ANTHARTIC Trial

Prevention of Early Ventilator-Associated Pneumonia after Cardiac Arrest

CRICS-TRIGGERSEP Network and the ANTHARTIC Study Group

Even though controversies persist, targeted temperature management at 32 to 36°C remains recommended in patients with out-of-hospital cardiac arrest. Since it has beneficial effects on morbidity and mortality. However, targeted temperature management is associated with an increased risk of secondary infections and constitutes an independent risk factor for early ventilator-associated pneumonia.

What’s known?

The key early study suggesting the benefit of single-day administration of cefuroxime in patients with coma was performed more than 20 years ago. Few studies have been conducted since, including one single-centre, small, unblinded, prospective, randomized, controlled trial8 and one prospective, nonrandomized cohort study. Moreover, several retrospective studies showed a decreased incidence of infectious complications and decreased related morbidity when antibiotic therapy was given early to patients receiving targeted temperature management after cardiac arrest.

A randomized trial has previously shown that short-term treatment with amoxicillin-clavulanate failed to prevent ventilator-associated pneumonia in patients with out of hospital cardiac arrest as compared with placebo (50% in both groups). This single-centre trial was underpowered and limited by the absence of adjudication of secondary infections.
Hypothesis – 

Authors hypothesized that systematic administration of empirical 2-day antibiotic therapy could prevent early ventilator-associated pneumonia and related complications in patients with out-of-hospital cardiac arrest treated with targeted temperature management, without clinically or biologically significant adverse effects, and could reduce intensive care unit (ICU) and hospital lengths of stay and medical costs related to ventilator-associated pneumonia.

Population: 

Adult patients (>18 years of age) hospitalized in the ICU after an out-of-hospital cardiac arrest with shockable rhythm and treated with 32–34°C from 16 French ICUs were eligible. 

Design: Double-blind, placebo-controlled, randomized controlled trial. Computer-generated, web-based randomization, stratified according to center, with a fixed block size of four units. Randomization within 6 h of return of spontaneous circulation 

Not included 

• Non-shockable rhythm
• In-hospital cardiac arrest 
• Presence of pneumonia on the initial evaluation, gross aspiration evident during laryngoscopy, confirmed on chest radiography
• Pre-existing lung disease interfering with accurate interpretation of chest radiographs
• Use of extracorporeal support
• Use of antibiotics in the previous week, colonization with MDR organisms
• Beta-lactam allergy; contraindication to amoxicillin-clavulanic acid
• Early limitation of care likely
• Moribund status
• Pregnancy

Intervention: Amoxycillin-clavulanic acid, 1.2 g, IV, 8 hourly for 2 days

Control: Normal saline as placebo

Common to both groups: Rapid induction of hypothermia (cold saline infusion); the temperature was maintained at 32 to 34°C for 24 to 36 hours. Sedation protocol, the use of neuromuscular blockade agents, and the method of targeted temperature management left to clinician preferences. Patients who developed secondary infections were treated with appropriate antibiotics according to local guidelines.  

Primary outcome: Early-onset ventilator-associated pneumonia defined as during the first 7 days of hospitalization. Clinical Pulmonary Infection Score and Sequential Organ Failure Assessment scores, chest radiography, arterial blood gas analysis, blood cultures, and quantitative sampling of the lower respiratory tract by BAL or tracheal aspirate culture before antibiotic administration. Three blinded senior intensivists with experience in VAP trials adjudicated on the presence of VAP. Clinical, radiological, and microbiological criteria. 

Sample size based on an incidence of 68%, 25% reduction; for 90% power and type I error of 5%, 163 patients. Death as a competing event increased the sample size by 15% to 192

When the cut off for “early” VAP was 5 days: 17% vs. 31%; hazard ratio: 0.53; 95% CI, 0.30 to 0.95; P = 0.03. 

Secondary outcomes (antibiotic vs. control): 

• Late ventilator-associated pneumonia (after day 7, until death or discharge from ICU): 4% vs. 5%
• Other nosocomial infections (including bloodstream and urinary tract infections)

• 28-d mortality: 39% overall; 41%  vs. 37% in the control group; no significant difference 
• Intestinal acquisition of multidrug-resistant bacteria on day 7 by detection of such bacteria in stool samples (enterobacteria resistant to third-generation cephalosporins, carbapenemase-producing enterobacteria, methicillin-resistant Staphylococcus aureus, and vancomycin-resistant enterococcus: No emergence of MDR bacteria
• % of days free of antibiotics (apart from prophylactic co-amoxiclav) during ICU stay: 23% vs. 50% 
• ICU LOS: 5 days among ICU survivors [interquartile range, 3.5 to 8.5] vs. 8 days [interquartile range, 3 to 11] and 7 days [interquartile range, 4 to 12] vs. 7 days [interquartile range, 5 to 9], in non-survivors
• Ventilator-free days until day 28: 21 [interquartile range, 0 to 26] vs. 19 [interquartile range, 0 to 25] days
• More than half the had a good neurologic outcome (Cerebral Performance Category of 1 or 2, on a scale from 1 [good cerebral performance] to 5 [death or brain death]) on day 28, and at 3 and 12 months of follow-up. 
• Costs consequence analysis 

Summary of main findings: Among patients resuscitated after of out-of-hospital cardiac arrests with a shockable rhythm who underwent targeted temperature management, two days of intravenous with amoxicillin-clavulanate significantly reduced the incidence of early ventilator-associated pneumonia (within 7 days) compared to placebo.

No difference in Late ventilator-associated pneumonia, other nosocomial infections, 28-d mortality, ICU LOS, ventilator-free days till day 28, intestinal acquisition of MDR organisms on the day. Percentage of antibiotic-free days tended to be lower in the group who had prophylactic antibiotics

Strengths- 

1. Multicentre double-blinded placebo-controlled RCT
2. Strict adherence to protocol
3. Good follow up
4. Adds significant information on the prognosis of this subset of patients.
5. Independent assessment of diagnosis
6. Similarity at baseline with APACHE of 24
7. Good baseline resuscitation effort with a median no-flow time of 2 min
8. Health economic analysis – Total costs were significantly lower by €7,193 (11,518; 2,967) in the antibiotic group due mostly to the lower rate of readmission following the initial admission.

Limitations – 

1. The results cannot be extrapolated to in-hospital cardiac arrests or those with shockable rhythms.
2. Patients with overt aspiration were excluded.
3. Microbiota analysis was designed to detect multidrug-resistant bacteria only and was not repeated after day 7. No data on what happens to gut flora or ecology.
4. whether the present results apply to patients with out-of-hospital cardiac arrest whose condition is managed with a different targeted temperature remains to be determined.
5. Definition of VAP is questionable and limits the generalisability of the results. Different continents use different definitions.
6. The hypothesis itself is questionable as the effect of Augmentin is supposed to last for 24-48 hours and we are looking for its effect lasting for 7 days.
7. The primary outcome is not a clinical endpoint:  committee-adjudicated incidence of VAP
8. No diagnostic criteria mentioned for secondary infections
9. The discrepancy between clinician-reported and adjudicator reported VAP (80 vs. 60 cases;) There was adjudicator consensus only on 78% of cases. 
10. VAP only on the bacteriological confirmation. 
11. Sample size calculation was based on an incidence of VAP of 68%; however, the actual incidence in the study was 29.4%; hence the study was underpowered.

Summary – 

Due to several limitations and underpowered sample size, this trial won’t change our practices.