| dc.description.abstract | INTRODUCTION:
Acute bacterial meningitis remains an important cause of morbidity and mortality in Uganda and worldwide and a noticeable number of survivors (22-28%) are left with marked neurological sequel despite adequate treatment.
Diagnosis of ABM among infants depends on signs and symptoms and CSF analysis with CSF culture as the gold standard, however, imaging investigations like cranial-ultrasound have been found to be extremely valuable in diagnosis of ABM and its complications.
OBJECTIVES:
The aim of this study is to establish the diagnostic accuracy of cranial ultra sound in acute bacterial meningitis among infants admitted in Mulago Hospital using CSF gramstain, protein and leucocyte count as operational gold standard.
DESIGN AND SETTING:
A Cross sectional study study was conducted in acute care unit of mulago hospital from December 2008 to April 2009.
STUDY POPULATION:
These were infants aged 0-12 months admitted with clinically suspected ABM and in whom CSF had been sent for analysis.
A Sample size of 227 infants was used. Recruited infants after undergoing routine lumbar puncture underwent a cranial ultrasound examination and findings were documented using a pre-coded questionnaire. Data was entered in a computer and checked for consistency and accuracy.
STATISTICAL ANALYSIS:
Results were analysed using SPSS II with the assistance of a statistician. Cross tabulation of cranial ultrasound scan findings with CSF results led to computation of sensitivity, specificity, predictive vales and likelihood ratios.. Univariate and bivariate analysis were conducted. Logistic regression analysis was done to determine sonographic features most predictive of ABM in infants.
RESULTS:
Cranial ultrasound scan has a sensitivity of 85%, specificity of 74.1%, positive predictive value 32.9%, negative predictive value 97.2%, positive diagnostic likelihood ratio 3.35 {95% CI (2.5-4.49)} and negative diagnostic likelihood ratio as 0.19 {95% CI(0.08-0.49)}. The most common sonographic abnormalities among proven cases of ABM were echogenic sulci (92.6%), ventriculomegaly (77.8%) and parenchymal abnormalities (62.9%). Sonographic findings most predictive of ABM among infants are echogenic sulci, ventriculitis and cerebral edema.
CONCLUSION:
Cranial ultrasound scan has a high sensitivity of 85% and high negative predictive value of 97.2%, making it a very useful screening and diagnostic test for acute bacterial meningitis.
RECOMMENDATION:
Cranial ultrasound scan should be considered as an initial diagnostic tool in management of acute bacterial meningitis followed by CSF analysis for confirmation if positive scan result is obtained. | en_US |
