Improving major amputation rates in the multicomplex diabetic foot patient: focus on the severity of peripheral arterial disease

Abstract

Objective:

Peripheral arterial disease (PAD), as well as diabetic neuropathy, is a risk factor for the development of diabetic foot ulcers. The aim of this study was to evaluate differences and predictors of outcome parameters in patients with diabetic foot by stratifying these subjects according to the severity of PAD.

Research design and methods:

In a prospective study, patients with new diabetic foot ulcers have been treated and investigated by structured healthcare. Subjects were recruited between 1 January 2000 and 31 December 2007. All study participants underwent a 2-year follow-up observation period. The patients underwent a standardized examination and classification of their foot ulcers according to a modification of the University of Texas Wound Classification System. The severity of PAD was estimated by measurement of the ankle brachial index (ABI) and the continuous wave Doppler flow curve into undisturbed perfusion (0.9 < ABI < 1.3), compensated perfusion (0.5 < ABI < 0.9), decompensated perfusion (ABI < 0.5) and medial arterial calcification.

Results:

A total of 678 patients with diabetic foot were consecutively included into the study (69% male, mean age 66.3 ± 11.0 years, mean diabetes duration 15.8 ± 10.2 years). Major amputations (above the ankle) were performed in 4.7% of the patients. 22.1% of these subjects had decompensated PAD. These subjects had delayed ulcer healing, higher risk for major amputation [odds ratio (OR) 7.7, 95% confidence interval (CI) 2.8–21.2, p < 0.001] and mortality (OR 4.9, 95 % CI 1.1–22.1, p < 0.05).

Conclusion:

This prospective study shows that the severity of PAD significantly influences the outcome of diabetic foot ulcers regarding to wound healing, major amputation and mortality.

Introduction

Foot ulcers are a major complication in patients with diabetes and remain one of the most common causes for hospitalization and the high costs associated with this disease. Data on the lifetime risk of developing a foot ulcer in diabetic patients are scarce but some studies suggesting that this risk may be as high as 25% [Singh et al. 2005].

Major amputation is one of the most feared complications. Structured healthcare is one of the promising ways to reduce major amputation in diabetic subjects [Weck et al. 2013; Prompers et al. 2008; Gershater et al. 2009]. Recently we showed that structured healthcare for patients with diabetic foot ulcers resulted in a significant reduction of major amputation rates by more than 75% compared with standard care [Weck et al. 2013].

However, prospective data on the predictors of outcome, especially with respect to the severity of peripheral arterial disease (PAD), in patients with diabetic foot are limited [Armstrong et al. 2011; Prompers et al. 2008]. Furthermore, there is no clear common definition of PAD which can be used to compare different groups of patients with diabetic foot. In particular subjects with medial arterial calcification (MAC) as an important subgroup of PAD are underrepresented or underreported in the literature [Faglia, 2011].

The aim of present study was therefore to obtain prospective data on outcome, especially major amputation, of individuals with a new diabetic foot ulcer in relation to the severity of PAD.

Study design and methods

Design

We organized in 2000 a contract regulating the organization of structured care for subjects with diabetic foot in the region near Dresden in southeast Germany. Briefly, patients with diabetic foot were referred to the interdisciplinary diabetic foot ward of the regional hospital (Weisseritztal-Kliniken Freital) by general practitioners, specialized diabetic outpatient departments or other specialists. At the interdisciplinary diabetic foot ward, diagnostic procedures were carried out and treatment started. Thereafter, patients were transferred to a rehabilitation clinic. After discharge from rehabilitation, a diabetic outpatient department carried out 6-monthly checkups, including all necessary further individual interventions.

In order to achieve a standardized clinical procedure, all participating medical institutions share a common set of diagnostic and therapeutic algorithms (Figure 1). A handbook explaining the standards was available for all of the professionals involved [Weck et al. 2006]. All instances were subject to supervision by senior specialists in diabetes (M.W., T.S., U.D.). The inclusion of subjects with new foot ulcers was scheduled over a time frame of 8 years with a follow-up investigation for each individual patient over 2 years.

Figure 1.

Clinical pathway of diagnosis and treatment of diabetic foot during the disease management programme.

DOAP, diabetic osteoarthropathy; PTA, percutaneous transluminal angioplasty; TCC, total contact cast.

Study population and procedure

A total of 1475 subjects were hospitalized because of diabetic foot ulceration between 1 January 2000 and 31 December 2007. Of this group, 736 patients covered by insurance from AOK (Germany’s largest health insurance company) and presenting with a recently manifested foot ulcer were enrolled consecutively into this observational and prospective study. Exclusion criteria were acute myocardial infarction or stroke within the last 6 months, terminal renal failure or any kind of cancer. Based on these criteria, 58 subjects out of the group insured by AOK were excluded from the study. All other patients were covered by other health insurance companies.

Management of diabetic foot ulcer

All patients were treated according to our internal protocol [Weck et al. 2006] based on the International Consensus of the Diabetic Foot [Schaper et al. 2003, 2012] and the guidelines of the Working Group on Diabetic Foot of German Diabetes Association [Morbach et al. 2008], which include offloading, diagnosis and treatment of infection, assessment of vascular status, treatment of PAD and regular wound debridement.

Patients with PAD were seen by the interventional angiologist and the vascular surgeon. If vascular reconstruction or interventional radiologic procedures were not possible, prostaglandins, low-dose urokinase or autologous bone marrow derived mononuclear cells (intramuscular application) were applied in order to improve perfusion [Weck et al. 2008. 2011; Amann et al. 2009]

Patients were transferred to a rehabilitation centre when acute treatment of the diabetic foot was completed. At the rehabilitation centre, the treatment procedures of foot ulcers were continued and combined with an intensified education programme. The patients were discharged if the diabetic foot was healed completely or an outpatient treatment was possible. Before discharge the subjects received definitive individual therapeutic footwear.

All subjects gave their written informed consent before participating in this study. The study was approved by the local ethics committee.

Potential predictive factors and comorbidities

According to EURODIALE (European Study Group on Diabetes and the Lower Extremity) and other studies we evaluated the following parameters: sex; age at baseline; duration of diabetes; HbA_1C; vibration perception and other parameters of peripheral polyneuropathy; parameters of cardiac autonomic neuropathy (CAN); background and proliferative retinopathy; renal impairment; prior amputation; coronary artery disease; hypertension; dyslipidaemia; and smoking (Appendix A).

Ulcer characteristics

All patients underwent a standardized examination and classification of their foot ulcers according to a modification of the University of Texas Wound Classification System (Wagner–Armstrong System). The extent of the foot wound is graded as described by Wagner [Wagner, 1981] and the staging of the wound situation is described by Armstrong and colleagues [Armstrong et al. 1998]. This descriptive system classifies foot ulcers according to the categories extent, depth, infection and perfusion. The outcome, especially with respect to major amputation, appears to deteriorate with increasing grade and stage of the foot wounds.

Perfusion

The assessment of perfusion included as first line evaluation the palpation of pedal pulses and the measurement of the ankle brachial index (ABI) using a handheld Doppler device. With this procedure we classified PAD as in Table 1.

Table 1.

Classification of PAD.

Measurement	Classification
ABI > 0.9	Undisturbed perfusion
ABI 0.5–0.9	Reduced but compensated perfusion
ABI < 0.5 or absolute closing pressure < 50–70 mmHg	Decompensated perfusion and CLI
noncompressible limb vessels, ABI > 1.3 and/or closing pressure > 200 mmHg	MAC, mediasclerosis

ABI, ankle-brachial index; CLI, critical limb ischemia; MAC, medial arterial calcification.

Because of false high ABI in subjects with MAC we used continuous wave Doppler sonography for more detailed analysis of perfusion. We considered compensated perfusion if the increasing wing of the Doppler curve was steeper than the decreasing wing. A gently inclining curve or flat curve was considered as decompensated perfusion or critical limb ischemia.

With these measures we subdivided the patients with diabetic foot into four categories:

1. Undisturbed perfusion

2. Disturbed but compensated perfusion

3. Decompensated perfusion and critical limb ischemia (CLI)

4. MAC

These categories were the basis for further analysis in addition to the modified University of Texas classification system.

Infection

Infection was diagnosed if two or more of the following signs were present: frank purulence; local warmth; erythema; lymphangitis; oedema; pain; fever; and foul smell.

Evaluation of sensation

The evaluation of sensation (peripheral neuropathy, PNP) included measurement of vibration sensation (128 Hz tuning fork on dorsum of the hallux), tactile sensation (cotton wisp on dorsum of the foot) and blunt/sharp discrimination (dorsum of the foot). PNP was diagnosed if two or more of these tests were abnormal.

Statistical analysis

The calculations were performed using SPSS version 17.0 (SPSS Inc., IL, USA). Summary tabulations were presented by the number of observations, mean and standard deviation, for continuous variables, and by frequency and percentage per category for categorical data.

Comparison of means from two independent groups was performed using two-sample t-test. Differences were considered statistically significant if p < 0.05.

The influence of the risk factors age, duration of diabetes, systolic and diastolic blood pressure, body mass index (BMI) and HbA_1c on perfusion was examined by one-way analysis of variance and then tested with a post hoc Student–Newman–Keuls procedure. To assess the effect of perfusion on major amputation and wound healing, odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated.

Results

Patient structure

A total of 736 subjects with newly detected diabetic foot, covered by the health insurance company AOK, were consecutively included into the study. Because of acute myocardial infarction or stroke within the last 6 months, terminal renal failure or any kind of cancer, 58 subjects were excluded from the study. Therefore, 678 subjects with diabetic foot ulceration were suitable for data analysis. The baseline data are shown at Table 2.

Table 2.

Baseline data (mean ± SD).

	Male (n = 470)	Female (n =2 08)	Total (n = 678)	P
Age (years)	64.4 ± 10.5	70.6 ± 11.1	66.3 ± 11.0	<0.001
Diabetes duration (years)	15.2 ± 10.0	17.2 ± 10.7	15.8 ± 10.2	n.s.
HbA1C (%)	7.9 ± 3.5	7.8 ± 1.8	7.9 ± 3.0	n.s.
BMI (kg/m2)	29.4 ± 5.3	30.2 ± 6.9	29.6 ± 5.8	n.s.
Blood pressure (mmHg)
systolic	139 ± 20	140 ± 22	139 ± 21	n.s.
diastolic	77 ± 11	76 ± 11	76 ± 11	n.s.

BMI, body mass index; n.s., not significant.

Men made up 69% of the patients. Female subjects were significantly older than male ones. Detailed data on diabetic complications and comorbidities of the study group are given at the Appendix A. The prevalence of wound severity by ulcer grade and stage according to the modified University of Texas system is outlined in Table 3.

Table 3.

Prevalence of wound severity by ulcer grade and stage (modified University of Texas Wound Classification System) at admission to the clinic, n = 678 (% of 678).

Modified University of Texas grade (Wagner)		1	2	3	4
University of Texas stage
A	n	1	0	0	0
	%	(0.1)
B	n	2	22	16	2
	%	(0.3)	(3.2)	(2.4)	(0.3)
C	n	12	36	9	4
	%	(1.8)	(5.3)	(1.3)	(0.6)
D	n	59	218	257	60
	%	(8.7)	(32.1)	(37.9)	(8.8)

At admission to the structured care, about 77% of all lesions were found to penetrate to tendon, capsule, bone or joint, had a local necrosis, and were infected and ischemic (Wagner grade 2, 3 and 4, University of Texas stage D).

Clinical outcome

At discharge from the clinic about 30% of all foot wounds were healed. Another 52% of foot wounds were improved to modified University of Texas grade 1. At the 2-year follow-up examination 74% of ulcers were healed completely and another 17% were in modified University of Texas grade 1.

A total of 32 subjects with diabetic foot underwent major amputation (amputation above the ankle) during hospital treatment (major amputation rate 4.7%). During the 2-year follow up, 22 subjects (3.2%) underwent major amputation and 215 patients (31.7%) experienced minor amputations (distal of the ankle). This high–low amputation rate according to Wrobel and colleagues [Wrobel et al. 2006] was about 1:7, indicating a favourable result.

The mortality rate during hospitalization was 2.5% (n = 17). After 2 years, 20.9% (n = 143) had died.

PAD status

Only a minority of the study population had an undisturbed perfusion (n = 43, 6.3%), 41.3% (n = 279) had disturbed but compensated perfusion, 22.1% (n = 150) had decompensated perfusion and CLI and 30.4% (n = 206) had the signs of MAC (Figure 2).

Figure 2.

Analysis of perfusion disturbances in the whole study population (n = 678).

CPI, critical limb ischemia; MAC, medial arterial calcification.

In the groups of patients with compensated and decompensated perfusion we performed 45 infrapopliteal bypass reconstructions and 81 iliacal or femoral arterial reconstructions. A total of 445 patients underwent percutaneous transluminal angioplasty (PTA) and another 116 patients were treated by prostaglandins, low-dose urokinase and ‘stem cells’.

Ulcer characteristics according to PAD status

The ulcer characteristics described by the modified University of Texas Wound Classification System and according to the PAD status are shown in Table 4.

Table 4.

Ulcer characteristics described by the Wagner grade and according to the PAD status of the diabetic foot subjects (n = 678).

PAD status	Ulcer grading according to Wagner
n (%)*	1 (n = 74)	2 (n = 256)	3 (n = 282)	4 (n = 66)
Undisturbed perfusion	3 (4.1%)	22 (8.6%)	16 (5.7%)	2 (3.0%)
Disturbed but compensated perfusion	37 (50.0%)	105 (41.0%)	114 (40.4%)	23 (34.8%)
Decompensated perfusion and CLI	12 (16.2%)	49 (19.1%)	64 (22.7%)	25 (37.9%)
MAC	22 (29.7%)	80 (31.3%)	88 (31.2%)	16 (24.2%)

^*Percentage of distinct PAD status according to the grading of the foot ulcer (Wagner).

CLI, critical limb ischemia; MAC, medial arterial calcification; PAD, peripheral arterial disease.

Undisturbed perfusion was a very rare PAD status among our population with diabetic foot ulcers. With increasing severity of ulcer grade, however, we observed an increasing severity of PAD status with the highest percentage of decompensated perfusion and CLI at Wagner grade 4. MAC was uniformly distributed at the different grades of ulcer severity according to Wagner.

Diabetic foot ulcer healing according to PAD status

When stratifying patients according to the severity of PAD, worse healing rates were observed with increasing severity of PAD: 45% of diabetic foot ulcers of subjects with undisturbed perfusion were healed at discharge from hospital; 28% subjects with disturbed but compensated perfusion; 12% of patients with decompensated perfusion and CLI; and 21% of subjects with MAC. The ulcer healing rate increased during the 2-year follow up but during follow up we observed the same pattern as at discharge from hospital, that is, worse healing rates with increasing severity of PAD, except MAC (Figure 3).

Figure 3.

Percentage of ulcers completely healed according to PAD status at discharge from hospital and during 2-year follow up.

CPI, critical limb ischemia; MAC, medial arterial calcification; PAD, peripheral arterial disease.

When stratifying patients with major amputations during hospitalization (n = 32; major amputation rate 4.7%) according to the severity of PAD, we observed no amputations in the group with undisturbed perfusion, 5 major amputations (15.6%) in subjects with disturbed but compensated perfusion, 19 major amputations (59.4%) in the group with decompensated perfusion and 8 major amputations (25%) in subjects with MAC.

In Table 5 potential risk factors of nonhealing and comorbidities of patients without PAD are compared with those with PAD. Subjects with diabetic foot ulcer and PAD were significantly older, had more severe and infected ulcers, and had a significantly higher proportion of diabetic nephropathy, hypertension and history of smoking.

Table 5.

Potential predictive factors and comorbidities according to the PAD status of the patients.

	Patients without PAD (n = 43)	Patients with PAD (n = 635)	p value
Age (years)	60.14 ± 9.43	66.69 ± 11.04	<0.001
Male	30 (69.8%)	440 (69.3%)	0.948
Deep and infected ulcer	0 (0.0%)	317 (49.9%)	<0.001
Polyneuropathy (< 4/8)	14 (32.6%)	392 (61.7%)	<0.001
CAN	14 (32.6%)	260 (40.9%)	0.278
Preproliferative retinopathy	28 (61.1%)	490 (65.1)	0.072
Proliferative retinopathy	3 (7.1%)	42 (6.6%)	0.926
Microalbuminuria (30–300 mg per 24 h)	12 (27.9%)	177 (27.9%)	0.996
Proteinuria (>1 g per 24 h)	0 (0.0%)	7 (1.1%)	0.489
Diabetic nephropathy	4 (9.3%)	178 (28.0%)	<0.05
History of amputation	9 (20.9%)	207 (32.6%)	0.112
BMI (kg/m2)	28.8 ± 5.8	29.7 ± 5.8	0.358
Arterial hypertension	36 (83.7%)	591 (93.1%)	<0.05
Hyperlipidaemia	25 (58.1%)	411 (64.7%)	0.383
Smoking	22 (51.2%)	203 (32.0%)	<0.05

BMI, body mass index; CAN, cardiac autonomic neuropathy; PAD, peripheral arterial disease.

The risk factors for major amputations are shown in Table 6. The only significant risk factor in these study patients was decompensated perfusion/CLI.

Table 6.

Risk factors for major amputation.

	OR	95% CI	p
Age	1.04	1.00–1.08	0.09
HbA1C	0.95	0.76–1.18	0.63
Decompensated perfusion	7.70	2.78–21.32	<0.001
MAC	2.61	0.84–8.14	0.10

CI, confidence interval; MAC, medial arterial calcification; OR, odds ratio.

When stratifying mortality according to the PAD status we found the following data. Mortality in the group with undisturbed perfusion was 0% during hospitalization and 5% during the 2-year follow up. In the group with compensated perfusion it was 2% during hospitalization and 24% during the follow-up period. With decompensated perfusion and CLI it was 6% during hospitalization and 28.4% during the follow-up period and, in the group with MAC, it was 3% during hospitalization and 15.6% during the follow-up period.

The significant risk factors for 2-year mortality were age, compensated perfusion and decompensated perfusion/CLI.

Discussion

Prospective long-term studies regarding the outcome of a structured care of the diabetic foot are scarce, especially with a sufficient number of included patients. We therefore conducted this prospective observational study with an 8-year inclusion and a 2-year post-treatment observation period. The central objective of the main project, a dramatic reduction of major amputation in diabetic foot patients, was achieved by introducing structured care for the diabetic foot [Weck et al. 2013].

The number of amputations above the ankle was 4.7% at the end of hospital treatment and therefore very low compared with another regional hospital [Weck et al. 2013] and literature data [Prompers et al. 2008; Richard et al. 2011; Faglia et al. 2001; 2006; 2009; Gershater et al. 2009; Uccioli et al. 2010]. An important limitation of our study regarding generalizability is that patients with acute ischemic events (myocardial infarction and stroke) and those with end-stage renal diseases were excluded. These are, of course, the subjects with worst prognosis regarding mortality and limb saving [Morbach et al. 2012].

Patients with diabetes and foot ulcer are suffering from a multiorgan disease with a complexity of factors related to outcome. PAD, infection of the foot and comorbidities are the most important factors predicting ulcer healing, major amputation and mortality. Therefore, we analysed these outcome predictors in a group of patients with diabetic foot ulcers successfully treated in a system of structured care.

PAD is one of the most important risk factors causing diabetic foot ulcers and in this way an important predictor of treatment outcome [Prompers et al. 2008; Pecoraro et al. 1990; McNeely et al. 1995; Adler et al. 1999; Boyko et al. 1999; Faglia, 2011; Apelqvist et al. 2011; Aerden et al. 2011; Game, 2012; Hanssen et al. 2012; Ince et al. 2007]. In our study cohort these data are reflected strikingly. Only 6.3% of subjects had normal ABI, 63% had PAD with an ABI <0.9 and 30% had signs of MAC, another typical feature of macrovascular disease. In the EURODIALE study with 1229 subjects with diabetic foot, Prompers and colleagues [Prompers et al. 2007, 2008] diagnosed PAD by the means of ABI <0.9 in 49% of this study cohort.

According to age, diabetes duration, glucose control (HbA_1C) and sex distribution, the EURODIALE cohort compares well with our study population. The extremely low number of patients without restricted perfusion in our study population might be partly explained by the study design in that the subjects of this study are a hospital-based cohort.

In the French OPIDIA study [Richard et al. 2011], in 291 patients with diabetic foot, 50–62% had PAD. The highest rates of PAD in diabetes foot subjects were reported from the Milan centre for diabetic foot with more than 80% having PAD [Faglia et al. 2001].

One of the largest and most comprehensive studies on outcome factors of diabetic foot comes from the Swedish centres in Lund and Malmö [Gershater et al. 2009]. A total of 2480 patients with diabetes and foot ulcer (Wagner grade 1–5) were prospectively followed and treated until healing was achieved or until death. According to the authors, 49% of the surviving patients had PAD and according to our calculation of the study data, about 11% had CLI. The primary healing rate of the whole cohort without amputation was 65%. The rate of proliferative retinopathy was extraordinary high at about 40%.

In the diabetic foot ulcers with PAD amputation was related to comorbidities, PAD and type of the ulcer. Age, sex, duration of diabetes, neuropathy, deformity and duration of ulcer or site of the ulcer did not evidently have an influence on the probability of amputation.

PAD as a risk factor for diabetic foot ulcers is well-known and extensively discussed elsewhere [Holstein, 1984; Larsson et al. 1993; Boulton, 1996; Adler et al. 1999; Boyko et al. 1999; Faglia, 2011] but data on outcome such as ulcer healing, amputation and death are relatively scarce.

In the EURODIALE study [Prompers et al. 2008], ulcer healing rates were significantly worse in subjects with PAD (69%) compared with patients without PAD (84%). Moreover, the combination of PAD and infection had a major impact on healing rates.

In our cohort of subjects with diabetic foot we subdivided PAD according to the severity of the disease by ABI and the continuous Doppler curve form. At discharge from acute clinical treatment, 45% of ulcers with undisturbed perfusion were completely healed compared with 28% of subjects with disturbed but compensated perfusion, 12% in patients with decompensated perfusion and CLI, and 21% with MAC.

If we compare our 1-year follow-up data with the 1-year follow-up results of the EURODIALE population we had nearly the same results: completely healed ulcers in 78.8% of patients without PAD against 62.5% in subjects with different severity of PAD (OR 2.2; 95% CI 0.94–5.31; p = 0.06).

One drawback of our study is that we did not use the PEDIS system for characterizing diabetic foot ulcers as in EURODIALE. We have used the modified University of Texas system for many years. With this wound classification system and the diagnostic tools for infection assessment (as described above), we observed extremely high rates of infected ulcers (University of Texas stage B and D); 91% of the ulcers were infected and 85% had a combination of PAD and infection, indicating a very bad initial wound and disease situation.

Data on the influence of the severity of PAD on major amputation rates in subjects with diabetic foot are extremely scarce. The most comprehensive and detailed data come from the Milan group [Faglia et al. 1998, 2001, 2006, 2009; Faglia, 2011]. The clinical pictures of PAD in patients with diabetes is characterized by the following facts [Faglia, 2011]

• - The occlusive disease is common.

• - The localization is mainly distal.

• - Arterial wall calcification occurs frequently

• - Occlusion occurs more frequently than stenosis.

To our experience and from our data, MAC is a typical feature of macrovascular disease. Its presence predisposes to a more severe PAD in the tibial and peroneal arteries, and is a strong predictor of cardiovascular morbidity and mortality in diabetics [Edmonds, 2000; Smith et al. 2008]. Moreover, MAC is present in approximately 40% of diabetic patients with CLI [Faglia et al. 2006] and an ABI >1.3 is indicative of excess mortality [Allison et al. 2008]. With respect to its pathogenesis there is a strong association between MAC and CAN in diabetic patients [Forst et al. 1995].

In our study population we found a high prevalence of subjects with ‘pure’ MAC without significant peripheral occlusive disease but we cannot exclude the possibility that there is a high proportion of MAC in the groups of patients with decompensated perfusion and CLI. We strongly believe that MAC is under-recognized in diagnostic and therapeutic strategies in patients with diabetic foot and PAD. However, accurate diagnosis of the different clinical pictures of PAD is essential for proper treatment and for prognosis with respect to limb salvage.

Moreover, the subjects of our study had an extremely high prevalence of CAN as measured by a standardized battery of tests of heart rate variability (Appendix A). According to literature data, CAN appears to be a predictor for poor clinical outcome, especially for mortality in diabetic subjects.

In our cohort of patients with diabetic foot we found a clear association between increasing severity of PAD and clinical outcome: worse ulcer healing rates; more major amputations; and higher mortality in these groups of patients. The association between MAC and clinical outcome is less clear.

No major amputation had to be carried out in diabetic patients with undisturbed perfusion. We believe that this can be an indicator for the effectiveness of our system of structured care of the diabetic foot. To our opinion and according to literature data, and by considering the severity of diabetic foot ulcer and the severity of PAD, the major amputation rates are generally low in this study and they are concentrated in those subjects with decompensated perfusion. This can and should be an indicator of the revascularization procedures we have described, including ‘stem cells’ and low-dose urokinase.

In the EURODIALE study [Prompers et al. 2008], a major amputation rate of 5.1% in 1229 patients was documented. Severe limb ischemia, as defined by an ABI of <0.5, was present in 12% of these patients. In our study decompensated perfusion was found in 22% of all patients.

Faglia and colleagues [Faglia et al. 1998] addressed the problem of occlusive PAD in subjects with diabetic foot ulcers. From 1993 to 1995, 121 patients with diabetic foot were admitted consecutively to Milan University Hospital. Angiography was carried out in 104 subjects. The most interesting data in this study were the extraordinary high rate of occlusive arterial disease. Only one out of 104 subjects did not have haemodynamically significant stenoses. Nearly half of the patients had stenoses in the popliteal and infrapopliteal axis only. Because neuropathy was also found to be very common (86.5%), the prevalent picture was neuroischemic diabetic foot. Similarly, more than 80% off all subjects in our study had neuroischemic diabetic foot. In 2009, the same group reported even more impressive data on 554 patients with CLI [Faglia et al. 2009]. In these patients, peripheral angioplasty (PTA) was performed in 75% and bypass graft (BPG) in 21%. Neither PTA nor BPG were possible in 5% of the subjects. The major amputation rate in this highly complicated group of patients was 13.4% (8% in PTA patients, 21% in BPG and 59% in the subgroup who received no revascularization). Comparable data were reported by Uccioli and colleagues [Uccioli et al. 2010]. It can be concluded from these studies, the EURODIALE [Prompers et al. 2008] and our data, that the degree of PAD has a significant effect on the outcome of diabetic foot.

Another study which compares well to our study design and cohort is the French OPIDIA study [Richard et al. 2011]. A total of 291 patients with diabetic foot were included, with grading and staging of the foot ulcer performed according to the University of Texas system. Age, sex distribution and diabetes duration also compared well. The percentage of PAD in the OPIDIA cohort was lower than in our group, but the description of PAD was less clear and conclusive. Revascularization procedures were performed much less than in our study cohort, but major amputation rates were comparable. Richard and colleagues [Richard et al. 2011] concluded that the major amputation risk is increased in University of Texas stages C and D, indicating PAD and infection. More than 80% of the major amputations in our study were done for diabetic foot ulcers in Wagner grade 3 or more and all major amputations were carried out for ulcers in University of Texas stage C and D.

PAD is an important prognostic marker for mortality in diabetic patients [Faglia, 2011]. ABI <0.9 approximately doubles the risk of mortality and the risk is four-fold higher in patients with ABI values <0.6 [Ankle Brachial Index Collaboration, 2008]. The highest mortality rates were reported in patients in whom revascularization was not possible [Faglia et al. 2009]. MAC is indicative of excess cardiovascular mortality [Allison et al. 2008]. Faglia et al. [2009] reported a mortality rate of about 50% in diabetic patients with CLI during a mean follow up of 5.93 years. Nevertheless, data on the association between the severity of PAD and mortality in patients with diabetic foot are rare.

In our study cohort, total mortality was 2.4% during hospital treatment, 11.1% died in the first year of follow up and 6.5% during the second year (20.0% 2-year mortality). We observed a clear correlation between the severity of PAD and mortality. The highest mortality was documented in the group with decompensated perfusion/CLI (28.4% during 2-year follow up).

We conclude that the severity of PAD has a significant influence on the outcome of the diabetic foot with respect to wound healing, major amputation and mortality in a system of structured care for the diabetic foot. Decompensated perfusion was the most important risk factor for outcome. The prognosis for patients with diabetic foot ulcers and PAD, especially CLI, is dependent mainly on excellent diagnostic and therapeutic accuracy of the underlying macrovascular disease. Centres for diabetic foot need to provide excellent revascularization strategies every time.

Appendix

Appendix A.

Diabetic complications of the subjects in the structured care programme for the diabetic foot (n = 678).

Parameter	n (%)
VPT < 4/8	654 (95.6)
CAN	289 (42.3)
Retinopathy
Background	561 (82.0)
Proliferative	55 (8.0)
Microalbuminuria	269 (39.2)
Creatinine >130 µmol/l	104 (15.2)
Prior amputation	249 (36.4)
Toe	152 (22.2)
Transmetatarsal	34 (5.0)
Below the knee	40 (5.8)
Above the knee	23 (3.4)
CAD	567 (82.9)
Prior AMI	47 (6.9)
Prior stroke	51 (7.5)
Hypertension	621 (90.8)
Dyslipidemia	540 (78.9)
Smoking	231 (33.8)

AMI, acute myocardial infarction; CAD, coronary artery disease; CAN, cardiac autonomic neuropathy; VPT, vibration perception threshold.