Poor Glucose Control in the Year Before Admission as a Powerful Predictor of Amputation in Hospitalized Patients With Diabetic Foot Ulceration

Although there is a strong association between lower-extremity amputation (LEA) and HbA_1c (A1C) in diabetic patients (1,2), little information is available on glucose control in the period preceding and following LEA (3). Our objective was to evaluate the predictors of LEA and to examine the role of blood glucose control during the year before and the year after admission. A total of 122 diabetic patients (82 men and 40 women aged 69.7 ± 10.9 years, disease duration 19.7 ± 10.4 years) consecutively admitted for diabetic foot ulcers to the Medicine Department of United Hospitals of Bergamo between January 2003 and December 2004 were enrolled in our observational study and assigned to one of three groups.

The prevalence of long-term diabetes complications was similar in each group; >75% had peripheral polyneuropathy, nephropathy, retinopathy, and ischemic cardiomyopathy, and >90% had peripheral vascular disease. Glycemic control was evaluated before and after admission using the mean of three A1C levels measured in the year preceding admission and in the year following discharge. Patients were reevaluated 1 year after admission and assigned to one of three groups: group A, major amputation (n = 28); group B, minor amputation (n = 44); and group C, no amputation (n = 50). Major amputation was defined as amputation above the ankle, and minor amputation was defined as amputation below the ankle.

Data are shown as means ± SD. Categorical variables were compared by χ² test, continuous variables by ANOVA.

In the year before admission, mean A1C levels were significantly higher in groups A (9.7 ± 1.5%) and B (9.1 ± 1.9%) than in the group C (7.9 ± 1.3%, P < 0.001 vs. A and P < 0.05 vs. B). Similarly at admission, A1C levels were higher in groups A (10.1%) and B (9.4 ± 1.8%) than in the group C (8.1 ± 1.5%, P < 0.001). A1C levels decreased in all groups after 1 year of follow-up: 8.4 ± 1% in group A, 7.9 ± 1.3% in group B, and 7.5 ± 1.2% in group C (P < 0.05 vs. A).

At admission, patients in groups A and B had a higher ulceration grade according to Wagner’s classification (3.9 ± 0.4 and 3.7 ± 0.4, respectively) than those in group C (2.7 ± 1.1, P < 0.001). The prevalence of patients with LDL cholesterol levels <100 mg/dl was significantly lower in groups A (38%) and B (40%) than in group C (65%; P < 0.01 vs. A and P < 0.05 vs. B). More patients in group C (75%, P < 0.05) were on aspirin therapy than in groups A (61%) and B (59%). The frequency of optimal blood pressure control (≤130/80 mmHg) was ∼50% in all groups. Patients in group A (21%, P < 0.001 vs. B and C) received less follow-up with a specialist in a diabetes clinic than patients in the groups B (50%) and C (56%).

Although hypertension and cholesterol levels were inadequately treated, and aspirin therapy underused in a large number of these patients, multivariate logistic regression analysis revealed that only A1C level and ulceration grade were independently associated with LEA.

In conclusion, poor metabolic control and inadequate treatment of modifiable risk factors are powerful predictors of LEA in patients hospitalized for diabetic foot ulcers. We suggest that all diabetic patients with poor metabolic control and risk factors associated with cardiovascular disease receive early referral for education and aggressive treatment with a diabetes clinic.