Nanosilver Dressings and 12-Week Healing in Diabetic Foot Ulcers: A Retrospective Propensity Score-Matched Study

Summary

This retrospective propensity score-matched study evaluates the association of nanosilver versus alginate dressing strategies with 12-week wound healing and related clinical outcomes in multidisciplinary diabetic-foot care.

Abstract

Diabetic foot ulcers (DFUs) are associated with prolonged morbidity and limb-related complications. This single-center retrospective cohort study compared nanosilver and alginate primary dressing strategies in patients managed within a multidisciplinary diabetic-foot program between January 2020 and December 2023. Among 393 screened patients, 361 formed the unmatched cohort, including 121 treated with nanosilver and 240 with alginate dressings; 1:1 propensity-score matching generated 115 matched pairs. The single primary endpoint was complete wound healing at week 12. Healing occurred in 87 of 115 patients (75.7%) receiving nanosilver and 65 of 115 (56.5%) receiving alginate dressings (matched OR, 2.47; 95% CI, 1.32–4.84; P = 0.003; matched RR, 1.34; 95% CI, 1.11–1.62). At least 50% wound-area reduction at week 4 occurred in 82.6% versus 67.0% (matched OR, 2.80; 95% CI, 1.32–6.46; Holm-adjusted P = 0.015), and median time to complete healing was 45 versus 58 days (HR, 1.58; 95% CI, 1.22–2.05; Holm-adjusted P < 0.004). The primary association remained similar in the extended measured-confounding sensitivity analysis (adjusted OR, 2.31; 95% CI, 1.17–4.56). Other clinical and safety outcomes were exploratory. Nanosilver dressing use was associated with greater 12-week complete healing and shorter time to closure in this matched tertiary-care cohort; prospective multicenter validation is required.

Introduction

Diabetic foot ulcers (DFUs) are a major complication of diabetes and are associated with prolonged morbidity, infection, hospitalization, lower-extremity amputation, and substantial healthcare burden¹. Despite advances in multidisciplinary care, a substantial proportion of ulcers remain unhealed at 12 weeks, and recurrence after wound closure is common^2,3. Even with progress in multidisciplinary care, as many as one-third of DFUs remain unhealed in 12 weeks, and over 40% recur in the first year. DFU is also associated with substantial systemic morbidity, and major lower-extremity amputation carries a poor long-term prognosis^4,5,6,7,8. Accordingly, improving wound closure while preventing limb-threatening complications remains an important goal of multidisciplinary diabetic-foot care^9,10,11.

Important components of contemporary DFU management include comprehensive wound care, encompassing sharp debridement, infection management, offloading, vascular assessment and management, and appropriate dressing selection^12,13,14. Calcium alginate and hydrocolloid dressings are commonly used because of their exudate-absorbing properties and ability to maintain a moist wound environment^9,15,16. However, conventional non-impregnated dressings primarily provide wound protection and exudate management and generally lack intrinsic antimicrobial or actively regenerative functions. Persistent bacterial burden, chronic inflammation, and oxidative stress can delay granulation and re-epithelialization, thereby prolong wound healing and increasing the risk of adverse limb outcomes. These limitations have stimulated the development of bioactive and multifunctional dressings designed to modulate the wound microenvironment and support tissue repair^17,18,19.

Nanosilver-containing dressings represent one bioactive approach intended to combine wound-environment management with local antimicrobial activity. These dressings release ionic silver (Ag⁺) in a regulated fashion, which results in sustained broad-spectrum antimicrobial activity, breaches of bacterial biofilm, and inhibitory effects on the geometric reaction of local inflammatory responses^{20,21,22,23,24,25}. Preclinical studies have suggested that silver-containing nanomaterials may influence fibroblast activity, angiogenesis, inflammatory signaling, and extracellular-matrix remodeling, although the extent to which these mechanisms translate into clinically meaningful differences in DFU healing remains uncertain. There are preliminary clinical studies that have reported increased healing rates and infection control in the case of diabetic and venous ulcers using nanosilver dressings. However, the clinical evidence is generally inconsistent. The majority of the current literature is hampered by small samples, mixed populations of patients, no standard endpoint measures, or a statistic lack of proper measures to control confounding variables, including ulcer severity, infection status, or peripheral arterial disease^26,27,28.

Moreover, there is limited research regarding the impact of nanosilver dressings on clinically relevant outcomes, e.g., time to complete healing, major amputation rate, or recurrence over time. Although there are case series and several randomized trials that can be positively influenced by something, their external validation is not definite because of the constraining criteria of inclusion and the limited period of follow-up. Practical evidence of treatment performance of large patient groups (especially when control variables of prognostic variables are considered) is still limited. Because DFU healing is influenced by systemic, vascular, infectious, and wound-related factors, comparative real-world studies with explicit control of measured baseline differences may help clarify the association between dressing strategy and clinically relevant healing outcomes^29,30,31,32.

Against these knowledge gaps, we conducted a real-world retrospective propensity score–matched cohort study comparing nanosilver and alginate dressings in patients with diabetic foot ulcers. The primary objective was to determine whether the use of nanosilver dressing was associated with a higher proportion of patients achieving complete wound healing at week 12. Key secondary objectives were to compare early wound-healing responses, complete-healing rates at earlier follow-up points, and time to complete wound closure. Other secondary objectives included comparisons of amputation, reinfection, recurrence, rehospitalization, repeat debridement, and treatment-related safety outcomes. Exploratory subgroup analyses were performed to assess the consistency of the association with 12-week complete healing across selected baseline clinical characteristics.

Protocol

The study was conducted in accordance with the Declaration of Helsinki and reported in accordance with the STROBE recommendations for observational studies. The Ethics Committee of the General Hospital of Northern Theater Command granted ethical approval (approval number NCH202508221). Because of the retrospective design and use of deidentified clinical data, the Ethics Committee waived the requirement for written informed consent. Only deidentified analytic data were used for the study analyses.

Study design and setting

This retrospective cohort study was conducted at the General Hospital of Northern Theater Command, a tertiary referral center in China with a dedicated multidisciplinary program for chronic diabetic wound care. The study period extended from January 2020 through December 2023. Consecutive adult patients with an active diabetic foot ulcer who received either a nanosilver or alginate primary dressing strategy during routine care were screened for eligibility. All patients were managed within the institutional multidisciplinary diabetic-foot care pathway described below.

Patient selection and eligibility criteria

Patients were eligible if they were aged ≥18 years, had type 1 or type 2 diabetes, and had an active diabetic foot ulcer suitable for conservative limb-preserving management. Wagner grade was retained to characterize ulcer depth and tissue involvement, but was not used as a surrogate measure of limb perfusion. Baseline vascular eligibility was assessed using clinical vascular examination and available objective perfusion measurements, including pedal Doppler waveforms, ankle–brachial index, and toe–brachial index or toe pressure when available.

Patients with untreated severe limb ischemia requiring urgent vascular evaluation or revascularization before wound-dressing management were excluded based on the available objective hemodynamic and clinical findings. No single ABI, TBI, or toe-pressure threshold was used as an automatic exclusion criterion. Instead, untreated severe ischemia requiring exclusion was determined from the overall vascular assessment, including objective hemodynamic measurements, pedal Doppler findings, clinical ischemic features, and documentation by the multidisciplinary vascular team that urgent vascular evaluation or revascularization was required before conservative dressing-based wound management. Patients who had undergone revascularization before cohort entry were eligible when post-revascularization assessment indicated adequate perfusion for conservative wound management. Patients requiring immediate major amputation at baseline were also excluded.

Additional exclusion criteria were non-diabetic or non-index ulcers; missing one or more variables required for the primary propensity-score model; insufficient follow-up information to ascertain whether the week-12 primary endpoint had occurred; use of adjunctive biologic or advanced wound therapies inconsistent with the study protocol; and inability to assign a consistent dressing exposure according to the predefined exposure definition. Missing primary-outcome information and insufficient follow-up were treated as a single mutually exclusive category because both prevented reliable ascertainment of the week-12 primary endpoint. Exclusion reasons were assigned hierarchically so that each screened patient was counted only once. Inability to reconstruct every component of the WIfI classification was not, by itself, an exclusion criterion for the primary matched analysis, provided that vascular eligibility, the primary matching variables, and the primary endpoint could be determined reliably. Among the 393 screened patients, seven were excluded for missing data-related reasons. Four patients lacked one or more variables required for the primary propensity-score model, and three had insufficient follow-up information to ascertain the week-12 primary endpoint. These seven patients did not enter the 361-patient unmatched cohort. Incomplete WIfI information was handled separately and was not an exclusion criterion for the primary analysis.

After application of these criteria, 361 patients were included in the unmatched cohort, comprising 121 patients treated with nanosilver dressings and 240 treated with alginate dressings. One-to-one propensity score matching generated 115 matched pairs for the primary comparative analyses.

Vascular assessment and WIfI classification

Baseline vascular status was assessed using pedal pulse examination, pedal Doppler waveforms, ankle systolic pressure, ankle–brachial index, and toe–brachial index or toe pressure when available. When toe pressure could not be obtained, transcutaneous oxygen pressure was used if it had been recorded during routine clinical care. No single vascular test was considered sufficient to characterize limb perfusion in isolation.

For an additional complete-case sensitivity analysis, the Wound, Ischemia, and foot Infection classification was retrospectively reconstructed using baseline information recorded before assessment of the study outcomes. The wound component was assigned according to ulcer depth, tissue involvement, and gangrene extent; the ischemia component was assigned using the available objective perfusion measurements; and the foot-infection component was assigned using the IWGDF/IDSA infection criteria.

Complete information for all three WIfI components was available for 108 of the 115 matched pairs. These 108 pairs constituted the WIfI-complete sensitivity cohort. The remaining seven matched pairs had incomplete information for one or more WIfI components but had sufficient data for vascular eligibility assessment, the primary propensity-score model, and primary-outcome ascertainment. They were therefore retained in the 115-pair primary analysis and excluded only from the WIfI sensitivity analysis. Two investigators independently reconstructed the WIfI component grades using baseline records, with disagreements resolved by consensus with a third investigator. Outcome information was not used during WIfI reconstruction.

Treatment protocols

Treatment allocation was not randomized. The choice between nanosilver and alginate dressings was made by the treating multidisciplinary wound-care team as part of routine clinical practice, based on the overall clinical assessment of the wound and the available treatment options. No prospective randomization procedure or mandatory institutional allocation algorithm required a patient to receive one dressing rather than the other. The retrospective medical records did not contain a standardized field documenting a single reason for dressing selection in every patient. Therefore, although treatment decisions were made within a common multidisciplinary care pathway, clinical judgment and other patient- or practice-related factors may have influenced dressing selection, creating the potential for confounding by indication.

All patients were managed within the same institutional multidisciplinary diabetic-foot care pathway, which specified the core components of wound-bed preparation, debridement, infection management, vascular assessment, offloading, glycaemic optimization, and follow-up scheduling. This common pathway was intended to reduce major differences in routine co-interventions between treatment groups. However, individual provider identity, provider-level protocol-fidelity measures, and quantitative indicators of wound-care quality or consistency were not systematically captured in the electronic health records and therefore could not be incorporated into the propensity-score model.

Multidisciplinary management was provided in accordance with the institutional diabetic-foot pathway and contemporary International Working Group on the Diabetic Foot recommendations. Necrotic tissue, devitalized tissue, and surrounding callus were sharply debrided when clinically indicated to obtain an adequately prepared wound bed. Vascular status was evaluated using clinical examination together with available objective perfusion measurements, including ankle–brachial index, pedal Doppler waveforms, ankle systolic pressure, and toe-based measurements when available. Patients requiring vascular intervention were assessed by the multidisciplinary team, and revascularization was performed when clinically indicated before or during the wound-care pathway according to vascular status.

Systemic antibacterial therapy was prescribed for ulcers with clinical evidence of infection and was not determined by dressing assignment. Ulcers without clinical evidence of infection were not routinely treated with systemic antibiotics. The antibiotic-management framework was standardized across the two dressing groups, but the specific antimicrobial agent, route of administration, and treatment duration were individualized rather than fixed. Initial empirical therapy, when required before microbiological results were available, was selected according to infection severity, previous antimicrobial exposure, available previous microbiological information, drug-allergy history, renal function, and routine institutional antimicrobial practice.

When clinically indicated, wound specimens were obtained for microbiological culture and susceptibility testing, preferably before or shortly after initiation of systemic therapy. Once microbiological results became available, antimicrobial treatment could be narrowed, broadened, or otherwise modified according to organism identification, susceptibility findings, infection evolution, renal function, treatment tolerance, and clinical response. Thus, both groups were treated according to the same antibiotic decision framework, although no single antibiotic regimen or mandatory duration was applied uniformly to all patients. Baseline systemic antibiotic exposure, initial intravenous therapy, empirical treatment, culture-guided modification, antimicrobial escalation or de-escalation, and total systemic antibiotic duration were retrospectively abstracted from medication and microbiological records for the antibiotic-related analyses.

After standardized wound-bed preparation, patients in the nanosilver group received ACTICOAT FLEX 3 Antimicrobial Barrier Dressing (Smith+Nephew), a nanocrystalline silver antimicrobial dressing. Dressings were generally changed every 48–72 h, with the interval adjusted according to exudate burden and the condition of the wound bed. Patients in the alginate group received calcium alginate dressings after the same wound-bed preparation and according to a comparable dressing-change schedule based on exudate and clinical assessment. Treatment exposure for the comparative analysis was defined according to the documented primary dressing strategy used consistently during the first 4 weeks after cohort entry.

Both groups received the same core supportive management, including glycaemic optimization, pressure offloading, infection treatment when indicated, vascular management, and scheduled wound reassessment. Offloading was provided using total-contact casting, removable or customized offloading devices, or custom orthopedic footwear according to ulcer location, limb condition, tolerance, and routine clinical assessment. Because actual longitudinal adherence to prescribed offloading was not systematically quantified in the electronic records, adherence could not be incorporated as a measured baseline covariate.

Adjunctive biologic or advanced wound therapies that were inconsistent with the study protocol, including negative-pressure wound therapy, platelet-rich plasma, stem-cell therapy, and other advanced regenerative interventions, were not permitted as concomitant index-ulcer treatments during the defined comparative treatment period; patients receiving such therapies were excluded according to the eligibility criteria. Routine use of additional topical antimicrobial or antiseptic agents outside the assigned dressing strategy was also avoided to reduce contamination between dressing groups.

Clinical assessments and outcome definitions

Baseline assessments and follow-up evaluations at weeks 4, 8, and 12, or at the time of complete healing, were performed independently by two wound-care specialists. Because nanosilver and alginate dressings were visually distinguishable and treatment information could be apparent from the clinical records, the assessors were not considered blinded to treatment allocation. To reduce subjectivity in outcome assessment, all evaluations followed prespecified outcome definitions and a standardized wound-imaging protocol. Wound area was measured by calibrated digital planimetry using standardized photographs and ImageJ software (National Institutes of Health, Bethesda, MD, USA). Baseline C-reactive protein (CRP) was defined as the value measured closest to cohort entry and before or on the index date.

The single primary endpoint was complete wound healing of the index ulcer at week 12. Complete wound healing was defined as full epithelialization of the index ulcer without drainage and without the need for wound dressing, confirmed at two consecutive assessments conducted at least two days apart.

Key secondary healing outcomes were the proportion of patients achieving at least 50% reduction in wound area at week 4, the proportion achieving complete wound healing at weeks 4 and 8, and time to complete healing. Time to complete healing was defined as the number of days from cohort entry to the first assessment documenting complete epithelialization that was subsequently confirmed at the next assessment. For consistency across the longitudinal healing analyses, complete healing at weeks 4, 8, and 12 was derived from the same patient-level time-to-event record using the corresponding 28-, 56-, and 84-day landmarks.

Other secondary clinical outcomes included minor amputation, defined as amputation below the ankle; major amputation, defined as amputation at or above the ankle; reinfection within 12 weeks according to the International Working Group on the Diabetic Foot/Infectious Diseases Society of America criteria; recurrence after complete healing; rehospitalization within three months after treatment initiation; and repeat debridement. Safety outcomes included local allergic reactions, worsening pain after dressing application, excessive wound exudation, a decline in estimated glomerular filtration rate of more than 25%, and a serum silver concentration above 10 µg/L.

Subgroup analyses of the primary endpoint were considered exploratory and examined the consistency of the association between dressing strategy and 12-week complete healing across selected baseline clinical characteristics. Follow-up completeness was assessed separately in weeks 4, 8, and 12. The primary endpoint was considered ascertainable when the medical record documented either complete healing before or at week 12 or persistence of an unhealed index ulcer at the week-12 assessment. Patients whose available records were insufficient to establish either status were classified as having insufficient follow-up and were excluded before cohort assembly. All 230 patients subsequently included in the primary propensity score–matched cohort had ascertainable week-12 primary-endpoint data. No additional loss to follow-up occurred after formation of the matched cohort.

Propensity-score matching

To reduce measured baseline confounding arising from the non-random selection of dressing strategies, propensity scores representing the probability of receiving nanosilver dressings were estimated using logistic regression. The primary propensity-score model included age, diabetes duration, HbA1c, Wagner grade, infection status, peripheral arterial disease/ankle–brachial index, and baseline ulcer area. These variables were selected because they were routinely recorded for the full primary cohort and were clinically associated with dressing selection or wound-healing prognosis.

One-to-one nearest-neighbor matching without replacement was performed using a caliper width of 0.2 standard deviations of the logit of the propensity score. Covariate balance was evaluated using standardized mean differences, with an absolute standardized mean difference <0.10 indicating adequate balance. For the seven prespecified covariates included in the primary propensity-score model—age, diabetes duration, HbA1c, Wagner grade, infection status, peripheral arterial disease/ankle–brachial index, and baseline ulcer area—absolute standardized mean differences were calculated both before and after matching. These values were displayed in the Love plot to quantify the improvement in covariate balance.

To provide a broader assessment of residual baseline imbalance, post-matching standardized mean differences were additionally calculated for all other available baseline characteristics reported in Table 1, even when those variables had not been included in the propensity-score model. These additional SMDs were used as balance diagnostics only and did not alter the original propensity-score model, matching algorithm, or composition of the 115 matched pairs. An absolute SMD <0.10 was considered indicative of minimal residual imbalance.

Propensity-score overlap was assessed before and after matching. This procedure generated 115 matched pairs, which formed the primary analytic cohort. The WIfI component grades were not included in the primary propensity-score model because complete WIfI information was unavailable for seven matched pairs. Instead, the WIfI components were incorporated in an additional complete-case sensitivity analysis restricted to the 108 matched pairs with fully reconstructable WIfI data.

Sample size justification

Because this was a retrospective cohort study, no prospective sample-size calculation was used to determine enrollment. All consecutive patients meeting the predefined eligibility criteria during the study period were considered for inclusion, and the final analytic sample was determined by the number of eligible patients available within the study period and by successful propensity score matching.

To provide a supplementary, clinically oriented assessment of sample-size adequacy for the single primary endpoint, we considered an absolute between-group difference of 20 percentage points in the week-12 complete-healing rate to be clinically meaningful. Assuming complete-healing rates of 55% in the alginate group and 75% in the nanosilver group, a two-sided α level of 0.05, 80% statistical power, and a 1:1 group ratio, a continuity-corrected comparison of two proportions indicated that approximately 94 patients per group would be required. After allowing for approximately 10% potential loss due to non-analyzable records or unsuccessful matching, a minimum of 105 patients per group was considered desirable. The final primary matched cohort included 115 patients per group and therefore exceeded this supplementary adequacy threshold.

Because reliable external estimates of the expected discordant-pair proportions were unavailable, the calculation used a two-proportion framework as a pragmatic benchmark rather than a matched-pair McNemar sample-size calculation. This supplementary assessment was not used to determine enrollment and should not be interpreted as a prospective sample-size calculation or as a post hoc observed-power analysis.

The sample-size assessment was intended only to evaluate the adequacy of the primary matched cohort for the single primary endpoint of complete wound healing at week 12. The study was not specifically powered for sparse secondary outcomes, safety events, subgroup interactions, or the WIfI-complete sensitivity analysis.

Statistical analysis

Continuous variables were summarized as mean ± standard deviation or median with interquartile range, according to their distribution, and categorical variables as number and percentage. Between-group descriptive comparisons used Student’s t test or the Mann–Whitney U test for continuous variables and Pearson’s χ² test or Fisher’s exact test for categorical variables, as appropriate. In the propensity score–matched cohort, baseline balance was assessed primarily using standardized mean differences rather than hypothesis-testing P values.

The primary analysis included all 115 matched pairs and compared complete wound healing at week 12 between the nanosilver and alginate groups. Binary outcomes were analyzed using methods that accounted for the matched-pair structure. Conditional logistic regression was used to estimate matched odds ratios (ORs) with 95% confidence intervals (CIs). Matched relative risks (RRs) were estimated using modified Poisson regression with robust sandwich standard errors clustered by matched pair, and absolute risk differences with 95% CIs were estimated using nonparametric bootstrap resampling at the matched-pair level. McNemar’s test was used for paired comparisons of common binary outcomes. For sparse outcomes or highly unbalanced discordant pairs, including major amputation and other uncommon safety events, exact conditional methods and the exact McNemar test were used.

Time to complete wound healing was summarized using Kaplan–Meier methods. The median time to complete healing was defined as the time at which the Kaplan–Meier estimate of cumulative complete healing first reached 50%. Hazard ratios (HRs) with 95% confidence intervals (CIs) and nominal P values were estimated using a Cox proportional hazards model with robust standard errors clustered by matched pair. Patients who had not achieved complete healing by day 84 were censored at day 84.

Three sensitivity analyses were performed for the single primary association. First, an extended conditional logistic regression model stratified by matched pair was fitted in all 115 matched pairs and additionally adjusted for body mass index, serum albumin, calendar year of treatment, ulcer duration, baseline CRP, and baseline systemic antibiotic use. These covariates were selected as parsimonious measures of physiological and nutritional status, temporal changes in clinical practice, ulcer chronicity, baseline inflammatory burden, and concomitant systemic antimicrobial treatment. The original propensity-score model and composition of the 115 matched pairs were retained unchanged.

Systemic antibiotic management was additionally examined as a potentially important concomitant intervention. Baseline systemic antibiotic use, empirical treatment, initial intravenous therapy, culture-guided modification, antimicrobial escalation or de-escalation, antimicrobial class, microbiological findings, and total treatment duration were summarized descriptively by dressing group. Exploratory analyses examined 12-week complete healing according to baseline systemic antibiotic exposure and tested a dressing-strategy-by-antibiotic-use interaction. Total antibiotic duration was not entered as a regression covariate because it accrued during follow-up and could itself have been influenced by infection resolution, clinical deterioration, or wound-healing trajectory. These analyses were intended to assess robustness and co-intervention patterns rather than the causal efficacy of antibiotic therapy.

Second, the primary matched analysis was repeated among the 108 matched pairs with complete WIfI information and was then additionally adjusted for the WIfI wound, ischemia, and foot-infection grades. The seven pairs with incomplete WIfI information remained in the primary 115-pair analysis and were excluded only from this complete-case sensitivity analysis.

Third, two extreme case analyses were performed for the three otherwise eligible patients whose week-12 primary outcome could not be ascertained, using the exposure-classified cohort with complete primary baseline covariates. In one scenario, the missing nanosilver outcome was classified as unhealed, and the two missing alginate outcomes healed; in the opposite scenario, these assignments were reversed. Treatment associations were re-estimated using multivariable logistic regression adjusted for the seven variables included in the primary propensity-score model.

Missing data were evaluated separately for variables required for the primary propensity-score model, the week-12 primary endpoint, additional baseline characteristics, and WIfI components. Four of the 393 screened patients lacked one or more variables required for the primary propensity-score model, and three had insufficient follow-up information to ascertain the week-12 primary endpoint; these seven patients were excluded before formation of the 361-patient unmatched cohort. Patients excluded for missing data-related reasons were compared descriptively with included patients using available baseline characteristics and standardized mean differences, without inferential P values because of the small, excluded group. No imputation was performed for the primary matched analysis.

To address multiplicity among secondary outcomes, a post hoc Holm step-down procedure was applied within clinically defined outcome families. The single primary endpoint of complete wound healing at week 12 was not multiplicity-adjusted and was evaluated using a two-sided α level of 0.05. The key secondary healing family comprised complete healing at weeks 4 and 8, at least 50% wound-area reduction at week 4, and time to complete healing. Major amputation, minor amputation, reinfection, recurrence after healing, rehospitalization, and repeat debridement formed a separate secondary clinical-outcome family, and safety outcomes were adjusted as an additional family. Both nominal and Holm-adjusted P values were reported for these secondary analyses. The extended measured-confounding, extreme-case, and WIfI-complete analyses were treated as sensitivity analyses of the primary association and were not included in the secondary-outcome multiplicity families. Antibiotic-stratified analyses were exploratory and were likewise not included in these multiplicity families.

Exploratory subgroup analyses of the primary endpoint were performed according to Wagner grade, infection status, peripheral arterial disease, HbA1c level, and ulcer duration. Treatment-effect heterogeneity was assessed using interaction terms, and the Holm procedure was applied across the interaction P values. Interpretation was based on the interaction tests rather than on statistical significance within individual subgroup strata. All tests were two-sided. Effect estimates and 95% CIs were emphasized when interpreting the magnitude and precision of associations, and secondary subgroup, and sensitivity analyses were regarded as supportive or exploratory rather than confirmatory.

Results

Patient characteristics

Between January 2020 and December 2023, 393 patients with diabetic foot ulcers (DFUs) were screened for eligibility, and cohort assembly is shown in Figure 1. Thirty-two patients were excluded before formation of the unmatched cohort: 8 had a non-diabetic or non-index ulcer, 4 lacked baseline variables required for the primary propensity-score model, 3 had insufficient follow-up to ascertain the week-12 primary endpoint, 6 ...

Discussion

The principal finding of this real-world retrospective propensity score–matched study was that nanosilver dressing use was associated with a higher proportion of complete wound healing at week 12 than alginate dressing use. The primary matched estimate was supported by the extended measured-confounding and WIfI-complete sensitivity analyses. Among the key secondary healing outcomes, complete healing at week 8, at least 50% wound-area reduction at week 4, and time to complete healing remained supportive after Holm a...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
0.9% Sodium Chloride Injection	Sichuan Kelun Pharmaceutical Co., Ltd.	N/A	Used for wound cleansing and routine clinical care
3M Scotchcast Plus Casting Tape	3M Health Care	82003	Fiberglass casting tape used for total-contact casting when clinically indicated
Accu-Chek Inform II Blood Glucose Monitoring System	Roche Diagnostics	Accu-Chek Inform II	Used for routine glucose monitoring during diabetic-foot care
ACTICOAT FLEX 3 Antimicrobial Barrier Dressing	Smith+Nephew	66800396	Nanocrystalline silver antimicrobial dressing used as the primary nanosilver dressing; changed every 48–72 h according to exudate and wound-bed condition
Agilent 7900 ICP-MS	Agilent Technologies	7900 ICP-MS	Used for serum silver ion measurement when clinically indicated
BD BBL CultureSwab	BD	220099	Used for microbiological sampling when clinically indicated
Canon EOS 80D digital camera	Canon Inc.	EOS 80D	Used for standardized serial wound photography
DARCO OrthoWedge Forefoot Offloading Shoe	DARCO International	OrthoWedge	Used for forefoot pressure redistribution and offloading when clinically indicated
Disposable surgical blade, No. 15	Kai Industries Co., Ltd.	B15	Used during sharp wound debridement
Disposable wound measuring ruler	Jiangsu Kangjin Medical Instrument Co., Ltd.	N/A	Used as a calibration reference for wound-size measurement in photographs
Handheld vascular Doppler system	Huntleigh Healthcare	Dopplex DMX	Used for ankle–brachial index and peripheral perfusion assessment
ImageJ software	National Institutes of Health	Version 1.53t	Used for calibrated digital planimetry and wound-area measurement
IWGDF/IDSA infection criteria	International Working Group on the Diabetic Foot/Infectious Diseases Society of America	2023 guideline	Used to define diabetic foot infection and reinfection
KALTOSTAT Calcium Sodium Alginate Dressing	ConvaTec	168212	Calcium sodium alginate primary dressing used as the comparator after standardized wound-bed preparation
MatchIt package	R package	Version 4.5.5	Used for 1:1 nearest-neighbor propensity-score matching
Neusoft Hospital Information System / Electronic Medical Record platform	Neusoft Medical Systems	N/A	Institutional HIS/EMR deployment; Used for retrospective extraction of clinical and outcome data
R software	R Foundation for Statistical Computing	Version 4.3.2	Used for statistical analyses, propensity-score matching, and time-to-event analyses
Resona 7 Diagnostic Ultrasound System	Mindray	Resona 7	Used for Doppler ultrasonography of lower-limb perfusion
Roche cobas c 702 Automated Chemistry Analyzer	Roche Diagnostics	cobas c 702	Used for serum creatinine, lipid profile, and routine biochemical tests
Sterile gauze and absorbent pads	Winner Medical Co., Ltd.	N/A	Used for routine wound cleaning and dressing support
Sterile surgical debridement kit	Jinhuan Medical Products Co., Ltd.	JH-SDK-01	Used for sharp debridement of necrotic tissue and callus
survival package	R package	Version 3.5	Used for Kaplan–Meier and Cox proportional hazards analyses
Sysmex XN-1000 Automated Hematology Analyzer	Sysmex Corporation	XN-1000	Used for routine blood-count assessment
Tosoh HLC-723G8 HbA1c Analyzer	Tosoh Corporation	HLC-723G8	Used for baseline glycated hemoglobin measurement
VITEK 2 Compact	bioMérieux	VITEK 2 Compact	Used for culture-based pathogen identification and antimicrobial susceptibility testing
Wagner ulcer classification	Not applicable	Not applicable	Clinical grading system used to characterize diabetic foot ulcer depth and tissue involvement