed longitudinally to monitor wound progress [54,55,56]. The type of debridement chosen for each DFU is patient and situation dependent.

Offloading

Offloading, or taking pressure away from the wounded portion of the foot, is paramount for wound healing. Without pressure relief, DFU are unlikely to heal [14, 57]. The likelihood of a DFU healing is increased with offloading adherence and current evidence favors the use of non-removable casts or fixed ankle walking braces as optimum offloading modality [58]. However, as succinctly noted by a consensus team evaluating offloading strategies in DFUs, there is a gap between what the evidence supports regarding the efficacy of DFU offloading and what is performed in clinical practice despite expert consensus on the standard of care [58]. Another study showed only 2.2% of 221,192 visits involving DFU documented offloading [59].

Topical Wound Dressings:

The number of wound care products to assist in DFU healing is astounding. The products available include alginates, foams, hydrocolloids, silicone, polyurethanes, astringents, hydrogels, honey, and bioengineered skin alternatives (BATs) to promote wound healing [60]. The products available maintain a moist wound environment to improve outcomes [61]. Several important trials demonstrating efficacy with various products to heal DFU are included in Table 2 [62,63,64,65,66,67,68,69,70].

Table 2 Summary of select trials reporting various treatment modalities on diabetic foot ulcers
Full size table

Detailed knowledge of utility of each dressing is necessary but cumbersome and may not have peer-review support. The International Working Group of the Diabetic Foot (IWGDF) evaluated such evidence [71]. Over 2100 studies they evaluated demonstrated poor methodological quality with lack of a blinded assessment and inadequate control. According to the IWGDF, with the exception of negative pressure wound therapy in post-operative wounds, there is little evidence to justify the use of newer therapies for DFUs [71].

Likewise, a Cochrane database evaluated 13 systematic reviews that accounted for 17 relevant randomized controlled trials (RCTs). They concluded that there is no robust evidence for differences between wound dressings for any outcome in DFUs regardless of setting [72].

Therefore, the need and call for more well-planned and executed research is desperately needed to improve our understanding and outcomes of utility of each wound care product. An algorithmic approach may be beneficial to help elucidate some of these differences yet determined by literature reviews.

Charcot Neuroarthropathy

Mechanisms:

Charcot neuroarthropathy is a devastating orthopedic condition that afflicts patients with diabetes. The true incidence or prevalence of this condition is not known. However, estimates demonstrate incidence to be between 0.1 and 0.9% [73,74,75]. Originally described by Jean Marie Charcot in the late nineteenth century as an end-stage sequelae of neurosyphilis, severe destruction of foot architecture was observed. This unfortunate disease is now most commonly seen in the diabetic patient due to neuropathy occurrence.

The mechanisms influencing the disease are multifactorial and are only beginning to be understood on a molecular level. Historically, Jean Marie Charcot described the condition as having a neurovascular etiology [76]. He thought vasomotor nerve centers caused an alteration of bone and joint nutrition [76]. Later, German scientists described the condition occurring because of neurotraumatic influences. As our understanding of the biochemical processes evolved, it is learned that these theories each are entangled and influenced by other systemic states like inflammation, bony regulation, neuropeptides, and more recently microparticles (Fig. 1).

Clinical Exam and Diagnosis

Physical Examination:

Instituting an annual (diabetic) foot exam can be used as a time to screen individuals for CN. By definition, patients with CN will have a component of peripheral neuropathy. In addition to LOPS, the CN patient will often complain of an erythematous and edematous foot. They may or may not have an ulcer present. In either case, we believe providers should include CN higher in the differential diagnosis when examining a warm and swollen foot.

Temperature differences between limbs can guide diagnosis. By using the patient’s contralateral limb, as a control, providers can narrow the location of a Charcot event using dermal thermometry. In a study with 39 patients with acute CN, the foot with CN was 8.8 ± 2.3 °F warmer [77]. The local calor was strongly correlated with radiographic changes [77, 78] and markers of bony turnover [79, 80]. However, there is no specific biomarker or laboratory test to confirm or suggest CN. The provider must rely on their clinical judgment and a full assessment with imaging modalities.

Imaging:

Radiographs have classically been used to recognize CN. CN has been classified according to radiographic acuity [81] and anatomic location [82]. Eichenholtz describes CN on presence of osseous destruction and disorganization in three stages [81]. A fourth stage was later added to include a prodromal state, known as stage 0, which was characterized by an absence of radiographic changes but clinically suspicious [83]. Most CN occurs in the TMTJ [82, 84] but in theory can occur throughout the foot.

Newer approaches include the use of magnetic resonance imaging (MRI) to diagnose the condition earlier. The most recent system accounts for clinical, MRI, radiographic, and histologic changes [85]. This system is useful to monitor progression during both inactive and actives phases of CN. However, its full implementation and reach is not widespread.

When MRI cannot determine whether the event is related to CN or osteomyelitis, a bone scan is useful. In particular, the addition of sulfur colloid to a technetium bone scan can differentiate between infection and CN [86, 87] .

Diagnosis Challenges:

Early Recognition

Early recognition of CN is important to reduce the risk of serious complication. Initially, the delay in diagnosis of CN arises from a lack of understanding and a lack of clinical expertise. Non-specialist provider knowledge of the condition is limited [7].

Another factor contributing to the delay of CN diagnosis is the presence of non-specific clinical findings. Acute CN may present with erythema, edema, and calor but chronic CN may not have inflammatory signs. Patients may also have pain which is counterintuitive given the concomitant diagnosis of peripheral neuropathy. Given the inconsistency of CN symptoms, the provider needs to be aware of any foot deformity which may warrant additional work-up and accommodation.

Thirdly, the differential diagnosis for CN includes infection, osteomyelitis, deep vein thrombosis, and cellulitis among others. The differential diagnosis may present concomitantly with the CN. Approximately 58% of patients with CN present with a Charcot-related DFU [88•]. As a result, the more frequent diabetic foot complication diagnoses delay the immediate recognition of CN and therefore postpone care [89]. This important distinction indicates that primary detection of the disease is difficult and indirect.

Realistic Expectations:

The main stay of treatment for CN is conservative therapy. Surgical intervention is reserved when obtaining a braceable and plantargrade foot is not feasible. Most patients with CN are managed with offloading of the affected foot. Offloading should entail placing these patients into TCC to reduce pedal pressures from the affected foot. Patients return to physical activity and walking once skin temperature difference as compared to the contralateral limb is < 4 °F for 2–4 consecutive weeks. Unfortunately, recurrence of the CN event or Charcot foot is approximately 15–30%. In fact, 21% have contralateral occurrence within the first 3 years. Pinzur et al. followed 198 patients (201 ft) with CN. Eighty-seven patients were treated non-surgically while 60 ft were treated surgically [90]. At a minimum of 1-year follow-up, 87 of the 147 midfoot Charcot patients achieved desired endpoint without surgery (59.2%). The remaining 40.8% required surgery [90].

CN treated in a structured, intensive, and non-operative manner was associated with an approximately 2.7% annual rate of amputation, a 23% risk of requiring a brace for > 18 months, and a 49% risk of recurrent ulceration. Limbs with open ulcers at initial presentation or chronically recurrent had an increased risk for higher level amputation [91].

Predictors of Failure

The factors that contribute to CN failure leading to amputation are poorly understood; however, several have been identified. First is the presence or absence of a neuropathic foot ulcer. In one study, 193 wounds were evaluated in patients with both CN and a DFU over a 7-year period. The group was able to heal 65% of wounds using both surgical and non-surgical approaches. Unfortunately, even with aggressive wound care and offloading attempts, 35% of patients underwent higher level amputation when both DFU and CN were present [92]. Patients with CN and a DFU demonstrate a worse prognosis and have up to a 12× increased risk of amputation as compared to those with CN alone.

Second, in addition to the presence of CN, DFU location is important. Hindfoot or ankle DFU in CN has an increased risk of amputation. For instance, in the aforementioned study, DFU occurring in the hindfoot or ankle accounted for 75% of all the higher level amputations performed [92].

Thirdly, obtaining regular interval radiographs in CN becomes important and a prognostic for success of non-surgical management. Wukich et al. evaluated 114 patients with midfoot CN with and without DFU and made nine reproducible radiographic measurements to differentiate the cohorts. Radiographs demonstrating destabilization of the lateral column of the foot, as measured by a decreased cuboid height (cuboid bone closer to ground upon weight bearing), decreased calcaneal inclination, and decreased lateral calcaneal fifth metatarsal angle, were more likely to be associated with ulceration and potential for loss of limb [93].

Therapeutic Approaches:

To modify the underlying mechanisms regulating the disease state, immobilization is used during active CN. Expert consensus and standard of care require patients with CN to be immobilized using total contact cast (TCC). TCC for the treatment of CN can stabilize greater than 50% of patients and avoid limb-threatening complications or unnecessary surgical intervention [94,95,96].

The authors believe that the mainstay of future management of CN will be dictated by a better understanding of the molecular mechanisms. The evidence supporting these molecules to challenge standard of care is not complete. A summary of pharmacologic trials in CN is described in Table 3. Brief descriptions are provided and mentioned below [97,98,99,100,101]. The targets which could become important in developing a more granular approach to CN may include the role of inflammation, bony composition, neuropeptides, and circulating microparticles on the disease state.

Table 3 Summary of pharmacologic trials on Charcot neuroarthropathy
Full size table

Inflammation

In combination with DPN, trivial amounts of trauma, such as activities of daily living, can lead to increased inflammation. Increased inflammation causes a proinflammatory state and the soft tissues respond with release of acute phase reactants like tumor necrosis factor-alpha (TNF-α), interleukin 1β (IL-1β), and receptor activator of nuclear factor-Kβ (RANK) ligand (RANK-L) [102, 103]. Briefly, because of the proinflammatory state, osteoclastogenesis is induced and there is dysregulation of bony turnover [102, 103].

Given this observation, a recent study demonstrated that the effect of a single dose of RANK-L antibody on acute CN is real [101]. Fracture resolution was significantly shorter after denosumab was given compared to standard of care cohort [101].

Bony Composition

In addition to inflammation, bony composition is altered. The osseous composition is different in patients with CN. Histologically, the bone in diabetes with CN displayed inflammatory and myxoid infiltrates with a disorganized trabecular pattern versus diabetics and non-diabetics [104]. The number of trabecular bone was not necessarily different from that of a diabetic patient without CN [104].

Two medications have been evaluated to regularize bony turnover: bisphosphonates and intranasal calcitonin [99, 105]. However, pharmacological benefit has not been fully demonstrated [105].

Neuropeptides

Nitric oxide (NO) and calcintonin gene-related peptide (CGRP) also influence CN. CGRP is well studied and is active in regulating osteoblastic activity. CGRP is expressed by osteoblasts endogenously and may have an autocrine loop [106]. CGRP is shown to inhibit proinflammatory cytokine production and increase the release of IL-10 by monocytes. Unfortunately, patients with CN are deprived of CGRP and local regulation is therefore compromised [38, 107] .

Similarly, nitric oxide is a free radical gas that functions in many pathways in biology. It can influence bone metabolism [108]. Again, the relative lack of NO available to cells in patients with CN is noted by the decreased expression of eNOS. This decreased level of NO leads to suppression of osteolcastic activity and contributes to an increase in the fragility of osteoporotic bone [108]. Furthermore, newer studies suggest that NO has a biphasic effect on osteoclasts, promoting osteoclastogenesis in low concentrations and acting inhibitory at higher concentrations [109]. No clinical study has been attempted to evaluate in vivo effects.

Microparticles

Circulating microparticles are major mediators of cardiovascular complications in patients with DM. A recent study revealed that microparticles from patients with CN have a high content of inflammatory cytokines G-CSF, GM-CSF, IL-1-ra, and IL-2 [110•] which increase the differentiation of monocytes into osteoclasts in vitro. The exact function or role of microparticles in vivo in patients with CN is not known currently. It could be an eventual therapeutic target or a diagnostic tool to enable a provider to liberate patient’s ambulation from protected weightbearing to full weightbearing [110•].

Conclusions

In conclusion, diabetic foot disease is fraught with complications including ulceration, infection, amputation, and death. To improve the success of treating DFU and CN depends on the astute provider to recognize the disease and treat accordingly [7]. Once these conditions are identified, focused intensive diabetic limb salvage care will potentially avoid both costly and untoward complications of these diseases.