- Email: [email protected]
Variable Significance of Brain MRI Findings in Infective Endocarditis and Its Effect on Surgical Decisions
ORIGINAL ARTICLE
Variable Significance of Brain MRI Findings in Infective Endocarditis and Its Effect on Surgical Decisions Tia Chakraborty, MD; Eugene Scharf, MD; Daniel DeSimone, MD; Abdelghani El Rafei, MD; Waleed Brinjikji, MD; Larry M. Baddour, MD; Walter Wilson, MD; James M. Steckelberg, MD; Jennifer E. Fugate, DO; Eelco F.M. Wijdicks, MD, PhD; and Alejandro A. Rabinstein, MD Abstract Objective: To determine how brain magnetic resonance imaging (MRI) findings impact clinical outcomes in patients with infective endocarditis (IE) and to propose a management algorithm for patients with neurologic symptoms who are candidates for valve surgery (VS). Patients and Methods: Data from our center were retrospectively reviewed for patients hospitalized with IE between January 1, 2007, and December 31, 2014. Outcomes were postoperative intracerebral hemorrhage (ICH), 6-month mortality, and functional outcome at last follow-up as described by the modified Rankin Scale (mRS) score. Good outcome was defined as an mRS score of 2 or less. Results: A total of 361 patients with IE were identified, including 127 patients (35%) who had MRI. One hundred twenty-six of 361 patients (35%) had neurologic symptoms, which prompted MRI in 79 of 127 patients (62%); 74 of 79 (94%) had acute or subacute MRI abnormalities. One patient with subarachnoid and multifocal ICH on MRI developed postoperative ICH. Patients with VS despite MRI abnormalities had lower 6-month mortality (odds ratio [OR], 0.17; 95% CI, 0.06-0.48; P<.001) and better functional outcome (OR, 4.43; 95% CI, 1.51-13.00; P¼.005). Irrespective of VS, lobar or posterior fossa ICH on MRI was associated with 6-month mortality (OR, 3.58; 95% CI, 1.22-10.50; P¼.02) and territorial ischemic stroke was inversely associated with good mRS (OR, 0.29; 95% CI, 0.13-0.66; P¼.002). In neurologically asymptomatic patients who had VS, MRI findings did not impact 6-month mortality or functional outcomes. Conclusion: Magnetic resonance imaging detects a large number of abnormalities in patients with IE. Preoperative lobar hematoma and large territorial stroke determine outcome irrespective of VS. When indicated, VS increases the odds of a good outcome despite MRI abnormalities. ª 2018 Mayo Foundation for Medical Education and Research
E
mboli to the brain are a frequent complication of infective endocarditis (IE). Both hemorrhagic and ischemic infarcts can be seen in as many as 80% of patients who undergo brain magnetic resonance imaging (MRI), and a sizable proportion are asymptomatic.1,2 As MRI has been shown to have a higher sensitivity for detecting abnormalities than does computed tomography (CT),3 MRI scans are commonly obtained, even in the absence of overt neurologic symptoms by cardiothoracic surgeons, cardiologists, infectious disease specialists, and
n
Mayo Clin Proc. 2019;nn(n):1-9
neurologists.4,5 Once these scans are acquired, cardiovascular surgeons may consider postponement of valve surgery (VS) if the patient does not have considerable valve regurgitation, cardiac failure, or uncontrolled infection. However, not all cerebrovascular abnormalities justify delaying or canceling VS because of the concern for perioperative intracerebral hemorrhage (ICH).6-12 Multiple studies have reported varying results regarding the safety of VS in patients with IE and ischemic stroke (IS) or ICH.8-10 Among patients with indication for VS, the
Mayo Clin Proc. n XXX 2019;nn(n):1-9 n https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org n ª 2018 Mayo Foundation for Medical Education and Research
From the Department of Neurology (T.C., E.S., J.E.F., E.F.M.W., A.A.R.), Department of Infectious Diseases (D.D., A.E.R., L.M.B., W.W., J.M.S.), Department of Cardiovascular Medicine (L.M.B.), and Department of Radiology (W.B.), Mayo Clinic, Rochester MN.
1
MAYO CLINIC PROCEEDINGS
intervention was not performed or it was delayed in up to 20% of patients with stroke.3,13-15 These studies address mortality and more recently timing of surgery, but there are limited data on postoperative functional outcomes related to the nature of cerebral lesions.16-18 In our previous study, we described clinical outcomes in patients who had VS on the basis of particular MRI findings.19 In this investigation, we reviewed how normal vs abnormal brain MRI with or without VS impacted clinical outcomes in patients with IE and aimed to propose a management algorithm for patients with neurologic symptoms or brain MRI lesions who are candidates for VS. PATIENTS AND METHODS Patient Population and Study Design The Mayo Clinic Institutional Review Board approved this study. From the electronic medical record, we identified 2123 patients with suspected or confirmed IE who received care at Mayo Clinic’s campus in Rochester, Minnesota, between January 1, 2007, and December 31, 2014. Patients were excluded if they did not have brain MRI, they did not meet the definition of possible or definite IE based on the modified Duke criteria,20 there was no assessment by a consultant neurologist, they were lost to follow-up, or there was incomplete clinical data (Figure 1). Neurologic symptoms, radiological studies (both pre- and postoperative head CT and MRI in patients who had VS), organism, anticoagulation other than perioperative heparin, renal failure as noted in the medical chart, VS, 6-month mortality, and functional outcomes as described by the modified Rankin Scale21 (mRS) score on the basis of information from the medical chart were recorded. Primary outcomes were postoperative ICH, 6-month mortality, and mRS score at last follow-up. Neurologic Assessment Patients underwent neurologic assessment and physical examination by neurologists throughout the hospital admission, prompted 2
Mayo Clin Proc.
n
by symptoms or for the indication of surgical clearance before VS. Patients were categorized as being asymptomatic or having focal deficits or alterations in the level or content of consciousness. Functional outcomes were categorized using the mRS at last follow-up. Favorable functional outcome was defined as an mRS score of 2 or less. Neuroimaging All images obtained by CT, MRI, magnetic resonance angiography (MRA), and conventional angiography were interpreted by a neuroradiologist. Both radiology reports and source images were further reviewed by one of the study authors blinded to clinical outcomes. For patients who had VS, all imaging information was recorded for imaging performed before and after surgery. Indications for brain imaging included neurologic symptoms and presurgical clearance in neurologically asymptomatic patients. The presence of IS, ICH, microhemorrhage (MH), subarachnoid hemorrhage (SAH), or mycotic aneurysm (MA) were recorded. Acute IS was detected as hyperintensity on diffusion-weighted imaging with hypointense correlate on apparent diffusion coefficient sequences. Small IS was defined as a punctate infarct in the cortex or as maximal diameter less than 20 mm in the basal ganglia or posterior fossa; all other ISs were defined as territorial IS (ie, involving entire territories of branches of a major intracranial artery). Intracerebral hemorrhage was categorized as lobar, deep gray matter, posterior fossa, or multifocal. Microhemorrhages were defined as T2* hypointensities present on gradient echo or susceptibility-weighted imaging sequences and having a diameter of 10 mm or less. Microhemorrhages were counted and categorized by number (<5, 5-10, and >10). Subarachnoid hemorrhage was defined by location as being in the central cisterns, sulcal, or both. The presence of MA was assessed using conventional cerebral angiography and/or MRA of the head. When present, aneurysms were categorized by size (2-16, 7-12, 13-24, and >25 mm), arterial distribution, and whether there was an association with ICH or SAH. XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
Additional Clinical Findings Organisms categorized as “other” included fastidious gram-negative bacteria such as Haemophilus, Aggregatibacter, Cardiobacterium, Eikenella, and Kingella species. Additional information was obtained on potential indications for VS, including valve failure, congestive heart failure, sepsis as defined by systemic inflammatory response system criteria, and IE vegetation size greater than 10 mm. Both valve failure (defined as moderate to severe valve regurgitation) and vegetation size were based on echocardiographic findings. Statistical Analyses Qualitative data were described as nominal or ordinal variables and as mean SD for continuous variables. Data were analyzed using JMP 10.0.0 software (SAS Institute Inc.). We used a t test or the chi-square test (or Fisher exact test for n<5) to assess associations among variables. Odds ratios (ORs) and 95% CIs were computed for nominal data. Statistical testing was done at the 2tailed a level of .05. RESULTS We identified 361 patients with IE during the study period, and 127 patients (35%) had MRI. The mean age was 6215 years, with a male predominance (249 of 361 [69%]). Of the 361 patients, 214 (59%) had native and 147 (41%) had prosthetic valve endocarditis. The most common infectious organisms were Staphylococcus aureus in 95 of 361 (26%), Haemophilus, Aggregatibacter, Cardiobacterium, Eikenella, and Kingella species in 70 (19%), Viridans group Streptococcus in 61 (17%), coagulase-negative staphylococci in 52 (14%), Enterococcus species in 46 (13%), other streptococci in 34 (9%), and fungal in 3 (1%) patients. One hundred seven of 361 patients (30%) were anticoagulated. The median follow-up time after hospital discharge was 10 weeks (range, 1-92 weeks). Overall, 126 of 361 patients (35%) had neurologic symptoms, which prompted MRI in 79 of 127 patients (62%); 74 of 79 (94%) had acute or subacute MRI abnormalities. Of the 127 patients, 38 (30%) had focal Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
2123 Hospitalized with IE
1762 Excluded: indefinite IE diagnosis, no neurologic assessment, lost to follow-up, incomplete patient data
361 With definite or possible IE according to the modified Duke criteria
234 Had no MRI
127 Had MRI
103 Had abnormal MRI
24 Had normal MRI
100 Had VS
78 Had VS
17 Had VS
3 Had postoperative ICH
1 Had postoperative ICH
0 Had postoperative ICH
FIGURE 1. Study design. All patients were admitted to Saint Marys Hospital, one of Mayo Clinic’s hospitals in Rochester, Minnesota, between January 1, 2007, and December 31, 2014. ICH ¼ intracerebral hemorrhage; IE ¼ infective endocarditis; MRI ¼ magnetic resonance imaging; VS ¼ valve surgery.
symptoms, 23 (18%) had altered consciousness, and 18 (14%) had both symptoms. Of the 127 patients who underwent MRI, 103 (81%) had cerebral abnormalities and 24 (19%) had normal findings. Of the 48 neurologically asymptomatic patients who had brain MRI, 29 (60%) had MRI abnormalities. A CT scan was obtained in only 89 of 361 patients (25%), and it was abnormal in 29 of 89 (33%). One hundred forty of 361 patients (39%) had no brain or cerebral vascular imaging throughout their acute hospitalization.
https://doi.org/10.1016/j.mayocp.2018.09.015
3
MAYO CLINIC PROCEEDINGS
FIGURE 2. Intracerebral hemorrhage and perioperative death due to a nonneurologic cause. This 65-year-old man with Streptococcus viridans endocarditis had intracerebral hemorrhage in the posterior fossa (right cerebellum) as well multifocal acute ischemic strokes. Magnetic resonance imaging sequences are shown: (A) T1 with gadolinium, (B) fluid attenuation inversion recovery, and (C) susceptibility-weighted imaging. He underwent valve surgery 8 days after endocarditis diagnosis, during which he died in the operating room because of uncontrollable perioperative hemorrhage at the site of attempted vessel anastomosis of the distal left main coronary artery.
Magnetic resonance imaging findings included IS in 91 of 127 patients (72%), including territorial infarctions in 9 of 91 (10%), small infarctions in 55 (60%), and a combination of territorial and small in 27 (30%). Intracerebral hemorrhage was seen in 22 of 127 patients (17%), and it was lobar in 18 of 22 of them (82%) (Figure 2). Microhemorrhages were present in 43 of 127 patients (34%), including 26 of 43 (61%) with 1 to 5 MHs, 9 (21%) with 5 to 10 MHs, and 8 (19%) with more than 10 MHs. Fifteen of 127 patients (12%) had SAH on MRI: sulcal in 14 cases and diffuse in 1 case. Three of 361 patients (1%) were found to have MA. In all cases, the aneurysm was identified on conventional cerebral angiography (1 also had MRA that failed to visualize the aneurysm). Two patients with MA presented with ICH and 1 with SAH, and they all had coexistent infarctions. One patient had the MA clipped before undergoing VS. We observed no instances of bleeding from an MA after its diagnosis (see Supplemental Material, available online at http://www. mayoclinicproceedings.org). Valve surgery was performed in 95 of 127 patients who had MRI (75%) and 100 of 234 patients without MRI (43%). Among patients with preoperative MRI, 78 of 95 (82%) had abnormal radiological findings (IS in 91, 4
Mayo Clin Proc.
n
ICH in 22, MH in 43, SAH in 15, and MA in 3). Of the 32 patients who had MRI (25%) and then did not have VS, 23 (72%) had neurologic symptoms and 25 (78%) had MRI abnormalities; 29 (91%) had indications for VS. Overall, the presence of MRI abnormalities was not associated with VS (OR, 1.28; 95% CI, 0.48-3.45; P¼.25). The mean time to VS was 5489 days in patients with MRI abnormalities vs 65154 days without MRI abnormalities (P¼.79). Of the 89 patients without preoperative MRI, the head CT scan was abnormal in 29 (33%). Of the 195 patients who had VS, 22 (11%) had postoperative CT; 17 of 22 (77%) had scans that were prompted by new or worsening neurologic symptoms, and 4 (18%) had new abnormalities including ICH in 1, SAH in 1, and IS in 2 patients. Six-month mortality in patients with and without VS was 76 of 361 (21%) and 159 of 361 (44%) patients had good mRS at last follow-up. Six-month mortality in patients with VS was 22 of 195 (11%) and 54 of 166 (33%) patients who did not have VS (OR, 0.26; 95% CI, 0.15-0.46; P<.001). Good mRS at last follow-up was present in 102 of 195 patients who had VS (52%) vs 57 of 166 patients who did not undergo VS (34%) (OR, 2.10; 95% CI, 1.36-3.21; P<.001). XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
FIGURE 3. Postoperative intracerebral hemorrhage. This 51-year-old woman with Staphylococcus aureus endocarditis initially presented with dysarthria, right hemiparesis, and right hemisensory changes. Magnetic resonance imaging sequences obtained after head computed tomography are shown: (A) fluid attenuation inversion recovery and (B) susceptibility-weighted imaging showed multifocal ischemic stroke with intracerebral hemorrhage in the left parietal lesion and sulcal subarachnoid hemorrhage. She underwent valve surgery 3 days after infective endocarditis diagnosis in the setting of sepsis. She was given 1 dose of warfarin 2 days after surgery when postoperative (C) head computed tomography prompted by persistent somnolence and decreased respiratory drive showed new acute right cerebellar hemorrhage and acute hemorrhagic conversion of multifocal ischemic stroke. She had a modified Rankin Scale score of 5 at last follow-up.
Four patients who underwent VS had perioperative symptomatic hemorrhage; only 1 of them had preoperative MRI abnormalities and the other 3 had not had MRI. One patient with SAH and multifocal ICH seen on the initial CT scan with subsequent MRI had postoperative hemorrhage found after postoperative encephalopathy and labored breathing developed (Figure 3). This patient’s time from IE diagnosis to surgery was 3 days, and the mRS score at last follow-up was 5. No patients with IS (regardless of size), MH (regardless of number), or MA with MRI who underwent VS had postoperative ICH. Three 100 patients who did not have MRI (3%) had postoperative ICH; 1 patient had SAH seen on CT and underwent VS 1 day after IE diagnosis with an mRS score of 0 at last follow-up. Another patient had posterior fossa ICH detected on CT who underwent VS 16 days after IE diagnosis and was dead at last follow-up beyond 6 months after diagnosis. The third patient had no symptoms or imaging, had neurologic symptoms 18 days after VS in the setting of heparin for dialysis catheter placement, and then had an acute subdural hematoma and SAH. This patient had VS 2 days Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
after IE diagnosis and was dead at last follow-up within 6 months of IE diagnosis. The Table lists mortality and functional outcome rates according to MRI and VS; overall, patients who had VS despite MRI abnormalities had better clinical outcomes. Irrespective of VS, lobar or posterior fossa ICH on MRI was associated with higher odds of 6-month mortality (OR, 3.58; 95% CI, 1.22-10.50; P¼.02) and displayed a trend toward worse functional outcomes at last follow-up (OR, 0.56; 95% CI, 0.21-1.49; P¼.24). Territorial infarctions on MRI were inversely correlated with favorable mRS at last follow-up (OR, 0.29; 95% CI, 0.130.66; P¼.002) and displayed a trend toward higher 6-month mortality (OR, 1.71; 95% CI, 0.64-4.57; P¼.28). Neurologic symptoms in patients who underwent MRI were not predictive of 6-month mortality (OR, 1.67; 95% CI, 0.59-4.57; P¼.34) or good mRS at last follow-up (OR, 0.58; 95% CI, 0.281.20; P¼.14). In neurologically asymptomatic patients who had VS, MRI findings did not impact 6-month mortality (OR, 0.91; 95% CI, 0.27-3.11; P>.99) or good mRS at last follow-up (OR, 1.50; 95% CI, 0.703.21; P¼.29).
https://doi.org/10.1016/j.mayocp.2018.09.015
5
MAYO CLINIC PROCEEDINGS
TABLE. MRI, VS, and Patient Outcomesa,b Outcome Abnormal MRI with VS Abnormal MRI without VS Normal MRI with VS Normal MRI without VS
6-mo mortality 9/78 (12)
Good mRS at last follow-up
P<.001 OR, 0.17 (95% CI, 0.06-0.48)
41/78 (53)
P¼.29c
14/17 (82)
11/25 (44) 0/17 (0)
5/25 (20)
1/7 (14)
No MRI with VS
13/100 (13)
No MRI without VS
42/134 (31)
P¼.005 OR, 4.43 (95% CI, 1.51-13.00) P¼1.00 OR, 0.77 (95% CI, 0.07-9.08)
6/7 (86) P¼.001 OR, 0.33 (95% CI, 0.16-0.65)
47/100 (47)
P¼.05 OR, 1.67 (95% CI, 0.99-2.88)
46/134 (34)
MRI ¼ magnetic resonance imaging; mRS ¼ modified Rankin Scale; OR ¼ odds ratio; VS ¼ valve surgery. Data are presented as n/N (percentage) or as P value, OR (95% CI). c No OR could be calculated because there were no patients with normal MRI and VS who had 6-mo mortality. a
b
DISCUSSION Our study confirmed that most patients with IE have cerebral lesions on MRI. Just 1 patient who had preoperative CT revealing hemorrhage with subsequent MRI before VS had postoperative symptomatic ICH, suggesting a low risk of peri- or postoperative hemorrhage irrespective of preoperative MRI findings. In our institution MRI was ordered often, but it is unclear how MRI abnormalities influenced surgical decisions on VS. Therefore, it is reasonable to conclude that the use of MRI is equivocal, and in contrast to current guidelines, we do not recommend brain MRI for screening of patients with IE, as its findings should not delay surgery in the absence of hemorrhage or large territorial IS visible on the CT scan.4,5 CT of the brain remains the first line of testing for patients with neurologic symptoms. On the basis of our experience, we suggest the approach shown in Figure 4 for the use and interpretation of neuroimaging in patients with IE. CT scans should suffice to evaluate for territorial infarction or hemorrhage before VS in patients with neurologic symptoms. The added value of MRI is unclear, but MRA or cerebral angiogram can be considered in a select group of patients. Once an infectious aneurysm is found, occlusion can be attempted. However, postoperative ICH is not always avoided by occluding the visible aneurysm because those patients already have pan-vasculitis and they can 6
Mayo Clin Proc.
n
bleed from a different vessel.22 Only 1 patient with hemorrhage on MRI in our cohort developed postoperative ICH despite the high proportion of patients with cerebral lesions on MRI. Furthermore, no patient with MH developed postoperative ICH, which may be a less common complication of VS than previously considered. Our detailed analysis of brain MRI findings in patients with IE also confirms that cerebrovascular abnormalities are common, including a high proportion of clinically silent ischemic and hemorrhagic lesions. Ninety-four percent of patients with neurologic symptoms had cerebral lesions on MRI in the present series. Ischemic stroke was the most common finding in 72%, followed by MH in 34%, ICH in 17%, and SAH in 12%. Previous studies have reported rates of IS ranging from 37% to 83% and rates of MH between 57% and 94%.1,15,23-25 The higher rate of MH in other studies may be attributable to the fact that hemosiderin-sensitive MRI sequences were not routinely used in the earlier years of our study time frame. In addition, Goulenok et al14 reported a higher rate of MA (23%) and SAH (17%) in their prospective study of 30 patients. This may be due to selection bias, with all patients undergoing MRI within the first week of admission, including MRA with subsequent catheter angiography in cases of suspected MA. Treatment of IE must ultimately weigh the benefit of VS against postoperative risk XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
Neurologic symptoms
Obtain CT head
Equivocal CT with unresponsiveness or territorial syndrome on examination
Normal, small IS, deep ICH, or MH
Obtain MRI scan Early VS
Territorial IS, or lobar or posterior fossa ICH
Weigh benefit, timing, and indications of VS against risk of postoperative ICH
SAH
Obtain MRA or conventional angiography to assess for MA
No MA
MA
Consider endovascular treatment
FIGURE 4. Approach to the patient with IE and neurologic symptoms. Patients without neurologic symptoms do not need neuroimaging unless symptoms arise. In patients with neurologic symptoms, head CT scans may initially be obtained. Based on CT findings, MRI may be considered and/or vessel imaging if MA is suspected as a source of hemorrhage seen on CT. When a mycotic aneurysm is detected in patients with IE, direct surgery may be an alternative. CT ¼ computed tomography; ICH ¼ intracerebral hemorrhage; IE ¼ infective endocarditis; IS ¼ ischemic stroke; MA ¼ mycotic aneurysm; MH ¼ microhemorrhage; MRA ¼ magnetic resonance angiography; MRI ¼ magnetic resonance imaging; SAH ¼ subarachnoid hemorrhage; VS ¼ valve surgery.
of hemorrhage, particularly in patients at higher risk of poor outcomes without surgery (such as those with severe heart failure, refractory sepsis, or large vegetations26). Patients with severe IE and cerebral embolism have a substantial risk of recurrent embolism if there is no prompt response to antibiotic therapy. In these cases, VS may be the optimal strategy to prevent recurrent brain lesions. Rates of 6-month mortality were lower (12% vs 44%) and favorable functional outcomes at last follow-up were higher (53% vs 20%) in patients with abnormal MRI scans who had VS than in those who did not have VS in our study. In comparison, our series reported a higher rate of favorable outcome (mRS score2) at last follow-up (52% at a median Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
follow-up of 10 weeks postdischarge) than did the previous series by Sonneville et al16 (29% with a mRS score of 3 at w12 weeks postdischarge). This discrepancy may be attributable to the fact that the cohort of Sonneville et al was restricted to patients admitted to the intensive care unit, unlike ours that also included noneintensive care unit patients. Our study adds information on prognostic factors. We found that territorial IS and lobar or posterior fossa ICH were associated with a poor outcome. Our study has several limitations. First, this study was observational and retrospective and the decision to proceed with surgery and surgical timing were determined by the treating team; thus, selection bias may have influenced the very low risk of perioperative
https://doi.org/10.1016/j.mayocp.2018.09.015
7
MAYO CLINIC PROCEEDINGS
cerebrovascular complications. Second, hemosiderin-sensitive sequences were not always acquired during the early years of the study period; the inclusion of some MRI scans lacking these sequences may have resulted in an underestimation of the number of MH. Third, not all patients with ICH or SAH had vessel imaging to investigate the presence of MA; our study may underestimate the number of patients with MA. Fourth, our study was limited to patients who had VS with available postoperative follow-up; our study may underestimate the number of patients with postoperative neurologic complications after hospital discharge if they did not have further followup care at our institution. Lastly, patients who were in the nonsurgical cohort may have included patients who had indications for VS but were too ill to tolerate surgery, leading to a poor clinical outcome in comparison with those with VS.
CONCLUSION Most patients with IE who underwent brain MRI had evidence of cerebrovascular complications, either symptomatic or silent. Threequarters of them underwent VS, a surgical rate that was much higher than that in patients who had normal MRI or did not have MRI. Thus, the additional value of brain MRI in the management of patients with IE remains uncertain. Postoperative hemorrhage may be a less common complication of VS than previously thought, and evidence of cerebrovascular lesions on brain MRI may not be generally considered contraindications for surgery, particularly when small, even if multiple (eg, multiple small ISs and multiple MHs). Although the risk of postoperative hemorrhage must always be weighed against the benefit of VS, brain MRI should not be considered indispensable for surgical clearance. SUPPLEMENTAL ONLINE MATERIAL Supplemental material can be found online at: http://www.mayoclinicproceedings.org. Supplemental material attached to journal articles has not been edited, and the authors take responsibility for the accuracy of all data. 8
Mayo Clin Proc.
n
Abbreviations and Acronyms: CT = computed tomography; ICH = intracerebral hemorrhage; IE = infective endocarditis; IS = ischemic stroke; MA = mycotic aneurysm; MH = microhemorrhage; MRA = magnetic resonance angiography; MRI = magnetic resonance imaging; mRS = modified Rankin Scale; OR = odds ratio; SAH = subarachnoid hemorrhage; VS = valve surgery
Potential Competing Interests: Dr Scharf has received small honorarium for MOC lecture series from the American Academy of Neurology. Dr Wijdicks has received honorarium and travel expenses for Symposium on Organ Donation Leuven, Belgium, and royalties from Oxford University Press. Dr Rabinstein has received consultancy fees from SpaceWorks and royalties from Oxford, Elsevier, and UpToDate. The rest of the authors report no competing interests. Correspondence: Address to Tia Chakraborty, MD, Department of Neurology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 ([email protected]).
REFERENCES 1. Hess A, Klein I, Iung B, et al. Brain MRI findings in neurologically asymptomatic patients with infective endocarditis. AJNR Am J Neuroradiol. 2013;34(8):1579-1584. 2. Cooper HA, Thompson EC, Laureno R, et al. Subclinical brain embolization in left-sided infective endocarditis: results from the evaluation by MRI of the brains of patients with left-sided intracardiac solid masses (EMBOLISM) pilot study. Circulation. 2009;120(7):585-591. 3. Duval X, Iung B, Klein I, et al; IMAGE (Resonance Magnetic Imaging at the Acute Phase of Endocarditis) Study Group. Effect of early cerebral magnetic resonance imaging on clinical decisions in infective endocarditis: a prospective study. Ann Intern Med. 2010;152(8):497-504. W175. 4. Baddour LM, Wilson WR, Bayer AS, et al; American Heart Association Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease of the Council on Cardiovascular Disease in the Young, Council on Clinical Cardiology, Council on Cardiovascular Surgery and Anesthesia, and Stroke Council. Infective endocarditis in adults: diagnosis, antimicrobial therapy, and management of complications: a scientific statement for healthcare professionals from the American Heart Association. Circulation. 2015;132(15):1435-1486 [published corrections appear in Circulation. 2015 and Circulation. 2016]. 5. Habib G, Lancelloti P, Antunes MJ, et al; ESC Scientific Document Group. 2015 ESC Guidelines for the management of infective endocarditis: the Task Force for the Management of Infective Endocarditis of the European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J. 2015;36(44): 3075-3128. 6. Gaca JG, Sheng S, Daneshmand MA, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg. 2011;141(1):98-106. e101-e102. 7. Chu VH, Park LP, Athan E, et al. International Collaboration on Endocarditis (ICE) Investigators. Association between surgical indications, operative risk, and clinical outcome in infective endocarditis: a prospective study from the International Collaboration on Endocarditis. Circulation. 2015;131(2):131-140. 8. Kang DH, Kim YJ, Kim SH, et al. Early surgery versus conventional treatment for infective endocarditis. N Engl J Med. 2012;366(26):2466-2473. XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
9. Anantha Narayanan M, Mahfood Haddad T, Kalil AC, et al. Early versus late surgical intervention or medical management for infective endocarditis: a systematic review and meta-analysis. Heart. 2016;102(12):950-957. 10. Okita Y, Minakata K, Yasuno S, et al. Optimal timing of surgery for active infective endocarditis with cerebral complications: a Japanese multicentre study. Eur J Cardiothorac Surg. 2016; 50(2):374-382. 11. Thuny F, Avierinos JF, Tribouilloy C, et al. Impact of cerebrovascular complications on mortality and neurologic outcome during infective endocarditis: a prospective multicentre study. Eur Heart J. 2007;28(9):1155-1161. 12. Wilbring M, Irmscher L, Alexiou K, Matschke K, Tugtekin SM. The impact of preoperative neurological events in patients suffering from native infective valve endocarditis. Interact Cardiovasc Thorac Surg. 2014;18(6):740-747. 13. García-Cabrera E, Fernández-Hidalgo N, Almirante B, et al; Group for the Study of Cardiovascular Infections of the Andalusian Society of Infectious Diseases; Spanish Network for Research in Infectious Diseases. Neurological complications of infective endocarditis: risk factors, outcome, and impact of cardiac surgery: a multicenter observational study. Circulation. 2013; 127(23):2272-2284. 14. Goulenok T, Klein I, Mazighi M, et al; IMAGE Study Group. Infective endocarditis with symptomatic cerebral complications: contribution of cerebral magnetic resonance imaging. Cerebrovasc Dis. 2013;35(4):327-336. 15. Iung B, Tubiana S, Klein I, et al; ECHO-IMAGE Study Group. Determinants of cerebral lesions in endocarditis on systematic cerebral magnetic resonance imaging: a prospective study. Stroke. 2013;44(11):3056-3062. 16. Sonneville R, Mirabel M, Hajage D, et al; ENDOcardite en REAnimation Study Group. Neurologic complications and outcomes of infective endocarditis in critically ill patients: the
Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
ENDOcardite en REAnimation prospective multicenter study. Crit Care Med. 2011;39(6):1474-1481. Yanagawa B, Pettersson GB, Habib G, et al. Surgical management of infective endocarditis complicated by embolic stroke: practical recommendations for clinicians. Circulation. 2016; 134(17):1280-1292. Hodges KE, Hussain ST, Stewart WJ, Pettersson GB. Surgical management of infective endocarditis complicated by ischemic stroke. J Card Surg. 2017;32(1):9-13. Chakraborty T, Scharf E, Rabinstein AA, et al. Utility of brain magnetic resonance imaging in the surgical management of infective endocarditis. J Stroke Cerebrovasc Dis. 2017;26(11): 2527-2535. Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis. 2000;30(4):633-638. Sulter G, Steen C, De Keyser J. Use of the Barthel index and modified Rankin scale in acute stroke trials. Stroke. 1999; 30(8):1538-1541. van de Beek D, Rabinstein AA, Peters SG, Cloft H, Wijdicks EF. Staphylococcus endocarditis associated with infectious vasculitis and recurrent cerebral hemorrhages. Neurocrit Care. 2008;8(1): 48-52. Klein I, Iung B, Labreuche J, et al; IMAGE Study Group. Cerebral microbleeds are frequent in infective endocarditis: a casecontrol study. Stroke. 2009;40(11):3461-3465. Morofuji Y, Morikawa M, Yohei T, et al. Significance of the T2*weighted gradient echo brain imaging in patients with infective endocarditis. Clin Neurol Neurosurg. 2010;112(5):436-440. Malhotra A, Schindler J, Mac Grory B, et al. Cerebral microhemorrhages and meningeal siderosis in infective endocarditis. Cerebrovasc Dis. 2017;43(1-2):59-67. Behrouz R. Preoperative cerebrovascular evaluation in patients with infective endocarditis. Clin Cardiol. 2015;38(7):439-442.
https://doi.org/10.1016/j.mayocp.2018.09.015
9
Variable Significance of Brain MRI Findings in Infective Endocarditis and Its Effect on Surgical Decisions Tia Chakraborty, MD; Eugene Scharf, MD; Daniel DeSimone, MD; Abdelghani El Rafei, MD; Waleed Brinjikji, MD; Larry M. Baddour, MD; Walter Wilson, MD; James M. Steckelberg, MD; Jennifer E. Fugate, DO; Eelco F.M. Wijdicks, MD, PhD; and Alejandro A. Rabinstein, MD Abstract Objective: To determine how brain magnetic resonance imaging (MRI) findings impact clinical outcomes in patients with infective endocarditis (IE) and to propose a management algorithm for patients with neurologic symptoms who are candidates for valve surgery (VS). Patients and Methods: Data from our center were retrospectively reviewed for patients hospitalized with IE between January 1, 2007, and December 31, 2014. Outcomes were postoperative intracerebral hemorrhage (ICH), 6-month mortality, and functional outcome at last follow-up as described by the modified Rankin Scale (mRS) score. Good outcome was defined as an mRS score of 2 or less. Results: A total of 361 patients with IE were identified, including 127 patients (35%) who had MRI. One hundred twenty-six of 361 patients (35%) had neurologic symptoms, which prompted MRI in 79 of 127 patients (62%); 74 of 79 (94%) had acute or subacute MRI abnormalities. One patient with subarachnoid and multifocal ICH on MRI developed postoperative ICH. Patients with VS despite MRI abnormalities had lower 6-month mortality (odds ratio [OR], 0.17; 95% CI, 0.06-0.48; P<.001) and better functional outcome (OR, 4.43; 95% CI, 1.51-13.00; P¼.005). Irrespective of VS, lobar or posterior fossa ICH on MRI was associated with 6-month mortality (OR, 3.58; 95% CI, 1.22-10.50; P¼.02) and territorial ischemic stroke was inversely associated with good mRS (OR, 0.29; 95% CI, 0.13-0.66; P¼.002). In neurologically asymptomatic patients who had VS, MRI findings did not impact 6-month mortality or functional outcomes. Conclusion: Magnetic resonance imaging detects a large number of abnormalities in patients with IE. Preoperative lobar hematoma and large territorial stroke determine outcome irrespective of VS. When indicated, VS increases the odds of a good outcome despite MRI abnormalities. ª 2018 Mayo Foundation for Medical Education and Research
E
mboli to the brain are a frequent complication of infective endocarditis (IE). Both hemorrhagic and ischemic infarcts can be seen in as many as 80% of patients who undergo brain magnetic resonance imaging (MRI), and a sizable proportion are asymptomatic.1,2 As MRI has been shown to have a higher sensitivity for detecting abnormalities than does computed tomography (CT),3 MRI scans are commonly obtained, even in the absence of overt neurologic symptoms by cardiothoracic surgeons, cardiologists, infectious disease specialists, and
n
Mayo Clin Proc. 2019;nn(n):1-9
neurologists.4,5 Once these scans are acquired, cardiovascular surgeons may consider postponement of valve surgery (VS) if the patient does not have considerable valve regurgitation, cardiac failure, or uncontrolled infection. However, not all cerebrovascular abnormalities justify delaying or canceling VS because of the concern for perioperative intracerebral hemorrhage (ICH).6-12 Multiple studies have reported varying results regarding the safety of VS in patients with IE and ischemic stroke (IS) or ICH.8-10 Among patients with indication for VS, the
Mayo Clin Proc. n XXX 2019;nn(n):1-9 n https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org n ª 2018 Mayo Foundation for Medical Education and Research
From the Department of Neurology (T.C., E.S., J.E.F., E.F.M.W., A.A.R.), Department of Infectious Diseases (D.D., A.E.R., L.M.B., W.W., J.M.S.), Department of Cardiovascular Medicine (L.M.B.), and Department of Radiology (W.B.), Mayo Clinic, Rochester MN.
1
MAYO CLINIC PROCEEDINGS
intervention was not performed or it was delayed in up to 20% of patients with stroke.3,13-15 These studies address mortality and more recently timing of surgery, but there are limited data on postoperative functional outcomes related to the nature of cerebral lesions.16-18 In our previous study, we described clinical outcomes in patients who had VS on the basis of particular MRI findings.19 In this investigation, we reviewed how normal vs abnormal brain MRI with or without VS impacted clinical outcomes in patients with IE and aimed to propose a management algorithm for patients with neurologic symptoms or brain MRI lesions who are candidates for VS. PATIENTS AND METHODS Patient Population and Study Design The Mayo Clinic Institutional Review Board approved this study. From the electronic medical record, we identified 2123 patients with suspected or confirmed IE who received care at Mayo Clinic’s campus in Rochester, Minnesota, between January 1, 2007, and December 31, 2014. Patients were excluded if they did not have brain MRI, they did not meet the definition of possible or definite IE based on the modified Duke criteria,20 there was no assessment by a consultant neurologist, they were lost to follow-up, or there was incomplete clinical data (Figure 1). Neurologic symptoms, radiological studies (both pre- and postoperative head CT and MRI in patients who had VS), organism, anticoagulation other than perioperative heparin, renal failure as noted in the medical chart, VS, 6-month mortality, and functional outcomes as described by the modified Rankin Scale21 (mRS) score on the basis of information from the medical chart were recorded. Primary outcomes were postoperative ICH, 6-month mortality, and mRS score at last follow-up. Neurologic Assessment Patients underwent neurologic assessment and physical examination by neurologists throughout the hospital admission, prompted 2
Mayo Clin Proc.
n
by symptoms or for the indication of surgical clearance before VS. Patients were categorized as being asymptomatic or having focal deficits or alterations in the level or content of consciousness. Functional outcomes were categorized using the mRS at last follow-up. Favorable functional outcome was defined as an mRS score of 2 or less. Neuroimaging All images obtained by CT, MRI, magnetic resonance angiography (MRA), and conventional angiography were interpreted by a neuroradiologist. Both radiology reports and source images were further reviewed by one of the study authors blinded to clinical outcomes. For patients who had VS, all imaging information was recorded for imaging performed before and after surgery. Indications for brain imaging included neurologic symptoms and presurgical clearance in neurologically asymptomatic patients. The presence of IS, ICH, microhemorrhage (MH), subarachnoid hemorrhage (SAH), or mycotic aneurysm (MA) were recorded. Acute IS was detected as hyperintensity on diffusion-weighted imaging with hypointense correlate on apparent diffusion coefficient sequences. Small IS was defined as a punctate infarct in the cortex or as maximal diameter less than 20 mm in the basal ganglia or posterior fossa; all other ISs were defined as territorial IS (ie, involving entire territories of branches of a major intracranial artery). Intracerebral hemorrhage was categorized as lobar, deep gray matter, posterior fossa, or multifocal. Microhemorrhages were defined as T2* hypointensities present on gradient echo or susceptibility-weighted imaging sequences and having a diameter of 10 mm or less. Microhemorrhages were counted and categorized by number (<5, 5-10, and >10). Subarachnoid hemorrhage was defined by location as being in the central cisterns, sulcal, or both. The presence of MA was assessed using conventional cerebral angiography and/or MRA of the head. When present, aneurysms were categorized by size (2-16, 7-12, 13-24, and >25 mm), arterial distribution, and whether there was an association with ICH or SAH. XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
Additional Clinical Findings Organisms categorized as “other” included fastidious gram-negative bacteria such as Haemophilus, Aggregatibacter, Cardiobacterium, Eikenella, and Kingella species. Additional information was obtained on potential indications for VS, including valve failure, congestive heart failure, sepsis as defined by systemic inflammatory response system criteria, and IE vegetation size greater than 10 mm. Both valve failure (defined as moderate to severe valve regurgitation) and vegetation size were based on echocardiographic findings. Statistical Analyses Qualitative data were described as nominal or ordinal variables and as mean SD for continuous variables. Data were analyzed using JMP 10.0.0 software (SAS Institute Inc.). We used a t test or the chi-square test (or Fisher exact test for n<5) to assess associations among variables. Odds ratios (ORs) and 95% CIs were computed for nominal data. Statistical testing was done at the 2tailed a level of .05. RESULTS We identified 361 patients with IE during the study period, and 127 patients (35%) had MRI. The mean age was 6215 years, with a male predominance (249 of 361 [69%]). Of the 361 patients, 214 (59%) had native and 147 (41%) had prosthetic valve endocarditis. The most common infectious organisms were Staphylococcus aureus in 95 of 361 (26%), Haemophilus, Aggregatibacter, Cardiobacterium, Eikenella, and Kingella species in 70 (19%), Viridans group Streptococcus in 61 (17%), coagulase-negative staphylococci in 52 (14%), Enterococcus species in 46 (13%), other streptococci in 34 (9%), and fungal in 3 (1%) patients. One hundred seven of 361 patients (30%) were anticoagulated. The median follow-up time after hospital discharge was 10 weeks (range, 1-92 weeks). Overall, 126 of 361 patients (35%) had neurologic symptoms, which prompted MRI in 79 of 127 patients (62%); 74 of 79 (94%) had acute or subacute MRI abnormalities. Of the 127 patients, 38 (30%) had focal Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
2123 Hospitalized with IE
1762 Excluded: indefinite IE diagnosis, no neurologic assessment, lost to follow-up, incomplete patient data
361 With definite or possible IE according to the modified Duke criteria
234 Had no MRI
127 Had MRI
103 Had abnormal MRI
24 Had normal MRI
100 Had VS
78 Had VS
17 Had VS
3 Had postoperative ICH
1 Had postoperative ICH
0 Had postoperative ICH
FIGURE 1. Study design. All patients were admitted to Saint Marys Hospital, one of Mayo Clinic’s hospitals in Rochester, Minnesota, between January 1, 2007, and December 31, 2014. ICH ¼ intracerebral hemorrhage; IE ¼ infective endocarditis; MRI ¼ magnetic resonance imaging; VS ¼ valve surgery.
symptoms, 23 (18%) had altered consciousness, and 18 (14%) had both symptoms. Of the 127 patients who underwent MRI, 103 (81%) had cerebral abnormalities and 24 (19%) had normal findings. Of the 48 neurologically asymptomatic patients who had brain MRI, 29 (60%) had MRI abnormalities. A CT scan was obtained in only 89 of 361 patients (25%), and it was abnormal in 29 of 89 (33%). One hundred forty of 361 patients (39%) had no brain or cerebral vascular imaging throughout their acute hospitalization.
https://doi.org/10.1016/j.mayocp.2018.09.015
3
MAYO CLINIC PROCEEDINGS
FIGURE 2. Intracerebral hemorrhage and perioperative death due to a nonneurologic cause. This 65-year-old man with Streptococcus viridans endocarditis had intracerebral hemorrhage in the posterior fossa (right cerebellum) as well multifocal acute ischemic strokes. Magnetic resonance imaging sequences are shown: (A) T1 with gadolinium, (B) fluid attenuation inversion recovery, and (C) susceptibility-weighted imaging. He underwent valve surgery 8 days after endocarditis diagnosis, during which he died in the operating room because of uncontrollable perioperative hemorrhage at the site of attempted vessel anastomosis of the distal left main coronary artery.
Magnetic resonance imaging findings included IS in 91 of 127 patients (72%), including territorial infarctions in 9 of 91 (10%), small infarctions in 55 (60%), and a combination of territorial and small in 27 (30%). Intracerebral hemorrhage was seen in 22 of 127 patients (17%), and it was lobar in 18 of 22 of them (82%) (Figure 2). Microhemorrhages were present in 43 of 127 patients (34%), including 26 of 43 (61%) with 1 to 5 MHs, 9 (21%) with 5 to 10 MHs, and 8 (19%) with more than 10 MHs. Fifteen of 127 patients (12%) had SAH on MRI: sulcal in 14 cases and diffuse in 1 case. Three of 361 patients (1%) were found to have MA. In all cases, the aneurysm was identified on conventional cerebral angiography (1 also had MRA that failed to visualize the aneurysm). Two patients with MA presented with ICH and 1 with SAH, and they all had coexistent infarctions. One patient had the MA clipped before undergoing VS. We observed no instances of bleeding from an MA after its diagnosis (see Supplemental Material, available online at http://www. mayoclinicproceedings.org). Valve surgery was performed in 95 of 127 patients who had MRI (75%) and 100 of 234 patients without MRI (43%). Among patients with preoperative MRI, 78 of 95 (82%) had abnormal radiological findings (IS in 91, 4
Mayo Clin Proc.
n
ICH in 22, MH in 43, SAH in 15, and MA in 3). Of the 32 patients who had MRI (25%) and then did not have VS, 23 (72%) had neurologic symptoms and 25 (78%) had MRI abnormalities; 29 (91%) had indications for VS. Overall, the presence of MRI abnormalities was not associated with VS (OR, 1.28; 95% CI, 0.48-3.45; P¼.25). The mean time to VS was 5489 days in patients with MRI abnormalities vs 65154 days without MRI abnormalities (P¼.79). Of the 89 patients without preoperative MRI, the head CT scan was abnormal in 29 (33%). Of the 195 patients who had VS, 22 (11%) had postoperative CT; 17 of 22 (77%) had scans that were prompted by new or worsening neurologic symptoms, and 4 (18%) had new abnormalities including ICH in 1, SAH in 1, and IS in 2 patients. Six-month mortality in patients with and without VS was 76 of 361 (21%) and 159 of 361 (44%) patients had good mRS at last follow-up. Six-month mortality in patients with VS was 22 of 195 (11%) and 54 of 166 (33%) patients who did not have VS (OR, 0.26; 95% CI, 0.15-0.46; P<.001). Good mRS at last follow-up was present in 102 of 195 patients who had VS (52%) vs 57 of 166 patients who did not undergo VS (34%) (OR, 2.10; 95% CI, 1.36-3.21; P<.001). XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
FIGURE 3. Postoperative intracerebral hemorrhage. This 51-year-old woman with Staphylococcus aureus endocarditis initially presented with dysarthria, right hemiparesis, and right hemisensory changes. Magnetic resonance imaging sequences obtained after head computed tomography are shown: (A) fluid attenuation inversion recovery and (B) susceptibility-weighted imaging showed multifocal ischemic stroke with intracerebral hemorrhage in the left parietal lesion and sulcal subarachnoid hemorrhage. She underwent valve surgery 3 days after infective endocarditis diagnosis in the setting of sepsis. She was given 1 dose of warfarin 2 days after surgery when postoperative (C) head computed tomography prompted by persistent somnolence and decreased respiratory drive showed new acute right cerebellar hemorrhage and acute hemorrhagic conversion of multifocal ischemic stroke. She had a modified Rankin Scale score of 5 at last follow-up.
Four patients who underwent VS had perioperative symptomatic hemorrhage; only 1 of them had preoperative MRI abnormalities and the other 3 had not had MRI. One patient with SAH and multifocal ICH seen on the initial CT scan with subsequent MRI had postoperative hemorrhage found after postoperative encephalopathy and labored breathing developed (Figure 3). This patient’s time from IE diagnosis to surgery was 3 days, and the mRS score at last follow-up was 5. No patients with IS (regardless of size), MH (regardless of number), or MA with MRI who underwent VS had postoperative ICH. Three 100 patients who did not have MRI (3%) had postoperative ICH; 1 patient had SAH seen on CT and underwent VS 1 day after IE diagnosis with an mRS score of 0 at last follow-up. Another patient had posterior fossa ICH detected on CT who underwent VS 16 days after IE diagnosis and was dead at last follow-up beyond 6 months after diagnosis. The third patient had no symptoms or imaging, had neurologic symptoms 18 days after VS in the setting of heparin for dialysis catheter placement, and then had an acute subdural hematoma and SAH. This patient had VS 2 days Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
after IE diagnosis and was dead at last follow-up within 6 months of IE diagnosis. The Table lists mortality and functional outcome rates according to MRI and VS; overall, patients who had VS despite MRI abnormalities had better clinical outcomes. Irrespective of VS, lobar or posterior fossa ICH on MRI was associated with higher odds of 6-month mortality (OR, 3.58; 95% CI, 1.22-10.50; P¼.02) and displayed a trend toward worse functional outcomes at last follow-up (OR, 0.56; 95% CI, 0.21-1.49; P¼.24). Territorial infarctions on MRI were inversely correlated with favorable mRS at last follow-up (OR, 0.29; 95% CI, 0.130.66; P¼.002) and displayed a trend toward higher 6-month mortality (OR, 1.71; 95% CI, 0.64-4.57; P¼.28). Neurologic symptoms in patients who underwent MRI were not predictive of 6-month mortality (OR, 1.67; 95% CI, 0.59-4.57; P¼.34) or good mRS at last follow-up (OR, 0.58; 95% CI, 0.281.20; P¼.14). In neurologically asymptomatic patients who had VS, MRI findings did not impact 6-month mortality (OR, 0.91; 95% CI, 0.27-3.11; P>.99) or good mRS at last follow-up (OR, 1.50; 95% CI, 0.703.21; P¼.29).
https://doi.org/10.1016/j.mayocp.2018.09.015
5
MAYO CLINIC PROCEEDINGS
TABLE. MRI, VS, and Patient Outcomesa,b Outcome Abnormal MRI with VS Abnormal MRI without VS Normal MRI with VS Normal MRI without VS
6-mo mortality 9/78 (12)
Good mRS at last follow-up
P<.001 OR, 0.17 (95% CI, 0.06-0.48)
41/78 (53)
P¼.29c
14/17 (82)
11/25 (44) 0/17 (0)
5/25 (20)
1/7 (14)
No MRI with VS
13/100 (13)
No MRI without VS
42/134 (31)
P¼.005 OR, 4.43 (95% CI, 1.51-13.00) P¼1.00 OR, 0.77 (95% CI, 0.07-9.08)
6/7 (86) P¼.001 OR, 0.33 (95% CI, 0.16-0.65)
47/100 (47)
P¼.05 OR, 1.67 (95% CI, 0.99-2.88)
46/134 (34)
MRI ¼ magnetic resonance imaging; mRS ¼ modified Rankin Scale; OR ¼ odds ratio; VS ¼ valve surgery. Data are presented as n/N (percentage) or as P value, OR (95% CI). c No OR could be calculated because there were no patients with normal MRI and VS who had 6-mo mortality. a
b
DISCUSSION Our study confirmed that most patients with IE have cerebral lesions on MRI. Just 1 patient who had preoperative CT revealing hemorrhage with subsequent MRI before VS had postoperative symptomatic ICH, suggesting a low risk of peri- or postoperative hemorrhage irrespective of preoperative MRI findings. In our institution MRI was ordered often, but it is unclear how MRI abnormalities influenced surgical decisions on VS. Therefore, it is reasonable to conclude that the use of MRI is equivocal, and in contrast to current guidelines, we do not recommend brain MRI for screening of patients with IE, as its findings should not delay surgery in the absence of hemorrhage or large territorial IS visible on the CT scan.4,5 CT of the brain remains the first line of testing for patients with neurologic symptoms. On the basis of our experience, we suggest the approach shown in Figure 4 for the use and interpretation of neuroimaging in patients with IE. CT scans should suffice to evaluate for territorial infarction or hemorrhage before VS in patients with neurologic symptoms. The added value of MRI is unclear, but MRA or cerebral angiogram can be considered in a select group of patients. Once an infectious aneurysm is found, occlusion can be attempted. However, postoperative ICH is not always avoided by occluding the visible aneurysm because those patients already have pan-vasculitis and they can 6
Mayo Clin Proc.
n
bleed from a different vessel.22 Only 1 patient with hemorrhage on MRI in our cohort developed postoperative ICH despite the high proportion of patients with cerebral lesions on MRI. Furthermore, no patient with MH developed postoperative ICH, which may be a less common complication of VS than previously considered. Our detailed analysis of brain MRI findings in patients with IE also confirms that cerebrovascular abnormalities are common, including a high proportion of clinically silent ischemic and hemorrhagic lesions. Ninety-four percent of patients with neurologic symptoms had cerebral lesions on MRI in the present series. Ischemic stroke was the most common finding in 72%, followed by MH in 34%, ICH in 17%, and SAH in 12%. Previous studies have reported rates of IS ranging from 37% to 83% and rates of MH between 57% and 94%.1,15,23-25 The higher rate of MH in other studies may be attributable to the fact that hemosiderin-sensitive MRI sequences were not routinely used in the earlier years of our study time frame. In addition, Goulenok et al14 reported a higher rate of MA (23%) and SAH (17%) in their prospective study of 30 patients. This may be due to selection bias, with all patients undergoing MRI within the first week of admission, including MRA with subsequent catheter angiography in cases of suspected MA. Treatment of IE must ultimately weigh the benefit of VS against postoperative risk XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
Neurologic symptoms
Obtain CT head
Equivocal CT with unresponsiveness or territorial syndrome on examination
Normal, small IS, deep ICH, or MH
Obtain MRI scan Early VS
Territorial IS, or lobar or posterior fossa ICH
Weigh benefit, timing, and indications of VS against risk of postoperative ICH
SAH
Obtain MRA or conventional angiography to assess for MA
No MA
MA
Consider endovascular treatment
FIGURE 4. Approach to the patient with IE and neurologic symptoms. Patients without neurologic symptoms do not need neuroimaging unless symptoms arise. In patients with neurologic symptoms, head CT scans may initially be obtained. Based on CT findings, MRI may be considered and/or vessel imaging if MA is suspected as a source of hemorrhage seen on CT. When a mycotic aneurysm is detected in patients with IE, direct surgery may be an alternative. CT ¼ computed tomography; ICH ¼ intracerebral hemorrhage; IE ¼ infective endocarditis; IS ¼ ischemic stroke; MA ¼ mycotic aneurysm; MH ¼ microhemorrhage; MRA ¼ magnetic resonance angiography; MRI ¼ magnetic resonance imaging; SAH ¼ subarachnoid hemorrhage; VS ¼ valve surgery.
of hemorrhage, particularly in patients at higher risk of poor outcomes without surgery (such as those with severe heart failure, refractory sepsis, or large vegetations26). Patients with severe IE and cerebral embolism have a substantial risk of recurrent embolism if there is no prompt response to antibiotic therapy. In these cases, VS may be the optimal strategy to prevent recurrent brain lesions. Rates of 6-month mortality were lower (12% vs 44%) and favorable functional outcomes at last follow-up were higher (53% vs 20%) in patients with abnormal MRI scans who had VS than in those who did not have VS in our study. In comparison, our series reported a higher rate of favorable outcome (mRS score2) at last follow-up (52% at a median Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
follow-up of 10 weeks postdischarge) than did the previous series by Sonneville et al16 (29% with a mRS score of 3 at w12 weeks postdischarge). This discrepancy may be attributable to the fact that the cohort of Sonneville et al was restricted to patients admitted to the intensive care unit, unlike ours that also included noneintensive care unit patients. Our study adds information on prognostic factors. We found that territorial IS and lobar or posterior fossa ICH were associated with a poor outcome. Our study has several limitations. First, this study was observational and retrospective and the decision to proceed with surgery and surgical timing were determined by the treating team; thus, selection bias may have influenced the very low risk of perioperative
https://doi.org/10.1016/j.mayocp.2018.09.015
7
MAYO CLINIC PROCEEDINGS
cerebrovascular complications. Second, hemosiderin-sensitive sequences were not always acquired during the early years of the study period; the inclusion of some MRI scans lacking these sequences may have resulted in an underestimation of the number of MH. Third, not all patients with ICH or SAH had vessel imaging to investigate the presence of MA; our study may underestimate the number of patients with MA. Fourth, our study was limited to patients who had VS with available postoperative follow-up; our study may underestimate the number of patients with postoperative neurologic complications after hospital discharge if they did not have further followup care at our institution. Lastly, patients who were in the nonsurgical cohort may have included patients who had indications for VS but were too ill to tolerate surgery, leading to a poor clinical outcome in comparison with those with VS.
CONCLUSION Most patients with IE who underwent brain MRI had evidence of cerebrovascular complications, either symptomatic or silent. Threequarters of them underwent VS, a surgical rate that was much higher than that in patients who had normal MRI or did not have MRI. Thus, the additional value of brain MRI in the management of patients with IE remains uncertain. Postoperative hemorrhage may be a less common complication of VS than previously thought, and evidence of cerebrovascular lesions on brain MRI may not be generally considered contraindications for surgery, particularly when small, even if multiple (eg, multiple small ISs and multiple MHs). Although the risk of postoperative hemorrhage must always be weighed against the benefit of VS, brain MRI should not be considered indispensable for surgical clearance. SUPPLEMENTAL ONLINE MATERIAL Supplemental material can be found online at: http://www.mayoclinicproceedings.org. Supplemental material attached to journal articles has not been edited, and the authors take responsibility for the accuracy of all data. 8
Mayo Clin Proc.
n
Abbreviations and Acronyms: CT = computed tomography; ICH = intracerebral hemorrhage; IE = infective endocarditis; IS = ischemic stroke; MA = mycotic aneurysm; MH = microhemorrhage; MRA = magnetic resonance angiography; MRI = magnetic resonance imaging; mRS = modified Rankin Scale; OR = odds ratio; SAH = subarachnoid hemorrhage; VS = valve surgery
Potential Competing Interests: Dr Scharf has received small honorarium for MOC lecture series from the American Academy of Neurology. Dr Wijdicks has received honorarium and travel expenses for Symposium on Organ Donation Leuven, Belgium, and royalties from Oxford University Press. Dr Rabinstein has received consultancy fees from SpaceWorks and royalties from Oxford, Elsevier, and UpToDate. The rest of the authors report no competing interests. Correspondence: Address to Tia Chakraborty, MD, Department of Neurology, Mayo Clinic, 200 First St SW, Rochester, MN 55905 ([email protected]).
REFERENCES 1. Hess A, Klein I, Iung B, et al. Brain MRI findings in neurologically asymptomatic patients with infective endocarditis. AJNR Am J Neuroradiol. 2013;34(8):1579-1584. 2. Cooper HA, Thompson EC, Laureno R, et al. Subclinical brain embolization in left-sided infective endocarditis: results from the evaluation by MRI of the brains of patients with left-sided intracardiac solid masses (EMBOLISM) pilot study. Circulation. 2009;120(7):585-591. 3. Duval X, Iung B, Klein I, et al; IMAGE (Resonance Magnetic Imaging at the Acute Phase of Endocarditis) Study Group. Effect of early cerebral magnetic resonance imaging on clinical decisions in infective endocarditis: a prospective study. Ann Intern Med. 2010;152(8):497-504. W175. 4. Baddour LM, Wilson WR, Bayer AS, et al; American Heart Association Committee on Rheumatic Fever, Endocarditis, and Kawasaki Disease of the Council on Cardiovascular Disease in the Young, Council on Clinical Cardiology, Council on Cardiovascular Surgery and Anesthesia, and Stroke Council. Infective endocarditis in adults: diagnosis, antimicrobial therapy, and management of complications: a scientific statement for healthcare professionals from the American Heart Association. Circulation. 2015;132(15):1435-1486 [published corrections appear in Circulation. 2015 and Circulation. 2016]. 5. Habib G, Lancelloti P, Antunes MJ, et al; ESC Scientific Document Group. 2015 ESC Guidelines for the management of infective endocarditis: the Task Force for the Management of Infective Endocarditis of the European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J. 2015;36(44): 3075-3128. 6. Gaca JG, Sheng S, Daneshmand MA, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg. 2011;141(1):98-106. e101-e102. 7. Chu VH, Park LP, Athan E, et al. International Collaboration on Endocarditis (ICE) Investigators. Association between surgical indications, operative risk, and clinical outcome in infective endocarditis: a prospective study from the International Collaboration on Endocarditis. Circulation. 2015;131(2):131-140. 8. Kang DH, Kim YJ, Kim SH, et al. Early surgery versus conventional treatment for infective endocarditis. N Engl J Med. 2012;366(26):2466-2473. XXX 2019;nn(n):1-9
n
https://doi.org/10.1016/j.mayocp.2018.09.015 www.mayoclinicproceedings.org
BRAIN FINDINGS IN INFECTIVE ENDOCARDITIS
9. Anantha Narayanan M, Mahfood Haddad T, Kalil AC, et al. Early versus late surgical intervention or medical management for infective endocarditis: a systematic review and meta-analysis. Heart. 2016;102(12):950-957. 10. Okita Y, Minakata K, Yasuno S, et al. Optimal timing of surgery for active infective endocarditis with cerebral complications: a Japanese multicentre study. Eur J Cardiothorac Surg. 2016; 50(2):374-382. 11. Thuny F, Avierinos JF, Tribouilloy C, et al. Impact of cerebrovascular complications on mortality and neurologic outcome during infective endocarditis: a prospective multicentre study. Eur Heart J. 2007;28(9):1155-1161. 12. Wilbring M, Irmscher L, Alexiou K, Matschke K, Tugtekin SM. The impact of preoperative neurological events in patients suffering from native infective valve endocarditis. Interact Cardiovasc Thorac Surg. 2014;18(6):740-747. 13. García-Cabrera E, Fernández-Hidalgo N, Almirante B, et al; Group for the Study of Cardiovascular Infections of the Andalusian Society of Infectious Diseases; Spanish Network for Research in Infectious Diseases. Neurological complications of infective endocarditis: risk factors, outcome, and impact of cardiac surgery: a multicenter observational study. Circulation. 2013; 127(23):2272-2284. 14. Goulenok T, Klein I, Mazighi M, et al; IMAGE Study Group. Infective endocarditis with symptomatic cerebral complications: contribution of cerebral magnetic resonance imaging. Cerebrovasc Dis. 2013;35(4):327-336. 15. Iung B, Tubiana S, Klein I, et al; ECHO-IMAGE Study Group. Determinants of cerebral lesions in endocarditis on systematic cerebral magnetic resonance imaging: a prospective study. Stroke. 2013;44(11):3056-3062. 16. Sonneville R, Mirabel M, Hajage D, et al; ENDOcardite en REAnimation Study Group. Neurologic complications and outcomes of infective endocarditis in critically ill patients: the
Mayo Clin Proc. n XXX 2019;nn(n):1-9 www.mayoclinicproceedings.org
n
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
ENDOcardite en REAnimation prospective multicenter study. Crit Care Med. 2011;39(6):1474-1481. Yanagawa B, Pettersson GB, Habib G, et al. Surgical management of infective endocarditis complicated by embolic stroke: practical recommendations for clinicians. Circulation. 2016; 134(17):1280-1292. Hodges KE, Hussain ST, Stewart WJ, Pettersson GB. Surgical management of infective endocarditis complicated by ischemic stroke. J Card Surg. 2017;32(1):9-13. Chakraborty T, Scharf E, Rabinstein AA, et al. Utility of brain magnetic resonance imaging in the surgical management of infective endocarditis. J Stroke Cerebrovasc Dis. 2017;26(11): 2527-2535. Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis. 2000;30(4):633-638. Sulter G, Steen C, De Keyser J. Use of the Barthel index and modified Rankin scale in acute stroke trials. Stroke. 1999; 30(8):1538-1541. van de Beek D, Rabinstein AA, Peters SG, Cloft H, Wijdicks EF. Staphylococcus endocarditis associated with infectious vasculitis and recurrent cerebral hemorrhages. Neurocrit Care. 2008;8(1): 48-52. Klein I, Iung B, Labreuche J, et al; IMAGE Study Group. Cerebral microbleeds are frequent in infective endocarditis: a casecontrol study. Stroke. 2009;40(11):3461-3465. Morofuji Y, Morikawa M, Yohei T, et al. Significance of the T2*weighted gradient echo brain imaging in patients with infective endocarditis. Clin Neurol Neurosurg. 2010;112(5):436-440. Malhotra A, Schindler J, Mac Grory B, et al. Cerebral microhemorrhages and meningeal siderosis in infective endocarditis. Cerebrovasc Dis. 2017;43(1-2):59-67. Behrouz R. Preoperative cerebrovascular evaluation in patients with infective endocarditis. Clin Cardiol. 2015;38(7):439-442.
https://doi.org/10.1016/j.mayocp.2018.09.015
9