The Incidence of Confounding Factors in Patients With Diabetes Mellitus Hospitalized for Diabetic Foot Ulcers

Introduction. Uncontrolled deformity, deep infection, and/or ischemia-hypoxia are highly associated with healing challenges of diabetic foot ulcers (DFUs). This paper reports the occurrences of these factors that the authors label the “Troublesome Triad” (TT) in a prospective series of 62 patients with diabetes mellitus (DM), who were hospitalized because of their DFUs. Materials and Methods. With Institutional Review Board approval, the authors gathered data in a prospective series of patients hospitalized because of lower extremity wounds. From this data, they analyzed the DFU cohort for the incidence of each of the components of the TT. The severity of the wound was graded with the authors’ 0 to 10 Wound Score in the patients who had components of the TT and compared with those who did not. Results. One or more components of the TT were observed in 57 patients (91.9%). As the number of confounders increased, mean Wound Scores decreased from 5.2 for 1 confounder to 2.9 for 3 confounders (P = 0.003). Most patients had 1 or 2 confounders (38.7% and 45.2%, respectively), while only 5 (8.1%) patients had all 3 confounders. Unresolved infection was the major confounder in 38 (61.3%) patients, uncontrolled deformity in 31 (50.0%), and ischemia-hypoxia in 26 (41.9%). Conclusion. For those patients with DM who were hospitalized because of DFUs, confounders that require remedial interventions were present in more than 90% of patients. Recognition and management of the TT eliminates wasteful uses of resources in an attempt to heal lower extremity wounds in patients with DM where the confounders need to be addressed first.

Diabetic foot ulcers (DFUs) are a common cause of morbidity and often require comprehensive, multidisciplinary management. The cost of care imposes substantial economic burdens on the health care system. It adds an additional $9 to $13 billion to the annual $245 billion spent on the care of patients with diabetes mellitus in the United States.¹ Diabetic ulcers also are a major risk factor for lower extremity amputations.² The progression from preulcer to ulceration to an infected, nonhealing wound is a common course of events and often leads to lower limb amputation.

The majority of DFUs in the cohort that did not require hospitalization healed uneventfully with management that included the appropriate wound dressing, debridement, and offloading. There are usually identifiable reasons why a DFU does not heal — often multifactorial. The authors have observed 3 confounding factors responsible for the failure of wound healing in more than 90% of cases (Figure 1). The authors term these factors the “Troublesome Triad” (TT), which include deep infection, uncontrolled deformity, and ischemia-hypoxia. 

Foot deformities are a cause of pressure concentrations and create biomechanical stresses that cause ulcerations.³ Deformities are considered uncontrolled when they are not adequately managed with offloading and protective footwear. Soft tissue breakdown follows repetitive cycles of pressure concentrations and/or shear stresses, especially in association with sensory neuropathy.⁴ Frequently observed deformities in the diabetic foot architecture include claw, hammer, and mallet toes; hallux valgus, varus, or rigidus; forefoot abduction/adduction; midfoot pronation/supination; plantarward extrusion of midfoot bones, equinus contracture, rocker bottom foot, or combinations of these problems. Bony prominences underlying mal perforans ulcers (MPU) often are a consequence of the deformities and are a reason DFUs do not heal. Charcot neuroarthropathy (CN) of the foot is consistently associated with these deformities.

Perfusion sufficient to meet oxygen requirements for wound healing and infection control is a second TT confounder. Pompers et al⁵ reported that 50% of DFUs exhibited a component of ischemia. The likelihood of wound healing without a major amputation is inversely related to the seriousness of the patient’s comorbidities, complexity of tissue involvement, and the severity of underlying peripheral arterial disease (PAD).⁶ The PAD evaluation starts with checking history for symptoms of intermittent claudication and rest pain. However, these may not be apparent because of sensory neuropathy. Components of the physical exam include checking for palpable pulses as well as secondary signs of perfusion such as assessment of pedal hair growth, skin quality, coloration, temperature, and toe capillary refill time. Based on clinical signs and symptoms, imaging studies and possibly juxta-wound transcutaneous oxygen measurements are done to assess the severity of PAD and provide justification for interventions. Nerone et al⁷ reported that PAD was the only significant risk factor for patients with diabetes, who had a partial foot amputation that subsequently failed and required a lower extremity amputation.

The third component of the TT is infection, which frequently occurs with DFUs. Infections pertinent to this component of the TT are those that are below the level of the skin and/or the base of the wound and include the involvement of bone, bursa, cicatrix or combinations of these. These deep infections typically do not respond to antibiotics and require surgical debridement to achieve healing. Inadequate debridement of deep infection is often the cause of nonhealing wounds. In a retrospective study of patients with diabetes and forefoot amputations, Synder et al⁸ reported that uncontrolled infection at the amputation site was the reason 37% of the patients subsequently required a transtibial or above-the-knee amputation.

A variety of wound dressing agents are used for wound management such as negative pressure wound therapy/subatmosphere wound dressings, bioengineered wound coverings, antimicrobials (including those with silver), medical grade honey, and agents that absorb secretions. When elements of the TT are present, failure of healing is likely even if the most advanced of these therapies are used. Consequently, the authors studied the cohort of patients with DFUs that required hospitalization because of nonhealing and/or concerns that an amputation is needed for the presence of 1 or more of the components of the TT. They also compared the severity of the patients’ wounds using their Wound Score with the number of components of the TT that were present. For the DFU, the authors feel it is incumbent the caregiver recognize the components of the TT and correct them before trying to achieve wound healing with prolonged, often costly interventions. Otherwise it is unlikely that healing will occur, and if it does occur the results will be durable while components of the TT persist.

With approval from the Institutional Review Board, the authors documented the severity of lower extremity wounds using a 0 to 10 Wound Score in a prospective series of 62 patients. A component of the study included recording the presence or absence of deep infection including bone, bursa, and cicatrix, inadequately controlled deformity and ischemia-hypoxia severe enough to interfere with healing. Exclusion criteria consisted of patients who did not agree to participate in the study, patients younger than 18 years of age, patients with tumors, or patients whose surgeons did not want their patients to be included in the study. Over an 18-month period, 62 patients who were admitted with DFUs were included in the database for this study.

The Wound Score is a 0 (worst possible) to 10 (best possible/healed) score utilizing the 5 essential assessments for evaluating a wound (Figure 2).⁹ Three of the 5 are derived from the most frequently used assessments to grade wounds, namely perfusion (Wagner¹⁰), wound depth (National Pressure Ulcer Advisory Panel¹¹) and infection (Infectious Diseases Society of America¹²). The other 2 most important assessments for evaluating a wound and predicting healing are appearance of the wound base and size of the wound. This provides 5 assessments for wound evaluation. By using a 3-grade continuum based on objective findings for each grade, with 2 points being optimal, 1 point being in-between, and 0 points being worst possible, the severity of the wound is accurately established. Wound scoring was done by the first author, an orthopedic foot and ankle surgeon who focuses on management of extremity wounds and deformities, an emergency medicine-hyperbaric medicine specialist, and by first- through third-year podiatry residents. 

The authors analyzed 62 hospitalized patients with DFUs. Ages in decades ranged from less than 40 to more than 80 with a mode of 51 to 60 (18 patients) and a mean of 58. Of these, 59% were male and 41% female. Ethnicities included Caucasian (38.1%), Hispanic (31.8%), African-American (22.2%), Asian (3.2%), and other (4.8%). The age, gender, and ethnic differences were not statistically significant with P values of 0.53, 0.38, and 0.41 for age, gender, and ethnicity, respectively.

One or more confounders were found in 91.9% of patients and 2 or more in 53.2% (Table 1). Only 5 (8.1%) of the 62 patients had no confounders. Residual deep infection was the most frequent confounder being present in 61.3% of the patients, followed by deformity in 50% and critical ischemia-hypoxia in 41.9%. The differences in penetrance of the 3 confounders were not statistically significant (P = 0.086). 

Wound Scores were matched with the number of confounders. The mean Wound Score (0 to 10 scale) was 4.9 (SD ± 1.71). Mean Wound Scores varied from 5.8 (SD ± 1.15) in the patients with no confounders to 2.9 (SD ± 1.02) for the patients with 3 confounders (Figure 3). The mean wound scores for 1, 2 and 3 confounders were 5.2 (SD 2.17), 4.9 (SD 1.34), and 2.9 (SD 1.02), respectively (Figure 2). Wound scores with confounders were compared with one-way analysis of variance (ANOVA), and it showed a mean difference in at least 1 pair of groups of confounders (F [3,55] = 2.77, P = 0.034). Tukey’s multiple comparison test was conducted to compare pairs of group means with Wound Scores. It showed the mean of 1 confounder group (5.2) was statistically significantly different (higher) from the mean of the group with 3 confounders (2.9) with P = 0.003 (95% CI, 0.177–4.4186). 

This prospective study examined the frequency of the components of the TT in 62 patients with DFUs. The components of the TT include uncontrolled deformity, deep infection (involving bursa, cicatrix, and/or bone), and ischemia/hypoxia that interfered with healing. The authors observed a high prevalence rate of components of the TT in patients with diabetes requiring hospitalization for DFUs. Each component of the TT requires specific interventions. The interventions should not be ignored or deferred in favor of using advanced therapies which fail to address the underlying problems of the TT.

Deep infection was the most common TT finding in the present study, being present in 38 (61.3%) of the 62 patients. Typical findings of deep infection include persistence of infection even with antibiotics and superficial debridement; induration and/or maceration around the wound margins; verrucous-cobblestone appearance of the ulcer base; persistent and/or recurrent fibrous membranes, hypertrophic granulation tissue, and/or biofilms; pale, atrophic appearance of the ulcer base; or combinations of these. The hallmark of deep infection is the persistence of the wound at repetitive clinic visits even when the wound appears improved after debridement. Often, these DFUs have been managed with a variety of advanced therapies before the decision is made for exploration and debridement surgery of the deep infection elements of bone, bursa, or cicatrix.

Plain X-rays and nuclear medicine scans using a combination of indium and technetium are useful in identifying the source of the deep infection. Magnetic resonance imaging (MRI) is helpful in delineating the soft tissue components of the infection, but tends to be over read since the interpretation of bone infection is based on bone edema. This typically involves more than the osteomyelitic bone because inflammation without infection extends to surrounding tissues including adjacent bone and is interpreted as such on the MRI.

Surgical management of deep infection requires removal of the infected bone, bursa, and cicatrix. Usually all 3 components are present in a nonhealing DFU, where infection is the reason the ulcer persists. With adequate debridement, the authors recommend antibiotics be continued for a minimum of 2 weeks after surgery to sterilize the soft tissues adjacent to the debrided bone and soft tissues. The indication for using hyperbaric oxygen as an adjunct to management of the deep infection problem requires careful consideration. If ischemia-hypoxia is evident (see later discussion on this component of the TT), there are concerns the host factors will not be able to sterilize any residual infection in the bone or soft tissues. This is because neutrophil oxidative killing of bacteria and bone resorption by osteoclasts are highly oxygen dependent. Hyperbaric oxygen should be used as an adjunct to the surgical and antibiotic managements in such situations.

Deformity was the second most frequent confounder in this series, present in 31 (50%) of the 62 patients. Although DFUs secondary to deformities often are a consequence of underlying bony prominences, deformities also occur secondary to CN bone destruction, motor components of peripheral neuropathy, malunion, cicatrix, and hypertrophic bursa formation. These complications often overshadow the extent of the underlying bony spur. Cicatrix and bursa formation develop as a defense mechanism by the body to provide padding over the deformity and can lead to a DFU.¹³ Often times this is self-defeating with the amount of cicatrix and/or bursa far exceeding the magnitude of the bone deformity. If the DFU is superficial, these deformities may only be a mechanical problem and not have a deep infection component, as previously discussed.

Consequently, the authors recommend surgical management of the deformity when the DFU persists after a trial of off-loading and/or protective footwear with as much attention being given to removing the cicatrix and bursa as to eliminating the underlying bony deformity. Although total contact casting (TCC) is recommended for outpatient management of DFUs, especially in forefoot locations, these ulcers typically occur because of underlying deformities Recurrence rates with TCC after healing and compliant use of protective footwear approach 50%.¹⁴ When the ulcers occur in the midfoot and hindfoot, TCC is usually not as effective as when ulcers occur in the forefoot. In these situations, surgery is required.

Mal perforans ulcers under the metatarsal (MT) heads are considered an important indication for TCC. Appreciation of the pathomechanics of the ulcer explains why surgical management should be proactive rather than reactive after the ulcer has progressed to a deep ulcer and osteomyelitis of the MT head, ascending tenosynovitis and/or progressive necrotizing soft tissue infection. The mechanics of this problem start with the MT head being depressed plantarward. This increases contact pressures between the plantar surface of the MT head and the forefoot skin and plantar pad. When the contact stresses exceed the integrity of the overlying soft tissues, a MPU develops from the inside (plantar surface of MT head) out. The primary reason the MT head becomes depressed is the diabetic motor neuropathy leads to clawing of the toes. The muscle imbalances lead to hyperextension at the MT-phalangeal joint, hiding of the plantar flexion creases of the toes (ie, hidden crease sign), hyperflexion at the toe interphalangeal joints, retraction of the toes dorsally on the forefoot, and subluxation of the proximal phalanges dorsally over the MT heads. The abnormal biomechanics result in increased plantar pressures on the MT head with standing and the stance phase of walking which can evolve to the MPU.

Staged management for the abnormal biomechanics should be done. When callus formation under the MT head is the only finding, the recurrent callus should be debrided and protective footwear to off-load the MT heads utilized. If a superficial ulcer even to the intact joint capsule is present, the MT head needs to be realigned. This can be done minimally invasively by percutaneously scoring the MT neck with a single drill hole through the skin and performing an osteoclasis through the drill hole in the bone to direct the MT head dorsally. The redirected MT head load shares with the adjacent MT heads while the fracture heals and prevents the transfer of MPU lesions to other toes. Typically the MPU heals in 3 to 4 weeks while full weight bearing with this technique. A more invasive technique to redirect the MT head is the Weil osteotomy which requires internal fixation and temporary non-weight bearing. If the MPU progresses to the point that osteomyelitis develops in the MT head, then debridement with MT head removal is required. This compromises load sharing with the other MT heads and subjects the patient to new transfer MPUs.

Ischemia-hypoxia is the third confounder that interferes with healing the DFU. In this 62-patient report, 26 (41.9%) had components of ischemia-hypoxia significant enough to interfere with healing their DFUs. This was based on failure for the wounds to improve and concomitant physical signs of ischemia in the foot. Hunt et al¹⁵ showed 30 mm Hg to 40 mm Hg oxygen tensions in the wound are required for healing to occur. At lesser oxygen tensions, the tissues may not die, but will be unable to heal the wound or eliminate the infection.¹⁶ The clinical detection of wound ischemia-hypoxia can be confirmed with Doppler ultrasound, which typically have poor quality signals and/or toenail bed capillary refill greater than 2 seconds. Other signs of ischemia-hypoxia include coolness or coldness and dusky or paleness discoloration of the skin of the distal forefoot and toes. Because of diabetic sensory neuropathy, ischemia-associated pain is often times not a complaint.

With failure for the wound to improve and clinical findings as just noted, angiography is the next step in the evaluation of critical limb ischemia-hypoxia. About one-third of patients with the ischemia-hypoxia confounder and associated wound healing problems have been revascularized and/or because of their advanced arterial disease are not candidates for revascularization. Juxta-wound transcutaneous oxygen tensions (TCOM) in room air and with hyperbaric oxygen provide objective criteria whether hyperbaric oxygen is needed for wound healing in these situations. If TCOM readings are > 40 mm Hg, wound oxygenation is sufficient for wound healing and failure to heal is likely due to one of the other confounders or an occasional biochemical problem (eg, matrix metalloproteinases). If  TCOMs are < 30 mm Hg, healing is not likely to occur. If juxta-wound TCOMs increase to > 200 mm Hg with a hyperbaric oxygen exposure, the authors observed healing in 88% of the study patients regardless of the room air TCOMs, and similar findings were reported in additional studies.^17,18

Other interventions to mitigate wound ischemia-hypoxia should not be overlooked in this cohort of patients. These include edema reduction, optimization of cardiac function, and use of medications to improve blood rheology. Edema increases the diffusion distance of oxygen from the capillary to the cell along a gradient.¹⁹ The benefits of improved cardiac function are self-explanatory in as much as this increases perfusion to the ischemic tissues. Finally, rheological agents such as aspirin, clopidogrel, warfarin, heparin, pentoxifylline/dextran, and similar agents improve perfusion through anticoagulation, decreasing slugging of erythrocytes in the microcirculation, and/or improving red blood cell deformity. When used as the only technique to improve perfusion, they will not likely be adequate to achieve wound healing of the ischemic-hypoxic wound. Consequently, they should be used in conjunction with the other methods to improve wound perfusion and oxygenation.

While the authors did not study correlations between the 3 wound healing confounders per se, they made several pertinent observations about associations between the confounders. First, patients with CN typically did not have ischemia-perfusion problems as the reason their wounds failed to heal. This supports the hypothesis that CN deformities develop as consequences of increased perfusion, mineral washout of the bones, and bone collapse secondary to osteopenia. Second, as expected, those patients with deformity-associated wounds typically had infected bursa and cicatrix, even though osteomyelitis was not usually present as long as the periostium and/or joint capsule were intact. Third, wound scores were not significantly lower in patients with 1 or 2 confounders, but were not so when all 3 confounders were present.

Several concerns may be raised in regards to this study. First, it was a prospective study of patients with DFUs serious enough to require hospitalization, but the study held no indication for blinding or randomization. However, almost all study patients had been previously treated with interventions that did not sufficiently address the confounders that required hospitalization. For example, off-loading techniques were used for deformities, but consideration was not given to removing the deformity and courses of antibiotics were used for infected wounds without thought being given to removing the infection focus. Second, outcomes, with respect to managing the confounders were not included in this paper, but will be a subject for future inquiry. Third, occasionally, either patients or their attending physicians did not give permission for the patient to be included in the study (patients were not included against their physicians’ wishes), and in 1 patient with diabetes, confounders were not recorded; however, these omissions were not felt to materially influence the authors’ observations of those patients who agreed to be studied. Fourth, although the authors offer generally accepted interventions in this paper to address the DFU confounders, management decisions for the DFUs were entirely those of the patients’ attending physicians/surgeons. Fifth, the seriousness of each confounder was not documented, merely its presence or absence was used in this analysis; however, the Wound Score tool was correlated with the number of confounders, which indirectly addressed the seriousness of the wound. Finally, the reliability of observations between 2 different investigators assessing the DFU confounders was not considered in this report. The objectivity of the findings of each confounder did not warrant reliability assessments between observers.

One or more components of the TT of deep infection, underlying deformity, and/or ischemia-hypoxia were present in 91.9% of the patients admitted to the authors’ hospital to manage DFUs. It is important to recognize and address the components of the TT. Deep infection was the most frequent confounder observed in this study. This was followed by deformity and, finally, ischemia-hypoxia. Patients with Charcot neuroarthropathy usually did not exhibit ischemia-hypoxia in their wounds; rather, deep infection and deformity were the causes for nonhealing. Each confounder has characteristic findings that can be confirmed with the examination and interventions that can remedy the situation. These include removal of the deformity and stabilization of the foot in a plantigrade position; debridement of the infected bone, bursa, and cicatrix; and improving perfusion with revascularization techniques, hyperbaric oxygen, and medical interventions. The Wound Score tool is useful for assessing the seriousness of the DFU and provides guidance as to whether salvage with the above interventions is feasible or a lower limb amputation is needed for definitive management.

The authors would like to thank Keith Panera, DPM for his assistance in formulating the IRB proposal, Diane Eisenstein for implementing the data processing, and to the following doctors of podiatric medicine for scoring wounds and collecting data: Karim Manji, Derrick Lew, Lisa Nhan, Jeremy Busch, Chris Jones, Anna Tan, and Susanna Chan. Their contributions were instrumental in bringing this paper to fruition even though none aided in the actual writing, data analysis, or editing of the manuscript.

Affiliations: Long Beach Memorial Medical Center, Long Beach, CA; and California State University, Long Beach, CA

Correspondence:
Michael B Strauss, MD
Long Beach Memorial Medical Center, Hyperbarics
2801 Atlantic Avenue 
Long Beach, CA 90801
mstrauss@memorialcare.org 

Disclosure: Drs. Strauss and Miller receive royalties from MasterMinding Wounds. Dr. Strauss also receives royalties from Diving Science.