Retrospective Cohort Study Evaluating Time to Graft Between Intact Fish Skin Grafts and a Synthetic Matrix Used in...

Introduction: While global burden is underappreciated, traumatic wounds account for 23 million annual emergency department visits in the United States, accounting for trillions in medical costs and work and quality of life loss. We hypothesized that the use of intact fish skin grafts (IFSGs) would result in faster time to obtain a graftable wound bed compared to a synthetic matrix (SM). Methods: This retrospective analysis of patients with full-thickness acute traumatic wound was conducted on a matched sample of patients receiving IFSG versus SM. We hypothesized that there would be at least a 10-day difference between the two treatments. Two surgeons reviewed electronic charts for primary outcomes. Incomplete charts were excluded. For sensitivity analysis, both simple linear regression and generalized linear mixed models were utilized to adjust the primary outcome to potential confounders. Results: Data were reviewed for 80 patients. The final cohort included 38 patients [IFSG (n = 12) and SM (n = 26)]. Demographics were not different between groups. The mean age was 42.4 ± 20.7 years, with 62.2% being male and equal proportions having burns treated vs. another traumatic wound. The most common site was the lower extremity. All IFSG patients had negative pressure wound therapy (NPWT) placed at the time of implantation vs. 50% of SM (p = 0.001). The SM group had a larger TBSA (p < 0.001) and wound size (p = 0.001). Differences were subsequently used to adjust primary outcomes via regression. Patients with IFSG had faster time to a graftable wound bed compared to SM (mean ± SD: 13.9 ± 5.3 vs. 29.1 ± 10.3, p < 0.001). Neither wound size, burn severity, presence of infection, nor use of NPWT significantly altered the inference. Conclusions: In this paired sample, the use of IFSG as a dermal template for full-thickness traumatic injuries was associated with faster time to obtain a graftable wound bed compared to an SM. However, regression analysis alone is not enough to conclude that the same difference would exist in a prospectively controlled experiment, which should be performed.