Editorial on low-dose acetylsalicylic acid treatment and impact on short-term mortality in Staphylococcus aureus bloodstream infection: a propensity score-matched cohort study

Abstract

The manuscript “Low-Dose Acetylsalicylic Acid Treatment and Impact on Short-Term Mortality in Staphylococcus aureus (S. aureus) Bloodstream Infection: A propensity Score-Matched Cohort Study” published in Critical Care Medicine by Osthoff et al. reported an association of aspirin intake with a reduced short-term mortality. Direct anti-microbial effects of aspirin and its metabolite salicylate were suggested in preclinical studies. Especially intriguing is the inclusion of a control group with Escherichia coli (E. coli) blood stream infections in this study, in which aspirin was not associated with an improved outcome. However, as other observational studies also reported benefits of aspirin in critically ill patients, randomized trials are needed to confirm the effects of low-dose aspirin.

Acetylsalicylic acid (ASA) at low doses is usually regarded as an anti-platelet substance. The mode of action is an irreversible inhibition of the enzyme cyclooxygenase (COX) by acetylation (1,2), which inhibits arachidonic acid induced platelet aggregation even after single oral doses of 162 mg of aspirin (3). This effect is limited to COX-1 within platelets at low doses. At higher doses ASA also inhibits COX-2, which explains its use as an anti-inflammatory and analgetic drug. The use of low-dose ASA in the prevention of thromboembolic events has remained essential up to now (4,5).

However, apart from these well-known effects other beneficial responses were assigned to ASA and subject to investigation. ASA-triggered lipoxins, like other lipid mediators regulated by prostaglandins and COX enzymes, are thought to play an important part in the resolution of inflammation (6,7). Interestingly, at higher doses ASA inhibits the NF-κB pathway (8).

Inhibition of platelets, nowadays identified as “immune” cells contributing to inflammation activation and regulation, may prove beneficial in severe systemic inflammatory responses as inhibition may reduce the platelets’ contribution to such a disease state (8,9). Various observational studies were performed showing overall positive effects of low-dose ASA in community-acquired pneumonia, in critically ill patients or in acute respiratory distress syndrome (10-14). Moreover, a possible role of low-dose ASA in prevention of cancer was suggested (15).

Finally, ASA was demonstrated to exert direct, antimicrobial effects. In a rabbit model of Staphylococcus aureus (S. aureus) endocarditis ASA at 8 mg/kg/day reduced vegetation weight, growth, bacterial density and embolic lesions (16). These observed benefits were at least partially diminished when other doses (4 and 12 mg/kg) were used. It’s important to note that salicylate levels at the dose of 8 mg/kg were ranging between 47 and 53 µg/mL over a period of 12 hours, which is much higher than peak levels measured in healthy volunteers after a single dose of 162 mg of ASA, which were approximately 7.6±1.4 µg/mL (3). After intake of 800 mg of different NO-aspirin formulations mean peak plasma levels ranged between 10 and 21 µg/mL (17). Considering these data from healthy volunteers, rather high doses of aspirin would be necessary to achieve the proposed doses. On the other hand, the interspecies differences have to be considered. Beneficial effects of ASA in combination with ticlopidine were also found in a rat endocarditis model caused by Enterococcus faecalis and Streptococcus gallolyticus (18). Salicylic acid, the main metabolite of ASA, affected the virulence of S. aureus in-vitro by activating a stress response regulon sigma factor β. This stress response reduces α-hemolysin (hla) and fibronectin-binding protein A (fnbA) gene expression (19,20). These results were confirmed in a study investigating not just the adhesive potential of S. aureus but also its invasiveness using human vascular endothelial cells. ASA reduced the virulence of S. aureus in this in vitro model (21). Again doses of 30 and 50 µg/mL were used. However, using a distinct, encapsulated S. aureus strain, ASA enhanced invasiveness in an in vitro model using bovine mammarian endothelial cells (MAC-T) (22). This suggests that the effects of ASA may differ between strains of S. aureus and more data are needed to better define the antimicrobial effects of ASA (22). Interestingly platelets were demonstrated to contribute to biofilm formation and inhibition of platelets by low-dose ASA may reduce biofilm formation and its resistance to antimicrobial substances (23).

In their manuscript Osthoff et al. investigated effects of low-dose ASA in S. aureus bloodstream infection (BSI) (24). Low-dose ASA was associated with a reduced mortality in this propensity score-matched cohort study. Of note, beside ASA intake another difference was evident between groups: in the ASA group significantly more patients were treated with statins compared to the non-ASA group. The intake of statins alongside the intake of low-dose ASA is not merely surprising taking the overlapping indications into account. Although statins were associated with reductions in mortality in other non-interventional studies (25), in the multivariate analysis of this study statins were not associated with an improved clinical outcome. Thus, although it cannot entirely be excluded, statin use should not interfere with the study’s endpoints.

How should the findings of the study be interpreted and integrated into the growing body of evidence that there may be more to ASA than inhibiting platelet aggregation? This retrospective study was well-designed with a large propensity score-matched cohort, groups were well-balanced, except for the statin use, and existing standard practices for treatment and diagnosis of BSI within the hospital improve the quality of data. There is one particular strength, which adds much to the validity of the results: the inclusion of the Escherichia coli (E. coli) BSI control group. Thus, it is tempting to assume that direct antimicrobial effects of ASA against S. aureus cause this benefit. Furthermore it suggests that platelet inhibition in BSI caused by E. coli does not improve survival. After all, it has to be emphasized that although this study certainly adds important information still many questions remain. Observational studies, no matter how well designed and performed, are prone to bias and confounders. To confirm the reported, potentially beneficial effects of treating patients with S. aureus BSI with low-dose ASA, or other reported positive effects of ASA in critical illness, randomized trials are necessary. As the number of observational studies suggesting positive effects of ASA increase the call for such studies becomes louder. It seems likely that realization of such a trial is subject to academic research, as ASA use does not offer enough financial incentives for industrial sponsoring. Until randomized trials are performed, it remains unknown whether the observed effects are true or subject to bias.