**Background:** The development of atrioventricular bioprostheses has accelerated in recent decades, but a major barrier to clinical translation is the lack of standardized preclinical large animal models. Preclinical in vivo studies are essential to evaluate safety, effectiveness, biocompatibility, durability, and hemodynamic performance before human trials. However, there is significant variability in study design, animal selection, and outcome reporting across studies. This systematic review aims to synthesize existing evidence and propose a framework for future preclinical research on atrioventricular valve bioprostheses.
**Methods:** A systematic review was conducted following PRISMA guidelines. Three reviewers searched PubMed, Embase, and Web of Science from inception to December 31, 2021, using MeSH terms including bioprosthesis, mitral valve, tricuspid valve, hydrodynamic, durability, and feasibility studies. Inclusion criteria were preclinical acute or chronic trials evaluating surgically implanted mitral or tricuspid valve bioprostheses in large animals. Studies on mechanical valves, transcatheter valves, xenografts, or those lacking design details were excluded. After screening 564 citations (558 duplicates removed), 541 irrelevant citations were excluded, and 11 more were excluded after full-text review, leaving 12 eligible studies comprising 199 animals. Data on study design, animal characteristics, prosthesis type, and outcomes (hemodynamic performance, calcification, gross and histological findings) were independently abstracted by three authors. Quality was assessed using GRADE, and risk of bias was evaluated per Cochrane handbook guidelines.
**Key Results:** Among the 12 included studies, 9 were chronic (survival) and 3 were acute (non-survival). Sheep were the most common animal model, used in 9 of 12 studies. Other species included calves (1 study), baboons (1 study), swine (1 study), and dogs (1 study). Animal ages ranged from 3 months to 5 years, with most studies using juvenile sheep. Mean animal weight varied by species: sheep typically weighed 30–70 kg, calves 79±9 kg, baboons 21.9±5.7 kg, and swine 100 kg. Follow-up for chronic studies ranged from 3 months to 12 months. An average of ~20 animals were used in chronic studies and ~5 in acute studies. Most bioprostheses were stented (8 of 9 chronic studies; 1 of 3 acute studies), with trileaflet designs predominating (10 of 12 studies). One chronic study used a unileaflet valve, and one chronic and one acute study used bileaflet valves. Prosthesis sizes ranged from 21 mm to 29 mm. Hemodynamic data (transvalvular pressure gradient, ΔP) were reported variably: e.g., Collatusso et al. reported ΔP 10±2 mmHg for a 23 mm valve; Flameng et al. reported ΔP 3.9±1.6 mmHg for a 25 mm valve; Hassoulas et al. reported ΔP 9.73±4.93 mmHg for a 21 mm valve. Calcification was reported in 6 studies, with rates ranging from 0 to 6 animals affected. Early termination rates were as high as 50% in two studies (Collatusso et al. and Thiene et al.). Sterilization methods varied: most used glutaraldehyde (0.2%–0.65%), while Flameng et al. used 100% ethylene oxide. The quality of evidence was assessed as not serious for risk of bias and indirectness, but inconsistency and imprecision were noted due to heterogeneous designs.
**Clinical Implications:** This review highlights the critical need for a standardized preclinical framework for atrioventricular valve bioprostheses. The predominance of sheep models reflects their anatomical and physiological similarity to humans, including comparable heart rate (60–120 bpm), systolic/diastolic pressure (~90–115 mmHg), and valve anatomy. However, the variability in animal age, species, follow-up duration, and outcome reporting complicates cross-study comparisons and regulatory evaluation. The authors propose a decision-making algorithm (Fig. 2) based on US FDA guidelines, emphasizing factors such as animal age (skeletally mature vs. juvenile), body weight, sex, and the need for hemodynamic and mineralization assessment. Sheep are recommended for chronic survival studies due to their docile nature, low cost, and suitability for imaging, but their high mineralization tendency must be considered. Swine models offer advantages in coronary anatomy and organ size but are prone to refractory arrhythmias. Canine models are less commonly used due to extensive collateral coronary circulation. The review underscores that no single ideal animal model exists, and the choice should be guided by the specific research question. Standardizing reporting of outcomes—including transvalvular gradient, effective orifice area, calcification, and survival—is essential for future meta-analyses and regulatory approval. This framework aims to improve the efficiency and translational value of preclinical studies, ultimately benefiting both animal welfare and clinical device development.