Indian Pediatrics 2003; 40:870-873
Unrecognised Ventriculitis/Meningitis Presenting as Hydrocephalus in Infancy
Vrajesh Udani, Soonu Udani, Rohan Merani and Manisha Bavdekar
From the Department of Pediatrics, P.D. Hinduja Hospital & Medical Research Center, Veer Savarkar Marg, Mumbai 400 016. India.
Hydrocephalus in infancy is either due to prenatal etiologies like aqueductal stenosis, and meningomyelocele or due to peri-postnatal etiologies like infection and hemorrhage(1-3) Epidemiological studies in the West indicate that approximately two-thirds of infantile hydrocephalus is prenatal(1).
Hydrocephalus usually presents with a progressive macrocrania, irritability and vomiting. In post-meningitic hydrocephalus there is usually an apparent history of symptomatic neonatal meningitis and the hydrocephalus is not unexpected. Over the last few years we have identified several infants with progressive macrocrania in early infancy, without a past history of overt neonatal or early infantile meningitis and who on subsequent investigation had confirmed active meningitis/ventriculitis (MV). We believe that these cases were due to unrecognized meningitis in the neonatal period where clinical symptoms were masked by the use of inadequate antibiotics.
Subjects and Methods
The relevant medical records between 1991 and 1998 were reviewed by searching under the ICD coding of hydrocephalus for infants. Etiology was determined using clinical, imaging and other relevant data. MV was diagnosed when lumbar and ventricular CSF showed pleocytosis with or without positive cultures. The study group includes only those patients with hydrocephalus where the diagnosis of MV was not clinically suspected and where there was no obvious history of prior meningitis. This group was then analyzed for presence or absence of risk factors for meningitis(4) i.e., pre-maturity, low-birth weight, multiple births, signs of sepsis in the neonatal period, prolonged neonatal hospitalization, and empirical antibiotic use in the neonatal period. We also analyzed the clinical details of the current illness, the CSF biochemistry and culture, duration of antibiotic use and the use of neurosurgical procedures such as external ventricular drainage (EVD) and VP shunt. Outcome was analyzed after follow-up of at least 6 months. Patients were classified as normal/near normal, mildly disabled and severely disabled. The evaluation was based on both routine neurological examination and the Denver Developmental Scale.
Analysis of records revealed that 39 out of 72 cases of hydrocephalus (54.2%) were the result of meningitis. 13/72 infants had confirmed prior neonatal meningitis and a further 13/72 had tuberculous meningitis. In 13/72, the patients had no obvious clinical MV and were diagnosed only by CSF examination and culture results. These 13 are reported here. The average age of the patients in the study group was 3 months with a range from 2 to 6 months of age. There were 12 males and 1 female.
All the patients in the study group had risk factors for meningitis(4) in the neonatal period. 8/13 were premature or low birth-weight, 3/13 were one of twins or triplets, 8/13 had clinical evidence of sepsis in the neonatal period. At least 5/13 had a history of prolonged hospital stay ranging from 11 days to 1.5 months. 1/13 had an invasive procedure (exchange transfusion). 10/13 had been given a course of antibiotics course in the neonatal period. In 6 of these 10, where details were available the duration of antibiotics was less than 2 weeks. Only one CSF study was performed in the neonatal period, and was abnormal (‘60 cells’).
Between the time of discharge from the newborn nursery to the time of presentation to us, 6/13 were perceived as "sick", and 6/13 were thought to be "well" by parents and the doctors following the children. 10/13 patients presented in early infancy with increasing head circumference (HC). All patients had macrocrania at presentation. Fever was present in only 5/13, seizures in 4/13, and vomiting in 3/13. Only 1/13 blood culture was positive, while all lumbar and ventricular CSF results were abnormal. 10/13 CSF cultures were positive and the samples grew a variety of organisms including Candida albicans, Staphylococcus aureus, Pseudomonas, Acinetobacter, Enterococcus, Steptococcus pneumoniae and Moxarella catarrhalis. Contrast CT scans revealed hydrocephalus in all 13 infants, 9 of whom were thought to be due to acqueductal obstruction. Ventriculitis was obvious in 2/13 and abscesses were seen in 1 case. Average total length of antibiotics given was 32.8 days. 4/13 patients required EVD and all were shunted between 1-5 weeks after admission. At follow-up, 7 patients were severely disabled, 2 mildly disabled, 2 were normal/near normal, and 2 were lost to follow-up.
The aim of this study was to highlight the fairly common occurrence of infantile hydrocephalus due to clinically unsuspected meningitis. This is probably a unique problem in our country because of non-standardized practices of neonatal care. In this 8 year study period 54.2% (39/72) of hydrocephalus were found to be due to infection, far higher than what is reported from developed nations. An ongoing Swedish study from 1967 to 1994, found a prenatal etiology to be the most common cause in full term infants(l) and IVH to be the most common cause in pre-term infants(2,3). Infection was seen in only about 5% of term(l) and 7% of pre-term infants(3). Infection seems to be a more common etiology in India not unexpectedly. In 3/13 study patients CSF culture was negative despite a CSF picture typical of MV. Two of these infants were premature, raising the possibility of a persistent chemical (from IVH) rather than infective MV. In our study, of the 26/72 patients with neonatal meningitis, only 13 were diagnosed in the neonatal period. The other 13, despite risk factors for meningitis, were empirically treated for "sepsis" with an inadequate course of antibiotics. Only one child had a neonatal CSF examination. As a result, probably no newborn was treated for the recommended period of 3 weeks(4). This possibly resulted in a partially treated MV manifesting later as hydrocephalus. Partially treated meningitis often results in communi-cating hydrocephalus due to a block at the level of the convexity subarachnoid space. Ventriculitis, however results typically in inflammatory acqueductal obstruction(5,6) resulting in obstructive tri-ventricular hydrocephalus that can be easily confused on imaging with congenital acqueductal stenosis. To confound matters further 8/13 infants in our series did not have the usual symptoms of infection and could have been thus misdiagnosed as being due to prenatal acqueductal stenosis. However in our patients the ventricular CSF was clearly abnormal and grew microorganisms in the majority confirming the etiology of this obstructive hydrocephalus. If the CSF had not been sampled it could have led to a shunting procedure in infected CSF (as was the case in our initial patients) with disastrous results.
There is controversy in the literature about the need for a CSF examination in newborns with clinical or microbiologically proven sepsis(7-10). Though 8/13 in our series had clinical evidence of sepsis, 5/13 were asymptomatic and were treated with routine antibiotics only because they were ‘high-risk’. This suggests that CSF examination should be done in any high-risk newborn given antibiotics regardless of symptomatology, as ventriculitis maybe completely masked(10).
Even with early recognition and prompt treatment, there are long-term consequences of neonatal bacterial meningitis(2). Fernell and Hagberg(3) reported a 42% rate of CP in the survivors in 1967 a rate that improved to 22% by 1994. We believe that the poor outcome in 50% of our study group was the result of delayed diagnosis and inadequate treatment. It seems likely that had the diagnosis of meningitis been made in the neonatal period with a CSF examination and had adequate therapy been given then, long-term morbidity could have been reduced to a minimum.
Finally, it may be a good practice to sample the CSF in every infant with hydrocephalus before a shunting procedure is performed especially in those who had risk factors for meningitis in the neonatal period. This might prevent shunting infected ventricles with potential disastrous consequences.
Contributors: VU conceptualized the study, provided the cases and checked clinical, radiological and follow up data. SU provided the cases, checked references and data and answered referees queries. RM prepared the database after studying the records and found, references. MB supervised the research fellow. The manuscript was written by VU and SU and approved by all authors.
Competing interests: None stated.