Wound healing, wound-free period and amputation-free survival in patients with Wound, Ischemia, and foot Infection (WIfI) stage 3 and 4 disease

Abstract

Background:

Recent publications in low-risk Wound, Ischemia, and foot Infection (WIfI) stage 1 to 2 chronic limb-threatening ischemia (CLTI) suggest wound care-first strategies are safe. Although revascularization is widely accepted as beneficial in high-risk WIfI stage 3 to 4 patients, its impact on wound-centered outcomes such as wound healing time (WHT) and wound-free period (WFP) has not been well-characterized.

Methods:

A two-center, retrospective analysis was performed. Baseline data included demographics, comorbidities, perfusion indices, WIfI, and treatment modality (endovascular, open, hybrid, wound care only). Outcomes were WHT, WFP, and amputation-free survival. WHT was defined as the interval from initial presentation to complete epithelialization, regardless of interim reinterventions; major amputation or death prior to healing were classified as unhealed episodes. WFP was defined as the healed interval until recurrence or censoring.

Results:

Over 9 years, 1483 patients were managed with CLTI, of whom 920 (1091 limbs) were WIfI stage 3 and 4 disease. These contributed 1158 wounds episodes (median follow-up, 29.0 months; interquartile range [IQR], 10.0–45.0 months). Most limbs underwent revascularization (n = 945; 81.6%), including endovascular (n = 713; 62%), open (n = 196; 17%), and hybrid (n = 36; 3%) approaches, whereas 18.4% were managed with wound care only. Overall, 550 wound episodes (47.6%) healed, with a median WHT of 248 days (IQR, 120-441 days). The median WHT did not differ meaningfully by revascularization strategy. Successful revascularization was independently associated with a higher likelihood of healing (adjusted odds ratio, 2.45; 95% confidence interval [CI], 1.71-3.50). Independent predictors of healing included successful revascularization (hazard ratio [HR], 2.45; 95% CI, 1.71-3.50; P < .001), obesity (HR, 1.52; P = .003), and hyperlipidemia (HR, 1.55; P = .044), whereas WIfI stage 4 predicted decreased healing (HR, 0.61; P < .001). Among 550 healed episodes, 79 (14.4%) developed recurrence or underwent major amputation at a median of 502 days (IQR, 163-1084 days). The median WFP was 502 days (95% CI, 350-654 days) and did not differ by treatment strategy on Kaplan-Meier analysis (P = .41). At last follow-up, 720 patients (62.2%) experienced major amputation/death. One-year limb salvage, survival, and amputation-free survival were consistently lower in the wound care-only group, whereas revascularized patients demonstrated more favorable early outcomes.

Conclusions:

In high-risk CLTI (WIfI 3-4), wound healing is prolonged (>6 months), and fewer than one-half of wound episodes heal. Revascularization was strongly associated with wound healing. Once healing occurred, durability (WFP) was similar across strategies. These findings highlight the value of wound-centered metrics (WHT and WFP) to complement traditional survival-based endpoints in this population

Chronic limb-threatening ischemia (CLTI) comprises the most severe chronic manifestation of peripheral artery disease.^1,2 Multidisciplinary care consisting of wound care, podiatry, and revascularization represents the modern gold standard of CLTI management.^1–3 In particular, this flexible, multi-modal care manages the wound, infection, medical comorbidities, and social complexities that frequently intersect with vascular insufficiency.

Ironically, frequently utilized outcome metrics in CLTI, such as amputation-free survival (AFS), fail to accurately quantify the morbidity associated with these various treatment strategies, and particularly, wound management.^4,5 Efforts to measure comparative and cost effectiveness are blunted due to the inability of traditional outcome metrics to capture patient-centric values such as wound healing. Moreover, our ability to guide shared decision-making with the patients and caregivers is impaired by the omission of wound-healing metrics in CLTI.^5–7 Especially important is the morbidity associated with wound management, as the wound severity is statistically significantly associated with inferior outcomes in CLTI.⁵ Brief Delphi consensus efforts to clarify patient-reported outcome measures in CLTI, as well as prior publications, ratify the consensus surrounding the inclusion of wound-related metrics in the evaluation of CLTI outcomes.^7–9

Wound healing time (WHT) and wound-free period (WFP) are emerging as reliable and precise patient-centric metrics in CLTI.^8,9 Inclusion of these metrics suggest that personalization of revascularization-first vs wound care-first strategies is possible and may result in similar efficacy.^9,10 Yet, prior series utilizing these metrics are limited to CLTI populations with milder risk profiles for major amputation,¹⁰ as stratified by the Wound, Ischemia, and foot Infection (WIfI) staging system.¹¹ Other trials describing the similarity between wound care-only vs revascularization-first strategies in CLTI were limited to small single-center retrospective cohorts without external validation.⁹ Although the benefit of revascularization in WIfI stage 3 and 4 CLTI is well-established, the wound-centered consequences of different treatment strategies—particularly with respect to WHT and WFP—remain incompletely characterized in the literature,⁶ limiting the ability of practitioners to personalize wound care with or without revascularization strategies in across the spectrum of CLTI presentations.

In this context, we aimed to quantify the WHT and WFP in a large, two-center, retrospective cohort of patients with WIfI stage 3 and 4 CLTI. We sought to quantify WHT and WFP in WIfI stage 3 and 4 CLTI and to examine how these metrics vary by treatment strategy. Our primary outcome metric was WHT. Secondarily, we describe the WFP, wound recurrence, AFS, limb salvage, and mortality.

METHODS

Study design.

We retrospectively studied all patients with WIfI stage 3 and 4 CLTI presenting to the Vascular Surgery services at two quaternary referral centers. We included patients with CLTI from August 1, 2014, through January 1, 2022. Institutional review board (H-50,772 & 18-1153) exemption was obtained at both sites, and due to the retrospective design, impracticality of obtaining consent, and minimal risk to patients, informed consent was waived at both sites.

Patient population.

Consistent with current guidelines^1,2 and prior publications,^9,10 we included consecutive adults (age ≥18 years) presenting with CLTI as defined by ≥2 weeks of ischemic rest pain and/or ulceration/gangrene with objective evidence of ischemia (ankle-brachial index [ABI] <0.4; ankle pressure <50 mm Hg, toe pressure [TP]<30 mm Hg, or flat pulse volume waveforms. Because the focus of this study was wound healing, only patients with WIfI stage 3 or 4 disease presenting with tissue loss (ulceration or gangrene) were included.^1,11 Patients with ischemic rest pain alone without a wound were excluded. All included patients had objective evidence of ischemia (abnormal ABI, TP, or pulse volume recordings); patients with purely nonischemic wounds (eg, neuropathic or pressure ulcers without ischemia) were excluded. Patients were also excluded if ischemia was due to nonatherosclerotic etiologies (such as vasculitis, Buerger’s disease, or radiation arteritis), venous ulceration, acute limb ischemia, trauma, or intermittent claudication only. Wound care included debridement, offloading, serial dressing changes, negative-pressure therapy, biologic grafts, secondary closures, and split-thickness skin grafting as clinically indicated, within a multidisciplinary care framework. Patients were offered revascularization if there was evidence of ischemia grade 3 (TP <30 mm Hg). A wound care-only treatment strategy was provided for those who either refused revascularization or were deemed unreconstructable and preferred wound care only instead of major amputation.

Data collection and treatment groups.

Patient-level variables included age, sex, race/ethnicity, and comorbidities (diabetes, obesity, hyperlipidemia, hypertension, coronary artery disease, myocardial infarction, congestive heart failure, cerebrovascular disease, chronic kidney disease, and end-stage renal disease [ESRD]). Limb-level measures included ABI, TP, toe-brachial index (TBI), and WIfI classification (subscores and overall stage). Procedural data included attempted and achieved revascularization, categorized as endovascular, open, hybrid, or wound care only. Episodes were categorized according to the index strategy that ultimately resulted in durable treatment. Failed endovascular attempts without subsequent revascularization remained in the wound care-only group, whereas failed endovascular attempts followed by successful open revascularization were classified in the open group. Patients were not reclassified based on subsequent loss of patency.

All revascularizations (open, endovascular, or hybrid) were performed at the discretion of board-certified vascular surgeons at their respective institutions. Similarly, podiatrists performed partial foot amputations (digital, transmetatarsal, Lisfranc, or Symes) at their discretion after a multidisciplinary discussion. Wound care procedures also included debridement, application of biologic dressings and grafts, secondary wound closures, vacuum-assisted closure, and split thickness grafts that also performed either by the vascular or podiatric surgeons at either institution.

Because the patients were at high risk of major amputation (above the knee or below the knee), the decisions regarding revascularization and/or wound care were performed in a multidisciplinary fashion. Reinterventions were defined as a subsequent ipsilateral vascular or podiatric procedures and were decided in a multidisciplinary fashion. There was not a uniform protocol for each patient across institutions, but standard practices of each team of vascular surgeons and podiatrists at each site are similar and attempt to be individually tailored for each patient. Patients were categorized according to the index revascularization strategy. The first group (wound care only) included those who did not undergo revascularization or had failed endovascular attempts resulting in no revascularization; these were considered conservatively managed patients who received medical, wound, and podiatric care only. The second group underwent endovascular revascularization, including balloon angioplasty, stenting, and/or atherectomy. The third group underwent open surgical revascularization, including bypass grafting or endarterectomy. The fourth group underwent hybrid procedures, defined as combined open and endovascular interventions performed during the same hospitalization or within 30 days.

Outcomes.

A wound episode was defined as the period from presentation of an ischemic wound until complete healing, major amputation, or death. Major amputation or death prior to healing were treated as competing events and classified as unhealed episodes.⁹ Overlapping wounds in the same limb were considered part of the same episode, whereas new nonoverlapping wounds were analyzed separately. Episodes were classified according to the strategy that ultimately resulted in durable treatment; failed endovascular attempts followed by successful open revascularization were classified in the open group, whereas episodes without successful revascularization remained in the wound care-only group. WHT was measured from initial presentation to complete epithelialization, regardless of interim reinterventions, consistent with prior publications.^8,9 The WFP was analyzed only among episodes that achieved complete healing and was defined as the interval from wound healing to the development of a new, nonoverlapping wound or major amputation, or censoring at last follow-up. WFP time was calculated in days from the wound healing date to the first recurrence event or censoring.

Traditional metrics of AFS, limb salvage, and mortality were also reported.^1,4 Briefly, AFS is defined as the composite duration of survivorship without either a major amputation or death. Limb salvage is the duration of time without a major amputation. Mortality is the duration of survivorship, regardless of the status of the limb.

Statistical analysis.

Continuous variables were summarized as median (interquartile range [IQR]) and compared using the Kruskal-Wallis test; categorical variables were compared using χ² or Fisher exact tests. Predictors of wound healing were assessed with logistic regression. WFP was analyzed using Kaplan-Meier methods with censoring at last follow-up to account for variable follow-up duration. Predictors of WFP were explored using multivariable linear regression. Candidate variables included age, sex, race/ethnicity, ESRD, WIfI ischemia and infection grades, revascularization type, and wound healing duration. Model fit was assessed using R², adjusted R², the F-test, and collinearity diagnostics. Predictors of wound recurrence were evaluated using Kaplan-Meier curves and Cox proportional hazards models. AFS was examined with Kaplan-Meier analysis and multivariable Cox regression, stratified by center. Hazard ratios (HRs), odds ratios (ORs), and regression coefficients (β) were reported with 95% confidence intervals (CIs). Statistical significance was defined as P < .05. Analysis was performed in IBM SPSS statistics version 31 (IBM Corp).

RESULTS

Study population and baseline characteristics.

Over 9 years, a total of 1483 patients with CLTI were screened. After excluding those with WIfI stage 1 and 2 or indeterminate staging, 920 patients (median age, 66 years; IQR, 58-75 years; 62% male) with WIfI stage 3 and 4 disease remained, contributing 1091 limbs and 1158 nonoverlapping wound episodes across two high-volume centers. The median follow-up was 29.0 months (IQR, 10.0-45.0 months).

Baseline demographic and comorbidity characteristics and limb characteristics are summarized in Tables I and II. The most frequent baseline comorbidities included a history of tobacco abuse, hyperlipidemia, hypertension, and coronary artery disease. Approximately one-half of the wounds at presentation were WIfI Wound grade 3. Most limbs presented with a WIfI Ischemia grade 3. Similarly, more the three-fourths of limbs were WIfI clinical stage 4 at presentation.

Table I.

Baseline demographic and comorbidity characteristics of patients with Wound, Ischemia, and foot Infection (WIfI) stage 3 and 4 chronic limb-threatening ischemia (CLTI), stratified by initial treatment strategy

Variable	Wound care only		Endovascular		Open		Hybrid		All		P1 (all)	P2 (endovascular vs open)
	n = 164		n = 561		n = 165		n = 30		n = 920
Age, years	69.0 (58-80)		66 (58-75)		66 (59-74)		69 (61-73)		66.4 (58-75)		.55	.77
Male sex	94	57.3	345	61.5	113	68.5	17	56.7	569	61.8	.18	.18
Non-Hispanic ethnicity	139	84.8	399	71.1	129	78.2	25	83.3	692	75.2	<.001	.04
Race
White	75	45.7	343	61.1	98	59.8	21	70.0	537	58.4	.12	.88
Black	72	43.9	193	34.5	58	35.2	7	23.3	330	35.9	.01	.74
Asian	0	0.0	8	1.4	5	3.0	1	3.3	14	1.5	.13	.15
Native	1	0.6	3	0.5	0	0.0	0	0.0	4	0.4	.77	.36
Unable to determine	16	9.8	14	2.5	4	2.4	1	3.3	35	3.8	<.001	.99
History of tobacco abuse	99	60.4	331	59.2	129	78.2	27	90.0	586	63.8	<.001	<.001
Obesity	36	22.0	162	29.0	35	21.2	5	16.7	238	26.0	.06	.05
Hyperlipidemia	102	63.0	513	91.4	152	92.7	30	100.0	797	86.9	<.001	.6
Hypertension	143	87.2	550	98.0	157	95.7	29	96.7	879	95.6	<.001	.09
Diabetes	62	37.8	432	77.0	95	57.6	23	76.7	612	66.5	<.001	<.001
Coronary artery disease	89	54.3	392	69.9	122	74.4	22	73.3	625	68.0	<.001	.26
Myocardial infarction	44	26.8	177	31.6	52	31.5	14	46.7	287	31.2	.18	.1
Congestive heart failure	73	44.5	287	51.2	77	46.7	13	43.3	450	48.9	.37	.3
Stroke/CVA	74	45.1	190	33.9	62	37.6	15	50.0	341	37.1	.03	.38
Chronic kidney disease	82	50.0	371	66.1	83	50.3	13	43.3	549	59.7	<.001	<.001
End-stage renal disease	38	23.2	207	36.9	40	24.2	6	20.0	291	31.6	<.001	.003
VTE	32	19.6	107	19.1	33	20.1	4	13.3	176	19.2	.86	.78
COPD	48	29.3	131	23.4	51	30.9	5	16.7	235	25.6	.09	.05

COPD, Chronic obstructive pulmonary disease; CVA, cerebrovascular accident; Endo, endovascular; ESRD, end-stage renal disease; VTE, venous thromboembolism.

Data are presented as number (%) or median (interquartile range).

P1, Comparison among all treatment groups using χ² or Kruskal-Wallis test.

P2, Pairwise comparison between endovascular and open cohorts.

Table II.

Wound, Ischemia, and foot Infection (WIfI) grades at initial wound presentation (baseline episode per limb)

Variable	Wound care only		Endovascular		Open		Hybrid		All		P1 (all)	P2 (endovascular vs open)
No.	180		682		193		36		1091
WIfI wound grade
1	24	13.3	216	31.7	56	29.0	14	38.9	310	28.4	<.001	.48
2	65	36.1	359	52.6	89	46.1	17	47.2	530	48.6	<.001	.11
3	90	50.0	107	15.7	48	24.9	5	13.9	250	22.9	<.001	.003
WIfI ischemia grade
0	1	0.6	3	0.4	0	0.0	0	0.0	4	0.4	008	.36
1	24	13.5	57	8.4	7	3.6	2	5.6	90	8.2	.29	.03
2	19	10.7	70	10.3	14	7.3	1	2.8	104	9.5	.002	.21
3	134	75.3	548	80.8	171	89.1	33	91.7	886	81.2	<.001	.008
WIfI foot infection grade
0	157	87.7	473	69.4	150	77.7	22	61.1	802	73.5	<.001	.02
1	7	3.9%	112	16.4	27	14.0	10	27.8	156	14.3	<.001	.41
2	14	7.8%	82	12.0	14	7.3	4	11.1	114	10.4	.15	.06
3	1	0.6%	15	2.2	2	1.0	0	0.0	18	1.7	.3	.30
WIfI clinical stage
3	42	23.3	273	40.0	64	33.2	15	41.7	394	36.1	<.001	.08
4	138	76.7	409	60.0	129	66.8	21	58.3	697	63.9	<.001	.08

Includes 1091 limbs (first wound episode per limb).

P1, Comparison across all groups.

P2, Pairwise comparison between endovascular and open revascularization.

Treatment groups.

Because wound management was analyzed at the episode level, procedural patterns are reported per wound episode (n = 1158) rather than per patient. Across all episodes, 183 (15.8%) had no revascularization attempt, whereas 973 episodes (84.2%) underwent a revascularization attempt based on the treating surgeon’s clinical judgment. Endovascular therapy accounted for the majority of all episodes (713 episodes; 61.7%), followed by open surgical revascularization (196 episodes; 17.0%) and hybrid procedures (36 episodes; 3.1%). Among endovascular attempts, 28 (2.4%) were technical failures. When combining episodes without a revascularization attempt and those with unsuccessful endovascular attempts, 213 wound episodes (18.4%) were ultimately managed with wound care only, whereas 945 episodes (81.6%) achieved durable arterial revascularization. Wounds episodes managed with wound care only more frequently belonged to patients of non-Hispanic ethnicity, Black race, with more advanced wounds, but a lower prevalence of diabetes, hyperlipidemia, hypertension, and renal disease (Table I). Sex was not statistically significantly different across the different revascularization and wound strategies. Tobacco use was most prevalent in the open (78.2%) and hybrid (90.0%) cohorts (P < .001). Diabetes was more common among revascularized patients (77% endovascular; 76.7% hybrid vs 37.8% wound care only; P < .001). Coronary artery disease and ESRD were also more prevalent in revascularized groups.

Wound grades were more severe in the wound care-only cohort (50.3% grade 3 vs 15.7% grade 3 in the endovascular cohort). Ischemia grades 3 and 4 predominated across all groups (75%-92%). Foot infection was more common in nonrevascularized patients (87.7% vs 61%-78% revascularized; P < .001). Clinical stage 4 was most frequent in the wound care-only group (76.7% vs ~60% revascularized) (Table II and Supplementary Table I (online only)).

Wound healing, wound healing time.

Wound healing was analyzed both for initial presentation and for all recurrent and de novo wound episodes. During the first wound episode, healing occurred in 510 limbs, with a median time of 215 days (IQR, 95-436 days) (Table III). When all 1158 episodes were included, 550 wound episodes achieved complete healing, with the lowest healing rate observed in the wound care-only group compared with revascularized strategies (P < .001). The median wound-healing time did not differ meaningfully by revascularization type. Limb loss with unhealed wounds occurred in approximately one-fourth of episodes across strategies. Death before healing was more frequent in wound care-only patients (27.0% vs 14%-8% in revascularized; P = .002) (Table III).

Table III.

Outcomes of wound healing and related events (per wound episode)

Variable	Wound-care—only	Endovascular	Open	Hybrid	All (N = 1158)	P1 (all)	P2 (endovascular vs open)
Initial wound episode healed	45 (25.0)	347 (50.9)	100 (51.8)	18 (50.0)	510 (46.7)	<.001	.98
Healing time, initial episode, days	139 (54-386)	229 (104-447)	240 (121-436)	210 (103-454)	215 (95-436)	.52	—
All wound episodes healed	69 (32.4)	363 (50.9)	100 (51.0)	18 (50.0)	550 (47.5)	<.001	.98
Healing time, all episodes, days	242 (40-486)	272 (133-491)	275 (173-428)	210 (128-437)	248 (120-441)	.35	.79
Limb loss with unhealed wound	54 (25.4)	145 (20.3)	49 (25.0)	9 (25.0)	257 (22.2)	.32	.16
Death with unhealed wound	57 (26.8)	128 (18.0)	28 (14.3)	3 (8.3)	216 (18.7)	.002	.23
Unhealed wound (lost to follow-up)	33 (15.5)	77 (10.8)	19 (9.7)	6 (16.7)	135 (11.7)	.15	.66
WFP, days	412 (58-869)	478 (193-1183)	666 (150-1128)	255 (7-869)	501 (163-1084)	.82	—
Wound events after healing	12 (17.6)	49 (13.5)	15 (15.0)	3 (16.7)	79 (14.4)	.82	—
Recurrence	12 (17.4)	41 (11.3)	14 (14.0)	3 (16.7)	70 (12.7)	.32	.46
Amputation after healing	0 (0)	8 (2.2)	1 (1.0)	0 (0)	9 (1.6)	.49	.44
Death after healing	2 (1.2)	2 (0.4)	0 (0)	0 (0)	4 (0.4)	.36	—
Major amputation (overall)	54 (25.4)	156 (21.9)	56 (28.7)	9 (25.0)	275 (23.8)	.23	.17
Absolute 1-year freedom from major amputation, (event-free)	133 (73.9)	556 (82.0)	156 (80.8)	30 (85.7)	895 (82.3)	.026	.30
Kaplan-Meier 1-year freedom from major amputation (95% CI)	71.2 (63.9-78.5)	82.8 (79.9-85.7)	80.0 (74.1-85.9)	83.4 (70.1-96.7)	80.4 (77.9-82.9)
Death (overall)	89 (54.3)	268 (47.9)	67 (40.4)	11 (36.7)	435 (47.3)	.05	.09
Absolute 1-year survival, (event-free)	108 (68.4)	400 (79.5)	131 (85.1)	20 (90.9)	659 (78.7)	<.001	.13
Kaplan-Meier 1-year survival (95% CI)	69.0 (61.7-76.3)	60.4 (55.9-64.9)	80.5 (77.2-83.8)	96.2 (88.8-100)	67 (63-71)
AFS event (major amputation or death)	140 (65.7)	445 (62.4)	116 (59.2)	19 (52.8)	720 (62.2)	.35	.35
Absolute 1-year AFS (event-free)	75 (47.5)	338 (64.8)	105 (67.3)	18 (75.0)	536 (62.3)	<.001	.56
Kaplan-Meier 1-year AFS (95% CI)	48.5 (40.9-56.1)	66.2 (62.3-70.1)	68.1 (60.9-75.4)	76.8 (60.5-93.1)	63.6 (60.5-66.7)

AFS, Amputation-free survival; CI, confidence interval; WFP, wound-free period.

Data are presented as number (%) or median (interquartile range) unless otherwise indicated.

Median healing times are reported for limbs that achieved complete wound healing.

WFP and post-healing events are reported only among healed limbs.

P1, Comparison among all treatment groups using χ² or Kruskal-Wallis tests.

P2, Pairwise comparison between endovascular and open revascularization.

Univariable analysis revealed that patients with healed wound episodes were more likely to be non-Hispanic White, obese, and have hyperlipidemia or diabetes, whereas nonhealed wound episodes were more frequent in patients with ESRD or prior stroke (Supplementary Table II (online only)). Healed wound episodes also demonstrated higher perfusion indices (ABI and TBI) and were associated with lower WIfI wound and overall stage. Revascularization success was strongly associated with healing, with 87.5% of healed wound episodes achieving durable flow restoration compared with 76.3% among nonhealed (P < .001). In unadjusted analyses, several factors were associated with healing; however, in multivariable modeling, successful revascularization remained the strongest independent predictor of wound healing. On multivariable logistic regression, independent predictors of complete wound episode healing included successful revascularization (OR, 2.45; 95% CI, 1.71-3.50; P < .001), obesity (OR, 1.52; 95% CI, 1.15-1.99; P = .003), and hyperlipidemia (OR, 1.55; 95% CI, 1.01-2.37; P = .044), whereas WIfI stage 4 independently predicted nonhealing (OR, 0.61; 95% CI, 0.47-0.78; P < .001) (Table IV).

Table IV.

Logistic regression for predictors of wound healing

Variable	Unadjusted OR (95% CI)	P value	Adjusted OR (95% CI)	P value
Age (per year)	0.99 (0.97-1.02)	.57	—	—
Male sex	1.19 (0.90-1.58)	.21	—	—
Diabetes	0.97 (0.52-1.79)	.91	—	—
ESRD	0.70 (0.42-1.19)	.19	—	—
Stroke/CVA	0.68 (0.52-0.89)	.001	— (NS in multivariable)	—
Obesity	1.52 (1.15-1.99)	.003	1.52 (1.15-1.99)	.003
Hyperlipidemia	1.55 (1.01-2.37)	.044	1.55 (1.01-2.37)	.044
WIfI stage 4 (vs 3)	0.61 (0.47-0.78)	<.001	0.61 (0.47-0.78)	<.001
Achieved revascularization	2.45 (1.71-3.50)	<.001	2.45 (1.71-3.50)	<.001

CI, Confidence interval; CVA, cerebrovascular accident; ESRD, end-stage renal disease; NS, not significant; OR, odds ratio; WIfI, Wound, Ischemia, and foot Infection.

Variables significant in univariate analyses (P < .10) were entered into multivariable modeling. Achieved revascularization, obesity, and hyperlipidemia independently predicted healing, whereas WIfI stage 4 predicted nonhealing.

Recurrence and wound-free period.

Among the 550 healed wound episodes, recurrence or post-healing amputation occurred in 79 episodes (14.4%), with nonsignificant variation across treatment strategies. Treatment-stratified analyses were limited to healed episodes with available classification and included 68 wound care only, 363 endovascular, 100 open, and 18 hybrid episodes (total n = 549). Recurrence was observed in 12 of 68 (17.6%), 49 of 363 (13.5%), 15 of 100 (15.0%), and 3 of 18 (16.7%) episodes, respectively. Accordingly, WFP was analyzed as a time-to-event outcome using Kaplan-Meier methods. The median WFP differed numerically but not significantly across strategies (log-rank P = .41). Patients receiving wound care only had a median WFP of 412 days (95% CI, 0-1033 days), compared with 499 days after endovascular repair (95% CI, 259-739 days), 755 days after open repair (95% CI, 240-1270 days), and 502 days after hybrid procedures (95% CI, 0-1294 days). Overall, the cohort demonstrated a median WFP of 502 days (95% CI, 350-654 days). WFP was analyzed as a time-to-event outcome using Kaplan-Meier methods. Kaplan-Meier analysis showed no significant difference among groups (log-rank P = .41). Patients with recurrence tended to be younger (median, 63 vs 68 years; P = .07) and had a higher proportion of WIfI stage 4 limbs (66.7% vs 54.8%; P = .06), although most baseline comorbidities and perfusion indices were similar between groups. The majority (90.7%) had achieved revascularization, without significant differences in treatment type (P = .49). On univariate analysis, no single variable significantly predicted recurrence, although numerically shorter WFPs were observed with older age, non-Hispanic ethnicity, longer initial healing time, and higher WIfI stage (data not shown). On multivariable linear regression (Table V), younger age (β = −6.4 days per year; P = .026) and male sex (β = +132 days; P = .041) were independently associated with longer wound-free periods, whereas endovascular, open, or hybrid revascularization showed a borderline trend toward improved durability (β = +92 days; P = .055).

Table V.

Multivariable linear regression for predictors of wound-free period (WFP)

Predictor	β coefficient, days	95% CI	P value
Age (per year)	−6.4	−12.1 to −0.8	.026
Male sex	+132.4	+5.7 to +259.1	.041
ESRD	−15.8	−155.8 to +124.3	.825
Revascularization type (endovascular/open/hybrid vs wound care only)	+92.3	−2.1 to +186.7	.055
Wound healing time (per day)	−0.17	−0.33 to 0.00	.051
Non-Hispanic ethnicity	−128.6	−266.0 to +8.7	.066
Race (Black/other vs White)	−35.6	−117.3 to +46.1	.392
WIfI stage (4 vs 3)	−84.2	−207.1 to +38.7	.179
WIfI foot infection	+61.1	−27.8 to +150.0	.178

CI, Confidence interval; ESRD, end-stage renal disease; WIfI, Wound, Ischemia, and foot Infection.

Positive β coefficients represent longer wound-free intervals. Younger age and male sex independently predicted prolonged WFP, whereas revascularization type and wound-healing duration showed borderline significance.

Median WFP was 502 days (95% CI, 350-654 days) and did not differ among revascularization strategies (P = .43). Survival by procedure type was longest after open revascularization (30.4 months) but not statistically significant (P = .18). WFP events (wound recurrence or amputation) occurred as follows: wound care only 18%, endovascular 14%, open 15%, and hybrid 17%. Kaplan-Meier curves (Fig 1) demonstrated no significant difference among groups (log-rank χ² = 2.95; df = 3; P = .399). Kaplan-Meier estimated WFP at 1, 2, and 3 years was 92.3%, 85.5%, and 69.3% for wound care only; 92.6%, 89.0%, and 85.7% for endovascular; 93.2%, 91.6%, and 84.3% for open; and 94.4%, 84.0%, and 84.0% for hybrid.

Fig 1.

Kaplan-Meier curves of wound-free period (WFP) after healing. Kaplan-Meier analysis comparing wound-free period after healing among treatment groups. No significant difference was observed (log-rank P = .43). Estimated 1-, 2-, and 3-year wound-free survival rates were: endovascular (Endo) 92.6%, 89.0%, and 85.7%; open 93.2%, 91.6%, and 84.3%; hybrid 94.4%, 84.0%, and 84.0%; and wound care-only 92.3%, 85.5%, and 69.3%.

Amputation-free survival, limb salvage, and survival.

Major amputation rates were similar across groups, whereas mortality and AFS were consistently worse in the wound care-only cohort. Consistent with these patterns, 1-year absolute and Kaplan-Meier freedom from major amputation were highest after any revascularization and lowest in the wound care-only group, with all revascularization types achieving similarly favorable limb salvage durability. One-year survival followed the same trend, with significantly lower survival in the wound care-only cohort and the best outcomes after open and hybrid procedures; Kaplan-Meier curves showed early mortality divergence in patients without durable flow restoration. Similarly, 1-year AFS was markedly worse in the wound care-only group, whereas all revascularization strategies demonstrated improved AFS, with endovascular and open repair performing comparably. Hybrid repair showed the highest 1-year Kaplan-Meier AFS, although this was limited by sample size. These differences were significant, with clear separation of AFS curves favoring revascularization.

The median AFS was shortest in the wound care-only group (11.0 months) and longer after endovascular (18.7 months), open (21.7 months), and hybrid (17.5 months). Kaplan-Meier analysis demonstrated significantly better AFS in revascularized limbs compared with wound care-only limbs. Endovascular and open surgery had comparable outcomes, whereas hybrid procedures followed overlapping trajectories but were limited by small sample size (Fig 2).

Fig 2.

(A) Kaplan-Meier estimated overall survival in patients with Wound, Ischemia, and foot Infection (WIfI) stage 3 and 4 chronic limb-threatening ischemia (CLTI), stratified by treatment strategy. Differences between the groups were statistically significant (log-rank P < .001). (B) Kaplan-Meier estimated freedom from amputation in patients with WIfI stage 3 and 4 CLTI (log-rank P < .05). (C) Kaplan-Meier estimated amputation-free survival (AFS) in patients with WIfI stage 3 and 4 CLTI (log-rank P < .001). Endo, Endovascular.

On univariate analysis, revascularization was associated with a significantly lower hazard of amputation or death compared with wound care only. Endovascular repair reduced risk by 28% (HR, 0.72; 95% CI, 0.56-0.91; P = .006), open repair by 38% (HR, 0.62; 95% CI, 0.46-0.84; P = .002), and hybrid procedures by 42% (HR, 0.58; 95% CI, 0.34-0.98; P = .041). Revascularization was associated with a significantly lower hazard of amputation or death compared with wound care only. After multivariable adjustment for age, sex, ESRD, and TP, these associations remained significant: endovascular HR of 0.66 (95% CI, 0.51-0.86; P = .002), open HR of 0.56 (95% CI, 0.41-0.77; P < .001), and hybrid HR of 0.50 (95% CI, 0.28-0.90; P = .021). Older age (HR, 1.02 per year; P < .001) and ESRD (HR, 1.86; P < .001) were independently associated with worse outcomes, whereas higher TP was protective (HR, 0.99 per mm Hg; P = .001). Female sex demonstrated a nonstatistically significant association with higher risk (HR, 1.18; P = .06). Fig 3 illustrates the adjusted HRs and CIs, and Supplementary Table III (online only) provides complete univariate and multivariable model results.

Fig 3.

Forest plot of Cox proportional hazards regression for amputation-free survival. Hazard ratios (HRs) and 95% confidence intervals (CI) are displayed for revascularization type (endovascular, open, hybrid) compared with the reference group (wound care only), as well as for patient-level covariates (age, sex, end-stage renal disease [ESRD], and toe pressure [TP]). Revascularization of any type was associated with significantly improved outcomes, whereas increasing age and ESRD predicted higher risk, and higher TP was protective.

DISCUSSION

In our cohort, traditional endpoints such as AFS, limb salvage, and mortality were within expected ranges and remain clinically important benchmarks in CLTI. However, these survival-based outcomes do not capture the prolonged morbidity associated with wound care. Accordingly, our analysis focuses on wound episode healing, WHT, and WFP as complementary, wound-centered metrics in a high-risk WIfI stage 3 and 4 population in which these patient-centered outcomes have been incompletely characterized. Among patients presenting with WIfI clinical stage 3 and 4, WHT for the index wound episode was prolonged, usually requiring more than 6 months. Less than one-half of the index wound episodes achieved complete healing by last follow-up, and more than one-half of patients experienced major amputation and/or death, underscoring the persistently poor prognosis of high-stage CLTI. Given the high competing risk of death and major amputation in WIfI stage 3 and 4 CLTI, composite endpoints such as AFS are heavily influenced by systemic disease burden; in contrast, WHT and WFP provide complementary wound-centered perspectives that should be interpreted independently of survival-based outcomes.

Criteria defining appropriateness of revascularization can serve to individualize patient and clinician shared decision-making.¹² Yet such criteria are lacking in CLTI, with significant practice pattern variation between clinicians and institutions.^13,14 In contrast to lower stage WIfI patients, revascularization is a critical pillar of achieving wound healing in WIfI stage 3 and 4 disease, and in our cohort, successful revascularization was independently associated with a substantially higher likelihood of wound healing. To avoid overstatement, we interpret these findings as associative rather than causal given the observational study design. In contrast to WIfI stage 1 and 2 populations, where delayed revascularization may be reasonable in centers with robust wound care infrastructure^9,10 the high ischemic burden of WIfI stage 3 and 4 disease frequently necessitates revascularization to enable healing. It is important to emphasize, however, that revascularization does not guarantee success; healing rates remained modest despite contemporary multidisciplinary planning and wound care. Providing data for estimated wound healing rate, WHT, and WFP are critical to set realistic expectations for patients and caregivers, and for patient-centric outcome adjudication between clinicians, clinical trialists, and other researchers.

Our work further highlights the utility of the WIfI clinical staging system to accurately risk-stratify for limb-related outcomes, in addition to its original intent to attempt to predict 1-year risk of major amputation, as well as the benefit of revascularization. Prior authors^15–17 have briefly explored wound-healing metrics, but did not describe WHT and WFP, which helps to quantify the morbidity associated with wounds beyond the binary metric of whether or not a wound heals. The progressive gradient of wound healing outcomes across WIfI stages observed in this cohort further validates its use as a contemporary framework for guiding treatment selection and wound-based assessments. Moreover, our focus upon late-stage WIfI 3 and 4 limbs highlights the heavy mortality and limb loss in comparison to WIfI stage 1 and 2 limbs, again validating the efficacy of the WIfI staging system and its subgroups. Also, the high mortality and limb loss rate in our subgroup emphasizes the need to focus on more patient-centric endpoints like WHT and recurrence. Future work will be required to clarify all of the care trajectories undertaken by patients, to better marry patient expectations and wishes with realistic outcomes. Based upon our work, we advocate for the inclusion of wound healing, WHT, and WFP as a core outcome metric^18,19 in CLTI quality assessments between institutions and practitioners and clinical trials, to effect a better understanding of patient-centric wound-related morbidity.

Although differences in procedural durability were expected, our analysis demonstrated no significant variation in wound-free survival after complete healing among endovascular, open, hybrid, or nonrevascularized limbs. At first glance, this finding may appear counterintuitive; however, it becomes clinically logical when recognizing that the analysis was limited to wounds that achieved closure. Once a wound episode heals, the determinants of long-term durability shift from macrovascular patency to local and systemic factors such as microcirculatory health, neuropathy, infection control, and adherence to preventive wound care.^20,21 Consequently, the revascularization technique may exert less influence on subsequent wound recurrence or limb loss. In this context, WFP likely reflects host physiology and long-term care quality more than procedural type. These results emphasize that, although revascularization is indispensable for achieving healing, maintaining a healed wound may depend more upon preventative measures, such as glycemic control, infection prevention, offloading, and vigilant limb surveillance.

Several findings require specific clarification. First, the wound care-only cohort demonstrated a shorter median WHT among the minority who healed, despite the lowest overall healing rate. This paradox likely reflects competing-risk and survivor bias: patients managed without revascularization more frequently experienced early major amputation or death, truncating observed healing times among those who might otherwise have had prolonged courses. Likewise, the association between obesity and improved healing should be interpreted cautiously and likely reflects residual confounding or survivorship bias rather than a true biologic protective effect. Similarly, hyperlipidemia may serve as a surrogate for statin use or greater health care engagement rather than a direct biologic effect on wound healing. In contrast, revascularized patients remained at risk longer, permitting both prolonged healing trajectories and a higher probability of eventual closure. Also, the wound care-only cohort was smaller than the revascularized groups, reflecting contemporary practice patterns. Accordingly, comparisons should be interpreted as associative rather than causal.

Second, baseline ABI patterns should be interpreted cautiously. ABI may overestimate true perfusion in patients with diabetes and medial arterial calcification, which may explain why the wound care-only group demonstrated a higher median ABI despite more advanced WIfI clinical stages and lower toe-based perfusion indices. Accordingly, ischemia assessment and clinical decision-making at both centers relied more heavily on TP and TBI than ABI alone, which argues against inadvertent inclusion of nonischemic wounds as an explanation for these findings. Finally, the observed trend favoring open revascularization at lower ABI values likely reflects greater anatomic complexity or multilevel disease. Because standardized anatomic classification systems (eg, Global Limb Anatomic Staging System [GLASS]¹) and angiographic endpoints were not available across both centers, future studies matching intervention strategy to disease distribution will be important to determine whether wound-centered outcomes vary when interventions are aligned with lesion complexity and target level.

Finally, our work highlights the increased morbidity and procedural volume required for limb salvage attempts with increasingly severe WIfI stage. Although intuitive to vascular surgeons, the paucity of explicit data showing the relationship between increasingly complex disease and procedural difficulty confounds attempts to advocate for favorable compensation models for vascular surgeons. Indeed, the Center for Medicare and Medicaid Services (CMS) rule CMS-1832-F, enacted in 2025 and effective as of January 1, 2026, introduces an “efficiency adjustment.” This reduces work relative value units by 2.5% for nontime-based codes.²⁰ Moreover, there is a 50% reduction in the practice expense component for in-hospital procedures, which will likely result in a net decrease in work relative value unit compensation for hospital-based procedures. These reductions occur because CMS does not account for increasing complexity with frequently performed procedures, such as those required for WIfI stage 3 and 4 patients. Our work provides much needed data for our advocates to hopefully improve compensation models for those undertaking increasing complex patients and wounds.

Several limitations merit consideration. First, this was a retrospective, nonrandomized analysis from two quaternary referral centers, which introduces potential selection bias and limits generalizability. Second, treatment assignment was not standardized and was determined by clinical judgment, anatomic considerations, and patient preference; consequently, the wound care-only cohort likely represents a higher risk and less reconstructable population, introducing selection bias and residual confounding that cannot be fully eliminated. Prior attempts to construct prospective controlled trials randomizing to wound care only vs revascularization in CLTI have been critiqued as unethical; hence, our retrospective cohort is likely the best data that will be published comparing wound care only vs early revascularization strategies, albeit with selection bias. Third, standardized anatomic/durability variables (inline flow, pedal target achievement, runoff scoring, GLASS classification,¹ and patency endpoints) were not available across both centers, limiting our ability to stratify outcomes by anatomic success or procedural durability; physiologic indices (ABI, TP, TBI) were used as surrogates and may not fully capture revascularization quality. Fourth, wound healing and recurrence assessment relied on chart abstraction and may introduce misclassification, although this risk was mitigated by dedicated data abstractors. Finally, some patients lacked complete longitudinal follow-up, particularly for wound-free survival, which may underestimate recurrence events. Moreover, other quality metrics and/or questionnaires championed in prior works^21,22 were not utilized due to the retrospective nature of the study and impracticality of implementing questionnaires. Nonetheless, the large sample size, episode-level analysis, and comprehensive inclusion of both revascularized and conservatively treated patients strengthen the robustness and clinical relevance of our wound-related findings.

CONCLUSIONS

In conclusion, among WIfI stage 3 and 4 patients with tissue loss, wound healing remains prolonged and overall prognosis remains poor in spite of modern multidisciplinary wound care and revascularization. Successful revascularization was strongly associated with achieving wound healing; however, among episodes that healed, the durability of closure—as measured by WFP—did not differ significantly across treatment modalities These findings show that, although revascularization remains critical to achieve healing, maintaining that healing depends primarily on patient-specific and systemic factors rather than procedural type. Together, these data support the incorporation of WHT and WFP as complementary wound-centered endpoints alongside traditional survival-based outcomes and emphasize the importance of comprehensive longitudinal care after wound closure to sustain limb preservation and minimize recurrence.

ARTICLE HIGHLIGHTS.

• Type of Research: Two-center, retrospective cohort

• Key Findings: Among 920 patients (1091 limbs; 1158 wound episodes) with Wound, Ischemia, and foot Infection stage 3 and 4 chronic limb-threatening ischemia, 82% underwent revascularization and 18% were managed with wound care only. Wound healing occurred in 47.5% overall with a median of 248 days. Revascularization was associated with a 2.5-fold higher likelihood of healing (P < .001). The wound-free period (median, 502 days) and durability of healing did not differ significantly by revascularization modality.

• Take Home Message: In high-risk Wound, Ischemia, and foot Infection stage 3 and 4 chronic limb-threatening ischemia, revascularization is associated with achieving wound healing. Achieving initial closure is critical, as recurrences are uncommon. Among episodes that heal, long-term durability appears to depend more on patient and systemic factors than on the revascularization method.