Scottish Foot Ulcer Risk Score Predicts Foot Ulcer Healing in a Regional Specialist Foot Clinic

The study was approved by the Tayside Ethics Committee. All patients had diabetes as defined by the World Health Organization and were registered on the web-based Tayside Regional Diabetes Register (21). Health care professionals undertaking foot examinations, such as diabetologists, podiatrists, general practitioners, practice nurses, district nurses, and others, had participated in an education program delivered by podiatrists or doctors with an expertise in diabetes and foot care. Health care professionals had been shown how to use the foot assessment tool and how to record the information electronically on the web-based database. The details of the foot risk score are described elsewhere (17). In brief, it integrates information on previous history of foot ulceration, assessment of pulses, sensation to 10-g monofilament, presence of foot deformity, and inability to self-care (Fig. 1). The software automatically calculated the patient's individual degree of foot risk and categorized them into low, moderate, and high risk of foot ulceration. Patients underwent foot risk assessment as part of routine clinical care in hospital and general practice–based diabetes clinics within Tayside. All electronic information was shared by health care professionals using the web-based shared electronic record, which is 97% sensitive with a positive predictive value of 97% for the diagnosis of diabetes (21). Foot risk data were collected from the year 2000 onwards.

Information on patients attending the specialist foot clinic with foot ulceration was collected from 1 April 2004 until 31 December 2006. A foot ulcer was defined as a full-thickness skin break below the level of the malleoli. The following information was collected on each patient: patient age and sex, ulcer site, and the most recent foot risk score. Ulcer size was not measured. Clinical examination was performed by consultants in diabetes and vascular surgery to record absence of pulses, absence of sensation to a 10-g monofilament, presence of foot deformity, depth of ulcer using the University of Texas ulcer classification system (22), and presence of sepsis. Sepsis was defined clinically as surrounding cellulites or presence of pus. Information on outcome of ulcer and duration of ulcer from arrival at the specialist foot clinic until final outcome was recorded. Healing was defined as complete re-epithelialization of the wound. Nonhealing was defined as any patient requiring minor or major amputation, patients with a nonhealing ulcer, or a patient who died with their ulcer. A retrospective analysis was performed to determine the predictive nature of the individual clinical criteria and the overall risk score on foot ulcer healing.

In the data analysis, the most recent risk stratification score was used, or the risk score immediately before the onset of any ulcer. The unit of analysis was the foot ulcer. Patient demographics and details of ulcer site and outcome were compared for risk factors using χ² tests for categorical variables, t tests for continuous data in dichotomous groups, and ANOVA for continuous data in more than two groups. Duration of ulcer had a skewed distribution and was presented by median and the interquartile range and analyzed using the Wilcoxon test. Potential predictors of a foot ulcer not healing (male sex, age, ulcer site, ulcer depth, high-risk score, absent pulses, neuropathy, foot deformity, and sepsis) were each assessed in a univariate regression analysis. Then all the variables were entered into a multivariate regression model. All analyses were carried out using SAS version 8.

Within Dundee and Angus, 7,184 patients with diabetes have had a foot risk assessment performed and recorded electronically (51.5% of all patients with diabetes). Of these, 62.9% were recorded as being at low risk, 23.9% at moderate risk, and 13.1% at high risk of developing foot ulceration. During the study period, there were 221 referrals to the specialist foot clinic for foot ulceration in 198 patients over 33 months. These were all separate episodes, and if there were simultaneous ulcers, the one that took longest to heal was included as the index ulcer. The mean age of patients was (means ± SD) 67.3 ± 12.7 years. Of these 221 ulcers, 165 (74.7%) healed, 7 (3.2%) became chronic nonhealing ulcers, 8 (3.6%) healed after minor amputation, 18 (8.1%) required major amputation, and 23 (10.4%) died with their ulcer. Further analysis was performed per referral (n = 221). If patients had multiple ulcers at referral, the ulcer that took the longest time to heal was classified as the index ulcer for further analysis. If a patient had a healed ulcer and represented with a new ulcer, they were categorized as a new ulcer event.

Of 221 referred ulcers, 97 were superficial (43.9%), 40 were deep (18.1%), and 84 were down to the bone (38.0%). In addition, 117 ulcers (52.9%) were infected, while 104 (47.1%) were not. Superficial ulcers were more likely to heal than deep ulcers or ulcers that went down to the bone (88, 68, and 63%, respectively; χ²; P < 0.001). Noninfected ulcers were more likely to heal than infected ulcers (81 vs. 69%; P = 0.02). Of all ulcers, 151 were in male subjects (68.3%) and 70 were in female subjects (31.7%). There was no difference in mean age of male (67.2 ± 11.8 years) and female (67.5 ± 14.6 years) subjects.

Of 221 ulcers, the most recent previous foot risk score was low risk in 5 subjects (2.3%), moderate risk in 52 subjects (23.5%), and high risk in 164 subjects (74.2%). Overall, 154 (69.7%) had absent foot pulses, 181 (81.9%) had absence of sensation to 10-g monofilaments, 33 (14.9%) had a foot deformity, and 88 (39.8%) had a previous foot ulcer. Of 221 ulcers, 108 (48.9%) were sited on a toe, 41 (18.6%) at a metatarsal head, 3 (1.4%) on the dorsum of the foot, 54 (24.4%) on the heel, and 15 (6.8%) elsewhere. The numbers in each risk category in the general population were 943 for high risk (13.1% of 7,184), 1,720 for moderate risk (23.9% of 7,184), and 4,521 (62.9%) for low risk. For patients in the community, the rate of attending the specialist foot clinic with a foot ulcer was 171 per 1,000 for high-risk patients (164 of 943), 30.2 per 1,000 for moderate-risk patients (52 of 1,720), and 1.1 per 1,000 for low-risk patients (5 of 4,521). The rate of referral to the specialist foot clinic with a foot ulcer for high-risk patients was 157 times higher than that of low-risk patients and 5.8 times higher than that of moderate-risk patients. Moderate-risk patients were 27 times more likely to be referred than low-risk patients.

Data on the impact of foot ulcer risk score on healing of foot ulcers is shown in Table 1. Patients at high risk of ulceration were less likely to heal than low- or moderate-risk patients (68 vs. 93%; P < 0.0001), although there was no difference in the site of ulceration for such patients (Table 1). Patients at high risk of ulceration were slightly older and took longer for their ulcers to heal (Table 1). The impact of absent foot pulses and presence of neuropathy on healing is shown in Table 2. The rate of healing for patients with absent pulses was similar to those with neuropathy (89.7 vs. 91.7%), but in patients who had both clinical criteria, the healing rate was significantly lower (61.2%; P < 0.0001). Patients with absent pulses were more likely to develop ulcers on their toes, while metatarsal head ulcers were more common in those with neuropathy (Table 2). Healing rates, ulcer sites, and time to healing were similar between those with and those without previous foot ulceration (Table 2). Although twice as many male than female subjects presented with foot ulceration, the rates of healing were similar between the sexes (Table 1). Male subjects were more likely to have toe ulcers, while female subjects were more likely to develop heel ulcers. For 33 patients with a foot deformity, 26 healed (78.8%), while in 188 patients without a foot deformity, 139 healed (73.9%; no significant difference).

For patients categorized as high risk of ulceration, absent foot pulses, neuropathy, increasing age, deep ulcer, or sepsis were all significant risk factors for a nonhealing ulcer on univariate analysis (Table 3). On multivariate analysis, absent pulses, neuropathy, increasing age, and ulcer depth remained significant.

Although the foot risk tool was developed (18) and validated (17) to predict the likelihood of foot ulceration in routine clinical practice, it appears that it is also a predictor of healing for patients who subsequently develop a foot ulcer. The healing rate was 68% in high-risk patients and 93% in low- and moderate-risk patients. The site of foot ulceration was no different between the different foot risk groups. Of five criteria comprising the risk score, only neuropathy and absent pulses were significant predictors of poor healing, showing that these were the main useful criteria within a high-risk score that predicted poor healing. Although neuropathy or absent foot pulses on their own were associated with healing rates of 93 and 88%, respectively, the combination of neuropathy and absent pulses was a good predictor of poor healing (61%). Our data show how a combination of neuropathy and ischemia predicts much poorer outcome than either criteria alone.

It was a surprise to find that having a previous foot ulcer or having a foot deformity were not predictors of poor healing. The majority of these patients had neuropathy but not vascular disease. This combination of features seems to put patients at risk of recurrent ulceration (22,23) but with ulcers that tend to heal well when managed appropriately. Although age is not part of the risk score, it was a predictor of poor ulcer healing, and nonhealing has previously been associated with age (23).

It was noticeable that 68% of referrals were male, when 53% of the diabetic population are male in Tayside. It has been shown that female subjects are more active in self-care and prevention for acute diabetes-related foot problems, while male subjects are more likely to seek help from others and take a passive approach (24). This may mean that female subjects develop fewer foot ulcers, and, when they do, they may be more likely to successfully self-manage simple ulcers. In our study, there was a nonsignificant trend toward female subjects having a higher prevalence of neuropathy, suggesting that female subjects may have more advanced diabetic foot disease when they present with foot ulceration. This may also explain why there was no sex difference in the healing rates of foot ulcers, if female subjects have more advanced disease but are better at caring for their own feet (24). Alternatively, foot ulcers in male subjects may just respond well when cared for by others.

The majority of patients referred to the specialist foot clinic with a foot ulcer had previously been identified as being at high risk of foot ulceration. Patients identified as being at high risk were 157 times more likely to attend the specialist foot clinic with an ulcer than low-risk patients. The corresponding value for moderate risk was 27 times that of increased risk. These relative risks are greater than those observed in our prospective community based follow-up of all ulcers, where which the risks were 83- and 6-fold increased, respectively (17). This may be because in the community study, many of the foot ulcers in the low-risk patients healed relatively rapidly and were not referred to the specialist foot clinic and therefore would not be included in the current analysis, which may exaggerate the relative difference in outcome between low- and high-risk patients. Of patients attending the foot clinic with a foot ulcer, 68% had already been identified as being at high risk and 98% at high or moderate risk. High-risk patients represent 12% of the diabetic population, and high-/moderate-risk patients represent 36% of them. This is reassuring that the foot risk tool is effective at identifying patients likely to ulcerate. In addition, high-risk groups of patients could be specifically targeted with preventative measures, such as education or more regular podiatry, to try and avoid future foot ulcers.

Ulcer healing rates were marginally better in this study at 75% compared with other reported series, with a healing rate of 65–67% (23,25–27), even though there was a high rate of arteriopathy (70%) and the average patient age was 67 years. This may be because we followed-up our patients until there was final outcome, rather than having a defined duration of follow-up. However, in our study the amputation rate of 12% and death rate of 10% were similar to the rates reported in the literature of 8–17% amputation and 4–17% deaths (23,25–27). These remain alarmingly high. We also know that patients with diabetes who undergo amputation are subsequently more likely to die and suffer further amputations than nondiabetic amputees (28). Overall, the patients seen in our specialist foot clinic are similar to those seen in other foot clinics. Predictors of an ulcer not healing were similar to previous studies. The main predictors of nonhealing appear to be ischemia (our data, 23,29–31), age (our data, 23), ulcer depth (our data, 29,30,32), and ulcer size (29,33,34). Sepsis has been associated with amputation, especially in the presence of ischemia (26), and in our study was a borderline predictor of a nonhealing ulcer. This may reflect a fairly aggressive use of antibiotics. Previously, neuropathy has not been a predictor of poor healing (23,29); however, it is a predictor of reulceration (17,22,23). In the current study, neuropathy did predict poor healing, but we demonstrate that that appears to be mainly when in the presence of ischemia rather than in patients with isolated neuropathy. Previously, it has been shown that toe ulcers heal better than foot ulcers (30), but our data showed similar healing rates.

There are other standardized clinical assessment tools that both monitor ulcer progression and predict healing (22,23). These are ideal in the setting of the specialist foot clinic. However, in the community setting, where many patients are cared for by nonspecialists, the current foot risk tool could be a useful initial guide to clinicians about foot ulcer healing, without the need to understand and remember new and more complex grading schemes. It could be recommended that patients with high-risk feet who develop ulcers should be prioritized toward specialist foot clinics at an early stage because of the lower chance of healing. This may be particularly helpful in areas with restricted availability of health care resources. Once in the specialist foot clinic, the more sophisticated schemes to monitor progression and predict healing of the ulcer would be the most appropriate to use.

This study highlights the difficulties in achieving foot ulcer healing in patients with diabetes. Increasing age and distal peripheral vascular disease are particularly difficult clinical problems. We have also demonstrated that the Scottish foot ulcer risk score predicts healing of foot ulcers, as it provides an integrated assessment of arteriopathy and neuropathy. The majority of patients referred to the specialist foot clinic with an ulcer have already been identified as being at high risk, providing an opportunity for targeted educational with such patients.

We thank all the clinical staff who performed clinical assessments.