Endocarditis in Adults With Bacterial MeningitisClinical Perspective
Background—Endocarditis may precede or complicate bacterial meningitis, but the incidence and impact of endocarditis in bacterial meningitis are unknown.
Methods and Results—We assessed the incidence and clinical characteristics of patients with meningitis and endocarditis from a nationwide cohort study of adults with community-acquired bacterial meningitis in the Netherlands from 2006 to 2012. Endocarditis was identified in 24 of 1025 episodes (2%) of bacterial meningitis. Cultures yielded Streptococcus pneumoniae in 13 patients, Staphylococcus aureus in 8 patients, and Streptococcus agalactiae, Streptococcus pyogenes, and Streptococcus salivarius in 1 patient each. Clues leading to the diagnosis of endocarditis were cardiac murmurs, persistent or recurrent fever, a history of heart valve disease, and S aureus as the causative pathogen of bacterial meningitis. Treatment consisted of prolonged antibiotic therapy in all patients and surgical valve replacement in 10 patients (42%). Two patients were treated with oral anticoagulants, and both developed life-threatening intracerebral hemorrhage. Systemic (70%) and neurological (54%) complications occurred frequently, leading to a high proportion of patients with unfavorable outcome (63%). Seven of 24 patients (29%) with meningitis and endocarditis died.
Conclusions—Endocarditis is an uncommon coexisting condition in bacterial meningitis but is associated with a high rate of unfavorable outcome.
Bacterial meningitis is a life-threatening disease that is associated with considerable mortality and morbidity.1–3 To prevent death and long-term disabling sequelae caused by bacterial meningitis conjugate, vaccines were developed.4 Vaccination resulted in a sharp decrease in meningococcal disease occurrence and a moderate decrease in pneumococcal meningitis.4 Currently, Streptococcus pneumoniae is responsible for 70% of the cases in Europe and the United States.5 Bacterial meningitis is often related to other foci of infection outside the central nervous system such as pneumonia, sinusitis, or otitis.1,2,5,6 An uncommon focus of bacterial meningitis is infective endocarditis.7 Endocarditis may precede or complicate bacterial meningitis, but the impact of coexisting endocarditis for patients with meningitis is unknown. We investigated the incidence, clinical features, treatment, and outcome of patients with both bacterial meningitis and endocarditis identified in a nationwide cohort study of adults with community-acquired bacterial meningitis.
Clinical Perspective on p 2062
In a prospective nationwide observational cohort study in the Netherlands, we included episodes of community-acquired bacterial meningitis confirmed by culture of cerebrospinal fluid (CSF) in adults. Methods have been described in detail previously.6 In summary, all patients were ≥16 years of age and were listed in the database of the Netherlands Reference Laboratory for Bacterial Meningitis from January 2006 to March 2012. This laboratory receives CSF isolates from ≈90% of all patients with bacterial meningitis in the Netherlands. Patients with negative CSF cultures, hospital-associated meningitis, or a neurosurgical device and patients who underwent a neurosurgical operation within 1 month before bacterial meningitis onset were excluded. Patients using immunosuppressive drugs and those with asplenia, diabetes mellitus, alcoholism, or infection with HIV were considered immunocompromised. Informed consent was obtained from all participating patients or their legally authorized representatives.
Clinical data were prospectively collected by means of an online case record form. The presence of endocarditis was scored as a standard question in the case record form. The diagnosis of endocarditis was confirmed by reanalysis of the results of cardiologic analyses and echocardiography, which were collected retrospectively. Endocarditis was defined as heart valve vegetations identified by echocardiography or autopsy or the fulfillment of the Duke criteria for infective endocarditis.8,9 At discharge, all patients underwent a neurological examination performed by a neurologist, and outcome was graded according to the Glasgow Outcome Scale. This is a well-validated measurement scale with scores varying from 1 to 5.10 A favorable outcome was defined as a score of 5; an unfavorable outcome, as a score of 1 to 4.
The Mann–Whitney U test was used to evaluate differences between bacterial meningitis patients with and without endocarditis with respect to continuous variables, and the χ2 test and Fisher exact test were used to compare categorical variables. Statistical analyses were performed with IBM SPSS Statistics, version 19, and values of P<0.05 were considered significant. The study was approved by the ethics committee of the Academic Medical Center, Amsterdam, the Netherlands.
From January 2006 to March 2012, 1025 episodes of community-acquired bacterial meningitis were included in the cohort, and endocarditis was identified in 24 of these patients (2.3%). Nine patients (38%) were female, and the median age was 63 years (quartiles 1–3, 50–71 years; Table 1). Causative organisms were S pneumoniae in 13 patients, Staphylococcus aureus in 8 patients, and Streptococcus agalactiae, Streptococcus pyogenes, and Streptococcus salivarius in 1 patient each. Patients with endocarditis were more likely to have S aureus as the causative pathogen compared with patients without endocarditis (8 of 24 [33%] versus 6 of 1001 [1%]; P<0.001; Table 2).
Endocarditis was identified on admission in 3 patients (12%; all had pneumococcal meningitis), during hospitalization in 19 patients (79%; median time between admission and detection, 8 days [minimum–maximum values, 2–57 days]), or after discharge (2 patients with pneumococcal meningitis, 8%). Both patients with endocarditis diagnosed after discharge had malaise, night sweats, and recurrent fever. Median time between admission and detection was shorter among patients with S aureus meningitis compared with those with S pneumoniae meningitis (21 patients diagnosed after admission: 3 days [quartiles 1–3, 3–8 days] versus 16 days [quartiles 1–3, 9–39 days]; P=0.015).
A cardiology consult was requested in 22 patients. Reasons for consulting a cardiologist were cardiac murmur (10 patients), persistent or recurrent fever (6 patients), preexisting heart valve disease (3 patients), secondary clinical deterioration (3 patients), S aureus as the causative pathogen (3 patients), atrial flutter (2 patients), splinter hemorrhages (1 patient), or a combination of factors (2 factors in 7 patients, 3 factors in 1 patient). In 3 patients, no clear reason was reported for consulting a cardiologist. Two patients were not seen by a cardiologist: 1 patient died a few hours after admission, and the other patient had no clinical suspicion of endocarditis (the diagnosis was made during autopsy). Endocarditis was confirmed by echocardiography in 21 patients (transesophageal in 16 patients and transthoracic in 5 patients) and by autopsy in 1 patient. Eight patients had aortic valve endocarditis; 9 patients had mitral valve endocarditis; and 1 patient had tricuspid valve endocarditis. Four patients had involvement of 2 valves (aortic and mitral valve endocarditis in 2 patients, aortic and tricuspid valve endocarditis in 1 patient, mitral and tricuspid valve endocarditis in 1 patient). Two patients fulfilled the Duke criteria for endocarditis without a positive echocardiogram. In 1 of these patients, no echocardiography was performed because the patient died a few hours after admission. In the other patient, 1 transesophageal and 2 transthoracic echocardiographies were performed without clear findings of endocarditis. Blood cultures were positive on admission in all patients with endocarditis. Typical endocarditic skin lesions were reported in 6 patients (25%) and occurred more often in patients with S aureus meningitis compared with patients with S pneumoniae meningitis (4 of 8 [50%] versus 1 of 13 [8%]; P=0.05).
Patients often had predisposing conditions for infective endocarditis. Immunocompromised state was identified in 8 patients, alcohol abuse in 4, and heart valve disease in 3; 1 patient had an intracardiac device.11 Five patients (21%) presented with the triad of meningitis, endocarditis, and pneumonia caused by S pneumonia, known as the Austrian syndrome. The majority of patients had a subacute presentation (18 of 24 [75%], defined as signs and symptoms >24 hours). Cranial computed tomography was performed on admission in 23 patients and showed abnormalities in 7 patients: cerebral infarction in 7 patients, brain edema in 2 patients, and mastoid opacification and cerebral aneurysm in 1 patient each.
The proportion of patients with at least 1 individual CSF finding predictive of bacterial meningitis (glucose level <34 mg/dL [1.9 mmol/L], ratio of CSF glucose to blood glucose <0.23, protein level >220 mg/dL, or leukocyte count >2000 per 1 mm3)12 was lower in patients with endocarditis compared with those without endocarditis (14 of 24 [58%] versus 877 of 1001 [88%]; P<0.001). Endocarditis patients with S aureus meningitis were less likely to have at least 1 individual CSF finding predictive of bacterial meningitis as described above compared with endocarditis patients with pneumococcal meningitis (1 of 8 [13%] versus 12 of 13 [92%]; P=0.001). CSF Gram staining was performed in 21 patients and showed bacteria in 14 patients (67%).
All patients received microbiologically adequate initial antimicrobial therapy. The median duration of antimicrobial treatment in surviving patients with endocarditis was 49 days (minimum–maximum values, 28–166 days; Table 3). Adjunctive dexamethasone therapy was administered before or with the first dose of antibiotics in 17 of 24 patients (71%), and all patients received dexamethasone 10 mg 4 times daily for 4 days. Two patients received adjunctive dexamethasone after the first dose of antibiotics, and the remaining 5 patients were not treated with adjunctive corticosteroid therapy.
The majority of patients developed complications during the clinical course (Table 3). The proportion of patients with respiratory failure (13 of 23 [57%] versus 247 of 966 [25%]; P=0.001), circulatory shock (9 of 23 [39%] versus 100 of 956 [11%]; P<0.001), and arthritis (4 of 22 [18%] versus 27 of 948 [3%]; P=0.004) was higher among patients with endocarditis compared with those without endocarditis. Two patients with endocarditis developed spondylodiscitis and a psoas abscess (S aureus and S pneumoniae in 1 patient each). Patients with endocarditis were more likely to develop cerebral infarctions (9 of 24 [38%] versus 222 of 1000 [22%]; P=0.08) and intracerebral hemorrhages (3 of 24 [13%] versus 18 of 1000 [2%]; P=0.01) compared with patients without endocarditis.
Two patients were treated with oral anticoagulants because of a history of heart valve replacement (both aortic and mitral valve replacement with mechanic valves in both patients), and both developed intracerebral hemorrhages. Cranial imaging was performed during admission in 20 patients and was abnormal in 9 patients. New abnormalities, not present on admission, were found 5 patients and consisted of cerebral infarction in 4 patients, intracerebral hemorrhage in 3 patients, and a brain abscess in 1 patient.
Seven of 24 patients (29%) with endocarditis died (median time after admission, 25 days; minimum–maximum values, 1–72 days). Unfavorable outcome occurred in 15 of 24 patients with endocarditis (63%) compared with 386 of 1001 meningitis patients without endocarditis (39%; P=0.019). Neurological sequelae were present on discharge in 8 of 17 survivors (47%), consisting of cognitive impairment in 3 (18%), hemiparesis in 3 (18%), aphasia in 2 (12%), and hearing loss in 2 patients (12%).
Cardiac surgery was performed in 11 of 24 patients (6 patients with S aureus and 5 patients with S pneumoniae meningitis) and consisted of a valve replacement in 10 patients (6 mechanical valves and 4 biotissue valves [3 animal tissues and 1 allograft]) and removal of an intracardiac device (an implantable cardioverter-defibrillator) in 1 patient. Echocardiographic and surgical data in 10 patients who underwent valve surgery are presented in Table 4. Mitral valve surgery was performed in 5 patients: 4 patients received a valve replacement (3 biotissue valves and 1 mechanical valve) and 1 patient received a valve repair. Four of these 5 patients with mitral valve surgery had large vegetations on the mitral valve with echocardiography; 1 patient also had an abscess on this valve; and the mitral valve was perforated in 2 patients. Aortic valve replacement was done in 6 patients (1 biotissue valve and 5 mechanical valves): 5 patients had large vegetations, and the aortic valve was completely destroyed in 1 patient.
The median time to cardiac surgery after diagnosis was 9 days (minimum–maximum values, 1–45 days). In 6 of 10 patients, surgery was performed after completion of the course of antibiotics for the meningitis. There was no difference in outcome between patients who were completely treated or partially treated for meningitis before cardiac surgery was performed.
Nine of 10 patients who received valve replacement (90%) survived compared with 8 of 14 patients without valve replacement (57%; P=0.17). One patient had a recurrence of prosthetic valve endocarditis 30 days after surgery, prompting surgical replacement of the prosthetic valve. Another patient underwent replacement of the prosthetic valve 5 months after surgery because the prosthetic device dehisced.
Endocarditis is an uncommon coexisting condition in bacterial meningitis identified in 2% of patients, but it is associated with high rates of unfavorable outcome (63%). The most common causative pathogens were S pneumoniae and S aureus, and clues leading to the diagnosis of endocarditis were cardiac murmurs, persistent or recurrent fever, a history of heart valve disease, and S aureus as the causative pathogen of bacterial meningitis. Therefore, cardiologic consultation should be a priority in patients with community-acquired meningitis presenting with clues for endocarditis. Because about half of patients with pneumococcal meningitis have either persistent or recurrent fever,13 many future patients with pneumococcal meningitis will need ancillary investigations to rule out or to establish endocarditis. The general recommendation for antibiotic treatment duration in patients with endocarditis and meningitis is 4 to 6 weeks, which is substantially longer than the standard 10 to 14 days for meningitis patients without endocarditis.14 Even longer courses of antibiotics are advised if the patient undergoes cardiac surgery.
Several studies suggest that combined antibiotic and surgical therapy for infective endocarditis reduces the risk of death resulting from any cause, especially among patients who have congestive heart failure, perivalvular invasive disease, or uncontrolled infection despite maximal antimicrobial therapy.15 In our series, although nonsignificant and confounded by indication, patients who underwent valve replacement had better survival compared with those without surgery (90% versus 57%; P=0.17). The timing and indications for surgical intervention in infective endocarditis remain controversial.16 A randomized, controlled trial of 76 patients with infective endocarditis showed that early valve surgery performed within 48 hours after diagnosis reduces the risk of death resulting from any cause or embolic events by reducing the risk of systemic embolism.16 In our 10 patients with valve replacement, surgery was performed a median of 9 days after diagnosis (minimum–maximum values, 1–45 days); only 3 patients underwent valve replacement within 48 hours.
Time between admission and detection of endocarditis was substantially shorter for patients with S aureus meningitis compared with those with S pneumoniae meningitis (3 versus 16 days). Previous studies have shown that patients with S aureus meningitis almost uniformly present with a primary infection focus, most commonly pneumonia or endocarditis.17 This suggests that, in case of S aureus infection, endocarditis precedes bacterial meningitis and that, in bacterial meningitis caused by S pneumonia, endocarditis is a complication. In case of S aureus infection, meningitis is caused by septic emboli originating from cardiac valve vegetations. Our finding that fewer patients with S aureus infection had individual CSF findings predictive of bacterial meningitis compared with patients with pneumococcal infection is in line with this hypothesis.
Ischemic stroke is a common complication in all patients with bacterial meningitis, occurring in 22% of bacterial meningitis patients without endocarditis in our study. In a previous study, we showed that cerebral infarction was present in 25% of 696 patients with bacterial meningitis.18 Ischemic stroke is also a major complication in endocarditis patients; in a previous cohort study of 1437 patients, 15.2% developed cerebral infarctions.19 In patients with both bacterial meningitis and endocarditis, there seems to be an additional risk at developing ischemic stroke; 9 of 24 patients (38%) with both bacterial meningitis and endocarditis developed cerebral infarctions in our study.
Pneumococcal meningitis patients more often present with typical clinical and CSF characteristics of bacterial meningitis, and the possibility of endocarditis is perhaps considered only after the development of complications indicative of endocarditis. Early detection of endocarditis in pneumococcal meningitis patients may lower the rate of complications and unfavorable outcome. Five of 13 patients with pneumococcal meningitis had coexisting endocarditis and pneumonia, also known as the Austrian syndrome.7,20 All patients with the Austrian syndrome had an unfavorable outcome, reflecting the severity of this condition. Whether the primary focus of infection is the meningitis or endocarditis remains difficult to distinguish because initial complaints of endocarditis can be nonspecific.
Patients with meningitis and endocarditis should not be treated with anticoagulant therapy. Intracranial hemorrhage is a rare but devastating complication in patients with bacterial meningitis, with high rates of mortality and unfavorable outcome (65% and 95%, respectively).21 A previous study showed a 5-fold increased risk of developing intracranial hemorrhage in patients with bacterial meningitis using anticoagulant therapy.21 Patients with S aureus meningitis and endocarditis are at even higher risk of intracerebral hemorrhage compared with other bacterial meningitis patients using anticoagulant therapy.15,21 In these patients, discontinuation of anticoagulant therapy should be considered until the patient has recovered from the acute phase of the bacterial meningitis episode.
This study has several limitations. First, several clinical characteristics of endocarditis patients were not scored in the case record form (eg, cardiac murmur, skin lesion, dental focus), which makes it difficult to determine their relevance in the diagnosis of endocarditis in meningitis patients. Furthermore, in patients with severe bacterial meningitis who died in the first days of admission, endocarditis may have been missed. This could have led to an underestimation of the incidence of endocarditis. Therefore, the provided incidence figures should be regarded as the minimal values. A further limitation of our study was that only patients with positive CSF cultures were included. Negative CSF cultures are estimated to occur in 11% to 30% of patients with bacterial meningitis.22,23 However, no significant differences in clinical presentation have been reported between patients with culture-positive bacterial meningitis and culture-negative bacterial meningitis. The design of our study, a prospective cohort study, precludes firm conclusions about the mechanisms explaining the association between meningitis and endocarditis; we did not have standard baseline (on admission) echocardiographic data of all our patients.
Endocarditis should be considered in patients with S aureus meningitis, a history of heart valve disease, or cardiac murmurs and in patients with clinical deterioration or persistent/recurrent fever during admission. Endocarditis in meningitis patients is associated with a high rate of neurological and systemic complications and requires prolonged antibiotic treatment and cardiac surgery in a selection of patients. Anticoagulant therapy is contraindicated in the acute phase of meningitis because the risk of intracranial hemorrhages is high. Despite optimal antibiotic treatment, 63% of patients die or have long-term neurological sequelae.
We are indebted to all the Dutch physicians who participated in the study.
Sources of Funding
This research has been supported by grants from the Netherlands Organization for Health Research and Development (Veni [916.76.023] and Vidi [016.116.358] to Dr van de Beek; Veni [916.13.078] to Dr Brouwer), the Academic Medical Center (Fellowship 2008 to Dr van de Beek), and the European Research Council (Starting Grant to Dr van de Beek).
- Received January 24, 2013.
- Accepted April 10, 2013.
- © 2013 American Heart Association, Inc.
- Brouwer MC,
- Tunkel AR,
- van de Beek D
- Li JS,
- Sexton DJ,
- Mick N,
- Nettles R,
- Fowler VG Jr.,
- Ryan T,
- Bashore T,
- Corey GR
- Kanakadandi V,
- Annapureddy N,
- Agarwal SK,
- Sabharwal MS,
- Ammakkanavar N,
- Simoes P,
- Sanjani HP,
- Nadkarni GN
- Dickerman SA,
- Abrutyn E,
- Barsic B,
- Bouza E,
- Cecchi E,
- Moreno A,
- Doco-Lecompte T,
- Eisen DP,
- Fortes CQ,
- Fowler VG Jr.,
- Lerakis S,
- Miro JM,
- Pappas P,
- Peterson GE,
- Rubinstein E,
- Sexton DJ,
- Suter F,
- Tornos P,
- Verhagen DW,
- Cabell CH
- Aronin SI,
- Mukherjee SK,
- West JC,
- Cooney EL
Bacterial meningitis is a serious and life-threatening disease. The predominant causative pathogen is Streptococcus pneumoniae, causing two thirds of cases. Mortality rates associated with bacterial meningitis are high, and neurological and systemic complications occur in a large proportion of patients. One of the potential complications is endocarditis. Endocarditis may precede or complicate bacterial meningitis, but the incidence and impact of endocarditis in bacterial meningitis were unknown. In a large, prospective, nationwide cohort study of adults with community-acquired bacterial meningitis, we analyzed the incidence and clinical characteristics of patients with meningitis and endocarditis. We found that although endocarditis is an uncommon coexisting condition in bacterial meningitis, it is associated with a high rate of unfavorable outcome. The most common causative pathogens were S pneumoniae and Staphylococcus aureus. Clues leading to the diagnosis of endocarditis were cardiac murmurs, persistent or recurrent fever, preexisting heart valve disease, secondary clinical deterioration, S aureus as the causative pathogen, atrial flutter, and splinter hemorrhages. This study indicates that cardiologic consultation should be a priority in patients with community-acquired meningitis presenting with clues for endocarditis. Because about half of patients with pneumococcal meningitis have either persistent or recurrent fever, many future patients with pneumococcal meningitis will need ancillary investigations to rule out or to establish endocarditis. In most patients, endocarditis was diagnosed during hospitalization or even after discharge, so early detection of endocarditis in patients with bacterial meningitis may lower the rate of complications and unfavorable outcome.