BRAIN ABSCESS CONTINUES to constitute one of the most important neurosurgical emergencies. Prior to 1975, the incidence of multiple brain abscesses was reported as 5–15% [1,2]. However, multiple brain abscesses were being diagnosed more frequently since the advent of the computerized tomography (CT) scan, with a 2–3 fold increase in their incidences [3–8]. Although some series reported the mortality in multiple brain abscesses as higher than in single abscess [4,8], other series reported the mortality of multiple brain abscesses as similar to rates reported for solitary abscess [3,5–7,9]. And although several series had developed some principles to guide the management of patients with multiple brain abscesses [5,7,8], the management of these patients had been subject to greater variability. For example, Dyste et al.  recommended the aspiration of one abscess for the purpose of diagnosis or in cases that had failed to respond to empiric antibiotic therapy, whereas Barsoum et al.  recommended the excision of superficial lesions and the drainage of deep-seated ones. Nonoperative management had been advocated by many, especially when abscesses were in the early cerebritis phase of development [1,7,9,10]. We report our experience of treating 12 patients with multiple brain abscesses and try to develop some therapeutic guidelines and improve the therapeutic strategy for this potentially fatal central nervous system infection.
Patients and Methods
Over a period of 14 years (1986–1999), 117 patients with brain abscess were treated at Kaohsiung Chang Gung Memorial Hospital, the largest medical center in the southern Taiwan. Of these 117 patients, 12 were identified as having multiple pyogenic brain abscesses. These 12 patients were diagnosed with having multiple brain abscesses due to: 1) classical clinical manifestations including headache, fever, localized neurological signs and/or altered consciousness; 2) CT and/or magnetic resonance imaging (MRI) showing characteristic findings of brain abscesses; 3) two or more lesions at a distance from each other, clearly separated by brain parenchyma in neuroimaging studies; and 4) resolution of the lesions following treatment. For each patient, presenting symptoms, underlying source of infection, site of abscess, therapeutic methods, and outcome were retrospectively reviewed. The neuro images were carefully reviewed to record the number, size, location, and mass effect of brain abscesses. Mass effect was defined as positive when the abscess caused midline shift or compression of ventricular system in CT scan. Special attention was paid to the medically treated patients to assess the evolutional changes of abscesses in the follow-up CT scans.
Treatment was judged according to clinical status, neuroradiographic findings, and therapeutic response. Surgical treatment consisted of either aspiration or evacuation. Aspiration was to aspirate the content of the abscess with ventricular catheter via a burr hole or a small craniotomy and leave the capsule alone. Craniotomy and resection of the abscess was defined as evacuation. In patients with positive cultures, the choice of antibiotic therapy was based on the susceptibility tests. If all cultures were negative, antibiotics were selected empirically. The duration of antibiotic treatment was dependent on the therapeutic response. Therapeutic outcome was assessed according to the Glasgow outcome scale (GOS): good recovery, moderate disability, severe disability, persistent vegetative state, and death. Mortality was defined as death by any cause occurring during this hospitalization.
Clinical and neuroradiographic findings, treatment, and therapeutic outcomes of patients are summarized in Tables 1, 2 and 3. Among them, 8 were male and 4 were female. Of these 12 patients, the source of infection could be identified in 10 patients. In other 2 patients, the source of infection could not be found, although extensive surveys were conducted. Hematogenous spread from a remote focus, the most common source of infection in our patients, was noted in 6 patients. Three patients were found to have gotten the infection from a parameningeal focus. Postoperative infection was noted in one patient.
Headache was the most common symptoms in our patients. Vomiting was noted in seven patients and fever in six. Only two patients were admitted with altered consciousness. Other clinical manifestations included seizure in three patients, and hemiparesis in two patients. Streptococcus species (Viridans streptococci in three, Peptostreptococcus in one and Viridans streptococci mixed with Peptostreptococcus in one) were the most frequently isolated organisms in these patients. Gram-negative bacilli, including Klebsiella (K.) pneumoniae, Escherichia (E.) coli, and Proteus mirabilis were the prevalent pathogens associated with hematogenous spread.
Except for one abscess located in the cerebellum, the locations of these abscesses were supratentorial. The most common location of the abscesses was the frontal lobe. The left hemisphere was involved more than the right hemisphere, with five in the left, three in the right and three bilaterally in both hemispheres. In two patients (Patients 9 and 10), more than five abscesses were found in initial CT scan. In six patients, the size of the largest abscess was larger than 3 cm, five of which caused significant mass effect. In the other six cases where the size of the abscesses was less than 3 cm, no mass effect was noted. The neuroradiographic findings on contrast-enhanced scans were ringenhancing lesions surrounded by perifocal edema in all patients. Clinical improvement preceded radiological improvement in three patients (Patients 10–12) with nonsurgical treatment, and the interval between abscess diagnoses to first decrease in size after antimicrobial treatments was 2 to 4 weeks. The duration of final resolution of all abscesses was 10 to 12 weeks (Table 3).
In these patients, eight received both surgical and antimicrobial therapy, and four received antimicrobial therapy alone. Among the eight patients who underwent surgical treatment, seven accepted evacuation, and one accepted aspiration. Patient 4 underwent two operations to resect the cerebral and cerebellar abscess, respectively. Patient 5 accepted a second operation to remove another abscess, which failed to respond to 2 weeks of antibiotics. Patient 7 accepted the operation because the abscesses failed to respond to 2 weeks of antibiotic treatment. The abscesses were small in size without any mass effect in three cases (Patients 10–12). They received nonsurgical treatment and followed by serial neuroimaging studies. Most patients accepted intravenous antibiotics for 4–6 weeks. The criteria for stopping antibiotics were improved clinical symptomatology and progressive resolution of abscess on neuroimaging studies. Penicillin was given in eight patients and third-generation cephalosporin was given to two patients. Others used antibiotics including ticarcillin, ampicillin, moxalactam, and chloramphenicol (Table 1). Patient 9 was admitted in a deep comatose state with septic shock, and finally died. Overall mortality rate was 8.3%. In the survivors, three had chronic epilepsy.
The incidence of multiple brain abscesses varies from 10–40% [3,5–8,11]. The observed incidence of 10.3% in the presented series is similar to a previously published report from India . Kratimenos et al.  reported an even distribution of cases throughout the age groups, as in our report. Male predominance has been a common feature in all the reported series on multiple brain abscesses [4–8], also as in our series.
The clinical features of our patients are similar to those of brain abscess in general and there is nothing specific to suggest multiple lesions. Headache is the most common symptom of our patients, which is similar to several reported series [3,5]. Although several series reported high incidence of altered consciousness [5,8,9], our patients seldom presented with altered consciousness. The presence of focal neurologic findings (e.g., hemiparesis, hemisensory deficits, and aphasia) depends on the location of the abscess, and are seen in approximately 25% (3/12) of our cases. The most common source of infection in our patients is hematogenous spread from a remote focus, which is similar to several reports [5,7]. Parameningeal spread is a rare cause in our patients, which is the opposite of another report .
The causative pathogens responsible for the infection vary with time period, geographic distribution, age, underlying medical and/or surgical condition, and mode of infection. In this study, brain abscesses due to gram-negative bacilli were not rare, and accounted for about 33% (4/12) of cases. It has been reported in Taiwanese studies that Klebsiella septicemia commonly causes metastatic septic infection such as arthritis, pericarditis, pyomyositis, brain and liver abscesses, and endophthalmitis, particularly in diabetic patients [12–14]. Streptococcus species are the most common isolated organisms in our series, which is similar to two reports [4,5]. But Sharma et al.  reported Staphylococcus aureus as the most common organism. Viridans streptococci are normal inhabits of the oral cavity, gastrointestinal tract, female genital tract, and upper respiratory tract, and they are often considered to be contaminants when isolated from blood culture. However, these pathogens have been known to cause infectious endocarditis [15,16], especially in individuals with preexisting valvular heart disease, as is found in Patients 11 and 12. Sterile abscess is noted in three cases of our report, which is similar to other reports [5,8,9]. This may be due to the consequence of previous antibiotic treatment.
A common problem is the difficulty in differentiating multiple pyogenic brain abscesses from metastatic tumor lesions , and though it is not found in our study, pathological confirmation may be required for definitive therapy if uncertainty remains. The predominant supratentorial location of our series is in accordance with other published reports [4,8]. More than half of our patients had 2 abscesses, whereas earlier reports had only one-third of patients with 2 abscesses [4,18]. Generally, clinical improvement precedes radiological improvement. In three patients treated medically, the duration of definite decrease in size and final resolution of lesions in CT scan is about 2–4 and 10–12 weeks, respectively, which is similar to another report . This may suggest that biweekly follow-up of CT scanning for at least 3 months may be necessary to document the therapeutic response and complete resolution of abscesses in patients with multiple brain abscesses.
In our series, eight patients accept combined surgical and medical treatment, and four patients accept medical treatment only. This decision is made by the correlation of clinical symptoms and CT scan data. In most of patients who accept surgery, larger size (≥ 3 cm) and significant mass effect were noted in initial CT scan. Mamelak et al.  suggested that all abscesses larger than 2.5 cm or those causing significant mass effect should merit surgical intervention. Another report  suggested surgically treating any abscesses larger than 3 cm or causing significant mass effect. They selected 2.5 or 3 cm of abscess diameter as a conservative cutoff of multiple lesions on the basis of a previous report [19,20] that demonstrated that solitary abscesses smaller than 3.0 cm could be treated nonsurgically, whereas those larger than 3 cm were more likely to require surgery for cure. In three patients who were treated successfully with antibiotic treatment, the size of all abscesses was about 1–2 cm. This supports the suggestion of previous reports [5,7] that medical treatment can be tried initially with broad-spectrum antibiotics if the diameter of abscesses is less than 3 cm and there is no major mass effect in initial CT scan. However, it is not advised when antibiotic susceptibility of organisms is unknown, especially if there is no systemic disease from which cultures can be obtained. In this situation, aspiration of most superficial or larger lesions producing neurological deficits is advocated, while treating the remaining lesions with antibiotics [19,20]. Evacuation is the major surgical method in our patients. Stereotactic surgery has been demonstrated to be a very effective technique in the treatment of multiple brain abscesses , although this technique is not used in our patients. The duration of intravenous antibiotic treatment in most of our patients was 4–6 weeks. Two reports [5,7,21,22] suggested that intravenous antibiotics should be continued for at least 6 weeks. We draw the conclusions that the duration of intravenous antibiotics should depend on the causative pathogens, therapeutic response, and neuro radiographic findings.
The observed mortality of 8.3% in this report is comparable to several reports [5,7]. The only expired patient was admitted with deep coma and poor hemodynamic status. In those stable patients with large size of abscesses and significant mass effect, combined surgery and antibiotic treatment still resulted in good outcome. This suggests that the clinical status at admission, not the number and size of abscesses, is the major determiner of outcome. Our favorable outcome is similar to the rate reported for solitary bacterial abscesses [5,10,21–23]. We believe that this result is largely attributable to the accuracy with which CT scan can detect and localize multiple abscesses, coupling with an appreciation of the need for the prompt surgical drainage of any large (> 3 cm) abscesses and those causing significant mass effect.
Multiple brain abscesses accounted for 10.3% of our patients with brain abscesses. Our series demonstrate that the prognosis for patients with multiple brain abscesses can be as good as that for patients with solitary brain abscesses. Any abscesses larger than 3 cm or those causing significant mass effect should be surgically treated and then combined with 6 weeks of antibiotic treatment. If the diameter of abscesses is less than 3 cm and there is no major mass effect, antibiotic treatment can be tried initially. Regular follow-up with CT should be performed to evaluate the therapeutic response. If any abscess enlarges after 2 weeks of antibiotics or fails to resolve after 3 to 4 weeks of antibiotics, surgical intervention is necessary. However, the treatment of these patients should be judged on the clinical status, neuroradiographic findings, and therapeutic response. Prognosis is favorable with early diagnosis and prompt treatment.
1. Barsoum AH, Lewis HC, Cannillo KL. Nonoperative treatment of multiple brain abscesses. Surg Neurol 1981; 16:283–287.
2. Carey ME, Chou SN, French LA. Experience with brain abscesses. J Neurosurg 1972; 36:1–9.
3. Dyste GN, Hitchon PW, Menezes AH, et al. Stereotaxic surgery in the treatment of multiple brain abscesses. J Neurosurg 988;69:188–194.
4. Kratimenos G, Crockard HA. Multiple brain abscess: a review of fourteen cases. Br J Neurosurg 1991; 5:153–161.
5. Mamelak AN, Mampalam TJ, Obana WG, et al. Improved management of multiple brain abscesses: a combined surgical and medical approach. Neurosurgery 1995; 36:76–86.
6. Mampalam TJ, Rosenblum ML. Trends in the management of bacterial brain abscesses: a review of 102 cases over 17 years. Neurosurgery 1988; 23:451–458.
7. Rousseaux M, Lesoin F, Destee A, et al. Developments in the treatment and prognosis of multiple cerebral abscesses. Neurosurgery 1985; 16:304–308.
8. Sharma BS, Khosla VK, Kak VK, et al. Multiple pyogenic brain abscesses. Acta Neurochir (Wein) 1995; 133:36–43.
9. Boom WH, Tuazon CU. Successful treatment of multiple brain abscesses with antibiotics alone. Rev Infect Dis 1985; 7:189–199.
10. Rosenblum ML, Hoff JT, Norman D, et al. Nonoperative treatment of brain abscesses in selected high-risk patients. J Neurosurg 1980; 52:217–225.
11. Morgan H, Wood MW, Murphey F. Experience with 88 consecutive cases of brain abscess. J Neurosurg 1973; 38:698–704.
12. Liliang PC, Lin YC, Su TM, et al. Klebsiella
brain abscess in adults. Infection 2001; 29:81–86.
13. Tang LM, Chen ST, Hsu WC, et al. Klebsiella
meningitis in Taiwan: an overview. Epidemiol Infect 1997; 119:135–142.
14. Wang LS, Lee FY, Cheng DL, et al. Klebsiella pneumoniae
bacteremia: analysis of 100 episodes. J Formosan Med Assoc 1990; 89:756–763.
15. Enting RH, de Gans J, Blankevoort JP, et al. Meningitis due to viridans streptococci in adult. J Neurol 1997; 244:435–38.
16. Koorevaar CT, Scherpenzeel PG, Neijens HJ, et al. Childhood meningitis caused by enterococci and viridans streptococci. Infection 1992; 20:118–21.
17. Coulam CM, Seshul M, Donaldson J. Intracranial ring lesions: can we differentiate by computed tomography? Invest Radiol 1980; 15:103–12.
18. Basit A, Bhatia R, Banerji K, et al. Multiple pyogenic brain abscesses: an analysis of 21 patients. J Neurol Neurosurg Psychiatry 1989; 52:591–594.
19. Rosenblum ML, Hoff JT, Norman D, et al. Decreased mortality from brain abscesses since advent of computerized tomography. J Neurosurg 1978; 49:658–668.
20. Sharma BS, Gupta SK, Khosla VK. Current concepts in the management of pyogenic brain abscess. Neurol India 2000; 48:105–11.
21. Chun CH, Johnson JD, Hofstetter M, et al. Brain abscess: a study of 45 consecutive cases. Medicine (Baltimore) 1986; 65:415–431.
22. Mathisen GE, Johnson JP. Brain abscess. Clin Infect Dis 1997; 25:763–81.
23. Yang SY. Brain abscess: a review of 400 cases. J Neurosurg 1981; 55:794–799.