S. E. Poos1, B. P. Hermans2, R. P. Ten Broek1, H. Van Goor1 1Radboud University Medical Centre, Surgery, Nijmegen, GELDERLAND, Netherlands 2Radboud University Medical Center, Cardio-Thoracic Surgery, Nijmegen, GELDERLAND, Netherlands
Introduction:
Seroma formation, fluid accumulation in subcutaneous cavities post-surgery, poses substantial challenges in surgical procedures such as mastectomy, abdominoplasty, and hernia repair. While various interventions have been employed in clinical practice, their notable drawbacks underscore the ongoing search for effective alternative interventions. Animal models have long been applied for the preclinical examination of new interventions, yet concerns regarding their translational value persist.
Method:
The objective of this systematic review was to examine methodological characteristics and outcome measures of animal seroma models used in literature and to analyze methodological shortcomings and biases of these models. Meta-analyses were performed to estimate the validity of negative control groups, as well as to estimate the effectiveness of interventions used in animal models on preventing seroma formation.
Results:
A total of 42 studies were identified, describing 6 different species, 8 different surgical applications, and 9 different intervention types. Considerable heterogeneity was revealed in terms of methodology and outcome measures. The most common reported outcome measures were qualitative seroma assessment (90%), histology (74%), and seroma volume (69%). Seroma volume was the most common outcome measure when seroma formation was the primary aim compared to a secondary aim of a study (93% vs 15%). Bias was present in all studies. Some risks of bias were regularly addressed in studies (e.g. 62% randomization, 52% blinding, 69% negative control group), while other risks of bias were often not mentioned (e.g. 12% confounders, 21% allocation concealment, 0% random housing). Negative control groups described an average seroma incidence rate of 89.48% (CI 82.70; 96.25); a funnel plot revealed the need for a sample size of ≥20 animals to obtain robust results (see Figure). Meta-analyzing all interventions, the average incidence rate ratio (IRR) and seroma volume standardized mean difference (SMD) were 2.27 (CI 1.53; 3.35) and 2.68 (CI 1.80; 3.56), respectively, reflecting significant heterogeneity (I2 IRR: 79%, I2 SMD: 86%) among studies. Fibrin glue was the most frequent intervention examined in animal seroma models with an IRR and SMD of 3.14 (CI 1.71; 5.75) and 2.04 (CI 0.13; 3.94) and considerable heterogeneity (I2 IRR: 80%, I2 SMD: 92%).
Conclusion:
This study demonstrates the disparities in methodology and outcome measures in animal seroma models, underscoring the need for method standardization. A standardized animal seroma model mandates a clear defect formation, well-defined negative control groups, and quantifiable outcome measures, for which recommendations can be derived from this review. This review contributes to improving translational value of interventions preventing seroma after surgery, impacting future design and conduct of clinical trials.
