OBJECTIVES Since bacterial resistance is a public health concern worldwide, more accurate and faster diagnosis methods are essential for appropriate therapy. Molecular methods provide rapid identification of the resistance profile, allowing faster therapeutic adjustments. The aim of this study was to conduct an overview to summarize the mortality information published data regarding methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant Enterococci (VRE) and carbapenem-resistant gram-negative bacilli (CRGNB).

METHODS To search for meta-analyses evaluating infections by MRSA, VRE and CRGNB in the hospital setting, the CoCoPop model was used: Condition - infection by one of these microorganisms; Context - hospital setting; Population - hospitalized patients with infection by one of these microorganisms. Studies were identified using PubMed and Scopus databases, and were included when mortality was reported. Articles published in non-Roman characters were excluded. All steps of the overview were performed by two independent reviewers.

RESULTS The initial search yielded 347 articles. After removing the duplicates, 308 studies were screened, of which 123 were selected for full-text review. Of these, 24 meta-analyses were selected according to the inclusion criteria. Ten reported data regarding MRSA, eight studies appraised VRE and six articles assessed CRGNB. The population included in MRSA studies have heterogeneous characteristics, due to different mortality rates in each subgroup. One MRSA study included only children and neonates and present a mortality rate of 26.9%. Considering studies that included just adults patients with critical disease the mortality rate is around of 10% (8.4 - 13.1%). Six studies about VRE features just United States and Taiwan data and infection-related mortality ranged from 11.8 to 36.2%. In general, critically population are availed in CRGNB studies and the mortality was between 38.7 and 56%.

CONCLUSIONS The mortality due to MRSA, VRE or CRGNB infection is important data influenced by country development and population characteristics.