1
Epidemiology and Economic Impact of Foot Ulcers

Edward J. Boyko1 and Matilde Monteiro-Soares2

1 VA Puget Sound Healthcare System and the University of Washington, Seattle, WA, USA

2 MEDCIDS and CINTESIS, Faculty of Medicine, University of Porto, Porto, Portugal

1.1 Introduction

Diabetes mellitus accounts for the majority of nontraumatic lower limb amputations [1], despite affecting a minority (9.4%) of the U.S. population [2]. A systematic review of risk of lower limb amputation in persons with and without diabetes in a defined population yielded 19 publications reporting results from Europe, UK and the US with relative risk of amputation in diabetes ranging from 7.4 to 41.3 [3]. Notably this range of relative risks for amputation exceeds that for diabetes and fatal coronary heart disease of 3.5 (women) and 2.1 (men) from a meta-analysis of 37 prospective studies [4], and is of similar magnitude to that seen for the association between ever smoking and lung cancer [5]. The chief antecedent to diabetic lower limb amputation is the nonhealing foot ulcer [6], which precedes approximately 80% of diabetic amputations, and which has been reported to lead to amputation in 15% of cases [7-9]. The problem of higher amputation risk in diabetes can be traced backwards in the causal chain of events to the development of a diabetic foot ulcer (DFU) [6].

1.2 Diabetic Foot Ulcer (DFU) Definition

Diabetic Foot Ulcer (DFU) is defined as a full thickness skin defect below the ankle that is slow to heal or non-healing. Amongst 1000 patients with DFU enrolled in the Eurodiale Study and followed for up to one year, median time to healing was 147?days for toe (95% CI 135-159?days), 188?days for midfoot (95% CI 158-218?days) and 237?days for heel ulcers (95% CI 205-269?days) [10], with healing success at one year of 79% for plantar and 73% for non-plantar ulcers. The available data demonstrate prolonged healing times for DFU and the frequent occurrence of healing failure.

The study of the epidemiology of DFU or any health condition requires a case definition to identify afflicted persons. There is no established case definition for DFU in general use. Definitions of DFU in the published literature may include features such as full thickness skin defect distal to the malleoli, but in some reports no case definition is provided with the identification of DFUs left to the judgement of foot care providers; or through review of podiatry records or electronic medical record data. Although delayed healing is a recognized feature of DFU, there is no established criterion for 'slow to heal' and rarely is a duration criterion included in the ulcer definition. The Seattle Diabetic Foot Study required a healing time greater than 14?days to meet the criteria for a DFU [11], whilst a clinical trial of custom footwear specifically mentioned that duration of ulceration was not be taken into consideration [12]. Consensus on a DFU definition would enable comparisons of the frequency of this complication across populations, regions, and over time.

1.3 DFU Classification

DFU classification systems function as a guide to the best treatment and as a predictor of the probability of wound healing and the need for amputation. The typical components of a wound classification system include wound depth, presence and severity of infection, gangrene, and ischemia. A systematic review of DFU classification systems identified 15 such systems reported in 25 articles [13]. Examples of the systems for which the greatest number of validation studies have been performed are the Meggitt-Wagner (9 validations), and the University of Texas and S(AD)SAD systems (5 validations each). Features of these three systems are seen in Table 1.1.

Comparisons of the ability of different classification systems have generally found them to be predictive of healing and amputation [14, 15].

Little research has been conducted comparing these classification systems using methods commonly employed for other diagnostic modalities such as comparison of area under receiver operating characteristic curves (AUROC), or by comparing test characteristics that alter pre-test probability, such as likelihood ratios [16, 17]. Jeon et al. compared five different DFU classification systems on the ability to predict amputation [18] that included the systems shown in Table 1.1, except that a simplified version of the S(AD)SAD system was used (SINBAD). All systems showed excellent ability to predict amputation with AUROC ranging from 0.85 to 0.89, and positive and negative likelihood ratios ranging from 4 to 18 and 0.21 to 0.41, respectively. Monteiro-Soares et al. compared 11 different prediction models for lower limb amputation in a prospective study of 293 patients with DFU and found positive and negative likelihood ratios ranging from 1.0 to 5.9, and 0.1 to 0.9, respectively, and AUROC curves ranging from 0.53 to 0.83, indicating that considerable diversity exists with some models showing significantly poorer prediction performance [19]. Further research is needed comparing existing classification systems to predict healing and amputation in patients with DFU.

DFUs are often described without using a classification system, but in reference to likely ulcer aetiology (neuropathic versus ischemic versus both) or foot location [20]. Ulcers usually develop over bony prominences that can be found on the plantar and dorsal foot surfaces especially if a structural deformity has developed resulting in abnormally high arch, prominent metatarsal heads, and clawing of the toes. The most common location for ulcer in the Eurodiale Study was on toes, with nearly equal division between plantar and non-plantar surfaces. Ulcers on the heel took longer and were less likely overall to re-epithelialize than other foot locations, but no difference was observed in overall healing success comparing plantar to non-plantar locations [10]. Similar healing success was seen amongst 405 patients with a neuropathic ulcer in plantar (n = 175, 91% healed) and non-plantar (n = 230, 94% healed) locations [21].

Table 1.1 Diabetic foot ulcer wound classification systems.

Wagner Stage 1 Superficial ulcer of skin or subcutaneous tissue 2 Ulcers extend into tendon, bone, or capsule 3 Deep ulcer with osteomyelitis, or abscess 4 Gangrene of toes or forefoot 5 Midfoot or hindfoot gangrene University of Texas Wound description No infection or ischemia Infection present Ischemia present Infection and ischemia present Superficial 1A 1B 1C 1D Penetrates to tendon or capsule 2A 2B 2C 2D Penetrates to bone or joint 3A 3B 3C 3D S(AD)SAD - Size (Area, Depth), Sepsis, Arteriopathy, Denervation SIZE Grade Area Depth Sepsis Arterial disease Neuropathy 0 Skin intact Skin intact None Pulses present Pin prick intact 1 < 1?cm2 Skin and subcutaneous Surface Pulses diminished or one missing Pin prick reduced 2 1-3?cm2 Tendon, joint Cellulitis Absence of pedal pulses Pin prick absent 3 > 3?cm2 Bone, joint space Osteomyelitis Gangrene Charcot deformity

1.4 DFU Incidence and Prevalence

A wide range of estimates is available for DFU incidence and prevalence. A recent systematic review of the global literature identified 67 publications from 33 different countries and 5 continents [22]. Foot ulcer prevalence ranged from 1.5 to 16.6% in populations that included inpatients, outpatients, diabetes clinics, and defined communities. Identification of foot ulcers was based on self-report, examination, medical record review, or electronic diagnostic codes (ICD-9). The report highlights the difficulties in assessing the frequency of this complication worldwide given the inconsistent methodologies employed.

The capture of this complication in a large population or nationally requires use of electronic diagnostic codes. The value of this information depends on the completeness and accuracy of such codes. The sensitivity and specificity of five such methods were recently estimated by comparison to medical record reviews of 512 patients...