to the editor: Tunkel and Scheld¹ suggest that when bacterial meningitis is suspected, “initiation of antimicrobial therapy optimally should begin within 30 minutes of the patient's presentation to the hospital.” The authors qualify the statement by noting that no definitive study has correlated a delay in the initiation of therapy with a worsened outcome. Little evidence has suggested that delays of one to two hours in antibiotic treatment lead to adverse outcomes.

Dr. Kallio and colleagues² compared the outcome of children three months to 15 years of age in whom the diagnosis of meningitis was made immediately with the outcome in those in whom the diagnosis was delayed up to four days. No difference was found in the outcomes measured at six months. The study did not address the interval from diagnosis to treatment. The statistical and medicolegal issues of this study were debated in a subsequent letter to the editor by Dr. Wenner, with a reply from Dr. Kallio and his associates.³

Dr. Meadow and colleagues⁴ investigated the actual time from patient presentation to administration of antibiotics. The study setting consisted of two university-affiliated pediatric emergency departments that were staffed by residents, fellows and attending physicians. The median time from presentation to administration of antibiotics in children with meningitis was two hours. Only one of 93 patients received treatment within 30 minutes, and only 15 received treatment within 60 minutes of initial presentation. No significant correlation was found between the time elapsed before treatment and outcome.

Drs. Talan and Zibulewsky⁵ performed a similar study of 122 patients, ranging in age from one month to 60 years of age or older, in two teaching hospitals. Patients were evaluated by pediatric, internal medicine or emergency medicine residents and/or emergency medicine attending physicians. The mean time from registration to antibiotic administration was 2.7 hours. The authors did not attempt to correlate treatment timing with outcome.

A literature review performed by Radetsky⁶ addressed delays in diagnosis and treatment. Evaluation included 22 studies of different design, involving a total of 4,707 patients. The findings showed that in patients with “insidious” or nonspecific symptoms at onset, a delay in treatment of three to five days did not result in adverse sequelae; in patients with “fulminant meningitis,” a delay in therapy was not connected with outcome; and in patients with “clinically overt nonfulminant meningitis,” an “excessive” delay in initiating therapy increased the risk for permanent injury. “Excessive” delay was not defined.

The goal of physicians should be to diagnose and initiate the proper treatment of meningitis as soon as possible. Failure to do so may possibly place the patient at risk for adverse outcome and the physician at risk for medicolegal inquiry. It would not be reasonable to delay the diagnosis and treatment of bacterial meningitis, but many factors contribute to the length of time from a patient's presentation to treatment. Standards of care should be based on data and should reflect the realities of typical practice settings.

in reply: We agree with Dr. Hash that no definitive study has correlated a delay in the initiation of antimicrobial therapy with a worsened outcome in patients with bacterial meningitis. He cites several studies to support this statement, which are all flawed given their retrospective nature. Even in the large retrospective review by Radetsky,¹ in the subgroup of patients with clinically apparent meningitis and documented neurologic findings, there were insufficient data to establish any correlation between duration of symptoms and outcome. Furthermore, Radetsky did not separate the data based on causative organism. This is critical since Haemophilus influenzae type b meningitis is associated with a mortality rate of 3 to 6 percent, whereas patients with meningitis caused by Streptococcus pneumoniae have a mortality rate ranging from 19 to 26 percent.²

Other studies have supported the concept of rapid administration of antimicrobial therapy. One such study³ was a retrospective analysis of 46 consecutive patients with meningococcal disease before presentation to the hospital; none of the 13 patients who were given parenteral penicillin by the referring physician died, while eight deaths occurred among 33 patients who were admitted without such initial treatment. Another retrospective study⁴ reviewed hospital notes and laboratory and public health medical department records of patients in Southwest England who were presumed to have meningococcal meningitis. Patients who were given parenteral benzylpenicillin by general practitioners before hospital admission had a lower mortality rate than patients treated later (5 percent versus 9 percent); the mortality rate was also lower in patients who presented with a hemorrhagic rash (5 percent versus 12 percent). Although these data are retrospective in nature and did not specifically review patients with meningitis, they suggest that early administration of antimicrobial therapy to patients with meningococcal disease improves outcome. Furthermore, in a retrospective review⁵ by the British Society for the Study of Infection of 305 patients with bacterial meningitis, mortality rates were lower in the patients who received antimicrobial therapy before admission than in those treated later (2 percent versus 12 percent).

Dr. Hash also makes the point that early administration (within 30 minutes) of antimicrobial therapy in patients with bacterial meningitis does not reflect the realities of typical practice settings; this was suggested by two of the trials he cited. However, these studies were also retrospective in nature. We feel they highlight the importance of educating physicians to recognize the likelihood of bacterial meningitis as soon as possible to begin appropriate therapy.

In the absence of clinical data and in view of the fact that ethical concerns preclude performance of a prospective, randomized trial of the time from presentation to administration of antimicrobial therapy in patients with bacterial meningitis, what is the optimal timing of administration of antimicrobial therapy that should be recommended? Our suggestion that antimicrobial therapy should optimally be administered within 30 minutes of the patient's presentation to the hospital was not meant as a recommendation for the standard of care in the medicolegal sense. We feel that this recommendation will serve to alert health care professionals to the importance of considering the diagnosis of bacterial meningitis in the right clinical setting and administering antimicrobials as rapidly as possible.

When patients with bacterial meningitis present for medical attention, all that is known is the duration of symptoms—not the duration of bacterial meningitis. Many patients who are eventually diagnosed with bacterial meningitis and who have had symptoms for more than three to five days may not have had bacterial meningitis for that entire period of time. In contrast, patients with fulminant presentations may have had bacterial meningitis long before presentation.

We agree with Dr. Hash that failure to rapidly diagnose and treat the patient with bacterial meningitis may only possibly place the patient at risk for an adverse outcome. However, we feel that the only reasonable course, in the best interest of the patient and in the absence of clinical data, is emergent diagnosis and rapid initiation of antimicrobial therapy.