At a time of an emerging threat of multi-resistant strains of pathogens,,  and acute awareness of the need for tight stewardship of antimicrobial administration should we be considering routine prophylactic antibiotic administration to patient populations who may benefit based on observational evidence alone? In this issue of the journal, Gagnon and colleagues report an association between prophylactic antibiotic administration and a significant reduction in pneumonia rate in cardiac arrest survivors from 54.9% to 12.6% in their retrospective analysis of data from the Northern Hypothermia Network.

Early onset pneumonia appears to be the most frequently occurring infectious complication in survivors of out-of-hospital cardiac arrest (OHCA), with a reported incidence of up to 65%. The reason this patient cohort is at particular risk of early pneumonia is multifactorial, resulting from both physical injury and host response factors. Aspiration of gastric and oro-pharyngeal contents is likely to occur at the time of collapse as protective airway reflexes are lost, this may be compounded by chest compressions and ventilation via a non-definitive airway device. Ischaemia–reperfusion injury and the post-cardiac arrest syndrome then lead to immunosuppression. Targeted temperature management (TTM) has also been independently associated as a risk factor for the development of early onset pneumonia after OHCA, although in the randomised controlled trial by Nielsen et al., there was no significant difference in the incidence of pneumonia between the 33?°C and 36?°C arms.

The development of pneumonia is clearly a common problem for OHCA survivors on the intensive care unit (ICU), but does it adversely affect survival as one may expect? Previous cohort studies have had conflicting findings showing both no association between the development of early-onset pneumonia and in-hospital mortality,  and an actual increase in survival associated with pneumonia, but this was most likely due to the confounding factor of survivors having a longer length of ICU stay and thus being more likely to develop infection.

Whether or not the development of pneumonia post-OHCA leads to increased in-hospital mortality, it does seem to be associated with both increased length of mechanical ventilation and ICU stay.,  In light of this burden, any means of reducing its incidence should be explored.

Gagnon et al. report a four-fold decrease in the incidence of pneumonia in cardiac arrest survivors when given antibiotic prophylaxis, but no difference in significant functional outcome or ICU length of stay. Given that this is a retrospective cohort study, this relationship can only be deemed to be an association. They may have been comparing different units who either routinely give antibiotic prophylaxis or not. The authors acknowledge this and state they could not control for clinician or hospital bias leading to the reduction of pneumonia. Another shortcoming in the retrospective analysis design of this study was that the registry used only had a limited data set; specific information about the type of antibiotic prophylaxis given, time of initiation and course duration were not available for analysis, nor were any microbiological aspirates recorded, making firm conclusions difficult to make.

Diagnosing the presence of any infective process post-OHCA is difficult. The use of inflammatory markers is unreliable, as C-reactive protein is almost universally high in the first 72?h for these patients and white cell count has been shown to be non-discriminatory, while the use of targeted temperature management means the presence of fever cannot be used during the temperature-controlled period. As these were the variables used to diagnose pneumonia in this study, along with the presence of new or progressive consolidation on a chest radiograph and purulent tracheobronchial secretions, it seems likely that the divide between prophylactic and therapeutic administration of antibiotics was difficult to determine. This is likely to remain an issue for any future non-randomised study into prophylactic antibiotic use post-OHCA.

So is there any other evidence currently available supporting the use of prophylactic antibiotics post-OHCA? Gagnon et al. refer to two randomised controlled trials of antibiotic prophylaxis in intubated patients with structural brain injury (note not OHCA), where the administration of two doses of cefuroxime and three days of ampicillin-sulbactam reduced the incidence of pneumonia from 50 to 24% and 57.9 to 21% respectively.,  This has led to a Level II recommendation by the Brain Trauma Foundation for peri-intubation prophylactic antibiotic administration in severe traumatic brain injury. A retrospective cohort study performed at a single regional cardiac arrest centre showed an improved hospital mortality rate (56.6% vs 75.3%) for patients who received antibiotics during the first seven days of their ICU stay. This may have been attributable to treatment bias for those patients who survived long enough to receive antibiotics, alternatively given the high incidence of bacteraemia (38%) in one prospective study of OHCA patients presenting to an emergency department, this mortality reduction may be to be expected.

Given this conflicting collection of limited quality evidence, the answer to the opening question of prophylactic antibiotics for all has to be ‘no’ for the time being. As Gagnon and colleagues state in their thought provoking paper, what is needed to answer the question is a definitive prospective randomised placebo-controlled trial and hopefully the ANTHARTIC study group will achieve this.

Conflict of interest statement

Dr Matthew Thomas has no financial conflicts of interest but does hold grants from the Resuscitation Council (UK) and the National Institute for Healthcare Research.

References