Continuation of statins after acute intracerebral haemorrhage and survival: the role of inflammation, infection, and indication bias

Abstract

Introduction: Some studies suggest an increased risk of intracerebral haemorrhage (ICH) with HMG-CoA reductase inhibitors (statins). However, other studies focusing on continuation of statins after ICH have shown consistent association with better outcomes. No clear mechanism has been identified, but reduced inflammation and/or infection have been hypothesised. We tested whether statin use, both before and after acute ICH, is associated with risk of death in our cohort. We then investigated potential mediators and tested for indication bias.

Methods: We conducted a secondary analysis of a prospective spontaneous ICH registry including patients between 22/12/2009 to 23/03/2019. Multifactorial logistic regression was used to test for an association between 30-day case fatality (collected by linkage to national databases) and statin exposure before the index ICH and in the 72 hours after onset. A mediation analysis was conducted including available markers of inflammation and infection, including C-reactive protein, white cell count, lymphocyte count, and red cell distribution width on admission and urinary tract infections and pneumonia in the seven days after arrival. A sensitivity analysis, including only those able to take a statin in the 72 hours after onset (not palliated, enteral access), was also performed.

Results: 1423 patients were included. Statin continuation after admission (vs. none) was independently associated with lower case fatality (odds ratio [OR] 0.28; 95% confidence interval [CI] 0.16 to 0.49; p<0.001) but statins taken prior to the ICH and not after (vs. none) were not (OR 0.77; 95% CI 0.51 to 1.15; p=0.2). Receiving a statin after the ICH, regardless of whether it was received before, was similarly associated with 30-day case fatality (OR 0.30; 95% CI 0.18 to 0.49; p<0.001). None of the potential mediators tested were significant. The association between statins and mortality was no longer significant in our sensitivity analysis, excluding those unable to take a statin within 72 hours of their ICH (n=1048; OR 0.62; 95% CI 0.37 to 1.05; p = 0.073).

Conclusion: Exposure to statins (vs. none) after ICH was associated with lower odds of death at 30 days after adjustment for confounders. Available inflammatory markers, urinary tract infections, and pneumonia did not mediate this association. Indication bias is likely to have contributed to the association observed between statins and mortality in prior studies.