OBJECTIVES: The mortality due to pneumococcal meningitis remains high, ranging from 15 to 60%. The purpose of this study was to analyze the prognosis of pneumococcal meningitis using a multivariate approach. The objectives were to select the most accurate initial mortality risk factors in pneumococcal meningitis and to determine an individual probability of death according to these prognostic factors. METHODS: The records of 105 consecutive cases of proven pneumococcal meningitis were retrospectively reviewed. The following parameters, recorded at presentation were considered for prognostic analysis: age, sex, duration of symptoms and antimicrobial chemotherapy before admission, underlying debilitating conditions, occurrence of seizures, pneumococcal bacteraemia or pneumonia, Glasgow coma scale score, Simplified Acute Physiological Score (SAPS), Cerebral Spinal Fluid (CSF) leucocyte count, CSF protein level, CSF glucose level and CSF/blood glucose ratio. Those parameters having a prognostic value after univariate logistic regression analysis were then entered into a multivariate stepwise logistic regression model. RESULTS: In multivariate analysis, a fatal outcome was best predicted by 4 variables. In descending order, these were the Glasgow coma scale score, age, CSF glucose level and concomitant pneumonia. The risk of dying was ten times higher in patients with Glasgow coma scale score < 7, seven times higher in patients older than 45 or having a CSF glucose level lower than 0.6 mmol/L and 4 times higher in patients with pneumonia. The logistic model provided an equation for the probability of dying that proved to have predictive values greater than 80% in determining the individual prognostic of pneumococcal meningitis. CONCLUSIONS: Our results are in agreement with most of previously published findings. We further demonstrated that the individual prognosis of pneumococcal meningitis could be accurately predicted with only 4 variables and highlighted the preeminent prognostic value of Glasgow coma scale score.
OBJECTIVES: The mortality due to pneumococcal meningitis remains high, ranging from 15 to 60%. The purpose of this study was to analyze the prognosis of pneumococcal meningitis using a multivariate approach. The objectives were to select the most accurate initial mortality risk factors in pneumococcal meningitis and to determine an individual probability of death according to these prognostic factors. METHODS: The records of 105 consecutive cases of proven pneumococcal meningitis were retrospectively reviewed. The following parameters, recorded at presentation were considered for prognostic analysis: age, sex, duration of symptoms and antimicrobial chemotherapy before admission, underlying debilitating conditions, occurrence of seizures, pneumococcal bacteraemia or pneumonia, Glasgow coma scale score, Simplified Acute Physiological Score (SAPS), Cerebral Spinal Fluid (CSF) leucocyte count, CSF protein level, CSF glucose level and CSF/blood glucose ratio. Those parameters having a prognostic value after univariate logistic regression analysis were then entered into a multivariate stepwise logistic regression model. RESULTS: In multivariate analysis, a fatal outcome was best predicted by 4 variables. In descending order, these were the Glasgow coma scale score, age, CSF glucose level and concomitant pneumonia. The risk of dying was ten times higher in patients with Glasgow coma scale score < 7, seven times higher in patients older than 45 or having a CSF glucose level lower than 0.6 mmol/L and 4 times higher in patients with pneumonia. The logistic model provided an equation for the probability of dying that proved to have predictive values greater than 80% in determining the individual prognostic of pneumococcal meningitis. CONCLUSIONS: Our results are in agreement with most of previously published findings. We further demonstrated that the individual prognosis of pneumococcal meningitis could be accurately predicted with only 4 variables and highlighted the preeminent prognostic value of Glasgow coma scale score.
Authors: David E Barroso; Daniel Godoy; Terezinha M P P Castiñeiras; Mary M Tulenko; Maria C Rebelo; Lee H Harrison Journal: Pediatr Infect Dis J Date: 2012-01 Impact factor: 2.129
Authors: Christian T Brandt; Jens D Lundgren; Søren Peter Lund; Niels Frimodt-Møller; Thomas Christensen; Thomas Benfield; Frank Espersen; David M Hougaard; Christian Østergaard Journal: Infect Immun Date: 2004-08 Impact factor: 3.441
Authors: Maria Karanika; Vasiliki A Vasilopoulou; Antonios T Katsioulis; Panagiotis Papastergiou; Maria N Theodoridou; Christos S Hadjichristodoulou Journal: PLoS One Date: 2009-07-29 Impact factor: 3.240