The frequency of VAP mainly varies according to the diagnostic criteria applied and to the type of ICU and population. According to the criteria applied in this study, the frequency of VAP was of 26.2%. Other authors have found incidences between 9% and 27%, and this incidence increases according to the duration of mechanical ventilation .
The mortality rate found in patients undergoing VAP in a few studies performed in Brazil varies between 32.1% and 70.9% [18–21], which agrees with a review by Chastre and Fagon who reported rates from 24% to 76% . The present study found a higher rate of mortality (78.8%) than previous studies, and that comorbidity was the only variable significantly associated with mortality in VAP patients. The presence of some underlying conditions or diseases, including cardiac surgery, acute lung injury, and immunocompromising conditions, significantly affect prognosis in VAP patients, and consequently, the mortality rates are high .
Distinct microorganisms are described as agents of VAP, and their frequency may differ according to many factors, including the population of patients in the ICU, duration of hospital stay, and methods of diagnosis. However, in many studies, Gram-negative bacteria were involved in more than 60% of VAP cases, with principally P. aeruginosa, Acinetobacter spp., and Enterobacteriaceae. Among Gram-positive bacteria, S. aureus is the predominant agent in VAP cases [4, 10, 11, 22]. Results from Latin American studies are similar to ones obtained in this study [18, 20, 21, 23, 24]. S. maltophilia has also been frequently described in VAP cases [11, 13]. Similar to these previous studies, in our study, the most frequent microorganisms were Acinetobacter spp., P. aeruginosa, Enterobacteriaceae, and S. aureus. Their frequencies were also significantly associated with death. Three isolates of Candida were detected, but they did not appear to be involved as etiologic agents because they were isolated in association with multiresistant bacteria, and because yeasts have been detected from respiratory tract specimens from patients without disease [25, 26].
Microbiological investigation of VAP is of great importance for developing appropriate antimicrobial therapy and for standardizing empirical therapies to be used in the future. This is because the local susceptibility profile of bacteria commonly associated with the disease would already be known. In this context, the culture of tracheal aspirate has similar importance for diagnosis compared with invasive techniques of bronchoalveolar wash and a protected specimen brush, and it is also a simpler and less expensive technique . However, quantitative cultures of tracheal aspirates have low specificity (48%–78%) and sensitivity (26%–65%) . In our study, the mortality rate was high, even though etiological agents were identified in 97% of the cases with quantitative culture of the tracheal aspirates, confirming that others factors are important in the prognosis of VAP patients, as previously described .
In our study we found a frequency of 59% of multiresistant bacteria associated with VAP. In another study in which the criteria applied for classifying multiresistant bacteria were similar to ours, the frequency of multiresistant bacteria was 27% . Among the most frequent multiresistant bacteria found in VAP are P. aeruginosa, Acinetobacter spp., and S. aureus[10, 12, 19]. This finding was confirmed in our study.
Because we analyzed the isolation frequency of those bacterial species that were multiresistant, a significant correlation with death was observed. However, no association was evident when we compared death in patients displaying multiresistant isolates with death in patients with VAP caused by non-multiresistant bacteria, which is the ideal comparison. The lack of association between antibiotic resistance and mortality has been previously reported in VAP patients whose agents were P. aeruginosa and MRSA [28, 29]. However, bacterial resistance was responsible for a high mortality rate in patients undergoing VAP in many reports [9, 30–32]. Such lack of association observed in the current study is probably due to the small sample size, which is a limitation of this study.
For bacterial resistance, it is difficult to determine the most appropriate antimicrobial therapy for VAP, and patients also stay for long periods in hospitals and ICUs, where use of antimicrobial agents is greatly required [24, 33]. Additionally, the likely multiresistant bacterial etiology of VAP increases the use of large spectrum empirical antibiotics and of combined therapies . The high levels of antibiotic resistance found among P. aeruginosa, Acinetobacter spp., and S. aureus against major antibiotics usually administered in patients infected with these microorganisms could have led to inadequate antibiotic therapy and a poor outcome in our patients.
In our study, we found that previous antibiotic therapy was used in 97% of patients with VAP. The antibiotics most used were cefepime, vancomycin, and cefalotin. According to the spectrum range of antimicrobial agents for Gram-negative bacteria, carbapenems present with the largest spectrum range of action, followed by cefepime, piperacillin/tazobactam, quinolones, and others . Depuydt et al. observed that patients who received antibiotics of at least two different classes showed a higher probability of getting infected by multiresistant bacteria . Previous use of antibiotic therapy also affects the resistance of etiological agents of VAP . Occurrence of multiresistant P. aeruginosa and MRSA has been significantly associated with previous exposure to ceftazidime, while Acinetobacter spp. are associated with previous exposure to piperacillin/tazobactam , which are similar findings to our study.