Indian Pediatrics 2002; 39:914-921
Diagnosis and Outcome of Acute Bacterial Meningitis in Early Childhood
Nandita Chinchankar, Meenakshi Mane*, Sheila Bhave, Swatee Bapat, Ashish Bavdekar, Anand Pandit, K.B. Niphadkar*, Anil Dutta**, Didier Leboulleux**
ACUTE bacterial meningitis (ABM) is an important disease of early childhood, with high case fatality and risk of neurologic handicaps(1). The community incidence of ABM in India is not known. The exact etiological diagnosis is often not possible, because of poor culture facilities(2,3). The three organisms commonly associated with ABM in early childhood in western countries are Hemophilus influenzae type b, S. Pneumoniae and Neisseria meningitidis. However, the etiology may vary in different parts of the world(1). Many of these infections are likely to be preventable in the near future(4,5).
We have prospectively examined the hospital based frequency of ABM in early childhood, especially in relation to its etiology. The clinical and diagnostic features, mortality, complications and especially long term sequelae were also analysed.
Subjects and Methods
The study was conducted in the Pediatric Department of KEM Hospital, Pune, which includes a 6 bedded Pediatric Intensive Care Unit (PICU) and a 40 bedded General Pediatric ward. All patients admitted with a clinical diagnosis of ABM during April 1997 to March 1999 were included if they satisfied the following criteria: (i) age between 1 month and 5 years; (ii) cerebrospinal fluid (CSF) showing protein>40 mg/dl, sugar <40 mg/dl and >10 neutrophils/hpf. Other investigations included complete blood counts, chest x-ray, renal and liver function tests, electrolytes and blood culture. C-reactive protein (CRP) was determined in the CSF and categorized as positive or negative. CSF antigen detection by latex agglutination (LAT) was done using Slidex Meningite kit 5 (Biomerieux, France) for the detection of soluble antigens of H. influenzae type b, S. pneumoniae and N. meningitidis groups A, B and C(6,7).
CSF was cultured on specific media: Trypticase soy agar, Chocolate agar with Polyvitex, Columbia sheep blood agar and Hemoline performance two phase aerobic (bottle). CSF cultures were further identified by: (i) specific antisera (DIFCO, France), specifically polyserum and type b serum for the H. influenzae, types A, B and C sera for meningococci, and S. pneumoniae; (ii) standard techniques for identification of E. coli, streptococcus, staphylococcus, etc. Antibiotics sensitivity testing was done by the Kirby-Bauger Method.
Ultrasound and CT scan of the head were carried out whenever required.
Following diagnosis, the patients were treated with antibiotics based on their age: (i) patients between 1 and 6 months’ old: cefotaxime with gentamicin or amikacin; (ii) >6 months’ old: ampicillin or penicillin with chloramphenicol, or cefotaxime or ceftriaxone. Intravenous dexamethasone was given in a dose of 0.15 mg/kg/dose every 6 hr for four days (the first dose given with or prior to the first dose of antibiotics). Antibiotics were changed (if required) as per the sensitivity results and clinical response. Appropriate supportive care including attention to fluids, electrolytes, ventilation, parenteral nutrition and neurosurgical intervention was provided.
Children who recovered were regularly followed up in the out-patients clinic and the Pediatric Rehabilitation Center. Home visits were arranged for patients who did not come for regular follow up. Evaluations on follow up included detailed neurological examination, audiometry, BERA and tests for intelligence and development quotients.
The study was approved by the Ethics Committee of the hospital and informed consent of the parents was obtained.
Of 3686 admissions, 54 children (1.5%) (age 1 month to 5 years) had ABM; 28 (54.9%) were males. Forty two patients were below the age of one year and 28 (51.9% of all cases) under the age of six months. Only 2 of the 54 children with ABM had been immunized with H. influenzae vaccine.
The presenting features included high fever (96%), altered sensorium (98%), refusal of feeds (83%) and convulsions (81%); two children had circulatory failure. Meningeal signs were elicited in 26% only. Twenty one patients (39%) had received antibiotics for 1-3 days prior to admission. The CSF was turbid in majority of cases (85%); CSF proteins ranged from 40-660 mg/dl, sugar from 8-72 mg/dl, and leukocytes from 40-10,000/cu mm. Majority of the leukocytes were neutrophils; 10 children (19%) showed 10 to 30% lymphocytes. CRP was positive in the CSF in 22 patients (41%), gram stain was postive in 36 cases (67%), but correlated with culture or positive LAT in 29 cases only (54%).
CSF cultures were positive in 27 (50%) and LAT was positive in 36 (67%). The etiology of ABM, on basis of CSF culture and LAT showed S. pneumoniae in 21 (39%) and H. influenzae type b in 14 (26%). Only one patient showed infection with N. meningitidis. The LAT was positive in all patients with positive CSF cultures for the above organisms.
Of the 18 patients (33%) who had a negative LAT, 8 were culture positive for other organisms (Pseudomonas, Staphylo-coccus and betahemolytic streptococci in 2 each and Streptococcus pyogenes and Citrobacter in 1 each). An etiologic diagnosis was not made in 10 (19%) patients.
Blood cultures were positive in 4 (H.influenzae, alpha-hemolytic strepto-coccus, Pseudomonas and S.aureus in one each). Positive blood culture, CSF culture and positive LAT was present only in one patient with H. influenzae infection. Patients with positive blood cultures for Pseudomonas and S. aureus had sterile CSF cultures.
The clinical and laboratory features of the three major groups of ABM are seen in Table I. There were no significant differences between the S. pneumoniae and H. influenzae groups at admission, except higher CSF cells in the latter (P = 0.01). Most cultures of S. pneumoniae and H. influenzae were sensitive to standard medications. Streptococci were resistant to cefotaxime, penicillin and ampicillin in 2 cases. Two patients with H. influenzae infection showed resistance to ampicillin and vancomycin.
During the course of treatment, antibiotics were changed as per protocol in 10 children (the added or changed drugs included vancomycin in 2, piperacillin and ceftazidime in 2 each, ceftriaxone in 4, tobramycin in 1 and chloramphenicol in 3).
Complications seen during hospitalization were subdural effusion (n=10), hemiplegia (n = 4), ventricular abscesses and hydrocephalus (n = 2), cerebral infarcts (n = 6), cranial nerve palsy (n = 3) and isolated hydrocephalus in one patient; 5 patients had more than one complication. The subdural effusions were treated conservatively. Both patients with ventricular abscesses required frequent tapping (one requiring craniotomy) and later developed hydrocephalus. Two of the 4 patients with hemiplegia showed no focal deficit at recovery.
Table I-Clinical and Laboratory Features of Acute Bacterial Meningitis
Figures in parenthesis indicate percentage. # N. meningitidis * Pseudomonas (2), Citrobacter (1), Staphylococcus (2), b hemolytic strep (2), Strep pyo (1).
Eighteen (41%) of 44 children with seizures continued to have fits beyond four days of admission; of these 6 patients died, while 10 showed persistent seizures. Of the 26 children who had seizures restricted to first four days of admission, 8 died, 16 improved (with no convulsions) and 2 had persistent seizures.
Of the 54 patients with ABM, 10 (19%) died in hospital within 1-7 days of admission (6 died within 48 hr of admission) and 5 were discharged against medical advice in a critical condition (3 of these died at home soon after discharge while two were lost to follow up). The others were discharged after 6 to 45 days (mean 12 days). Four patients died at home, 2 each due to complications and 2 of unrelated causes. Six patients could not be traced for further follow up.
31 patients therefore, were available for long term follow up. Of these, 14 (45%) had a good outcome with no neurological sequelae. In others, sequelae ranged from isolated hearing loss to severe developmental delay with a combination of motor, hearing and visual defects (Table II). Twelve children had persistent seizures. On analysis of immediate complications with long term outcome of the 10 patients with subdural effusion, 5 had no sequelae, one had isolated hearing loss, 3 had moderate to severe developmental delay and 1 died. Two of the 6 children with cerebral infarcts died, one had isolated hearing loss and 3 had moderate to severe developmental delay with multiple disabilities. Of the 2 children with cerebral abscess one died and the other was left with multiple handicaps and hydrocephalus. Patients with pneumococcal meningitis had higher mortality, risk of complication and neurological sequelae compared to H. influenzae infection (P = 0.04, Fischer exact test).
ABM accounted for 1.5% of all our pediatric admissions during the study period. We restricted our study to the age group of 1 month to 5 years to focus on the distinct clinical group excluding neonatal meningitis and meningitis in older children. Kabra, et al. reviewing a survey of pyogenic meningitis in major centers in India reported a frequency of 0.5 to 2.6% of hospital admissions(2). The community prevalence in international studies have been quoted between 3/100,000 in USA(9), 16/100,000 in UK(10) to 45.8/100,000 in Brazil(9). Though all these numbers are relatively small, the importance of ABM is chiefly because of the associated high mortality and serious morbidity.
In our study, of a total of 54 cases, 17 (31.5%) died in hospital or soon after discharge. The case fatality rate in India and other developing countries has been quoted as 16 to 30%(1,3,11-13). Approximately one-third of all deaths in our study occurred in the first 48 hr of hospitalization, reflecting the critical condition of the patients at admission. Even in developed countries, inspite of availability of all facilities, the case fatality rates of bacterial meningitis in early childhood approaches 10%(9).
The complication rate in ABM too, is high inspite of aggressive management(3,9,14). Atleast 40% of our patients had acute complications, including subdural effusions, cerebral infarcts, ventricular abscesses, hydrocephalus, cranial nerve palsies, repeated seizures and motor deficits. Even more disturbing than the high mortality (and complications) were the long term sequelae and disabling handicaps in the survivors. On follow up, only one-fourth of the children tested were neurodevelopmentally normal. A significant proportion were left with hearing deficits and moderate to severe mental retardation with multiple disabilities causing considerable financial and emotional burden to the family. Early and careful follow up with neurodevelopmental and auditory testing is important because more than half of these disabled children appeared normal at hospital discharge. Similar long term sequelae of ABM have been described from both developing(14-16) and developed countries(17-18), though such data from India is lacking.
Early diagnosis, prompt initiation of therapy and supportive care are important for improving the long term outcome(2,9,11). Unfortunately, as seen in the present study, signs of meningitis cannot be used for making an early diagnosis. Altered sensorium and convulsions are late features. Hence a high index of suspicion is necessary to suspect meningitis and perform a lumbar puncture. A presumptive diagnosis of ABM is usually possible on the basis of biochemical analysis of CSF. Difficulties arise if patients are already treated with antibiotics as CSF may show normal sugar content and cells may be predominantly lymphocytes. In such cases, one has to rely on other CSF parameters and clinical clues. Unfortunately what seems to be difficult especially in our country is accurate bacteriological diagnosis of meningitis(2,3,19). Gram-staining, though a cheap and easy technique, can identify organisms at best in 60% of cases(3,11,20), whereas CSF culture results are positive in not more than 15-35% patients(2,3,19). The culture results can perhaps be improved with the use of special media and special techniques especially for H. influenzae as in the present study. However, most western series quote a culture positivity of upto 90%(9). The reasons for low yields of cultures in our country are not clear but may include poor quality of culture media and use of antibiotics prior to hospitalization(3,6). Other quick and sensitive techniques of diagnosis of ABM obviously need to be urgently explored. Of great promise is the CSF latex agglutination test (LAT) which various authors have confirmed as simple with superior sensitivity and specificity and unaffected by previous antibiotics therapy(6,7,11). In the present study, more than two-third patients were positive for LAT whereas cultures were positive in only one-half. However the LAT kits are expensive and available only for common organisms and hence not suitable as the lone diagnostic technique in ABM. Besides, the kits cannot provide information on antibiotics sensitivity and hence both techniques (LAT and culture sensitivity) should be used together. In our series, 10 patients were negative on culture and LAT, but the final diagnosis was made on CSF biochemistry and clinical features. In such cases, it is difficult to rule out tuberculous and viral meningitis in the initial stages.
Until recently the commonest organism associated with ABM in early childhood was H. influenzae(9). However, most Indian studies have quoted a low isolates of the organism(2,3,19). Whether this is because the organism is difficult to grow or whether the incidence is genuinely low is not clear. In our earlier studies, we have shown a high susceptibility of infants to infection with H. influenzae(4). In the present study, the contribution by various organisms was H. influenzae type b in 26%, S. pneumoniae in 39% and N. meningitidis in 2%. The low incidence of infection with N. meningitidis and relatively high incidence of pneumococcal infection has been noted by other Indian workers(3,17).
Table II-Outcome in Acute Bacterial Meningitis
Figures in parenthesis indicate percentage. * More than 1 complication in 5 children
The management of ABM includes a suitable combination of antibiotics, dexamethasone for first few days, and importantly, intensive care therapy especially for shock and raised intracranial pressure(2,5,9). Treatment of complications includes antiepileptic drugs and neurosurgical procedures. Rehabilitation programmes are necessary for the handicapped. All these add to the tremendous financial and emotional burden on the family. Measures to prevent ABM are thus extemely important.
The vaccine against H. influenzae has reduced infection with one of the most important causes of ABM in advanced countries. This is now being introduced as an optional vaccine in India. However, to be effective against H. influenzae meningitis it should be given early in infancy as majority of cases of ABM occur in the first six months of life. Theoretically, pneumococcal meningitis too is a vaccine preventable disease, but unfortunately, the currently available polysaccharide vaccine is not effective in children under two years of age(5,9). Universal use of meningococcal vaccine is unattractive as the disease is generally sporadic and vaccine is of limited immunogenicity(9). However if H. influenzae and pneumococcal vaccines become a reality, the incidence of ABM in early childhood can reduce by more than 60%, with considerable reduction in the financial and emotional cost burden of the disease.
Special thanks are due to Dr. Bhavana Doshi and Ms. Manjusha Rajarshi, Consultants, Aventis Pasteur, Mumbai for their help in the study.
Contributors: NC and SwB were involved in clinical data collection. MM and KBN carried out the bacteriological investigations. ShB co-ordinated the study, drafted the paper and will act as guarantor. AB co-ordinated the study and analysed the data. AP supervised the study and reviewed the manuscript. DL prepared the study design and AD monitored data collection.
Funding: Aventis Pasteur, France.
Competing interests: None stated.