Questions with inconclusive answers

To the Editor:

I read with great interest the study by Prats et al. 1 entitled “Effects of antibiotics on protected specimen brush sampling in ventilator-associated pneumonia” recently published in the European Respiratory Journal. The study demonstrated that the effect of antimicrobial treatment on the diagnostic yield of the bronchoscopically retrieved protected specimen brush (PSB) must be identified according to the microorganism involved. Haemophilus influenzae and Streptococcus pneumoniae seemed particularly vulnerable to antimicrobial treatment, whereas Staphylococcus aureus, Pseudomonas aeruginosa and Acinetobacter baumanii seemed more resistant. But what do these findings implicate in terms of the usefulness of PSB in patients with suspected ventilator-associated pneumonia (VAP)? Are we to withhold antimicrobial treatment prior to diagnostic evaluation? Should we rely more widely on PSB after having corrected the apparent biases in the literature evaluating the diagnostic yield of bronchoscopic sampling?

The issue of antimicrobial pretreatment should be honestly discussed in a realistic clinical scenario. Although it is clearly inadequate to initiate antimicrobial treatment prior to diagnostic sampling when the corresponding facilities are established, these facilities may not be available under defined circumstances (e.g. during night hours and weekends). Moreover, most regimens of antimicrobial pretreatment are present because of earlier septic episodes unrelated to suspected VAP. The indication for these antimicrobial treatment courses may be debatable in some cases but will be a matter of fact in most instances. The implications of these two types of antimicrobial treatment are completely different: although antimicrobial pretreatment initiated for suspected VAP clearly reduces the diagnostic yield, particularly in pathogens involved in early onset pneumonia, it does not when initiated for other reasons 2, 3. In fact, it is reasonable to assume that in patients developing VAP during antimicrobial treatment, the diagnostic yield may even be increased since underlying pathogens would not have caused pneumonia if these had been appropriately covered by the current antimicrobial regimen.

With these considerations in mind, any diagnostic sampling will be most rewarding in the presence of a high amount of resistant pathogens not covered by any given initial antimicrobial treatment policy. This conclusion is in line with the fact that an institution reporting a rate of 59% resistant microorganisms in patients with suspected VAP has continuously found the best operative figures of bronchoscopic sampling 4. However, from these associations we should learn that bronchoscopic sampling should not be continuously implemented in the routine practice of any intensive care unit (ICU). The challenge of reducing overall resistance rates by reducing antimicrobial selection pressure cannot be resolved by withholding antimicrobial treatment and implementing invasive diagnostic measures in order to establish the presence of VAP. It should be remembered that antimicrobial pretreatment is not the only confounder when applying diagnostic techniques based on quantitative cultures 5, 6. Instead, the preferred strategy when facing this challenge is the definition and implementation of: 1) an appropriate policy of empiric initial antimicrobial treatment; 2) an infection control programme, including strict measures in hospital environments with excess resistance rates; and 3) a critical judgment, including clinical and microbiological data, when estimating the probability of pneumonia in individual patients. Additional clues emerge from studies evaluating strategies to reduce selection pressure by short-course antimicrobial treatment regimen in patients at low risk of mortality 7.

Therefore, in my view, reopening the debate about the usefulness of bronchoscopically sampled protected specimen brush (and bronchoalveolar lavage) in patients with suspected ventilator-associated pneumonia would mean dealing with questions that have been shown to have inconclusive answers.