Correlation of Time to Soft Tissue Coverage With Clinical Outcome in Lower Extremity Trauma in the Modern Era: A Single-Center Retrospective Review

Lower extremity salvage in the setting of severe trauma requires the consideration of multiple surgical specialties and treatment algorithms. Advances in bony fixation methods, automobile safety, wound management, antibiotics, and decreases in tobacco use have yet to definitively improve severe lower extremity trauma outcomes over the past 30 years. The addition of negative pressure wound therapy (NPWT) as well as improvements in the reliability of free tissue transfer also have yet to demonstrate obvious improvements in long-term outcomes to these patients.^1-8

A multidisciplinary approach to the most serious of lower extremity injuries, beginning from the assessment in the trauma bay, is warranted.^9-11 Nevertheless, a line of care wherein the orthopedic trauma specialist irrigates, debrides, and fixates fractures in the acute setting, subsequently communicating with the reconstructive surgery team regarding need for soft tissue coverage, frequently manifests. Inefficient and costly, this method of care pushes the reconstructive surgeon to perform soft tissue reconstruction or else bear the responsibility of recommending amputation of an extremity for which much time, effort, and money has already been expended salvaging.

We wanted to understand how, in the modern era, time to soft tissue coverage impacts meaningful lower extremity outcomes. We hypothesized that time to first ambulation, ambulation without an assistive device, chronic osteomyelitis, and delayed amputation were not affected by the time to soft tissue coverage in Gustilo IIIB and IIIC fractures at our institution.

Institutional review board approval was obtained before performing patient data collection. Patients were identified by ICD-9 code for open tibia and open fibula fractures (823.10, 823.12, 823.30, 823.32, 823.90, 823.92). All patients meeting criteria from January 1, 2007, to August 30, 2017, were included. Each patient chart was then individually reviewed for data collection. Patients who required any form of soft tissue coverage to the lower extremity during their initial hospitalization and had at least 30 days follow-up from time of hospital discharge were included. Patients who underwent soft tissue coverage at a subsequent admission or outside institution were excluded to limit those receiving coverage for osteomyelitis or other nonacute complications.

Baseline patient characteristics included age, sex, body mass index (BMI), smoking status, and psychiatric illness, in addition to other comorbidities. Modified Gustilo severity,¹⁶ time to initial irrigation and debridement, use and length of NPWT treatment, time to soft tissue coverage (from the date of injury), and method of soft tissue reconstruction were recorded; patients were classified according to the most complex form of soft tissue reconstruction received. By definition, all Gustilo IIIC patients received vascular reconstruction. Acute surgical complications and infections requiring a return to the operating room were counted if they occurred during the initial hospitalization. Long-term complications included chronic osteomyelitis (either diagnosed intraoperatively or by magnetic resonance imaging), delayed amputation, and deep vein thrombosis.

Data were tabulated on length of follow-up and functional outcome defined as independent ambulation or ambulation requiring an assistive device, either documented in the chart by the orthopedic or plastic surgery team or physical therapist. Return-to-work status was initially proposed as an outcome measure, but this was rarely documented definitively and was abandoned as a result. Primary outcomes were time to first ambulation, ambulation without an assistive device, chronic osteomyelitis, and delayed amputation. Secondary outcomes were any complication requiring return to the operating room at the initial hospitalization, cellulitis, abscess, deep venous thrombosis, pulmonary embolism, and flap compromise or loss (partial or total).

The typical protocol at our institution involved consultation by Orthopedic Surgery when, at their discretion, soft tissue coverage was required. Initial consultation could vary from as early as in the trauma bay, to 5 to 7 days after admission once bony fixation was completed.  Plastic Surgery would evaluate the wound on the day of consultation and formulate a plan. The typical goal was to achieve bony and/or hardware coverage within 7 days of consultation. In certain instances, at the orthopedists’ discretion, the patient would be transferred to an outside institution in an effort to obtain soft tissue coverage sooner.

Descriptive statistics were performed as indicated for categorical and continuous variables. Regression analysis was computed using the Pearson correlation for 2 nominal variables, whereas the Spearman rank correlation was utilized for ordinal variables (ie, type of coverage ranked in order of increasing severity of Gustilo grade). Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated after adjusting for risk factors to evaluate the impact of time to soft tissue coverage on chronic osteomyelitis, 90-day return to any ambulation, or 180-day return to ambulation without assistive devices. Variables included in the multivariable analysis were age, sex, BMI, BMI > 30, BMI > 35, diabetes, coronary artery disease (CAD), immunosuppression, smoking status, psychiatric diagnosis, substance abuse, Gustilo grade (I, II, IIIA, IIIB, and IIIC), vessel injury, number of NPWT days, number of operating room debridements, and time to soft tissue coverage. Univariable analysis was performed using an α value of 0.10 as the threshold for inclusion into the multivariable models, which were subsequently analyzed with significant factors at an α of 0.05. A power analysis was not conducted. All statistical analyses were performed using JASP version 0.8.6.

A search of our electronic medical record system returned 905 patients who presented with open tibia fractures from January 1, 2007, to August 1, 2017. After chart review and removal of duplicates, 575 remained. Of these, 89 required soft tissue coverage. Nine additional patients were transferred to an outside facility for definitive coverage. A flowchart based on Strengthening the Reporting of Observational Studies in Epidemiology criteria is presented in Figure 1. The cohort was consistent with the previously published literature: most patients were male (74.1%) and aged 20 to 39 years (44.9%), with a bimodal peak of the elderly aged 60 to 69 years (16.8%). Active smokers made up 33% of the cohort, 25% had documented substance abuse, 16% had a psychiatric diagnosis, 12% had diabetes, 9% had CAD, and 7% were immunosuppressed. Of the 9 patients transferred to an outside institution for soft tissue coverage, 8 (89%) were male and 3 (33%) were active smokers; the average age of this subgroup was 39.4 years (standard deviation, 19.0), similar to our cohort. Methods of soft tissue coverage by injury grade can be found in Table 1.

Methods of fixation were consistent throughout. Of the 89 patients requiring soft tissue coverage, 39 (43.8%) received an intramedullary rod only, 30 (37.5%) received plate fixation, 11 (12.4%) received plate fixation and an intramedullary rod, and 40 patients (44.9%) received external fixation. Also of note, the overwhelming majority of patients (n = 86; 97%) were treated with NPWT. No patient had a documented tibial nerve transection.

Among the Gustilo IIIB fractures, 1 acute amputation was performed. Median days of NPWT were 8 (interquartile range [IQR], 14.3), and median wound debridements before definitive soft tissue coverage was 3 (IQR, 2). Median time to soft tissue coverage was 7.5 days (IQR, 12.8). A total of 14 patients (25.9%) with grade IIIB fractures developed chronic osteomyelitis; of these, 4 were a skin graft only, 7 were treated with a local or pedicled flap, and 3 were treated with free flaps (Table 2). Median days of NPWT and washouts in this group before soft tissue coverage were 7.0 (IQR, 6.5) and 3.0 (IQR, 3.0), respectively, compared with 8.5 NPWT days (IQR, 15) and 3.0 washouts (IQR, 2) for grade IIIB fractures that did not develop osteomyelitis (Table 3). Median time to soft tissue coverage without development of chronic osteomyelitis was 8 days (IQR, 13.5) compared with 6 days (IQR, 12) for those with chronic osteomyelitis.

Among 18 Gustilo IIIC fractures, 1 acute amputation was performed. Median days treated with NPWT and wound debridements before definitive soft tissue coverage were 11 (IQR, 14.3) and 3 (IQR, 2), respectively. Median time to soft tissue coverage was 8.5 days (IQR, 7). A total of 6 patients (33%) developed osteomyelitis. Among the 6 IIIC fractures that developed chronic osteomyelitis, 2 were skin grafted, 1 received a local flap, and 3 required a free flap, 1 of which resulted in a flap loss. Median NPWT days and wound debridements were 10.5 (IQR, 9) and 4.0 (IQR, 2.5), respectively. Median NPWT days and wound debridements for patients with grade IIIC fracture without chronic osteomyelitis were 9 (IQR, 17.8) and 3.0 (semi-interquartile range, 1). Median time to coverage without development of osteomyelitis was 8.5 days (IQR, 1) compared with 10.5 days (IQR, 9) for those who developed osteomyelitis.

Among 40 patients (45%) who underwent external fixation, 14 (35%) developed osteomyelitis, and this approached significance on univariable analysis (P = .079), Table 4. Additionally, 33 patients underwent definitive fixation via open reduction internal fixation (ORIF), and they were found to have a significantly lower rate of osteomyelitis (n = 7 [22%]; P = .003) than those who received an intramedullary nail (IMN; 13 of 40; 33%) or both (IMN and ORIF, 3 of 12; 25%). This was also found to be significant on multivariable analysis when controlling for coronary artery disease and external fixation (P = .028).

Univariable analysis regarding method of fixation and time to ambulation (ie, ambulation at 270 days from the time of injury with assistive device, 365 days from the time of injury without assistive device) demonstrated there was no significant difference among those who received external fixation nor by method of definitive reconstruction with respect to time to ambulation with or without an assistive device (Table 5 and Table 6).

Regarding the 9 patients transferred to an outside institution, 6 patients had information available that allowed for adequate analysis. No patient had an acute or delayed amputation. All patients were treated with NPWT initially. The median number of NPWT days was 8 (IQR, 8.5), and median number of debridements was 3.5 (IQR, 2.5). Median time to coverage was 22 days (IQR, 8). Four patients (67%) had a free flap (2 latissimus dorsi, 2 anterolateral thigh). One patient in this cohort developed osteomyelitis; this patient had a IIIB fracture, received 7 days of NPWT, and underwent free latissimus dorsi transfer for coverage 22 days after the date of the injury. He ambulated with an assistive device 11 months after his injury and ambulated without an assistive device 13 months after his injury.

Table 7, Table 8, and Table 9 detail the risk factors associated with chronic osteomyelitis and return to ambulation. Only CAD was significantly associated with chronic osteomyelitis on univariable analysis, but this factor did not influence time to soft tissue coverage on multivariable analysis (OR, 1.01; 95% CI 0.99-1.04; P = .29). No factors were associated with return to first ambulation without assistive device after 90 days, including time to soft tissue coverage (OR, 0.99; 95% CI, 0.96-1.02; P = .50). Only vessel injury was significantly associated with return to first ambulation without assistive device after 180 days on univariable analysis, but this factor did not influence time to soft tissue coverage on multivariable analysis (OR, 0.99; 95% CI, 0.96-1.02; P = .55).

After evaluating 89 patients who presented with Gustilo IIIB and IIIC fractures treated over a recent 10-year period, we found that time to soft tissue coverage did not affect time to first ambulation, ambulation without an assistive device, chronic osteomyelitis, or delayed amputation. This may further substantiate that timing to obtaining soft tissue coverage may not affect functional or meaningful outcomes in patients with severe lower extremity trauma. This study is one of only a select few multivariable analyses to evaluate time to soft tissue coverage as an independent variable for lower extremity functional outcomes.^17,18

Several high-level-of-evidence studies have suggested that, regardless of timing, method of wound care, or method of soft tissue coverage, meaningful outcomes (eg, osteomyelitis, need for assistive devices, and return to work) are more related to the severity of the injury and patient characteristics than they are to treatment method and timing.^1,14,19-21 Although there are no high-level-of-evidence studies evaluating this correlation, several lower-level-of-evidence studies suggest that soft tissue coverage timing directly affects lower extremity outcomes.²² Although outcomes measured vary, they are infrequently a measure of meaningful function, and many studies evaluated injuries sustained before the use of NPWT and the development of modern broad-spectrum antibiotics (impenem and ceftazidime in 1985, linezolid in 1996, daptomycin in 2003, and tigecycline in 2005).^9,23-27 This is in addition to the literature demonstrating NPWT has been of unclear benefit.^7,8

Our institution is notable for having a low rate of acute amputations (2.2% of Gustilo IIIB and IIIC injuries from 2007 to 2017) and thus performs a proportionally high rate of open tibia fracture soft tissue coverage procedures.^12-15 The decision to suggest or even pursue acute amputation in the setting of lower extremity trauma is an ethically and personally difficult one for many physicians treating these injuries.²⁸ A leap forward in severe lower extremity trauma care and salvage appeared on the horizon with the advent of free tissue transfer, reinvigorated along the way by antibiotic development and advances in wound management. Yet, during this time of great development, criteria developed to assist with the decision to amputate have become more obscure than lucid.^29-33

This study is one of only a few outside of the LEAP trial database^30,38,39 that has attempted to evaluate meaningful lower extremity function as an outcome in limb salvage.^17,18  Although our cohort was small, we added 89 patients to the literature and accounted for the influence of several variables on limb salvage and ambulation, with only CAD being found to have an impact on outcomes, specifically chronic osteomyelitis. Other variables measured, such as age, smoking status, and severity of injury, all likely impact lower extremity trauma outcomes, highlighting the limitations of our study.^17,18,39 Additionally, we excluded all patients transferred outside of our institution, further limiting our sample size. We did not feel that these patients represented the typical lower extremity trauma cohort, and they may have received free tissue transfer at a higher rate (66% of transferred patients, with available data) than had they not been transferred. Our patient cohort was nearly identical to that of several previous studies from the United States and Europe, highlighting the possible generalizability of our findings.

This study is limited by its small sample size, including an even smaller sample size for those requiring complex reconstruction and free tissue transfer; this indicates the severity of the defects may not be as grave despite the technicality of soft tissue loss. Additional weaknesses include incomplete long-term data and the retrospective nature of the analysis, which allows for possible confounding, recall, and selection bias in addition to limiting the ability to evaluate such granular details as exposed hardware, bone, accurate defect size, and detailed description of what service is performing the debridements. Data being obtained from a single institution with multiple reconstructive surgeons, limiting generalizability due to our specific population and institutional practices, as well as operator bias are also limitations.  Additionally, Gustilo grading is inherently subjective and was either documented by the orthopaedic or plastic surgeon or calculated based on data available from the operative notes describing the area (cm²) of soft tissue coverage required.

Meaningful demographic measures that we did not evaluate include preoperative socioeconomic status, level of education, and social support.  Meaningful clinical measures we did not evaluate included presence of exposed hardware, operative time, operative blood loss, intraoperative flap thrombosis, and specific defect location in the lower extremity (proximal, middle, or distal third).

The strengths of this study include the multivariable analysis for multiple outcomes of interest, the relatively long average follow-up time, especially for those with documented ambulation status, and the lack of reliance on self-reported outcomes (ie, “do you have pain”).

It remains difficult to definitively prove that time to soft tissue coverage meaningfully impacts lower extremity outcomes. We argue this study joins a series of others in the modern era (since the advent of improved wound care and methicillin-resistant Staphylococcus aureus treatments around the year 2000) that have demonstrated timing of soft tissue reconstruction does not dictate meaningful lower extremity trauma outcomes. Modifiable risk factors that remain consistent seem to be time to irrigation and debridement, adequate debridement, and the quality of the soft tissue transfer (either local or distant).^24,36,40,41 We do not dispute one modifiable risk factor: the timing of free tissue transfer as it relates to flap outcomes. Specifically, it seems that when free tissue transfer is performed greater than 10 to 21 days after injury, flap loss rates increase,^23,24,42 although this has also been called into question by some studies in the modern era.^1,35,37

As always, future larger studies need to be performed. A recreation of the LEAP study is warranted. Until this is undertaken, we will continue to debate what influences lower extremity trauma outcomes in a substandard fashion.

Affiliations: ¹Division of Plastic Surgery, Department of Surgery, University of California Davis Medical Center, Sacramento, CA; ²California Northstate University College of Medicine, Elk Grove, CA; ³University of California Davis Medical School, Sacramento, CA; ⁴Plastic and Reconstructive Surgeons, Renton, WA

Correspondence: Chad M Bailey, MD; ch.bailey@proliancesurgeons.org

Disclosures: The authors disclosed no conflicts of interest.