Neurological Outcomes of Patients With Mycotic Aneurysms in Infective Endocarditis

Abstract

Background and Purpose:

Mycotic aneurysms (MA) are rare neurovascular complications of infective endocarditis (IE). The natural history and outcomes of MA under contemporary medical therapy have not been well characterized. The purpose of this study is to describe treatments and outcomes of patients with ruptured and unruptured MA in IE, specifically in relation to medical versus surgical/endovascular treatment.

Methods:

Retrospective chart review was performed at 3 US academic medical centers of adult patients with IE and MA. Information was collected regarding risk factors, imaging, treatments, and outcomes, including ischemic stroke, intracerebral hemorrhage, MA size changes, and inhospital mortality.

Results:

Thirty-five patients with IE had 63 MA. Nineteen patients had at least one ruptured MA; 13 patients underwent invasive treatment and 6 received antibiotics alone. Of 19 patients on antibiotics alone (6 with at least one ruptured MA and 13 with unruptured MA), 14 underwent repeat imaging and 5 had enlarging MA. Of 16 patients treated invasively, 2 had unruptured MA initially treated with antibiotics but ultimately underwent intervention. No MA ruptured after aneurysm discovery. Fifteen patients underwent cardiothoracic surgery (CTS), of which 11 had unsecured MA and 4 had secured MA. No patients suffered perioperiative neurological events attributable to their MA. Three patients treated with antibiotics alone and 3 patients treated invasively died from causes unrelated to their MAs.

Conclusions:

For patients with unruptured MA, treatment with antibiotics alone may have similar outcomes to invasive treatment. Further investigation is warranted to determine the risk of undergoing CTS with unsecured MA.

Introduction

Mycotic aneurysms (MA), also called infectious intracranial aneurysms or microbial aneurysms, are rare neurovascular complications of systemic or intracranial infections, such as infective endocarditis (IE).¹ Feared complications of MA include subarachnoid hemorrhage (SAH) or intracerebral hemorrhage (ICH), particularly when exposed to anticoagulant drugs during open heart surgery.² Clinical practice is guided by expert opinion-based algorithms that generally recommend treatment with antibiotics and case-by-case consideration for neurosurgical or endovascular procedures.^3-6 Some MA respond to prompt antimicrobial treatment alone.^6-8 However, the contemporary natural history and outcomes of MA treated with medical versus invasive treatment have not been well characterized. We sought to describe treatments and outcomes of patients with ruptured and unruptured MA in a multicenter study.

Methods

Study Design

We performed a multicenter, retrospective cohort study at New York-Presbyterian/Columbia University Irving Medical Center (NYP/CUIMC), Montefiore Medical Center/Albert Einstein College of Medicine (MMC/AECOM), and University of Florida Health (UFH). We included adult patients with IE and MA. The Institutional Reviews Boards of each institution approved this study and granted a waiver of informed consent.

Study Patients

Methods for NYP/CUIMC data collection have been published.⁹ Patients at NYP/CUIMC were identified using International Classification of Diseases, Ninth Edition codes for IE from 1996 to 2014 and at UFH from 2006 to 2014. Patients at MMC/AECOM were identified by automated language analysis of neuroradiological studies from 2012 to 2019 searching for “mycotic,” “mycotic aneurysm,” “cortical SAH,” and “peripheral SAH” in the indication and body of reports; patients records were then manually reviewed to confirm IE and MA. Additional patients with MA and IE from 2015 to 2019 were included from authors’ personal records. Each patient met modified Duke criteria for IE.¹⁰

Measurements

Authors performed manual data extraction from the electronic medical record and collected information regarding risk factors, imaging, treatments, and outcomes. Aneurysm size, location, and rupture status were tabulated. All aneurysms included in this study were presumed to be MA since no patients were known to have prior aneurysms or SAH. Aneurysms were detected by computed tomography (CT) angiography (CTA), magnetic resonance (MR) angiography (MRA), or digital subtraction angiography (DSA). Aneurysm location, shape, and size (maximal dimension) were extracted from radiological or neurosurgical/endovascular operative reports. An aneurysm was considered secured if it was clipped, coiled, embolized, or had parent vessel occlusion. Aneurysms were categorized as ruptured and unruptured; rupture was defined as CT or MR imaging evidence of hemorrhage in the area of the aneurysm. Appropriate antibiotics were defined as endocarditis-targeted antibiotics, such as broad-spectrum or organism-specific. Outcomes were clinically diagnosed ischemic stroke, ICH, SAH, change in MA size, and inhospital mortality. Missing data were omitted from analysis.

Statistical Analysis

We used standard descriptive statistics to summarize patient and MA characteristics. Wilcoxon median 2-sample test was used to compare ruptured and unruptured aneurysm sizes using SAS version 9.4.

Results

Patient Characteristics

From 1357 patients with IE reviewed at NYP/CUIMC and UFH, 18 (1.3%) had at least one MA. A total of 35 patients with IE and MA were identified from our 3 centers (Figure 1). The median age was 49 years (interquartile range [IQR], 42-64), and 14 (40%) were women (Table 1). Eight patients were immunosuppressed and 8 were known intravenous drug users. Twenty-nine had positive cultures; 1 had Candida endocarditis and the remainder had bacterial endocarditis. One culture-negative patient had Bartonella henselae. Mycotic aneurysms were identified due to the presence of neurological deficits or symptoms in 31 patients; 4 patients had asymptomatic MA discovered on screening imaging. Among 18 patients with ruptured MA and known medications at time of rupture, 4 were on antiplatelet medication (22%) and none on therapeutic anticoagulation. Among patients with ruptured MA, median Glasgow Coma Scale (GCS) on presentation to the hospital was 15 (IQR, 15-15), and at time of MA rupture was 14 (IQR, 13-15), with neurological examination ranging from comatose, to focal deficits such as inattention, aphasia, and hemiplegia, to normal.

Figure 1.

Flow chart describing patient identification process at 3 university hospitals. Charts were reviewed from 3 centers to identify 35 patients with infective endocarditis (IE) and at least 1 mycotic aneurysm. ICD-9, International Classification of Diseases, Ninth Edition; MA, mycotic aneurysm; MMC/AECOM, Montefiore Medical Center/Albert Einstein Medical College; NYP/CUIMC, New York-Presbyterian Hospital/Columbia University Irving Medical Center; UFH, University of Florida Health.

Table 1.

Patient Characteristics Grouped by MA Rupture Status and Treatment Strategy.

Characteristicsa	All patients with MA, N = 35	≥ 1 ruptured MA, n = 19	Only unruptured MA, n = 16	Patients treated invasively n = 16	Patients treated medically n = 19
Age (years)	49 (42-64)	51 (41-62)	49 (47-65)	56 (43-64)	47 (41-63)
Male	21 (60%)	14 (74%)	7 (44%)	11 (69%)	10 (53%)
Nonwhite	16 (47%)	8 (44%)	8 (40%)	7 (47%)	9 (47%)
Hypertension	18 (51%)	10 (53%)	8 (50%)	10 (63%)	8 (42%)
Atrial fibrillation	5 (14%)	3 (16%)	2 (13%)	3 (19%)	2 (11%)
Ever smoker	18 (51%)	11 (58%)	7 (44%)	8 (50%)	10 (53%)
Ever intravenous drug user	8 (23%)	5 (26%)	3 (19%)	1 (6%)	7 (37%)
Immunosuppressed	8 (23%)	5 (26%)	3 (19%)	5 (31%)	3 (16%)
Positive cultures	29 (83%)	16 (84%)	13 (81%)	13 (81%)	16 (84%)
MA found incidentally	4 (11%)	1 (6%)	3 (19%)	2 (13%)	2 (11%)
Timeline (number of days from)
Presentation to appropriate antibiotics	0 (0-3)	1 (0-3)	0 (0-1)	2 (1-4)	0 (0-1)
Presentation to MA diagnosis	4.5 (1-9)	2.5 (1-10)	6.5 (3-9)	3 (1-10)	5 (2-9)
MA diagnosis to repeat vessel imaging	6 (3.5-14)	4 (4-9)	11.5 (5-24)	4.5 (3-12)	7 (4-15)
MA diagnosis to neurosurgery/EVT	0 (0-3)	0 (0-1)	25 (3-131)	0 (0-3)	-
MA diagnosis to CTS	6 (4-18)	5 (3-20)	7.5 (4-14)	5 (4-6)	11.5 (4-16)
Intracranial arterial diagnostic method
CT angiography only	7 (20%)	3 (16%)	4 (25%)	3 (19%)	4 (21%)
MR angiography only	7 (20%)	2 (11%)	5 (31%)	1 (6%)	6 (32%)
Digital subtraction angiography only	5 (14%)	3 (16%)	2 (13%)	4 (25%)	1 (5%)
Two or more modalities	15 (43%)	11 (58%)	4 (25%)	8 (50%)	7 (37%)
Initial imaging finding
Infarction without hemorrhage	11 (31%)	0 (0%)	11 (69%)	2 (13%)	9 (47%)
Hemorrhage without infarction	12 (34%)	11 (58%)	1 (6%)	8 (50%)	4 (21%)
Infarction and hemorrhage	10 (29%)	8 (42%)	2 (13%)	5 (31%)	5 (26%)
No infarction or hemorrhage	2 (6%)	0 (0%)	2 (13%)	1 (6%)	1 (5%)
Median # of MA per patient	1 (1-2)	1.5 (1-2)	1 (1-3)	1 (1-2)	1.5 (1-3)
Median MA size on first scan (mm)	3 (2-5)	3.3 (2-6)	3.0 (3-4)	5.9 (4-7)	3.0 (2-4)
Number of patients with ≥ 1 ruptured MA	19 (54%)	19 (100%)	0 (0%)	13 (81%)	6 (32%)

Abbreviations: CT, computed tomography; CTS, cardiothoracic surgery; EVT, endovascular therapy; ICH, intracerebral hemorrhage; IQR, interquartile range; MA, mycotic aneurysms; MR, magnetic resonance; SAH, subarachnoid hemorrhage.

^a Data presented as median (IQR, 25th-75th percentile) or number (%) unless otherwise stated.

Aneurysm Characteristics

The number and size of MA were available in 33 patients. There were at least 64 MA (22 ruptured among 19 patients, 42 unruptured among 29 patients), with median of 1 MA per patient (IQR, 1-2; range, 1-6) and median maximum dimension 3.0 mm (IQR, 2-5; range, 0.5-15.0 mm; Table 1). Among 19 patients with at least 1 ruptured MA, 13 had subarachnoid hemorrhage (SAH), of which 6 had convexity SAH, 6 had convexity and deep SAH, none had only deep SAH, 1 had unknown SAH location, and 17 had ICH. Mycotic aneurysms were predominantly in the anterior circulation, especially involving the middle cerebral artery (MCA). Mycotic aneurysms were characterized with CTA only in 21%, MRA only in 21%, DSA only in 15%, and multiple modalities in 44%; in total, 51% underwent DSA. Of 21 ruptured MA with known morphology, 8 (38%) were saccular, 7 (33%) sacculofusiform, and 6 (29%) fusiform. Of 35 unruptured MA with known morphology, 15 (43%) were sacculofusiform, 8 (23%) saccular, and 12 (34%) fusiform. Of 43 MA with known sizes, the median ruptured MA was 4.5 mm (IQR, 2-6) versus 3.0 mm (IQR, 3-4) for unruptured MA (P = 0.20; Figure 2). Two patients had multiple unruptured MA without documentation of exact number or sizes.

Figure 2.

Distribution of mycotic aneurysm of known sizes at presentation, based on rupture status. Ruptured and unruptured mycotic aneurysms were of similar size.

Management and Outcomes

Of 19 patients with at least 1 ruptured MA, 6 patients (32%) received antibiotics alone. Of 16 patients with only unruptured MA, 13 (81%) received antibiotics alone. Of these 19 patients on antibiotics alone, 14 (74%) underwent repeat imaging; 5 patients had enlarging MA, although none ruptured or were intervened upon (Table 2, Figure 3).

Table 2.

Patient Outcomes Grouped by MA Rupture Status and Treatment Strategy.

Outcomea	All patients with MA, N = 35	≥ 1 ruptured MA, n = 19	Only unruptured MA, n = 16	Patients treated invasively, n = 16	Patients treated medically, n = 19
≥ 1 enlarging MA	5 (14%)	3 (16%)	2 (13%)	2 (13%)	3 (16%)
MA rerupture	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
Death during hospitalization	6 (17%)	4 (21%)	2 (13%)	3 (19%)	3 (16%)
Underwent CTS for IE	15 (43%)	5 (26%)	10 (63%)	6 (38%)	9 (47%)
Underwent CTS before MA secure	11 (73%)	2 (40%)	9 (90%)	2 (33%)	9 (100%)
Death after CTS	1 (7%)	0 (0%)	1 (10%)	0 (0%)	1 (11%)
Neurological outcome after CTS
Perioperative ischemic stroke	1 (7%)	0 (0%)	1 (10%)	0 (0%)	1 (11%)
Perioperative hemorrhagic stroke	0 (0%)	0 (0%)	0 (0%)	0 (0%)	0 (0%)
No neurological event	14 (93%)	5 (100%)	9 (90%)	6 (100%)	8 (89%)
Events unrelated to CTS
Ischemic stroke	4 (11%)	3 (16%)	1 (6%)	3 (19%)	1 (5%)
Hemorrhagic stroke	3 (9%)	2 (11%)	1 (6%)	3 (19%)	0 (0%)

Abbreviations: CTS, cardiothoracic surgery; IE, infectious endocarditis; MA, mycotic aneurysms.

^a Data presented as number (%).

Figure 3.

Mycotic aneurysm size changes over time among patients treated with antibiotics alone. Figure depicts evolution of mycotic aneurysm treated with antibiotics alone and had follow-up vessel imaging that provided size information. This illustrates 30 mycotic aneurysms of 15 unique patients. Each line represents a unique mycotic aneurysm. *One mycotic aneurysm was reimaged at 391 days. It was 9 mm in maximum diameter at diagnosis and was 8 mm at day 391; this mycotic aneurysm’s final size is not shown to allow visualization of patients reimaged within 50 days.

Of 16 patients invasively treated, 2 had unruptured MA initially treated with appropriate antibiotics that ultimately underwent intervention due to enlarging MA, 1 while receiving 3 out of 6 weeks of antibiotics, and the other after 6 weeks of antibiotics. Examples of radiographic images of ruptured mycotic aneurysms before and after invasive intervention are provided in Figure 4. No MA ruptured after aneurysm discovery. However, one patient had recurrent SAH attributed to infectious arteritis rather than MA rerupture as his DSA showed resolution of prior MA and there was fusiform dilation of the MCA in the region of the SAH, which was interpreted as an area of infectious arteritis rather than MA. Three patients had hemorrhagic stroke after MA discovery, which were deemed to be separate from hemorrhage attributable to MA rupture based on location (Table 2).

Figure 4.

Radiographic images of 2 patients with ruptured mycotic aneurysms before and after invasive intervention. A, Patient 1, axial CT head without contrast showing right frontotemporal intracerebral hemorrhage (red arrow) with edema and mass effect. B, Patient 1, axial MRA after decompressive hemicraniectomy showing 9 mm saccular aneurysm in right middle cerebral artery (yellow arrow). C, Patient 1, axial CTA with contrast showing hemicranieotomy and aneurysm clip (green arrow). D, Patient 2, axial CT head without contrast showing left hemispheric convexity subarachnoid hemorrhage (pink arrow) with mild edema. E, Patient 2, axial CTA with contrast showing 7 mm saccular aneurysm in left posterior communicating artery (yellow arrow). F, Patient 2, axial CTA with contrast showing coil artifact in area of aneurysm after successful endovascular coiling. CT indicates computed tomography; CTA, CT angiography, MRA, magnetic resonance angiography.

Fifteen (43%) out of 35 patients underwent cardiothoracic surgery (CTS), 11 patients with unsecured MA (2 of whom had ruptured MA) and 4 patients with secured MA (3 of whom had ruptured MA). The median number of days on appropriate antibiotics prior to CTS for unsecured patients was 8 days (IQR, 6-23) compared to 14 days (IQR, 11-16) for secured patients. The neurovascular intervention and number of days prior to CTS of the 4 patients who had secured MA were endovascular therapy (EVT) 1 day prior, EVT 2 days prior, EVT 6 days prior, and neurosurgery 59 days prior to CTS. Two patients had their MA secured after CTS (neurosurgery 21 days and EVT 231 days after CTS), both of whom had thin convexity SAH, minimal neurological deficits, and unsecured ruptured MA that were stable or reduced in size prior to CTS.

One patient with unsecured MA had an ischemic stroke after CTS, while no others had any perioperative neurological events. This patient with postoperative ischemic stroke awoke from surgery with a left ICA occlusion causing ischemic stroke unrelated to her unruptured MA and died after withdrawal of care. One patient with enlarging unruptured, unsecured MA underwent urgent CTS 15 days after MA was initially discovered, and she suffered no neurological complications during or after CTS. Of 19 patients treated with antibiotics only, 3 (16%) died inhospital from causes unrelated to their MA. Similarly, of 16 patients with secured MA, 3 (19%) died inhospital, of which 2 died after withdrawal of care due to lack of neurological recovery from the original MA rupture. Of 19 patients with ruptured MA, 16 required intubation. Of 19 patients with ruptured MA, at time of discharge or death, median GCS was 15 (IQR, 8-15).

Discussion

Given the rarity of MA and absence of prospective and randomized controlled trials of MA management in IE, we conducted this retrospective observational study to assess contemporary epidemiology, management, and outcomes of MA in IE, particularly in relation to rupture status and outcomes after CTS. To our knowledge, this study of 35 patients is the largest multicenter retrospective observational study to date of patients with MA and IE.

In this study of 35 patients with MA and IE, 19 patients were treated with medical care alone, while 16 patients were treated with medical and invasive therapy using neurosurgical or endovascular interventions. No patient suffered aneurysmal rupture after MA discovery and antibiotic initiation, including 11 patients who underwent CTS with unsecured MA, 2 of which had initially ruptured MA. Outcomes in patients with medically versus invasively treated MA appeared similar.

Few studies highlight outcomes of patients with IE and MA after CTS even though treatment of MA in IE often depends on the urgency of cardiac valve repair for management of IE.² It has been suggested that prior to CTS, MA should be treated with antibiotics, monitored for resolution, and secured if persistent and time allows, either with neurosurgical clipping which requires a craniotomy and often 2 to 3 week delay prior to CTS given anticoagulation exposure, or EVT which allows even same-day CTS.^2-6,11 A scientific statement from the American Heart Association reports it is reasonable in patients who require CTS to secure MA with EVT even if no rupture occurred.³ However there has been a dearth of data on contemporary treatments and outcomes for patients with IE and MA, particularly those who are treated medically and undergo CTS. A systematic review of 384 patients with IE and MA showed 21% underwent valve surgery, of which 15% were unsecured, although their rupture status and outcomes were not reported.¹² Patient outcomes for those treated with antibiotics in our study were more favorable than reported in a recent single-center study, which found unfavorable angiographic outcomes for 44% of aneurysms.¹³ However, only one patient in that single-center study had rupture of a MA, which is consistent with our finding: few patients had poor clinical outcomes attributable to antibiotic treatment of MA.

In a national registry study of 339 patients with IE and MA, 18% underwent CTS, of which 92% had unsecured medically treated MA despite 93% having had ruptured MA.¹⁴ Of patients who underwent CTS, 39% had poor outcomes, and mortality was 14%, which was nominally better than the 43% with poor outcome and 27% with mortality among patients who did not undergo CTS. The national registry patients had high rates of initial MA rupture with low rates of neurosurgical/endovascular intervention and showed a trend toward improved outcome and mortality in those who underwent CTS; however, it is not clear if patients suffered rerupture or other neurological or medical sequel to explain the overall high rates of poor outcome and mortality because this was not the focus of their study. In contrast, our study featured granular data on patients with MA who were medically treated, including 11 of 15 who underwent CTS without MA-related complications.

In a single-center retrospective study featuring 14 patients with IE and MA who underwent CTS, 4 of 10 patients with unsecured MA suffered postoperative cerebral hemorrhage (4-57 days after CTS), in contrast to 4 patients with secured MA who did not have postoperative cerebral hemorrhage.¹⁵ Mycotic aneurysms rupture status was not entirely clear since they defined cerebral hemorrhage as SAH, ICH, and hemorrhagic infarction, and did not report whether the hemorrhage was in proximity of MA. Among their 9 patients with MA and preoperative cerebral hemorrhage, 6 underwent CTS prior to securing MA, of which 2 suffered postoperative cerebral hemorrhage; among 5 patients without initial hemorrhage, 4 had CTS without secured MA, of which 2 had postoperative cerebral hemorrhage. These findings differ from our study’s 15 patients with IE and MA who underwent CTS, 11 of whom were unsecured, and none suffered postoperative MA rupture regardless of whether MA was secured or not. Their study did not describe whether patients were on appropriate antimicrobials prior to cerebral hemorrhage nor the evolution of MA size; we speculate these factors could relate to risk for MA rupture.

Our study demonstrated that MA of all sizes and rupture status did not rupture/rerupture or cause further neurological deterioration after initiation of appropriate antibiotics. It is unclear what the risks are of enlarging MA as 24% of our patients treated with initial medical management had enlarging MA; of these 5 patients with enlarging MA, 2 underwent neurosurgery or EVT, but the other 3 did not rupture, undergo further intervention, or become symptomatic. Our results contrast with historical retrospective studies where, for example, 29% of patients with MA on antibiotics ruptured/reruptured during or after therapy.¹⁶ We speculate improved time to diagnosis of IE, prompt initiation of antimicrobials, and improved medical management of aneurysms, such as blood pressure control contribute to the decrease in poor neurological outcomes compared to historical studies.

There are several limitations to this study. As a retrospective study, there may be selection bias due to our dependence on accurate coding and documentation in the health record. Other limitations include the lack of standardized follow-up vessel imaging to track the evolution of MA size, as well as detailed outcome data such as postdischarge functional status and postdischarge morbidity. Patients were not randomized to MA treatment or CTS, so indication bias with respect to both MA treatment and CTS are possible. Additionally 2 patients with reported multiple unruptured MA did not have vessel imaging or documentation of exact number of MA available to review, which limited data collection and analysis and underestimates the true number of MA in this cohort. However, this study provides a large sample from 3 centers with data on rupture status, interventions, and inhospital outcomes.

These limitations notwithstanding our data raise the possibility that some MA can be treated with antibiotics alone, even if undergoing CTS. Further investigation is warranted to determine the risk of undergoing CTS with MA treated with antibiotics alone.