ABC | Volume 112, Nº3, March 2019

Original Article Kiyose et al A systematic review for valvular prostheses Arq Bras Cardiol. 2019; 112(3):292-301 Figure 1 – Study search and selection processes. References identified using electronic database search (n = 7,725) References identified by manual search of other sources (n = 2) References after duplicate removal (n = 6,383) Screening by title and abstract (n = 6,383) Complete articles analysed (n = 19) Excluded references (n = 6,364) Complete articles excluded from analysis and reasons (n = 15) • 3 Not randomised • 6 Without comparison between biological and mechanical valve • 6 Multiple publication Included studies (n = 4) 1,528 participants Identification Selection Eligibility Inclusion Vallejo et al. 13 randomized 110 mitral valve replacement candidates, from 1975 to 1979, into one of three groups: Angell-Shiley porcine bioprosthesis, Björk-Shiley mechanical prosthesis, and Lillehei-Kaster mechanical prosthesis. The mean follow-up time was approximately two years. 13 The Veterans Affairs Cooperative Study randomized 575 patients between 1977 and 1982. 14-17 This study included men who received a Hancock first generation porcine bioprosthesis or Björk-Shiley mechanical single spherical 60 degrees disc prosthesis. Most patients (70%) underwent aortic valve replacement. The mean follow-up time was 15 years. Bloomfield et al randomized 533 patients of both genders to receive either a mechanical Byork-Schiley 60 degrees spherical tilting disk or a porcine bioprosthetic valve. Between 1975 and 1977 the patients assigned to a bioprosthesis received a Hancock prosthesis and after January 1977 to 1979 such patients received a Carpentier- Edwards prosthesis. 18-20 Approximately half of the patients underwent aortic valve replacement, and half underwent mitral valve replacement. The mean follow-up time was 20 years. Stassano et al. 21 randomized 310 patients, who required aortic valve replacement between 1995 and 2003, into a biological prosthesis group and a mechanical prosthesis group. 21 Carpentier-Edwards porcine or Carpentier-Edwards bovine pericardial prostheses were used in the bioprosthesis group. In the group allocated to mechanical prostheses, Carbomedics or St. Jude double disc prostheses were used. The mean follow-up time was 8.8 years. Risk of Bias Characteristics related to the risk of bias of the studies are presented in Table 2. None of the studies described how the random list was generated. The trials were at low risk of bias for all the other domains including allocation concealment, blinding of outcome assessors, and incomplete outcome data. None of the studies used blinding of patients and health professionals, which is not feasible in this scenario. Clinical outcomes There was no statistically significant difference in the risk of death between biological or mechanical prosthesis, although most of the confidence interval favours the latter (RR = 1.07; 95% CI 0.99-1.15) (Figure 2). In addition, mortality was similar in the subgroups of patients receiving prostheses in the aortic or mitral positions or in both positions simultaneously. The effect estimates from different studies were reasonably homogeneous (I² = 22%). 294