Journal Club

Utility of Gram Stain Results Guiding Empiric Antibiotic Therapy for Patients With Ventilator-Associated Pneumonia

By Milana Bogorodskaya, MD

In a recent study published in JAMA Network Open, Yoshimura et al. conducted a multicenter, open-label, noninferiority randomized clinical trial in Japan from 2018 to 2020 to evaluate whether use of Gram stain (GS) results to guide empiric antibiotic therapy for ventilator-associated pneumonia (VAP) was noninferior to guideline-directed empiric therapy.

Patients ≥15 years of age with VAP (and without diagnoses of COVID-19, heart failure, or atelectasis) were randomized to receive standard of care empiric therapy based on guidelines or empiric therapy guided by the GS result performed on an endotracheal aspirate at time of diagnosis. Anti-methicillin-resistant Staphylococcus aureus (MRSA) therapy and an anti-Pseudomonal beta-lactam agent was/were administered if GS showed gram-positive cocci in clusters and/or gram-negative bacilli, respectively. Both anti-MRSA and anti-Pseudomonal therapy were administered if GS did not show any microorganisms.

Out of 206 randomized patients, only 3 to 4 percent were immunocompromised, and only 3 to 6 percent had septic shock. Both the GS-guided arm and the guideline-guided arm had low rates of GS without any microorganisms (5 percent vs. 9 percent). Among bacteria isolated from the endotracheal aspirates, S. aureus was the most common, followed by Klebsiella spp. and Haemophilus influenzae. The clinical response rate in the GS-guided arm was noninferior to the guideline-guided arm (76.7% vs. 71.8%, P < .001), with a noninferiority margin of 20%. Similarly, there were no differences in 28-day mortality, ventilator-free days, or ICU-free days. However, those in the GS-guided group were significantly less likely to receive anti-MRSA (61.2% vs. 100%, P < .001) and anti-Pseudomonal (69.9% vs. 100%, P < .001) agents compared with the guideline-guided group. Although escalation of therapy based on culture results occurred in more patients in the GS-guided group than in the guideline-based group (6.8% vs. 1.0%, P = .03), it does not appear to have led to worse clinical outcomes.

Limitations of the study include the low number of patients with septic shock and/or immunocompromising conditions, as well as the low prevalence of multidrug-resistant organisms in the participating institutions, making the results difficult to generalize to these other settings. Additionally, it may be logistically difficult to obtain GS results in such a timely fashion after diagnosis to have an effect on empiric antibiotic therapy. Nevertheless, this study supports the notion that de-escalation of empiric therapy may be considered based on GS results in certain patients with VAP.

(Yoshimura et al. JAMA Netw Open. 2022;5(4):e226136.)