Journal of Clinical Neuroscience 22 (2015) 800–806
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Review
Intracranial bacterial infections of oral origin Alan A. Moazzam a,⇑, Sowmya M. Rajagopal b, Parish P. Sedghizadeh b, Gabriel Zada c, Mina Habibian b a
Department of General Internal Medicine, Keck Hospital of University of Southern California, 1500 San Pablo Street, Los Angeles, CA 90033, USA Ostrow School of Dentistry of University of Southern California, Los Angeles, CA, USA c Department of Neurosurgery, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA b
a r t i c l e
i n f o
Article history: Received 5 October 2014 Accepted 8 November 2014
Keywords: Brain abscess CNS abscess Dental infections Intracranial abscess Oral infections Periodontitis
a b s t r a c t Brain abscesses are rare but potentially deadly complications of odontogenic infections. This phenomenon has been described mainly in the form of case reports, as large-scale studies are difficult to perform. We compiled a total of 60 previously published cases of such a complication to investigate the predisposing factors, microbiology, and clinical outcomes of intracranial abscesses of odontogenic origin. A systematic review of the literature using the PubMed database was performed. Men accounted for 82.1% of cases, and the mean age was 42.1 years. Caries with periapical involvement and periodontitis were the two most common intra-oral sources, and wisdom tooth extraction was the most common preceding dental procedure. In 56.4% of cases, there were obvious signs of dental disease prior to development of intracranial infection. Commonly implicated microorganisms included Streptococcus viridans (especially the anginosus group), Actinomyces, Peptostreptococcus, Prevotella, Fusobacterium, Aggregatibacter actinomycetemcomitans and Eikenella corrodens. There was an 8.3% mortality rate. Intracranial abscesses can form anywhere within the brain, and appear unrelated to the side of dental involvement. This suggests that hematogenous spread is the most likely route of dissemination. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Brain abscesses are life-threatening infections that require immediate neurosurgical attention. The incidence is between one and eight per 100,000 patients per year in the USA [8]. Brain abscesses most frequently occur due to bacterial dissemination from a primary lesion at a distant site. They can also occur from direct contiguous invasion from an adjacent site of infection. The oral cavity harbors a diverse and abundant microflora. As part of the Human Microbiome Project, culture-independent molecular approaches have identified more than 1200 different types of microbes in the human mouth [7]; 350 different bacterial strains have been isolated in marginal periodontitis and 150 in endodontic infections [5]. Bacteria can disseminate from these locations due to a variety of dental conditions such as gingivitis and periodontitis, and also from procedures such as tooth extractions, endodontic treatments, and oral surgery. Even simple tooth brushing has been shown to induce transient bacteremia in 38.5% of cases [8]. Currently, the Council of Dental Therapeutics of the American Dental Association has recommended against routine antibiotic prophylaxis for dental procedures specifically to ⇑ Corresponding author. Tel.: +1 858 342 0874. E-mail address: [email protected] (A.A. Moazzam). http://dx.doi.org/10.1016/j.jocn.2014.11.015 0967-5868/Ó 2014 Elsevier Ltd. All rights reserved.
prevent brain abscesses given the relative low incidence of this complication and the potential for development of resistant strains [8]. Due to the rarity of this occurrence, to our knowledge no studies have specifically investigated this phenomenon in a comprehensive manner. The literature primarily consists of individual case reports. We systematically reviewed the literature to identify and analyze all cases representing central nervous system (CNS) infections of oral origin. Our goal was to provide a synthesis of available knowledge on the natural history, pathogenesis, microbiology, interventions and outcomes of this potentially fatal pathology.
2. Materials and methods Our review followed recommendations as outlined in the AMSTAR guidelines for systematic reviews [17]. We used an a priori research design that would allow us to better understand this clinical problem and provide insight into PICO-related questions (Patient problem or population, Intervention, Comparison and Outcome) in the context of brain abscesses of oral origin. A protocol specifying all aspects of the review method was developed before commencing the review. This protocol included ascertainment criteria for considering studies for this review, search methods for identification of studies, data collection and analysis.
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
The protocol was planned to minimize the effect of author bias on the review and the potential to alter or misinterpret data presented on study findings. We critically appraised the selected works and synthesized the data according to a predetermined and explicit methodology. This included quality assessment of reviewed papers, determination of levels of evidence, and evaluation of outcome measures but no statistical analyses given the nature of the available data and literature which we describe shortly. The PubMed electronic database was searched for all Englishlanguage case reports of CNS abscesses that were the result of primary oral infections. A broad search strategy was employed using any combination of the Medical Search Heading keywords from the following two groups: Group 1 = ‘‘brain abscess,’’ ‘‘intracranial abscess,’’ ‘‘CNS infection’’, ‘‘meningitis’’ and ‘‘cavernous sinus thrombosis’’ and Group 2 = ‘‘odontogenic source,’’ ‘‘dental treatment,’’ ‘‘tooth extraction’’, ‘‘periodontitis’’, and ‘‘oral infection.’’ Search results were screened by title and abstract to select for reports of intracranial abscesses arising from oral-related pathology. In addition, references from screened case reports were reviewed to identify additional cases for inclusion. Articles that met the following criteria were included in our data: (1) cases involving a confirmed intracranial infection; (2) cases describing a presumed dental origin of the infection; (3) a well-described clinical course, including symptoms at presentation and diagnostic work-up; and (4) manuscripts published after 1980. Included articles were subsequently read and data were extracted into an electronic database for analysis. When available, these data included information pertaining to patient demographics, prior medical history, preceding dental procedure or pathology, location of both intra-oral and intracranial infection, time course of symptoms to presentation, results from bloodstream, intra-oral, and intracranial cultures, treatments provided or performed, and clinical outcomes. Authors M.H. and S.M.R. independently reviewed these cases for extraction of all relevant dental and neurosurgical data. As part of our assessment to determine whether dental pathology had spread to CNS locations via direct venous or hematogenous dissemination, we counted the number of times the side of dental involvement was concordant with the side of intracranial pathology. Ipsilateral dental and intracranial pathology would imply direct venous drainage, whereas discordant laterality could imply hematogenous spread. Bilateral dental pathology/procedures were considered concordant with intracranial pathology of either side (tongue piercing procedures were treated as midline, therefore ‘‘bilateral’’, procedures), while unilateral dental pathology/procedures were considered concordant only with intracranial pathology on the same side. Midline intracranial lesions were considered concordant with dental pathology/procedures from either side. Some cases recorded abscesses that spanned across multiple areas within the brain, such as ‘‘fronto-parietal’’; these cases were counted as both frontal and parietal involvement.
3. Results 3.1. Patient demographics Figure 1 shows the search strategy and results. Twenty-two manuscripts were excluded from the analysis (Supp. Table 1). One manuscript did not report evidence of obvious intracranial infection, one reported numerous alternative sources of bacterial infection, and one did not report any obvious intra-oral pathology [16]. Twelve manuscripts were published before 1980 and were therefore excluded. Seven additional manuscripts were excluded because they reported intra-oral bacterial dissemination to spinal structures; we could not discern whether this represented
801
infection of bony column structures or true trans-dural bacterial translocation, and therefore these manuscripts were excluded. A total of 55 manuscripts were identified that met our criteria (Supp. Table 2). From these manuscripts, 60 individual patients were identified and were included in data analysis. No cohort studies or prospective studies were found, and thus all included works represented a relatively low level of evidence. The distribution of publication dates are presented in Figure 2. Of the cases that reported patient sex, males and females accounted for 82.1% (46/ 56) and 17.8% (10/56) of patients, respectively. The mean patient age was 42.1 years, with a range of 3 to 70 years. Medical co-morbidities were poorly reported; most manuscripts did not mention chronic underlying medical conditions, and of those that did, the descriptions were not well-elaborated. No cases reported active immunocompromised conditions; however two cases did report diabetes mellitus, three reported active alcohol abuse, and two reported active polysubstance abuse. Five patients had anatomic cardiac anomalies (three patent foramen ovale, and two unspecified congenital cardiac anomalies), and three patients had hereditary hemorrhagic telangiectasia. 3.2. Precipitating dental pathology or procedures Preceding dental procedures were reported in 41.7% (25/60) of patients and underlying intra-oral pathology was noted in 86.7% (52/60) of patients. In patients without underlying intra-oral pathology, the brain abscess was assumed to be the result of the preceding dental procedure. Many patients had more than one dental pathology. The frequencies of each intra-oral procedure and pathology are presented in Table 1. The majority of procedures were related to molar teeth. Tooth extraction accounted for 60.0% (15/25) of precipitating dental procedures. Dental symptoms or procedures preceded the onset of neurologic symptoms in 40.0% (24/60) of patients. The average time between the performance of a dental procedure and the onset of neurologic symptoms was 17.6 days. 3.3. Microbiology In 13 patients, cultures were obtained from various dental locations, such as the extraction socket, the periodontal pocket, dental plaque, or purulent fluid from intra-oral abscesses. Of these, only eight publications reported micro-organisms that matched those grown from intracranial cultures. Fifty-five cases reported intracranial culture results; two were negative, 49.1% (27/55) were monomicrobial and 47.3% (26/55) were poly-microbial. The frequencies of grown pathogens are presented in Table 2. 3.4. Route of CNS dissemination The locations of reported intracranial involvement are presented in Table 3. As stated in the Methods section, we investigated the route of spread of intra-oral infections into the brain by analyzing the laterality of intra-oral and intracranial infections and whether they were concordant (implying direct venous dissemination) or discordant (implying systemic hematogenous dissemination). Seventy percent (42/60) of cases reported intracranial pathology to be on the same side as underlying intra-oral pathology. Many cases reported generalized intra-oral conditions (for example, gingivitis or periodontitis) and were therefore considered to have ‘‘bilateral’’ intra-oral pathology. Even if these cases are excluded (that is, only taking into account cases reporting clearly defined unilateral intra-oral pathology), again 70.0% (14/20) reported concordant laterality of intra-oral and intracranial pathology. When only taking into account cases with intracranial locations most specifically associated with direct venous drainage
802
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
Fig. 1. Flow diagram showing the case report selection.
Fig. 2. Distribution of dates of publication of case reports included in our review.
into the cranial cavity (ipsilateral frontal lobe involvement, ipsilateral subdural empyema in the anterior fossa, or isolated meningitis) [11], 26.7% (16/60) reported these circumstances. Results of blood cultures were reported in 17 patients; 88.9% of these were negative. Thirteen cases reported results of echocardiography and none of these cases revealed endocardial heart valve vegetations. 3.5. Clinical interventions and outcomes Almost all patients received antibiotic therapy of some sort. The range of antibiotics used was diverse and represented a myriad of
agents. Given this heterogeneity, we have not presented these data. Additionally, antibiotic duration times were rarely reported accurately, therefore they are not presented. Neurosurgical interventions were reported in 85.0% (51/60) of patients. The neurosurgical interventions were often vaguely described; the frequency of these interventions as best interpreted by author A.A.M. are presented in Table 4. Tooth extractions were performed in 31.7% (19/60) of patients as part of treatment (these procedures are separate from dental procedures performed preceding the onset of intracranial disease). Clinical outcomes of patients are presented in Table 5.
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
803
Table 1 Location and frequency of dental procedures and intra-oral pathology Location Left Right Bilateral None
18 (35.3%) 2 (3.9%) 28 (54.9%) 3 (5.9%)
Mandibular Maxillary Both None
12 (24.0%) 10 (20.0%) 34 (68.0%) 3 (6.0%)
Incisor Canine Premolar Molar Full mouth involvement
3 (5.0%) 3 (5.0%) 6 (10.0%) 19 (31.7%) 20 (33.3%)
Dental procedure
Frequency (as percentage of total cases that reported a preceding dental procedure, n = 25)
Extraction Tongue piercing Orthodontic procedure Root canal Scaling and root planning Prophylactic cleaning Maxillofacial surgery Unspecified procedures
15 (60.0%) 2 (8.0%) 2 (8.0%) 1 (4.0%) 1 (4.0%) 1 (4.0%) 1 (4.0%) 2 (8.0%)
Dental symptoms/procedures precede onset of neurologic symptoms? Yes No Not obvious
34 (56.7%) 21 (35.0%) 5 (8.3%)
Intra-oral pathology
Frequency (as percentage of total cases, n = 60)
Caries Caries with periapical involvement Caries total Gingivitis Periodontitis Jaw osteomyelitis Gum injury Oral candidiasis Unspecified ‘‘tooth infection’’ Associated cavernous sinus involvement Associated peri-oral soft tissue abscess Associated paranasal sinus involvement No dental pathology (cases involving a preceding dental procedure without dental pathology: orthodontic procedures, tongue piercings, etc)
15 (25.0%) 21 (35.0%) 36 (60.0%) 2 (3.3%) 26 (43.3%) 2 (3.3%) 2 (3.3%) 1 (1.7%) 2 (3.3%) 3 (5.0%) 7 (11.7%) 6 (10.0%) 8 (13.3%)
Percentages are based on the frequency of each variable (for example, left versus right, mandible versus maxilla) compared to the total number of patients within our collection who reported that stated variable. For frequency of tooth involvement, many patients reported multiple involved teeth within each case, therefore percentages are stated as the frequency of each tooth type relative to all reported patients (n = 60). For instance, 5.0% of all patients reported incisor involvement.
4. Discussion This study utilized a systematic review methodology to overcome the inherent difficulties of finding a sizeable cohort of patients that present with this rare clinical situation. The results of this review provide some degree of insight into the pathogenesis, clinical presentation, microbiology, and clinical outcomes of this disease. The parameters that were selected for evaluation, such as demographics, microbiology, or preceding procedures were selected for their utility to inform practitioners of the important clinical features of this potentially fatal clinical situation.
4.1. How do bacterial infections spread from the oral cavity to the CNS? Odontogenic infections can theoretically enter the cranial vault via one of four routes: (1) systemic hematogenous bacteremia; (2) direct venous drainage via the two main venous networks leading to the cavernous sinus, the facial and the pterygoid vein systems; (3) inoculation via contiguous extension or by introduction of
foreign objects; and (4) lymphatic drainage. Understanding which of these routes is predominant in the pathogenesis of odontogenic CNS infections has important clinical implications on issues such as the necessity of echocardiography, the interpretation of peripheral blood culture results, and the significance of right-to-left shunts. We counted the prevalence of cavernous sinus involvement and the concordance of intra-oral and intracranial laterality in order to analyze the likelihood of intra-oral infections draining via the facial and pterygoid systems and emissary veins into the cranial vault. Discordant laterality or posterior intracranial involvement would imply spread occurred via a hematogenous pathway rather than this direct method. In our study, of the total 57 patients with reported laterality of intra-oral pathology, 26.7% (16/60 patients overall) had involvement of the ipsilateral frontal lobe (whether by parenchymal abscess or subdural/epidural empyema) or had bacterial meningitis. Overall, only 31.7% had frontal lobe involvement, and 5.0% had clinical evidence of cavernous sinus involvement. We believe that if direct venous drainage played a predominate role in intracranial dissemination, the frequency of these specific parameters would be much higher. Therefore, hematogenous spread appears to be
804
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
Table 2 Frequency of reported pathogenic micro-organisms Negative cultures
2
Gram positive cocci Streptococcus Viridans, total S. Viridans unspecified species S. Mitis group (i.e. S. Oralis) S. Mutans group S. Anginosus group
48 33 2 1 1 17
S. Sanguinis group Streptococcus Pneumoniae Unspecified Streptococcus species Actinomyces
3 1 8 7
Staphylococcus
5
Enterococcus
2
Gram positive bacillus Corynebacterium Arcanobacterium hemolyticum Propionibacterium Bacillus
8 2 1 2 2
Anaerobes Fusobacterium
31 12
Prevotella
6
Peptostreptococcus
8
Bacteroides
3
Eubacterium Porphyromonas gingivalis
1 1
Gram negative bacillus Aggregatibacter actinomycetemcomitans Eikenella corrodens Campylobacter
20 5 4 2
Veillonella
2
Wolinella species Escherichia coli Enterobacter cloacae Capnocytophaga ochracea Burkholderia cepacia
1 1 1 1 1
Unspecified subspecies S. Anginosus S. Intermedius S. Constellatus
4 3 8 2
Unspecified subspecies A. israelii A. meyeri A. odontolyticus S. aureus S. epidermidis S. saprophyticus S. warneri
3 1 1 2 1 2 1 1
Includes all cases reported as S. Milleri
Includes one patient reported as ‘‘diphtheroids species’’
B. subtilis B. circulans
1 1
Unspecified subspecies F. nucleatum F. polyphorum Unspecified subspecies P. melaninogenica P. denticola Unspecified subspecies P. tetradius P. anaerobius P. micros Unspecified subspecies B. fragilis
4 7 1 1 4 1 3 1 1 3 2 1
Formerly known as Bacteroides melaninogenica and Bacteroides oralis
Includes one patient reported as unspecified gram negative rod Formerly known as Actinobacillus actinomycetemcomitans, Haemophilus aphrophilus, Haemophilus paraphrophilus C. gracilis C. ureolyticus Unspecified subspecies V. alcalescans
the more important pathophysiological route of spread. Since involvement of posterior intracranial structures (13.3% of cases) are less likely to be due to direct venous drainage, this further suggests that hematogenous spread is the culprit of intracranial dissemination. 4.2. Which dental lesions or procedures are at higher risk for odontogenic CNS infections? Molar teeth are most commonly associated with CNS infections, and represent the largest proportion of implicated teeth. There appeared to be no predilection for mandibular versus maxillary involvement. This finding corroborates those made by Haymaker et al., who demonstrated that maxillary and mandibular foci for infection had equal incidence in relation to intracranial infections [10].
1 1 1 1
Formerly known as Bacteroids ureolytticus
Almost half of reports document caries and periodontitis as an underlying intra-oral pathology. The majority of cases reporting caries also reported periapical involvement, however many of the cases describing isolated caries used vague terms such as ‘‘decayed teeth’’ and did not elaborate on the extent of involvement. It is possible that many of these patients also likely had some degree of periapical involvement. Therefore, we propose that periodontitis or caries with periapical involvement, particularly of the molar teeth, have the highest risk of causing a CNS infection. 4.3. Discussion of microbiology Forty-seven percent of CNS infections in our study were polymicrobial, which also correlates with previous studies, as 32–60% of brain abscesses have been shown to be poly-microbial [1,4]. Intracranial cultures correlated poorly with both peripheral blood
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806 Table 3 Locations of central nervous system infections Location
Frequency
Frontal lobe Parietal lobe Temporal lobe Posterior circulation (brainstem, cerebellum and occipital lobes) Subdural/epidural empyema without intra-parenchymal involvement Thalamus Pituitary gland Isolated meningitis/ventriculitis Multiple intracranial locations Left Right Bilateral or midline lesions
19 (31.7%) 13 (21.7%) 8 (13.3%) 8 (13.3%) 5 (8.3%) 1 (1.7%) 1 (1.7%) 3 (5.0%) 10 (16.7%) 23 (38.3%) 20 (33.3%) 17 (28.3%)
Many cases reported more than one intracranial location, therefore each location was counted separately, and percentages are as proportion of total patients (n = 60).
Table 4 Clinical interventions and outcomes Neurosurgical procedures
Frequency
Craniotomy Burr hole aspiration Stereotactic-guided biopsy Unspecified neurosurgical procedure
31 17 2 1
Table 5 Frequencies of clinical outcomes (percentage calculated from total reported outcomes, n = 55) Died
Survived with neurologic sequelae
Survived without neurologic sequelae
Outcome not reported
5 (8.3%)
17 (28.3%)
33 (55.0%)
5 (8.3%)
and dental/oral cultures. Common bacteria reported included gram positive organisms such as Streptococcus viridans, particularly the anginosus subgroup, and Actinomyces. Common anaerobes included Peptostreptococcus, Prevotella, and Fusobacterium. Common gram negative bacteria included Aggregatibacter actinomycetemcomitans and Eikenella corrodens. Standard bacterial culturing may not serve as an effective method for pathogen identification in some cases. Many oral bacteria are fastidious pathogens that may not necessarily grow in culture sufficiently to allow easy identification. Actinomycotic infections serve as a good example. These bacteria cause infections almost exclusively in the presence of other pathogenic bacteria, socalled ‘‘companion’’ or ‘‘associate’’ pathogens. It is believed actinomycotic bacteria rely on other bacteria to decrease the oxygen concentration of the local tissue, thereby creating the ideal anaerobic milieu for actinomycotic growth. In return, the actinomycotic organisms aggregate around their companion bacteria and confer some protection against phagocytosis and antibiotics [9]. Actinomycotic bacteria are difficult to culture and have to be taken to the laboratory immediately under strict anaerobic conditions [15]. For these reasons, more advanced methods of microbiological investigation such as gene amplification and sequencing technologies are being used. For instance, within our collection of cases, several authors described using 16S rRNA polymerase chain reaction techniques to identify pathogenic bacteria [14,18,19]. These techniques in lieu of standard bacterial culture may be useful for several reasons. Although they do not test for individual antibiotic susceptibility, they nonetheless may help identify which class of the pathogen (gram positive versus gram negative), which may subsequently broaden or narrow empiric
805
antibiotic coverage. Secondly, if intra-oral and intracranial pathogens can be matched, this may help identify the source of an otherwise cryptogenic intracranial abscess. Mueller et al. attempted to correlate brain abscesses with subgingival microflora by use of conventional culturing techniques [13]. Of the 11 patients investigated, only six had correlating subgingival pathogens with isolates taken from brain culture. Gene amplification techniques may serve as better modalities for correlating these micro-organism populations. Da Silva et al. used three genetic fingerprinting techniques – restriction fragment length polymorphisms, ribotyping, and random amplified polymorphic DNA – to successfully match similarities of bacteria in the periodontal pocket with those found in brain abscesses [6]. 4.4. When should oral sources be suspected? Of all the patients, 40.0% (20/60) had intra-oral symptoms or a recent dental procedure prior to the onset of neurologic symptoms. When procedures were performed, the average time until the onset of neurologic symptoms was approximately 17.6 days. Site of intracranial involvement, results of blood cultures, and the age of the patient do not appear to correlate well with the presence of intra-oral pathology. Ewald et al. have proposed criteria for establishing the diagnosis of odontogenic brain abscesses. Three criteria must be fulfilled to make the diagnosis: (1) no alternative source of bacteremia must be found; (2) microbiological studies reveal organisms typically found in oral microflora; and (3) clinical or radiographic signs of active dental or periodontal disease must be present [8]. Although we agree with these criteria, we would add several caveats. Dental procedures even in the absence of dental pathosis can potentially cause CNS infections, particularly if performed within 1 to 4 weeks prior to the onset of disease. If standard culturing techniques are used, results may not necessarily reveal all bacteria present in a CNS lesion. Finally, the presence of endocarditis should not detract from investigation into the mouth, as this is a low cost screening study that may provide key information pertaining to the infectious source. 4.5. Which interventions should be performed? Treatment of odontogenic brain abscesses should follow established evidence-based recommendations for treatment of brain abscesses in general [2]. The results of our study will not fundamentally change most of these recommendations; however they may be used to refine some clinical decisions. For instance, described odontogenic organisms include gram positive, gram negative, and anaerobic pathogens, therefore empiric coverage should be broad spectrum when oral sources of origin are suspected. Blood and intra-oral cultures correlated poorly with intracranial pathogens, therefore these should not be used to guide antibiotic selection. Given the heterogeneity of antibiotics used, durations of therapy, and neurosurgical procedures performed, it is not possible to make conclusions of the most appropriate respective interventions. One group of authors proposed normalization of C-reactive protein and white blood cell count, and complete resolution of the inflammatory changes on MRI as adequate criteria for termination of antibiotics [8]. Some authors such as Milli et al. advocated closure of intra-cardiac anatomic anomalies, especially patent foramen ovale, in patients who develop brain abscesses of odontogenic origin [12]. Without formal studies, it remains unclear if such interventions would have meaningful clinical benefits. Our study has several inherent limitations. Primary among these was the heterogeneity of the reported cases and the predilection for publication bias. Included manuscripts varied significantly in
806
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
format, quality, and reported information, and the nature of the current study cannot preclude the inherent publication or reporting bias associated with this type of review. Details of various parameters were not consistently reported, therefore we had to standardize some of the data collected. For example, if a manuscript reported ‘‘widespread periodontal disease’’ we would interpret this as bilateral oral disease with involvement of both maxilla and mandible. As a result, the data in our review fall subject to the biases of our interpretations of the reported cases. For example, we encountered two unusual discoveries. Firstly, there was a higher proportion of males as compared to females, which is contrary to the known nationwide prevalence of dental caries or periodontitis between sexes [3]. Secondly, we observed an increased proportion of left-sided intraoral pathology with 18 (35.3%) left versus two (3.9%) right. In both these observations, we see no obvious reason for these disparities and believe these may be anomalous findings due to the nonrandom nature of our study. Given the heterogeneity of interventions and antibiotics used, we were unable to investigate and answer several clinical questions, such as the efficacy of outcomes of early versus late dental intervention, the ideal duration of antibiotic use, or the ideal antibiotics agents to use. 5. Conclusion Intracranial abscesses are a life-threatening and largely preventable condition. In 40.0% of cases, there is evidence of dental pathology before the onset of neurologic symptoms. Caries with periapical involvement and periodontitis are the most common precipitating factors. Dental procedures can also precede such infections, especially extractions of wisdom teeth. There is no predilection for maxillary or mandibular teeth. The average time between the performance of a procedure and the development of neurologic symptoms is approximately 2–3 weeks. Intracranial infection location does not necessarily correlate with the side of the originating intra-oral source, suggesting that bacteria most likely enter the cranial vault via hematogenous spread rather than by venous drainage. Common bacterial organisms include Streptococcus viridans (especially anginosus group), Actinomyces, Peptostreptococcus, Prevotella, Fusobacterium, Aggregatibacter actinomycetemcomitans and Eikenella corrodens. Further studies are required to determine the ideal timing of dental intervention after intracranial infections have developed, and the optimal antibiotic regimen and course. Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication.
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jocn.2014.11.015. References [1] Andersen WC, Horton HL. Parietal lobe abscess after routine periodontal recall therapy. Report of a case. J Periodontol 1990;61:243–7. [2] Arlotti M, Grossi P, Pea F, et al. Consensus document on controversial issues for the treatment of infections of the central nervous system: bacterial brain abscesses. Int J Infect Dis 2010;14:S79–92. [3] Beltrán-Aguilar ED, Barker LK, Canto MT, et al. Surveillance for dental caries, dental sealants, tooth retention, edentulism, and enamel fluorosis–United States, 1988–1994 and 1999–2002. MMWR Surveill Summ 2005;54:1–43. [4] Carpenter J, Stapleton S, Holliman R. Retrospective analysis of 49 cases of brain abscess and review of the literature. Eur J Clin Microbiol Infect Dis 2007;26:1–11. [5] Corson M, Postlethwaite K, Seymour R. Are dental infections a cause of brain abscess? Case report and review of the literature. Oral Dis 2001;7:61–5. [6] da Silva RM, Caugant DA, Josefsen R, et al. Characterization of Streptococcus constellatus strains recovered from a brain abscess and periodontal pockets in an immunocompromised patient. J Periodontol 2004;75:1720–3. [7] Dewhirst FE, Chen T, Izard J, et al. The human oral microbiome. J Bacteriol 2010;192:5002–17. [8] Ewald C, Kuhn S, Kalff R. Pyogenic infections of the central nervous system secondary to dental affections—a report of six cases. Neurosurg Rev 2006;29:163–7. [9] Haggerty CJ, Tender GC. Actinomycotic brain abscess and subdural empyema of odontogenic origin: case report and review of the literature. J Oral Maxillofac Surg 2012;70:e210–3. [10] Haymaker W. Fatal infections of the central nervous system and meninges after tooth extraction: with an analysis of twenty-eight cases. Am J Orthod Oral Surg 1945;31:A117–88. [11] Hollin SA, Hayashi H, Gross SW. Intracranial abscesses of odontogenic origin. Oral Surg Oral Med Oral Pathol 1967;23:277–93. [12] Milli B, Rocci A, Paganelli E, et al. Brain abscess of odontogenic origin in a man with interatrial defect. Acta Biomed Ateneo Parmense 2010;81:225. [13] Mueller AA, Saldamli B, Stübinger S, et al. Oral bacterial cultures in nontraumatic brain abscesses: results of a first-line study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;107:469–76. [14] Rahamat-Langendoen JC, van Vonderen MG, Engström LJ, et al. Brain abscess associated with Aggregatibacter actinomycetemcomitans: case report and review of literature. J Clin Periodontol 2011;38:702–6. [15] Roth M, Montone KT. Actinomycosis of the paranasal sinuses: a case report and review. Otolaryngol Head Neck Surg 1996;114:818–21. [16] Roy S, Ellenbogen JM. Seizures, frontal lobe mass, and remote history of periodontal abscess. Arch Pathol Lab Med 2005;129:805–6. [17] Shea BJ, Hamel C, Wells GA, et al. AMSTAR is a reliable and valid measurement tool to assess the methodological quality of systematic reviews. J Clin Epidemiol 2009;62:1013–20. [18] Stepanovic´ S, Tosic´ T, Savic´ B, et al. Brain abscess due to Actinobacillus actinomycetemcomitans. APMIS 2005;113:225–8. [19] Wang H-K, Chen Y-C, Teng L-J, et al. Brain abscess associated with multidrugresistant Capnocytophaga ochracea infection. J Clin Microbiol 2007;45:645–7.
Contents lists available at ScienceDirect
Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn
Review
Intracranial bacterial infections of oral origin Alan A. Moazzam a,⇑, Sowmya M. Rajagopal b, Parish P. Sedghizadeh b, Gabriel Zada c, Mina Habibian b a
Department of General Internal Medicine, Keck Hospital of University of Southern California, 1500 San Pablo Street, Los Angeles, CA 90033, USA Ostrow School of Dentistry of University of Southern California, Los Angeles, CA, USA c Department of Neurosurgery, Keck School of Medicine, University of Southern California, Los Angeles, CA, USA b
a r t i c l e
i n f o
Article history: Received 5 October 2014 Accepted 8 November 2014
Keywords: Brain abscess CNS abscess Dental infections Intracranial abscess Oral infections Periodontitis
a b s t r a c t Brain abscesses are rare but potentially deadly complications of odontogenic infections. This phenomenon has been described mainly in the form of case reports, as large-scale studies are difficult to perform. We compiled a total of 60 previously published cases of such a complication to investigate the predisposing factors, microbiology, and clinical outcomes of intracranial abscesses of odontogenic origin. A systematic review of the literature using the PubMed database was performed. Men accounted for 82.1% of cases, and the mean age was 42.1 years. Caries with periapical involvement and periodontitis were the two most common intra-oral sources, and wisdom tooth extraction was the most common preceding dental procedure. In 56.4% of cases, there were obvious signs of dental disease prior to development of intracranial infection. Commonly implicated microorganisms included Streptococcus viridans (especially the anginosus group), Actinomyces, Peptostreptococcus, Prevotella, Fusobacterium, Aggregatibacter actinomycetemcomitans and Eikenella corrodens. There was an 8.3% mortality rate. Intracranial abscesses can form anywhere within the brain, and appear unrelated to the side of dental involvement. This suggests that hematogenous spread is the most likely route of dissemination. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction Brain abscesses are life-threatening infections that require immediate neurosurgical attention. The incidence is between one and eight per 100,000 patients per year in the USA [8]. Brain abscesses most frequently occur due to bacterial dissemination from a primary lesion at a distant site. They can also occur from direct contiguous invasion from an adjacent site of infection. The oral cavity harbors a diverse and abundant microflora. As part of the Human Microbiome Project, culture-independent molecular approaches have identified more than 1200 different types of microbes in the human mouth [7]; 350 different bacterial strains have been isolated in marginal periodontitis and 150 in endodontic infections [5]. Bacteria can disseminate from these locations due to a variety of dental conditions such as gingivitis and periodontitis, and also from procedures such as tooth extractions, endodontic treatments, and oral surgery. Even simple tooth brushing has been shown to induce transient bacteremia in 38.5% of cases [8]. Currently, the Council of Dental Therapeutics of the American Dental Association has recommended against routine antibiotic prophylaxis for dental procedures specifically to ⇑ Corresponding author. Tel.: +1 858 342 0874. E-mail address: [email protected] (A.A. Moazzam). http://dx.doi.org/10.1016/j.jocn.2014.11.015 0967-5868/Ó 2014 Elsevier Ltd. All rights reserved.
prevent brain abscesses given the relative low incidence of this complication and the potential for development of resistant strains [8]. Due to the rarity of this occurrence, to our knowledge no studies have specifically investigated this phenomenon in a comprehensive manner. The literature primarily consists of individual case reports. We systematically reviewed the literature to identify and analyze all cases representing central nervous system (CNS) infections of oral origin. Our goal was to provide a synthesis of available knowledge on the natural history, pathogenesis, microbiology, interventions and outcomes of this potentially fatal pathology.
2. Materials and methods Our review followed recommendations as outlined in the AMSTAR guidelines for systematic reviews [17]. We used an a priori research design that would allow us to better understand this clinical problem and provide insight into PICO-related questions (Patient problem or population, Intervention, Comparison and Outcome) in the context of brain abscesses of oral origin. A protocol specifying all aspects of the review method was developed before commencing the review. This protocol included ascertainment criteria for considering studies for this review, search methods for identification of studies, data collection and analysis.
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
The protocol was planned to minimize the effect of author bias on the review and the potential to alter or misinterpret data presented on study findings. We critically appraised the selected works and synthesized the data according to a predetermined and explicit methodology. This included quality assessment of reviewed papers, determination of levels of evidence, and evaluation of outcome measures but no statistical analyses given the nature of the available data and literature which we describe shortly. The PubMed electronic database was searched for all Englishlanguage case reports of CNS abscesses that were the result of primary oral infections. A broad search strategy was employed using any combination of the Medical Search Heading keywords from the following two groups: Group 1 = ‘‘brain abscess,’’ ‘‘intracranial abscess,’’ ‘‘CNS infection’’, ‘‘meningitis’’ and ‘‘cavernous sinus thrombosis’’ and Group 2 = ‘‘odontogenic source,’’ ‘‘dental treatment,’’ ‘‘tooth extraction’’, ‘‘periodontitis’’, and ‘‘oral infection.’’ Search results were screened by title and abstract to select for reports of intracranial abscesses arising from oral-related pathology. In addition, references from screened case reports were reviewed to identify additional cases for inclusion. Articles that met the following criteria were included in our data: (1) cases involving a confirmed intracranial infection; (2) cases describing a presumed dental origin of the infection; (3) a well-described clinical course, including symptoms at presentation and diagnostic work-up; and (4) manuscripts published after 1980. Included articles were subsequently read and data were extracted into an electronic database for analysis. When available, these data included information pertaining to patient demographics, prior medical history, preceding dental procedure or pathology, location of both intra-oral and intracranial infection, time course of symptoms to presentation, results from bloodstream, intra-oral, and intracranial cultures, treatments provided or performed, and clinical outcomes. Authors M.H. and S.M.R. independently reviewed these cases for extraction of all relevant dental and neurosurgical data. As part of our assessment to determine whether dental pathology had spread to CNS locations via direct venous or hematogenous dissemination, we counted the number of times the side of dental involvement was concordant with the side of intracranial pathology. Ipsilateral dental and intracranial pathology would imply direct venous drainage, whereas discordant laterality could imply hematogenous spread. Bilateral dental pathology/procedures were considered concordant with intracranial pathology of either side (tongue piercing procedures were treated as midline, therefore ‘‘bilateral’’, procedures), while unilateral dental pathology/procedures were considered concordant only with intracranial pathology on the same side. Midline intracranial lesions were considered concordant with dental pathology/procedures from either side. Some cases recorded abscesses that spanned across multiple areas within the brain, such as ‘‘fronto-parietal’’; these cases were counted as both frontal and parietal involvement.
3. Results 3.1. Patient demographics Figure 1 shows the search strategy and results. Twenty-two manuscripts were excluded from the analysis (Supp. Table 1). One manuscript did not report evidence of obvious intracranial infection, one reported numerous alternative sources of bacterial infection, and one did not report any obvious intra-oral pathology [16]. Twelve manuscripts were published before 1980 and were therefore excluded. Seven additional manuscripts were excluded because they reported intra-oral bacterial dissemination to spinal structures; we could not discern whether this represented
801
infection of bony column structures or true trans-dural bacterial translocation, and therefore these manuscripts were excluded. A total of 55 manuscripts were identified that met our criteria (Supp. Table 2). From these manuscripts, 60 individual patients were identified and were included in data analysis. No cohort studies or prospective studies were found, and thus all included works represented a relatively low level of evidence. The distribution of publication dates are presented in Figure 2. Of the cases that reported patient sex, males and females accounted for 82.1% (46/ 56) and 17.8% (10/56) of patients, respectively. The mean patient age was 42.1 years, with a range of 3 to 70 years. Medical co-morbidities were poorly reported; most manuscripts did not mention chronic underlying medical conditions, and of those that did, the descriptions were not well-elaborated. No cases reported active immunocompromised conditions; however two cases did report diabetes mellitus, three reported active alcohol abuse, and two reported active polysubstance abuse. Five patients had anatomic cardiac anomalies (three patent foramen ovale, and two unspecified congenital cardiac anomalies), and three patients had hereditary hemorrhagic telangiectasia. 3.2. Precipitating dental pathology or procedures Preceding dental procedures were reported in 41.7% (25/60) of patients and underlying intra-oral pathology was noted in 86.7% (52/60) of patients. In patients without underlying intra-oral pathology, the brain abscess was assumed to be the result of the preceding dental procedure. Many patients had more than one dental pathology. The frequencies of each intra-oral procedure and pathology are presented in Table 1. The majority of procedures were related to molar teeth. Tooth extraction accounted for 60.0% (15/25) of precipitating dental procedures. Dental symptoms or procedures preceded the onset of neurologic symptoms in 40.0% (24/60) of patients. The average time between the performance of a dental procedure and the onset of neurologic symptoms was 17.6 days. 3.3. Microbiology In 13 patients, cultures were obtained from various dental locations, such as the extraction socket, the periodontal pocket, dental plaque, or purulent fluid from intra-oral abscesses. Of these, only eight publications reported micro-organisms that matched those grown from intracranial cultures. Fifty-five cases reported intracranial culture results; two were negative, 49.1% (27/55) were monomicrobial and 47.3% (26/55) were poly-microbial. The frequencies of grown pathogens are presented in Table 2. 3.4. Route of CNS dissemination The locations of reported intracranial involvement are presented in Table 3. As stated in the Methods section, we investigated the route of spread of intra-oral infections into the brain by analyzing the laterality of intra-oral and intracranial infections and whether they were concordant (implying direct venous dissemination) or discordant (implying systemic hematogenous dissemination). Seventy percent (42/60) of cases reported intracranial pathology to be on the same side as underlying intra-oral pathology. Many cases reported generalized intra-oral conditions (for example, gingivitis or periodontitis) and were therefore considered to have ‘‘bilateral’’ intra-oral pathology. Even if these cases are excluded (that is, only taking into account cases reporting clearly defined unilateral intra-oral pathology), again 70.0% (14/20) reported concordant laterality of intra-oral and intracranial pathology. When only taking into account cases with intracranial locations most specifically associated with direct venous drainage
802
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
Fig. 1. Flow diagram showing the case report selection.
Fig. 2. Distribution of dates of publication of case reports included in our review.
into the cranial cavity (ipsilateral frontal lobe involvement, ipsilateral subdural empyema in the anterior fossa, or isolated meningitis) [11], 26.7% (16/60) reported these circumstances. Results of blood cultures were reported in 17 patients; 88.9% of these were negative. Thirteen cases reported results of echocardiography and none of these cases revealed endocardial heart valve vegetations. 3.5. Clinical interventions and outcomes Almost all patients received antibiotic therapy of some sort. The range of antibiotics used was diverse and represented a myriad of
agents. Given this heterogeneity, we have not presented these data. Additionally, antibiotic duration times were rarely reported accurately, therefore they are not presented. Neurosurgical interventions were reported in 85.0% (51/60) of patients. The neurosurgical interventions were often vaguely described; the frequency of these interventions as best interpreted by author A.A.M. are presented in Table 4. Tooth extractions were performed in 31.7% (19/60) of patients as part of treatment (these procedures are separate from dental procedures performed preceding the onset of intracranial disease). Clinical outcomes of patients are presented in Table 5.
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
803
Table 1 Location and frequency of dental procedures and intra-oral pathology Location Left Right Bilateral None
18 (35.3%) 2 (3.9%) 28 (54.9%) 3 (5.9%)
Mandibular Maxillary Both None
12 (24.0%) 10 (20.0%) 34 (68.0%) 3 (6.0%)
Incisor Canine Premolar Molar Full mouth involvement
3 (5.0%) 3 (5.0%) 6 (10.0%) 19 (31.7%) 20 (33.3%)
Dental procedure
Frequency (as percentage of total cases that reported a preceding dental procedure, n = 25)
Extraction Tongue piercing Orthodontic procedure Root canal Scaling and root planning Prophylactic cleaning Maxillofacial surgery Unspecified procedures
15 (60.0%) 2 (8.0%) 2 (8.0%) 1 (4.0%) 1 (4.0%) 1 (4.0%) 1 (4.0%) 2 (8.0%)
Dental symptoms/procedures precede onset of neurologic symptoms? Yes No Not obvious
34 (56.7%) 21 (35.0%) 5 (8.3%)
Intra-oral pathology
Frequency (as percentage of total cases, n = 60)
Caries Caries with periapical involvement Caries total Gingivitis Periodontitis Jaw osteomyelitis Gum injury Oral candidiasis Unspecified ‘‘tooth infection’’ Associated cavernous sinus involvement Associated peri-oral soft tissue abscess Associated paranasal sinus involvement No dental pathology (cases involving a preceding dental procedure without dental pathology: orthodontic procedures, tongue piercings, etc)
15 (25.0%) 21 (35.0%) 36 (60.0%) 2 (3.3%) 26 (43.3%) 2 (3.3%) 2 (3.3%) 1 (1.7%) 2 (3.3%) 3 (5.0%) 7 (11.7%) 6 (10.0%) 8 (13.3%)
Percentages are based on the frequency of each variable (for example, left versus right, mandible versus maxilla) compared to the total number of patients within our collection who reported that stated variable. For frequency of tooth involvement, many patients reported multiple involved teeth within each case, therefore percentages are stated as the frequency of each tooth type relative to all reported patients (n = 60). For instance, 5.0% of all patients reported incisor involvement.
4. Discussion This study utilized a systematic review methodology to overcome the inherent difficulties of finding a sizeable cohort of patients that present with this rare clinical situation. The results of this review provide some degree of insight into the pathogenesis, clinical presentation, microbiology, and clinical outcomes of this disease. The parameters that were selected for evaluation, such as demographics, microbiology, or preceding procedures were selected for their utility to inform practitioners of the important clinical features of this potentially fatal clinical situation.
4.1. How do bacterial infections spread from the oral cavity to the CNS? Odontogenic infections can theoretically enter the cranial vault via one of four routes: (1) systemic hematogenous bacteremia; (2) direct venous drainage via the two main venous networks leading to the cavernous sinus, the facial and the pterygoid vein systems; (3) inoculation via contiguous extension or by introduction of
foreign objects; and (4) lymphatic drainage. Understanding which of these routes is predominant in the pathogenesis of odontogenic CNS infections has important clinical implications on issues such as the necessity of echocardiography, the interpretation of peripheral blood culture results, and the significance of right-to-left shunts. We counted the prevalence of cavernous sinus involvement and the concordance of intra-oral and intracranial laterality in order to analyze the likelihood of intra-oral infections draining via the facial and pterygoid systems and emissary veins into the cranial vault. Discordant laterality or posterior intracranial involvement would imply spread occurred via a hematogenous pathway rather than this direct method. In our study, of the total 57 patients with reported laterality of intra-oral pathology, 26.7% (16/60 patients overall) had involvement of the ipsilateral frontal lobe (whether by parenchymal abscess or subdural/epidural empyema) or had bacterial meningitis. Overall, only 31.7% had frontal lobe involvement, and 5.0% had clinical evidence of cavernous sinus involvement. We believe that if direct venous drainage played a predominate role in intracranial dissemination, the frequency of these specific parameters would be much higher. Therefore, hematogenous spread appears to be
804
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
Table 2 Frequency of reported pathogenic micro-organisms Negative cultures
2
Gram positive cocci Streptococcus Viridans, total S. Viridans unspecified species S. Mitis group (i.e. S. Oralis) S. Mutans group S. Anginosus group
48 33 2 1 1 17
S. Sanguinis group Streptococcus Pneumoniae Unspecified Streptococcus species Actinomyces
3 1 8 7
Staphylococcus
5
Enterococcus
2
Gram positive bacillus Corynebacterium Arcanobacterium hemolyticum Propionibacterium Bacillus
8 2 1 2 2
Anaerobes Fusobacterium
31 12
Prevotella
6
Peptostreptococcus
8
Bacteroides
3
Eubacterium Porphyromonas gingivalis
1 1
Gram negative bacillus Aggregatibacter actinomycetemcomitans Eikenella corrodens Campylobacter
20 5 4 2
Veillonella
2
Wolinella species Escherichia coli Enterobacter cloacae Capnocytophaga ochracea Burkholderia cepacia
1 1 1 1 1
Unspecified subspecies S. Anginosus S. Intermedius S. Constellatus
4 3 8 2
Unspecified subspecies A. israelii A. meyeri A. odontolyticus S. aureus S. epidermidis S. saprophyticus S. warneri
3 1 1 2 1 2 1 1
Includes all cases reported as S. Milleri
Includes one patient reported as ‘‘diphtheroids species’’
B. subtilis B. circulans
1 1
Unspecified subspecies F. nucleatum F. polyphorum Unspecified subspecies P. melaninogenica P. denticola Unspecified subspecies P. tetradius P. anaerobius P. micros Unspecified subspecies B. fragilis
4 7 1 1 4 1 3 1 1 3 2 1
Formerly known as Bacteroides melaninogenica and Bacteroides oralis
Includes one patient reported as unspecified gram negative rod Formerly known as Actinobacillus actinomycetemcomitans, Haemophilus aphrophilus, Haemophilus paraphrophilus C. gracilis C. ureolyticus Unspecified subspecies V. alcalescans
the more important pathophysiological route of spread. Since involvement of posterior intracranial structures (13.3% of cases) are less likely to be due to direct venous drainage, this further suggests that hematogenous spread is the culprit of intracranial dissemination. 4.2. Which dental lesions or procedures are at higher risk for odontogenic CNS infections? Molar teeth are most commonly associated with CNS infections, and represent the largest proportion of implicated teeth. There appeared to be no predilection for mandibular versus maxillary involvement. This finding corroborates those made by Haymaker et al., who demonstrated that maxillary and mandibular foci for infection had equal incidence in relation to intracranial infections [10].
1 1 1 1
Formerly known as Bacteroids ureolytticus
Almost half of reports document caries and periodontitis as an underlying intra-oral pathology. The majority of cases reporting caries also reported periapical involvement, however many of the cases describing isolated caries used vague terms such as ‘‘decayed teeth’’ and did not elaborate on the extent of involvement. It is possible that many of these patients also likely had some degree of periapical involvement. Therefore, we propose that periodontitis or caries with periapical involvement, particularly of the molar teeth, have the highest risk of causing a CNS infection. 4.3. Discussion of microbiology Forty-seven percent of CNS infections in our study were polymicrobial, which also correlates with previous studies, as 32–60% of brain abscesses have been shown to be poly-microbial [1,4]. Intracranial cultures correlated poorly with both peripheral blood
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806 Table 3 Locations of central nervous system infections Location
Frequency
Frontal lobe Parietal lobe Temporal lobe Posterior circulation (brainstem, cerebellum and occipital lobes) Subdural/epidural empyema without intra-parenchymal involvement Thalamus Pituitary gland Isolated meningitis/ventriculitis Multiple intracranial locations Left Right Bilateral or midline lesions
19 (31.7%) 13 (21.7%) 8 (13.3%) 8 (13.3%) 5 (8.3%) 1 (1.7%) 1 (1.7%) 3 (5.0%) 10 (16.7%) 23 (38.3%) 20 (33.3%) 17 (28.3%)
Many cases reported more than one intracranial location, therefore each location was counted separately, and percentages are as proportion of total patients (n = 60).
Table 4 Clinical interventions and outcomes Neurosurgical procedures
Frequency
Craniotomy Burr hole aspiration Stereotactic-guided biopsy Unspecified neurosurgical procedure
31 17 2 1
Table 5 Frequencies of clinical outcomes (percentage calculated from total reported outcomes, n = 55) Died
Survived with neurologic sequelae
Survived without neurologic sequelae
Outcome not reported
5 (8.3%)
17 (28.3%)
33 (55.0%)
5 (8.3%)
and dental/oral cultures. Common bacteria reported included gram positive organisms such as Streptococcus viridans, particularly the anginosus subgroup, and Actinomyces. Common anaerobes included Peptostreptococcus, Prevotella, and Fusobacterium. Common gram negative bacteria included Aggregatibacter actinomycetemcomitans and Eikenella corrodens. Standard bacterial culturing may not serve as an effective method for pathogen identification in some cases. Many oral bacteria are fastidious pathogens that may not necessarily grow in culture sufficiently to allow easy identification. Actinomycotic infections serve as a good example. These bacteria cause infections almost exclusively in the presence of other pathogenic bacteria, socalled ‘‘companion’’ or ‘‘associate’’ pathogens. It is believed actinomycotic bacteria rely on other bacteria to decrease the oxygen concentration of the local tissue, thereby creating the ideal anaerobic milieu for actinomycotic growth. In return, the actinomycotic organisms aggregate around their companion bacteria and confer some protection against phagocytosis and antibiotics [9]. Actinomycotic bacteria are difficult to culture and have to be taken to the laboratory immediately under strict anaerobic conditions [15]. For these reasons, more advanced methods of microbiological investigation such as gene amplification and sequencing technologies are being used. For instance, within our collection of cases, several authors described using 16S rRNA polymerase chain reaction techniques to identify pathogenic bacteria [14,18,19]. These techniques in lieu of standard bacterial culture may be useful for several reasons. Although they do not test for individual antibiotic susceptibility, they nonetheless may help identify which class of the pathogen (gram positive versus gram negative), which may subsequently broaden or narrow empiric
805
antibiotic coverage. Secondly, if intra-oral and intracranial pathogens can be matched, this may help identify the source of an otherwise cryptogenic intracranial abscess. Mueller et al. attempted to correlate brain abscesses with subgingival microflora by use of conventional culturing techniques [13]. Of the 11 patients investigated, only six had correlating subgingival pathogens with isolates taken from brain culture. Gene amplification techniques may serve as better modalities for correlating these micro-organism populations. Da Silva et al. used three genetic fingerprinting techniques – restriction fragment length polymorphisms, ribotyping, and random amplified polymorphic DNA – to successfully match similarities of bacteria in the periodontal pocket with those found in brain abscesses [6]. 4.4. When should oral sources be suspected? Of all the patients, 40.0% (20/60) had intra-oral symptoms or a recent dental procedure prior to the onset of neurologic symptoms. When procedures were performed, the average time until the onset of neurologic symptoms was approximately 17.6 days. Site of intracranial involvement, results of blood cultures, and the age of the patient do not appear to correlate well with the presence of intra-oral pathology. Ewald et al. have proposed criteria for establishing the diagnosis of odontogenic brain abscesses. Three criteria must be fulfilled to make the diagnosis: (1) no alternative source of bacteremia must be found; (2) microbiological studies reveal organisms typically found in oral microflora; and (3) clinical or radiographic signs of active dental or periodontal disease must be present [8]. Although we agree with these criteria, we would add several caveats. Dental procedures even in the absence of dental pathosis can potentially cause CNS infections, particularly if performed within 1 to 4 weeks prior to the onset of disease. If standard culturing techniques are used, results may not necessarily reveal all bacteria present in a CNS lesion. Finally, the presence of endocarditis should not detract from investigation into the mouth, as this is a low cost screening study that may provide key information pertaining to the infectious source. 4.5. Which interventions should be performed? Treatment of odontogenic brain abscesses should follow established evidence-based recommendations for treatment of brain abscesses in general [2]. The results of our study will not fundamentally change most of these recommendations; however they may be used to refine some clinical decisions. For instance, described odontogenic organisms include gram positive, gram negative, and anaerobic pathogens, therefore empiric coverage should be broad spectrum when oral sources of origin are suspected. Blood and intra-oral cultures correlated poorly with intracranial pathogens, therefore these should not be used to guide antibiotic selection. Given the heterogeneity of antibiotics used, durations of therapy, and neurosurgical procedures performed, it is not possible to make conclusions of the most appropriate respective interventions. One group of authors proposed normalization of C-reactive protein and white blood cell count, and complete resolution of the inflammatory changes on MRI as adequate criteria for termination of antibiotics [8]. Some authors such as Milli et al. advocated closure of intra-cardiac anatomic anomalies, especially patent foramen ovale, in patients who develop brain abscesses of odontogenic origin [12]. Without formal studies, it remains unclear if such interventions would have meaningful clinical benefits. Our study has several inherent limitations. Primary among these was the heterogeneity of the reported cases and the predilection for publication bias. Included manuscripts varied significantly in
806
A.A. Moazzam et al. / Journal of Clinical Neuroscience 22 (2015) 800–806
format, quality, and reported information, and the nature of the current study cannot preclude the inherent publication or reporting bias associated with this type of review. Details of various parameters were not consistently reported, therefore we had to standardize some of the data collected. For example, if a manuscript reported ‘‘widespread periodontal disease’’ we would interpret this as bilateral oral disease with involvement of both maxilla and mandible. As a result, the data in our review fall subject to the biases of our interpretations of the reported cases. For example, we encountered two unusual discoveries. Firstly, there was a higher proportion of males as compared to females, which is contrary to the known nationwide prevalence of dental caries or periodontitis between sexes [3]. Secondly, we observed an increased proportion of left-sided intraoral pathology with 18 (35.3%) left versus two (3.9%) right. In both these observations, we see no obvious reason for these disparities and believe these may be anomalous findings due to the nonrandom nature of our study. Given the heterogeneity of interventions and antibiotics used, we were unable to investigate and answer several clinical questions, such as the efficacy of outcomes of early versus late dental intervention, the ideal duration of antibiotic use, or the ideal antibiotics agents to use. 5. Conclusion Intracranial abscesses are a life-threatening and largely preventable condition. In 40.0% of cases, there is evidence of dental pathology before the onset of neurologic symptoms. Caries with periapical involvement and periodontitis are the most common precipitating factors. Dental procedures can also precede such infections, especially extractions of wisdom teeth. There is no predilection for maxillary or mandibular teeth. The average time between the performance of a procedure and the development of neurologic symptoms is approximately 2–3 weeks. Intracranial infection location does not necessarily correlate with the side of the originating intra-oral source, suggesting that bacteria most likely enter the cranial vault via hematogenous spread rather than by venous drainage. Common bacterial organisms include Streptococcus viridans (especially anginosus group), Actinomyces, Peptostreptococcus, Prevotella, Fusobacterium, Aggregatibacter actinomycetemcomitans and Eikenella corrodens. Further studies are required to determine the ideal timing of dental intervention after intracranial infections have developed, and the optimal antibiotic regimen and course. Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication.
Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jocn.2014.11.015. References [1] Andersen WC, Horton HL. Parietal lobe abscess after routine periodontal recall therapy. Report of a case. J Periodontol 1990;61:243–7. [2] Arlotti M, Grossi P, Pea F, et al. Consensus document on controversial issues for the treatment of infections of the central nervous system: bacterial brain abscesses. Int J Infect Dis 2010;14:S79–92. [3] Beltrán-Aguilar ED, Barker LK, Canto MT, et al. Surveillance for dental caries, dental sealants, tooth retention, edentulism, and enamel fluorosis–United States, 1988–1994 and 1999–2002. MMWR Surveill Summ 2005;54:1–43. [4] Carpenter J, Stapleton S, Holliman R. Retrospective analysis of 49 cases of brain abscess and review of the literature. Eur J Clin Microbiol Infect Dis 2007;26:1–11. [5] Corson M, Postlethwaite K, Seymour R. Are dental infections a cause of brain abscess? Case report and review of the literature. Oral Dis 2001;7:61–5. [6] da Silva RM, Caugant DA, Josefsen R, et al. Characterization of Streptococcus constellatus strains recovered from a brain abscess and periodontal pockets in an immunocompromised patient. J Periodontol 2004;75:1720–3. [7] Dewhirst FE, Chen T, Izard J, et al. The human oral microbiome. J Bacteriol 2010;192:5002–17. [8] Ewald C, Kuhn S, Kalff R. Pyogenic infections of the central nervous system secondary to dental affections—a report of six cases. Neurosurg Rev 2006;29:163–7. [9] Haggerty CJ, Tender GC. Actinomycotic brain abscess and subdural empyema of odontogenic origin: case report and review of the literature. J Oral Maxillofac Surg 2012;70:e210–3. [10] Haymaker W. Fatal infections of the central nervous system and meninges after tooth extraction: with an analysis of twenty-eight cases. Am J Orthod Oral Surg 1945;31:A117–88. [11] Hollin SA, Hayashi H, Gross SW. Intracranial abscesses of odontogenic origin. Oral Surg Oral Med Oral Pathol 1967;23:277–93. [12] Milli B, Rocci A, Paganelli E, et al. Brain abscess of odontogenic origin in a man with interatrial defect. Acta Biomed Ateneo Parmense 2010;81:225. [13] Mueller AA, Saldamli B, Stübinger S, et al. Oral bacterial cultures in nontraumatic brain abscesses: results of a first-line study. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 2009;107:469–76. [14] Rahamat-Langendoen JC, van Vonderen MG, Engström LJ, et al. Brain abscess associated with Aggregatibacter actinomycetemcomitans: case report and review of literature. J Clin Periodontol 2011;38:702–6. [15] Roth M, Montone KT. Actinomycosis of the paranasal sinuses: a case report and review. Otolaryngol Head Neck Surg 1996;114:818–21. [16] Roy S, Ellenbogen JM. Seizures, frontal lobe mass, and remote history of periodontal abscess. Arch Pathol Lab Med 2005;129:805–6. [17] Shea BJ, Hamel C, Wells GA, et al. AMSTAR is a reliable and valid measurement tool to assess the methodological quality of systematic reviews. J Clin Epidemiol 2009;62:1013–20. [18] Stepanovic´ S, Tosic´ T, Savic´ B, et al. Brain abscess due to Actinobacillus actinomycetemcomitans. APMIS 2005;113:225–8. [19] Wang H-K, Chen Y-C, Teng L-J, et al. Brain abscess associated with multidrugresistant Capnocytophaga ochracea infection. J Clin Microbiol 2007;45:645–7.
