Skip to content

Do antibiotics affect CSF results?


Share on facebook
Share on twitter
Share on linkedin
Share on whatsapp

Paediatricians often have to make a decision about whether to just go ahead and give antibiotics in suspected meningitis, or wait for a lumbar puncture (LP) – this could be due to parental refusal, an unstable patient, or a failed attempt.

There is often a discussion about repeating the LP later that day, or even the following day. We all know that having had antibiotics might affect the results. But what effect does it actually have?

Here, I summarise three key papers looking at this very question – do antibiotics affect cerebrospinal fluid (CSF) results in bacterial meningitis?

Paper 1 - Michael et al (2010)

View paper

Who were the patients?

Patients were adults from a large UK district hospital and were identified retrospectively through a coding diagnosis of meningitis.


How was bacterial meningitis defined for inclusion criteria?

Patients had to have clinical features consistent with meningitis and had to have had an LP with a cell count of >4 cells/ml.


How many patients were included?

They had 92 patients included in the study.

They had been diagnosed with meningitis and had an LP with >4 cells/ml.

They all received antibiotics prior to the LP.


What did they find?

What they concluded from the analysis was that once antibiotics have been started, an LP within 4 hours of antibiotic administration is still likely to be culture positive.  After the 4 hour mark the proportion of positive CSF cultures dwindled. 

Paper 2 - Kengaye et al (2001)

View paper

Who were the patients?

The cohort was drawn from all patients discharged from San Diego Children’s Hospital during a 4 year period.

The patient group was identified by a coding diagnosis of bacterial or suspected bacterial meningitis.


How was bacterial meningitis defined for inclusion criteria?

CSF culture positive with bacteria; CSF WCC >10/mm3 + CSF antigen or Gram stain positive; CSF WCC >100/mm3 + blood culture positive; or CSF WCC >4000/mm3 in the absence of positive cultures.


How many patients were included?

There were 128 patients included.

43% had an LP both pre- and post-antibiotics, 30% had antibiotics prior to LP, and 27% had LP prior to antibiotics.


What did they find?

There were far less positive CSF cultures in post-antibiotic LPs.

In particular N. meningitides was sterilized earlier than Strep. penumoniae or Group B Strep. meningitis.

No N. meningitides CSF cultures were positive by 2 hours post-antibiotics.

Their conclusion was that negative cultures occurred in 44% of post-antibiotic LPs and only 8% of pre-antibiotic LPs.  And that meningococcal meningitis is very quick to sterilize.

Paper 3 - Nigrovic et al (2008)

View paper

Who were the patients?

This was a retrospective cohort study across twenty Emergency Departments in US paediatric centres.

Paediatric patients were identified through a coding diagnosis of bacterial meningitis or unspecified meningitis; and a review of positive CSF cultures for bacteria.


How was bacterial meningitis defined for inclusion criteria?

CSF culture for positive for a bacterial pathogen; CSF WCC >=10 cells/microL with positive blood culture +/- positive CSF agglutination study results.


How many patients were included?

245 patients were included.

159 (65%) had the LP before antibiotic treatment and 85 (35%) had the LP after antibiotic treatment.

Of those who had received treatment prior to LP: 24% had oral antibiotics; 69% had IV antibiotics; 7% had both oral and IV antibiotics.


What did they find?

CSF culture results were significantly more likely to be negative after receiving antibiotics.

4 hours post-antibiotics: CSF WCC was not affected by the administration of antibiotics; but the CSF glucose was significantly higher; and the CSF protein lower (although not significantly).

This was more marked (and more significant) 12 hours post antibiotics.

What should we take from this for our daily practice?

I find it hard to draw any useful conclusions from the Benedict et al study. There are three major flaws with it:

  1. Every single patient had antibiotics before having their LP.  There is no comparison to the group that had the LP first (apparently there were none in this category) and so to draw any conclusion about the effect of the antibiotics on the CSF results seems a stretch.
  2. Patients were actually excluded if their CSF had <5 cells/ml and the culture was negative.  This seems to hugely skew the results.  It could be that there were thousands of (excluded) patients who had antibiotics prior to LP and that all their CSF sample showed no WCC and were culture negative.  This would vastly change the results.  It’s also in adults which makes it difficult to draw paediatric conclusions.
  3. The patients were split into viral and bacterial meningitis groups, and part of the way this decision was made was by looking at the CSF results.  It’s self-fulling spiral.

But, it is fair to say that in the patient group they looked at, the CSF cultures were still positive even after antibiotic administration as long as it was within 4 hours.  By the time there was an 8 hour gap post-antibiotics, none of the CSF cultures were positive.

All the studies were retrospective and relied on correct coding diagnosis.  The retrospective nature also made it difficult to accurately assess timing of lumbar puncture and antibiotics administration.  Deciding the inclusion criteria for bacterial meningitis in a study about the effect on CSF results is fraught with difficulties.

Kanageye’s paper, however, does indicate that CSF culture results are affected by antibiotic administration (even within a couple of hours) and so repeating the lumbar puncture the following day may well give false reassurance. And Nigrovic’s paper reinforces this finding, and adds that CSF glucose will increase, and CSF protein will decrease, post antibiotics (especially 12 hours post antibiotics).

Although the accuracy of the timing measurement is potentially flawed, this is something to bear in mind.  Often in paediatrics the LP is unsuccessful, the patient is treated anyway, and the LP will be repeated the following day.  This can give falsely reassuring results.  Be wary of making decisions around length or choice of antibiotic, based on a post-antibiotic lumbar puncture



Michael B, Menezes BF, Cunniffe J, Miller A, Kneen R, Francis G, Beeching NJ, Solomon T. Effect of delayed lumbar punctures on the diagnosis of acute bacterial meningitis in adults. Emerg Med J. 2010 Jun;27(6):433-8. 

Kanegaye JT, Soliemanzadeh P, Bradley JS. Lumbar puncture in pediatric bacterial meningitis: defining the time interval for recovery of cerebrospinal fluid pathogens after parenteral antibiotic pretreatment. Pediatrics. 2001 Nov;108(5):1169-74.

Nigrovic LE, Malley R, Macias CG, Kanegaye JT, Moro-Sutherland DM, Schremmer RD, Schwab SH, Agrawal D, Mansour KM, Bennett JE, Katsogridakis YL,Mohseni MM, Bulloch B, Steele DW, Kaplan RL, Herman MI, Bandyopadhyay S, Dayan P, Truong UT, Wang VJ, Bonsu BK, Chapman JL, Kuppermann N;American Academy of Pediatrics, Pediatric Emergency Medicine Collaborative Research Committee. Effect of antibiotic pretreatment on cerebrospinal fluid profiles of children with bacterial meningitis. Pediatrics. 2008 Oct;122(4):726-30.



About the authors

  • Tessa Davis is a Consultant in Paediatric Emergency Medicine at the Royal London Hospital and a Senior Lecturer at Queen Mary University of London.


High flow therapy – when and how?

Chest compressions in traumatic cardiac arrest

Searching for sepsis

The missing link? Children and transmission of SARS-CoV-2

Don’t Forget the Brain Busters – Round 2

An evidence summary of Paediatric COVID-19 literature


The fidget spinner craze – the good, the bad and the ugly

Parenteral Nutrition

Leave a Reply

Your email address will not be published.



We use cookies to give you the best online experience and enable us to deliver the DFTB content you want to see. For more information, read our full privacy policy here.