A 4-year Study of Neonatal Meningitis: Clinical and Microbiological Findings

M. ADHIKARI ET AI_ A 4-year Study of Neonatal Meningitis: Clinical and Microbiological Findings by M. Adhikari, MB, ChB, FCP (Paed) (SA), MD, Y. M. Coovadia, MB, ChB, FF, Path (Micro), and D. Singh Department of Paediatrics & Child Health, Faculty of Medicine, University of Natal, PO Box 39, Congella, 43, Durban, South Africa Summary The clinical and microbiological data of 6 neonates, 3 from the Neonatal Unit (Group I) and 3 (Group II) from the General Paediatric Wards with meningitis are presented. The overall prevalence/ was significantly lower in Group I (36) than in Group II (.; /*<.). This low incidence follows the introduction of amikacin for the treatment of the ill neonate in 96. Streptococcus agalactiae (3 per cent), KlebsieUa pneumoniae ( per cent) and E. coli ( per cent) were the commonest pathogens accounting for per cent of the cases. Streptococcus agalactiae isolates were uniformly susceptible to penicillin and chloramphenicol. Gram negative isolates showed resistance to ampicillin, chlorampbenicol and sulphamethoxazole-trimethoprim. In addition K. pneumoniae isolates showed resistance to gentamycin and amikacin. All isolates were fully susceptible to cefotaxime. Recently, four of six cases of A', pneumoniae in the Neonatal Unit were resistant to amikacin. Low birth weight, additional clinical problems, and ultrasound changes on cranial scanning carried a poor prognosis. Emphasis should be placed on close collaboration between clinicians and microbiologists in the choice of antimicrobial agents and aseptic techniques for the care of neonates. Introduction Neonatal meningitis, an uncommon illness presents an ongoing challenge with regard to appropriate therapy and attempts at reduction of morbidity and mortality. Relatively few studies from either the African continent or other developing countries have addressed this problem specifically.'" In the early 9s gentamycin-resistant KlebsieUa was shown to be the major pathogen isolated in neonatal meningitis at King Edward Hospital. Following the introduction of cefotaxime as the antibiotic of choice in Gramnegative meningitis and amikacin as empirical therapy in the ill neonate a reduction in the mortality and incidence respectively occurred.* The purpose of this study was to review our data on neonatal bacterial meningitis for the 4-year period 9-99. features, outcome, antimicrobial therapy, and susceptibility patterns of isolates were analysed. Early onset disease was defined as presentation within the first days of life and late onset disease as that presenting thereafter. Hospital acquired infection was regarded as infection presenting after 3 days postnatal age (nursery cases only). Bacteriology Bacterial isolates were identified according to their cultural and biochemical characteristics, and by the API E System (API, France). 3 Antimicrobial susceptibility to penicillin, ampicillin, chloramphenicol, sulphamethoxazole-trimethoprim [SMZ-TMP], gentamicin, amikacin, and cefotaxime was determined by the Stokes Agar Disc Diffusion Method. 6 Streptococcus agalactiae antigen was detected in CSF samples by the Wellcogen Kit (Wellcome Diagnostics, England). Downloaded from http://tropej.oxfordjournals.org/ at Pennsylvania State University on April 9, 6 Methods Patients The clinical and laboratory records of all neonates from the Neonatal Nursery and General Paediatric Wards at King Edward VIII Hospital, Durban, who had either positive cerebrospinal'fluid (CSF) bacterial cultures or positive, Streptococcus agalactiae CSF antigen tests, during the period January 9 to December 99 were reviewed. Information on age at the time of onset of meningitis, birth weight, clinical Statistical analysis This was performed using the x test. Significance was taken at the per cent level. Therapy Routine supportive treatment was given to all babies. A combination of penicillin and cefotaxime or amikacin were prescribed once a diagnosis of meningitis was made. This followed from microbiological data as reported in the previous study. 4 Journal of Tropical Pediatrics Vol.4 April 99 Oxford University Press 99

M. ADHIKARI ET AL. Results Sixty neonates had culture proven (n = ) or S. agalactiae antigen positive (n= ) meningitis. Of these, 3 were from the Neonatal Unit (Group I) and 3 from the General Paediatric Wards (Group II). However, clinical data could only be obtained on 44 neonates. Eight ( per cent) of the babies weighed less than <. kg, (34 per cent) between. and. kg, and (4 per cent) >. kg. The incidence rates and case mortality rates (CM R) of meningitis in Groups I and II are indicated in Table. Prevalence (Table ) The overall prevalence/ was significantly lower in Group I (,36) than in Group II (,) (/><.). The rate in the Newborn Nursery fell dramatically from (,9) (9-96) to (,) (9) and in the present study (9-99) it was (,36) (/ <.). This follows the introduction of amikacin for the ill neonates in 96. Case mortality rates (CMR) (Table ) The overall CMR for the period of the study was (,4) (/) in Group I and (,44) (/3) in Group II. The CMR for the different isolates were:. agalactiae (Group B streptococcus).4 (/), Klebsiella pneumoniae.4 (/), E. coli.9 (/), S. pneumoniae.33 (/3) and other Gram-negative bacteria. (4/). CMR for early onset disease was.3 (4/3) and that for late onset disease. (6/) [p NS]/The CMR of. (/) in neonates weighing less than. kg was significantly higher than that of.4 (6/) and.9 (6/) in those weighing.-. kg (P<.) and >. kg, respectively (P<.6). Aetiological agents (Table ) The aetiological agents in groups I and II respectively were:. agalactiae (44 per cent) and (3 per cent), K. pneumoniae (3 per cent) and ( per cent), and E. coli 4 ( per cent) and 6 (6 per cent). These three pathogens accounted for per cent (4/6) of all cases. Not a single case of L. monocytogenes or S. aureus meningitis was encountered during the study period. Pseudomonas aeruginosa, Salmonella sp. Proteus mirabilis, Enterobacter cloacae and Acinetobacter anitratus were detected in only one or two cases each year. Gram staining detected per cent (43/) of all culture positive CSF samples with TABLE Annual prevalence and case mortality rates (CMR) of neonatal meningitis (9 99) in the newborn nursery (Group I) and general paediatric wards (Group II) Year 9-96 9 9 99 99 Overall Group I Prevalence.9..4.6 44.36 CMR.9 6..33.4..44 Group II Prevalence _..3.4 CMR -.43.. - TABLE Aetiological agents of bacterial meningitis in neonates (9-99) at King Edward VIII Hospital, Durban Organism 9 99 99 99 Total % Downloaded from http://tropej.oxfordjournals.org/ at Pennsylvania State University on April 9, 6 Streptococcus agalactiae Klebsiella pneumoniae Escherichia coli Streptococcus pneumoniae Pseudomonas aeruginosa Salmonella spp. Proteus mirabilis Enterobacter cloacae Acinetobacler anitratus Total 4 4 3 3 3»(3) () () 4() (3) (3) (3) (3) (3) 6() 'Includes three cases in Group and which were culture negative, but positive on Ag. testing. Journal of Tropical Pediatrics Vol.4 April 99

M.ADHIKARETAL. sensitivities varying from per cent (4/4) for Streptococcus pneumoniae, per cent (4/6) S. agalactiae, per cent (3/) K. pneumoniae, 6 per cent (6/) E. coli, and other Gram-negative bacteria per cent (6/). Antimicrobial susceptibilities The S. agalactiae and. pneumoniae isolates were uniformly susceptible, to penicillin, ampicillin and chloramphenicol. The K. pneumoniae isolates, in contrast were uniformly resistant to ampicillin, 9 per cent to chloramphenicol, 4 per cent SMZ-TMP, 4 per cent to gentamicin, and 3 per cent to amikacin. Eighty per cent of E. coli were also resistant to ampicillin, 4 per cent to chloramphenicol, 6 per cent to SMZ-TMP, but were fully susceptible to gentamycin and amikacin. All the Gram-positive and Gramnegative isolates were, however, fully susceptible to cefotaxime. All isolates with the exception of P. aeruginosa and E. cloacae were also fully sensitive to cefuroxime. Comparison of the antimicrobial susceptibility pattern of isolates between groups I and II, respectively, revealed the major difference was higher resistance in Group I to SMZ-TM P ( v. 3 per cent), gentamycin ( v. 3 per cent) and amikacin ( v. per cent) amongst K. pneumoniae isolates. The E. coli isolates in group II, however, were significantly more resistant to chloramphenicol (6 v. per cent) (P values were not significant for any of the above because of the small numbers compared in each group). The changing pattern of pathogens causing meningitis in the Newborn Nursery for the period 9-9 has been previously reported. 4 For the first m olate to >. rato o n a o 6 «i log- s' - 6- - 4-3- - - \ A. \ / \ \ / A\ years following the introduction of amikacin in 96 for the empirical therapy of suspected sepsis in the ill neonate, only a single case of AT. pneumoniae meningitis was encountered. However, in 99 there were four cases of K. pneumoniae meningitis, but none in 99. With the exception of 9, when no cases of S. agalactiae meningitis were encountered, this organism has emerged as the leading cause of meningitis in the newborn nursery. The number of cases due to E. coli has remained at approximately one case/year since the introduction of amikacin. Furthermore, besides two cases of P. aeruginosa meningitis, cases due to other Gram-negative bacteria have not been encountered in the nursery since 96. Age at presentation Thirty-six (6 per cent) of neonates presented at days of age, with S. agalactiae accounting for 3 (36 per cent) of those cases (Table 3). Klebsiella pneumoniae and E. coli were distributed more or less equally between early and late onset cases. However, using 3 days as the dividing line between maternal and hospital acquired infection, the vast majority of K. pneumoniae ( per cent), S. agalactiae ( per cent), and E. coli ( per cent) cases of meningitis presented after day 3. Presentation and clinical course All the babies presented with unstable temperatures or hypothermia. Apnoeic or cyanotic attacks occurred in per cent (/3) v. 6 per cent (/), convulsions 4 per cent (/3) v. 44 per cent (/), abdominal distention 3 per cent (4/3) v. per cent (/), ^ \ VAX \' V / * 'i /A\ / / A\ w \\ \\ A > V i(v A/ --A / \ A/ /\\ A / h Klebsiella ~ S. agalactiae E.cdl Other Gram. neg. bacilli 9 9 93 94 9 96 9 9 99 99 99 t t Cefottxtma Amikacin ICU Years FIG.. Distribution of CSF bacterial pathogens in the neonatal nursery (group I) for the period 9-99. Journal of Tropical Pediatrics Vol.4 April 99 3 Downloaded from http://tropej.oxfordjournals.org/ at Pennsylvania State University on April 9, 6

M. ADHIKARI ET AL TABLE 3 Bacterial pathogens in relation ot age at the time of onset of meningitis in all neonates Bacteria S3 days ^ 3 days <, days ^ days Streptococcus agalacliae Klebsiella pneumoniae Escherichia coli Streptococcus pneumoniae Pseudomonas aeruginosa Salmonella spp. Proteus mirabilis Enterobacter cloacae Acinetobacler anitratus Total hypotonia percent (/3) v. per cent (6/) in the early and late onset cases, respectively. Hydrocephalus occurred in 44 percent (/) of the late compared to per cent (/3) of those presenting early. Clinical improvement occurred within 3 days with a mean of.6 days. Improvement was assessed on temperature response, improved neurological status and general well being of the patients. The disease ran a fulminating course in nine patients, eight of whom died within 4 hours of commencing appropriate antimicrobial therapy. Death occurred in patients.-6 days after diagnosis with a mean of 6.4 days. Associated conditions Sixteen ( per cent) of the babies who died had the following associated problems: septicaemia (6), pneumonia (6, two of these had lumbar meningomyelocoeles), jejunal atresia (), severe bilateral purulent conjunctivitis (), and pulmonary haemorrhage (). Two babies had clinically significant patent ductus arteriosus. In contrast, only 9 (39 per cent) of 3 who survived had associated problems [pneumonia (), septicaemia (), conjunctivitis (), and severe oral moniliasis (3)]. Complications associated with meningitis Four babies developed hydrocephalus, K. pneumoniae was the aetiological agent in two of these, and P. mirabilis and P. aeruginosa in one each. A further baby with P. mirabilis meningitis developed an abscess in the right parietal region which on resolving formed a communicating porencephalic cyst. None of the babies developed subdural fluid collections. Cranial ultrasonography Data concerning cranial ultrasound findings was available on of the 3 Nursery patients. Eighteen of the were scanned and eight had normal scans. Three other babies died before the scans were performed. Both babies with lumbar meningomyelocoeles (one 6 4 9 4 each with Klebsiella and Pseudomonas meningitis) had mildly dilated lateral ventricles with normal third and fourth ventricles. Three babies had grade I to II intraventricular haemorrhages associated with. agalactiae and P. aeruginosa all three died. Five babies with K. pneumoniae or E. coli meningitis also had dilated ventricles. In two, this was transient and both survived; the remaining three babies had ventriculitis, one of whom had cerebral oedema. Antimicrobial therapy One baby with jejunal atresia, did not receive any antibiotics and died within 3 hours after birth. Fortythree other babies received a combination of two antibiotics [penicillin and cefotaxime (36), penicillin and amikacin (6), and ceftazidime and vancomycin ()]. Of the seven who did not receive cefotaxime, four had S. agalactiae and three E. coli meningitis. 3 3 36 Discussion The findings of this study demonstrate the importance of close liaison between clinicians and microbiologists, particularly with respect to choosing the most appropriate antimicrobial agent for the empirical therapy of the ill neonate. As reported previously, 4 the selection of cefotaxime, a safe and effective antimicrobial agent for Gram-negative bacillary meningitis, in 94 led to a significant reduction in mortality from this disease. Following this initial decline,* the CMR has remained unchanged in the present study. This has occurred in spite of the emergence of. agalactiae as the leading cause of neonatal meningitis and may reflect on the possible virulence of this organism. Studies reported in the mid-eighties indicated that there was little difference in the outcome between babies suffering Gram-negative bacillary meningitis 9 or. agalactiae meningitis. A continuing decline in number of cases is shown in this study and in the developing world. The incidence of neonatal meningitis is quoted as declining from : to :3 in 93 to :3 to :6 in Downloaded from http://tropej.oxfordjournals.org/ at Pennsylvania State University on April 9, 6 4 Journal of Tropical Pediatrics Vol.4 April 99

M. ADHIKARI ET AL. 99." The decline in the incidence of nosocomially acquired Gram-negative organisms (gentamycinresistant K. pneumoniae) follows the introduction of amikacin and may, in addition, be attributed to a very strict regimen of hand washing and hand spraying with per cent alcohol. However, of concern is the emergence in recent years of amikacin-resistant K. pneumoniae as a cause of meningitis in our nursery after a -year period of exclusive amikacin usage. Our experience is therefore contrary to that reported from institutions in the more developed countries, where resistance to amikacin did not increase during a period of exclusive amikacin usage. Streptococcus agalactioe together with K. pneumoniae and E. coli accounted for per cent of all cases of neonatal meningitis both in the nursery and in the general pediatric wards. This is in marked contrast to reports from other developing countries where Salmonella or S. aureus are the leading causes of the disease. - 3 It is of interest that in the present study not a single case of L. monocytogenes, which is rare in our situation, 3 or S. aureus* was encountered, and Salmonella sp. was isolated from only a single patient. The. agalactiae isolates were uniformly susceptible to penicillin, ampicillin, and chloramphenicol, whereas amongst the Gram-negative bacilli, resistance to these antibiotics as well as to SMZ-TMP was high. Klebsiella pneumoniae isolates, in addition, also showed increased resistance to gentamycin and amikacin. Although cefotaxime has been used in our unit for the past years, resistance amongst CSF isolates as yet has not emerged. Little information on neonatal meningitis is available from developing countries " 3 and it is therefore difficult to compare the findings in this study with a comparable situation elsewhere. Little is also written on the clinical course and the expected rate of clinical improvement. However, the clinical presentation of early and late onset disease was typical, the earlier presenting with generalized and respiratory signs and the late with dominant neurological signs. 4 Clinical recovery in the survivors occurred within 3- days and the time of death in the non-fulminating cases (after 4 hours) was of a similar period. Of significance was the number of associated clinical problems in those that died compared to the surviving babies. The high mortality of the low birth weight infant with meningitis has been documented 3 and is not surprising. Five babies developed serious intracranial complications, all of whom had Gram-negative bacillary meningitis. Despite 9 per cent receiving appropriate and adequate antibiotic therapy the mortality remained significant, and a number died rapidly. Of interest were the cranial ultrasound associations of intraventricular haemorrhage, ventriculitis a known feature of neonatal meningitis 6 and cerebral oedema with poor prognosis. In summary, although the incidence of neonatal meningitis has decreased the mortality remains unchanged over the past 4 years. Low birth weight, accompanying clinical conditions and cranial ultrasound changes carried a poor prognosis. Furthermore, we believe that close collaboration between clinicians and microbiologists, careful attention to aseptic procedure and meticulous hand care practices played a critical role in reducing the incidence of bacterial disease in our unit. References. Mcdonald R. Purulent meningitis in newborn babies: Observations and comments based on a series of patients. Clin Paediat 9; : 4-4.. Barclay B. High frequency of Salmonella species as a cause of neonatal meningitis in Ibadan Nigeria. Acta Scan Paediat 9; 6: 4-4. 3. Longe AC, Omene JA, Okolo AA. Neonatal meningitis in Nigerian infants. Acta Scan Paediat 94; 3: 4-. 4. Coovadia Y, M Mayosi B, Adhikari M, Solwa Z, van den Ende J. Hospital acquired neonatal bacterial meningitis: The impact of cefofaxime usage on mortality and of amikacin usage on incidence. Ann Trop Paediat 99; 9: 33-9.. Cowan ST. Cowan and Steel's manual for the identification of medical bacteria, nd edn. Cambridge: Cambridge University Press, 94: 3. 6. Stokes EJ, Waterworth PM. Antibiotic sensitivity tests by diffusion methods. Association of Clinical Pathologists Broadsheet: No., revised December 9.. Naqvi SH, Maxwell MA, Dunkle LM. Cefotaxime therapy of neonatal gram-negative bacillary meningitis. Paediat Infect Dis 9; 4: 499-.. Edwards MS, Rench MA, Haffar AAM, Murphy MA, Desmond MM, Baker CJ. Long term sequelae of Group B streptococcal meningitis in infancy. J Paediat 9; 6: -. 9. McCkracken GH. New Developments in the management of children with bacterial meningitis. Paediat Infect Dis 94; 3 (Suppl): 3-.. Klein JO, Marcy SM. Bacterial sepsis and meningitis. In: Remington JS, Klein JO (Eds). Infectious diseases of the fetus and newborn, nd edn. Philadelphia: WB Saunders, 93: 69-3.. de Louvois J, Blackbourn J, Hurley R, Harvey J. Infantile meningitis in England and Wales: A two year study. Arch Dis Childh 99; 66: 63-.. Gerding DN, Larson TA. AminoglycoskJe resistance in Gram negative bacilli during increased amikacin use. Am J Med 9; 9: -. 3. Lalloo UG, Coovadia YM, Adhikari M, Dulton MF. Listeria monocytogenes meningitis at King Edward VIII Hospital, Durban. A year experience 9-99. S Afr Med J 99; : -9. 4. Volpe JJ. Bacterial and fungal intracranial infections. In: Volpe JJ (Ed.) Neurology of the newborn, nd edn. Philadelphia: WB Saunders, 9: 96-63.. Mulder CJJ, Zanan HC. A study of cases of neonatal meningitis in the Netherlands. J Infect 94; 9: -4. 6. Salmon JH. Ventriculitis complicating meningitis. Am J Dis Childh 9; 4: 3-4. Downloaded from http://tropej.oxfordjournals.org/ at Pennsylvania State University on April 9, 6 Journal of Tropical Pediatrics Vol.4 April 99