Diabetic Foot Ulcer–Associated Versus Other Lower‐Extremity Necrotising Soft‐Tissue Infections: A Retrospective Cohort Study

ABSTRACT

To compare clinical characteristics, microbiology, treatment burden and outcomes between lower‐extremity necrotising soft tissue infections associated with diabetic foot ulcers (DFU) and lower‐extremity necrotising soft tissue infections not associated with DFU. We retrospectively reviewed 86 surgically treated patients at a single tertiary hospital from 2015 to 2025. Twenty patients had diabetic foot ulcer–associated infections and 66 had infections from other causes. Baseline characteristics, laboratory findings, microbiology, resource utilisation, wound healing, amputation and mortality were compared between the cohorts. The diabetic foot ulcer‐associated group had a longer diabetes duration, higher glycated haemoglobin and a greater prevalence of hypertension, coronary heart disease and chronic kidney disease. Microbiological profiles showed overlapping distributions of major pathogens. The proportion of patients who were alive with a healed wound at 90 days was lower in the DFU‐associated group (35.0% vs. 80.3%), and overall amputation was more frequent (55.0% vs. 15.2%), primarily driven by minor amputation (40.0% vs. 1.5%). Major amputation (15.0% vs. 13.6%) and mortality (10.0% vs. 4.5%) did not differ statistically, although estimates were imprecise. Lower‐extremity necrotising soft tissue infections associated with DFU were associated with greater comorbidity, resource utilisation, minor amputation and a lower proportion of patients alive with a healed wound at 90 days. This retrospective study cannot determine whether minor amputation improves survival or prevents major amputation.

Key Points

• Worse Short‐Term Outcomes: Patients with diabetic foot ulcer‐associated necrotizing soft tissue infections (DF/NSTI) demonstrated a significantly lower proportion of patients alive with a healed wound at 90 days (35.0% vs. 80.3%) and greater healthcare resource utilization compared to those with non‐DFU‐associated NSTI.

• Amputation for Source Control: The DF/NSTI cohort experienced a significantly higher overall amputation rate (55.0% vs. 15.2%), which was primarily driven by minor amputations (40.0% vs. 1.5%) necessitated by the distal anatomical origin and the need for localized surgical source control.

• Systemic Comorbidity Burden: DF/NSTI patients presented with a heavier comorbidity burden, including a significantly longer duration of diabetes, higher glycated hemoglobin (HbA1c) levels, and higher prevalences of hypertension, coronary heart disease, and chronic kidney disease.

• Microbiological Profile: While major pathogen distributions overlapped between the two cohorts, exploratory analyses indicated that Pseudomonas aeruginosa was more frequently detected in the DF/NSTI group.

• Clinical Implications: A concurrent diabetic foot ulcer in NSTI serves as a late‐stage clinical marker of cumulative systemic metabolic burden rather than merely a local factor, requiring an integrated treatment strategy of prompt surgical debridement and multidisciplinary medical management.

1. Introduction

1.1. Current Status of Necrotising Soft Tissue Infections and Diabetic Foot Ulcers

At the intersection of the modern treatment modalities for acute severe infections and endocrine metabolic diseases, clinical decision‐making relies heavily on established guidelines. However, when necrotising soft tissue infections (NSTI)—an extreme form of soft tissue infection—overlaps with a diabetic foot ulcer (DFU)—a model of chronic ischemia and neuropathy—in the same individual, existing unidimensional clinical paradigms for managing acute severe infections face significant challenges.

The core pathophysiological hallmark of NSTI is the fulminant spread of pathogens along the superficial fascial planes, frequently complicated by extensive microvascular thrombosis and progressive tissue necrosis [1]. Epidemiological data indicate that despite significant advancements in modern intensive care medicine, the mortality rate remains as high as 20%–30% and can exceed 50% if complicated by toxic shock syndrome [2, 3]. Consequently, surgical intervention for classic NSTI adheres to a time‐dependent principle, executing an early and radical debridement strategy [4]. Conversely, as a chronic complication with a reported global prevalence of approximately 6.3%, the clinical management of DFU is deeply rooted in the paradigm of chronic wound care [5]. Because these patients frequently present with concurrent peripheral neuropathy and microvascular occlusion, the core objective emphasises staged, conservative debridement based on precise vascular assessment [6].

Both NSTI and DFU management are clinically guided by tissue viability and the need for timely source control while preserving viable tissue. When NSTI arises from a diabetic foot ulcer, the distal anatomical origin of infection, chronic ulceration, diabetes‐related comorbidities and possible vascular impairment may influence operative options and short‐term wound healing. However, the effect of any specific amputation strategy cannot be determined from retrospective observational data.

1.2. Research Objectives

This retrospective cohort study aimed to compare clinical characteristics, microbiological findings, treatment burden and 90‐day outcomes between DFU‐associated and non‐DFU‐associated lower‐extremity NSTI. Minor and major amputations were evaluated as observed clinical outcomes rather than as predefined treatment strategies.

2. Patients and Methods

2.1. Study Population

This retrospective cohort study included 86 patients diagnosed with lower‐extremity NSTI at a tertiary academic hospital between January 2015 and December 2025. Given the retrospective observational nature of this study, the requirement for written informed consent was waived by the ethics committee. All included cases were confirmed at surgical exploration based on grey devitalised fascia, loss of normal fascial resistance, lack of fresh bleeding, extensive undermining or dishwater‐like exudate. Histopathological examination was used as a supportive adjunct when available and was not required for inclusion.

2.1.1. Grouping Criteria

Based on the aetiology of the infection and pathological characteristics, patients were categorised into two groups:

DF/NSTI Group (n = 20): This group was defined as NSTI directly resulting from the proximal spread of a diabetic foot ulcer (DFU). The inclusion criteria required all of the following:

1. Presence of a DFU as defined by the IWGDF (2023) guidelines (i.e., a full‐thickness skin defect occurring distal to the ankle in a patient with diabetes, with or without infection) [7].

2. The onset of NSTI was directly linked to the foot ulcer, with confirmed proximal spread of the infection along the fascial planes from the ulcer to the lower leg or thigh.

3. Intraoperative findings met the aforementioned universal diagnostic criteria for NSTI. (e.g., lack of fascial resistance to blunt dissection, presence of grey necrotic fascia and foul‐smelling ‘dishwater’ exudate).

   (Note: The LRINEC score was concurrently recorded as a laboratory diagnostic adjunct for all included patients.) [8].

Non‐DFU‐associated NSTI Group (n = 66): This group comprised patients with lower‐extremity NSTI diagnosed in accordance with the World Society of Emergency Surgery (WSES) guidelines [9]. To differentiate the clinical aetiology from the DF/NSTI cohort, patients were assigned to the non‐DFU‐associated NSTI group based on the following criteria:

1. The primary infection did not originate from a DFU; established triggers included trauma, insect bites, skin abscesses or idiopathic infections.

2. Patients with concurrent diabetes mellitus were assigned to this group if the origin of the infection was unrelated to a foot lesion (e.g., NSTI secondary to direct calf trauma) or if the primary entry portal was located proximal to the ankle. These cases were classified as non‐DFU‐associated NSTI on the basis of a documented non‐DFU portal of entry; this etiologic classification does not imply that diabetes‐related neuropathy, vascular disease or other systemic effects were absent.

Common Exclusion Criteria: (1) Non‐infectious ulcers: necrosis caused by malignancies (e.g., squamous cell carcinoma), vasculitis (e.g., pyoderma gangrenosum) or radiation‐induced ulcers; (2) Immune interference: patients receiving long‐term, high‐dose corticosteroids or immunosuppressive therapy; (3) Extremis state: patients presenting with irreversible shock upon admission, rendering them unfit for surgery or those whose families abandoned treatment leading to automatic discharge against medical advice.

2.2. Study Methods

2.2.1. Clinical Data Collection

Baseline data within 24 h of admission for all included patients were systematically extracted from the hospital's electronic medical record system. The collection dimensions included:

1. Demographic and anthropometric characteristics: Age, sex and body mass index (BMI).

2. Diabetes‐specific parameters: Duration of diabetes, treatment regimens and complications (with a focus on macrovascular diseases such as hypertension and coronary heart disease).

3. Admission status: Vital signs (temperature and blood pressure), baseline Wagner classification of the wound and initial laboratory results (complete blood count, biochemical profiles, inflammatory markers, etc.).

2.2.2. LRINEC Score and Diagnostic Adjunct

The LRINEC score was recorded as a laboratory‐based diagnostic adjunct and was not used as a measure of disease severity or to determine operative management. This score is calculated based on six weighted indicators: CRP, white blood cell count, haemoglobin, serum sodium, creatinine and glucose [8]. A total Score > 6 suggests a high risk of NSTI. Given that diabetic patients frequently present with baseline hyperglycemia and diabetic nephropathy (elevated creatinine), the baseline LRINEC score may be artificially elevated (false positive). Therefore, in this study, the definitive diagnosis of NSTI was established based on intraoperative surgical exploration findings, supported by corresponding radiological features (e.g., subcutaneous emphysema) and clinical signs (e.g., crepitus).

2.2.3. Microbiological Specimen Collection

To ensure the fidelity of the microbiological outcomes, microbiological samples were obtained via intraoperative deep tissue retrieval or sterile wound swabs during bedside debridement, depending on the clinical scenario, under strict aseptic conditions:

1. Timing of collection: Specimens were preferentially obtained during the initial surgical debridement following admission. For patients whose surgical intervention was temporarily deferred, collection occurred during the first bedside dressing change.

2. Collection methodology: Prior to sampling, the wound surface was thoroughly irrigated with sterile saline to remove superficial necrotic debris. For open or incised wounds, tissue specimens from the deep wound bed or adjacent viable tissue were preferentially obtained for culture. For intact wounds, deep purulent exudate was collected using a sterile swab.

3. Management of transferred patients with prior antimicrobial exposure: For patients transferred from external facilities who had already undergone local debridement or systemic antibiotic therapy, to mitigate the confounding interference of prior interventions on the microecological profile and microbial distribution, the routine re‐collection of deep tissue specimens during the initial surgical exploration at our institution was mandated.

2.2.4. Comprehensive Treatment Strategy

1. Multidisciplinary Medical Management and Antimicrobial Therapy: A multidisciplinary team (MDT) approach was initiated immediately upon admission. Alongside foundational treatments such as intensive insulin therapy for glycemic control, hemodynamic stabilisation and the correction of hypoproteinemia, the antimicrobial strategy adhered to a ‘de‐escalation’ principle. In the early admission phase, empirical broad‐spectrum antibiotics covering both aerobic and anaerobic bacteria were administered in strict accordance with the IDSA guidelines [10]. Once the results of local microbiological cultures and antimicrobial susceptibility testing were available, the regimen was promptly adjusted to targeted, susceptible antibiotics.

2. Principles of Surgical Intervention: All patients routinely underwent emergency surgical intervention following the completion of preoperative evaluation and preparation. The surgical procedure mandated the thorough debridement of necrotic fascia, liquefied tissue and devitalised muscle, while preserving skin and soft tissue with viable blood supply to the greatest extent possible. Postoperatively, based on the wound condition, adjunctive negative pressure wound therapy (NPWT/VSD) or the implantation of antibiotic‐impregnated polymethyl methacrylate (PMMA) beads or blocks was utilised [6, 11]. The application of PMMA was selectively indicated for patients presenting with concurrent chronic osteomyelitis or deep osseous destruction (e.g., post‐traumatic avascular necrosis with infection). Following thorough debridement and sequestrectomy, PMMA was placed within the residual bone defect to obliterate the dead space and provide a sustained, high‐concentration local release of antibiotics. For patients whose infections were solely confined to the soft tissues and fascial layers without deep bone involvement, PMMA was not utilised.

3. Amputation Decision Criteria: Amputation (including both major and minor amputations) was performed primarily to achieve infection source control. While a MDT approach was utilised for preoperative resuscitation, the determination of the amputation level was made intraoperatively by the senior surgeon. This assessment was based on macroscopic tissue viability, fascial bleeding and the proximal extent of necrosis to facilitate immediate source control. Intraoperative frozen section histology was utilised as an adjunct in equivocal cases to assess margin viability, with surgical progression prioritised over awaiting pathological confirmation.

In this study, ‘minor amputation’ was defined as any amputation at or distal to the ankle joint (e.g., toe, transmetatarsal or partial foot amputation); ‘major amputation’ was defined as any amputation proximal to the ankle joint (e.g., below‐knee or above‐knee amputation). Indications for amputation included the following scenarios: (1) severe infection contributing to septic shock or multiple organ failure; (2) extensive ischemic necrosis of the affected limb that could not be revascularised; and (3) extensive soft tissue loss resulting in a loss of limb function not amenable to reconstructive techniques.

Amputation decisions were based on the documented intraoperative extent of non‐viable tissue, ischemia and the feasibility of reconstruction. By relying on these objective clinical and anatomical findings, the study aimed to apply consistent intraoperative decision‐making principles across both cohorts. Nevertheless, given the retrospective design and inherent variations in the initial anatomical site of infection, these standardised intraoperative criteria cannot entirely mitigate structural selection biases or residual confounding in prognostic evaluations.

2.2.5. Follow‐Up and Outcome Determination

Clinical outcomes were assessed at 90 days after the index operation. Ninety‐day vital status, amputation status and wound status were ascertained for all patients, and no patient was lost to follow‐up. Any amputation performed within 90 days was classified according to the highest anatomical level reached. All four criteria were required to be fulfilled simultaneously for a wound to be classified as healed: (1) complete epithelialisation of the final wound; (2) absence of purulent drainage; (3) absence of sinus formation; (4) resolution of local signs of infection. Patients who died before Day 90 were classified as not having achieved wound healing. Consequently, the primary analysis for wound healing was strictly based on the full cohort as a composite endpoint of ‘alive and healed at 90 days’. Conditional analyses based solely on survivors were not performed to prevent survival bias.

2.3. Statistical Analysis

All statistical analyses were performed using SPSS 22.0 (IBM Corp., Armonk, NY, USA) and R software (version 4.3.0). Continuous variables were initially assessed for normality using the Shapiro–Wilk test: normally distributed data are presented as mean ± standard deviation and were compared using the independent Student's t‐test or Welch's t‐test; non‐normally distributed data are expressed as median and interquartile range (IQR) and were compared using the Mann–Whitney U test. Categorical variables are described as frequencies and percentages (n [%]), and intergroup comparisons were conducted using Pearson's chi‐square test or Yates' continuity correction (or Fisher's exact test when expected frequencies were < 5). p values were reported as auxiliary metrics to describe data distributions rather than serving as absolute determinants of clinical significance. All statistical tests were two‐tailed, and p < 0.05 was considered statistically significant. For selected categorical clinical outcomes, unadjusted risk ratios and 95% confidence intervals were calculated using the Katz log method. Organism‐level comparisons were exploratory, and no adjustment for multiple comparisons was performed.

3. Results

3.1. Baseline Characteristics and Clinical Features

3.1.1. Patient Clinical Profiles

A total of 157 patients were initially screened for this study. After applying the exclusion criteria, 71 patients were excluded: 45 did not meet the intraoperative diagnostic criteria for fascial necrosis, 14 presented with non‐infectious ulcers, 7 experienced pre‐surgical care withdrawal, and 5 had incomplete medical records. This resulted in a final cohort of 86 patients. Among them, 20 patients were in the DF/NSTI group and 66 patients were in the non‐DFU‐associated NSTI group. There were no statistically significant differences between the two groups regarding demographic characteristics such as age, gender distribution and smoking/alcohol history (all p > 0.05). However, among patients with concurrent diabetes, the DFU‐associated group had a significantly longer duration of diabetes compared to the non‐DFU‐associated NSTI group (median 9.50 [IQR, 6.75, 13.25] vs. 6.00 [IQR, 4.00, 8.00] years, p = 0.001). Similarly, the DFU‐associated group had significantly elevated glycated haemoglobin (HbA1c) levels among patients with available data (11.03 ± 4.99 vs. 7.94 ± 2.32, p = 0.015). Additionally, the prevalence of hypertension (55.0% vs. 21.2%, p = 0.008), coronary heart disease (45.0% vs. 16.7%, p = 0.020) and chronic kidney disease (65.0% vs. 16.7%, p < 0.001) were significantly higher in the DF/NSTI group (Table 1).

TABLE 1.

Baseline characteristics and clinical features of the study population.

Variables	DF/NSTI (n = 20)	Non‐DFU‐associated NSTI (n = 66)	p
Demographics
Age (years)	54.90 ± 11.32	55.65 ± 14.12	0.808
Male, n (%)	14 (70.0)	48 (72.7)	1.000
Smoking history, n (%)	4 (20.0)	18 (27.3)	0.770
Drinking history, n (%)	8 (40.0)	21 (31.8)	0.591
Clinical and metabolic parameters
Concurrent diabetes, n (%)	20 (100.0)	42 (63.6%)	0.004*
Diabetes duration, years, among patients with diabetes (n = 20 and n = 42) a	9.50 (6.75, 13.25)	6.00 (4.00, 8.00)	0.001*
HbA1c (%), among patients with available data (n = 20 and n = 42) a	11.03 ± 4.99	7.94 ± 2.32	0.015*
Admission temperature (°C)	38.25 (37.70,38.60)	37.90 (37.50, 38.38)	0.068
BMI (kg/m2)	23.50 ± 2.62	23.32 ± 2.56	0.783
Comorbidities, n (%)
Hypertension	11 (55.0)	14 (21.2)	0.008*
Coronary heart disease	9 (45.0)	11 (16.7)	0.020*
Chronic kidney disease	13 (65.0)	11 (16.7)	< 0.001*

Note: Data are presented as mean ± standard deviation (SD) for normally distributed continuous variables, median and interquartile range [IQR] for non‐normally distributed continuous variables, and numbers (percentages) for categorical variables.

Abbreviations: BMI, body mass index; DF/NSTI, diabetic foot ulcer‐associated necrotising soft tissue infection; HbA1c, glycated haemoglobin.

^a Diabetes duration and HbA1c were analysed exclusively within the restricted all‐diabetic cohort: DF/NSTI (n = 20) vs. non‐DFU‐associated NSTI (n = 42).

^* p < 0.05 indicates statistical significance.

3.1.2. Clinical Characteristics and Laboratory Indicators

Analysis of diagnostic adjuncts showed a significantly elevated LRINEC score in the DF/NSTI group (median 9.50 [8.00, 10.20] vs. 8.00 [7.00, 9.75], p = 0.018). Laboratory tests revealed that admission blood glucose was significantly higher in the DF/NSTI group (median 14.60 [12.38, 18.86] vs. 11.98 [7.19, 15.10] mmol/L, p = 0.005). Additionally, the DF/NSTI group exhibited more severe anaemia, with haemoglobin levels significantly lower than the non‐DFU‐associated NSTI group (87.55 ± 17.72 vs. 98.14 ± 25.76 g/L, p = 0.044). Systemic inflammatory markers (CRP, WBC) and tissue enzyme indicators (CK, LDH) did not demonstrate statistically significant differences between the two groups (all p > 0.05) (Table 2).

TABLE 2.

Clinical characteristics and laboratory indicators at admission.

Variables	DF/NSTI (n = 20)	Non‐DFU‐associated NSTI (n = 66)	p
Laboratory diagnostic adjunct
LRINEC score	9.50 (8.00, 10.2)	8.00 (7.00, 9.75)	0.018*
Inflammatory markers
WBC (×109/L)	17.76 (13.95, 25.44)	15.97 (10.63, 21.26)	0.092
CRP (mg/L)	234.50 (178.14, 269.98)	198.55 (179.76, 236.80)	0.230
Metabolic and nutritional indicators
Blood glucose (mmol/L)	14.60 (12.38, 18.86)	11.98 (7.19, 15.10)	0.005*
Haemoglobin (g/L)	87.55 ± 17.72	98.14 ± 25.76	0.044*
Tissue damage markers
CK (U/L)	62.00 (29.00, 136.25)	44.50 (24.50, 121.25)	0.886
LDH (U/L)	226.50 (195.00, 280.50)	238.00 (190.25, 334.75)	0.874
Renal function and electrolytes
Creatinine (μmol/L)	69.50 (46.00, 96.00)	64.00 (50.25, 95.75)	0.874
Sodium (mmol/L)	132.87 ± 4.50	135.19 ± 4.85	0.060
Liver function and coagulation
TBil (μmol/L)	12.40 (9.65, 16.62)	15.60 (11.25, 24.07)	0.067
PT (s)	12.90 (11.93, 14.33)	13.20 (12.20, 14.60)	0.335

Note: Continuous variables are presented as mean ± SD or median (interquartile range). Variables were compared using the independent samples t‐test or Mann–Whitney U test, as appropriate.

Abbreviations: CK, creatine kinase; CRP, C‐reactive protein; DF/NSTI, diabetic foot ulcer–associated necrotising soft tissue infection; LDH, lactate dehydrogenase; LRINEC, Laboratory Risk Indicator for Necrotising Fasciitis; PT, prothrombin time; TBil, total bilirubin; WBC, white blood cell.

^* p < 0.05 indicates statistical significance.

3.2. Microbiological Characteristics

Microbiological evaluation was based on a combination of intraoperative deep tissue specimens and bedside sterile swabs. Overall, positive cultures were obtained in 86 patients (100.0%), with polymicrobial infections present in 33 cases (38.4%). The mean number of isolated organisms per patient was 1.63. Anaerobic cultures were not systematically performed.

Staphylococcus aureus was the most frequently detected Gram‐positive pathogen, with no significant difference between groups (40.0% vs. 24.2%, p = 0.169). Regarding Gram‐negative bacteria, Klebsiella pneumoniae was detected in 30.0% of the DF/NSTI group and 15.2% of the non‐DFU‐associated NSTI group (p > 0.05). Notably, Pseudomonas aeruginosa was detected in 5 of 20 patients (25.0%) in the DF/NSTI group and 4 of 66 patients (6.1%) in the non‐DFU‐associated NSTI group (Risk Ratio = 4.13, 95% CI: 1.22–13.90; unadjusted p = 0.029). Given the small event counts and multiple organism‐level comparisons performed, this finding should be considered exploratory (Table 3, Figure 1).

TABLE 3.

Microbiological profile of the study population.

Pathogens, n (%)	DF/NSTI (n = 20)	Non‐DFU‐associated NSTI (n = 66)	Risk ratio (95% CI)	p
S. aureus	8 (40.0)	16 (24.2)		0.169
K. pneumoniae	6 (30.0)	10 (15.2)		0.188
P. aeruginosa	5 (25.0)	4 (6.1)	4.13 (1.22–13.90)	0.029*
A. baumannii	2 (10.0)	10 (15.2)		0.724
E. cloacae	2 (10.0)	5 (7.6)		0.662
E. coli	2 (10.0)	5 (7.6)		0.662
Proteus spp.	2 (10.0)	4 (6.1)		0.620

Note: Data are presented as n (%). Groups were compared using Pearson's chi‐square test or Fisher's exact test, as appropriate based on expected cell counts.

Abbreviations: A. baumannii, Acinetobacter baumannii; DF/NSTI, diabetic foot ulcer–associated necrotising soft‐tissue infection; E. cloacae, Enterobacter cloacae; E. coli, Escherichia coli; K. pneumoniae, Klebsiella pneumoniae; P. aeruginosa, Pseudomonas aeruginosa; S. aureus, Staphylococcus aureus .

^* p < 0.05 indicates statistical significance.

FIGURE 1.

Pathogen distribution between the DF/NSTI group and the non‐DFU‐associated NSTI group. DF, diabetic foot; DFU, diabetic foot ulcer; NSTI, necrotising soft tissue infection.

3.3. Treatment Burden and Clinical Outcomes

Resource utilisation analysis demonstrated that the DF/NSTI group had a significantly longer hospital stay (median 47.50 [30.50, 62.25] days vs. 29.00 [20.00, 41.75] days, p = 0.013) and higher hospitalisation costs (median 9.55 [5.35, 12.27] vs. 5.21 [3.50, 7.72] × 10^4 CNY, p = 0.005) compared to the non‐DFU‐associated NSTI group. Furthermore, a greater reliance on intraoperative frozen section pathology was observed in the DF/NSTI cohort (60.0% vs. 19.7%, p < 0.001). While PMMA application was also more frequent in the DF/NSTI group (40.0% vs. 4.5%, p < 0.001), this adjunctive use was strictly driven by the specific indication of concurrent deep osseous destruction or osteomyelitis; thus, it reflects the underlying structural pathology rather than an independent measure of excess treatment burden.

Regarding prognostic outcomes, no statistically significant between‐group differences were detected in all‐cause mortality (10.0% vs. 4.5%, p = 0.329) or major amputation rates (15.0% vs. 13.6%, p = 1.000); however, event numbers were small and the statistical estimates were too imprecise to establish clinical equivalence. The overall amputation rate was significantly higher (55.0% vs. 15.2%, p < 0.001). This difference in amputation risk was primarily driven by the minor amputation rate, which was 40.0% in the DF/NSTI group compared to 1.5% in the non‐DFU‐associated NSTI group (p < 0.001) (Table 4, Figure 2).

TABLE 4.

Healthcare resource utilisation and clinical outcomes.

Variables	DF/NSTI (n = 20)	Non‐DFU‐associated NSTI (n = 66)	Risk ratio (95% CI)	p
Healthcare resource utilisation
Length of hospital stay (days)	47.50 (30.50, 62.25)	29.00 (20.00, 41.75)		0.013*
Hospitalisation costs (×104 CNY)	9.55 (5.35, 12.27)	5.21 (3.50, 7.72)		0.005*
Number of surgeries	7.25 ± 3.95	5.23 ± 3.83		0.043*
Intraoperative adjuncts, n (%)
Intraoperative frozen section pathology	12 (60.0)	13 (19.7)		< 0.001*
Use of PMMA	8 (40.0)	3 (4.5)		< 0.001*
Clinical outcomes, n (%)
Overall amputation	11 (55.0)	10 (15.2)	3.63 (1.81–7.27)	< 0.001*
Minor amputation	8 (40.0)	1 (1.5)	26.40 (3.51–198.57)	< 0.001*
Major amputation	3 (15.0)	9 (13.6)	1.10 (0.33–3.68)	1.000
Alive with a healed wound at 90 days	7 (35.0)	53 (80.3)	0.44 (0.24–0.80)	< 0.001*
All‐cause mortality	2 (10.0)	3 (4.5)	2.20 (0.39–12.26)	0.329

Note: Data are presented as median and interquartile range [IQR] for continuous variables, and numbers (percentages) for categorical variables. All prognostic endpoints and wound healing status were evaluated at 90 days after the index operation. Amputation classification was applied hierarchically based on the highest anatomical level performed during surgical procedures. For categorical outcomes, effect estimates are reported as risk ratios (RR) with 95% confidence intervals (CI). Due to small event numbers in mortality and major amputation, estimates reflect exploratory observations.

Abbreviations: CI, confidence interval; CNY, Chinese Yuan; DF/NSTI, diabetic foot ulcer–associated necrotising soft tissue infection; PMMA, polymethyl methacrylate (antibiotic‐loaded bone cement).

^* p < 0.05 indicates statistical significance.

FIGURE 2.

Clinical outcomes between the DF/NSTI group and the non‐DFU‐associated NSTI group. (A) Amputation rates. Overall and minor amputation rates were significantly higher in the DF/NSTI group (p < 0.001). (B) Ninety‐day alive‐and‐healed status and mortality. The non‐DFU‐associated NSTI group demonstrated a significantly higher proportion of patients alive with a healed wound at 90 days (80.3% vs. 35.0%, p < 0.001), whereas mortality rates did not differ significantly between the groups (p = 0.329). Group 1: DF/NSTI group (n = 20); Group 2: Non‐DFU‐associated NSTI group (n = 66). *p < 0.05.

3.4. Descriptive Comparison in the Restricted All‐Diabetic Cohort (n = 62)

To reduce heterogeneity related to diabetes status, a restricted descriptive analysis was conducted among patients with concurrent diabetes (n = 62, comprising 20 DF/NSTI cases and 42 diabetic non‐DFU‐associated NSTI cases). Within this subset, a total of 18 amputation events occurred (11 in the DF/NSTI group and 7 in the non‐DFU‐associated NSTI group). The descriptive comparison demonstrated that the overall and minor amputation rates remained significantly higher in the DF/NSTI subgroup compared to diabetic patients with non‐DFU‐associated NSTI (55.0% vs. 16.7%, p = 0.003; and 40.0% vs. 2.4%, p < 0.001, respectively). In contrast, major amputation rates (15.0% vs. 14.3%, p = 1.000) and all‐cause mortality (10.0% vs. 4.8%, p = 0.584) showed no statistically significant between‐group differences, though event counts were small. Furthermore, traditional inflammatory markers and laboratory parameters, including albumin, C‐reactive protein and white blood cell count, exhibited comparable distributions across these subgroups (all p > 0.05) (Table 5).

TABLE 5.

Descriptive comparison of clinical and laboratory characteristics in the restricted all‐diabetic cohort (n = 62).

Variables	DF/NSTI group (n = 20)	Diabetic non‐DFU‐associated NSTI group (n = 42)	p
Overall amputation	11 (55.0)	7 (16.7)	0.003*
Minor amputation	8 (40.0)	1 (2.4)	< 0.001*
Major amputation	3 (15.0)	6 (14.3)	1.000
All‐cause mortality	2 (10.0)	2 (4.8)	0.584
Albumin (g/L)	29.8 ± 4.2	31.5 ± 4.6	0.182
C‐reactive protein (mg/L)	234.5 (178.14, 269.98)	200.08 (175.20, 247.99)	0.299
White blood cell count (109/L)	17.76 (13.95, 25.44)	15.97 (10.88, 22.42)	0.168

Note: This restricted descriptive analysis was conducted exclusively on patients with concurrent diabetes (n = 62) to evaluate variable distributions within a restricted all‐diabetic subset, reducing heterogeneity related to diabetes status. Continuous variables with normal distributions are presented as mean ± SD, while non‐normally distributed variables are expressed as median [IQR].

Abbreviations: DF/NSTI, diabetic foot ulcer‐associated necrotising soft tissue infection; IQR, interquartile range; SD, standard deviation.

^* p < 0.05 indicates statistical significance.

3.5. Representative Case

A 66‐year‐old female with a 15‐year history of diabetes and chronic kidney disease presented with septic shock and extensive wet gangrene of the right foot. Erythema, edema and purulent exudate extended to the distal tibia, accompanied by palpable crepitus and severe systemic inflammation (WBC 31.95 × 10⁹/L, CRP 243.29 mg/L, LRINEC Score 9) (Figure 3). Emergent exploratory fasciotomy yielded copious ‘dishwater‐like’ exudate, revealing macroscopic fascial necrosis propagating to the mid‐thigh, beyond superficial demarcations (Figure 4). Intraoperative frozen section and subsequent histopathology corroborated the NSTI diagnosis, demonstrating extensive deep fascial necrosis, dense neutrophilic infiltration, microabscess formation and diffuse microvascular thrombosis (Figure 5). Because viable tissue margins were inadequate at the proximal tibia, the initially planned below‐knee amputation was deemed unfeasible; consequently, an emergent right above‐knee amputation was performed to ensure complete source control, leaving the stump open for serial assessment (Figure 6). Following serial debridement and negative pressure wound therapy (NPWT) for wound bed preparation, delayed stump closure was performed. By postoperative Day 45, the incision remained intact without erythema, purulent drainage or wound dehiscence (Figure 7).

FIGURE 3.

Preoperative clinical presentation of the affected foot. (A) Dorsal view showing extensive moist gangrene with purplish‐black discoloration, local skin ulceration, and serosanguineous exudate. Diffuse swelling of the dorsal soft tissues with ill‐defined borders is present. (B) Plantar view demonstrating an irregular deep ulcer with purulent exudate, accompanied by marginal erythema and desquamation.

FIGURE 4.

Intraoperative findings during emergent debridement. Extensive incision and decompression reveal diffuse liquefactive necrosis of the deep fascial layer, characteristic ‘dishwater‐like’ purulent exudate, and undermining necrotic cavities separating the fascia from the underlying skeletal muscle.

FIGURE 5.

Histopathological features of necrotising soft tissue infections (H&E staining). (A) Low‐magnification image (10×) demonstrating severe vascular and fascial damage. Asterisk (*) indicates hyaline degeneration and thickening of the vessel wall; black arrows indicate basophilic dystrophic calcification; white arrowheads indicate dense inflammatory infiltration and microthrombosis. Scale bar = 200 μm. (B) High‐magnification detail (40×) showing the anucleate hyalinised wall (*), calcified plaque (black arrows), and intraluminal fibrin‐leukocyte meshwork (white arrowheads). Scale bar = 100 μm.

FIGURE 6.

Local wound appearance following open debridement. Following serial debridement, the wound base demonstrates fresh granulation tissue formation and adequate drainage, preparing the site for delayed stump closure.

FIGURE 7.

Appearance of the surgical site during postoperative outpatient follow‐up. Clinical examination of the residual limb following delayed stump closure reveals an intact, well‐healed incision. No erythema, oedema or purulent exudate is observed in the surrounding soft tissues.

4. Discussion

4.1. Interpretation of Amputation and Other Clinical Outcomes

We observed that DFU complicated by necrotising soft tissue infections (DF/NSTI) represent a clinical subpopulation with unique pathophysiological characteristics. Notably, while DF/NSTI patients faced a significantly higher risk of overall amputation (55.0% vs. 15.2%, p < 0.001), their all‐cause mortality (10.0% vs. 4.5%, p = 0.329) and major amputation rates (15.0% vs. 13.6%, p = 1.000) did not differ significantly from those with non‐DFU‐associated NSTI (Table 4). However, it is crucial to note that the event numbers for these specific endpoints were small; therefore, this absence of statistical significance reflects data uncertainty and imprecise estimates rather than definitive clinical equivalence.

The increased hospitalisation duration (median 47.50 vs. 29.00 days, p = 0.013) and costs (9.55 vs. 5.21 × 10⁴ CNY, p = 0.005) in the DF/NSTI group indicate the complexity of perioperative management. However, the statistical disparity in overall amputations must be interpreted through an anatomical lens. Minor amputation was more frequently performed in the DFU‐associated group, reflecting the distal anatomical origin of infection and intraoperative assessment of nonviable tissue. Conversely, for non‐DFU‐associated NSTI originating more proximally (e.g., lower leg or thigh), minor amputations are anatomically inapplicable. Therefore, while DF/NSTI indicates severe local tissue destruction, the heightened ‘overall amputation risk’ heavily reflects the localised surgical intervention pattern characteristic of the diabetic foot, rather than representing an elevated biological vulnerability to whole‐limb loss or death [10, 12]. Unlike non‐DFU‐associated NSTI, which often originates from acute trauma in previously intact skin, the distal onset in DF/NSTI influences this specific surgical endpoint.

4.2. Subgroup Analysis of the All‐Diabetic Cohort

To further evaluate data patterns while reducing heterogeneity related to diabetes status, a descriptive subgroup analysis was conducted exclusively among patients with concurrent diabetes (n = 62). Within this subset, the DF/NSTI group maintained a significantly higher frequency of overall (55.0% vs. 16.7%, p = 0.003) and minor amputations (40.0% vs. 2.4%, p < 0.001) compared to diabetic patients with non‐DFU‐associated NSTI, whereas major rates (15.0% vs. 14.3%, p = 1.000) and all‐cause mortality rates (10.0% vs. 4.8%, p = 0.584) showed no statistically significant differences.

Within the diabetes‐only subgroup, patients with DFU‐associated NSTI had a longer duration of diabetes and higher HbA1c levels than diabetic patients with non‐DFU‐associated NSTI. These findings suggest that the presence of a DFU may serve as a clinical marker of longer‐standing diabetes and greater cumulative glycemic burden, rather than representing an isolated local factor. Accordingly, the observed differences in infection‐related outcomes should not be attributed to the presence of a DFU alone; the systemic effects of chronic diabetes may also have contributed to the observed clinical pattern. However, because of the retrospective design, small sample size, limited number of outcome events and absence of adjusted analyses specifically evaluating diabetes duration and HbA1c, their independent associations with clinical outcomes could not be determined. These observations should therefore be considered hypothesis‐generating. Future prospective studies should examine whether diabetes duration and cumulative glycemic burden independently influence infection severity, wound healing and amputation outcomes [13] (Table 5).

4.3. Analysis of Systemic Inflammatory Markers and Tissue Injury Enzymes

In our study population, baseline levels of acute inflammatory markers (CRP and WBC) and tissue injury enzymes (CK and LDH) showed overlapping distributions between the DF/NSTI and non‐DFU‐associated NSTI groups (all p > 0.05) (Table 2). Given the small sample size and the absence of multivariable‐adjusted analyses, these findings do not establish biological equivalence, nor do they definitively identify the mechanisms underlying the observed outcome differences.

Notably, baseline CK and LDH distributions overlapped between the DF/NSTI and non‐DFU‐associated NSTI groups. From a clinical perspective, this observational finding highlights an important diagnostic pitfall: these markers should not be used in isolation to exclude NSTI. Early evaluation must prioritise macroscopic and local clinical signs, and operative assessment remains decisive.

Additionally, we observed a statistically higher LRINEC score in the DF/NSTI cohort upon admission. As a laboratory diagnostic adjunct, the LRINEC score incorporates variables such as glucose and haemoglobin. In general clinical contexts, chronic metabolic derangements—such as baseline hyperglycemia or diabetes‐related anaemia—may inherently elevate this score independently of the acute infectious severity. However, because our study did not perform a component contribution analysis, this remains a general clinical hypothesis rather than a confirmed mechanism for our specific cohort.

4.4. Microbial Characteristics

Our study observed that P. aeruginosa was more frequently detected in the DF/NSTI group compared to the non‐DFU‐associated NSTI group (25.0% vs. 6.1%, p = 0.029) (Table 3). Given the small event counts and the nature of multiple organism‐level comparisons, this observation should be considered exploratory. While chronic wound exposure, prior antimicrobial therapy and frequent healthcare contact are plausible clinical explanations for this microbial distribution, these variables were not systematically measured in the present cohort. Consequently, we cannot draw definitive conclusions regarding the selective pressure of prior antibiotic use or the prevalence of multidrug‐resistant pathogens in these patients. Because our findings are exploratory and p values were unadjusted for multiple comparisons, these data cannot support any recommendations for altering current empirical antimicrobial strategies. Future prospective studies with standardised sampling protocols are required to clarify the exact microbial aetiology in this specific subpopulation [14].

4.5. Limitations and Future Directions

This study has several limitations. First, the small sample size of the DF/NSTI subgroup (n = 20) restricts the findings to exploratory descriptive comparisons, precluding multivariable adjustments. In addition, the limited sample size and event counts precluded a reliable assessment of whether diabetes duration or HbA1c independently influenced wound healing, amputation or mortality. Second, the retrospective design introduces selection bias. By excluding patients deemed inoperable or those whose families withdrew care, our cohort assembly inherently leads to an underestimation of the true overall mortality and the incidence of rapidly fatal outcomes. Third, the 11‐year study period may have introduced calendar‐time effects related to evolving antimicrobial strategies, surgical techniques and adjunctive treatments. Additionally, hospitalisation costs were not adjusted for inflation.

Several clinical and diagnostic confounders also exist. Fourth, an anatomical opportunity bias influences the amputation endpoints. Because DFU‐associated infections originate distally, the observed higher frequency of minor amputations primarily reflects localised surgical treatment patterns rather than a direct causal increase in systemic amputation risk. Fifth, we did not systematically report the baseline prevalence of chronic osteomyelitis; consequently, the difference in PMMA application is subject to confounding by indication. Sixth, standardised peripheral vascular assessments were not systematically performed prior to source control, leaving peripheral arterial status as an unadjusted confounder. Seventh, our microbiological profile is incomplete; because anaerobic cultures were not systematically performed, our data strictly reflect aerobic and facultative anaerobic pathogens. Finally, follow‐up endpoints focused primarily on short‐term perioperative outcomes. The absence of structured functional assessments restricts the evaluation of long‐term rehabilitation. Future multicentre prospective studies are warranted to evaluate the full clinical trajectory of these interventions.

5. Conclusion

In this single‐centre retrospective cohort of surgically treated patients, diabetic foot ulcer‐associated lower‐extremity necrotising soft‐tissue infection was associated with a greater comorbidity burden, greater healthcare resource use, more frequent minor amputation and a lower proportion of patients alive with a healed wound at 90 days than infection arising from other causes. Estimates for major amputation and mortality were imprecise and should not be interpreted as demonstrating equivalence between groups. Because treatment selection was clinically driven, this study cannot determine whether minor amputation reduces mortality or prevents major amputation. Larger prospectively characterised cohorts are needed to confirm these findings.

Funding

This work was supported by the National Natural Science Foundation of China (No. 81700695). Yantai City Science and Technology Innovation Development Plan (2024JCYJ053). Yantai Natural Science Foundation Biomedical Innovation Joint Funding (No. ZR2025LZ047).

Ethics Statement

This retrospective study was approved by the Institutional Review Board (IRB) of Yantai Yuhuangding Hospital (Approval No. K2026‐238).

Consent

Written informed consent for surgery was obtained from all patients or their legal surrogates.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.