A 28-year-old male presents to the emergency department with fever and severe headache. He has had a fever for 24 hours. Headache began that morning. On exam, he has marked nuchal rigidity, with a temperature of 103.5° F. He is fully oriented and has a nonfocal neurologic examination.

What would you do?

A) Lumbar puncture (LP).

B) CT scan, then LP.

C) Antibiotics, CT scan, then LP.

D) MRI, then LP.

E) Antibiotics, MRI, then LP.

The practice of obtaining a head CT scan (or MRI) before performing a lumbar puncture (LP) is commonplace in emergency departments. The concern is that if a lumbar puncture is done in a patient with increased intracranial pressure, then brain herniation could occur. How common is brain herniation, are there useful clinical indicators that make a CT scan not helpful, and does this concern reach myth status?

The routine use of head CT before LP in cases of suspected meningitis is common in many emergency departments.1,2 It is hard to get an absolute risk assessment on brain herniation following lumbar puncture. It does occur, but causality is hard to prove.3 Patients with rapidly developing, space-occupying lesions (hematomas, abscesses, recent large infarcts) can herniate without LP, so it is difficult to prove that herniations following LPs in case series were due to the LP.

In a retrospective study well before CT scans were available, 401 patients with brain tumors who had LP were reviewed, and 32% of the patients had papilledema.3 There was only one poor outcome because of the LP. This would be considered a high-risk group for complications, because many of these patients had clear focal neurologic signs, and one-third of them had papilledema.

There have been a number of studies looking at whether there is utility in obtaining a CT scan prior to LP in patients with suspected acute bacterial meningitis.

Dr. N.D. Baker of Brigham and Women’s Hospital, Boston, and colleagues retrospectively reviewed the records of 112 patients who had a routine CT before LP for suspected meningitis.2 Regardless of CT scan result, all patients received a lumbar puncture. Four patients had mass lesions on CT scan, though none of these patients had an increased opening pressure. No patient had an adverse outcome from LP.

Dr. Paul Greig and Dr. D. Goroszeniuk of Horton General Hospital, Banbury, England, reported on a retrospective study of all patients over a 6-month period considered for a LP.4 A total of 64 LPs were considered; 54 of these patients had a CT before LP. No patients had a bad outcome from the lumbar puncture.

The sensitivity and negative predictive value of a normal neurologic exam were very good in this study, leading to the conclusion from the authors that a normal neurologic exam and fundoscopic exam is an accurate predictor of a normal CT scan. The authors were dismayed that only 45% of the patients in the study received a fundoscopic exam.

Rodrigo Hasbun, MD, formerly of Yale University, New Haven, Conn., and his colleagues did a prospective study of 301 patients with suspected meningitis to see if clinical signs and symptoms could help guide utilization of CT scans.5 A total of 235 patients got CT scans, and 24% were abnormal.

The clinical features associated with higher likelihood of an abnormal CT were age greater than 60 years, immunosuppression, known CNS disease, and a seizure within the past week. On exam, altered mental status, aphasia, and focal neurologic findings were associated with higher likelihood of an abnormal CT scan.

Of the 96 patients who did not have any of these features, 93 had a normal CT exam, giving a negative predictive value of 97%.

No patients in the study had herniation from lumbar puncture. The only patients in the study who had brain herniations were two patients who had severe mass effect on CT, and did not receive lumbar punctures.

I think the patient in this case should have a lumbar puncture and does not need any imaging. I think that there is no reason to get neuroimaging in patients with suspected meningitis who are alert and have a nonfocal neurologic exam, and do not have papilledema.


1. Br J Radiol. 1999 Mar;72(855):319 .

2. J Emerg Med. 1994 Sep-Oct;12(5):597-601 .

3. AMA Arch Neurol Psychiatry. 1954 Nov;72(5):568-72 .

4. Postgrad Med J. 2006 Mar;82(965):162-5 .

5. N Engl J Med. 2001 Dec 13;345(24):1727-33 .

Dr. Paauw is professor of medicine in the division of general internal medicine at the University of Washington, Seattle, and he serves as third-year medical student clerkship director at the University of Washington. Contact Dr. Paauw at dpaauw@uw.edu.