Factors Influencing Major Amputation and Death Following Limb Salvage Surgery in a Diabetic Population: Systematic Review and Real-World Comparison

Abstract

Diabetic foot ulcers drive non-traumatic lower-limb amputation; limb salvage surgery is often pursued to preserve function and survival. Predictors of adverse outcomes remain incompletely defined, and evidence for multidisciplinary team (MDT) care is heterogeneous. We aimed to clarify risk factors for major amputation and death after diabetic limb salvage and evaluate MDT impact. We systematically reviewed 49 studies (2020–2025) reporting major amputation or mortality after limb salvage in diabetes (PROSPERO CRD420251044859). Risk factors spanned demographic, clinical, and surgical domains (e.g., older age, male sex, renal/cardiovascular disease, ischemia, osteomyelitis, advanced ulcer classification). MDT models generally showed lower amputation rates and improved wound healing, with occasional survival benefits; heterogeneity precluded meta-analysis. As a real-world comparator, the Mid Essex Diabetes Amputation Reduction Plan (MEDARP) treated 72 high-risk patients using a “toe and flow” MDT. Major amputation occurred in 6.9% and mortality in 12.5%, both at or below published ranges, with gains in patient-reported outcomes. Findings support MDT-based strategies, but conclusions should be interpreted cautiously given the predominantly observational evidence, and highlight the need for standardized outcome definitions and reporting.

1. Introduction

Diabetes mellitus is one of the most pressing global health challenges of the 21st century, with an estimated 537 million adults affected worldwide, a number projected to rise substantially in coming decades [1]. Among its most devastating complications is the development of diabetic foot ulcers (DFUs), which occur in up to one-quarter of individuals with diabetes during their lifetime [2,3]. DFUs are associated with infection, peripheral arterial disease, and neuropathy, and they remain the leading cause of non-traumatic lower-extremity amputations [4,5]. Major amputation, defined as below- or above-knee limb loss, carries profound consequences, including loss of independence, reduced quality of life, high healthcare costs, and a five-year mortality risk that rivals many cancers [2,6]. In diabetic patients, lower extremity arterial disease (LEAD) tends to manifest earlier and progress more rapidly to critical limb ischemia, reflecting the combined influence of systemic inflammation, endothelial dysfunction, macroangiopathy, and microangiopathy. This interplay contributes to accelerated vascular deterioration and poorer outcomes compared with non-diabetic populations [7,8].

Efforts to prevent amputation have led to the widespread adoption of limb salvage surgery, encompassing vascular and reconstructive procedures, surgical debridement, partial foot amputation, and multidisciplinary perioperative management. These approaches can preserve function and improve survival, but outcomes remain inconsistent across health systems and patient populations [2,6,9]. In some cases, initial limb salvage attempts culminate in delayed major amputation, a sequence often associated with worse morbidity and mortality than primary amputation [2,10]. Identifying factors that predispose patients to poor outcomes following salvage surgery is therefore essential for guiding surgical decision-making, optimizing patient selection, and tailoring perioperative care.

Although individual studies have examined risk factors such as age, renal disease, cardiovascular comorbidities, infection severity, wound characteristics, and socioeconomic determinants, the literature is fragmented and heterogeneous. Furthermore, while multidisciplinary team (MDT) models—including vascular, plastic, orthopedic, podiatric, and infectious disease specialists—have been promoted under frameworks such as the “toe and flow” concept, the strength of evidence supporting their impact on amputation-free survival and mortality remains unclear [11,12,13]. Recent observational studies suggest MDTs can reduce amputation rates and improve survival, but their implementation and reported outcomes vary widely [9].

To address these knowledge gaps, we conducted a study with two complementary components: a systematic review of the recent literature and a real-world comparison using data from the Mid Essex Diabetes Amputation Reduction Plan (MEDARP). The systematic review sought to: (Aim 1) identify the demographic, clinical, and surgical factors most consistently associated with major amputation and mortality following limb salvage surgery in patients with diabetes; and (Aim 2) examine how multidisciplinary care models for diabetic foot surgery have been structured in the recent literature and what evidence supports their impact on amputation and survival outcomes. The real-world comparison (Aim 3) evaluated whether these findings were mirrored in outcomes from a United Kingdom-based transformational program (i.e., MEDARP) implementing a multidisciplinary “toe and flow” model. Together, these three aims clarify risk factors, assess the influence of multidisciplinary care, and demonstrate how published evidence translates into practice.

2. Materials and Methods

2.1. Systematic Review—Search Strategies

A comprehensive and systematic literature search was conducted using the PubMed, Scopus, and CINAHL databases to identify studies evaluating major amputation and mortality outcomes in diabetic patients who underwent limb salvage surgery. The search strategy was developed to capture clinical, demographic, and socioeconomic risk factors associated with these outcomes. Searches were limited to articles published between 2020 and 2025, in English, and involving adult patients (≥18 years) with Type 1 or Type 2 Diabetes Mellitus. The 2020–2025 timeframe was deliberately chosen to capture the most up-to-date literature, reflecting contemporary surgical techniques, outcome reporting standards, and the rapid evolution of multidisciplinary team (MDT) models in diabetic limb salvage.

The PubMed search string was structured as follows: (“Diabetes Mellitus” [Title/Abstract] OR “Diabetic foot” [Title/Abstract] OR “Diabetic ulcer” [Title/Abstract] OR “Diabetic complications” [Title/Abstract] OR “Diabetic neuropathy” [Title/Abstract]) AND (“Limb salvage” [Title/Abstract] OR “Limb preservation” [Title/Abstract] OR “Foot surgery” [Title/Abstract] OR “Amputation prevention” [Title/Abstract] OR “Limb-sparing” [Title/Abstract] OR “Limb-saving” [Title/Abstract]) AND (“Amputation” [Title/Abstract] OR “Major amputation” [Title/Abstract] OR “Mortality” [Title/Abstract] OR “Survival” [Title/Abstract] OR “Risk factors” [Title/Abstract] OR “Predictors” [Title/Abstract] OR “Outcomes” [Title/Abstract] OR “Overall survival”) AND (“Clinical factors” [Title/Abstract] OR “Demographic factors” [Title/Abstract] OR “Surgical outcomes” [Title/Abstract] OR “Socioeconomic factors” [Title/Abstract] OR “Multidisciplinary care” [Title/Abstract] OR “Toe and Flow” [Title/Abstract] OR “HbA1c” [Title/Abstract] OR “Comorbidities” [Title/Abstract])

In PubMed, free-text Title/Abstract terms were used without MeSH restrictions, as this ensured capture of the most recent publications not yet indexed with MeSH. Equivalent free-text strategies were used in Scopus and CINAHL.

Following de-duplication, titles and abstracts were screened using a custom prompt developed for ChatGPT-4o (OpenAI, May 2024) to assist with applying the predefined eligibility criteria (Appendix A). Screening was conducted independently by three reviewers (K.F., S.M.A., and D.M.) who compared their screening results with that of ChatGPT. Discrepancies were reviewed and resolved by a fourth author (E.K.). The same AI-assisted process was used for full-text screening to determine final study inclusion.

The review process adhered to the PRISMA-ScR (Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews) checklist to ensure methodological transparency and rigor [14]. A PRISMA flow diagram (Figure 1) illustrates the identification, screening, and selection process. The review protocol was registered with PROSPERO (CRD1044859) [15].

Figure 1. PRISMA flow diagram depicting the process of literature selection and screening.

2.2. Systematic Review—Inclusion and Exclusion Criteria

Studies were included if they were peer-reviewed research articles, published in English between 2020 and 2025, and focused on adult patients (≥18 years) with Type 1 or Type 2 Diabetes Mellitus who had undergone limb salvage surgery for diabetic foot complications, such as ulceration, infection, or ischemia. Eligible studies reported at least one quantifiable clinical outcome, specifically major amputation (below- or above-knee) or mortality (perioperative or long-term). Studies were also included if they examined risk factors (clinical, surgical, demographic, or socioeconomic), amputation-free survival, wound-free days, postoperative quality of life, or the impact of multidisciplinary care models (e.g., toe and flow). Eligible study designs included randomized controlled trials, prospective and retrospective cohort studies, and case–control studies. Large case series involving 10 or more patients were also considered if outcome data were clearly reported and relevant to the study objectives. Studies including both Type 1 and Type 2 diabetes were eligible, as the vast majority of published cohorts (and our real-world dataset) comprised patients with Type 2 diabetes. Very few studies reported outcomes separately by diabetes type, precluding meaningful subgroup analyses.

Studies were excluded if they were systematic reviews, meta-analyses, editorials, narrative reviews, conference abstracts, or case reports/case series with fewer than 10 patients. Additional exclusion criteria included studies involving pediatric populations (<18 years), those focusing exclusively on primary amputation without discussion of limb salvage, and studies of non-diabetic or mixed populations where diabetes-specific outcomes were not reported separately. Research focused solely on short-term wound healing or involving experimental therapies not addressing major amputation or mortality risk was also excluded. Finally, studies conducted in highly specific or non-generalizable settings (e.g., military hospitals, rare genetic disorders) were not considered for inclusion.

2.3. Systematic Review—Data Extraction and Quality Assessment

Authors D.M., K.F., and S.M.A. independently performed data extraction and quality assessment for all included studies. Extracted variables included: first author, year of publication, study title, study design, population characteristics (e.g., age, diabetes type, comorbidities), details of limb salvage procedures, incidence of major amputation and mortality, reported risk factors, and outcomes such as amputation-free survival, wound-free days, and post-surgical quality of life. Each study was assigned a level of evidence according to the classification system from The Journal of Bone & Joint Surgery [16], which categorizes studies as follows:

• Level I: High-quality randomized controlled trials (RCTs) or systematic reviews/meta-analyses of RCTs.
• Level II: Lesser-quality RCTs or prospective cohort studies with comparison groups.
• Level III: Retrospective cohort studies or case–control studies.
• Level IV: Case series or retrospective studies without comparison groups.
• Level V: Expert opinion, narrative reviews, or basic science without clinical correlation.

To assess risk of bias, the National Heart, Lung, and Blood Institute (NHLBI) Study Quality Assessment Tools were applied, matched to the design of each study (e.g., cohort, case–control, or case series) [17]. Each study was rated across standardized criteria, including study population, comparability of groups, exposure and outcome assessment, statistical adjustment for confounding, and loss to follow-up. Based on the overall number and significance of methodological strengths and weaknesses, studies were categorized as follows:

• Good (low risk of bias): Most or all criteria met; minimal threats to internal validity.
• Fair (moderate risk of bias): Some criteria not met or unclear; potential for bias that could affect results.
• Poor (high risk of bias): Many criteria unmet or unclear; significant concerns regarding validity or reliability of findings.

Disagreements in data extraction or quality ratings were resolved through discussion among the reviewers until consensus was achieved.

2.4. Heterogeneity and Rationale for Narrative Synthesis

Considerable heterogeneity across the 49 included studies precluded formal meta-analysis. Patient populations varied from relatively young surgical cohorts undergoing elective reconstructions to elderly patients with chronic limb-threatening ischemia managed in tertiary referral centers. The scope of procedures was similarly diverse, encompassing partial foot amputations, vascular and endovascular interventions, complex flap reconstructions, and combined orthoplastic approaches.

Definitions of outcomes were inconsistent. For example, “major amputation” was variously defined as any amputation proximal to the ankle, below-knee only, or aggregated with higher-level amputations; mortality endpoints ranged from in-hospital events to survival at 30 days, 1 year, or 5 years. Reporting of complications was likewise variable, with some studies presenting composite outcomes (e.g., “surgical failure”) while others reported specific events such as wound dehiscence, osteomyelitis, or graft failure. Follow-up intervals differed widely, from short-term perioperative assessments to long-term survival extending beyond five years.

MDT models were also described with differing structures and emphases, ranging from consultative collaborations to co-located limb salvage services, with outcome measures that included limb salvage, amputation-free survival, wound healing, or cost reduction.

Given these clinical and methodological differences—spanning populations, procedures, outcome definitions, MDT structures, and follow-up timeframes—pooling effect estimates would have risked generating misleading summary statistics. A narrative synthesis was therefore undertaken to preserve the nuance of individual study findings and highlight recurring patterns across heterogeneous evidence. We nevertheless performed exploratory meta-analyses to quantify heterogeneity. Between-study heterogeneity was assessed using I² statistics, which consistently exceeded 80%, confirming considerable heterogeneity. These analyses, along with forest plots (Supplementary Figures S1 and S2), illustrate the wide variability in outcomes and informed our decision to rely on narrative synthesis.

2.5. Mid Essex Diabetes Amputation Reduction Plan—Description

A total of 72 unique surgical patients were managed through the MEDARP between December 2021 and March 2024. These patients accounted for 84 admissions and 328 procedures, reflecting the fact that some individuals required more than one admission during the study period. Where repeat admissions occurred, patients were counted once in the overall total, while each admission and procedure was included in the respective counts. Theater activity and its data were captured prospectively utilizing “dropped” vascular theater lists.

Patient-level variables of interest included gender, age, type of diabetes (Type 1 vs. Type 2), surgical stage of diabetic limb salvage, ulcer duration, prior surgery, and relevant blood markers.

The MEDARP at Broomfield Hospital was developed to improve diabetic foot ulcer management in line with National Institute for Health and Care Excellence (NICE) standards. The service model drew on the principles of the Manchester Amputation Reduction Strategy (MARS), which pioneered the “toe and flow” approach in the UK [11].

The Broomfield program established a dedicated MDT, comprising a full-time podiatrist and a part-time consultant podiatric surgeon, working in close collaboration with vascular surgery, plastic surgery, and other members of the existing diabetic foot service. The initiative extended beyond theater lists to include structured outpatient MDT clinics, inpatient ward rounds, multidisciplinary reviews, and on-call support to the Emergency Department (A&E). Over the course of the program, the team managed over 1000 patient contacts and provided targeted education to medical and nursing staff, reinforcing system-wide improvement in diabetic foot care delivery.

Importantly, MEDARP was commissioned through the NHS Diabetes Transformation Fund, a national program designed to expand specialist diabetes care in England. The project operated under strict criteria of time and financial constraints, requiring delivery of measurable improvements in outcomes within a defined period [18].

2.6. Mid Essex Diabetes Amputation Reduction Plan—Data Collection

Prospective service-evaluation data were systematically collected for all patients managed within MEDARP during the study period. Data capture was contemporaneous, using a secure local database, and included:

• Demographics: age, sex, type and duration of diabetes, frailty, comorbidities.
• Surgical details: stage of surgery, type of revision procedure, and use of adjunctive biomaterials.
• Clinical outcomes: major amputation, mortality, reintervention, and readmission.
• Patient-reported outcomes (PROMs): Musculoskeletal Health Questionnaire (MSK-HQ; 0–56, higher = better), Manchester-Oxford Foot Questionnaire (MOXFQ; 0–100, higher = worse) across all subscales, and self-reported physical activity frequency (days/week).

Data were entered prospectively at baseline, 6-month, and ~11-month follow-up visits. The structured use of PROMs allowed the team to capture functional recovery and quality of life alongside surgical and survival outcomes. In addition, at the time of manuscript preparation (June 2025), authors EK and MG retrospectively cross-checked patient records against the prospectively maintained database to verify accuracy and confirm the major study endpoints of major amputation and death. No new data were generated during this process.

2.7. Mid Essex Diabetes Amputation Reduction Plan—Data Analysis

Data were entered prospectively into a secure service-evaluation database and exported for analysis. Descriptive statistics were used to summarize patient demographics, comorbidities, surgical stage, and use of adjunctive materials. Continuous variables (e.g., age, MSK-HQ, MOXFQ subscales, physical activity days/week) were summarized as means with standard deviations. Categorical variables (e.g., sex, diabetes type, Fontaine stage, frailty category, revision surgery) were expressed as counts and percentages. Clinical outcomes of interest were the incidence of major amputation and all-cause mortality during follow-up. For patient-reported outcomes (MSK-HQ, MOXFQ subscales, physical activity), within-patient changes from baseline to 6-month and ~11-month follow-up were assessed using Wilcoxon signed-rank tests. Results are presented as means with standard deviations and corresponding p-values. Given the modest cohort size, these analyses were treated as exploratory and hypothesis-generating. Also, given the small number of outcome events (5 major amputations, 9 deaths), formal multivariate logistic regression was not feasible as it would risk model overfitting. Instead, exploratory univariate associations between baseline risk factors and outcomes were examined, with results interpreted descriptively.

3. Results

3.1. Study Selection and Characteristics

A total of 49 studies published between 2020 and 2025 met the eligibility criteria (Table 1). The majority were retrospective cohort studies (n = 28, 57%), followed by prospective cohort studies (n = 10, 20%), case–control studies (n = 5, 10%), case series (n = 5, 10%), and one randomized controlled trial (2%). Most originated from Europe (n = 18), Asia (n = 12), and North America (n = 11), with additional studies from the Middle East and Africa.

Table 1. Summary of Included Studies in the Systematic Review. Characteristics of the 49 studies included in the systematic review, ordered by level of evidence and then by year of publication. Information includes study identifiers, authorship, country, study design, sample size, level of evidence [16], quality assessment [17], and publication details.

Ref.	Authors/Year	Country	Level of Evidence	Quality Assessment	Study Type	Journal	Article Title
[19]	Davis et al. (2020)	United States	Level I	Fair	Randomized clinical trial	Wound Repair and Regeneration	Randomized clinical study to compare negative pressure wound therapy with simultaneous saline irrigation and traditional negative pressure wound therapy for complex foot infections.
[20]	Lava et al. (2024)	United States	Level II	Fair	Prospective cohort study	Plastic and Reconstructive Surgery—Global Open	A Comparative Analysis of Patient-Reported Outcomes Following Free Tissue Transfer, Partial Foot Amputation, and Below-Knee Amputation in High-Risk Limb Salvage Patients.
[21]	Lo et al. (2023)	Singapore	Level II	Fair	Prospective cohort study	International Wound Journal	Diabetic Foot in Primary and Tertiary (DEFINITE) Care: A Health Services Innovation in Coordination of Diabetic Foot Ulcer (DFU) Care within a Healthcare Cluster—18-Month Results from an Observational Population Health Cohort Study.
[22]	Aragón-Sánchez et al. (2023)	Spain and Costa Rica	Level II	Fair	Prospective cohort study	The International Journal of Lower Extremity Wounds	Clinical Features, Inflammatory Markers, and Limb Salvage in Older Adults with Diabetes-Related Foot Infections.
[23]	Weissler et al. (2021)	United States	Level II	Fair	Prospective cohort study	Vascular and Endovascular Surgery	Characteristics and Outcomes of Patients with Diabetes Mellitus Undergoing Peripheral Vascular Intervention for Infrainguinal Symptomatic Peripheral Artery Disease.
[24]	Azhar et al. (2021)	Egypt	Level II	Fair	Prospective cohort study	The International Journal of Lower Extremity Wounds	Prevalence of Peripheral Arterial Disease in Diabetic Foot Ulcer Patients and Its Impact in Limb Salvage.
[25]	Ge et al. (2024)	Singapore	Level II	Fair	Prospective cohort study	International Wound Journal	Multi-disciplinary diabetic limb salvage programme in octogenarians with diabetic foot ulcers is not futile: An observational study with historical controls.
[26]	Draper et al. (2025)	Canada	Level II	Good	Prospective case–control study	Diabetology	Identification of Inflammatory Biomarkers for Predicting Peripheral Arterial Disease Prognosis in Patients with Diabetes.
[27]	Rohrich et al. (2025)	United States	Level III	Fair	Retrospective cohort study	Annals of Plastic Surgery	Angiosome-Guided Revascularization in Local Flap Reconstruction of the Foot and Ankle: Comparable Outcomes With Both Direct and Indirect Revascularization.
[28]	Shin & Yoo (2025)	South Korea	Level III	Fair	Retrospective cohort study	Medicina	Predictive Factors of Wound Healing and Limb Salvage After Successful Below-the-Knee Endovascular Angioplasty in Patients with Diabetic Foot Ulcer: A Retrospective Study.
[29]	Almadwahi et al. (2024)	Yemen	Level III	Fair	Retrospective cohort study	Journal of Emergency Medicine, Trauma & Acute Care	Acute Limb Ischemia and Its Predictive Factors After Revascularization: A Single-Center Retrospective Study from a Resource-Limited Setting.
[30]	Biz et al. (2024)	Italy	Level III	Fair	Retrospective cohort study	Foot & Ankle International	Minimally Invasive Distal Metatarsal Diaphyseal Osteotomy (MIS-DMDO) for the Prevention and Treatment of Diabetic Forefoot Ulcers.
[31]	Ormaechevarria et al. (2024)	Spain	Level III	Fair	Retrospective cohort study	EJVES Vascular Forum	Critical Limb Ischaemia in Octogenarians: Treatment Outcomes Compared With Younger Patients.
[32]	Li et al. (2024)	United States	Level III	Fair	Retrospective cohort study	Journal of Clinical Medicine	A Multidisciplinary Approach to End-Stage Limb Salvage in the Highly Comorbid Atraumatic Population: An Observational Study.
[33]	Amarasena et al. (2024)	United Kingdom	Level III	Fair	Retrospective cohort	Practical Diabetes	Biochemical and patient factors as predictors for multiple surgeries in diabetic foot disease.
[34]	Huffman et al. (2024)	United States	Level III	Fair	Retrospective cohort study	Journal of Reconstructive Microsurgery	Muscle versus Fascia Free Tissue Transfer for Treatment of Chronic Osteomyelitis in the Comorbid Population.
[35]	Kostiuket al. (2024)	United States	Level III	Fair	Retrospective cohort study	Vascular	Long-term limb salvage and functional outcomes for patients undergoing partial calcanectomy.
[36]	Fujii et al. (2023)	Japan	Level III	Fair	Retrospective cohort study	The International Journal of Lower Extremity Wounds	Predictive Factors for Limb Salvage and Foot Ulcer Recurrence in Patients with Chronic Limb-Threatening Ischemia After Multidisciplinary Team Treatment: A 6-Year Japanese Single-Center Study.
[37]	Ron et al. (2023)	Israel	Level III	Good	Retrospective cohort study	Journal of Bone and Joint Surgery	Risk Factors for a Failed Transmetatarsal Amputation in Patients with Diabetes.
[38]	Liu et al. (2023)	China	Level III	Fair	Retrospective cohort study	Advances in Wound Care	Study on the Effect of the Five-in-One Comprehensive Limb Salvage Technologies of Treating Severe Diabetic Foot.
[39]	Ragghianti et al. (2023)	Italy	Level III	Fair	Retrospective cohort study	Journal of Wound Management	Effects of local antibiotics in calcium-sulphate granules for the treatment of diabetic forefoot osteomyelitis: A propensity-matched observational study.
[40]	Zamzam et al. (2023)	Canada	Level III	Fair	Retrospective cohort study	International Wound Journal	A novel Canadian multidisciplinary acute care pathway for people hospitalised with a diabetic foot ulcer.
[41]	Zhu et al. (2023)	China	Level III	Fair	Retrospective cohort study	Diabetology & Metabolic Syndrome	Effect of Peripheral Nerve Block versus General Anesthesia on the Hemodynamics and Prognosis of Diabetic Patients Undergoing Diabetic Foot Surgery.
[42]	Yammine et al. (2023)	Lebanon	Level III	Fair	Retrospective cohort study	International Wound Journal	Basic haematological tests as inflammatory performance markers of patients treated either by conservative surgery or minor amputation for infected diabetic foot ulcers.
[43]	Bobirca et al. (2022)	Romania	Level III	Fair	Retrospective cohort study	Life	The Outcome of Surgical Treatment for the Neuropathic Diabetic Foot Lesions—A Single-Center Study.
[44]	Abu El Hawa et al. (2022)	United States	Level III	Fair	Retrospective cohort study	Advances in Wound Care	Early Diagnosis and Surgical Management of Necrotizing Fasciitis of the Lower Extremities: Risk Factors for Mortality and Amputation.
[45]	Huang et al. (2022)	China	Level III	Fair	Retrospective cohort study	International Wound Journal	Association of Time in Range with Postoperative Wound Healing in Patients with Diabetic Foot Ulcers.
[46]	Kim et al. (2022)	United States	Level III	Fair	Retrospective cohort study	Open Forum Infectious Diseases	Association Between Foot Surgery Type and Subsequent Healing in Veterans With Moderate-to-Severe Diabetic Foot Infections.
[47]	Nigam et al. (2022)	United States	Level III	Fair	Retrospective cohort study	Plastic and Reconstructive Surgery	Expanding Criteria for Limb Salvage in Comorbid Patients with Nonhealing Wounds: The MedStar Georgetown Protocol and Lessons Learned after 200 Lower Extremity Free Flaps.
[48]	Schmidt et al. (2022)	United States	Level III	Fair	Retrospective cohort study	The Diabetic Foot Journal	Bedside versus operating room debridement of osteomyelitis of a phalanx of a given toe.
[49]	Haug et al. (2021)	United States and Germany	Level III	Good	Retrospective cohort study	Journal of Plastic, Reconstructive & Aesthetic Surgery	Combined (endo-)vascular intervention and microsurgical lower extremity free flap reconstruction: A propensity score matching analysis in 5386 ACS-NSQIP patients.
[50]	Thai et al. (2021)	South Korea	Level III	Fair	Retrospective cohort study	Journal of Orthopaedic Surgery and Research	Factors Affecting the Outcome of Lower Extremity Osteomyelitis Treated with Microvascular Free Flaps: An Analysis of 65 Patients.
[51]	Koivunen et al. (2021)	Finland	Level III	Fair	Retrospective cohort study	Scandinavian Journal of Surgery	Chronic Limb Threatening Ischemia and Diabetes Mellitus: The Severity of Tibial Atherosclerosis and Outcome After Infrapopliteal Revascularization.
[52]	Burmeister et al. (2021)	United States	Level III	Fair	Retrospective cohort study	International Wound Journal	Postoperative healing in the diabetic foot is impacted by discharge destination.
[53]	Kim et al. (2021)	South Korea	Level III	Fair	Retrospective cohort study	BioMed Research International	Reconstruction of Foot and Ankle Defects Using Free Lateral Arm Flap: A Retrospective Review of Its Versatile Application.
[54]	Piaggesi et al. (2020)	Italy	Level III	Fair	Retrospective cohort study	Diabetes Research and Clinical Practice	Diabetic Foot Surgery “Made in Italy”: Results of 15 Years of Activity of a Third-Level Centre Managed by Diabetologists.
[55]	Joyce et al. (2020)	United Kingdom	Level III	Fair	Retrospective cohort study	The Foot	Transmetatarsal Amputation: A 12-Year Retrospective Case Review of Outcomes.
[56]	Meloni et al. (2020)	Italy	Level III	Fair	Retrospective cohort study	Journal of Clinical Medicine	Prevalence, Clinical Aspects and Outcomes in a Large Cohort of Persons with Diabetic Foot Disease: Comparison between Neuropathic and Ischemic Ulcers.
[57]	Piwnica-Worms et al. (2020)	United States	Level III	Fair	Retrospective cohort study	Journal of Reconstructive Microsurgery	Risk Factors for Lower Extremity Amputation Following Attempted Free Flap Limb Salvage.
[58]	Lo et al. (2022)	Singapore	Level III	Fair	Case–control study	International Wound Journal	Clinical and economic outcomes of a multidisciplinary team approach in a lower extremity amputation prevention programme for diabetic foot ulcer care in an Asian population: A case-control study.
[59]	Seçkin et al. (2022)	Turkey	Level III	Fair	Case–control study	The Journal of Foot & Ankle Surgery	Predictive Factors and Amputation Level for Reamputation in Patients With Diabetic Foot.
[60]	Brodell et al. (2020)	United States	Level III	Fair	Case–control study	Foot & Ankle International	Intraoperative Site Vancomycin Powder Application in Infected Diabetic Heel Ulcers With Calcaneal Osteomyelitis.
[61]	Alsabbagh et al. (2024).	United States	Level IV	Fair	Retrospective case series	Annals of Vascular Surgery	Deep venous arterialization in critical limb-threatening ischemia (CLTI): Case series and literature review.
[62]	Brekelmans et al. (2023)	Netherlands	Level IV	Fair	Retrospective case series	International Wound Journal	Recurrent diabetic foot ulcers: Results of a maximal multidisciplinary approach including reconstructive foot/ankle surgery.
[63]	Altuntaş et al. (2023)	Turkey	Level IV	Fair	Retrospective case series	The International Journal of Lower Extremity Wounds	A Simple and Convenient Alternative for the Reconstruction of Lower Extremity Soft Tissue Defects due to Different Types of Etiologies: Bipedicled Flap.
[64]	Vasukutty et al. (2022)	United Kingdom	Level IV	Fair	Retrospective case series	The Diabetic Foot Journal	Limb salvage surgery in diabetic foot infection: encouraging early results with a local antibiotic carrier.
[65]	Frykberg et al. (2021)	United States	Level IV	Fair	Retrospective case series	The Diabetic Foot Journal	Keller arthroplasty: A cure for the chronic hallux ulceration, hallux limitus and degenerative hallux valgus—A retrospective study.
[66]	Black et al. (2020)	United States	Level IV	Fair	Retrospective case series	Plastic and Reconstructive Surgery	Limb Salvage Rates and Functional Outcomes Using a Longitudinal Slit Arteriotomy End-to-Side Anastomosis for Limb-Threatening Defects in a High-Risk Patient Population.
[67]	Ersin et al. (2020)	Turkey	Level IV	Fair	Retrospective case series	Acta Orthopaedica	Mid-term results of hindfoot arthrodesis with a retrograde intramedullary nail in patients with diabetic Charcot neuroarthropathy.

Levels of evidence were predominantly low to moderate: one study was Level I, seven were Level II, 33 were Level III, and eight were Level IV. Risk of bias assessment rated five studies as “Good” quality and the remainder as “Fair”; no study was judged “Poor.” Sample sizes varied widely, ranging from 27 to over 3000 patients.

3.2. Incidence and Risk Factors for Major Amputation and Mortality

Reported incidence of major amputation varied widely across the included studies, ranging from 0% in small surgical series to as high as 67.7% in high-risk cohorts with extensive ischemia or infection. In larger prospective and retrospective cohorts, typical values clustered around 4–8%: for example, Lo et al. (2023) reported 5.1% among 3475 patients [21], Meloni et al. (2020) 4.6% among 1198 patients [56], and Piaggesi et al. (2020) 4.9% among 1857 patients [54]. Across all studies that reported this outcome, the median incidence was approximately 7.6%, highlighting substantial heterogeneity across populations and designs.

Mortality was less consistently reported. When available, perioperative mortality was generally low (<5%), whereas 1-year mortality typically ranged between 9 and 18% (e.g., 9.1% in Lo et al. (2023) [21]; 15% in Joyce et al. (2020) [55]. Longer-term follow-up showed markedly higher cumulative mortality, with Joyce et al. (2020) reporting 43% at five years [55]. These findings underscore both the variation in outcome definitions and follow-up intervals, and the recurring association of advanced age, renal disease, and cardiovascular comorbidity with adverse survival.

Forest plots of major amputation and mortality incidence, stratified by level of evidence, are provided in Supplementary Figures S1 and S2. Studies lacking numerical outcome data were excluded. These descriptive plots highlight the heterogeneity of reported outcomes and precluded formal meta-analysis. As shown in Supplementary Figures S1 and S2, forest plots of major amputation and mortality incidence revealed wide dispersion across studies, with I² values exceeding conventional thresholds for meaningful pooling. These descriptive plots provide a quantitative depiction of heterogeneity while highlighting the inconsistency in definitions and follow-up intervals across included studies.

Risk factors were heterogeneous but could be categorized into demographic, clinical, and surgical domains (Table 2):

Table 2. Reported Risk Factors and Interventions Associated with Major Amputation and Mortality. Summary of demographic, clinical, and surgical risk factors for major amputation and mortality following limb salvage surgery in diabetes, as reported in included studies. Interventions described in the literature to mitigate risk are also noted. Abbreviations: CKD, chronic kidney disease; PAD, peripheral arterial disease; ESRD, end-stage renal disease; IHD, ischemic heart disease; HbA1c, glycated hemoglobin; CCI, Charlson Comorbidity Index; NPWT, negative pressure wound therapy; PVD, peripheral vascular disease.

Ref.	Authors/Year	Level of Evidence	Number of Patients Included	Major Amputation Incidence	Mortality Incidence	Type of Risk			Interventions to Reduce Risk
						Demographic	Clinical	Surgical
[19]	Davis et al. (2020)	Level I	90	6.7–13.3%	-	-	CKD associated with worse healing	Wound size, infection type	Early NPWT with irrigation
[20]	Lava et al. (2024)	Level II	200	39%	-	Age, sex	Diabetes, CKD, PVD	Procedure type	Standardized surgical pathways
[21]	Lo et al. (2023)	Level II	3475	5.1%	1-year mortality: 9.1%	Prior minor LEA, Charlson Index	PVD, ESRD, IHD, HbA1c	Wagner grade, baseline LEA, CCI	Wound surveillance, LEAPP clinic, revascularization, glucose and lipid control
[22]	Aragón-Sánchez et al. (2023)	Level II	200	7.0% (total); 7.4% (<66 years old); 5.9% (>66 years old)	-	Age, sex	PAD, eGFR, creatinine, BUN, CRP, albumin, retinopathy	Necrosis, severe infection, Staph aureus (protective), CRP, ESR	Antibiotics, revascularization
[23]	Weissler et al. (2021)	Level II	1189	3.7% (total); 5.2% (DM); 1.2% (Non-DM)	7.5% (total); 8.6% (DM); 5.9% (Non-DM)	-	Diabetes	-	Endovascular therapy
[24]	Azhar et al. (2021)	Level II	392 PAD+: 172 PAD−: 22	32.7% (total); 51.7% (PAD+); 17.7% (PAD−)	-	Sex, smoking	PAD, IHD, neuropathy, HbA1c, ABPI, hemodialysis	Wagner classification, PAD presence	Debridement, revascularization
[25]	Ge et al. (2024)	Level II	406	2.2%	34.7%	Age	Access to multidisciplinary care	-	Multidisciplinary care team
[26]	Draper et al. (2025)	Level II	392	-	-	-	Elevated TNF-alpha levels	-	-
[27]	Rohrich et al. (2025)	Level III	33	15.2%	-	Age, BMI	ESRD, PAD, CCI, neuropathy	Flap type, location, direct vs. indirect revascularization	Balloon angioplasty before flap; no free flaps
[28]	Shin & Yoo (2025)	Level III	85	29.4%	10.6%	ESRD, age, ambulation	ESRD on dialysis, low serum albumin, low toe pressure	Wagner grade, UT grade and stage	EVT, antibiotics, wound care
[29]	Almadwahi et al. (2024)	Level III	91	23.1%	1.1%	Age, sex	DM, HTN, AF, dialysis, malignancy	Symptom duration > 36 h, CRP > 5, Rutherford stage, fasciotomy	Anticoagulation, thromboembolectomy or endarterectomy
[30]	Biz et al. (2024)	Level III	60	0%	0%	Sex	High HbA1c	Mechanical load, pressure points	Surgical offloading via osteotomy
[31]	Ormaechevarria et al. (2024)	Level III	512	23.8% (total); 21% (<80 years old); 28% (≥80 years old)	35.9% (total); 27% (<80 years old); 49% (≥80 years old)	Age	Heart disease, CKD, CVD	Revascularization strategy (ET vs. OS vs. conservative), distal disease	Medical therapy, debridement, minor or major amputation
[32]	Li et al. (2024)	Level III	300	-	7.7%	Age, sex, BMI, CCI, ambulation, ESR	Diabetes, CKD, PVD, hypercoagulability	Wound size, location, flap type, infection control, ischemia, flap monitoring	Preop debridement, angiography, antibiotics, glucose control
[33]	Amarasena et al. (2024)	Level III	70	-	-	-	Elevated CRP increased risk for multiple surgeries	Pre-op infection and inflammatory markers associated with multiple surgeries	-
[34]	Huffman et al. (2024)	Level III	48	-	12.5%	Age, sex, BMI	Diabetes, PAD, ESRD, neuropathy, foreign body	Wound location, culture positivity, flap type	Debridement, NPWT, pre-op angiography, antibiotics
[35]	Kostiuket al. (2024)	Level III	157	44	22.7%	Older age, Black race	PAD, ESRD, CAD, high HgbA1c	Extent of tissue loss, osteomyelitis	Revascularization (endovascular and bypass)
[36]	Fujii et al. (2023)	Level III	84	10.6%	8.9%	-	Diabetes mellitus, heart disease, osteomyelitis, dyslipidemia, autoimmune disease	Ischemia grade, CRP, albumin, duration of antibiotics, number of bacterial types, ambulatory status, autoimmune disease, steroid use, infection, skin perfusion pressure	Multidisciplinary team intervention
[37]	Ron et al. (2023)	Level III	341	54.3%	61.3%	Age, sex	Renal disease (dialysis, transplant, eGFR), high creatinine	Prior vascular procedure, Norton Score, hemoglobin	Standard debridement, antibiotics
[38]	Liu et al. (2023)	Level III	120	5% (total); 1.4% (experimental); 10.4% (control)	-	Gangrene, OM, ischemia	Lack of revascularization linked to lower healing	-	Debridement, NPWT, revascularization, skin graft/flap
[39]	Ragghianti et al. (2023)	Level III	55	0%	3.6% (total); 0% (case); 5.9% (control)	-	Use of local antibiotics during minor amputation	Use of local antibiotics during minor amputation	Use of local antibiotics during minor amputation
[40]	Zamzam et al. (2023)	Level III	57	6.1%	-	-	Access to multidisciplinary care	-	Implementation of a multidisciplinary acute care pathway
[41]	Zhu et al. (2023)	Level III	262	-	32.9%	Charlson index, ASA score	PVD, nephropathy, comorbidities	Wagner grade, anesthesia type	Ultrasound-guided PNB vs. standard GEA
[42]	Yammine et al. (2023)	Level III	134	-	-	-	Neutrophil count, NLR, WCC, RDW	Osteomyelitis associated with amputation	-
[43]	Bobirca et al. (2022)	Level III	251	14.2% (total); 7.6% (BKA); 6.6% (AKA)	-	-	Anemia (OR 2.975), nephropathy (OR 3.565), hyperglycemia	Gangrene (OR 2.230), infected Charcot foot (OR 5.316)	-
[44]	Abu El Hawa et al. (2022)	Level III	62	67.7% (total); 43.5% (BKA); 1.6% (AKA); 12.9% (TMA); 19.4% (toe); 3.2% (calcanectomy)	16.1%	Age, thrombocytopenia, hypotension, pain, erythema	DM, PVD, necrotic skin, high LRINEC	High LRINEC, late debridement, PVD	Surgical debridement, negative pressure wound therapy, staged amputation
[45]	Huang et al. (2022)	Level III	260	-	1.5%	-	Low TIR, high WBC, low albumin, HDL, high Wagner score, insulin use, type of surgery	Amputation + grafting increased risk	Strict perioperative glycemic control (time in range)
[46]	Kim et al. (2022)	Level III	90	-	-	Age	HbA1c, PAD, osteomyelitis, CKD (none confounded outcome)	Infection location (toe vs. foot), surgery type (bone resection vs. debridement)	-
[47]	Nigam et al. (2022)	Level III	200	-	-	Age, sex, BMI	DM, PVD, ESRD, thrombophilia	Flap type, calcified arteries, vessel runoff, venous reflux	Preop angiography, thrombophilia panel, venous ultrasound
[48]	Schmidt et al. (2022)	Level III	70	-	-	Age, BMI	CKD, smoking	Debridement in OR (rather than bedside)	Debridement procedure performed at bedside (rather than OR)
[49]	Haug et al. (2021)	Level III	1230	-	3.5% (total); 3.6% (LXTR); 3.4% (VascLXTR)	Age, BMI, ASA	Diabetes, CHF, dialysis, cancer	Wound classification, prolonged OR time, albumin, smoking	Revascularization with angioplasty, bypass, AV loops
[50]	Thai et al. (2021)	Level III	65	3.1%	-	Age, sex, smoking	ESRD (p = 0.011), PAD, diabetes	Flap type, use of NPWT, antibiotic beads	Antibiotics, NPWT, angioplasty if needed
[51]	Koivunen et al. (2021)	Level III	497	39% (total); 46.9% (IT-DM); 17.5% (NIT-DM); 35.6% (non-DM)	18.7% (1-year); 39% (3-year)	Age, sex	Diabetes (insulin vs. non-insulin), CAD, MI, HF, CKD	Revascularization type, CIx IV, lesion location	Bypass with saphenous vein; endovascular PTA
[52]	Burmeister et al. (2021)	Level III	175	2.2% (total); 4% (discharged home); 10% (discharged to SNF)	5.5% (home); 20.6% (SNF); p = 0.012	-	Charlson Comorbidity Index compared	Preoperative infection present in all cases	Direct surgical–SNF communication; SNF staff trained in limb-salvage postop care; Track staff–patient ratios; SNF discharge = higher postop complexity
[53]	Kim et al. (2021)	Level III	20	5% (total); 16.7% (DM); 0% (Non-DM)	-	-	Diabetes linked to some flap complications	Flap design and pedicle length considerations	Free flap reconstruction to avoid amputation
[54]	Iacopi et al. (2020)	Level III	1857	4.9%	27.9%	Age, Charlson Comorbidity Index	Retinopathy, renal disease	Surgical type, era of surgery, revascularization rates	Revascularization, debridement, reconstructive surgery
[55]	Joyce et al. (2020)	Level III	47	11%	0% (1 month); 15% (1 year); 43% (5 years)	Age, sex	Renal disease, PVD, glycemic control	Healing influenced by wound closure, infection extent, vascular supply	Antibiotic beads, NPWT, tendon rebalancing, gastrocnemius recession
[56]	Meloni et al. (2020)	Level III	1198	4.6% (total); 0.5% (neuropathic); 6.6% (ischemic)	7.7% (total); 1.1% (neuropathic); 11% (ischemic)	Age, sex	Nephropathy, IHD, ESRD, HF	Ulcer size, depth, infection, revascularization failure	Antibiotics, revascularization, infection control
[57]	Piwnica-Worms et al. (2020)	Level III	129	7.8%	-	Diabetes	Residual osteomyelitis, flap failure, pre-op debridements	Arterial injury (esp. PT), poor vessel runoff, recipient artery pathology	Angiography, flap choice consideration
[58]	Lo et al. (2022)	Level III	220	7.3% (total); 3% (LEAPP); 8.7% (Non-LEAPP)	-	-	-	-	Use of multidisciplinary team approach
[59]	Seçkin et al. (2022)	Level III	114	48.2%	-	-	Albumin, duration of diabetes, smoking, hypertension	Number of debridements, distal amputation level	-
[60]	Brodell et al. (2020)	Level III	35	22.9%	-	Age	Charlson Comorbidity Index, CRP, PAD	Extent of ulcer, vancomycin use, clean vs. contaminated margin	Intraoperative vancomycin powder (vs none)
[61]	Alsabbagh et al. (2024)	Level IV	11	36.4%	18.2%	-	-	-	DVA prevented major amputation in some patients with no-option CTLI: 6 out of 12 limbs (50%) achieved full or partial healing during follow-up
[62]	Brekelmans et al. (2023)	Level IV	35	8.6%	11%	-	Access to multidisciplinary care	-	Multidisciplinary treatment approach including reconstructive surgery
[63]	Altuntaş et al. (2023)	Level IV	42	0%	-	-	Diabetes, PAD associated with higher risk	Defect size, ischemic comorbidity	Bipedicled flap, serial debridement, antibiotics
[64]	Vasukutty et al. (2022)	Level IV	47	6.4%	-	-	-	-	Use of CERAMENT G antibiotic biocomposite
[65]	Frykberg et al. (2021)	Level IV	57	0%	-	-	Ulceration, hallux limitus/valgus, peripheral neuropathy	-	Surgical deformity correction (Keller arthroplasty)
[66]	Black et al. (2020)	Level IV	115	16.5%	-	Age, sex	DM, HTN, hypercoagulability	Intraoperative thrombosis, vessel calcification, reanastomosis	Debridement, vascular imaging, hypercoagulation workup, vein graft patching
[67]	Ersin et al. (2020)	Level IV	24	0%	-	-	-	-	Arthrodesis with retrograde intramedullary nailing, second-generation IM nails

• Demographic factors. Increasing age consistently predicted poor outcomes, with markedly higher amputation and mortality risk in patients over 80 years. Male sex and Black race were also identified as independent predictors in selected cohorts.
• Clinical comorbidities. Renal dysfunction emerged as one of the strongest predictors: chronic kidney disease, dialysis dependence, and reduced eGFR were consistently associated with higher amputation and mortality rates. Cardiovascular comorbidities-including peripheral arterial disease, ischemic heart disease, and congestive heart failure-were also frequently linked to adverse outcomes. Laboratory markers such as hypoalbuminemia, elevated C-reactive protein, and poor glycemic control (high HbA1c) further stratified risk. Interestingly, colonization with Staphylococcus aureus was reported as protective in one cohort, underscoring the variability of findings across studies.
• Surgical and wound-related factors. Larger wound size, higher ulcer severity (Wagner/University of Texas classifications), osteomyelitis, and ischemia consistently predicted limb salvage failure. Procedural determinants included flap type, need for repeat debridement, and choice of revascularization strategy. Discharge destination also proved influential: patients discharged to skilled nursing facilities demonstrated higher risks of delayed healing, repeat hospitalization, and major amputation compared with those discharged home.
• Mortality outcomes. Mortality was less frequently reported than amputation, but when available, the same risk profile was observed. Advanced age, renal disease, ischemic heart disease, and insulin-dependent diabetes were recurrent predictors. Reported mortality ranged from 8% to 18% at one year, with long-term survival falling substantially in patients with combined diabetes and chronic limb-threatening ischemia.

3.3. Multidisciplinary Care Models

Seventeen studies described multidisciplinary care pathways for diabetic limb salvage (Table 3). Team composition and structure varied, ranging from orthoplastic collaborations (vascular and plastic surgeons) to comprehensive services integrating vascular surgery, podiatry, endocrinology, infectious diseases, wound care, rehabilitation, nursing, and case management.

Table 3. Multidisciplinary Care Models for Diabetic Limb Salvage. Descriptions of multidisciplinary team (MDT) models reported across 17 included studies. Details include MDT structure and composition, clinical specialties involved, and reported patient outcomes. Abbreviations: MDT, multidisciplinary team; Vasc, vascular surgery; Endo, endocrinology/diabetes medicine; Podi, podiatry; Ortho, orthopedic surgery; Plast, plastic surgery; ID, infectious diseases; Rehab, rehabilitation medicine; PT, physical therapy; LOS, length of stay; LEA, lower-extremity amputation; SNF, skilled nursing facility.

Ref.	Authors/Year	Level of Evidence	Number of Patients Included	MDT Model/Type	Team Composition	Reported Outcomes
[20]	Lava et al. (2024)	Level II	200	Unspecified		Interdisciplinary care noted, with preoperative optimization and patient selection emphasized.
[21]	Lo et al. (2023)	Level II	3475	Multi-site	Primary to Tertiary, Vasc, Endo, Podi	Diabetic foot coordinators; 80% ↓ minor, 35% ↓ major amputations.
[22]	Aragón-Sánchez et al. (2023)	Level II	200	Consultative	Ortho, Vasc, ID	Joint evaluation for revascularization, resection, or amputation in DFO patients
[25]	Ge et al. (2024)	Level II	406	Limb salvage pathway	Vasc, Endo, ID, Ortho, Wound	Improved healing, ↓ amputations, ↑ 1-yr survival.
[27]	Rohrich et al. (2025)	Level III	33	Orthoplastic	Vasc, Plast	Joint operative planning and revasc-first strategy; high limb salvage, functional recovery
[29]	Almadwahi et al. (2024)	Level III	91	Consultative	Vasc, Ortho, ID, Endo	Revasc-first, debridement, ABX stewardship; ↓ amputations, ↑ healing.
[31]	Ormaechevarria et al. (2024)	Level III	512	Vascular-led	Vasc, Med, Geri, Anes	Board-style decisions on revasc vs. amputation; safe outcomes in octogenarians
[32]	Li et al. (2024)	Level III	300	Clinic	Vasc, Endo, ID, Podi, Wound	↓ Amputations, ↓ mortality, faster healing
[33]	Amarasena et al. (2024)	Level III	70	Surgical	Ortho, Vasc, Anes, Micro	Periop co-management; CRP > 84 mg/L predicted repeat surgery
[34]	Huffman et al. (2024)	Level III	48	Rehab-focused	Vasc, Rehab, PT	Joint rounds, stump care, early rehab, discharge planning
[35]	Kostiuket al. (2024)	Level III	157	Consultative	Vasc, Ortho, Podi, ID	PC within limb salvage program; ~33% major amputation at follow-up
[36]	Fujii et al. (2023)	Level III	84	Diabetic foot program	Vasc, Endo, Podi, ID, Wound	↓ Major amputations, ↑ limb salvage.
[37]	Ron et al. (2023)	Level III	341	Consultative	Vasc, Ortho, ID	Preop decision on TMA vs. higher amputation
[38]	Liu et al. (2023)	Level III	120	‘Five-in-One’ pathway	Vasc, Ortho, Plast, Endo, ID, Wound	↑ Healing, ↓ time, ↓ amputations, ↓ recurrence.
[40]	Zamzam et al. (2023)	Level III	57	Acute pathway	Chiropody, Vasc, Med, EM, Nsg	↓ LOS, ↓ costs, ↑ discharge home.
[43]	Bobirca et al. (2022)	Level III	251	LEAPP clinic	Vasc, Podi, Endo, ID	↓ Amputations, ↓ mortality, faster healing.
[44]	Abu El Hawa et al. (2022)	Level III	62	Limb salvage service	Vasc, Podi, Plast, Ortho, ID	↓ Amputations (36→9%), ↑ limb salvage (62→84%)
[46]	Kim et al. (2022)	Level III	90	VA limb preservation	Vasc, Podi, ID	Bone resection ↑ healing (69% vs. 38%), esp. toes (RR 4.5)
[47]	Nigam et al. (2022)	Level III	200	Limb salvage service	Vasc, Plast, Podi, ID	59% limb salvage, 76% survival at 1 yr after failed revasc.
[52]	Burmeister et al. (2021)	Level III	175	Discharge planning	Surg, Med, PT, CaseMgr	SNF discharge ↑ risk: slower healing, ↑ rehospitalization, major amputation, mortality
[54]	Iacopi et al. (2020)	Level III	1857	Limb salvage program	Vasc, Podi, Plast, Ortho, ID	78% limb salvage, 64% amputation-free survival at 2 yrs
[55]	Joyce et al. (2020)	Level III	47	GVG guideline consensus	Vasc, Podi, Endo, ID, Wound	PLAN framework; MDT cornerstone of CLTI care.
[56]	Meloni et al. (2020)	Level III	1198	Specialized DFU unit	Endo, Podi, Vasc, Ortho, ID, Wound	65% healed, 8% amputation, 18% mortality (1 yr)
[57]	Piwnica-Worms et al. (2020)	Level III	129	Orthoplastic program	Plast, Ortho, Trauma, Vasc, Podi, Wound	7.8% secondary amputation.
[58]	Lo et al. (2022)	Level III	220	LEAPP clinic	Vasc, Podi, Endo	↓ Minor amputation (14→3%), ↓ major LEA (–33% adj.), ↓ mortality (19.2→7.5%); USD $1.86M cost avoidance
[59]	Seçkin et al. (2022)	Level III	114	Council	Ortho, Plast, Endo, ID, Cardio	Determined amputation level; re-amputation risks = distal level, low albumin, smoking, HTN, diabetes duration, debridements
[60]	Brodell et al. (2020)	Level III	35	Ortho	Foot, ankle	ID (post-op abx), Vasc (ABI referral); partial/total calcanectomy; vancomycin powder ↑ RTOR (85% vs. 36%), no benefit; limb salvage 83%.
[62]	Brekelmans et al. (2023)	Level IV	35	MDT	11 specialists, HIPA model: Endo, ID, Vasc, Podi, Surg, Foot, Ankle recon	↓ Recon surgery for CN/DFU; wound closure 69%, ulcer-free 358d, success (mobility w/o ulcer/amputation) 77%, LEA 14% (3 major, 2 minor)
[64]	Vasukutty et al. (2022)	Level IV	47	Endo/Ortho/Vasc/Podi/ID-Micro/Radiol; 2-center UK, debridement ± recon + local gentamicin		Healing 88%, limb salvage 94%, mobility 83%, 3 major LEA (all heel OM)

Abbreviations: General Terms: MDT—Multidisciplinary Team; LOS—Length of Stay; SNF—Skilled Nursing Facility; PC—Palliative Care; LEA—Lower-Extremity Amputation; RTOR—Return to Operating Room; CN—Charcot Neuroarthropathy; DFU—Diabetic Foot Ulcer; DFO—Diabetic Foot Osteomyelitis; OM—Osteomyelitis; GVG—Global Vascular Guidelines; PLAN framework—Patient risk, Limb severity, ANatomic complexity (GVG decision tool); HIPA model—Hyperglycemia, Infection, Pressure, Arterial flow (Dutch MDT model); Clinical Specialties: Vasc—Vascular Surgery; Endo—Endocrinology/Diabetes Medicine; Podi—Podiatry; ID—Infectious Disease; ID-Micro—Infectious Disease/Microbiology (lab-based antibiotic guidance); Ortho—Orthopedic Surgery; Plast—Plastic Surgery; Trauma—Trauma Surgery; Surg—Surgery (general/unspecified); Med—Internal Medicine; Geri—Geriatrics; Anes—Anesthesiology; Micro—Microbiology; Radiol—Radiology (esp. musculoskeletal imaging); Rehab—Physical Medicine & Rehabilitation; PT—Physical Therapy; CaseMgr—Case Management; Chiropody—Chiropody/Chiropodist (foot health practitioner); EM—Emergency Medicine; Nsg—Nursing; Treatments/Measures: Revasc—Revascularization; ABX—Antibiotics; ABI—Ankle–Brachial Index; TMA—Transmetatarsal Amputation; CERAMENT-G—Calcium sulphate/hydroxyapatite bone void filler loaded with gentamicin (local antibiotic carrier).

Despite heterogeneity, several consistent findings emerged. Implementation of MDTs was generally associated with reduced major amputation rates, improved healing, and enhanced survival. For example, one multi-site program reported a 35% reduction in major amputations, while LEAPP (Lower Extremity Amputation Prevention Program) clinics in North America demonstrated both reduced amputation and lower mortality, alongside significant cost savings. Acute pathway and rehabilitation-focused MDTs emphasized improved discharge outcomes and functional recovery. Nevertheless, all evidence was derived from observational cohorts, with outcome definitions and follow-up periods varying widely, limiting definitive conclusions regarding causality.

3.4. Real-World Comparison: Mid Essex Diabetes Amputation Reduction Plan

The MEDARP served as a real-world comparator to the systematic review findings (Table 4a–c). Implemented at Broomfield Hospital between December 2021 and March 2024, MEDARP adopted a multidisciplinary “toe and flow” model integrating vascular, podiatric, plastic surgical, and allied health support.

Table 4. (a–c) Baseline Characteristics, Patient-Reported Outcomes, and Clinical Outcomes in the MEDARP Cohort. Table 4a. Baseline demographic and clinical characteristics. Continuous variables are shown as mean ± SD unless otherwise specified; diabetes duration and Rockwood frailty are presented as median (range). Table 4b. Patient-reported outcomes (PROs) measured at baseline, 6 months, and ~11 months. Higher MSK-HQ scores indicate better function; higher MOXFQ subscale scores indicate greater impairment. Asterisks (*) denote statistically significant improvements from baseline (Wilcoxon signed-rank, p < 0.001). Table 4c. Clinical outcomes during follow-up, including incidence of major amputation and all-cause mortality.

(a)
Characteristic		Value
Patients, n		72
Age, mean (SD)		68.1 (14.0)
Sex: Male		21 (84.0%)
Sex: Female		4 (16.0%)
Diabetes type 1		14 (18.2%)
Diabetes type 2		63 (81.8%)
Diabetes duration, mean years (median)		20.0 (18.9)
Rockwood frailty, median (range)		5.0 (3–9)
Fontaine stage, worst (I/II/III/IV, n)		7/20/8/31
HbA1c, mean (SD)		63.8 (16.7)
eGFR, mean (SD)		59.2 (19.1)
(b)
Measure	Pre-op	6 Months	11 Months
MSK-HQ	29.3 (11.1)	44.2 (14.1) *	49.4 (9.9) *
MOXFQ Walking/Standing	68.0 (27.0)	16.8 (27.6) *	13.7 (22.8) *
MOXFQ Pain	42.5 (27.6)	6.0 (11.4) *	6.2 (11.6) *
MOXFQ Social Interaction	56.6 (31.0)	15.8 (24.5) *	11.8 (20.4) *
Physical activity (days/week)	1.5 (2.2)	3.3 (3.0) *	-
(c)
Outcome		n Patients
Major amputation (any mention)		5
Death (any mention)		9

* = Statistically significant change from baseline by Wilcoxon signed-rank test (p < 0.001).

A total of 72 patients (84 surgical admissions) were treated during this period. The cohort was older (mean age 68.1 ± 14.0 years), predominantly male (84%), and largely affected by long-standing type 2 diabetes (81.8%) with advanced vascular disease (Fontaine stage III–IV in 59%). Patients also demonstrated significant frailty (median Rockwood score 5, range 3–9). Procedural staging reflected complex presentations, with most classified as Stage 3 (73%) or Stage 4 (24%), and revision surgery was common. Vascular (23%), podiatric (4%), and orthopedic (5%) prior interventions were documented, with adjuvant biomaterials (e.g., Stimulan, Cerement, dermal grafts) used selectively.

Prior to implementation of MEDARP at our facility, major amputation in patients presenting with need for class 4, emergent, diabetic foot surgery (procedures undertaken to limit progression of acute limb-threatening infection) were almost universally associated with progression to major amputation [11]. In MEDARP, 24% of the caseload comprised class 4 surgery. While no control group was available, the observed incidence of 6.9% major amputation and 12.5% mortality lies at the lower end of ranges reported in contemporary cohorts. These findings are associative and should be interpreted in that context.

Patient-reported outcomes also improved significantly over time. MSK-HQ scores increased from 29.3 ± 11.1 at baseline to 44.2 ± 14.1 at 6 months and 49.4 ± 9.9 at ~11 months (Wilcoxon signed-rank p < 0.001 for both comparisons). MOXFQ Walking/Standing scores declined from 68.0 ± 27.0 to 16.8 ± 27.6 at 6 months and 13.7 ± 22.8 at ~11 months (p < 0.001). Similar improvements were observed in MOXFQ Pain (42.5 ± 27.6 → 6.0 ± 11.4 → 6.2 ± 11.6; p < 0.001) and Social Interaction (56.6 ± 31.0 → 15.8 ± 24.5 → 11.8 ± 20.4; p < 0.001). Physical activity frequency also rose significantly, from 1.5 ± 2.2 to 3.3 ± 3.0 days/week at 6 months (p < 0.001). Exploratory univariate analyses are presented in Supplementary Table S1. Patients who experienced major amputation were more likely to have advanced Fontaine stage and ASA ≥ 3, while those who died more often demonstrated higher frailty scores, lower eGFR, and anemia. These associations were descriptive only, as the small number of events precluded multivariate regression, but they align with the risk factors identified in the systematic review.

Taken together, these results suggest that the MEDARP program not only reduced adverse clinical events but was also associated with statistically and clinically meaningful improvements in function and quality of life, providing practical confirmation that locally adapted MDT models can translate into measurable benefits in real-world practice.

4. Discussion

4.1. Summary of Main Findings

This study combined a systematic review with a real-world comparison. This systematic review identified 49 studies published between 2020 and 2025 evaluating outcomes of major amputation and mortality following limb salvage surgery in patients with diabetes. Most studies were observational with moderate risk of bias, underscoring the limited availability of high-level evidence in this field. Across studies, risk factors for poor outcomes clustered into demographic (older age, male sex, Black race), clinical (renal disease, cardiovascular disease, poor glycemic control, inflammatory markers), and surgical (large wounds, severe ischemia, osteomyelitis, higher ulcer classification, complex reconstructions) domains. Mortality outcomes were less frequently reported, but consistently associated with advanced age, renal disease, and cardiovascular comorbidity. In our MEDARP cohort, exploratory univariate associations (Supplementary Table S1) suggested that frailty, impaired renal function, anemia, advanced Fontaine stage, and higher ASA class may contribute to adverse outcomes, consistent with the risk profile identified in the review. We considered but did not conduct multivariate regression due to the very limited number of outcome events, which precluded stable modeling. Larger multicenter cohorts will be required to test independent predictors more robustly.

Nearly 30 studies described MDT models for diabetic limb salvage. Despite heterogeneity in structure and composition, MDT implementation was generally associated with reduced rates of major amputation, improved wound healing, shorter length of stay, and in some programs, improved survival and cost savings. The real-world experience of the MEDARP provided further evidence that locally adapted MDT models can translate published findings into measurable reductions in amputation burden.

These findings were broadly consistent with exploratory analyses from the MEDARP cohort, where patients with higher frailty, poorer renal function, anemia, and impaired musculoskeletal health were more likely to die, and ASA ≥ 3 and advanced Fontaine stage were associated with major amputation. While underpowered for definitive inference, these real-world signals align with the literature and reinforce the multidimensional risk profile of this population.

4.2. Comparison with Previous Literature

Our synthesis corroborates longstanding observations that renal dysfunction, peripheral arterial disease, and broader cardiovascular comorbidity are among the strongest predictors of limb loss and mortality in diabetes [68,69]. Although our review was limited to studies published since 2020, we acknowledge that earlier foundational work established many of the key risk factors and principles of MDT care (e.g., Armstrong et al., 2017 [2]; Rogers et al., 2010 [11]). Our review builds upon this groundwork by emphasizing how these predictors and strategies have been evaluated in the most recent five years, a period that coincides with advances in endovascular techniques, orthoplastic collaboration, and integrated care pathways. The recurrent association we observed between poor nutritional status (e.g., hypoalbuminemia), systemic inflammation (elevated CRP), and suboptimal glycemic control (higher HbA1c) with adverse outcomes is also consistent with prior reviews and cohort studies in diabetic limb care, which emphasize the compounded risk carried by metabolic and inflammatory dysregulation [70,71,72]. Likewise, wound severity (including osteomyelitis), ischemia, and higher ulcer classification have repeatedly been linked to limb-salvage failure across settings [73].

Where our findings add nuance is in two areas. First, discharge destination emerged in our dataset as a pragmatic signal of downstream risk—patients discharged to skilled nursing facilities experienced worse outcomes than those discharged home. While this has been less emphasized in earlier surgical series, related work in postoperative care and social determinants research suggests that functional dependency, staffing ratios, and protocol variability may all contribute to the observed gradient in outcomes [74,75]. Second, we observed a counter-intuitive signal suggesting lower failure among patients colonized with Staphylococcus aureus. This contradicts the established literature, in which S. aureus, particularly MRSA, is recognized as a major pathogen driving infection and adverse outcomes. The most plausible explanation is unmeasured confounding: colonization may have triggered more aggressive prophylaxis, closer surveillance, or it may represent a statistical anomaly. This isolated finding contradicts established evidence and should be interpreted with extreme caution; it likely reflects unmeasured confounding rather than a true protective effect [76,77,78,79]. Our findings on MDT care are broadly concordant with prior vascular, podiatric, and orthoplastic literature: structured, team-based pathways tend to be associated with reduced major amputation, improved healing, and, in some programs, better survival and lower costs [7,80,81]. As in earlier reviews, however, the methodological substrate remains predominantly observational, with substantial heterogeneity in MDT composition (e.g., vascular-led vs. orthoplastic collaborations vs. integrated clinics), referral triggers, timing and availability of revascularization, and outcome definitions (amputation-free survival vs. amputation incidence vs. composite endpoints) [7,82]. This heterogeneity, coupled with variable follow-up durations, likely explains the spread of reported effect sizes across programs and limits formal meta-analytic inference. Our formal heterogeneity assessments reinforce this conclusion: I² values consistently exceeded 80%, and outcome definitions varied widely (e.g., ‘major amputation’ defined as proximal-to-ankle vs. below-knee only; mortality reported perioperatively, at 30 days, or up to five years). These findings underscore the need for standardized outcome definitions and reporting frameworks in future research. Finally, the under-reporting of longer-term mortality in the included studies contrasts with historical data showing high 3–5-year mortality after major amputation in diabetes [5,68,83]. Our review therefore aligns with previous calls for standardized outcome sets that pair limb-related endpoints with survival and readmission, enabling clearer benchmarking across institutions and care models [75,83].

In comparison to these published findings, outcomes observed in the MEDARP cohort were favorable despite the high-risk profile of enrolled patients. Among 72 patients undergoing 84 admissions, the incidence of major amputation was 6.9% (5/72) and mortality from any cause was 12.5% (9/72) during follow-up. These rates lie at the lower end of the ranges reported across the literature (major amputation 0–68%, median ~7.6%; mortality typically ~13–20% at one year in recent syntheses [68,83], rising to ~40–50% at five years overall [2,83] and substantially higher after major amputation [2]), suggesting that the integrated “toe and flow” multidisciplinary approach at Broomfield Hospital achieved outcomes comparable to or better than those of larger published series. In addition to reducing adverse events, patient-reported outcomes improved substantially: MSK-HQ scores rose from 29.3 at baseline to 49.4 at 11 months, while MOXFQ subscales for walking/standing, pain, and social interaction all declined markedly, reflecting less pain and greater functional independence. While interpretation is limited by sample size and follow-up duration, the concordance between MEDARP’s results and the risk factors identified in our review supports the translational relevance of MDT-based models in real-world clinical practice.

4.3. Implications for Clinical Practice

Risk stratification and shared decision-making. In patients considered for limb salvage, routine assessment should integrate renal function (eGFR/dialysis status), vascular status (including ischemia grading), nutritional markers (albumin), glycemic control (HbA1c), and inflammatory burden (CRP) alongside wound severity and infection extent. These factors-consistently associated with amputation and mortality in our dataset-can be combined to create a transparent risk conversation with patients and caregivers, inform expectations, and, where appropriate, prompt discussion of early primary amputation versus prolonged salvage attempts in those with very high predicted failure or mortality risk [70,71,72,73,79,84,85].

Perioperative optimization and sequencing. Programs should prioritize revascularization-first strategies where feasible, timely source control for infection (debridement, targeted antibiotics), glucose optimization, and nutrition support. For complex reconstructions, thoughtful selection of flap type, minimizing repeat debridements where possible, and standardized protocols for osteomyelitis can reduce downstream failure. These elements align with patterns in our review (wound/ischemia severity and osteomyelitis predicting failure; procedural decisions influencing outcomes) and echo prior best-practice recommendations [73,86].

MDT. Given consistent associations between MDT care and improved outcomes across diverse models, health systems should implement a core MDT at minimum (vascular surgery, podiatry/foot surgery, infectious diseases/microbiology, wound nursing), with extended roles (endocrinology/diabetes education, plastics/orthopedics, rehabilitation, case management) added based on local resources. Practical implementation features linked to benefit in published programs include:

• Fast-track access to vascular imaging and revascularization (“toe and flow” paradigm) [87,88].
• Co-located clinics or virtual boards that enable same-day cross-specialty decisions and reduce time-to-definitive treatment [88,89].
• Standardized order sets and care bundles (perioperative antibiotics, glycemic targets, off-loading, dressing protocols) to reduce unwarranted variation [86,90].
• Defined post-discharge pathways with early (e.g., 72-h) follow-up and rapid re-entry for wound setbacks [91,92].

Transitions of care and discharge planning. Because discharge to skilled nursing facilities was associated with worse outcomes in our dataset, teams should treat SNF discharge as a risk flag rather than a routine endpoint. Where SNF placement is unavoidable, programs can mitigate risk by: (i) communicating procedure-specific wound and off-loading protocols to SNF staff; (ii) establishing named points of contact for escalation; (iii) scheduling early clinic review; and (iv) incorporating telehealth/photo triage to identify deterioration sooner [92,93,94]. These measures reflect the system-level signals identified in our review and align with broader evidence on reducing post-acute complications.

Measurement and quality improvement. To strengthen local practice and comparability, centers should track a minimum outcome set: major amputation, amputation-free survival, time-to-revascularization, reoperation, readmission, and mortality at standardized intervals (e.g., 30/90 days, 1 year). When feasible, adding patient-reported outcomes and economic metrics (e.g., bed days, re-admissions avoided) can clarify the value of MDT pathways and inform resourcing decisions [75,95]. In the context of increasing pressures across the healthcare system, leaders and clinicians must also remain attentive to vulnerabilities in funding streams. The sustainability of MDT programs depends not only on their clinical effectiveness but also on continuous demonstration of value to commissioners and policymakers. Maintaining up-to-date, rolling data collection on patient outcomes, readmissions avoided, and cost savings is therefore essential. Such data strengthen the case for ongoing investment, help protect service lines from future contraction, and ensure that program benefits are visible to both patients and health systems. In line with recent initiatives to develop a core outcome set for diabetic foot research [75], we recommend that future studies consistently report a minimum standardized set of outcomes, including amputation-free survival at 1 and 5 years, wound healing time, and overall survival. Adoption of such a core outcome set would enhance comparability across studies, reduce heterogeneity in reporting, and accelerate translation of evidence into practice.

4.4. Strengths and Limitations

The strengths of this review include adherence to a pre-registered protocol, comprehensive literature search, and structured risk-of-bias assessment. We also incorporated a novel real-world comparator (i.e., MEDARP), enhancing the translational relevance of our findings.

Several limitations should be acknowledged. First, the evidence base remains dominated by retrospective cohorts, with only one randomized controlled trial identified. Second, definitions of outcomes (major amputation, limb salvage, amputation-free survival) were inconsistent, and follow-up periods varied widely, precluding pooled quantitative analysis. Third, reporting of mortality was infrequent, limiting conclusions about long-term survival after salvage procedures. Fourth, we did not perform separate analyses for Type 1 versus Type 2 diabetes, as most published studies did not stratify outcomes by diabetes type and our real-world cohort was overwhelmingly Type 2. This remains an important gap in the literature that future multicenter studies should address. Finally, while we attempted to capture multidisciplinary models comprehensively, variations in terminology and reporting may have led to underestimation of MDT impact.

4.5. Future Directions

Future research should prioritize prospective multicenter cohort studies and randomized evaluations of limb salvage strategies. Standardized outcome definitions and uniform reporting of both amputation and survival endpoints would improve comparability. More rigorous evaluation of MDT models-including economic analyses and patient-reported outcomes-will also be essential to establish best practices. Finally, studies exploring the role of social determinants, discharge pathways, and community-based follow-up may provide new avenues for reducing amputation burden in high-risk populations.

5. Conclusions

This study, which combined a systematic review and a real-world comparison, demonstrates that major amputation and mortality are strongly influenced by renal, cardiovascular, metabolic, and wound-related factors. Multidisciplinary team models consistently correlate with improved outcomes in the literature, and the MEDARP program provides real-world confirmation that such models can be adapted to local contexts, bridging the gap between research and practice. Sustaining these gains in an era of constrained resources will require proactive attention to funding vulnerabilities and ongoing data collection that demonstrates both improved patient outcomes and economic value. These conclusions should be interpreted with caution given that the current evidence base is dominated by observational and single-center studies, with very few randomized or multicenter trials available. As such, our recommendations highlight consistent associations rather than definitive causal predictors, and underscore the need for future high-quality prospective research to strengthen the evidence base.