Predicting limb amputation in trauma-related necrotizing fasciitis of the extremities: insights from a 10-year cohort Authors: Shuo Yang, Zhihao Ren, Yiran Li, Zihang Zhao, Changsheng Sun, Lin Liu, Lin Jin, Zhiyong Hou Published: 2026 ABSTRACT AbstractBackgroundNecrotizing fasciitis (NF) is a rapidly progressive, life-threatening soft tissue infection that in most cases requires aggressive surgical management. Major limb amputation is a devastating limb-related outcome. We retrospectively analyzed NF cases to quantify the rate of amputation and identify associated risk factors.MethodsWe retrospectively reviewed the medical records of all patients with trauma-related extremity NF treated at Hebei Medical University Third Hospital from Jan 2014 to Dec 2024. Demographics, comorbidities, admission laboratories and key time intervals were compared between patients who did and did not undergo major limb amputation. Multivariable logistic regression identified independent predictors, and receiver operating characteristic (ROC) analysis assessed the discriminative power of continuous variables.ResultsOf 134 patients, 22 (16.4%) required amputation. Compared with the non-amputation group, amputees had a markedly longer pre-hospital delay (median 242 h vs. 65 h), more multiple injuries (63.6% vs. 13.4%) and a higher prevalence of diabetes (59.1% vs. 21.4%). Admission lactate dehydrogenase (LDH) was substantially higher (median 589 vs. 182 U/L). In multivariable analysis, elevated LDH, prolonged time from traumatic injury to hospital admission, multiple injuries and diabetes were independently associated with amputation. LDH displayed excellent discrimination for amputation (AUC = 0.91), while pre-hospital delay showed good accuracy (AUC = 0.84).ConclusionsRoughly one in six patients with trauma-related extremity NF required limb amputation in this single-center cohort. High LDH on admission, delayed presentation, multiple injury sites and diabetes reliably identify the patients at greatest risk. Prompt recognition and aggressive management of these high-risk individuals are crucial for improving limb-salvage outcomes.Clinical trial numberNot applicable.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00068-026-03220-6.Keywords: Necrotizing fasciitis, Amputation, Lactate dehydrogenase, Diabetes mellitus, Time from traumatic injury to hospital admission Introduction Necrotizing fasciitis (NF) is a rapidly progressive, life-threatening infection of the fascia and subcutaneous tissues, often leading to widespread necrosis of the soft tissues. Although rare, NF is associated with high morbidity and mortality, with case fatality rates of 20–30% reported in multiple studies [1–2]. Survivors often require extensive surgical debridement, and in severe cases, limb amputation to halt infection spread [2]. The disease is usually driven by toxin-producing bacteria—most commonly Streptococcus pyogenes, Klebsiella pneumoniae, or Staphylococcus aureus—whose enzymes (e.g., hyaluronidase, collagenase) promote fascial destruction, thrombosis, and micro-vascular occlusion. Resultant tissue ischemia and systemic inflammatory response syndrome (SIRS) can progress to multiorgan failure [3]. In severe cases, rapid infection spread threatens vital organs, often necessitating amputation to prevent systemic collapse and preserve life. Amputation, while sometimes life-saving, results in permanent disability. It profoundly reduces mobility, quality of life, and psychological well-being, and imposes long-term rehabilitation costs [4]. Identifying clinical factors that predict the need for amputation is therefore critical for timely counselling and management. Prior studies have characterized risk factors for poor outcomes in NF, primarily focusing on mortality predictors such as advanced age, streptococcal toxic shock syndrome, and delays in treatment [3, 5–6]. However, specific predictors of limb loss (as opposed to death) have been less extensively studied. It is known that certain patient factors and clinical features may predispose to more severe presentations of NF. For example, diabetes mellitus is a common predisposing condition in NF, present in approximately 40–60% of cases [7–8]. Similarly, delayed treatment initiation has also been implicated in worse outcomes: NF is a surgical emergency, and a prolonged interval before adequate surgical debridement can allow unchecked infection spread, potentially necessitating more radical surgery including amputation [9]. A recent systematic review emphasised the paucity of data on limb-specific outcomes, particularly from Asian populations [1]. To date, only a few studies have specifically examined predictors of limb amputation in NF, and data are particularly limited in the context of the Chinese population. Therefore, we conducted a 10-year retrospective study at a tertiary referral center to determine the rate of major limb amputation in extremity NF and to identify independent clinical predictors of limb loss in this population. Methods Study design and setting We conducted a retrospective observational study of patients diagnosed with trauma-related NF at Hebei Medical University Third Hospital (Shijiazhuang, China), a large tertiary care center. The study period spanned 10 years, from January 2014 through December 2024. The hospital’s institutional review board approved the study protocol (2025-088-1), and patient consent was waived due to the retrospective nature. Patient selection All patients admitted with a diagnosis of trauma-related NF of the upper or lower extremities during the study period (January 2014 to December 2024) were eligible. All included patients underwent urgent surgical exploration and debridement, and NF was diagnosed based on the gold standard of intraoperative findings demonstrating grayish necrotic fascia and lack of bleeding in affected tissues, with histopathological confirmation (when available) of necrosis and bacterial invasion [10]. Clinical presentation (rapidly progressing soft tissue infection with signs such as severe pain, erythema, and systemic symptoms) and imaging (e.g., CT showing soft tissue gas or fascial thickening) were used to support the diagnosis when available. Patients were excluded if they (1) were < 18 years old; (2) had incomplete medical records; (3) were transferred after any prior surgical intervention for the index episode at an outside hospital (e.g., surgical exploration, debridement, fasciotomy, or amputation, including partial/limited debridement); (4) had non-NF soft tissue infections (e.g., cellulitis without fascial involvement). All included patients underwent their initial surgical exploration and definitive surgical management at our institution. Accordingly, 134 patients met the inclusion criteria, and complete data were available for all analyzed variables (Fig. 1). Flowchart of patient screening, exclusion, and inclusion in the final analysis Data collection Using a standardized data extraction form, we collected variables for each patient from the medical records. These included: (1) Demographics: age and sex; (2) Comorbidities: diabetes mellitus, hypertension, coronary heart disease, and history of smoking or alcohol use; (3) Injury and Presentation Characteristics: whether the infection was associated with an open fracture, presence of multiple injuries, time from injury occurrence to hospital admission (in hours), and whether the patient presented with septic shock before amputation (defined by hypotension requiring vasopressors and lactate > 2 mmol/L); (4) Treatment Timing: time from hospital admission to initial surgical intervention (hours); (5) Microbiology: whether a polymicrobial infection was identified (based on wound culture results); and (6) Laboratory Results on Admission: complete blood count (including white blood cell count, hemoglobin, hematocrit, platelet count), differentials (neutrophils, lymphocytes, eosinophils, etc.), coagulation profile (prothrombin time, INR, APTT, fibrinogen, D-dimer), serum biochemical markers (including albumin, liver enzymes alanine aminotransferase [ALT] and aspartate aminotransferase [AST], alkaline phosphatase [ALP], gamma-glutamyl transferase [GGT], globulin, glucose, creatinine, urea, electrolytes, and lactate dehydrogenase [LDH]), and inflammatory markers (high-sensitivity C-reactive protein [CRP]). These values were typically measured at admission. Outcome measures The primary outcome of interest was major limb amputation during the acute hospital stay. Major limb amputation was defined as amputation involving the affected upper or lower extremity at or above the wrist or ankle, excluding isolated digit amputations [11]. Amputations were classified as primary or secondary. Primary amputation was defined as major limb amputation performed at the initial operative intervention after admission, whereas secondary amputation was defined as major limb amputation performed after one or more prior debridement procedures during the same hospitalization. For patients who underwent major amputation, the timing of amputation, the number of debridements performed before amputation, and the documented indications for amputation were reviewed. The main indications for major amputation were uncontrolled infection despite repeated debridement and progressive irreversible tissue necrosis resulting in a non-salvageable limb. Patients were divided into two groups for analysis: those who underwent major limb amputation (amputation group) and those who did not (non-amputation group). Statistical analysis We first compared the major amputation and non-amputation groups using univariate analysis. For continuous variables, normality was assessed with the Shapiro–Wilk test. Variables with a normal distribution are presented as mean ± standard deviation and were compared using Student’s t-test, whereas non-normally distributed variables are presented as median (interquartile range [IQR]) and were compared using the Mann–Whitney U test. Categorical variables are presented as counts (percentages) and were compared using the chi-square test or Fisher’s exact test, as appropriate. A two-sided p value < 0.05 was considered statistically significant. Variables with p < 0.05 in univariate analysis were considered candidates for multivariable analysis. Binary logistic regression was performed to identify factors independently associated with major amputation. Given the relatively small number of major amputation events (n = 22), the number of covariates included in the model was restricted to reduce the risk of overfitting. Because of collinearity and overlap among several clinically related candidate variables, a forward stepwise selection approach (likelihood ratio method) was used to construct the final multivariable model. Variables with p < 0.05 were entered and those with p > 0.10 were removed at each step. Septic shock was excluded from the multivariable model because it exhibited complete separation, with all patients in the major amputation group presenting with shock, rendering regression estimates unstable. Adjusted odds ratios (ORs), 95% confidence intervals (CIs), and p values were calculated for variables retained in the final model. Continuous predictors were entered into the logistic regression model in their original measurement units. Accordingly, the OR for LDH represents the change in odds associated with each 1 U/L increase, and the OR for injury-to-admission interval represents the change in odds associated with each 1-hour increase. For continuous predictors identified in the multivariable analysis, receiver operating characteristic (ROC) curve analysis was performed to evaluate discriminatory performance. Optimal cutoff values were determined using the Youden index, and the corresponding sensitivity and specificity were calculated. All analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA). This study was reported in accordance with the STROBE guidelines. Results Patient characteristics and amputation rate A total of 134 patients with NF were included, of whom 22 patients (16.4%) required a major limb amputation during treatment. The remaining 112 patients (83.6%) underwent debridement and other limb-salvage therapies without amputation. Among the 22 major amputations, 18 involved the lower extremity and 4 involved the upper extremity, with amputations performed at different limb levels according to the extent of infection and tissue necrosis. All major amputations were secondary rather than primary. Before amputation, 7 patients underwent 1 debridement, 9 underwent 2 debridements, 3 underwent 3 debridements, 1 underwent 4 debridements, and 2 underwent 5 debridements. The median number of debridements before amputation was 2 (range, 1–5). Major amputation was ultimately performed because of uncontrolled infection despite repeated debridement, with progressive irreversible tissue necrosis and loss of limb salvageability. The median age of the entire cohort was 52 years (IQR 40–66 years), and 79.9% were male. Patients who underwent amputation tended to be slightly younger (median 45.5 vs. 53.5 years in non-amputation group), but this age difference was not statistically significant (p = 0.127). The gender distribution was similar between groups (86.4% male in the amputation group vs. 77.7% male in non-amputation, p = 0.530). Baseline body mass index (BMI) was also comparable (median ~ 25 kg/m2 in both groups, p = 0.63). Overall in-hospital mortality was 9.0% (12/134). Ten deaths occurred in the non-amputation group and two after major amputation. Apart from these baseline comparisons, several specific risk factors differed markedly between groups. Clinical risk factors and comorbidities Several clinical factors differed markedly between the amputation and non-amputation groups (Table 1). Time from injury to hospital admission was significantly longer in patients who ultimately required amputation. The amputation group had a median pre-hospital delay of 241.9 h (approximately 10 days) compared to 64.9 h (~ 2.7 days) in the non-amputation group (p < 0.001). In fact, over half of the amputation patients had delays exceeding one week from injury to presentation, whereas most non-amputation patients presented within a few days of injury. Notably, 21 of 22 (95.5%) patients in the amputation group had an initial open fracture that preceded the NF, compared to 58 of 112 (51.8%) in the non-amputation group (p < 0.001). Similarly, having multiple injuries (for example, multiple trauma sites or extensive wound areas) was strongly associated with amputation: 63.6% of amputee patients had multiple injury sites, versus only 13.4% of non-amputation patients (p < 0.001). This suggests that a greater initial injury burden or more extensive tissue involvement predisposed patients to more severe infection outcomes. Besides, time from hospital admission to initial debridement was similar in both groups. Table 1Comparison of baseline characteristics between patients who did and did not undergo amputationCharacteristicsAmputation group (n = 22)Non-Amputation group (n = 112) p Age, years45.50 (37.25 ~ 61.00)53.50 (41.75 ~ 67.00)0.127Gender, n (%)0.529Male19 (86.4%)87 (77.7%)Female3 (13.6%)25 (22.3%)BMI, kg/m225.33 (22.66 ~ 25.87)25.36 (23.00 ~ 27.76)0.630Time from injury to admission, hours241.92 (110.88 ~ 745.92)64.86 (26.61 ~ 119.75)< 0.001*Open fracture, n (%)< 0.001*Yes21 (95.5%)58 (51.8%)No1 (4.5%)54 (48.2%)Multiple injuries, n (%)< 0.001*Yes14 (63.6%)15 (13.4%)No8 (36.4%)97 (86.6%)ASA classification, n (%)0.736I1 (4.5%)5 (4.5%)II8 (36.4%)55 (49.1%)III10 (45.5%)41 (36.6%)IV3 (13.6%)11 (9.8%)Smoking history, n (%)0.907Yes6 (27.3%)35 (31.2%)No16 (72.7%)77 (68.8%)Alcohol history, n (%)0.205Yes3 (13.6%)33 (29.5%)No19 (86.4%)79 (70.5%)Multiple bacterial infections, n (%)0.088Yes10 (50.0%)30 (27.8%)No12 (50.0%)82 (72.2%)Blisters, n (%)0.617Yes1 (4.5%)12 (10.7%)No21 (95.5%)100 (89.3%)Time from admission to surgery, hours3.00 (2.00 ~ 7.75)24.00 (3.00 ~ 78.00)0.05Hypertension, n (%)0.118Yes3 (13.6%)37 (33.0%)No19 (86.4%)75 (67.0%)Coronary heart disease, n (%)0.184Yes1 (4.5%)21 (18.8%)No21 (95.5%)91 (81.2%)Diabetes, n (%)< 0.001*Yes13 (59.1%)24 (21.4%)No9 (40.9%)88 (78.6%)BMI = body mass index; ASA = American Society of Anesthesiologists; values are presented as the number (%) or the median (interquartile range). *p < 0.05 Comparison of baseline characteristics between patients who did and did not undergo amputation BMI = body mass index; ASA = American Society of Anesthesiologists; values are presented as the number (%) or the median (interquartile range). *p < 0.05 Regarding comorbid conditions, diabetes mellitus showed a significant association with amputation. Thirteen of 22 amputee patients (59.1%) were diabetic, compared to 24 of 112 (21.4%) in the non-amputation group (p < 0.001). In contrast, other comorbidities like hypertension (13.6% vs. 33.0%, p = 0.118) and coronary heart disease (4.5% vs. 18.8%, p = 0.18) were not significantly different between groups. A history of smoking or alcohol use similarly did not differ significantly. Septic shock on admission was present in all 22 patients (100%) in the major amputation group and in 41 of 112 patients (36.6%) in the non-amputation group. Laboratory findings On admission, several laboratory values differed significantly between groups (Supplementary Table S1). Lactate dehydrogenase (LDH) levels were elevated in patients who underwent amputation. The median LDH in the amputation group was 589.1 U/L (IQR 424.7–887.3), compared to 182.3 U/L (IQR 145.8–262.7) in the non-amputation group (p < 0.001). Inflammatory and hematologic parameters provided additional insight. Inflammatory markers (e.g., WBC, CRP) were elevated in both groups but did not significantly differ between amputees and non-amputees. However, eosinophil counts were significantly lower in the amputation group (p = 0.005), consistent with the phenomenon of stress eosinopenia in severe infection [12]. Notably, the platelet count (PLT) was much lower in patients requiring amputation: median platelet count was 133 × 10^9/L versus 242 × 10^9/L in non-amputation patients (p = 0.001). Coagulation times (PT, APTT, INR) were similar between groups on admission, although a trend toward prolonged thrombin time (TT) in the amputation group was observed (p = 0.051). Other lab differences included lower serum globulin levels in amputees (p < 0.001), potentially indicating protein consumption or dilution in acute shock, and a non-significant trend toward higher serum creatinine and urea (renal function markers) in the amputation group, possibly due to worse dehydration or shock (p = 0.650). Serum electrolytes did not significantly differ on admission, aside from a trend to slightly lower sodium in amputees. Importantly, polymicrobial infections (defined by culture of multiple organisms) were somewhat more common in the amputation group (50% vs. 27.8%), but this difference did not reach statistical significance (p = 0.088). Multivariate analysis – independent predictors of amputation In the multivariable logistic regression analysis, elevated LDH on admission (OR = 1.0074 per U/L, 95% CI 1.0032–1.0116), a longer injury-to-admission interval (OR = 1.0092 per hour, 95% CI 1.0014–1.0171), multiple injuries (OR = 8.43, 95% CI 1.89–37.56), and diabetes mellitus (OR = 12.32, 95% CI 1.74–87.37) remained independently associated with major amputation (Table 2). In contrast, PLT and AST were significant in univariate analysis but did not remain significant after multivariable adjustment. Table 2Binary logistic regression analysis of variables associated with amputationCharacteristicsOR95%CI p LDH1.00741.0032–1.0116< 0.001Time from injury to admission1.00921.0014–1.01710.021Multiple injuries8.431.89–37.560.005Diabetes12.321.74–87.370.012LDH = lactate dehydrogenase. For continuous variables, odds ratios are expressed per 1 U/L increase in LDH and per 1-hour increase in injury-to-admission interval. p < 0.05 indicates statistical significance Binary logistic regression analysis of variables associated with amputation LDH = lactate dehydrogenase. For continuous variables, odds ratios are expressed per 1 U/L increase in LDH and per 1-hour increase in injury-to-admission interval. p < 0.05 indicates statistical significance ROC curve analysis We further evaluated the two continuous variables from our model – LDH and time to admission – with ROC curves to assess their predictive performance for amputation (Fig. 2). ROC analysis showed excellent discrimination for LDH (AUC = 0.912, 95% CI = 0.851–0.963) and good discrimination for injury-to-admission time (AUC = 0.843, 95% CI = 0.752–0.927); an LDH cut-off of ≈ 290 U/L yielded 86.4% sensitivity and 83.9% specificity, whereas a pre-hospital delay of ≈ 8 days (193 h) provided 68.2% sensitivity and 87.5% specificity for predicting amputation (Fig. 2). ROC curves for LDH and time-to-admission as predictors of amputation Discussion In this 10-year retrospective study of trauma-related NF, we found that approximately one in six patients (16.4%) required major limb amputation as part of their management. This amputation rate is within the range reported in the international literature, where limb loss in NF patients varies but has been documented between about 10% and 30% [13–14]. Our analysis identified four key risk factors that were independently associated with amputation: elevated LDH levels on admission, delayed presentation (long time from injury to hospital admission), presence of multiple injuries, and diabetes mellitus. The ROC analysis further quantified their clinical utility: an LDH threshold of ≈ 290 U/L predicted amputation with an AUC of 0.91 (sensitivity 86.4%, specificity 83.9%), whereas a pre-hospital delay of ≈ 8 days (193 h) yielded an AUC of 0.84 (sensitivity 68.2%, specificity 87.5%). Together, these objective cut-offs provide actionable bedside markers that, alongside clinical context (polytrauma and diabetes), identify patients at greatest risk for eventual limb loss. Recognizing these risk factors early can alert clinicians to patients who may require especially aggressive intervention. In extreme cases, this may even include consideration of early amputation to control infection and save the patient’s life. Our overall in-hospital mortality was 9.0% (12/134), which is at the lower end of the 10–30% range reported in broader NF cohorts [9, 15–16]. This may be partly explained by two factors: first, the present study included only extremity NF, whereas truncal or perineal infections are generally associated with higher mortality; second, the sample size was relatively modest (22 major amputations and 112 limb-salvage cases), such that a small number of deaths could substantially affect the observed mortality rate. Although more in-hospital deaths occurred in the non-amputation group than in the major amputation group (10 vs. 2), death and major amputation represent two distinct severe outcomes and should not be viewed as directly interchangeable. Major amputation was performed for limb-related indications, whereas some patients in the non-amputation group may have died from overwhelming sepsis or multiorgan failure despite treatment. Given the retrospective design, the clinical decision-making process in individual fatal cases could not be fully reconstructed. Delayed presentation and importance of early intervention In our series, a median delay of 242 h in the amputation group versus 65 h in the non-amputation group (p < 0.001) remained independently predictive (OR 1.009 per hour). ROC analysis showed that a delay > ≈ 8 days (193 h) predicted amputation well (AUC = 0.84; 68% sensitivity, 88% specificity), underscoring how even a few extra days can convert a salvageable limb into one destined for loss. This aligns with the principle that NF outcomes are highly time-sensitive – the infection can advance rapidly, and any hold-up in treatment allows for greater tissue destruction [17–18]. Prior studies have emphasized that early surgical debridement is crucial to reducing mortality in NF. For instance, Kalaivani et al. reported that initiating surgery more than 24 h after hospital admission was associated with a dramatic increase in mortality (from 6% to 25%) [19]. While our study focused on limb loss rather than mortality, the same logic applies: a delay before adequate treatment correlates with more extensive infection, thereby increasing the likelihood that the affected limb cannot be salvaged. In many cases, patients with very long delays (several days to weeks) presented with overwhelming infection and septic shock, such that amputation became necessary as a life-saving measure. Our findings reinforce recommendations that any patient with suspected NF should be transferred and treated at a surgical center without delay, and that even a few days’ lag can make the difference between limb salvage and amputation. Public and professional awareness of NF, especially in at-risk populations (e.g. diabetics), should be improved to encourage earlier hospital presentation. Additionally, rapid diagnosis and surgical intervention upon hospital arrival remain paramount; although in our series in-hospital time to surgery was not a delaying factor for amputees, in general, prompt debridement is known to reduce the need for more extensive surgery later [17–18]. In patients with delayed presentation and other high-risk features, these findings also support consideration of earlier transfer to specialized centers and closer postoperative surveillance, including a lower threshold for planned re-exploration. Role of diabetes mellitus Diabetes mellitus was independently associated with major amputation in our cohort, with an adjusted OR of 12.32. Diabetic patients accounted for 59.1% of the major amputation group, compared with 21.4% of the non-amputation group. This finding is broadly consistent with previous reports identifying diabetes as one of the most important comorbidities associated with necrotizing fasciitis and with worse limb-related outcomes [20–21]. Several mechanisms may plausibly contribute to this association. Chronic hyperglycemia is known to impair neutrophil function and cellular immunity, while diabetic microvascular and macrovascular disease may compromise tissue perfusion and reduce delivery of antibiotics and immune cells to infected tissues [22–24]. In addition, diabetic neuropathy may delay recognition of minor trauma or early soft-tissue infection, allowing disease progression before presentation. Together, these factors may increase the likelihood of more extensive infection and reduce the chance of limb salvage once NF has developed. However, this finding should be interpreted with caution. In the present study, diabetes was recorded only as a binary variable, and more detailed diabetes-related information was not consistently available in the retrospective records. We were unable to reliably distinguish Type 1 from Type 2 diabetes, assess duration of disease, evaluate chronic glycemic control using HbA1c, document antidiabetic treatment, or determine the presence of diabetes-related complications such as peripheral vascular disease or diabetic foot disease. Therefore, although diabetes was independently associated with major amputation in this cohort, the magnitude and clinical meaning of this association may have been influenced by residual heterogeneity within the diabetes group and by unmeasured confounding. Our results support diabetes as an important clinical risk marker, but they do not allow more specific conclusions regarding the relative contribution of diabetes subtype, severity, or control status. Accordingly, diabetic status should be regarded as a clinically relevant marker of increased amputation risk in extremity NF, while more granular prospective data are needed to clarify how diabetes control and diabetes-related complications modify this risk. Lactate dehydrogenase as a marker of severity A notable finding of our study was the association between admission LDH levels and major amputation. LDH is released during tissue hypoxia, cellular injury, and necrosis, and elevated levels may therefore reflect extensive tissue damage in necrotizing soft tissue infection [25]. In our cohort, median LDH was 589 U/L in the major amputation group versus 182 U/L in the non-amputation group, and in multivariable analysis each 1-U/L increase in LDH was associated with an adjusted OR of 1.0074 (95% CI 1.0032–1.0116, p < 0.001). ROC analysis also showed that LDH had good discriminatory performance in this cohort (AUC = 0.91). This observation is consistent with previous reports suggesting that LDH tends to be elevated in more severe necrotizing infections [26–29]. For example, Wu et al. found that wound-fluid LDH concentrations were higher in patients with NSTI than in those with severe cellulitis, and Park et al. reported that non-survivors with NSTI had higher serum LDH levels than survivors [27–28]. These findings support the biological plausibility of LDH as a marker of disease severity. However, the predictive value of LDH should be interpreted cautiously. Our study was based on a single-center retrospective cohort, and neither the observed effect size nor the proposed cutoff value was externally validated. Therefore, although LDH appears to be a promising marker for risk stratification in trauma-related extremity NF, its clinical utility and generalizability remain uncertain. In particular, the cutoff identified in this cohort should be regarded as exploratory rather than definitive. Further prospective and multicenter studies are needed to validate LDH as a prognostic marker before it can be incorporated into routine clinical decision-making. Multiple injuries and trauma as risk factors In our cohort, patients with multiple injury sites had an OR of 8.43 for amputation, reinforcing that a more extensive initial insult greatly increases limb loss risk. Of the 29 cases classified as “multiple injuries,” 21 involved polytrauma (e.g., major crush injuries with several wound sites) and 8 represented multifocal infection (concurrent NF in more than one region). This variable likely captures the concept of extensive initial tissue insult. Extensive areas of devitalised or heavily contaminated tissue (e.g., crush wounds packed with dirt) provide an ideal anaerobic milieu for synergistic, polymicrobial (type I) necrotising fasciitis, which is typically more aggressive and difficult to control [30]. Each additional injury site could act as a separate focus of infection, compounding the overall burden. There is limited literature directly addressing “multiple injuries” as a risk factor for NF outcomes, but related evidence exists: for instance, combat-related blast injuries with large areas of tissue damage have been associated with necrotizing soft tissue infections often requiring amputations [31–32]. In civilian contexts, patients with extensive skin and muscle injuries (such as those in motor vehicle accidents or with large wounds) are at risk for NF, and such NF may be more severe simply due to the volume of tissue involved. Our data support the intuitive idea that NF affecting a larger territory (or arising from a high-energy mechanism causing multiple wounds) is harder to manage without limb sacrifice. Clinicians should be particularly vigilant in patients who develop NF in the setting of major trauma or multiple wounds – these patients may need early extensive debridement. While that decision is always difficult, recognizing the high risk in this subgroup can guide surgical aggressiveness. Additional observations In addition to the variables retained in the multivariable model, septic shock deserves particular attention. All patients who ultimately underwent major amputation presented with septic shock at admission. Although septic shock could not be included in the multivariable model because of complete separation, this finding is clinically highly relevant and suggests that septic shock may be a major determinant of poor limb outcome in this cohort, likely reflecting overwhelming systemic involvement, advanced infection, and markedly reduced potential for limb salvage. This observation is consistent with previous reports showing that NF patients who present with shock tend to have poorer outcomes and higher risks of severe complications [2–4]. Thus, in trauma-related extremity NF, septic shock at presentation should be regarded as an especially alarming sign of poor limb prognosis. We also observed significantly lower platelet counts in amputees, which is consistent with disseminated intravascular coagulation or platelet consumption in severe sepsis. Thrombocytopenia has been identified as an indicator of poor prognosis in critical illnesses generally, and in NF it may signal that the infection is overwhelming the host’s coagulative balance. These factors (shock, coagulopathy) reinforce that the amputation group was the subset of NF patients with the most severe systemic involvement. Conversely, traditional risk factors for poor outcomes of NF like older age or renal impairment did not show a clear difference in our limb loss analysis. Clinical implications From a practical standpoint, our findings may help clinicians stratify patients with suspected extremity NF at presentation and prioritize early senior surgical involvement. For example, an NF patient with diabetes who presents after a prolonged delay (e.g., 2 weeks) in shock and with a markedly elevated LDH (e.g., 1000 U/L) would represent a very high-risk phenotype in our cohort, suggesting a substantially increased likelihood of eventual amputation. In such scenarios, surgical teams may consider earlier, more decisive source control, closer monitoring, earlier escalation of care including ICU involvement when clinically indicated, and timely referral to centers with expertise in complex limb-threatening infection when such resources are not immediately available. These patients may also warrant a lower threshold for a structured second-look strategy, including planned re-exploration or repeat debridement over a short interval rather than a purely reactive approach. In addition, our findings may help clinicians identify patients in whom limb salvage is unlikely, thereby facilitating earlier multidisciplinary discussion and more realistic counselling of patients and families, while recognizing that amputation remains a last resort when limb-sparing attempts fail or when the infection is anatomically beyond salvage and limb viability is compromised. Conversely, a non-diabetic patient who presents promptly (within 1–2 days of injury) with only moderate LDH elevation may be more likely to be managed successfully with limb-sparing debridement. Our study supports aggressive initial management—broad-spectrum antibiotics and urgent surgical exploration with thorough debridement of all necrotic tissue—particularly in patients with clinical and laboratory features suggestive of severe disease. Multidisciplinary care is often required, including critical care for shock and, in diabetic patients, strict glycemic control and consideration of vascular assessment or revascularization when peripheral arterial disease is suspected. Importantly, because the LDH threshold and risk estimates were derived from a single-center retrospective cohort, they should be viewed as hypothesis-generating and used as an adjunct to—not a substitute for—clinical judgment; external validation in independent, ideally multicenter, cohorts is needed before these cutoffs can be generalized across different populations, care pathways, and laboratory assay platforms. Nevertheless, clinicians should recognize early when infection is not responding to limb-sparing surgery or is beyond salvage; in such cases, timely amputation can be life-saving and should not be unduly delayed, as prolonged hesitation may permit systemic deterioration. Limitations We acknowledge several limitations of this study. First, as a single-center retrospective study, our sample size (especially the amputation subgroup of 22 patients) is relatively small, which may limit generalizability. Some of our estimates (e.g., the OR for diabetes with a wide CI) have inherent uncertainty due to sample size. Larger multi-center studies or pooled analyses would help confirm these risk factors. Second, there may have been selection bias in terms of which patients were referred to our tertiary center – possibly more severe cases (with higher amputation likelihood) were transferred to us, influencing the amputation rate. Third, due to the retrospective design, we relied on the accuracy of medical records for timing variables and laboratory values. The “time from traumatic injury to hospital admission” can be subject to patient recall error. Fourth, we did not delve deeply into microbiological data in this report; pathogen virulence (e.g., group A Streptococcus producing superantigens) can also influence outcomes. It would be useful in future analyses to see if certain organisms are more often associated with amputations (perhaps polymicrobial synergistic infections in diabetics vs. monomicrobial streptococcal infections). Fifth, several clinically relevant covariates could not be captured in sufficient detail. In particular, diabetes was recorded only as a binary variable (yes/no); we were unable to reliably distinguish Type 1 from Type 2 diabetes, assess diabetes duration, chronic glycemic control (e.g., HbA1c), antidiabetic treatment, or diabetes-related complications such as peripheral vascular disease and diabetic foot disease. Therefore, residual heterogeneity and unmeasured confounding within the diabetes group cannot be excluded, and the observed association between diabetes and major amputation should be interpreted with caution. Sixth, our definition of “multiple injuries” was somewhat broad and based on documentation; a more granular assessment of anatomical infection extent (e.g., percentage of limb involvement, compartment involvement) could better quantify how infection extent correlates with amputation risk. In addition, microbiological data were not available in sufficient depth or consistency for a detailed analysis of pathogen distribution, virulence characteristics, or the association between specific organisms and major amputation. Therefore, the potential contribution of microbiological factors to limb loss could not be reliably evaluated in this retrospective cohort. Finally, because only 22 amputation events were available, we applied a forward stepwise logistic model to maintain parsimony. Although data-driven selection can be sensitive to sampling variation and may slightly inflate effect estimates, we mitigated this by: (i) pre-screening to keep the candidate pool small; (ii) restricting the model to four covariates, consistent with the event count; and (iii) confirming that the same predictors remained significant when all candidates were forced into the model. The resulting coefficients are therefore reasonable but should still be validated in larger, external cohorts. Strengths Despite these limitations, our study has notable strengths. It is one of the few to concentrate specifically on limb amputation in NF, an outcome of considerable importance to survivors. We had a comprehensive dataset with detailed lab measurements, allowing analysis of novel predictors like LDH. The consistency of our findings with pathophysiological reasoning (e.g., delayed treatment and diabetes making infections worse) lends credibility to the results. Also, the 10-year span increases the relevance across potential changes in practice; during this period, we improved our hospital’s NF protocols (such as faster surgical response times and use of negative pressure wound therapy), yet the identified risk factors remained prominent determinants of outcome. The use of objective cut-off analyses (ROC) for LDH and time-to-admission provides clinically usable thresholds that can be tested prospectively. Conclusion In this retrospective cohort of patients with trauma-related extremity NF, major limb amputation occurred in 16.4% of cases. Elevated LDH on admission, a longer injury-to-admission interval, multiple injuries, and diabetes mellitus were independently associated with major amputation. These findings suggest that readily available clinical and laboratory factors may help identify patients at increased risk of limb loss. However, because these predictors were derived from a single-center retrospective cohort, prospective validation in larger and ideally multicenter studies is required before broader clinical application. Electronic Supplementary Material Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We appreciate the great help from Hebei Medical University Third Hospital. Abbreviations Necrotizing fasciitis Receiver operating characteristic Area under ROC curve Odds ratio Confidence interval Interquartile range Systemic inflammatory response syndrome Basophil Eosinophil Hematocrit Hemoglobin Lymphocyte Mean corpuscular hemoglobin Mean corpuscular hemoglobin concentration Monocyte Mean platelet volume Neutrophil Platelet Red blood cell White blood cell Activated partial thromboplastin time Antithrombin III D-dimer Fibrinogen International normalized ratio Prothrombin time Thrombin time Albumin Alkaline phosphatase Alanine aminotransferase Aspartate aminotransferase AST/ALT ratio Calcium Cholinesterase Chloride Creatinine Direct bilirubin Anion gap Gamma-glutamyl transferase Globulin Glucose High-sensitivity C-reactive protein Indirect bilirubin Potassium Lactate dehydrogenase Magnesium Sodium Osmolality Phosphate Total bile acids Total bilirubin Carbon dioxide Triglycerides Total protein Uric acid Urea Author contributions SY, ZHR, and YRL were responsible for study concept and writing the article. ZHZ, CSS, and LL were responsible for screening the abstracts and reviewing the article. LJ and ZYH were responsible for revising and finalizing the article. Funding The research was supported by the Key R&D Program of the China Ministry of Science and Technology (2024YFC2510600); Key Project of Natural Science Foundation of Hebei Province (Class A) (H2024206134), Study on Microenvironment Regulation and Repair Mechanism of Acute Compartment Syndrome; Project of Hebei Provincial Department of Finance: “Research Team on Mechanistic Biological Signal Transduction Pathways of Acute Compartment Syndrome”. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Code availability Not Applicable. Declarations Ethics approval and consent to participate This retrospective study was approved by the Institutional Review Board of Hebei Medical University Third Hospital (2025-088-1) before collecting data. Written informed consent was waived by the IRB due to the retrospective design of the study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Not Applicable.