By Vibhu Sharma, MD

Assistant Professor of Medicine, University of Colorado, Denver

Dr. Sharma reports no financial relationships relevant to this field of study.

SYNOPSIS: In this prospective, randomized trial, intravenous amoxicillin-clavulanate (dosed three times daily and given for two days) administered to patients admitted with out-of-hospital cardiac arrest due to a shockable rhythm reduced the incidence of early ventilator-associated pneumonia.

SOURCE: François B, Cariou A, Clere-Jehl R, et al; CRICS-TRIGGERSEP Network and the ANTHARTIC Study Group. Prevention of early ventilator-associated pneumonia after cardiac arrest. N Engl J Med 2019;381:1831-1842.

This prospective, randomized controlled trial assessed whether prophylactic antibiotic therapy with intravenous amoxicillin-clavulanate (IVAC) administered to adult patients undergoing targeted temperature management (TTM) to 32-34°C after out-of-hospital cardiac arrest due to a shockable rhythm would reduce the incidence of ventilator-associated pneumonia (VAP). A total of 1,116 patients were assessed for eligibility, and 198 underwent randomization. Most exclusions were because of either a non-shockable rhythm (30%) or in-hospital cardiac arrest (20%). Other reasons for exclusion included preexisting pneumonia, abnormal chest X-ray at presentation, ongoing antibiotic therapy, witnessed aspiration during initial intubation, moribund status, or ≥ 6 hours between return of spontaneous circulation and randomization. The method of TTM was not standardized, but rapid achievement of hypothermia was required. All patients received a bundle of interventions to prevent VAP (e.g., head of the bed elevation, daily spontaneous awake trial [SAT], and daily spontaneous breathing trial [SBT]). Patients were randomized to IVAC 1 g or intravenous (IV) placebo three times daily for two days.

The primary endpoint was the incidence of VAP and required clinical criteria, imaging, and microbiological criteria. The Clinical Pulmonary Infection Score (CPIS) was calculated to determine whether VAP was present, with a score > 6 implying a higher probability of VAP. This score incorporates several criteria with points assigned for each: tracheal secretions characterized as rare/abundant/abundant and purulent, chest X-ray findings, fever or hypothermia, white blood cell count, PaO2/FiO2 ratio, radiographic progression, and exclusion of congestive heart failure (CHF) and acute respiratory distress syndrome (ARDS). A CPIS > 6 was required based on these criteria. In addition, two or more of the following were required: auscultatory findings of pneumonia or consolidation, ventilator changes reflecting worsening ventilation/perfusion (V/Q) mismatch, or worsening PaO2/FiO2 ratio. Radiographic criteria included new or worsening consolidation, and microbiologic criteria included a positive respiratory culture at prespecified thresholds of 106 colony forming units (CFU) for endotracheal aspirates and 104 CFU for bronchoalveolar lavage (BAL) specimens. The specific sampling technique of the lower respiratory tract was “at the discretion of the attending physician.” Blood cultures were obtained upon suspicion for VAP as well. VAP was defined as early if the occurrence was noted within seven days and late if it occurred after seven days of arrest. A final diagnosis of VAP was based on the clinical, radiologic, and microbiologic criteria, with all information available to a committee composed of two experienced intensivists. In the event of disagreement with respect to the diagnosis, a third intensivist arbitrated the final diagnosis.

The investigators reported 80 cases of VAP; however, the adjudicating committee only reported cases with pathogen documentation (60 of the 80). The initial rate of agreement on a VAP diagnosis among these 60 patients was 78%, requiring an adjudicator in the remaining 22%. Among patients receiving the antibiotic intervention, 33% of the sampled lower respiratory tract secretions grew bacterial pathogens, whereas among those receiving placebo, 62% of the sampled secretions grew pathogenic bacteria. Cases deemed to be “colonization” based on culture results were excluded for analysis by the adjudication committee. IVAC reduced the risk of early VAP but not late VAP, with a hazard ratio (HR) of 0.53 (95% confidence interval [CI], 0.31 to 0.92; P = 0.03). There was no difference in ICU length of stay, nonpulmonary secondary infections, or the development of multidrug-resistant organisms out to seven days after treatment with IVAC. Mortality was unaffected by treatment with IVAC.


Patients undergoing TTM are at increased risk for infections, particularly pneumonia. This is hypothesized to be because of the prolonged activation of NF-κβ and augmented generation of cytokines in the setting of hypothermia.1 An impaired immune response to gram-negative bacteria after cardiac arrest also has been hypothesized to play a role regardless of body temperature.2 The patients in this study were a relatively young (median age approximately 60 years), majority male, and presumably community-dwelling group (although this is not specified).

This study demonstrates the difficulty of diagnosing a simple condition for the purposes of a clinical trial. Confirmation of VAP required multiple clinical criteria as detailed earlier, some of which have a questionable interrater reliability (e.g., auscultation). The CPIS is cumbersome to compute and is based on several criteria, some of which are observer-dependent, and, therefore, subjective. Tracheal secretion characteristics and progression of pulmonary opacities are examples. One study found that the specificity of a CPIS > 6 for diagnosing VAP on day 3 was only 47%, with a sensitivity of 89%.3 The authors of the ANTHARTIC study accepted that, “The diagnosis of VAP remains complex owing to considerable heterogeneity in its definition.” A total of 80 patients were reported by the investigators to have VAP based on several criteria; however, only 60 were diagnosed with VAP by the adjudication committee. The initial rate of agreement by the two adjudicators for these 60 patients was 78%, implying that for 33 of the original 80 patients (41%) diagnosed as VAP by the investigators, the diagnosis was uncertain at best.

This study assumes that all VAP is diagnosed only if microbiologic cultures are positive and at a certain clinical threshold. The microbiologic thresholds as described are not used routinely in clinical practice, and the panel making updated recommendations for hospital-acquired pneumonia (HAP)/VAP in 2016 recommended against using them.4 Arguably, from a practical perspective, VAP may be diagnosed and treated in the absence of clearly positive microbiological cultures if multiple other clinical criteria point to the diagnosis. Cases with culture results deemed to be “colonization” were excluded from the analysis and were not reported. The flow diagram for the trial documented only one case with colonization excluded prior to randomization, but the number excluded after randomization due to “colonization” was not reported.

With respect to the bacterial pathogens reported, the majority were susceptible to IVAC and, while this may be true of the pathogens in intensive care units (ICUs) in Europe, this may not be true of ICU settings in other parts of the world. About one-third (35%) of all bacteria isolated were those that colonize the upper respiratory tract and included Haemophilus influenzae, Streptococcus pneumoniae, streptococcal species, Neisseria species, and Moraxella, suggesting that at least some of these events were simply due to aspiration of upper airway secretions that had not been apparent at the time of enrollment. Most of these pathogens were sensitive to the antibiotic being tested. The authors did not report on the number of events (i.e., VAP) diagnosed within the first two days of intubation; these would be classified more typically as community-acquired pneumonia (CAP). The authors also did not report on the proportion of Staphylococcus aureus species isolated that were resistant to methicillin. While staphylococci were overrepresented in the control group (14% vs. 9% in the intervention group), the relative proportions of methicillin-resistant S. aureus (MRSA) would be important for purposes of determining antibiotic efficacy. Finally, the investigators did not define what the diagnoses were in the 20% of cases deemed not to be VAP.

In summary, while the authors made a valiant attempt to diagnose VAP with high certainty, there were shortcomings. A more pragmatic trial may have been to randomize all 80 patients initially diagnosed with VAP. Until more outcome data are available, it may be prudent to prophylactically treat only those post-arrest comatose patients who are younger and community-dwelling (such as those enrolled in this study) and those with hypothermia targeted to 33° C rather than targeted normothermia, given some evidence of immunoplegia in the setting of hypothermia. IVAC is not available for use in the United States. Intravenous ampicillin/sulbactam may be a reasonable alternative once consideration has been given to institution/community-specific antibiograms.


  1. Fairchild KD, Singh IS, Patel S, et al. Hypothermia prolongs activation of NF-kappaB and augments generation of inflammatory cytokines. Am J Physiol Cell Physiol 2004;287:C422-C431.
  2. Beurskens CJ, Horn J, de Boer AM, et al. Cardiac arrest patients have an impaired immune response, which is not influenced by induced hypothermia. Crit Care 2014;18:R162.
  3. Luyt CE, Chastre J, Fagon JY. Value of the clinical pulmonary infection score for the identification and management of ventilator-associated pneumonia. Intensive Care Med 2004;30:844-852.
  4. Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by the Infectious Diseases Society of America and the American Thoracic Society. Clin Infect Dis 2016;63:e61-e111.