CSF Cell Count Interpreter
CSF Cell Count Interpreter
Read a CSF white cell count, differential, glucose ratio and protein together rather than one line at a time. The four textbook columns — bacterial, viral, tuberculous, fungal — overlap more than the table suggests, and this page is built around the overlap: where the pattern decides, it says so, and where it does not, it names the test that does.
What is this CSF profile telling you?
Count + differential + glucose + protein → patternA 63-year-old presents with fever, headache and confusion over twelve hours, having had no antibiotics. The CSF is turbid: white cells 1800/µL with 90% neutrophils, CSF glucose 0.9 mmol/L against a simultaneous serum glucose of 5.6 mmol/L, and CSF protein 3.4 g/L.
What each line of the profile is actually worth
The differential is the most quoted discriminator and the one that fails most often — early viral disease is neutrophilic, and tuberculous, listerial and partially treated bacterial disease are lymphocytic.
The glucose ratio is the most reliable single line. Viral meningitis leaves it alone; bacterial, tuberculous and fungal meningitis lower it. It is also the first thing a dose of antibiotic rescues.
The protein rises in almost everything and separates almost nothing, but a protein above 2.2 g/L is one of Spanos’ four single findings carrying 99% certainty of bacterial over viral disease.
The host outranks all four. An immunocompromised patient can have meningitis with a normal count.
Worked example
A 63-year-old presents with fever, headache and confusion over twelve hours, having had no antibiotics. The CSF is turbid: white cells 1800/µL with 90% neutrophils, CSF glucose 0.9 mmol/L against a simultaneous serum glucose of 5.6 mmol/L, and CSF protein 3.4 g/L.
Count 1800/µL — above the 1000 threshold this page uses, and above Spanos' 2000/µL figure only when you look at the neutrophil count: 90% of 1800 is 1620/µL, which clears his 1180/µL threshold for 99% certainty on its own
Glucose ratio: 0.9 / 5.6 = 0.16, well below 0.4 and below Tamune's 0.36
Protein 3.4 g/L, above Spanos' 2.2 g/L threshold — a second single finding carrying 99% certainty
Neutrophil predominant, count high, ratio low: all three point the same way, so the first rule fires
Action: empirical antibiotics immediately, with dexamethasone per the local guideline, plus Gram stain and culture, blood cultures, molecular meningitis panel and the CSF lactate interpreter. Cover Listeria — at 63 this patient is in the age group a cephalosporin alone does not protect
Change one entry and the page changes with it. Raise the CSF glucose to 3.4 mmol/L, so the ratio becomes 0.61, and the same count and differential read as a neutrophilic pleocytosis with a preserved ratio — which still gets treated, because the glucose is the line a single dose of antibiotic rescues first
The four columns, and what each of them shares with the others
| Pattern | Count (cells/µL) | Differential | Glucose ratio | Protein | What it is confused with |
|---|---|---|---|---|---|
| Bacterial | Often 1000+, but frequently in the low hundreds | Neutrophil | Low | High | Early viral disease, which is also neutrophilic; a parameningeal focus |
| Viral | Usually under 1000 | Lymphocyte, but neutrophilic early | Normal | Normal or mildly raised | Early bacterial disease; partially treated bacterial disease |
| Tuberculous | 5 to 1000, median 50 to 450 | Lymphocyte, but may be neutrophilic early | Low, median about 0.27 | 0.5 to 3 g/L | Fungal and listerial meningitis; partially treated bacterial disease |
| Fungal | Variable, often modest; may be normal | Lymphocyte | Low or normal | Raised | Tuberculous meningitis; in advanced HIV, a normal profile altogether |
| Partially treated bacterial | Unchanged by antibiotics | Any | Rescued upwards | Falls | Viral and tuberculous meningitis |
| Immunocompromised host | May be normal in any of the above | Any | Any | Any | A normal CSF |
What a dose of antibiotic before the tap does to each line
| Line of the profile | Effect of prior antibiotics | Consequence |
|---|---|---|
| White cell count | Not significantly changed in Nigrovic’s 245-child series | The count is the line that survives partial treatment best |
| CSF glucose | Rises, more markedly beyond 12 hours of pretreatment | A rescued glucose makes a bacterial CSF look viral |
| CSF protein | Falls, more markedly beyond 12 hours | Same direction of error |
| Culture | Significantly more likely to be negative | Meningococcus can be sterilised within about 2 hours, pneumococcus within about 4 |
| CSF lactate | Sensitivity falls to 0.49 (95% CI 0.23–0.75) in Sakushima’s meta-analysis | Lactate is the line MOST degraded by pretreatment, not the one that survives it |
| Molecular panel | Largely unaffected — nucleic acid persists after the organism is dead | This is the test to send when antibiotics came first |
Why four tidy columns would be teaching a confidence the fluid does not support
The classic CSF table sets bacterial, viral, tuberculous and fungal meningitis side by side with a distinct profile in each column, and read that way it is a good teaching device and a poor decision aid. Every one of its defining features is shared with at least one neighbouring column. Early viral meningitis is neutrophilic. Early tuberculous meningitis is neutrophilic. Listerial meningitis is lymphocytic in a substantial minority of cases and is not covered by the cephalosporin most services reach for first. Cryptococcal meningitis in advanced HIV may produce almost no cellular response at all. And a single dose of antibiotic given in primary care the day before moves the glucose and the protein towards normal while leaving the count where it was.
That last point is worth stating precisely, because the folklore around it is wrong in a specific way. The common teaching is that partial treatment turns a bacterial CSF lymphocytic. The one prospective study that measured it — Nigrovic and colleagues, 245 children with bacterial meningitis across twenty emergency departments, 85 of them tapped after antibiotics had been given — found that the total white cell count was not significantly affected. What changed was the chemistry: the glucose rose, the protein fell, and both effects became more marked beyond twelve hours of pretreatment. Cultures were significantly more likely to be sterile. So partial treatment does not rearrange the cells so much as rescue the numbers either side of them, and it removes the organism that would have settled the question.
The same study explains why the sequence of tests matters more than the arithmetic. A molecular meningitis and encephalitis panel detects nucleic acid that persists after the organism is dead, so it is the test least damaged by a prior antibiotic and the one worth sending first when the history includes one. CSF lactate, by contrast, is the line most damaged by it: Sakushima’s meta-analysis of thirty-three studies found a pooled sensitivity of 0.93 overall but only 0.49 in patients pretreated with antibiotics. A reassuring lactate in someone who has already had a dose of ceftriaxone is worth roughly half what the same number would be worth in an untreated patient, and reading it as though the two situations were equivalent is the specific error this page exists to prevent.
What survives all of this is a way of working rather than a threshold. Read the glucose as a ratio, not as an absolute, because a CSF glucose of 3.0 mmol/L means different things at a serum of 5 and a serum of 12. Read the differential as a pointer rather than a verdict, and expect it to be wrong early in the illness in both directions. Give the host more weight than any of the numbers, because an immunosuppressed patient can have meningitis with a normal count and the count will not tell you so. And when the profile is ambiguous, the useful next step is almost never more arithmetic on the same sample — it is a molecular panel, a repeat tap twelve to twenty-four hours later, or the specific test for the organism you are worried about.
Frequently asked questions
Can the CSF profile distinguish bacterial from viral meningitis on its own?
Sometimes, at the extremes. Spanos and colleagues found that any one of a CSF glucose below 1.9 mmol/L, a protein above 2.2 g/L, a white cell count above 2000/µL or a neutrophil count above 1180/µL carried at least 99% certainty of bacterial rather than viral meningitis. Outside those extremes it cannot. Early viral meningitis is neutrophilic, early bacterial meningitis can be modest in count, and partially treated bacterial meningitis has a rescued glucose and a sterile culture. The profile narrows the field; a molecular panel, a culture and the clinical course close it.
Does a lymphocytic CSF exclude bacterial meningitis?
No. Listerial meningitis is lymphocytic in a substantial minority of cases, and Listeria is not covered by a third-generation cephalosporin — which is why the empirical regimen adds amoxicillin or ampicillin in the over-50s, the pregnant and the immunosuppressed. Partially treated bacterial meningitis is the other reason: a dose of antibiotic given before the tap rescues the glucose and sterilises the culture, and the resulting profile can be read as viral by anyone who does not ask about the preceding week’s prescriptions.
Is CSF lactate more reliable than the differential when antibiotics have already been given?
No — this is the most commonly inverted fact on the CSF profile. CSF lactate is an excellent discriminator in an untreated community-acquired presentation, with a pooled sensitivity of 0.93 and specificity of 0.96 in Sakushima’s meta-analysis of thirty-three studies. In patients pretreated with antibiotics, the pooled sensitivity in that same analysis was 0.49. Lactate is the line most degraded by prior antibiotics, not the one that survives them. What survives best is the white cell count, which Nigrovic’s prospective series found was not significantly changed by pretreatment, and the molecular panel, which detects nucleic acid after the organism is dead. The CSF lactate interpreter covers the threshold and the other contexts in which it shifts.
Why does the page ask for the glucose ratio rather than the CSF glucose?
Because the absolute figure is meaningless without the serum it came from. A CSF glucose of 3.0 mmol/L is a normal 0.60 against a serum of 5.0 and a markedly low 0.25 against a serum of 12.0, and it is the absolute figure that gets reported and acted on. The ratio is also unit-free, so it is the same number whether the laboratory reports in mmol/L or mg/dL. The CSF to serum glucose ratio interpreter covers the timing of the paired sample and the effect of hyperglycaemia on what the expected ratio should be, and the CSF to serum glucose ratio calculator does the arithmetic.
The patient is immunocompromised and the CSF cell count is normal. Is that reassuring?
No, and this is the pattern most likely to kill. The pleocytosis is the host’s response, so a host who cannot mount one produces a normal-looking CSF while the infection proceeds. In advanced HIV, cryptococcal meningitis frequently presents with minimal or absent pleocytosis and is diagnosed by cryptococcal antigen testing rather than by the cell count. Send the full panel regardless of the count — Gram stain and culture, cryptococcal antigen, a molecular meningitis panel, mycobacterial and fungal cultures — record the opening pressure, and treat empirically while they are pending if the patient is unwell.
How does a traumatic tap change the interpretation?
It carries the patient’s own blood, and with it white cells and protein, into the sample, so both are unreliable in the same direction. The correction formulas in circulation are not agreed and none of them is endorsed by a major guideline. The glucose and the glucose ratio survive a blood-stained tap better than the count or the protein, so read those and send the culture and the molecular panel rather than trying to rescue the count arithmetically. The traumatic tap white cell correction calculator and the traumatic tap CSF protein correction calculator set out the formulas and their limits.
Related calculators
References
- van de Beek D, Cabellos C, Dzupova O, et al. ESCMID guideline: diagnosis and treatment of acute bacterial meningitis. Clin Microbiol Infect. 2016;22(Suppl 3):S37–S62.
- Spanos A, Harrell FE Jr, Durack DT. Differential diagnosis of acute meningitis. An analysis of the predictive value of initial observations. JAMA. 1989;262(19):2700–2707.
- Nigrovic LE, Malley R, Macias CG, et al. Effect of antibiotic pretreatment on cerebrospinal fluid profiles of children with bacterial meningitis. Pediatrics. 2008;122(4):726–730.
- Sakushima K, Hayashino Y, Kawaguchi T, Jackson JL, Fukuhara S. Diagnostic accuracy of cerebrospinal fluid lactate for differentiating bacterial meningitis from aseptic meningitis: a meta-analysis. J Infect. 2011;62(4):255–262.
- Marais S, Thwaites G, Schoeman JF, et al. Tuberculous meningitis: a uniform case definition for use in clinical research. Lancet Infect Dis. 2010;10(11):803–812.
- Tamune H, Takeya H, Suzuki W, et al. Cerebrospinal fluid/blood glucose ratio as an indicator for bacterial meningitis. Am J Emerg Med. 2014;32(3):263–266.
Medical Disclaimer: The tools and content provided here are for educational and reference purposes only. They are not intended to substitute for professional medical advice, diagnosis, or treatment. Clinical decisions should always be based on the comprehensive assessment of a qualified healthcare professional.