Relationship of clinical presentation to time to antibiotics for the emergency department management of suspected bacterial meningitis.

STUDY HYPOTHESIS

The acuity and specificity of the clinical presentation of bacterial meningitis are significantly associated with the time to antibiotic administration.

DESIGN

Retrospective case series.

SETTING

Seven hundred-bed university and 1,000-bed community hospital.

PARTICIPANTS

One hundred twenty-two children and adults primarily evaluated in the emergency department and admitted with the diagnosis of suspected bacterial meningitis.

METHODS

The ED chart was reviewed for demographic, historical, physical examination, and time data and sequence of interventions. In addition, we categorized patient presentations as "sick" or not and as "classic" or not based on the following predetermined definitions. A "sick" presentation was defined as at least two of the following: temperature of more than 40 C, lethargic or comatose mental status, hypotension, or tachycardia. A "classic" presentation was defined as temperature of more than 39 C and at least one of the following: nuchal rigidity, bulging fontanelle, or abnormal mental status. Association of clinical variables and management practices to time to antibiotics was analyzed by analysis of variance and regression.

RESULTS

The geometric mean time from ED registration until antibiotic initiation was 2.7 hours (range, 0.5 to 18 hours). Clinical factors that were associated independently with less time to antibiotics (hours less, P value) were a history of vomiting (0.5 hour, P = .06), no history of headache (0.8 hour, P = .01), hypotension (1.0 hour, P = .02), a bulging fontanelle (0.9 hour, P = .01), and a "sick presentation" (0.5 hour, P = .06). Management scenarios in which antibiotics were not administered until after return of results of computed tomography head scan or laboratory cerebrospinal fluid analysis and the practice of initiation of antibiotics on the ward compared with in the ED were associated independently with even greater delays (1.7 to 1.8 hours, P < .0001).

CONCLUSION

Certain clinical factors, particularly those associated with acute illness compared with those that suggest the specific diagnosis, are associated with less time to antibiotics. Management practices, such as the order of interventions and the site of initiation of antibiotic therapy, appear to be of much greater importance in predicting antibiotic timeliness and represent an area of potentially avoidable delay for the ED management of suspected bacterial meningitis.