Share this article on:

How Best to Rule Out Bacterial Meningitis? Neurologists Weigh in on New Guidelines

Valeo, Tom

Back to Top | Article Outline


  • ✓ A new study reports that the Bacterial Meningitis Score is still helpful for helping to diagnose and distinguish children with bacterial versus viral meningitis. But some neurologists say better and quicker diagnostic methods will make this methodology obsolete.

Although most children in the US today are vaccinated against some bacteria that cause a possibly deadly form of meningitis, guidelines to determine which children have less serious viral infections work just as well, according to a study published in the Journal of the American Medical Association (2007;297:52–60). But some neurologists who specialize in neuroinfectious disease believe the guidelines could soon be obsolete.

The Bacterial Meningitis Score, as the guidelines are called, was devised to help physicians diagnose those children with viral meningitis, so they could avoid unnecessary hospital admission and a round of antibiotics, which would be ineffective against the disease and expose them to possible adverse reactions.

Lise E. Nigrovic, MD, MPH, of the division of emergency medicine at Children's Hospital in Boston and Harvard Medical School, and two colleagues, Nathan Kupperman, MD, MPH, and Richard Malley, MD, developed the Bacterial Meningitis Score in 2002. The score was developed to help assess children with CSF pleocytosis (Pediatrics 2002;110:712–719). These children usually have flu-like symptoms — fever, anorexia, headache, and stiff neck. The symptoms tend to be similar in both bacterial and the far less serious viral meningitis, but only about 1 case in 25 is bacterial or fungal.

Culturing the CSF to detect bacteria provides a definite diagnosis, but results take two to three days. Waiting that long to begin treatment could be catastrophic, so physicians often hospitalize the children immediately and administer broad-spectrum antibiotics, either by injection or intravenous drip, just to be safe.

“I started as a resident in pediatrics, and I was admitting a lot of kids for meningitis and giving them antibiotics, but I realized most cases were viral, not bacterial, so the antibiotics were ineffective,” Dr. Nigrovic told Neurology Today. “I wondered if I could identify which cases were which to avoid admitting low-risk patients.”

The Bacterial Meningitis Score she developed classifies patients at very low risk of bacterial meningitis if they lack all of the following criteria:

  • A CSF Gram stain for bacteria;
  • CSF absolute neutrophil count (ANC) of at least 1,000 cells/ÌL;
  • CSF protein of at least 80 mg/dL;
  • Peripheral blood ANC of at least 10,000 cells/ÌL;
  • A history of seizure before or at the time the patient is seen.

In the original study, none of the 144 patients classified as very low risk had bacterial meningitis. Since 2000, however, the heptavalent pneumococcal conjugate vaccine has been recommended for children under 2. Subsequently, the incidence of bacterial meningitis in children dropped, so Dr. Nigrovic and her colleagues set out to validate the Bacterial Meningitis Score to see if the widespread vaccination would alter the results.

“In the past, about two-thirds of (bacterial meningitis) cases were caused by pneumococcus,” Dr. Nigrovic said. “Now it's only about 40 percent. So the bacteria haven't gone away. Also there are many types of pneumococcal bacteria, and the vaccine covers only seven of them.”



Back to Top | Article Outline


Dr. Nigrovic and her colleagues included 3,295 children ranging in age from 29 days to 19 years who appeared at emergency departments at 20 US academic medical centers with CSF pleocytosis between January 2001 and June 2004. Of these patients, 121, or 3.7 percent, had bacterial meningitis, while 3,174 had aseptic, or nonbacterial, meningitis. Of the 1,714 patients categorized as very low risk according to the Bacterial Meningitis Score, only 2 had bacterial meningitis. Both were younger than 2 months, and both had E. coli meningitis and urinary tract infections, but the urinalysis in each case came back negative. As a result, the authors warn that the Bacterial Meningitis Score for infants younger than 2 months may be less accurate.

Of 1,189 patients categorized as not very low risk according to the Bacterial Meningitis Score, 119, or 10 percent, actually had bacterial meningitis.

Dr. Nigrovic and her colleagues attempted to refine the Bacterial Meningitis Score by emphasizing three variables in the following order: CSF protein level of 80 mg/dL, positive CSF Gram stain for bacteria, and peripheral ANC of at least 10,000 cells/ÌL. However, of the 1,786 patients who had none of these variables, 3, or 0.2 percent, actually had bacterial meningitis (the two infants plus another child). Therefore, the authors concluded that the original Bacterial Meningitis Score rather than the proposed refinement is better for identifying children with CSF pleocytosis who are at very low risk for bacterial meningitis. They also suggest hospital admission and immediate antibiotics for patients who have at least one risk factor, or who are younger than two months.

“Our low-risk model does not say that kids should go without antibiotics,” Dr. Nigrovic said, “just that physicians might consider outpatient management with strong consideration of a long-acting antibiotic such as ceftriaxone.”

Back to Top | Article Outline


Some neurologists, who were not involved in the current study, doubt that the guidelines will be of much use, however. “Neurologists are not going to be comfortable with this Bacterial Meningitis Score,” said Karen Roos, MD, John and Nancy Nelson professor of neurology and professor of neurosurgery at the University of Indiana School of Medicine. “They're not going to change their standard of care. I think it might be important for the pediatric emergency medicine group, but not for neurologists because bacterial meningitis is an emergency, and if you miss the diagnosis, the child could die.”

Dr. Roos — who wrote Neurologic Infectious Diseases, and Meningitis: 100 Maxims in Neurology—predicted that the Bacterial Meningitis Score would soon be obsolete because new tests, such as the polymerase chain reaction (PCR) test, can identify bacterial meningitis quickly and reliably.

“The routine availability of newer technology will change clinical decision making,” she said. “In the next few months the laboratory here will be able to give me the results of a PCR for bacterial nucleic acid on blood in under 30 minutes. This, combined with CSF Gram stain, culture and bacterial PCR, and CSF enteroviral PCR, all readily available now, will guide the management of children with suspected meningitis. Bacterial meningitis is a neurological emergency. We do not want to ‘predict’ its presence or absence.”

Dr. Roos takes a dim view of clinical predictors of disease. “They've been around for a century,” she said. “The idea is if you can predict the likelihood of the absence of disease, you can actually streamline treatment and make it less expensive.”

But the dangers posed by bacterial meningitis, in her opinion, are so great that the risks outweigh the benefits of saving on antibiotic costs or avoiding unnecessary hospitalization.

“If patients survive bacterial meningitis, they may have neurological sequelae, and antibiotics are no problem,” she said. “We routinely order neuroimaging when patients don't need it at a cost of $10,000 or so. It might cost $500 to give antibiotics.”

Back to Top | Article Outline


Larry E. Davis, MD, neurology service chief at the New Mexico VA Health Care System in Albuquerque, NM, agreed that costs of the various diagnostic tests can exceed the cost of routine antibiotics. But he noted that the antibiotics are almost always given intravenously. “Thus, you have to hospitalize the patient, and the hospital costs rapidly exceed the costs of the diagnostic tests,” he said.

He also noted that the ability to detect viral meningitis quickly might be a more effective way to quickly rule out bacterial meningitis. But, he added, many hospitals currently do not have rapid laboratory methods to detect bacteria or viral nucleic acid.



“Simultaneous bacterial and viral meningitis is very rare,” he said. “Many hospital physicians test CSF for enterovirus and herpes simplex. If the CSF is positive for either virus, they know it's very unlikely to be bacterial meningitis. Then you can discharge the patient back to the family with close follow-up.”

Dr. Davis admired the way the creators of the Bacterial Meningitis Score reversed the algorithm — instead of seeking evidence indicating that patients probably have bacterial meningitis, they developed criteria to identify patients that probably don't.

But he found it strange that the creators of the Score, after demonstrating that their criteria were highly effective in detecting the absence of bacterial meningitis, recommended giving a long-acting antibiotic anyway.

“They didn't cite a single study that says one dose of a long-acting antibiotic will kill a broad enough spectrum of bacteria sufficient to always cure the patient with bacterial meningitis,” he said. “Anytime a doctor discharges a patient with any form of meningitis, it is important that a responsible individual caring for the patient be present who could return that patient to the hospital if he or she worsens. Sending the patient home with one dose of an antibiotic could lead to a false sense of security.”

Dr. Davis also noted that the study excluded many patients at higher risk for misclassification when they have symptoms of meningitis, and these exclusions should be taken into consideration by anyone using the Bacteria Meningitis Score with patients in the ER.

“While the Bacterial Meningitis Score has some value as a screening tool,” he said, “I doubt that its use will prevent most of the aseptic meningitis patients from being admitted to the hospital for at least observation and likely antibiotic coverage until the CSF culture report returns.”

The Bacterial Meningitis Score should always be used cautiously, Dr. Nigrovic emphasized. “This rule was derived and validated with kids who were not looking critically ill,” she said. “They didn't have low blood pressure; they had no difficulty breathing; they did not have neurosurgical problems or suppressed immune systems. None of the patients had been pretreated with antibiotics prior to lumbar puncture. This was a population of kids who might very well have been sent home. We don't think every low-risk patient will go home, but in the low-risk group, the risk is extremely low.”

Immunosuppressed children were excluded from the study, “but they aren't candidates for going home,” she said.

In short, Dr. Nigrovic retains her confidence in the Bacterial Meningitis Score. “We were wondering how the performance of this decision-making tool would be now that vaccines are used,” she said. “The epidemiology has changed, but the rule performed just as well.”

Back to Top | Article Outline


• Nigrovic LE, Kuppermann N, Malley R, et al. for the Pediatric Emergency Medicine Collaborative Research Committee of the American Academy of Pediatrics. Clinical prediction rule for identifying children with cerebrospinal fluid pleocytosis at very low risk of bacterial meningitis. JAMA 2007;297:52–60.
    ©2007 American Academy of Neurology