Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Cerebrospinal fluid (CSF) lactate for distinguishing acute bacterial meningitis from acute viral meningitis in children

Three Part Question

In [non-neonatal children undergoing lumbar punctures for possible meningitis] what are [the sensitivity and specificity of CSF lactate] for [diagnosing acute community-acquired bacterial meningitis]?

Clinical Scenario

You are evaluating a pediatric patient in the emergency department in whom you are considering the diagnosis of acute community-acquired bacterial meningitis. You perform a lumbar puncture and when ordering CSF studies wonder if sending a CSF lactate would help distinguish acute bacterial meningitis (ABM) from acute viral meningitis (AVM).

Search Strategy

PubMed
((("bacteria"[mesh] OR "bacterial infections"[mesh]) AND ("meningitis"[mesh:noexp])) OR "meningitis, bacterial"[mesh] OR (mening*[tw] AND (bacteria*[tw] OR escherichia[tw] OR haemophilus[tw] OR listeria[tw] OR meningococc*[tw] OR neisseria[tw] OR pneumococc*[tw] OR streptococc*[tw]))
AND (“cerebrospinal fluid”[mesh] OR “cerebrospinal fluid”[subheading] OR “cerebrospinal fluid”[tw] OR “CSF”[tw]) AND (“lactates”[mesh] OR lacta*[tw] OR lacti*[tw]))
AND english[lang]

EMBASE
(((‘bacterium’/exp OR ‘bacterial infection’/exp) AND ‘meningitis’/exp) OR (mening*:ab,ti,kw AND (bacteria*:ab,ti,kw OR ‘escherichia’:ab,ti,kw OR ‘haemophilus’:ab,ti,kw OR listeria:ab,ti,kw OR meningococc*:ab,ti,kw OR ‘neisseria’:ab,ti,kw OR pneumococc*:ab,ti,kw OR streptococc*:ab,ti,kw)))
AND ((‘cerebrospinal fluid'/exp OR ‘cerebrospinal fluid’:ab,ti,kw OR ‘CSF’:ab,ti,kw) AND (‘lactic acid'/exp OR lact*:ab,ti,kw))
AND [english]/lim
Searches were most recently run on December 19, 2018.

Search Outcome

PubMed -- 518 results
EMBASE -- 658 results
Total -- 742 unique results

108 full text articles were reviewed after screening titles/abstracts from which 38 original articles and two systematic reviews were ultimately selected for inclusion. For completeness, the original articles included in the two systematic reviews are listed below as there appear to be several minor discrepancies in the data abstracted in the systematic reviews.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Briem et al.
1983
Sweden		Single hospital Mean 44 vs. 29y ABM vs. AVM, 52% male n=45 ABM, n=102 AVMCohortLactate cutoff3.5mMExcluded ABM without organism identified
Sensitivity100%
Specificity96.1%
LR+25.5
LR-0
Brook et al.
1978
USA		Single hospital 54% adults n=25 ABM, n=26 AVMCohortLactate cutoff3.9mM using chromatographyExcluded ABM without organism identified
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Chen et al.
2012
China		Single hospital Mean 43±15 vs. 32±17y, 45% vs. 67% male ABM vs. AVM n=21 ABM, n=30 AVMCohortLactate cutoff3.3mMIncluded patients pretreated with antibiotics/steroids
Sensitivity71.4%
Specificity100%
LR+Infinity
LR-0.286
Curtis et al.
1981
UK		Two hospitals Unknown demographics n=10 ABM, n=12 AVMCohortLactate cutoff2.8mMExcluded ABM without organism identified
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Ellis and Oei
1979
USA		Single hospital Unknown demographics n=10 ABM, n=12 "CNS inflammation" including AVMCohortLactate cutoff5.0mMExcluded ABM without organism identified "CNS inflammation" group included not only AVM but also leukemia, ICH, MS, and "hypoxic seizures"
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Ferguson and Tearle
1977
UK		Single hospital Unknown demographics n=8 ABM, n=15 AVMCohortLactate cutoff5.6mM using chromatographyUnclear definition of ABM Included patients pretreated with antibiotics
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Gastrin et al.
1979
Sweden		Single hospital Median 33 vs. 54 vs. 24y ABM vs. presumed ABM vs. AVM n=38 ABM, n=5 presumed ABM, n=17 AVMCohortLactate cutoff3.5mM using chromatographyIncluded patients pretreated with antibiotics
Sensitivity95.3%
Specificity82.4%
LR+5.40
LR-0.056
Hurd et al.
1978
USA		Single affiliated adult/children hospital 77% adult n=11 ABM, n=80 comparator group including AVMCohortLactate cutoff3.4mM using chromatographyExcluded ABM without organism identified Comparator group included not only AVM but also ICH, cancer, and post-neurosurgical patients
Sensitivity100%
Specificity97.5%
LR+40.0
LR-0
Jordan et al.
1983
USA		Single hospital Unknown demographics n=7 ABM, n=3 AVMCohortLactate cutoff3.0mMExcluded ABM without organism identified n=3 AVM makes the specificity extraordinarily fragile
Sensitivity100%
Specificity66.7%
LR+3.00
LR-0
Landaas and Von Der Lippe
1985
Norway		Single hospital 47% vs. 52% male ABM vs. AVM n=19 ABM, n=24 AVMCohortLactate cutoff6.0mMIncluded patients pretreated with antibiotics Excluded ABM without organism identified
Sensitivity84.2%
Specificity100%
LR+Infinity
LR-0.158
Lester et al.
1985
Denmark		Multicenter Unknown demographics n=16 ABM, n=15 AVMCohortLactate cutoff3.5mMIncluded patients pretreated with antibiotics
Sensitivity93.8%
Specificity86.7%
LR+7.03
LR-0.072
Lindquist et al.
1988
Sweden		Multicenter Unknown demographics n=63 ABM, n=616 comparator group including AVMCohortLactate cutoff3.5mMABM group included TB meningitis Comparator group included not only AVM but also "unclassified meningitis", other infectious CNS conditions, seizures, and strokes More favorable results listed correspond to patients <=15y, however, unknown number of patients <=15y
Sensitivity88.9%/97.0%
Specificity96.4%/99.0%
LR+24.9/97.0
LR-0.115/0.030
Mandal et al.
1983
UK		Single hospital Unknown demographics n=20 ABM, n=59 AVMCohortLactate cutoff4.3mM
Sensitivity100%
Specificity98.3%
LR+59.0
LR-0
Ponka et al.
1983
Finland		Single hospital Unknown demographics n=11 ABM, n=27 AVMCohortLactate cutoff3.0mMExcluded ABM without organism identified
Sensitivity90.9%
Specificity96.3%
LR+24.5
LR-0.094
de Almeida et al.
2009
Brazil		Single hospital Mean 21±20y, 53% male n=20 ABM, n=41 AVMCohortLactate cutoff3.5mM using Date Behring analyzer
Sensitivity80%
Specificity97.6%
LR+32.8
LR-0.205
Berg et al.
1982
Sweden		Multicenter Mean 30 vs. 21y ABM vs. AVM n=18 ABM, n=121 AVMCohortLactate cutoff3.0mMExcluded ABM without organism identified
Sensitivity88.9%
Specificity92.6%
LR+12.0
LR-0.120
Huy et al.
2010
N/A		25 studies included (both adults and children) n=783 ABM, n=909 AVMSystematic review and meta-analysisAUC CSF lactate0.984Excluded non-English studies No gray literature Excluded studies with n<8 for either ABM or AVM Excluded studies with comparator groups including but not exclusively comprised of AVM
AUC CSF WBC0.948
AUC CSF glucose0.881
AUC CSF/plasma glucose0.952
AUC CSF protein0.862
Sakushima et al.
2011
N/A		33 studies included (both adults and children) n=934 ABM, n=951 AVMSystematic review and meta-analysisSensitivity93%Excluded non-English studies No gray literature Excluded studies with comparator groups including but not exclusively comprised of AVM
Specificity96%
LR+22.9
LR-0.07
Juarez Aragon et al.
1979
Mexico		Single hospital Unknown demographics n=23 ABM, n=21 AVMCohortLactate cutoff2.2mMUnclear definition of ABM
Sensitivity100%
Specificity95.2%
LR+21.0
LR-0
Bland et al.
1974
USA		Single hospital 12 vs. 18mo ABM vs. AVM n=15 ABM, n=25 AVMCohortLactate cutoff4.5mMExcluded ABM without organism identified
Sensitivity93.3%
Specificity100%
LR+Infinity
LR-0.067
Boon and Heng
1978
Singapore		Single hospital Unknown demographics n=6 ABM, n=13 AVMCohortLactate cutoff2.4mM
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Cameron et al.
1993
UK		Single hospital Unknown demographics n=11 ABM, n=9 AVMCohortLactate cutoff4.1mMExcluded ABM without organism identified
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Controni and Rodriguez
1976
USA		Single hospital Unknown demographics n=21 ABM, n=171 "non-bacterial cases" including AVMCohortLactate cutoff3.4mM using chromatographyExcluded ABM without organism identified Comparator group included not only AVM Likely overlap in patients with Controni et al., 1977
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Controni et al.
1977
USA		Single hospital Unknown demographics n=55 ABM, n=15 AVMCohortLactate cutoff2.8mMIncluded patients pretreated with antibiotics Excluded ABM without organism identified Likely overlap in patients with Controni and Rodriguez, 1976
Sensitivity96.4%
Specificity100%
LR+Infinity
LR-0.036
Donald and Malan
1986
South Africa		Single hospital Median 17 vs. 60mo ABM vs. AVM n=43 ABM, n=23 AVMCohortLactate cutoff2.85mMExcluded ABM without organism identified
Sensitivity93.0%
Specificity100%
LR+Infinity
LR-0.070
Dwivedi and Reddy
1983
USA		Single hospital Mean 2y vs. 7y ABM vs. AVM n=9 ABM, n=3 AVMCohortLactate cutoff3.9mMn=3 AVM makes the specificity extraordinarily fragile
Sensitivity100%
Specificity66.7%
LR+3.00
LR-0
Eross et al.
1981
Australia		Single hospital Median 1 vs. 5 vs. 4y ABM vs. AVM vs. presumed AVM n=66 ABM, n=31 AVM, n=58 presumed AVMCohortLactate cutoff3.9mMIncluded patients pretreated with antibiotics Excluded ABM without organism identified
Sensitivity97.0%
Specificity100%
LR+Infinity
LR-0.030
Knight et al.
1981
USA		Single ED Unknown demographics n=68 ABM, n=20 AVMCohortLactate cutoff3.4mMExcluded culture negative ABM
Sensitivity100%
Specificity85%
LR+6.67
LR-0
Low et al.
1986
Singapore		Single hospital Mean 21±31 vs. 46±43mo, 38% vs. 74% male ABM vs. AVM n=22 ABM, n=54 AVMCohortLactate cutoff2.8mMIncluded patients pretreated with antibiotics Excluded ABM without organism identified
Sensitivity86.4%
Specificity85.2%
LR+5.83
LR-0.160
Filho et al.
2014
Brazil		Single ED Mean 5y, 68% male n=40 ABM, n=411 AVMCohortLactate cutoff3.0mMExcluded ABM without organism identified No mention of lactate assay method and no reply from author Excluded patients with "critical illness" and concurrent alternative infection requiring parenteral antibiotics
Sensitivity95%
Specificity93.7%
LR+15.0
LR-0.053
Murata and Uemura
1981
Japan		Single hospital Unknown demographics n=2 ABM, n=18 AVMCohortLactate cutoff3.4mM using chromatographyExcluded ABM without organism identified Excluded AVM without organism identified n=2 ABM makes the sensitivity extraordinarily fragile
Sensitivity100%
Specificity94.4%
LR+18.0
LR-0
Nazir et al.
2018
India		Single ED Mean 14 vs. 63m, 30% vs. 40% male ABM vs. AVM n=60 ABM, n=156 AVMCohortLactate cutoff3.0mM using Instrumentation Laboratory analyzerExcluded patients with "critical illness" and concurrent alternative infections (e.g. pneumonia)
Sensitivity90%
Specificity100%
LR+Infinity
LR-0.100
Nelson et al.
1986
Sweden		Single hospital 59% male n=11 ABM, n=28 AVMCohortLactate cutoff2.4mMExcluded ABM without organism identified
Sensitivity100%
Specificity89.3%
LR+9.33
LR-0
Rutledge et al.
1981
USA		Single hospital Unknown demographics n=29 ABM, n=13 AVMCohortLactate cutoff3.0mMIncluded patients pretreated with antibiotics Excluded ABM without organism identified
Sensitivity93.1%
Specificity69.2%
LR+3.03
LR-0.100
Ruuskanen et al.
1985
Finland		Single hospital Unknown demographics n=32 ABM, n=30 AVMCohortLactate cutoff3.0mMExcluded ABM without organism identified
Sensitivity93.8%
Specificity93.3%
LR+14.1
LR-0.067
Pit et al.
1987
Malaysia		Single hospital Unknown demographics n=15 ABM, n=4 AVMCohortLactate cutoff3.4mM using chromatographyExcluded ABM without organism identified and cases with organisms thought to represent contamination n=4 AVM makes the specificity extraordinarily fragile
Sensitivity100%
Specificity100%
LR+Infinity
LR-0
Dashti et al.
2017
Iran		Single hospital Mean 43mo, 60% male n=12 ABM, n=38 AVMCohortLactate cutoff3.4mM
Sensitivity91.7%
Specificity89.5%
LR+8.71
LR-0.093
Shaltout
1989
Kuwait		Single hospital Unknown demographics n=15 ABM, n=9 AVMCohortLactate cutoff3.0mM
Sensitivity93.3%
Specificity88.9%
LR+8.40
LR-0.075
Uduman et al.
2000
UAE		Single hospital Median 5 vs. 18mo ABM vs. AVM n=23 ABM, n=42 comparator group including AVMCohortLactate cutoffNot reportedComparator group included not only AVM but also febrile seizures, GBS, peripheral neuropathies, and metabolic disorders
Sensitivity95.7%
Specificity92.9%
LR+13.4
LR-0.047
Vanprapar et al.
1983
Thailand		Single hospital Mean 25 vs. 9 vs. 46mo, 85% vs. 44% vs. 55% male ABM vs. presumed ABM vs. AVM n=13 ABM, n=9 presumed ABM, n=18 AVMCohortLactate cutoff3.9mM
Sensitivity90.9%
Specificity100%
LR+Infinity
LR-0.091

Comment(s)

Distinguishing acute bacterial meningitis from acute viral meningitis can be challenging. While a strategy of treating all possible cases of ABM with broad spectrum antibiotics until CSF cultures result is commonly employed, this strategy may lead to antibiotic resistance and unnecessary hospitalization. A combination of CSF white blood cell count/differential, CSF glucose, and CSF protein is commonly used to distinguish ABM from AVM. Unfortunately these tests all suffer from suboptimal sensitivity and specificity. The ideal test would be maximally sensitive and specific, rapidly available, and cheap. Despite not being commonly used, CSF lactate has been widely studied and appears to satisfy all three domains. The median cutoff used in the included studies was 3.4mM (~30mg/dL), with the majority of studies using enzymatic assays. As with most CSF tests including the gold-standard culture, CSF lactate's utility is somewhat limited in patients pretreated with antibiotics. In such cases, a lower cutoff should be used to maintain sufficient sensitivity.

Clinical Bottom Line

CSF lactate at a cutoff of 3.5mM appears to be both a sensitive and specific test for distinguishing acute bacterial meningitis from acute viral meningitis in non-neonatal children. With nearly 40 studies supporting its use over the past 40 years, it appears ready for routine use.

Level of Evidence

Level 1 - Recent well-done systematic review was considered or a study of high quality is available.

References

  1. Briem et al. Comparison between cerebrospinal fluid concentrations of glucose, total protein, chloride, lactate, and total amino acids for the differential diagnosis of patients with meningitis. Scand J Infect Dis 1983; 277-284
  2. Brook et al. Measurement of lactic acid in cerebrospinal fluid of patients with infections of the central nervous system. J Infect Dis 1978; 384-390
  3. Chen et al. The clinical diagnostic significance of cerebrospinal fluid d-lactate for bacterial meningitis. Clinica Chimica Acta 2012; 1512-1515
  4. Curtis et al. Cerebrospinal fluid lactate and the diagnosis of meningitis. J Infect 1981; 159-165
  5. Ellis and Oei Cerebrospinal fluid lactic acid: a valuable aid in the of diagnosis septic meningitis. J Indiana State Med Assoc 1979; 530-531
  6. Ferguson and Tearle Gas liquid chromatography in the rapid diagnosis of meningitis. J Clin Pathol 1977; 1163-1167
  7. Gastrin et al. Rapid diagnosis of meningitis with use of selected clinical data and gas-liquid chromatographic determination of lactate concentration in cerebrospinal fluid. J Infect Dis 1979; 529-533
  8. Hurd et al. Use of cerebrospinal fluid lactate levels in the diagnosis of bacterial meningitis. Am J Med Technol 1978; 11-13
  9. Jordan et al. CSF lactate in diseases of the CNS. Arch Intern Med 1983; 85-87
  10. Landaas and Von Der Lippe Chemical analyses for early differential diagnosis between bacterial and viral meningitis. Scand J Clin Lab Invest 1985; 525-529
  11. Lester et al. Value of CSF lactate in the differential diagnosis between bacterial meningitis and other diseases with meningeal involvement. Acta Pathol Microbiol Immunol Scand B 1985; 21-25
  12. Lindquist et al. Value of cerebrospinal fluid analysis in the differential diagnosis of meningitis: a study in 710 patients with suspected central nervous system infection. Eur J Clin Microbiol Infect Dis 1988; 374-380
  13. Mandal et al. How useful is cerebrospinal fluid lactate estimation in differential diagnosis of meningitis? J Infect 1983; 231-237
  14. Ponka et al. The differential diagnosis of bacterial and aseptic meningitis using cerebrospinal fluid laboratory tests. Infection 1983; 129-131
  15. de Almeida et al. Quantitation of cerebrospinal fluid lactic acid in infectious and non-infectious neurological diseases. Clin Chem Lab Med 2009; 755-761
  16. Berg et al. Cerebrospinal fluid lactate in the diagnosis of meningitis. Diagnostic value compared to standard biochemical methods. Scand J Infect Dis 1982; 111-115
  17. Huy et al. Cerebrospinal fluid lactate concentration to distinguish bacterial from aseptic meningitis: a systemic review and meta-analysis. Crit Care 2010; R240
  18. Sakushima et al. Diagnostic accuracy of cerebrospinal fluid lactate for differentiating bacterial meningitis from aseptic meningitis: A meta-analysis. J Infect 2011; 255-262
  19. Juarez Aragon et al. Assessment of five laboratory tests for differential diagnosis in bacterial and viral meningoencephalitides. Arch Invest Med 1979; 111-119
  20. Bland et al. Cerebrospinal fluid lactic acid level and pH in meningitis. Aids in differential diagnosis. Am J Dis Child 1974; 151-156
  21. Boon and Heng Bacterial meningitis and cerebrospinal fluid lactate. J Singapore Paediatr Soc 1978; 117-120
  22. Cameron et al. Cerebrospinal fluid lactate in meningitis and meningococcaemia. J Infect 1993; 245-252
  23. Controni and Rodriguez Diagnosis of pyogenic meningitis. Lancet 1976; 257
  24. Controni et al. Cerebrospinal fluid lactic acid levels in meningitis. J Pediatr 1977; 379-384
  25. Donald and Malan Cerebrospinal fluid lactate and lactate dehydrogenase activity in the rapid diagnosis of bacterial meningitis. S Afr Med J 1986; 39-42
  26. Dwivedi and Reddy Diagnostic use of cerebrospinal fluid lactic acid levels in meningitis. J Med 1983; 395-403
  27. Eross et al. Cerebrospinal fluid lactic acidosis in bacterial meningitis. Arch Dis Child 1981; 692-698
  28. Knight et al. Early (chemical) diagnosis of bacterial meningitis--cerebrospinal fluid glucose, lactate, and lactate dehydrogenase compared. Clin Chem 1981; 1431-1434
  29. Low et al. Cerebrospinal fluid lactate--its diagnostic value in septic meningitis. J Singapore Paediatr Soc 1986; 243-246
  30. Filho et al. Cerebrospinal fluid lactate level as a diagnostic biomarker for bacterial meningitis in children. Int J Emerg Med 2014; 14
  31. Murata and Uemura The diagnostic value of cerebrospinal fluid lactic acid levels in meningitis. Folia Psychiatr Neurol Jpn 1981; 175-179
  32. Nazir et al. Cerebrospinal fluid lactate: a differential biomarker for bacterial and viral meningitis in children. Jornal de Pediatria 2018; 88-92
  33. Nelson et al. The diagnostic and predictive value of cerebrospinal fluid lactate in children with meningitis. Its relation to current diagnostic methods. Acta Paediatr Scand 1986; 52-57
  34. Rutledge et al. Is the CSF lactate measurement useful in the management of children with suspected bacterial meningitis? J Pediatr 1981; 20-24
  35. Ruuskanen et al. CSF lactate in bacterial meningitis with minimal CSF abnormalities. Acta Paediatr Scand 1985; 292-293
  36. Pit et al. Cerebrospinal fluid lactic acid concentration in bacterial meningitis. Med J Malaysia 1987; 302-305
  37. Dashti et al. Diagnostic value of lactate, procalcitonin, ferritin, serum-C-reactive protein, and other biomarkers in bacterial and viral meningitis. Medicine 2017; e7637
  38. Shaltout Cerebrospinal fluid lactate is useful in differentiating viral from bacterial meningitis. Med Principles Pract 1989; 12-16
  39. Uduman et al. Haemophilus influenzae type b still remains a leading cause of meningitis among unvaccinated children--a prospective CSF analysis study. J Trop Pediatr 2000; 331-334
  40. Vanprapar et al. Cerebrospinal fluid lactate level in childhood purulent meningitis. Southeast Asian J Trop Med Public Health 1983; 186-190