Validation of the Glasgow Meningococcal Septicemia Prognostic Score

Abstract

OBJECTIVE: To derive performance characteristics for the Glasgow Meningococcal Septicemia Prognostic Score (GMSPS). DESIGN: Retrospective case-note study. SETTING: Two children's hospitals with Regional Intensive Care Unit. PATIENTS: One hundred twenty-three children with proven meningococcal septicemia (some with concurrent meningitis) from January 1, 1977 to December 31, 1986. MEASUREMENTS AND MAIN RESULTS: All 14 children who died after arrival scored greater than or equal to 8 either on admission (n = 8) or afterward (n = 6). Of 109 survivors, five scored greater than or equal to 8 (two were postictal at the time of scoring). A GMSPS of greater than or equal to 10 at or after admission predicted death with sensitivity 100%, specificity 98%, and positive predictive value of 88%; for GMSPS of both greater than or equal to 8 or 9, the values were 100%, 95%, and 74%, respectively. CONCLUSIONS: The GMSPS is a rapid clinical score that performs well in identifying children with poor prognosis who might benefit from early intensive care. It should be studied prospectively and compared with other scoring systems.