Author(s):
Tateum Mattingly; Connor King; Jordan Baker; Lynn Chollet-Hinton; Christopher Guidry

Background:

Pneumonia remains the most common ICU-acquired infection. The diagnosis remains difficult as many non-infectious causes can masquerade as pneumonia. A recent randomized trial evaluated antibiotic timing but only evaluated the first infectious episode.

Hypothesis:

The purpose of this study is to identify risk factors for potential recurrent pneumonia.

Methods:

We performed a retrospective analysis of the recent Trial of Antibiotic Restraint in Presumed Pneumonia (TARPP). Which was a multicenter trial of antibiotic initiation strategies. Patients were categorized as having either a single or multiple episodes of suspected pneumonia which was defined as the primary team sending a culture for a potential pneumonia.  Demographics, comorbidities, and outcomes were reviewed. Standard univariate statistical analysis was performed.

Results:

TARPP enrolled 186 patients with 47 patients (25.3%) having at least one additional episode of suspected pneumonia. There were no differences in age or race between the two groups. Patients with recurrent episodes of suspected pneumonia were more likely to identify as Hispanic or Latino (17% vs 4.4%; p=0.009) or to speak Spanish as their primary language (12.8% vs 2.9%; p=0.03). There were no differences in comorbidities, ISS score, or rate of other infection when admitted to the ICU. Patients with recurrent suspected episodes had longer ICU LOS, total days of antibiotics, and longer ventilator days. Patients with recurrent episodes had higher overall rates of culture positivity (97.9% vs 84.2%; p=0.01), however there was no difference in the rate of infection due to non-fermenting gram negatives. Patients with recurrent episodes had lower mortality rates overall compared to those with a single infectious episode (8.5% vs. 21.6%; p=0.045). There was no difference in recurrence based on grouping by Immediate or Specimen-Initiated antibiotics.

Conclusions:

This retrospective analysis suggests that ethnicity and language barriers may be associated with recurrent suspected pneumonia highlighting potential disparities in care. While higher rate of culture positivity is associated potential recurrence, the lower mortality rates in this group suggest a survivorship bias. More work is needed to evaluate the risks for recurrent pneumonia in the ICU.