852 Care Volume 38, May 2015

Predictors of Lower-Extremity Kristy Pickwell,1 Volkert Siersma,2 Marleen Kars,1 Jan Apelqvist,3 in Patients With an Karel Bakker,4 Michael Edmonds,5 Per Holstein,6 Alexandra Jirkovska,´ 7 Infected Edward Jude,8 Didac Mauricio,9 Alberto Piaggesi,10 Gunnel Ragnarson 11 12 Diabetes Care 2015;38:852–857 | DOI: 10.2337/dc14-1598 Tennvall, Heinrich Reike, Maximilian Spraul,13 Luigi Uccioli,14 Vilma Urbancic,15 Kristien van Acker,16 Jeff van Baal,17 and Nicolaas Schaper1

1Division of Endocrinology, Department of Inter- nal , Maastricht University Medical Centre, Maastricht, the Netherlands 2Research Unit for General Practice and Section of General Practice, Department of Public OBJECTIVE Health, University of Copenhagen, Copenhagen, Infection commonly complicates diabetic foot ulcers and is associated with a poor Denmark 3 outcome. In a cohort of individuals with an infected diabetic foot ulcer, we aimed Department of Endocrinology, University of Malmo,¨ Malmo,¨ Sweden to determine independent predictors of lower-extremity amputation and the 4International Diabetes Federation, Consultative predictive value for amputation of the International Working Group on the Di- Section and International Working Group on the abetic Foot (IWGDF) classification system and to develop a risk score for predicting Diabetic Foot, Heemstede, the Netherlands 5 amputation. Diabetic Department, Kings College Hospital, London, U.K. 6Copenhagen Wound Healing Centre, Bispebjerg RESEARCH DESIGN AND METHODS Hospital, Copenhagen, Denmark We prospectively studied 575 patients with an infected diabetic foot ulcer pre- 7Diabetes Centre, Institute for Clinical and Exper- senting to 1 of 14 diabetic foot clinics in 10 European countries. imental Medicine, Prague, Czech Republic 8Diabetes Centre, Tameside General Hospital, RESULTS Ashton-under-Lyne, U.K. 9Department of Endocrinology and Nutrition, Among these patients, 159 (28%) underwent an amputation. Independent risk Hospital de Sant Pau, Autonomous University factors for amputation were as follows: periwound edema, foul smell, (non)pu- of Barcelona, Barcelona, Spain rulent exudate, deep ulcer, positive probe-to-bone test, pretibial edema, fever, 10Sezione Dipartimentale Piede Diabetico, Dipar- and elevated C-reactive protein. Increasing IWGDF severity of infection also in- timento di Area Medica, Azienda Ospedaliero- Universitaria Pisana, Pisa, Italy dependently predicted amputation. We developed a risk score for any amputation 11Swedish Institute for Health Economics, Lund, and for excluding the lesser toes (including the variables sex, pain on Sweden 12 palpation, periwound edema, ulcer size, ulcer depth, and peripheral arterial dis- Innere Abteilung, Mariannen Hospital, Werl, Germany ease) that predicted amputation better than the IWGDF system (area under the 13Mathias-Spital, Diabetic Department, Rheine, ROC curves 0.80, 0.78, and 0.67, respectively). Germany 14Policlinico Tor Vergata, Department of Internal PATHOPHYSIOLOGY/COMPLICATIONS CONCLUSIONS Medicine, Rome, Italy 15 For individuals with an infected diabetic foot ulcer, we identified independent Department of Endocrinology, University Med- fi ical Centre, Ljubljana, Slovenia predictors of amputation, validated the prognostic value of the IWGDF classi ca- 16Department of Endocrinology, H Familie tion system, and developed a new risk score for amputation that can be readily Ziekenhuis and Centre de Sante´ des Fagnes, used in daily clinical practice. Our risk score may have better prognostic accuracy Rumst and Chimay, Belgium 17 than the IWGDF system, the only currently available system, but our findings need Department of Surgery, Twenteborg Ziekenhuis, Almelo, the Netherlands to be validated in other cohorts. Corresponding author: Kristy Pickwell, k.pickwell@ mumc.nl. Infection is a frequent of diabetic foot ulcers, with up to 58% of ulcers Received 3 July 2014 and accepted 7 January being infected at initial presentation at a diabetic foot clinic, increasing to 82% in 2015. patients hospitalized for a diabetic foot ulcer (1). These diabetic foot infections This article contains Supplementary Data online (DFIs) are associated with poor clinical outcomes for the patient and high costs at http://care.diabetesjournals.org/lookup/ suppl/doi:10.2337/dc14-1598/-/DC1. for both the patient and the health care system (2). Patients with a DFI have © 2015 by the American Diabetes Association. a 50-fold increased risk of hospitalization and 150-fold increased risk of lower- Readers may use this article as long as the work extremity amputation compared with patients with diabetes and no foot infection is properly cited, the use is educational and not (3). Among patients with a DFI, ;5% will undergo a major amputation and 20–30% a for profit, and the work is not altered. care.diabetesjournals.org Pickwell and Associates 853

minor amputation, with the presence of score for amputation in individuals the hallux, excluding amputations of dig- peripheral arterial disease (PAD) greatly with a DFI from characteristics of the its 2–5. We performed analyses for both increasing amputation risk (4–6). Fur- foot (ulcer), the leg, and the patient of these outcomes, as amputations of thermore, lower-limb amputation is asso- that can readily be assessed by the clini- digits 2–5 could be considered function- ciated not only with significant morbidity cian at presentation to guide further ally less important for mobility and qual- and mortality but also with major psycho- management. ity of life and as part of treatment rather social and financial consequences (7–9). than a measure of poor outcome. As infection of a diabetic foot wound RESEARCH DESIGN AND METHODS Statistical Analyses heralds a poor outcome, early diagnosis Study Design Unadjusted associations between the and treatment are important. Unfortu- Between 1 September 2003 and 1 October 1-year amputation incidence and patient, nately, systemic signs of inflammation 2004, all patients with diabetes presenting leg, ulcer, and infection characteristics such as fever and leukocytosis are often with a new foot ulcer to any of 14 diabetic were tested with x2 tests (categorical absent even with a serious foot infection foot centers in 10 European countries characteristics) and t tests (continuous (10,11). As local signs and symptoms were included. Both in- and outpatients characteristics). The characteristics stud- of infection are also often diminished, were included. The design and rationale ied included sex, age, immobility, serum because of concomitant peripheral neu- of the multicenter, observational, prospec- creatinine, HbA1c, presence of PAD and ropathy and ischemia (12), diagnosing tive Eurodiale study have previously been polyneuropathy, ulcer size, depth, dura- and defining resolution of infection can described in detail (17). We excluded pa- tion and location, periwound redness, be difficult. Amputation, instead of res- tients treated in the participating centers (periwound) edema, pain, foul smell, olution of symptoms or signs of infec- for an ulcer of the ipsilateral foot during exudate/pus, increased local skin tem- tion, could therefore be a reliable the previous 12 months and those with a perature, lymphadenitis/lymphangitis, outcome measure. life expectancy of ,1 year. fever, and C-reactive protein (CRP) levels. The system developed by the Interna- Patients were followed monthly until The relation of the different character- tional Working Group on the Diabetic healing of the foot ulcer(s), major ampu- istics with amputation was analyzed in Foot (IWGDF) and the Infectious Dis- tation, or deathdup to a maximum of multivariable Cox proportional hazards eases Society of America (IDSA) provides 1 year. Healing was defined as complete regression models. In these analyses, a criteria for the diagnosis of infection epithelialization of the whole foot at hazard ratio (HR) .1 indicates a higher of ulcers and classifies it into three cate- two consecutive visits. All patients amputation rate for the corresponding gories: mild, moderate, or severe. The were treated according to protocols category of a certain characteristic com- system was validated in three relatively based on the International Consensus pared with a baseline category. For in- small cohorts of patients with a DFI on the Diabetic Foot (19). stance,anHRof2meansthatatany (13–15), and increasing severity of infec- Ulcer characteristics were described ac- time point during follow-up, the patients tion was associated with toe amputa- cording to the PEDIS system of the IWGDF, in the corresponding category are twice tions in one large cohort (16). While the which classifies foot ulcers according to as likely to undergo an amputation as the system seems valid in predicting ampu- five categories: perfusion, extent, depth, patients in the baseline category. tation, it does not take into account the infection, and sensation (19,20). The patient demographics and foot, ul- results of diagnostic tests such as probing Ankle-brachial index (ABI) ,0.9 was cer, and infection characteristics were used to bone and does not include more gen- considered to represent the presence to construct risk scores for any amputation eral patient characteristics that may of PAD. A deep ulcer was defined as a and for amputations excluding lesser toes have a major impact on the outcome lesion of the skin extending through the in patients with an infected ulcer. Firstly, a of a DFI. subcutis with visible muscle, tendon, or score corresponding to a full model was The European Study Group on Diabe- bone, whereas a superficial ulcer was defined as the linear term of a logistic re- tes and the Lower Extremity (Eurodiale) limitedtotheskin.Inpatientswitha gression model on the 1-year amputation prospectively studied a large cohort of deep ulcer, the probe-to-bone test was incidence that included all demographics, patients with a diabetic foot ulcer (17), performed, yielding two categories of foot and ulcer characteristics, and infection enabling us to determine the prognostic deep ulcers: those with a positive and characteristics. Thereafter, this full model value of the IWGDF system for clinically negative probe-to-bone test. Infection was reduced by iterative (backward) elim- relevant lower-extremity amputations. was diagnosed according to the IWGDF ination of the least significant predictor un- A previously published study of this co- system by the presence of two or more til all predictors were significant (P , 0.05); hort identified determinants of minor of the following signs: frank purulence, the final risk scores are a version of the amputation but did not specifically fo- local warmth, periwound redness, lym- reduced models where the parameters cus on patients with an infected diabetic phangitis, periwound edema, pain or are rounded to increase usability. The full foot ulcer (18). The aims of the current tenderness on palpation, or fever (21). models and the risk scores were evaluated, study were to 1) identify which signs of For the current study, only the data of together with the IWGDF/IDSA classifica- infection independently predict ampu- patients with an infected foot ulcer di- tion, with receiver operating characteristic tation within 1 year in patients from agnosed by the IWGDF/IDSA guidelines (ROC) curves. the Eurodiale cohort presenting with a were analyzed. We defined two sepa- DFI, 2) determine the predictive value of rate outcomes: 1) any amputation, includ- RESULTS the IWGDF system for amputation in this ing all minor and major amputations, and Five hundred and ninety-one (57%) of cohort, and 3) develop an alternative risk 2) amputations proximal to and including 1,033 patients who completed the 854 Amputation in Infected Diabetic Foot Ulcers Diabetes Care Volume 38, May 2015

Eurodiale study had an infected ulcer, variables were associated with ampu- independent predictors of any amputa- and data were complete for 575 tations excluding the lesser toes, with tion were as follows: periwound edema, (97.3%) of them. One hundred and the exception of pus and lymphadenitis/ HR 2.01 (95% CI 1.33–3.03); foul smell, ninety-nine patients had a grade 2 lymphangitis. Incidence of amputation HR 1.74 (1.17–2.57); purulent and non- (mild) infection, 338 a grade 3 (moder- was also directly related to increasing purulent exudate, HR 1.67 (1.17–2.37) ate), and 38 a grade 4 (severe). Ampu- severity of infection according to the and 1.49 (1.02–2.18), respectively; tations were performed on 159 (28%) IWGDF system (P , 0.01), with 50% of deep ulcer, HR 3.49 (1.84–6.60); patients (126 minor and 33 major) patients with systemic inflammatory positive probe-to-bone test, HR 6.78 within the year of follow-up; 103 pa- signs (IWGDF grade 4 infection) undergo- (3.79–12.15); pretibial edema, HR 1.53 tients (18%) underwent amputations ing any amputation and 40% undergoing (1.02–2.31); fever, HR 2.00 (1.15–3.48); proximal to and including the hallux. amputations excluding the lesser toes. elevated CRP levels but less than three Immobility, serum creatinine, HbA1c, times the upper limit of normal, HR 2.74 Variables Associated With presence of polyneuropathy, and ulcer (1.40–5.34); and elevated CRP levels Amputation duration and location were not associ- more than three times the upper limit, Several general patient and ulcer char- ated with amputation. HR 3.84 (2.07–7.12). acteristics were univariately associated The independent predictors of ampu- with any amputation as depicted in Signs of Infection Independently tations proximal to and including the Table 1. The incidence of amputation Predicting Amputation hallux were as follows: the presence of increased with increasing redness, peri- Table 2 shows the results of Cox regres- either periwound or pretibial edema, HR wound or pretibial edema, the presence sion analysis corrected for all of the var- 1.84 (95% CI 1.14–2.98) and 1.89 (1.17– of pus, lymphadenitis/lymphangitis, iables as listed in RESEARCH DESIGN AND 3.05), respectively; a deep ulcer, HR 4.12 , fever (all P 0.01), and elevated METHODS as well as for center (unadjusted (2.01–8.47); positive probe-to-bone CRP levels (P =0.01).Thesame analyses in Supplementary Table 1). The test, HR 3.98 (2.05–7.71); and CRP levels more than three times the upper limit, HR 2.71 (1.31–5.61). In summary, — Table 1 Association between patient, leg, ulcer, and infection characteristics and the same variables predicted any ampu- lower-extremity amputation tation and amputations excluding the Any amputation Amputation excluding lesser toes lesser toes, with the exception of foul No Yes P* No Yes P* smell, nonpurulent and purulent exu- Total patients 74.1 25.9 82.1 17.9 date, and fever, which predicted any Patient characteristics amputation only. Female sex 40.6 29.5 0.02 40.5 25.2 ,0.01 In comparison with mild infection, the Age, years (mean) 65.4 66.2 0.49 65.4 66.4 0.50 presence of a moderate infection in- Leg characteristics ,0.01 ,0.01 creased the hazard for any amputation PAD 37.4 55.8 38.4 59.4 by a factor of 2.15 (95% CI 1.25–3.71) PAD with ABI ,0.5 6.8 13.6 7.0 15.8 and 3.01 (1.51–6.01) for amputations Ulcer characteristics excluding the lesser toes. For severe in- Ulcer size (cm2) ,0.01 ,0.01 1–5 51.8 59.7 52.7 59.2 fection, the hazard for any amputation .5 13.7 22.2 13.6 26.2 increased by a factor of 4.12 (1.99–8.51) Deep ulcer ,0.01 ,0.01 and for amputations excluding the Without probing to bone 22.7 26.5 22.2 30.4 lesser toes by a factor of 5.40 (2.20– With probing to bone 24.1 60.5 29.1 53.9 13.26). Larger ulcer size and presence Infection characteristics of PAD were also independent predic- Periwound redness (cm) ,0.01 0.01 tors of both any amputation and ampu- – 0.5 2 49.1 35.8 47.9 35.3 tations excluding the lesser toes, with .2 39.4 57.4 41.1 57.8 Periwound edema 52.7 72.5 ,0.01 54.8 71.8 ,0.01 HRs between 1.81 and 3 (and 95% CIs Pain or tenderness to palpation 50.7 54.4 0.44 49.8 60.2 0.06 between 1.05 and 6.6). Local skin temperature increased 60.0 65.8 0.21 61.2 63.1 0.71 Pus 29.7 41.6 ,0.01 32.3 35.0 0.60 Eurodiale Risk Scores Versus IWGDF , Lymphangitis/lymphadenitis 8.2 18.8 0.01 10.2 14.6 0.20 System for Predicting Amputation , , Body temperature 0.01 0.01 The risk score for any amputation that Fever (.38.08C) 5.2 13.4 5.7 14.6 included all the studied variables as re- Not recorded 12.3 14.1 12.1 15.5 IDSA severity ,0.01 ,0.01 ported in RESEARCH DESIGN AND METHODS had 2(mild)† 41.1 16.1 39.2 13.6 an area under the ROC curve of 0.84; the 3 (moderate)† 54.7 70.5 55.9 71.8 risk score for amputations excluding 4 (severe)† 4.2 13.4 4.9 14.6 lesser toes had an area under the ROC Data are percent unless otherwise indicated. *Pearson x2 test (categorical variables) or Student t curve of 0.85. To increase the applicabil- test (continuous variables). †Mild, infection of the skin/subcutaneous tissue; moderate, mild ity of the scores, we reduced them as with erythema .2 cm or infection of structures deeper than the subcutaneous tissue; severe, described above to the ones presented infection with systemic signs of inflammation. in Table 3, which include the variables care.diabetesjournals.org Pickwell and Associates 855

Table 2—HRs for amputation for individual infection characteristics Any amputation Amputation excluding lesser toes HR (95% CI) P value P value HR (95% CI) P value P value Periwound redness (cm) 0.11‡ 0.34‡ 0.5–2(vs.,0.5) 1.13 (0.54–2.36) 0.75† 1.05 (0.45–2.44) 0.91† .2(vs.,0.5) 1.68 (0.82–3.45) 0.15† 1.49 (0.65–3.38) 0.34† Periwound edema 2.01 (1.33–3.03) ,0.01 1.84 (1.14–2.98) 0.01 Pretibial edema 1.53 (1.02–2.31) 0.04 1.89 (1.17–3.05) ,0.01 Pain or tenderness to palpation 1.04 (0.70–1.54) 0.84 1.36 (0.84–2.18) 0.21 Local skin temperature increased 1.15 (0.79–1.68) 0.47 1.00 (0.64–1.58) 1.0 Increased exudate 1.49 (1.02–2.18) 0.04 0.97 (0.62–1.53) 0.91 Pus 1.67 (1.17–2.37) ,0.01 1.10 (0.70–1.72) 0.67 Lymphangitis/lymphadenitis 1.41 (0.84–2.37) 0.19 0.85 (0.43–1.68) 0.64 Body temperature 0.05‡ 0.16‡ Fever (.38.08C) 2.00 (1.15–3.48) 0.01† 1.80 (0.94–3.47) 0.08† Not recorded 1.21 (0.58–2.51) 0.61† 0.82 (0.37–1.83) 0.62† IDSA severity ,0.01‡ ,0.01‡ 3 (vs. severity 2)§ 2.15 (1.25–3.71) ,0.01† 3.01 (1.51–6.01) ,0.01† 4 (vs. severity 2)§ 4.12 (1.99–8.51) ,0.01† 5.40 (2.20–13.26) ,0.01†

Analysis adjusted for the following variables: sex; age; immobility; serum creatinine; HbA1c; presence of PAD and polyneuropathy; ulcer size, depth, and location; periwound redness; (periwound) edema; pain; foul smell; presence of exudate/pus; increased local skin temperature; lymphadenitis/ lymphangitis; fever; CRP levels; and center. †P value of test of subcategory of variable vs. baseline. ‡P value of test of all categories of variable combined. §Severity 2 (mild), infection of the skin/subcutaneous tissue; severity 3 (moderate), mild with erythema .2 cm or infection of structures deeper than the subcutaneous tissue; severity 4 (severe), infection with systemic signs of inflammation. sex, PAD, pain or tenderness on pal- was 0.67 in our population. Six percent CONCLUSIONS pation, ulcer size and ulcer depth, and of patients in the lowest tertile under- Independent Risk Factors for periwound edema yielding area under went any amputation; this increased to Amputation the ROC curves of 0.80 and 0.78 for 50% of patients in the highest tertile. Previously published studies that have any amputation and amputations ex- The respective percentages for amputa- aimed to identify independent risk fac- cluding the lesser toes, respectively. tions excluding the lesser toes were 1 tors for lower-extremity amputation in By way of comparison, the area under for the lowest tertile and 39 for the patients with a DFI have noted an asso- theROCcurvefortheIWGDFsystem highest tertile (Fig. 1). ciation with older age (5,22), the pres- ence of fever (5), elevated acute-phase reactants (5,22,23), higher HbA1c levels Table 3—Risk scores for amputation in patients with an infected diabetic foot (24), and renal insufficiency (5,22). In ulcer the current study, we identified the fol- Any amputation Amputation excluding lowing independent risk factors for (points) lesser toes (points) lower-extremity amputation in patients Sex with an infected diabetic foot ulcer (in Female 0 0 decreasing order of HR): positive probe- Male 0.5 1 to-bone test, deep ulcer, elevated CRP PAD levels, and the presence of periwound No 0 0 or pretibial edema. The presence of in- PAD 1 1 PAD with ABI ,0.5 1.5 2 creased (non)purulent exudate, foul Pain or tenderness on palpation smell, and fever independently pre- No 0 dicted any amputation but not amputa- Yes 0.5 tions excluding the lesser toes. Thus, we Ulcer size (cm2) confirmed some of the previously noted ,10risk factors but could not confirm 1–50.5others. This may be explained in part .51by several differences between our Ulcer depth study and those previously published. fi Super cial 0 0 Lipsky et al. (5) retrospectively studied Deep without probing to bone 1.5 1 Deep with probing to bone 2 2 amputation risk in 3,018 patients hospi- talized for a skin or soft tissue infection, Periwound edema No 0 but only 16% of these patients had an Yes 0.5 infected diabetic foot ulcer. Among three other studies that focused on 856 Amputation in Infected Diabetic Foot Ulcers Diabetes Care Volume 38, May 2015

Figure 1—ROC curves of risk scores for any amputation (A) and for amputations excluding the lesser toes (B) in patients with an infected diabetic foot ulcer.

amputation risk in patients with an in- risk scores we developed for any ampu- unclear why periwound edema is a risk fected diabetic foot ulcer, one ana- tation, and amputations excluding the factor for any amputation but not for am- lyzed only patients with large (.3 3 lesser toes had higher prognostic capa- putations excluding the lesser toes and 3 cm) and deep ulcers (23), while we bility, based on the area under the ROC why pain on palpation is a risk factor for included all ulcers, and the numbers of curve (0.80 and 0.78, respectively), than amputations excluding the lesser toes but patients included in the other two the IWGDF system (0.67). However, it not for any amputation. studies (22,24) were relatively small should be noted that these Eurodiale Some experts prefer an early toe am- (n = 100 and 165) compared with our scores were developed based on the putation to avoid long-term antibiotic cohort (n = 575). There were few patients available data of our cohort, and they treatment, and we therefore expected a with significant renal insufficiency in our will need to be validated in other pop- large variation between centers in the se- cohort, which may explain why we did ulations before any firm conclusions can lection of patients for toe amputations. not find an association between renal in- be drawn. The advantage of these newly However, when we added center to the sufficiency and amputation. Periwound developed scores is that they are easier risk score for any amputation, the area and pretibial edema as well as the pres- for clinicians to perform, as they do not under the ROC curve only marginally im- ence of increased exudate have not pre- require additional laboratory tests, as proved, from 0.80 to 0.84, contradicting viously been found to be associated with does the IWGDF system. There are rela- our assumption. amputation. Periwound edema may be tively few signs and symptoms of infec- related to more severe infection, and Limitations of the Current Study tion in the Eurodiale compared with the Our study has several limitations. In 37% pretibial edema may be associated with IWGDF risk scores, and several items of worse overall health status, leading to a of cases, patients had been treated by a the new scores are known risk factors medical specialist for their ulcer before greater risk of amputation. The presence – for poor outcome (25 27). Theoretically, they were enrolled. As these patients of increased exudate and foul smell pre- these new scores might predict the am- dicted any amputation, but not amputa- likely suffered from severe or complex putation risk of any ulcer, infected or not. tions excluding the lesser toes. It is diabetic foot disease, our results may In line with Louis Pasteur (“the terrain is possible that clinicians would consider not be applicable to patients with milder everything, the bacteria is nothing”), the an amputation of one of the digits 2–5 disease. While all patients included in characteristics of the individual in whom in the presence of these signs but would the Eurodiale cohort were treated ac- an infection develops might be more im- not consider it such a poor prognostic cording to IWGDF protocols, it is possi- portant than the infection itself. sign as to consider a larger amputation. ble that treatment differed in the Sex, PAD, and ulcer depth are in- various centers. Thus, we corrected all Eurodiale Risk Score Versus IWGDF cluded in the risk scores for both any the analyses for center. As mentioned System for Predicting Amputation amputation and amputations excluding previously, the risk score we created Previous studies have shown that with the lesser toes. Ulcer size was not a risk needs to be validated in other cohorts increasing grades of IWGDF infection se- factor for any amputation in our cohort, of patients with DFI. Our data were ob- verity, the risk of amputation increases probably because ulcers on digits 2–5 tained in 2003–2004, and the incidence (13–16,20,21). Also, in our cohort the are per definition small ulcers, diluting of lower-extremity amputation in patients number of amputations increased with the effect of ulcer size on amputation risk with diabetic foot ulcers has decreased increasing IWGDF risk score. The new when all amputations were analyzed. It is since then in several countries (28). care.diabetesjournals.org Pickwell and Associates 857

However, studies have observed that the disease in Europe. Baseline results from the diabetic foot ulcer classification system for re- relative risk of amputation for these pa- Eurodiale study. Diabetologia 2007;50:18–25 search purposes to predict lower extremity am- – tients has not changed (29,30), which is 2. Peters EJG, Lipsky BA. Diagnosis and man- putation. Int J Diabetes Dev Ctries 2009;29:1 5 agement of infection in the diabetic foot. Med 17. Prompers L, Huijberts M, Apelqvist J, et al. why we believe our results are still valid Clin North Am 2013;97:911–946 Optimal organization of health care in diabetic and applicable. 3. Lavery LA, Armstrong DG, Wunderlich RP, foot disease: introduction to the Eurodiale Mohler MJ, Wendel CS, Lipsky BA. Risk factors study. Int J Low Extrem Wounds 2007;6:11–17 Summary for foot infections in individuals with diabetes. 18. van Battum P, Schaper N, Prompers L, et al. Of the several clinical signs that we stud- Diabetes Care 2006;29:1288–1293 Differences in minor amputation rate in diabetic ied, deep ulcers, the presence of in- 4. Armstrong DG, Lavery LA, Harkless LB. Vali- foot disease throughout Europe are in part ex- fi creased exudate, foul smell, periwound dation of a diabetic wound classi cation sys- plained by differences in disease severity at pre- tem. The contribution of depth, infection, and sentation. Diabet Med 2011;28:199–205 and pretibial edema, and fever and ischemia to risk of amputation. Diabetes Care 19. Schaper NC, Apelqvist J, Bakker K. The in- elevated CRP levels were identified as 1998;21:855–859 ternational consensus and practical guidelines independent risk factors for amputation 5. Lipsky BA, Weigelt JA, Sun X, Johannes RS, on the management and prevention of the di- in patients with infected diabetic foot Derby KG, Tabak YP. Developing and validating a abetic foot. Curr Diab Rep 2003;3:475–479 fi ulcers. We confirmed the prognostic risk score for lower-extremity amputation in pa- 20. Schaper NC. Diabetic foot ulcer classi ca- tients hospitalized for a diabetic foot infection. tion system for research purposes: a progress value of the IWGDF system for predict- Diabetes Care 2011;34:1695–1700 report on criteria for including patients in re- ing amputation, and we developed two 6. Prompers L, Schaper N, Apelqvist J, et al. Pre- search studies. Diabetes Metab Res Rev 2004; new risk scores, one for any amputation diction of outcome in individuals with diabetic 20(Suppl. 1):S90–S95 and one for amputations excluding the foot ulcers: focus on the differences between 21. Lipsky BA, Berendt AR, Deery HG, et al.; In- lesser toes, with better prognostic accu- individuals with and without peripheral arterial fectious Diseases Society of America. Diagnosis disease. The EURODIALE Study. Diabetologia and treatment of diabetic foot infections. Clin racy than this system, which is currently 2008;51:747–755 Infect Dis 2004;39:885–910 the only one in use for infected diabetic 7. Tentolouris N, Al-Sabbagh S, Walker MG, 22. Aziz Z, Lin WK, Nather A, Huak CY. Predictive foot ulcers. These newly developed risk Boulton AJM, Jude EB. Mortality in diabetic factors for lower extremity amputations in di- scores can be readily used in daily clini- and nondiabetic patients after amputations abetic foot infections. Diabet Foot Ankle 2011; performed from 1990 to 1995: a 5-year fol- 2: DOI: 10.3402/dfa.v2i0.7463 cal practice without the necessity of ob- low-up study. Diabetes Care 2004;27:1598– 23. Akinci B, Yener S, Yesil S, Yapar N, taining additional laboratory testing. 1604 Kucukyavas Y, Bayraktar F. Acute phase reactants Future studies should be aimed at vali- 8. Meatherall BL, Garrett MR, Kaufert J, et al. predict the risk of amputation in diabetic foot in- dating our findings in other cohorts of Disability and quality of life in Canadian aboriginal fection. J Am Podiatr Med Assoc 2011;101:1–6 patients with an infected diabetic foot and non-aboriginal diabetic lower-extremity 24. Chiu CC, Huang CL, Weng SF, Sun LM, Chang amputees. Arch Phys Med Rehabil 2005;86: YL, Tsai FC. A multidisciplinary diabetic foot ulcer ulcer. 1594–1602 treatment programme significantly improved the 9. Willrich A, Pinzur M, McNeil M, Juknelis D, outcome in patients with infected diabetic foot Lavery L. Health related quality of life, cognitive ulcers. J Plast Reconstr Aesthet Surg 2011;64: Acknowledgments. The authors thank Prof. function, and depression in diabetic patients 867–872 B. Lipsky, Emeritus Professor of Medicine, with foot ulcer or amputation. A preliminary 25.PschererS,DippelF-W,LauterbachS, University of Washington; Visiting Professor, study. Foot Ankle Int 2005;26:128–134 Kostev K. Amputation rate and risk factors in Infectious Diseases, University of Geneva; and 10. Eneroth M, Apelqvist J, Stenstrom¨ A. Clini- type 2 patients with diabetic foot syndrome un- Teaching Associate, Green Templeton College, cal characteristics and outcome in 223 diabetic der real-life conditions in Germany. Prim Care University of Oxford, for critically reviewing the patients with deep foot infections. Foot Ankle Diabetes 2012;6:241–246 manuscript. Int 1997;18:716–722 26. Winkley K, Stahl D, Chalder T, Edmonds ME, Funding. The Eurodiale study was supported 11. Armstrong DG, Perales TA, Murff RT, Ismail K. Risk factors associated with adverse out- by the Fifth Framework Programme of the Edelson GW, Welchon JG. Value of white blood comes in a population-based prospective cohort European Commission. cell count with differential in the acute diabetic study of people with their first diabetic foot ulcer. fl Duality of Interest. No potential con icts of foot infection. J Am Podiatr Med Assoc 1996;86: J Diabetes Complications 2007;21:341–349 interest relevant to this article were reported. 224–227 27. Yesil S, Akinci B, Yener S, et al. Predictors of Author Contributions. K.P. and V.S. re- 12. Lipsky BA, Peters EJG, Senneville E, et al. amputation in diabetics with foot ulcer: single searched data and wrote the manuscript. M.K. Expert opinion on the management of infec- center experience in a large Turkish cohort. Hor- reviewed and edited the manuscript and contrib- tions in the diabetic foot. Diabetes Metab Res mones (Athens) 2009;8:286–295 uted to the discussion. J.A., K.B., M.E., P.H., A.J., Rev 2012;28(Suppl. 1):163–178 28. Schaper NC, Apelqvist J, Bakker K. Reducing E.J., D.M., A.P., G.R.T., H.R., M.S., L.U., V.U., K.v.A., 13. Lavery LA, Armstrong DG, Murdoch DP, lower leg amputations in diabetes: a challenge and J.v.B. planned the study, collected data, and Peters EJG, Lipsky BA. Validation of the Infec- for patients, healthcare providers and the reviewed and edited the manuscript. N.S. tious Diseases Society of America’s diabetic foot healthcare system. Diabetologia 2012;55: planned the study, collected data, and wrote infection classification system. Clin Infect Dis 1869–1872 the manuscript. K.P., V.S., and N.S. are the 2007;44:562–565 29. Kennon B, Leese GP, Cochrane L, et al. Re- guarantors of this work and, as such, had full 14. Wukich DK, Hobizal KB, Brooks MM. Severity duced incidence of lower-extremity amputations access to all the data in the study and take of diabetic foot infection and rate of limb salvage. in people with diabetes in Scotland: a nationwide responsibility for the integrity of the data and Foot Ankle Int 2013;34:351–358 study. Diabetes Care 2012;35:2588–2590 the accuracy of the data analysis. 15. Wukich DK, Hobizal KB, Raspovic KM, 30. Vamos EP, Bottle A, Edmonds ME, Valabhji Rosario BL. SIRS is valid in discriminating be- J, Majeed A, Millett C. Changes in the incidence References tween severe and moderate diabetic foot infec- of lower extremity amputations in individuals 1. Prompers L, Huijberts M, Apelqvist J, et al. tions. Diabetes Care 2013;36:3706–3711 with and without diabetes in England between High prevalence of ischaemia, infection and se- 16. Widatalla AH, Mahadi SE, Shawer MA, 2004 and 2008. Diabetes Care 2010;33:2592– rious comorbidity in patients with diabetic foot Elsayem HA, Ahmed ME. Implementation of 2597