Grand Rounds: “Bacterial Meningitis”

September 11, 2008

Bellevue Amphitheater

Commentary by  Stephen Carpenter MD, PGY-2

Dr. Vincent Quagliarello, Professor of Medicine at Yale University, delivered his Medicine Grand Rounds lecture entitled “Bacterial Meningitis in Adults: Principles and Practice in 2008” on Wednesday, September 3rd, 2008. He began with some of the history of the disease and noted that despite all the medical advances, bacterial meningitis still has a mortality rate of 27% and an even higher rate of 54% in poorer nations where high rates of HIV infection confound the problem. A quick review of the etiologies based on age groups ensued, with some discussion of the emergence of resistant-strains of S. pneumoniae. Currently, 25% of S. pneumoniae are penicillin resistant in the United States and 10% are multi-drug resistant.

Dr. Quagliarello then discussed some of his research evaluating the efficacy of clinical signs in patients with bacterial meningitis. Although a retrospective review of older published case series (JAMA, 1999) showed a combined sensitivity of 70% for nuchal rigidity, 57% for Kernig’s sign, and 97% for Brudzinski’s sign, a prospective contemporary observational study at Yale showed sensitivities of 30%, 5%, and 5%, respectively for diagnosis (Clin Infec. Dis, 2002). The conclusion was that these bedside diagnostic signs have limited diagnostic accuracy in the contemporary spectrum of patients seen in clinical practice, and that better bedside diagnostic signs are needed.

Furthermore, there has been a move towards performing non-contrast head CTs prior to LPs over the past 30 years. This is another area that Dr. Quagliarello has done extensive research. Proponents of doing CT scans argue that it can help to identify patients at risk for herniation; however, those opposed state that it can delay time to treatment. Based on his research, Dr. Quagliarello was able to show that the risk for herniation was so low as to be undetectable in an observational study of 301 patients, and only patients with specific risk factors were at risk and therefore deserving of a CT scan prior to LP. Those risk factors were: age >60, immunocompromised status, a known CNS mass, seizure within 1 week, decreased level of consciousness, inability to follow commands and focal findings on neurologic exam. If none of these features were present, there was a 97% chance that the CT scan would be completely normal, and those that were abnormal would not have significant mass effect to preclude a lumbar puncture (NEJM, 2001).

Once the diagnosis of bacterial meningitis is suspected, their prognosis can be determined based on 3 risk factors; namely, altered mental status, hypotension and seizures. If none of those risk factors are present, patients have an approximate 8% risk of adverse outcomes (i.e., death or neurological morbidity present at discharge); if 1 risk factor is present, the risk increases to 30% and if 2 or more are present, the risk is 63%.

At this point, Dr. Quagliarello discussed factors related to the actual invasion of the CSF. He noted four major steps required by the identified organism to invade the CSF: mucosal colonization, bacteremia, crossing the blood-brain barrier, and CSF survival. Of those, the critical step of crossing the blood-brain barrier appears to be mediated, at least in part, by the platelet-activating factor receptor for S. pneumoniae. Those observations have led to the use corticosteroids in bacterial meningitis, with the thinking that they could attenuate the breach of the blood-brain barrier. In a large clinical trial in the Netherlands, adjunctive dexamethasone reduced mortality and morbidity, but the overall effect appeared limited to those with S. pneumoniae disease and those with Glasgow Coma Scale of <11. Current recommendations for empiric treatment of bacterial meningitis require bactericidal antibiotics. For immunocompetent adults with suspected bacterial meningitis, the recommendations are Ceftriaxone plus Vancomycin (+Ampicillin if age >60 since Listeria is considered in this age group); however, if corticosteroids are given then Rifampin is recommended in place of Vancomycin, because dexamethasone inhibits the penetration of vancomycin, but not rifampin, into the CSF. In immune deficient or neurosurgical patients, double coverage of gram-negative organisms is necessary.

In summary: 1. An LP and initiation of antibiotics should be administered as soon as possible, but the impact of antibiotic timing on clinical outcome is driven more by the level of disease severity at the time of antibiotic administration, not by any arbitrary time delay from ER arrival, 2. LP should be performed without head CT if none of the high risk features were present to expedite diagnosis, and 3. treatment with dexamethasone appears to improve outcomes if they are administered before or simultaneously with antibiotics and if the patient’s Glasgow Coma Scale value is <11. The benefits of dexamethasone in developing nations is unproven, and two recent clinical trials did not reveal the overall benefit seen in the trial from the Netherlands.