Neurologic Complications of Infective Endocarditis

Abstract & Commentary

Source: García-Cabrera E, et al. Neurological complications of infective endocarditis: Risk factors, outcomes, and impact of cardiac surgery:
A multicenter observational study. Circulation 2013;127:2272-2284.

The relationship between neurologic complications of infective endocarditis (IE), urgent valve surgery, and outcomes is poorly understood. Thus, these investigators performed a retrospective analysis of a prospectively collected database of IE cases from 1984-2009 in seven Spanish hospitals. After excluding patients with isolated right heart and pacemaker lead infections, a total of 1345 patients with left-sided IE who had complete data and follow-up were available. Using the Duke criteria, 93% had definite IE and 7% had possible IE. A neurological event was experienced by 25%. Ischemic and presumably embolic events occurred in 56% of these patients. The majority of these events were small (72%). Cerebral hemorrhage occurred in 18%, encephalitis or meningitis in 24%, and brain abscess in 4%. Multivariate analysis showed that neurologic complications were related to vegetation size > 3 cm (hazard ratio [HR], 1.9; 95% confidence interval [CI], 1.07-3.43; P = 0.029), Staphylococcus aureus infection (HR, 2.47; 95% CI, 1.95-3.15; P < 0.001), anticoagulant therapy (HR, 1.3; 95% CI, 1.00-1.72; P = 0.048), and mitral valve involvement (HR, 1.29; 95% CI, 1.02-1.61; P = 0.03). Anticoagulant therapy was most associated with cerebral hemorrhage (HR, 2.71; 95% CI, 1.54-4.76; P = 0.001). In 86% of the patients with neurologic events, it occurred during the first week after diagnosis of IE — about one-third before antibiotic treatment and two-thirds after the onset of therapy. Hemorrhagic events were more common after therapy had begun. Early mortality (up to 30 days) was 30% and was not influenced by surgical therapy. Neurologic events increased mortality, 45% vs 24% in those without neurologic complications, but only moderate-to-severe ischemic events and hemorrhagic events were significant predictors. Also, mortality was higher if surgery was performed within 30 days of a hemorrhagic event. The authors concluded that moderate-to-severe ischemic stroke and cerebral hemorrhage complicating IE increased mortality. Early antibiotic therapy reduces cerebral events and stopping or not starting anticoagulant therapy may prevent cerebral hemorrhages.


This relatively large database study helps refine the use of anticoagulants and surgery in IE and details the impact of various neurologic complications on management. They show that overall, one-fourth of patients have neurologic complications, most of which are vascular embolic events. About one-third occur before the diagnosis of IE is made and most of the rest occur during the first week of therapy. The most important risk factors for cerebral complications overall are S. aureus infection (HR 2.47) and vegetation size (HR 1.9), but only with vegetations > 3 cm in diameter. Older, smaller studies have shown that vegetation size > 1 cm predicted poor outcomes, suggesting that such vegetations constituted an indication for surgery. However, this was never widely accepted as a criterion for surgery, perhaps because our experience didn’t support the gravity of a 1 cm vegetation vs the risk of surgery in IE. A vegetation of > 3 cm is a different matter and probably will drive surgery. Thus, strong indications for surgery will be a > 3 cm vegetation, failure to control infection, or hemodynamic deterioration that cannot be effectively managed medically. The critical period for infection control is the first week because few cerebral complications occur thereafter.

One-fifth of those with cerebral complications had cerebral hemorrhage. The major risk factor for this subgroup was anticoagulant therapy (HR 2.71). The major reasons for anticoagulant treatment were pre-existing conditions, such as prosthetic valves and atrial fibrillation, and treatment of the neurological complication, especially if it was present at initial presentation. Their results suggest that early anticoagulation for neurologic presentations is unwise if IE is suspected. Also, they recommend that oral anticoagulation be suspended when the diagnosis of IE is made, especially in those at highest risk of cerebral complications. Although this is not a controlled trial, this advice seems prudent.

Surgeons are often reluctant to operate when neurologic complications have occurred, even when clear indications for surgery are present. The results of this study suggest that mild neurological events do not affect the outcomes of surgery and should not preclude or delay surgery if indicated. On the other hand, moderate-to-severe cerebral events clearly increased the risk of surgery and the authors suggest that surgery be delayed for at least 2 weeks in such cases. For cerebral hemorrhage, they suggest delaying 4 weeks.

Overall mortality in this series remained high at 30% in 1 month. If cerebral complications were present, it was 45% and if absent 24%. So clearly, prevention of cerebral events is important. Strategies suggested by this study include early effective antibiotic therapy, judicious use of anticoagulants, and consideration of early surgery with very large vegetation sizes.