TY - JOUR

T1 - Long-term Risk of Hemorrhagic Stroke in Patients With Infective Endocarditis

T2 - A Danish Nationwide Cohort Study

AU - Klein, Christine F

AU - Gørtz, Sanne

AU - Wohlfahrt, Jan

AU - Munch, Tina N

AU - Melbye, Mads

AU - Bundgaard, Henning

AU - Iversen, Kasper K

PY - 2019/2/1

Y1 - 2019/2/1

N2 - Background: The present study aimed to investigate the long-term risk of hemorrhagic stroke (HS) in patients with infective endocarditis (IE).Methods: Using a register-based nationwide cohort of 9 million Danes, we performed propensity score matching between patients with left-sided IE from 1977 to mid-2015 and IE-free individuals (1:10). Follow-up started 1 year after the IE diagnosis. Hazard ratios (HRs) for HS in patients with IE compared with the matched cohort were estimated using Cox regression.Results: During follow-up of 5735 patients with left-sided IE from 1 year after IE diagnosis and up to 37.5 years (median, 6.3 years), 103 cases of HS were observed. Compared with the matched cohort, patients with IE had a higher long-term risk of HS (HR, 1.47; 95% confidence interval, 1.20-1.80; P < .001). The risk of HS was particularly increased in patients within the lowest propensity score quartile (HR, 2.60; 95% confidence interval, 1.89-3.58). Mediation analyses suggested that the increased HS risk could be explained by an indirect effect of mechanical heart valve insertion, atrial fibrillation, or treatment with anticoagulants. The cumulative risk of HS 30 years after start of follow-up was 3.0% in patients with IE.Conclusions: IE does not directly increase the long-term risk of HS. The apparent excess risk of HS in patients with previous IE was explained by mediating factors, including mechanical heart valve insertion, atrial fibrillation, and anticoagulation medication.

AB - Background: The present study aimed to investigate the long-term risk of hemorrhagic stroke (HS) in patients with infective endocarditis (IE).Methods: Using a register-based nationwide cohort of 9 million Danes, we performed propensity score matching between patients with left-sided IE from 1977 to mid-2015 and IE-free individuals (1:10). Follow-up started 1 year after the IE diagnosis. Hazard ratios (HRs) for HS in patients with IE compared with the matched cohort were estimated using Cox regression.Results: During follow-up of 5735 patients with left-sided IE from 1 year after IE diagnosis and up to 37.5 years (median, 6.3 years), 103 cases of HS were observed. Compared with the matched cohort, patients with IE had a higher long-term risk of HS (HR, 1.47; 95% confidence interval, 1.20-1.80; P < .001). The risk of HS was particularly increased in patients within the lowest propensity score quartile (HR, 2.60; 95% confidence interval, 1.89-3.58). Mediation analyses suggested that the increased HS risk could be explained by an indirect effect of mechanical heart valve insertion, atrial fibrillation, or treatment with anticoagulants. The cumulative risk of HS 30 years after start of follow-up was 3.0% in patients with IE.Conclusions: IE does not directly increase the long-term risk of HS. The apparent excess risk of HS in patients with previous IE was explained by mediating factors, including mechanical heart valve insertion, atrial fibrillation, and anticoagulation medication.

U2 - 10.1093/cid/ciy512

DO - 10.1093/cid/ciy512

M3 - Journal article

C2 - 29920590

VL - 68

SP - 668

EP - 675

JO - Clinical Infectious Diseases

JF - Clinical Infectious Diseases

SN - 1058-4838

IS - 4

ER -