Despite significant advances in care, survival to discharge for patients who experience cardiac arrest remains dismal, at 10.5% for out-of-hospital cardiac arrest (OHCA) and 23.5% for in-hospital cardiac arrest (IHCA).1 Minimizing infectious complications in patients following cardiac arrest in the ICU is crucial to reducing the risks of morbidity and mortality.
Pneumonia is the most common infection in patients postarrest, with attributable mortality rates of between 21% and 45%.2 Empiric broad-spectrum antibiotic administration for patients postarrest is not well-supported by the current literature, however, and may contribute to antibiotic resistance.
The use of prophylactic antibiotics in patients who are mechanically ventilated has a mixed history, with interventions such as inhaled antibiotics and selective digestive decontamination failing to reduce mortality on a consistent basis. Given the particularly high risk of pneumonia following IHCA and OHCA, there remains an interest in finding a role for prevention in this vulnerable patient population.
Antibiotics in postarrest care
The 2019 Antibiotherapy During Therapeutic Hypothermia to Prevent Infectious Complications (ANTHARTIC) trial randomized patients following cardiac arrest with shockable rhythms to a two-day course of amoxicillin-clavulanate vs placebo, leading to a reduction in rates of ventilator-associated pneumonia in the intervention arm. However, ANTHARTIC failed to show a reduction in overall mortality or ICU length of stay.3
More recently, the Prevention of Early Ventilation-Acquired Pneumonia in Brain-Injured Patients by a Single Dose of Ceftriaxone (PROPHY-VAP) trial identified a mortality benefit in patients who are neurologically injured who require mechanical ventilation from a single dose of prophylactic ceftriaxone following endotracheal intubation.4 Independent of the PROPHY-VAP trial but similar in concept, the Ceftriaxone to PRevent pneumOnia and inflammatTion aftEr Cardiac arresT (PROTECT) trial studied prophylactic ceftriaxone following OHCA. PROTECT randomized 52 patients with any underlying initial heart rhythm who had suffered an OHCA to receive 2 grams intravenously every 12 hours for three days vs a matching placebo dose of 0.9% sodium chloride.2 The primary outcome was to evaluate the development of early-onset pneumonia using a standardized definition. Notable secondary outcomes included late-onset pneumonia (more than four days following intubation), nonpulmonary infections, ICU-free days, ventilator-free days, and mortality at 28 days.
PROTECT failed to show a statistically significant decrease in the rate of early-onset pneumonia. However, there was a significant reduction in the use of open-label antibiotics; the treatment arm also notably had significantly fewer antibiotic resistance genes to common ICU antibiotics identified on surveillance testing when compared with the placebo arm. In-hospital mortality was also lower in the ceftriaxone group (42% vs 73%; RR, 0.58, 95% CI, 0.35-0.96). It is noteworthy that the treatment group had a notably higher rate of bystander CPR (85% vs 62%) and a notably lower average APACHE IV score (71 vs 102), potentially confounding these results.
Given the results of PROPHY-VAP, ANTHARTIC, and PROTECT, we may soon reach a point where it is reasonable to provide a brief duration of prophylactic, relatively narrow-spectrum antibiotics to prevent pneumonia and reduce mortality in patients who are neurologically injured or have recently suffered cardiac arrest. Conceptually, it may be useful to consider this not as empiric therapy but rather as peri-procedural prophylaxis, similar to peri-operative cefazolin before a major surgical procedure. The doses and durations used vary across these studies, although the long half-life of ceftriaxone may lend itself to a single dose for prophylaxis. Further studies will be needed to confirm and refine these findings.
Biomarkers in postarrest care
Another potentially practice-changing update in postarrest management is the use of neuron-specific enolase (NSE) as an adjunct for neuroprognostication. NSE is an isomer of enolase found in neurons that correlates with the degree of neuronal apoptosis secondary to hypoxic-ischemic cerebral injury. In the current European guidelines on postarrestcare, an NSE level > 60 at 48 to 72 hours postarrest is recommended as a predictor of poor neurologic outcomes.5 However, the use of NSE is not currently widespread in North America; most neuroprognostication relies on physical examination, neuroimaging, and electroencephalography (EEG).
NSE has shown reasonable predictive value in the postarrest period. In a retrospective study looking at both patients who experienced IHCA and OHCA with an initial GCS of < 8, the combination of malignant EEG findings and an NSE of > 42 was found to be 100% specific in predicting an unfavorable outcome.5 Even with a benign EEG pattern, an NSE of > 78.2 remained 100% specific for predicting an unfavorable neurologic outcome. Similarly, a prospective observational study published in 2025 including 120 postarrest cases reported that an NSE of > 68 ng/mL at 72 hours postarrest was 82% sensitive and 90% specific for predicting an unfavorable outcome.6
While potentially valuable, it remains unclear to what extent NSE could improve our existing practices in patients postarrest. Biomarkers have received a great deal of attention in other areas of critical care such as sepsis diagnosis, but these biomarkers have had variable impacts on actual management. It remains unclear, for example, whether an elevated NSE value in a patient who is persistently comatose with malignant EEG findings would meaningfully alter such a patient’s prognosis (or persuade a family to alter goals of care). However, there are few decisions more fraught in the ICU than neuroprognostication, and it could be that every additional piece of data would help.
A new frontier
While caring for patients following cardiac arrest in the ICU poses significant challenges, ongoing research continues to improve our understanding and hopefully our management. Prophylactic pneumonia coverage may present a potential mortality benefit without significantly increasing antibiotic resistance within the ICU; and advances in neuron-derived biomarkers, such as NSE, may provide an objective measurement for neuroprognostication.
It is easy to fall into the trap of fatalism when caring for patients postarrest, but the advances in cardiac care over the last several years have nonetheless been meaningful: fever avoidance and temperature management, hemodynamic and respiratory optimization, and—above all else—improvements in CPR and advanced cardiac life support. Prophylactic antibiotics and biomarkers may also have a role to play as we learn more about these most challenging and vulnerable cases.
This article was originally published in the Summer 2026 issue of CHEST Physician.
References
1. Del Rios M, Bartos JA, Panchal AR, et al. Part 1: Executive summary: 2025 American Heart Association guidelines for cardiopulmonary resuscitation and emergency cardiovascular care. Circulation. 2025;152(16_suppl_2):S284-S312. doi:10.1161/CIR.0000000000001372
2. Gagnon DJ, Burkholder KM, Weissman AJ, et al. Ceftriaxone to prevent early-onset pneumonia in comatose patients following out-of-hospital cardiac arrest: a pilot randomized controlled trial and resistome assessment (PROTECT). Chest. 2026;169(1):115-127. doi:10.1016/j.chest.2025.08.007
3. François B, Cariou A, Clère-Jehl R, et al. Prevention of early ventilator-associated pneumonia after cardiac arrest. N Engl J Med. 2019;381(19):1831-1842. doi:10.1056/NEJMoa1812379.
4. Dahyot-Fizelier C, Lasocki S, Kerforne T, et al. Ceftriaxone to prevent early ventilator-associated pneumonia in patients with acute brain injury: a multicentre, randomised, double-blind, placebo-controlled, assessor-masked superiority trial. Lancet Respir Med. 2024;12(5):375-385. doi:10.1016/S2213-2600(23)00471-X
5. Pelle J, Pruvost-Robieux E, Dumas F, et al. (2025). Personalized neuron-specific enolase level based on EEG pattern for prediction of poor outcome after cardiac arrest. Ann Intensive Care. 2025;15(1):11. doi:10.1186/s13613-024-01406-y
6. Badila R, Sava M, Bereanu AS, Neamtu S, Mustatea S, Roman-Filip C. Serum neuron-specific enolase as a predictor of neurological outcomes at hospital discharge in post-cardiac arrest patients: A prospective study. Cureus. 2025;17(12):e99748. doi:10.7759/cureus.99748
