Lower Limb Amputation Rates in Patients With Diabetes and an Infected Foot Ulcer: A Prospective Observational Study

Chronic wounds have become a significant and rapidly increasing global issue due to increased cases of chronic diseases (eg, diabetes, cardiac and vascular diseases, and obesity) in addition to an aging population. A large percentage of chronic wounds are diabetic foot ulcers (DFU) and pressure injuries.¹ DFU and its complications can lead to lower extremity amputations and present significant patient, family, and socioeconomic concerns. In persons with diabetes, a cumulative lifetime incidence of up to 25% has been reported for the development of DFU, of which approximately 24% of ulcers progress to require lower limb amputation.^2-4 A 10-year cohort study by Johannesson et al⁵ stated that the incidences of contralateral amputation and re-amputation per 100 amputee-years in female amputees with diabetes were 15 and 16, respectively; in male amputees with diabetes, these figures were 18 and 21, respectively. Several nationwide studies have been done in India in the past, but all had some type of shortcoming (eg, not including the rural population). This includes a survey conducted in the 1970s by the Indian Council of Medical Research, which  sampled only 6 regions of the country.⁶

Known factors responsible for increased risk of infection and amputation in patients with diabetic foot ulceration are the presence of peripheral neuropathy, peripheral arterial disease (PAD), elevated levels of hyperglycemia-related advanced glycation end-products, dysfunctional neutrophils, continuous inflammation, and cell death.^7-13

The aim of the current study was to determine risk factors associated with lower extremity amputation in patients with infected DFU. The objectives were to assess the relationship between the rate of lower extremity amputation in persons with infected DFU and characteristics related to patients (demographic and clinical data), vascular and ulcer variable (Perfusion, Extent, Depth, Infection and Sensation [PEDIS] score), and biochemical, microbiological, and radiological findings.

Study setting. A prospective observational study was conducted at the Department of General Surgery, All India Institute of Medical Sciences, Rishikesh, India, over a period of 18 months (from December 2018 to May 2020).

Inclusion and exclusion criteria. The inclusion criteria were age from 18 to 65 years, a diagnosis of diabetes, and a foot ulcer showing clinical signs of infection, such as peri-wound redness, peri-wound edema, pain or tenderness on palpation, and raised local skin temperature. Patients who showed evidence of osteomyelitis of the foot were excluded. Based on a previous similar study,¹⁴ patients with pathologies resulting in restriction of mobility (being bedridden or wheelchair-bound) also were excluded to reduce potential confounding factors. All consecutive patients with diabetes who presented with foot ulcers to both outpatient and inpatient departments were evaluated for study eligibility. All patients who met the inclusion criteria and provided informed consent were included in the study.

Ethical considerations. The study was initiated following approval by the ethics committee of the institution. All variables were recorded manually on a predesigned form. Patient names were not included on the data sheet.

Methodology. Once recruited, the patient’s demographic and clinical data were recorded. A detailed history was obtained that included age, sex, duration of ulcer-related complaint, duration of diabetes, presence of comorbidities such as chronic renal disease, history of previous amputation, and treatment of diabetes. The ulcer details included site, number, size, depth, edge, and wound base. Every ulcer was scored based on the PEDIS classification provided by the International Working Group on the Diabetic Foot.^15,16 Each criterion in the PEDIS classification system is given a score of 0 to 3, following which the score is summed. Signs of infection noted were peri-wound redness, peri-wound edema, pain or tenderness on palpation, local skin temperature measured with a handheld thermometer (based on previous studies),¹⁷ presence of pus/exudate, and body temperature. These signs were recorded as being present or absent.

The study population underwent medical tests that included complete blood counts, random as well as fasting and postprandial blood glucose levels, liver function tests, kidney function tests, and viral markers. In addition, an ankle brachial pressure index (ABPI) was recorded and bilateral lower limb arterial color Doppler ultrasound, radiography of the foot, hemoglobin A1c (HbA1c), and wound swab culture and sensitivity tests were performed. These investigations were done only on the first visit and recorded manually into the form.

Participants visited the outpatient department every third week, during which time the characteristics of the ulcer were reassessed and documented. Participants remained part of the study for 6 months or until they reached one of 3 predetermined outcomes: healing of the foot ulcer(s), major or minor amputation, or death. Ulcer healing was defined as the presence of epithelialization with no clinical signs of infection. All amputations were classified as either major or minor. Amputations involving digits 2 to 5 were considered minor, keeping in mind that the great toe has a major role in direction control and balance. Amputation of the great toe hampers mobility to a larger degree in comparison to amputation of the other 4 digits. Previous similar studies have also considered amputation of the great toe to be a major amputation.^18,19

Data management and statistical analysis. The data for each patient were recorded on the clinical study form, and the data were then converted to a master chart in Microsoft Excel. All variables were considered for testing association with all outcomes. Ulcer depth was categorized based on the deepest layer involved (1, skin; 2, subcutaneous; 3, fascia; 4, muscle and tendon; and 5, bone). Arterial Doppler ultrasound findings were categorized according to the most proximal vessel involved (0, normal study; 1, dorsalis pedis artery; 2, up to posterior tibial artery; 3, up to anterior tibial artery; 4, up to popliteal artery; and 5, up to femoral artery).

The Statistical Package of Social Science (SPSS) version 23 (IBM) was used for statistical analysis. The Shapiro-Wilk test was used for testing the normality of the data collected. Significance was considered as P < .05. Tests used to analyze the association between variables and outcomes were the Wilcoxon test, Fisher’s exact test, Chi-square test, and t-test. Parametric data were evaluated by t-test, and nonparametric data were evaluated by Chi-square test and Wilcoxon test. Fisher’s exact test was used to assess association in contingency tables where more than 20% of the total number of cells had an expected count of less than 5.

Eighty-one (81) patients were eligible to participate; 61 (75.3%) were men and 20 (24.7%) were women. Age ranged from 24 to 65 years (mean, 54.58 ± 9.04) (Table 1). One (1) patient had type 1 diabetes, and 80 patients had type 2 diabetes. The mean duration of the ulcer was 23.44 ± 21.79 days. The comorbidities recorded (percentage of the study population) were hypertension (16%), renal disorders (eg, chronic renal failure) (11.1%), cardiac disorders (eg, coronary artery disease) (12.3%), and tuberculosis (3.7%). The mean duration of diabetes was 7.45 ± 5.78 years. Of the 81 patients, 64 were taking oral hypoglycemic agents and 27 were receiving insulin injections. The mean PEDIS score was 7.16 ± 1.50, mean hemoglobin level was 10.36 ± 2.52 g/dL, and mean total lymphocyte count (TLC) was 14756.19 per mcL ± 7271.67). The mean fasting blood glucose level was 209.05 ± 79.37 mg/dL, mean postprandial blood glucose was 270.73 ± 87.70 mg/dL, and mean hemoglobin A1C (HbA1c) (%) was 9.67 ± 2.31. Growth on the wound swab culture was found in 57 (70.4%) patients. Escherichia coli was the most common organism isolated in the growth culture (22 [38.6%]). Maximum sensitivity was found to carbapenems (43.2%). The next most common was beta-lactam, to which 38.3% of participants had culture sensitivity. All growth cultures (100%) were resistant to penicillins. A total of 29 participants (35.8%) had normal findings on arterial Doppler ultrasound on initial examination. The majority of participants (18.5%) with arterial involvement showed involvement up to the anterior tibial artery. The mean ABPI was 0.92 ± 0.34.

Seventeen (17) patients underwent major amputation, 24 patients underwent minor amputation, 2 patients died, 33 patients had ulcers that healed, and 5 patients did not reach one of the study outcomes after 6 months (Table 2).

Variables that were found to be significantly different between patients who did and did not have a major amputation were high PEDIS Score (6.92 ± 1.36; P = .005) (Table 3), elevated HbA1c (%) (9.43 ± 2.19; P = .049), presence of growth on culture (41 [64.1%]; P = .016), culture sensitivity to beta lactam (20 [31.2%]; P = .012), and presence of PAD on arterial Doppler  ultrasound (P < .001) (Table 4). Minor amputation rates were significantly higher among patients who were male compared with patients who were female; that is, 22 males (91.7%) compared to 2 females (8.3%) underwent minor amputation (P = .027) (Table 5). Another variable that was found to be significantly associated with minor amputation was the presence of PAD on ABPI: a total of 14 patients (70%) of those who underwent minor amputation had evidence of PAD (P = .005) (Table 6). The mean ABPI in the group that underwent minor amputation was 0.69 ± 0.33. There was a significant difference between the 2 groups in terms of ABPI (t = -3.953; P = < .001), with the mean ABPI being highest in the group that did not undergo minor amputation.

The presence of renal disorders, which included chronic kidney disease and acute kidney injury, was significantly different between patients who did and did not die during the study. A total of 9 patients (11.1%) had renal disorders. Both patients who died during the study had a renal disorder (P = .011), high PEDIS score (P = .016), low hemoglobin level (g/dL) (P = 0.017), and elevated TLC (per mcL) (P = .048); the mean PEDIS score of the patients who died was 10.50 ± 0.71; mean hemoglobin level (g/dL) was 5.50 ± 0.71, and mean TLC level (per mcL) was 24725.00 ± 2439.52.

No other significant differences were found between patient outcomes and the variables examined.

The current study showed that various patient variables, laboratory and radiological test results played a role in increasing the rate of lower extremity amputation in persons with infected DFUs. Age did not affect the rate of major or minor amputations. This is in contrast to findings by Leung et al,²⁰ in which a total of 340 patients (of whom 45 had type 1 diabetes) were studied in Regional Public Hospital, Hong Kong, and older age (mean age of amputees, 74 years) was found to be a statistically significant factor affecting the risk of major amputation in patients with DFU. This contrast may be because the current study had only 1 patient with type 1 diabetes and a small sample size.

When considering minor amputations, male sex was a statistically significant factor affecting the rate of amputation in the current study, with 22 of 39 men (56%) compared to 2 of 20 (10%) women undergoing minor amputation (P = .027). This is in accordance with the findings of Hämäläinen et al,²¹ which was a follow-up study of 998 patients, with 100 patients (50 men and 50 women) in each 10-year age group between 10 and 79, conducted in Finland. Those authors also found that male sex was independently related to the rate of amputation.

The duration of ulcer (days) was not found to have a significant effect on the rate of minor or major amputations. This is in contrast to Ugwu et al,²² who conducted a multi-center observational study in Nigeria between March 2016 and April 2017 (n = 366), in which ulcers of > 1 month’s duration prior to hospitalization had a significant effect on lower extremity amputation. A probable reason for the difference in our results could be that a majority of patients had previously sought a primary opinion elsewhere and presented to the authors after being referred.

There were 2 deaths in the study population; both participants had acute or chronic kidney disease. This variable was found to be significant. This is similar to the findings of Jeyaraman et al,²³ who conducted a study in the Northern Territory of Australia (N = 513) to study the effect of comorbidities on mortality in patients with DFU. These findings might be explained by the fact that diabetic nephropathy is the leading cause of mortality and morbidity in people with diabetes.²⁴

A total of 27.2% of the current study population had a history of previous amputation but that did not affect the rate of major or minor amputation. Shojaiefard et al²⁵ conducted a study consisting of 146 patients in Tehran, Iran, and concluded that previous amputation was a statistically significant factor when considering amputation in patients with DFU. Their study also stated that previous amputation is not an independent factor, but rather depends on the duration of diabetes, poor glycemic control, presence of neuropathy, and PAD. The reason for our result can be multifactorial. For example, several patients in the study population underwent traumatic amputation (amputation due to noninfectious factors such as road traffic or occupational accidents), while a majority of patients sought primary medical care from elsewhere and were seen by the authors after referral.

Our study showed no statistically significant difference in outcomes related to duration of diabetes. However, glycemic control is a significant risk for amputation⁶ and in our study there was a significant difference between the group that underwent major amputation and the one that did not in terms of HbA1c (%), with the median HbA1c (%) being highest in the major amputation group. HbA1c is a measure of blood glucose levels over 3 months; therefore, an elevated HbA1c signifies poorly controlled blood glucose levels, which indirectly signifies poor healing capacity.

Several patients in this study were recently diagnosed with diabetes following the development of foot ulceration and most likely had diabetes for a much longer period of time than indicated in the records.

Patient outcomes did not differ significantly between persons with different BMI scores. Gunasekaran et al²⁷ conducted a prospective study at a hospital in Bengaluru, India, involving 102 patients with DFUs and found that BMI and HbA1c affected the prognosis of DFUs. The mean BMI in their study was 25.1 kg/m², whereas in our study it was 20.28 kg/m². The reason behind the current finding could possibly be the lower incidence of obesity in the population of the northern hilly regions,²⁸ an area where the majority of our study study population resides.

PEDIS scores had a significant impact on major amputation and mortality rates. A retrospective study by Chuan et al²⁹ conducted in Chongqing, China, which included 364 patients, showed that when a DFU is classified with the use of the PEDIS system, the prognosis worsened with the increasing severity of each subcategory. This, and other studies, show that PEDIS scores have predictive validity.

The median hemoglobin values were highest in the 2 patients who died. Aziz et al³⁰ conducted a prospective study of 100 patients with DFUs treated at the National University Hospital of Singapore and reported that limb loss occurred in 51.6% of patients with hemoglobin levels ≤ 10.0 g/dL. This pointed toward a significant relationship between limb loss and hemoglobin levels (P < .001).

When comparing the group of patients who died with those who survived, a significant difference was found in terms of TLC, with the median TLC being highest in the patients who died. This is in alignment with the fact that TLC acts as a marker for ongoing sepsis in the body. TLC levels were not significantly related to amputation rate in this study.

Wound culture growth was present in 94.1% of patients in the major amputation group compared 64% of patients who did not undergo a major amputation. Fejfarová et al³¹ stated that the presence of pathogens resistant to all oral antibiotics and, especially of resistant Staphylococcus species, was significantly higher in patients with diabetes with lower limb amputations in comparison to patients without amputations.

When the population was studied for the presence of PAD based on arterial Doppler ultrasound, there was a significant difference in terms of distribution of arterial disease. Patients with more proximal vessel involvement had a higher rate of major amputation. Only 17.6% of participants in the major amputation group had no signs of PAD. Vasculopathy plays a central role in impaired wound healing in diabetes mellitus. Similar findings were found by Aziz et al.³⁰ The current study found a statistically significant difference in the rate of minor amputation in patients with low compared to high ABPI values, with the majority of amputees having an ABPI < 0.9. Aziz et al³⁰ stated that patients who underwent major amputation (38.8%) were likely to have ABPI  < 0.8 (indicating vasculopathy), whereas those with ABPI > 0.8 were less likely to undergo amputation (6.1%). Thus, ABPI < 0.8 can be a predictor for lower limb amputation.

The first limitation of this study was the small sample size, in addition to only 1 patient having type 1 diabetes. Because the study was time-bound, power analysis was not done to obtain sample size. Another limitation was that 75% of the study population was male. Patients included in the study were selected on the basis of clinical evidence of infection, and no objective criteria (eg, microbiological evidence) was obtained before initiation of the study. Healing was measured on the basis of epithelization of the ulcer, not the consideration of objective measures such as changes in dimensions. There was also a limited number of confounding factors that influence amputation in patients with DFU. Finally, no multivariate analysis was done to ascertain which variables were independent and which were dependent. Future studies should take these limitations into consideration and explore them further.

In this prospective observational study, the rate of major amputation was significantly higher in patients with PEDIS score (> 8), HbA1c values (> 8%), growth of wound culture, and presence of PAD (on the basis of Doppler study). Regarding minor amputation, significant factors were male sex and ABPI (< 0.9). Mortality was dependent on factors that included the presence of renal disorders, PEDIS score (> 10), hemoglobin (< 8 g/dL), and TLC (> 11 000 per mcL). Patients with the above-mentioned factors should be provided aggressive, customized treatment, and early intervention should be planned to improve prognoses. PEDIS score should be calculated for every patient with DFU and taken into serious consideration. Routine investigations should not only include arterial Doppler ultrasound, but also ABPI calculation.

Dr. Chaudhary is a postgraduate student, Dr. Huda is an additional professor, Dr. Roshan is a postgraduate student, Professor Basu is the department head, Dr. Rajput is an assistant  professor, and Dr. Singh is an assistant professor, Department of General Surgery, All India Institute of Medical Sciences, Rishikesh, India. Address all correspondence to: Farhanul Huda, Additional Professor, Department of General Surgery, All India Institute of Medical Sciences, Rishikesh, District Dehradun Uttarakhand 249203, India; tel: +91 9997533211; email: farhanul1973huda@gmail.com.