current as of August 5, 2008. Online article and related content http://jama.ama-assn.org/cgi/content/full/296/16/2012 . 2006;296(16):2012-2022 (doi:10.1001/jama.296.16.2012) JAMA Sharon E. Straus; Kevin E. Thorpe; Jayna Holroyd-Leduc Results to Diagnose Bacterial Meningitis? How Do I Perform a Lumbar Puncture and Analyze the Correction Contact me if this article is corrected. Citations Contact me when this article is cited. This article has been cited 7 times. Topic collections Contact me when new articles are published in these topic areas. Clinical Examination Neurology; Diagnosis; Headache; Meningitis; Neurology, Other; The Rational Quality of Care; Quality of Care, Other; Infectious Diseases; Bacterial Infections; CME course Online CME course available. the same issue Related Articles published in . 2006;296(16):2050. JAMA Janet M. Torpy et al. Lumbar Puncture Related Letters . 2007;297(8):811. JAMA Sharon E. Straus et al. In Reply: . 2007;297(8):810. JAMA Stephen Piwinski et al. . 2007;297(8):810. JAMA Jeffrey L. Kaufman. How to Perform a Lumbar Puncture http://pubs.ama-assn.org/misc/permissions.dtl [email protected]Permissions http://jama.com/subscribe Subscribe [email protected]Reprints/E-prints http://jamaarchives.com/alerts Email Alerts at Northwestern University on August 5, 2008 www.jama.com Downloaded from
Transcript
current as of August 5, 2008. Online article and related content
Sharon E. Straus; Kevin E. Thorpe; Jayna Holroyd-Leduc
Results to Diagnose Bacterial Meningitis?How Do I Perform a Lumbar Puncture and Analyze the
Correction Contact me if this article is corrected.
Citations Contact me when this article is cited.
This article has been cited 7 times.
Topic collections
Contact me when new articles are published in these topic areas.Clinical Examination Neurology; Diagnosis; Headache; Meningitis; Neurology, Other; The Rational Quality of Care; Quality of Care, Other; Infectious Diseases; Bacterial Infections;
CME course Online CME course available.
the same issueRelated Articles published in
. 2006;296(16):2050.JAMAJanet M. Torpy et al. Lumbar Puncture
Related Letters
. 2007;297(8):811.JAMASharon E. Straus et al. In Reply:
. 2007;297(8):810.JAMAStephen Piwinski et al. . 2007;297(8):810.JAMAJeffrey L. Kaufman.
How Do I Perform a Lumbar Punctureand Analyze the Results to DiagnoseBacterial Meningitis?Sharon E. Straus, MD, MSc, FRCPCKevin E. Thorpe, MMathJayna Holroyd-Leduc, MD, FRCPC
PATIENT SCENARIOA previously healthy 70-year-oldwoman presents to the emergency de-partment with a 3-day history of fe-ver, confusion, and lethargy. She is un-able to cooperate with a full physicalexamination, but she has neck stiff-ness upon neck flexion. Her score onthe Glasgow Coma Scale is 13 (eye, 4;verbal, 4; motor, 5). The findings froma chest radiograph and urinalysis arenormal.1 You seek consent from herhusband to perform a lumbar punc-ture (LP).
Why Is This Diagnostic ProcedureImportant?
In a previous Rational Clinical Exami-nation article, Attia and colleagues1 dis-cussed the above scenario and recom-mended proceeding to LP for definitivetesting of the cerebrospinal fluid (CSF).Cerebrospinal fluid is a clear, color-less fluid that fills the ventricles andsubarachnoid space surrounding thebrain and spinal cord.2 Lumbar punc-ture allows this fluid to be sampled,facilitating the diagnosis of variousconditions.
Since it was first described byQuincke3 in 1891, the LP has become
See also Patient Page.
CME available online atwww.jama.com
Author Affiliations: Division of General Internal Medi-cine,UniversityofCalgary,Alberta(DrStraus);KnowledgeTranslation Program, University of Toronto/St Mi-chael’sHospital,Toronto,Ontario(DrsStrausandHolroyd-Leduc and Mr Thorpe); Department of Public HealthSciences, University of Toronto, Toronto, Ontario (MrThorpe); Division of General Internal Medicine, Univer-sityHealthNetwork,Toronto,Ontario(DrHolroyd-Leduc).
Corresponding Author: Sharon E. Straus, MD, MSc,FRCPC,FoothillsMedicalCentre140329thStNW,Cal-gary, Canada AB T2N 2T9 ([email protected]).The Rational Clinical Examination Section Editors:David L. Simel, MD, MHS, Durham Veterans AffairsMedical Center and Duke University Medical Center,Durham, NC; Drummond Rennie, MD, Deputy Edi-tor, JAMA.
Context Diagnostic lumbar punctures (LPs), commonly used to rule out meningitis,are associated with adverse events.
Objective To systematically review the evidence about diagnostic LP techniques thatmay decrease the risk of adverse events and the evidence about test accuracy of cere-brospinal fluid (CSF) analysis in adult patients with suspected bacterial meningitis.
Data Sources We searched the Cochrane Library, MEDLINE (using Ovid and PubMed)from 1966 to January 2006 and EMBASE from 1980 to January 2006 without lan-guage restrictions to identify relevant studies and identified others from the bibliog-raphies of retrieved articles.
Study Selection We included randomized trials of patients aged 18 years or olderundergoing interventions to facilitate a successful diagnostic LP or to potentially re-duce adverse events. Studies assessing the accuracy of biochemical analysis of the CSFfor possible bacterial meningitis were also identified.
Data Extraction Two investigators independently appraised study quality and ex-tracted relevant data. For studies of the LP technique, data on the intervention andthe outcome were extracted. For studies of the laboratory diagnosis of bacterial men-ingitis, data on the reference standard and test accuracy were extracted.
Data Synthesis We found 15 randomized trials. A random-effects model was usedfor quantitative synthesis. Five studies of 587 patients compared atraumatic needles withstandard needles and found a nonsignificant decrease in the odds of headache with anatraumatic needle (absolute risk reduction [ARR], 12.3%; 95% confidence interval [CI],−1.72% to 26.2%). Reinsertion of the stylet before needle removal decreased the risk ofheadache (ARR, 11.3%; 95% CI, 6.50%-16.2%). The combined results from 4 studiesof 717 patients showed a nonsignificant decrease in headache in patients who were mo-bilized after LP (ARR, 2.9%; 95% CI, −3.4 to 9.3%). Four studies on the accuracy ofbiochemical analysis of CSF in patients with suspected meningitis met inclusion criteria. ACSF–blood glucose ratio of 0.4 or less (likelihood ratio [LR], 18; 95% CI, 12-27]), CSFwhite blood cell count of 500/µL or higher (LR, 15; 95% CI, 10-22), and CSF lactatelevel of 31.53 mg/dL or more (�3.5 mmol/L; LR, 21; 95% CI, 14-32) accurately diag-nosed bacterial meningitis.
Conclusions These data suggest that small-gauge, atraumatic needles may de-crease the risk of headache after diagnostic LP. Reinsertion of the stylet before needleremoval should occur and patients do not require bed rest after the procedure. Futureresearch should focus on evaluating interventions to optimize the success of a diag-nostic LP and to enhance training in procedural skills.JAMA. 2006;296:2012-2022 www.jama.com
an important diagnostic tool, particu-larly when considering the diagnosis ofmeningitis. Evaluation of CSF can helpestablish a diagnosis and guide antimi-crobial therapy. Less commonly, LP isused as part of the diagnostic workupof patients with suspected subarach-noid hemorrhage, demyelinating dis-ease, and leptomeningeal metastasis.4,5
What Adverse Events Can ResultFrom an LP?
Bier was the first to report the tech-nique of spinal anesthesia and also pro-vided (through personal experience!)the first description of post-LP head-ache.6 Headache and backache are themost frequently reported adverse eventsassociated with LP. Headache can oc-cur in up to 60% of patients who un-dergo the procedure, although esti-mates vary due to differences ininclusion criteria and definitions ofheadache.7-9 Headache can be severe anddebilitating10 and is believed to occurbecause of CSF leakage through the du-ral puncture site.11 Backache is less com-mon but can occur in up to 40% of pa-tients following LP.7 Rare adverse eventsinclude cerebral herniation, intracra-nial subdural hemorrhage, spinal epi-dural hemorrhage, and infection.11
What Are the Contraindicationsto Performing an LP?
Clinicians often worry that an undetec-ted mass lesion or ventricular obstruc-tion causing raised intracranial pres-sure pose risks for cerebral herniationfollowing an LP.4 However, no conclu-sive evidence supports that the risk canbe reduced with universal neuroimag-ing prior to LP. Instead of universal neu-roimaging, clinicians can use the clini-cal examination to guide the decision toobtain neuroimaging. In a prospectivestudy, 113 patients were examined byinternal medicine residents (overseen byemergency physicians) prior to under-going computed tomography (CT) of thebrain and subsequent LP.12 The me-dian age of patients was 42 years, 36%were immunocompromised, and 46% ofpatients had altered mentation. Alteredmentation (likelihood ratio [LR], 2.2;
95% confidence interval [CI], 1.5-3.2),focal neurological finding (LR, 4.3; 95%CI, 1.9-10), and papilledema (LR, 11;95% CI, 1.1-115) increased the likeli-hood of an intracranial lesion.12 Over-all clinical impression (not defined in thestudy) was able to identify patients witha CT-defined contraindication to LP (LR,19; 95% CI, 4.8-43). In a second pro-spective study of 301 patients with sus-pected meningitis, 235 underwent a CTscan prior to LP.13 The mean age of pa-tients was 40 years (16% were �60years), 25% were immunocompro-mised, and 27% of patients had a Charl-son comorbidity score of more than 1.Patients were assessed clinically by anemergency physician or general inter-nist. The absence of a number of clini-cal features at baseline was able to iden-tify those who were unlikely to have anabnormal CT result (LR, 0.10; 95% CI,0.03-0.31). The absence of all of the fol-lowing baseline characteristics was as-sociated with this low LR: age 60 yearsor older, immunocompromised state,history of central nervous system dis-ease, and seizure within 1 week of pre-sentation. In addition, there could benone of the following physical exami-nation findings: abnormal level of con-sciousness, inability to answer 2 ques-tions correctly, inability to follow 2consecutive commands correctly, gazepalsy, abnormal visual fields, facial palsy,arm drift, leg drift, and abnormal lan-guage. Using the pretest probability ofan abnormal CT finding from this study(23.8%), the absence of all of these fea-tures would reduce the probability of anabnormal finding to 3.0%. The find-ings from these 2 studies have not beenvalidated prospectively in other inde-pendent populations.
Local infection at the puncture siteis also a contraindication to complet-ing an LP but this occurs infre-quently.4 More frequently, clinicians areconcerned about coagulation defectsand use of anticoagulants, which mayincrease the risk of epidural hemor-rhage. In 1 study of post-LP complica-tions, outcomes were compared in 166patients receiving anticoagulation with171 of those who were not receiving
therapy. There was a trend toward in-creased risk of paraparesis in the anti-coagulated patients (relative risk, 11.0;95% CI, 0.60-199) with 5 patients inthe anticoagulation group experienc-ing an adverse event compared withnone in the control group. In all pa-tients who experienced paraparesis, an-ticoagulation had been started withinan hour of the procedure.14 A survey of246 pediatric and adult neurology de-partment chairpersons and residencyprogram directors found that 45% of re-spondents ordered platelet and antico-agulation studies prior to LP.15 We wereunable to find any data evaluating thesafety of LP in patients with low plate-let counts. In a case series of 66 pa-tients with acute leukemia, patientswith lower platelet counts (�50�103/µL) had higher risk of a traumatic pro-cedure as defined by the presence in theCSF of more than 500 red blood cellsper high-powered field.16 However, LPswere not performed in patients withplatelet counts lower than 20�103/µLin this study.
We conducted a systematic review toidentify studies of interventions that en-hance the success of an LP and that mini-mize adverse events. Based on review ofthe evidence and its integration with ex-pert opinion, we provide a best-practice approach for LP in adults. Be-cause a clinician’s interpretation of theLP results is tightly coupled to the clini-cal examination findings, we also re-viewed the literature that addresses theaccuracy of common CSF tests for bac-terial meningitis. Tests for diagnosing vi-ral meningitis were not included in thisreview. Although there are other indi-cations for LP and CSF analysis, this ar-ticle focuses on CSF analysis for sus-pected bacterial meningitis because itrequires immediate action and is one ofthe most common diagnoses that gen-eralist physicians consider when per-forming this procedure.
METHODSSearches of the Cochrane Library,MEDLINE (using Ovid and PubMed)from 1966 to January 2006, andEMBASE from 1980 to January 2006
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
were completed to identify relevantstudies. The search strategy used theterms lumbar puncture, spinal punc-ture, dural puncture, headache, head-ach*, spinal needle*, cerebrospinal fluid,spinal fluid, and meningitis. Interven-tion studies were limited to random-ized controlled trials using the termsrandomized controlled trial, controlledclinical trial, clinical trial, random allo-cation, and random*. No language re-strictions were used. Additional ar-ticles were identified from searching thebibliographies of retrieved articles. De-tails on the search strategies are avail-able on request.
Randomized trials of patients (�18years) undergoing interventions to po-tentially reduce headache and back-ache at the time of diagnostic LP wereincluded. However, if no randomizedstudies of a particular intervention wereidentified, studies of lower quality—including cohort, case-control, and caseseries—were retrieved. Studies assess-ing patients undergoing LP during spi-nal anesthesia or myelography were ex-cluded because these procedures areclinically different from a diagnostic LP.Smaller amounts of fluid are removedduring spinal anesthesia and myelogra-phy than with diagnostic LP and fluidsare inserted during these other proce-dures. Moreover, the risk of headacheis greater with diagnostic LP than withspinal anesthesia.7 Interventions of in-terest included those that could be usedat the time of LP, such as immediate mo-bilization, atraumatic needles, and re-insertion of the stylet. We also at-tempted to identify studies that assessedthe impact of positioning of the patientand experience of the operator. The out-come of interest included headache oc-curring up to 7 days after LP.
To examine the accuracy of CSFanalysis in patients with suspected acutebacterial meningitis, we included stud-ies of predominantly adult popula-tions and those that described use of anappropriate reference standard (eg, CSFculture or bacterial antigen) in all pa-tients. In addition, primary data or ap-propriate summary statistics had to beavailable in the studies.
Two reviewers (S.E.S. and J.M.H-L.) independently reviewed and se-lected relevant publications that met theinclusion criteria from the search re-sults. Disagreements were resolved byconsensus. In cases of doubt, full-textarticles were retrieved for review anddiscussion. Full-text articles of all ab-stracts that met the inclusion criteriawere retrieved.
The 2 reviewers independently readall full-text articles to confirm that in-clusion criteria were met. The investi-gators also assessed study quality. Forintervention studies, a specially de-signed data collection form was used toextract data on study quality includ-ing the method of randomization, thepresence of blinding, and the methodused for outcome assessment. Data werealso extracted on the intervention andthe dichotomous outcome variable ofpost-LP headache. The minimum in-clusion criteria for randomized stud-ies of interventions to prevent adverseevents were the description of random-ization and the ability to extract rel-evant patient data. For studies of testaccuracy, data were extracted on the ref-erence standard, the presence of blind-ing, the index test, and the populationcharacteristics. Minimum inclusion cri-teria for studies of test accuracy in pa-tients with suspected meningitis werethe completion of an appropriate ref-erence standard in all patients and theability to extract relevant data. Differ-ences in assessment by the reviewerswere resolved through discussion, anda third investigator (K.E.T.) was avail-able if necessary.
For the intervention studies, statis-tical heterogeneity was assessed usingthe method described by Woolf.17 Arandom-effects model (DerSimonianand Laird) was used for quantitativedata. For the studies of test accuracy,LRs were calculated using the random-effects model. Statistical analyses wereconducted using R: A Language and En-vironment for Statistical Computing andthe rmeta contributed package. R is anopen-source dialect of the S language(S was developed by AT&T) that ismaintained by a core team (http://www
.r-project.org). A 2-tailed P value of�.05 was considered statistically sig-nificant.
RESULTSWe found 537 citations of potential in-terventions to optimize LP technique.Review of these led to retrieval of 22full-text articles for assessment, 15 ofwhich were subsequently identified forinclusion. Reasons for excluding trialswere lack of randomization (5 stud-ies18-22), repeat publication (1 study23),and inability to obtain outcomes data(1 study24). Studies were categorized byintervention including needle type,needle size, reinsertion of stylet, mo-bilization after LP, and use of supple-mental fluids. No studies of other in-terventions—such as positioning of thepatient during LP, direction of bevel,volume of CSF removed, or prophy-lactic use of an epidural blood patch—met the inclusion criteria.
Description of Studies
Fifteen randomized trials were identi-fied with sample sizes ranging from 44to 600 people. Eight studies had samplesizes of 100 patients or fewer.
Performing the Procedure
Experience of Operator. We were un-able to identify any randomized stud-ies that evaluated the impact of the ex-perience of the clinician performing LPson clinical outcomes. Some studies weidentified included experienced neu-rologists,23 whereas others involved stu-dents under the supervision of physi-cians.25 In a case series of LPs performedat an urban university-affiliated hospi-tal, the incidence of traumatic LP was15% using a definition of more than 400red blood cells per high-powered fieldand 10% using a definition of more than1000 red blood cells.26 However, thelevel of training and specialty of all phy-sicians were not recorded. One retro-spective study compared the inci-dence of traumatic LP at the end of theresident academic year when house-staff are more experienced with that atthe start of the next year when newhousestaff begin training. Using a cut-
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
off of 1000 red blood cells/µL, there wasno difference in risk of traumatic LP be-tween experienced housestaff (14%)and inexperienced housestaff (12%).27
In a prospective cohort of 501 patientswho underwent LP either by a nurse,physician, resident, or medical stu-dent, there was no significant differ-ence in the risk of post-LP headacheamong the 3 groups.28 We found no dataon the number of LPs required to dem-onstrate or maintain proficiency.
Positioning of Patient. We were un-able to identify any studies that evalu-ated the success of LP with different pa-tient positions or the impact of patientpositioning on the risk of adverseevents. One study assessed the inter-spinous distance to determine the im-pact of positioning. Measurement of theinterspinous distance was conducted in16 patients who were placed in 3 po-sitions (lateral recumbent with kneesto chest; sitting and bent forward overan adjustable bedside stand; and sit-ting with feet supported and chest rest-ing on the knees).29 The interspinousdistance was greatest when the patientwas placed in the sitting position withfeet supported.
Needle Choice and Number ofAttempts. Five studies with data from587 patients compared atraumaticSprotte or Pajunk needles with stan-dard Quincke needles (FIGURE 1) dur-ing diagnostic LP.30-34 One of these stud-ies32 described the randomizationmethod and 4 studies30-33 described theuse of blinded outcomes assessment.Three studies provided data for inten-tion-to-treat analysis.30-32
There was a nonsignificant decreasein the risk of headache among patientswho underwent diagnostic LP with anatraumatic needle (absolute risk reduc-tion [ARR], 12.3%; 95% CI, −1.72% to26.2%]). There was statistically signifi-cant heterogeneity among these trials(�2
4=13.3, P�.01). The heterogeneity ap-peared to be due primarily to the smallstudy (n = 61) by Lenaerts and col-leagues (FIGURE 2) with only 9 out-come events.30 One study32 includeddata on severe headache and found thisrisk significantly decreased with atrau-
matic needles (ARR, 23%; 95% CI, 6%-40%). There are no data available onhow often an introducer was used withatraumatic needles or its impact.
Three of these studies, which in-volved 296 patients, included data onthe number of attempts required tocomplete the LP when using an atrau-matic needle.31-33 There was no signifi-cant heterogeneity among these stud-ies (�2
2= 0.46, P = .80). There was anonsignificant increase in the risk of re-quiring 2 or more attempts when an
atraumatic needle was used (ARI, 4.9%;95% CI, −13% to 3.4%).31-33 One studyfound an increased risk of requiring 4attempts with an atraumatic needlecompared with standard needle (ARI,14%; 95% CI, 3.1%-25%).32 This studyalso included data on backache andfound no increased risk with the atrau-matic needle (ARR, 7.4%; 95% CI, −12%to 27%) despite requiring more at-tempts with an atraumatic needle.These data on backache and number ofattempts required with an atraumatic
Figure 1. Types of Lumbar Puncture Needles
StyletNeedle
Standard Needle
Atraumatic Needle
Two types of lumbar puncture needles are available—the atraumatic (Sprotte or Pajunk) needle and the stan-dard (Quincke) needle. Either the 22-gauge or 20-gauge atraumatic needle, with or without an introducer,can be used for diagnostic lumbar puncture. Use of an atraumatic needle compared with a standard needleand use of a 26-gauge standard needle compared with a 22-gauge standard needle have been shown to beassociated with reduced risk of headache after lumbar puncture.37,38
Figure 2. Atraumatic vs Standard Needles and Occurrence of Any Headache
No. of Patients With Headache/
Total No. of PatientsAtraumatic
NeedleBetter
StandardNeedleBetter
10.001.000.10
Odds Ratio
AtraumaticNeedle
StandardNeedleSource
Odds Ratio(95% Confidence Interval)
3/49 16/50Kleyweg et al,33 1998 0.14 (0.04-0.51)
7/26 2/35Lenaerts et al,30 1993 6.08 (1.14-32.28)
5/50 15/50Muller et al,31 1994 0.26 (0.09-0.78)
14/115 28/115Strupp et al,34 2001 0.43 (0.21-0.87)
21/49 31/48Thomas et al,32 2000 0.41 (0.18-0.93)
50/289 92/298Overall 0.46 (0.19-1.07)
The size of the data markers reflects the size of the study.
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
needle were secondary outcomes of thisstudy and should be further evaluatedin larger trials.
Recently published guidelines fromthe American Academy of Neurologysupport the use of atraumatic needleswhen completing diagnostic LPs to re-duce the risk of post-LP headache.35 Inan earlier version of this guideline, theyreported that LP trays containing theSprotte needle are the same price asthose containing the Quincke.36
One study of 100 patients com-pared use of a 26-gauge Quincke needlevs a 22-gauge Quincke needle.37 Blindedoutcomes observers were used and datawere provided for intention-to-treatanalysis. The risk of headache was sig-nificantly reduced with a smaller needle(ARR, 26%; 95% CI, 11%-40%).
Reinsertion of Stylet
Strupp and colleagues studied 600 pa-tients and compared the effects of rein-sertion of the stylet before removing theatraumatic needle with no reinser-
tion.38 No details were provided on themethod of randomization or the use ofblinded outcomes assessment. Fewer pa-tients who underwent LP with reinser-tion of the stylet experienced headache(ARR, 11%; 95% CI, 6.5%-16%). It ispostulated that a strand of arachnoidcould enter the needle along with the outflowing CSF and if the stylet is not re-placed, the strand may be threaded backthrough the dura during removal of theneedle, producing prolonged leakage ofthe CSF. By replacing the stylet beforeremoving the needle, the strand wouldbe pushed out and cut, reducing the riskof continued leakage and the resultingheadache.38
Bed Rest After the Procedure
Four studies with data from 717 pa-tients compared immediate mobiliza-tion with bed rest lasting 4 hours forreducing post-LP headache.25,39-41 Onestudy provided some details on themethod of randomization.25 This studyalso described use of blinded out-comes assessors. Two of the 3 studiesprovided data for an intention-to-treat analysis.25,40
There was no significant heterogene-ity among these studies (�2
3 = 1.04,P=.79). There was a nonsignificant de-crease in the risk of headache in pa-tients who were mobilized after LP (ARR,2.9%; 95% CI, ARI 3.4%-ARR 9.3%];FIGURE 3) Three studies looked at headpositioning during bed rest, but giventhat there was no significant effect from
bed rest, the results of these studieswere excluded from further analysis.42-45
Supplementary FluidsIn a study of 100 patients undergoingdiagnostic LP, Dieterich assessed the ef-fects of drinking 1.5 L vs 3 L of fluidsper day on risk of post-LP headache.45
No details were provided on the methodof randomization or the use of blindedoutcomes assessors. The number of out-comes events was too small to detect adifference between these groups (riskdifference, 0.0; 95% CI, ARI 18.8%-ARR 18.8%). Sudlow and Warlow46
conducted a systematic review of mo-bilization and fluids for preventingpost-LP headache and found no effecton post-LP headache.
Interpreting the Results
Maneuvers During the Procedure. Nor-mal resting CSF pressure is assumed tobe 60 to 180 mm of H2O or 6 to 14mm Hg.47,48 In the single identifiedstudy, CSF pressure changed little (�1.1mm of water) with flexion of the lowerextremities.49 Various maneuvers, suchas compressing the abdomen or thejugular vein (Queckenstedt’s maneu-ver50), can increase CSF pressure.51 Anobstruction to CSF flow prevents thenormal rise and fall in pressure (posi-tive Queckenstedt), but we were un-able to find any studies describing theaccuracy of this maneuver for detec-tion of CSF outflow obstruction.
Laboratory Tests
Although CSF samples can be sub-jected to various analyses, we focusedon test results immediately relevant anduseful to generalist physicians whenevaluating a patient suspected of hav-ing bacterial meningitis. Normal CSFvalues are listed in TABLE 1, but thesevalues may vary across different labo-ratories.52 We found 460 diagnostic ar-ticles in our literature search, and 6 metinclusion criteria (TABLE 2). One ex-pert suggests that the white blood cellcount be corrected for the presence ofred blood cells by subtracting 1 whiteblood cell from the total white bloodcell count in the CSF for every 700 red
Figure 3. Mobilization vs Bed Rest and Occurrence of Any Headache
No. of Patients With Headache/
Total No. of Patients
MobilizationBetter
Bed RestBetterMobilization Bed RestSource
Odds Ratio(95% Confidence Interval)
34/82 34/78Dieterich and Brandt,40 1985 0.92 (0.49-1.72)
4/26 2/23Johansson et al,41 1992 1.91 (0.32-11.54)
70/150 79/150Vilming et al,25 1998 0.79 (0.5-1.24)
15/104 19/104Vimala et al,39 1998 0.75 (0.36-1.58)
123/362 134/355Overall 0.84 (0.6-1.16)
10.001.000.10
Odds Ratio
The size of the data markers reflects sample size of the study.
Table 1. Reference Values for AdultCerebrospinal Fluid52
CerebrospinalFluid Variable Value
Pressure 60 to 180 mm H2O;6 to 14 mm Hg
Glucose 45-80 mg/dL (2.5-4.4mmol/L)
Protein 15-45 mg/dLBlood-glucose ratio �0.6White blood cell count �5/µLGram stain No organismsCytology No atypical cells
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
blood cells.2 He also states that a singlepolymorphonuclear cell in the CSF witha white blood cell count of less than 5µL is considered normal. Steele and col-leagues suggest rapid analysis of CSFand noted that neutrophil counts candecrease by 50% within 2 hours ofcollection.59
We found 3 studies that met the in-clusion criteria and that described theaccuracy of CSF Gram stain for diag-nosing bacterial meningitis.53-55 Noneof these studies appeared to be pro-spective. All studies reported sensitiv-ity of Gram stain and 1 study reportedspecificity of this test53 (TABLE 3). If bac-teria are seen on Gram stain, it helpsdiagnose bacterial meningitis but if thistest is negative, bacterial meningitis can-not be ruled out.
We identified 4 studies that met theinclusion criteria and that reported onthe accuracy of biochemical analysis ofCSF in patients with suspected centralnervous system infection55-58(Table 2and TABLE 4). Only 1 study of CSFwhite blood cell counts met our strictinclusion criteria. A CSF white bloodcell count of 500/µL or higher in-creases the likelihood of meningitis (LR,15; 95% CI, 10-22), whereas a countless than 500/µL lowers the likelihood(LR, 0.3; 95% CI, 0.2-0.4).56
A CSF–blood glucose ratio of 0.4 orless was accurate for diagnosing bac-terial meningitis (LR, 18; 95% CI, 12-27), whereas a normal CSF–blood glu-cose ratio made this diagnosis less likely(LR, 0.31; 95% CI, 0.21-0.45).56,57 ACSF lactate level of 31.53 mg/dL ormore (�3.5 mmol/L) was accurate fordiagnosing bacterial meningitis (LR, 21;95% CI, 14-3255-57; Table 4), whereasa CSF lactate level of less than 31.53mg/dL (�3.5 mmol/L) makes the di-agnosis of bacterial meningitis less likely(LR, 0.12; 95% CI, 0.07-0.23).
Prediction Models
Spanos and colleagues60 developed aprediction rule for diagnosing bacte-rial meningitis. This rule (TABLE 5) wasderived from a retrospective chart re-view of patients with a final diagnosisof acute meningitis. The sample was di-
vided into a derivation and validationset, but data from a large number ofcharts (94/214) in the derivation set hadmissing data and were excluded fromthe analysis. And, the technique usedfor CSF cell count changed during thestudy period, which could influence theresults of this study. The accuracy asmeasured by the area under the re-ceiver operating curve (AUC) was 0.97for the validation set.
Hoen and colleagues61 attempted tovalidate the above rule in a retrospec-tive review and used the same data togenerate their own decision rule(Table 5) that included 4 different clini-cal variables. In their validation of thework by Spanos and colleagues, the AUCwas 0.98 while that for their derivedequation was 0.99. In another retrospec-tive review of patients with meningitis,Leblebicioglu and colleagues62 assessedthe rules by Hoen and Spanos and theAUC was 0.99 and 0.95, respectively.
McKinney and colleagues63 obtainedsimilar results in their retrospectivereview.
In the only prospective validationthat we were able to identify, Baty andcolleagues64 assessed Hoen’s rule in asample of 109 patients aged 1 to 85years with acute community-acquiredmeningitis. Data are only available inpatients with bacterial and viral men-ingitis, and thus the specificity of themodel cannot be calculated. The sen-sitivity of their computed model was80% for the diagnosis of bacterial men-ingitis. For this decision rule to be rec-ommended in clinical practice, it needsto be validated prospectively in largerpopulations with broader diseasespectrum.65
Brivet and colleagues66 completed aretrospective study and found that thepresence of at least 1 sign of severity ofdisease at the time referral and a CSFneutrophil count of more than 1000/µL
Table 2. Studies Assessing Cerebrospinal Fluid Analysis in Patients With Suspected CentralNervous System Infection
Source Study Design Sample Size Age Reference Standard
Dunbar et al,53
1998Retrospective 2635 CSF
samplesAdults Positive CSF culture
Wasilauskas andHampton,54
1982
Retrospective 80 CSFsamples
Adults Positive CSF cultureor bacterial antigen
Lannigan et al,55
1980Cohort
(not clear ifprospective)
434 16-86 y Positive CSF culture
Lindquist et al,56
1988Prospective
cohort710 �2 mo but
majority adultsPositive CSF culture
or bacterial antigen
Briem,57 1983 Cohort(not clear ifprospective)
266 90% 15 y or older Positive CSF cultureor bacterial antigen
Komorowskiet al,58 1986
Retrospective 562 Adults Positive CSF culture
Abbreviation: CSF, cerebrospinal fluid.
Table 3. Accuracy of Cerebrospinal Gram Stain in Patients With Suspected BacterialMeningitis*
StudySensitivity, %
(95% CI)Specificity, %
(95% CI)
LikelihoodRatio for
Positive Test(95% CI)
LikelihoodRatio for
Negative Test(95% CI)
Dunbar et al,53 1998 86 (74-92.6) 100 (lower 95confidencelimit 99.7)
737 (230-2295) 0.14 (0.08-0.27)
Wasilauskas andHampton,54 1982
60 (47-71) . . . . . . . . .
Lannigan et al,55 1980 56 (34-75) . . . . . . . . .Abbreviation: CI, confidence interval.*Ellipses indicate that data are not available.
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
were predictors of bacterial meningi-tis. Severity was defined by the pres-ence of at least 1 of the following: al-tered consciousness, seizure, focalneurological findings, and shock. Be-cause this study was retrospective, rel-evant laboratory data was not avail-able for all patients. And, this model hasnot been prospectively validated in anindependent population.
How Should the ProcedureBe Performed?
Ideally, a successful LP should meet thefollowing criteria: (1) obtain suffi-cient CSF on the first attempt, (2) oc-curs without trauma (ie, CSF contain-ing �1000 red blood cells per highpowered field), (3) occurs with mini-mal discomfort to the patient duringand after the procedure, and, (4) oc-curs without serious adverse eventssuch as cerebral herniation. The fol-lowing description of the method toperform an LP considers the best avail-able evidence and expert opinion to fa-cilitate successful LP completion.
The procedure and its risks shouldbe explained to the patient and in-formed consent obtained if relevant inthe practice setting. The descriptionshould include how the procedure willbe performed, why it is being per-formed, what complications may oc-cur, and how these can be treated. Forexample, patients can be told that on
average, 6 out of 10 people may de-velop a transient headache after LP andthat up to 4 out of 10 people can ex-perience temporary backache. Pa-tients should be asked if they are aller-gic to any medications including localanesthetic. For the anxious patient,some experts suggest that a small doseof anxiolytic (eg, lorazepam) may begiven prior to the procedure if the pa-tient wishes.
In the absence of any focal neuro-logical findings, altered mentation orpapilledema, the LP can be performedwithout first completing a CT scan.12,13
If a CT scan is requested before the LPwhen bacterial meningitis is sus-pected, antibiotic therapy should bestarted immediately and should notawait completion of the CT scan. If pos-sible, blood cultures should be takenprior to starting antibiotics.
The LP is usually completed with thepatient in the lateral recumbent posi-tion with his/her back at the edge of thebed to minimize curving of the spine(FIGURE 4). Both legs should be flexedtoward the chest and the neck shouldalso be slightly flexed. The patient’sshoulders and pelvis should be verti-cal to the bed. In this position, an imagi-nary line connecting the patient’s pos-terior superior iliac crests would crossthe L4-L5 interspace (Figure 4). Lum-bar puncture can occur in the L3-L4,L4-L5, or L5-S1 interspace.67 It should
not be attempted at higher levels in or-der to avoid the conus medullaris. Thereis no evidence to guide the clinicianabout whether the L3-L4 or L4-L5 in-terspace is the optimal site for the ini-tial attempt, which is a topic high-lighted for future research. Theperformance of LP may occur at L5-S1because there are fewer nerve roots anda relatively larger interspace. Thespinous process superior to the cho-sen interspace should be palpated. Theneedle should be inserted about 1 cminferior to the tip of this process.68
Wearing sterile gloves and a mask,the clinician should cleanse the punc-ture site with an antiseptic solution byapplying it in a circular motion thatstarts at the center of where the punc-ture will occur. Sterile drapes can beapplied, leaving the puncture siteexposed. Palpate the identif iedspinous process again and with 2 to 3mL of local anesthetic (eg, lidocaine),infiltrate the patient’s skin and deepertissues allowing 1 to 2 minutes for thisto take effect. An atraumatic needledoes not have the same cutting edge asa standard needle so it may be prefer-able to use an introducer to puncturethe skin prior to insertion of theneedle if this needle type is used.Introduce the spinal needle (using thesame track that was used for the anes-thetic) and advance it horizontallywhile aiming toward the umbilicus toa depth of about 2 cm. If bone isencountered, withdraw the needle tothe subcutaneous position and reinsertat a slightly different angle. Continueto advance the needle until a pop isfelt, indicating penetration of the liga-mentum flavum. The needle shouldnow be in the subarachnoid space.When the stylet is withdrawn, clearfluid should drip. If no fluid emerges,rotate the needle to ensure that no flapof dura is blocking flow of CSF. Ifthere is still no fluid, reinsert the styletand advance the needle slightly, with-drawing the stylet after each move-ment. Pain radiating down either legindicates that the needle is too lateraland has touched nerve roots. If thisoccurs, immediately withdraw the
Table 4. Accuracy of Cerebrospinal Fluid Biochemical Analysis in Patients With SuspectedBacterial Meningitis
CSF Test
PositiveLikelihood
Ratio (95% CI)
NegativeLikelihood
Ratio (95% CI)
White blood cell count �500/µL56 15 (10-22) 0.30 (0.20-0.40)
needle almost to the skin, recheck thepatient’s position, and reinsert theneedle in the mid line. If this processfails, move down 1 interspace and tryagain. If this fails, the procedureshould be attempted by another per-son; alternatively it can be done underfluoroscopic guidance. If the LPattempt is unsuccessful with thepatient in the lateral position, it maybe attempted with the patient sittingupright.50 However, this position doesnot permit accurate measurement ofthe CSF pressure.50,69
WhenflowofCSFisseen,pressurecanbe measured by connecting a manom-eter directly to the needle or via a 2- or3-way stopcock. The 0 mark on themanometer should be held at the levelof thespinalneedleandthetubeheldver-tically. Normally, there will be variationin the pressure with respiration.
The manometer contents can be re-leased for collection and analysis. Ad-ditional CSF can be collected for the re-quired investigations. Typically, theexaminer prepares 3 to 4 collectiontubes. The initial CSF sample is placedinto a tube labeled for biochemistry, andtube number 2 is for bacterial studies(Gram stain, culture, and sensitivity).Cell counts can be done on tube num-ber 3, and the fourth tube can be usedfor cytology or for other tests that mightbe done when considering other diag-noses. These tests will vary dependingon the differential diagnosis for the in-dividual patient. Once CSF collection
has been completed, the stylet shouldbe reinserted prior to removal of theneedle. A bandage may be applied to thepuncture site and the patient allowedto ambulate.
How Should This ProcedureBe Taught or Learned?
Recent surveys of clinical practice sug-gest variation in LP technique. A 1996survey of senior registrars in all depart-ments of neurology and neurosurgeryin the United Kingdom found that 15%of respondents reported using atrau-matic needles and that 73% recom-mended bed rest for up to 6 hours af-ter completion of the LP.70 A similarsurvey of 2287 practicing neurolo-gists in the United States found that 2%used atraumatic needles.70 Neurolo-gists did not use atraumatic needles dueto lack of knowledge of them or be-cause the needles were not available foruse in their institution.71
The American Board of InternalMedicine recommends 3 to 5 LPs as aminimum standard for ensuring com-petence in completing LPs.72 How-ever, trainees report that they neededto complete 6 to 10 procedures to feelcomfortable with their performance.72
Based on the evidence reviewed inthis article, some targets could be sug-gested to help clinicians assess the qual-ity of their LP performance. A success-ful LP would be indicated by obtainingsufficient CSF fluid for analysis on thefirst attempt and by achieving risks of
post-LP headache of less than 60% andof backache of less than 40%. How-ever, these figures represent the aver-age risk of these events and thus can-not be considered as benchmarks.73
To date, there is little evidence toguide the teaching of this procedure.Simulators have been developed, butthey have not been rigorously evalu-ated.74 In one randomized trial of aWeb-based educational tool, 14 nov-ice trainees were randomized to re-ceive Web-based training or no train-ing.75 The Web-based training moduleprovided virtual simulation of an LP.Both groups completed pre- and post-LPs on a synthetic mannequin. Inves-tigators found an improvement in per-formance of LP with completion of thetraining module. There was no assess-ment of the procedural skills on pa-tients or of the risk of post-LP compli-cations. This evidence highlights theneed to evaluate teaching the perfor-mance of LP to ensure clinical compe-tence is achieved and maintained.
SCENARIO RESOLUTIONOur patient had a fever, confusion, andneck stiffness. Many clinicians mightproceed immediately to LP to rule outmeningitis. However, we weren’t ableto complete a full neurological exami-nation because our patient was unableto cooperate with the assessment. Asdiscussed in this article, evidence sug-gests that her altered mental state andage and our inability to confirm the ab-
Table 5. Studies That Developed and/or Validated Prediction Rules for Diagnosing Bacterial Meningitis*
Study Study DesignNo. of
Patients Age Model Accuracy
Spanos et al,60 1989 Retrospective chartreview
Derivationset, 120
�1 mo AUC 0.97
Hoen et al,61 1995 Retrospective chartreview
500 �1 mo AUC 0.99
Leblebicioglu et al,62
1996Retrospective chart
review40 NA Assessed models by Hoen and Spanos AUC 0.95 by Spanos§
AUC 0.99 by Hoen
McKinney et al,63 1994 Retrospective chartreview
170 �17 y Assessed model by Spanos AUC 0.98
Baty et al,64 2000 Prospective� 109 1-85 y Assessed model by Hoen Sensitivity 80%Abbreviations: AUC, area under the curve; CSF, cerebrospinal fluid; NA, not available; PABM, probability of acute bacterial meningitis; PMN, polymorphonuclear leukocytes.*Reference standard: bacterial meningitis is present when CSF, urine, or blood tests positive or when bacterial antigen is detected in blood, urine, or CSF.†Where L = 0.52 � number of months from August 1 −12.76 � CSF: blood glucose ratio (if ratio �0.6, use 0.6) � 0.341 � (PMNs in CSF � 106/1)0.333 � 2.29 � age � 2.79
(if age �1 y), −2.71 � age � 7.79 (if 1 y �age �2 y), −0.159 � age � 2.69 (if 2 y �age �22 y), or � 0.1 � age = 3.01 (if age �22 y).‡Where L = 32.13 � 10−4 � CSF PMN count (106/L) � 2.365 � CSF protein (g/L) � 0.6143 � blood glucose (mmol/L) � 0.2086 � leukocyte count (109/L) −11.§Note the units used in the equation: glucose 1 mg/dL. To convert to mmol/L, multiply by 0.0555.�Excluded those with encephalitis and in those for whom the diagnosis was obvious (eg, cloudy CSF or positive Gram stain).
1(1 + e−L)
PABM = †
1(1 + e−L)
PABM = ‡
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
sence of a focal neurological finding areappropriate indications for a CT scanto rule out an intracranial lesion. Thissituation describes a common clinical
conundrum even though the inci-dence of intracranial lesions with suchpresentations is low. Because our clini-cal findings were strongly suggestive of
meningitis, we obtained blood work in-cluding complete blood count, glu-cose, and cultures and initiated imme-diate antibiotic therapy. We decided to
Figure 4. Anatomical Considerations During Lumbar Puncture
SpinalCord
ConusMedullaris
CaudaEquina
Cauda Equina
CerebrospinalFluid
Vertebral Body
L4
L3
L2
L1
L5
2 cm
Patient Position and Identification of Bony LandmarksA
Hand Position, Needle Direction, and Anatomical Relationships B
Line Between Superior Borderof the Posterior Iliac Crests
Intersecting L4 Spinous Process
Dura Mater
Arachnoid
LigamentumFlavum
EpiduralSpace
Nerve Root
Supraspinous Ligament(Overlying Spinous Process)
InterspinousLigament
Lumbar PunctureNeedle
SpinousProcess
SubarachnoidSpace
T O P V I E W
Possible Needle Insertion Sites
L4L3 L5 S1
L4
1 cm
Vertical Alignment ofShoulders and Pelvis
SUPER
IOR
ANTER IO R
Lumbar puncture is usually performed with the patient in the lateral recumbent position. To avoid rotation of the vertebral column, align the patient’s shoulders andpelvis in a plane perpendicular to the bed. A line connecting the superior border of the posterior iliac crests intersects the L4 spinous process or the L4-L5 interspace.Insert the lumbar puncture needle in the midline of the L3-L4, L4-L5 (most commonly), or L5-S1 vertebral interspace. These interspaces are below the end of the spinalcord, which terminates at the level of L1. Angle the needle towards the patient’s umbilicus and advance it slowly. The needle will penetrate the ligamentum flavum,dura, and arachnoid to enter the subarachnoid space, where cerebrospinal fluid is located.
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
obtain a CT scan (on which no mass le-sion was noted) before performing anLP. Given her confusion and uncer-tain neurological status, we suggestedambulation with assistance. We used anatraumatic needle and completed theprocedure with the patient in the leftlateral recumbent position. Cerebrospi-nal fluid was sent for cell count, Gramstain, culture, protein, glucose, and lac-tate. Results showed a white blood cellcount of 5000/µL and a CSF–blood-glucose ratio of 0.2, helping us to di-agnose bacterial meningitis.
BOTTOM LINELumbar punctures to assess meningi-tis in adults should be performed un-der sterile conditions with the patientplaced in the lateral recumbent posi-tion and their knees flexed. Alterna-tively, the patient could sit up and leanforward with his/her feet supported toincrease the interspinous space. Formost patients, a CT scan of the head torule out mass lesion is not required be-fore the LP and the clinical examina-tion can guide the decision about neu-roimaging.
The following procedures may de-crease the risk of post-LP headache:
1. Use of small-gauge atraumaticneedles for diagnostic LPs is preferredbut may require more needle pass at-tempts,
2. Reinsertion of the stylet prior tothe removal of the spinal needle,
3. Mobilization of patients after com-pleting the LP.
In patients suspected of having bac-terial meningitis, the following labora-tory tests can be considered along withappropriate cultures, Gram stain and se-rologic studies:
1. CSF–blood glucose ratio of 0.4 orless (LR positive, 18; 95% CI, 12-27; LRnegative, −0.31; 95% CI, 0.21-0.45).
2. CSF white blood cell count of500/µL or higher (LR positive, 15; 95%CI, 10-22; LR negative, 0.30; 95% CI,0.20-0.40).
3. CSF lactate level of 31.5 mg/dL orless (�3.5 mmol/L; LR positive, 21;95% CI, 14-32; LR negative, 0.12; 0.07-0.23).
Author Contributions: Dr Straus had full access to allof the data in the study and takes responsibility forthe integrity of the data and the accuracy of the dataanalysis.Study concept and design: Straus.Acquisition of data: Straus, Holroyd-Leduc.Analysis and interpretation of data: Straus, Thorpe,Holroyd-Leduc.Drafting of the manuscript: Straus.Critical revision of the manuscript for important in-tellectual content: Straus, Thorpe, Holroyd-Leduc.Statistical analysis: Straus, Thorpe.Administrative, technical, or material support: Straus.Financial Disclosures: None reported.Funding/Support: Dr Straus is supported by a Tier 2Canada Research Chair; Mr Thorpe and Dr Holroyd-Leduc are supported by the Knowledge TranslationProgram at the University of Toronto.Role of the Sponsor: There was no external fundingobtained for the design and conduct of the study, thecollection, management, analysis, or for the interpre-tation of the data; or for the preparation, review, orapproval of the manuscript.Acknowledgment: We thank Darlyne Rath, LaurePerrier, and Desiree Chanderbhan for their assis-tance in the search and retrieval of relevant articlesand the preparation of the manuscript as part of theirnormal duties. Richard Bedlack, MD, PhD, Division ofNeurology, and Sheri Keitz, MD, PhD, Division of Gen-eral Internal Medicine, Duke University, Durham, NC,and Kaveh Shojania, MD, Department of Medicine,University of Ottawa, Ottawa, Ontario, who pro-vided useful feedback and suggestions on earlier draftsof the manuscript. We also thank Janis Miyasaki, MD,FRCPC, Department of Neurology, for comments onthe figure. They received no compensation for theirreviews of the manuscript.
REFERENCES
1. Attia J, Hatala R, Cook DJ, Wong JG. Does this adultpatient have acute meningitis? JAMA. 1999;282:175-181.2. Conly JM, Ronald AR. Cerebrospinal fluid as a di-agnostic body fluid. Am J Med. 1983;75:102-107.3. Quincke H. Ueber Hydrocephalus. Verhandl CongInn Med. 1891;10:321.4. Report of the Quality Standards Subcommittee ofthe American Academy of Neurology. Practice pa-rameters: lumbar puncture. Neurology. 1993;43:625-627.5. Health and Public Policy Committee, American Col-lege of Physicians. The diagnostic spinal tap: positionpaper. Ann Intern Med. 1986;104:880-885.6. Bier A. Versuche uber Cocainisirung desRuckenmarkes. Dt Zeitschr Chirurg. 1899;51:361-369.7. Flaatten H, Krakenes J, Vedeler C. Post-dural punc-ture related complications after diagnostic lumbar punc-ture, myelography, and spinal anaesthesia. Acta Neu-rol Scand. 1998;98:445-451.8. Kuntz KM, Kokmen E, Stevens JC, et al. Post-lumbar puncture headaches: experience in 501consecutive procedures. Neurology. 1992;42:1884-1887.9. Hindley NJ, Jobst KA, King E, et al. High accept-ability and low morbidity of diagnostic lumbar punc-ture in elderly subjects of mixed cognitive status. ActaNeurol Scand. 1995;91:405-411.10. Tohmo H, Vuorinen E, Muuronen A. Prolongedimpairment in activities of daily living due to postdu-ral puncture headache after diagnostic lumbarpuncture. Anaesthesia. 1998;53:299-302.11. Evans RW. Complications of lumbar puncture.Neurol Clin. 1998;16:83-105.12. Gopal AK, Whitehouse JD, Simel DL, Corey GR.Cranial computed tomography before lumbar puncture.Arch Intern Med. 1999;159:2681-2685.13. Hasbun R, Abrahams J, Jekel J, Quagliarello VJ.
Computed tomography of the head before lumbarpuncture in adults with suspected meningitis. N EnglJ Med. 2001;345:1727-1733.14. Ruff RL, Dougherty JH. Complications of lumbarpuncture followed by anticoagulation. Stroke. 1981;12:879-881.15. Waldman W, Laureno R. Precautions for lumbarpuncture: a survey of neurologic educators. Neurology.1999;52:1296-1299.16. Vavricka SR, Walter RB, Irani S, et al. Safety oflumbar puncture for adults with acute leukemia andrestrictive prophylactic platelet transfusion. AnnHematol. 2003;82:570-573.17. Woolf B. On estimating the relation between bloodgroup and disease. Ann Hum Genet. 1955;19:251-253.18. Carbaat PA, Van Crevel H. Lumbar puncture head-ache: controlled study on the preventive effect of 24hours’ bed rest. Lancet. 1981;2:1133-1135.19. Aamodt A, Vedeler C. Complications after LP re-lated to needle type: pencil-point versus Quincke. ActaNeurol Scand. 2001;103:396-398.20. Braune HJ, Huffmann G. A prospective double-blind clinical trial comparing the sharp Quincke needle(22G) with an “atraumatic” needle (22G) in the in-duction of post-lumbar puncture headache. Acta Neu-rol Scand. 1992;86:50-54.21. Kovanen J, Sulkava R. Duration of postural head-ache after lumbar puncture: effect of needle size.Headache. 1986;26:224-226.22. Spriggs DA, Burn DJ, Cartlidge NE, Bates D. Is bedrest useful after diagnostic lumbar puncture? Post-grad Med J. 1992;68:581-583.23. Strupp M, Brandt T. Should one reinsert thestylet during lumbar puncture? N Engl J Med. 1997;336:1190.24. Engelhardt A, Oheim S, Neundorfer B. Post-lumbar puncture headache: experiences with Sprotte’satraumatic needle. Cephalalgia. 1992;12:259.25. Vilming ST, Schrader H, Monstad I. Post-lumbarpuncture headache: the significance of body posture.Cephalalgia. 1988;8:75-78.26. Shah KH, Richard KM, Nicholas S, Edlow JA. In-cidence of traumatic lumbar puncture. Acad EmergMed. 2003;10:151-154.27. Williams C, Fost N. Doctors’ experience and trau-matic lumbar punctures. Lancet. 1994;344:1086-1087.28. Kuntz KM, Kokmen E, Stevens JC, et al. Post-lumbar puncture headaches: experience in 501 con-secutive procedures. Neurology. 1992;42:1884-1887.29. Sandoval M, Shestak W, Sturmann K, Hsu C. Op-timal patient position for lumbar puncture, measuredby ultrasonography. Emerg Radiol. 2004;10:179-181.30. Lenaerts M, Pepin J-L, Tombu S, Schoenen J. Nosignificant effect of an “atraumatic” needle on inci-dence of post-lumbar puncture headache or trau-matic tap. Cephalalgia. 1993;13:296-297.31. Muller B, Adelt K, Reichmann H, Toyka K. Atrau-matic needle reduces the incidence of post-lumbarpuncture syndrome. J Neurol. 1994;241:376-380.32. Thomas SR, Jamieson DRS, Muir KW. Random-ized controlled trial of atraumatic versus standardneedles for diagnostic lumbar puncture. BMJ. 2000;321:986-990.33. Kleyweg RP, Hertzberger LI, Carbaat PA. Signifi-cant reduction in post-lumbar puncture headache us-ing an atraumatic needle. Cephalalgia. 1998;18:635-637.34. Strupp M, Schueler O, Straube A, et al. “Atrau-matic” Sprotte needle reduces the incidence ofpost-lumbar puncture headaches. Neurology. 2001;57:2310-2312.35. Armon C, Evans RW. Addendum to assessment:prevention of post-lumbar puncture headaches.Neurology. 2005;65:510-512.36. Evans RW, Armon C, Frohman EM, Goodin DS.Assessment: prevention of post-lumbar punctureheadaches. Neurology. 2000;55:909-914.37. Tourtellotte WW, Henderson WG, Tucker RP,
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
Gilland O, Walker JE, Kokeman E. A randomizeddouble-blind clinical trial comparing the 22 versus 26gauge needle in the production of the post-lumbarpuncture syndrome in normal individuals. Headache.1972;12:73-78.38. Strupp M, Brandt T, Muller A. Incidence of post-lumbar puncture syndrome reduced by reinserting thestylet: a randomized prospective study of 600 patients.J Neurol. 1998;245:589-592.39. Vimala J, Peter JV, Jeyaseelan L, et al. Post lum-bar puncture headache: is bed rest essential? J AssocPhysicians India. 1998;46:930-932.40. Dieterich M, Brandt T. Is obligatory bed rest af-ter lumbar puncture obsolete? Eur Arch Psychiatry Neu-rol Sci. 1985;235:71-75.41. Johannsson G, Bertholds E, Hakansson S. Plan-lage efter lumbalpunktion onodigt och tidskravande.Lakartidningen. 1992;98:4030-4031.42. Handler CE, Perkin GD, Smith FR, Rose FC. Pos-ture and lumbar puncture headache: a controlled trialin 50 patients. J R Soc Med. 1982;75:404-408.43. Smith FR, Perkin GD, Rose FC. Posture and head-ache after lumbar puncture. Lancet. 1980;1:1245.44. Hilton-Jones D, Harrad RA, Gill MW, Warlow CP.Failure of postural manoeuvers to prevent lumbar punc-ture headache. J Neurol Neurosurg Psychiatry. 1982;45:743-746.45. Dieterich M, Brandt T. Incidence of post-lumbarpuncture headache is independent of daily fluid in-take. Eur Arch Psychiatry Neurol Sci. 1988;237:194-196.46. Sudlow C, Warlow C. Posture and fluids for pre-venting post-dural puncture headache [Cochrane Li-brary on CD-ROM]. Oxford, England, Update Soft-ware; 2005;Issue 1.47. Sternbach G. Lumbar puncture. J Emerg Med.1985;2:199-203.48. Clough C, Pearce JMS. LProcedures in practice:lumbar puncture. Br Med J. 1980;280:297-299.49. Abbrescia KL, Brabson TA, Dalsey WC, et al. Theeffect of lower-extremity position on cerebrospinal fluidpressures. Acad Emerg Med. 2001;8:8-12.50 . Queckens ted t H. Zur D iagnose de rRuckenmarkskompression. Dtsch Z Nervenheilkd.1916;55:325-333.51. Deliyannakis E. Influence of the position of the
head on the cerebrospinal fluid pressure. Mil Med.1971;136:370-372.52. Smith SV, Forman DT. Laboratory analysis of cere-brospinal fluid. Clin Lab Sci. 1994;7:32-38.53. Dunbar S, Eason RE, Musher DM, Clarridge J. Mi-croscopic examination and broth culture of cerebrospi-nal fluid in diagnosis of meningitis. J Clin Microbiol.1998;36:1617-1620.54. Wasilauskas BL, Hampton KD. Determination ofbacterial meningitis: a retrospective study of 80 cere-brospinal fluid specimens evaluated by four in vitromethods. J Clin Microbiol. 1982;16:531-535.55. Lannigan R, MacDonald MA, Marrie TJ, HaldaneEV. Evaluation of cerebrospinal fluid lactic acid levelsas an aid in differential diagnosis of bacterial and vi-ral meningitis in adults. J Clin Microbiol. 1980;11:324-327.56. Lindquist L, Linne T, Hansson LO, et al. Value ofcerebrospinal fluid analysis in the differential diagno-sis of meningitis: a study of 710 patients with sus-pected central nervous system infection. Eur J Clin Mi-crobiol Infect Dis. 1988;7:374-380.57. Briem H. Comparison between cerebrospinal fluidconcentrations of glucose, total protein, chloride, lac-tate, and total amino acids for the differential diag-nosis of patients with meningitis. Scand J Infect Dis.1983;15:277-284.58. Komorowski RA, Farmer SG, Knox KK. Compari-son of cerebrospinal fluid C-reactive protein and lac-tate for diagnosis of meningitis. J Clin Microbiol. 1986;24:982-985.59. Steele RW, Marmer DJ, O’Brien MD, et al. Leu-kocyte survival in cerebrospinal fluid. J Clin Microbiol.1986;23:965-966.60. Spanos A, Harrell FE, Durach DT. Differential di-agnosis of acute meningitis. JAMA. 1989;262:2700-2707.61. Hoen B, Viel JF, Paquot C, et al. Multivariate ap-proach to differential diagnosis of acute meningitis.Eur J Clin Microbiol Infect Dis. 1995;14:267-274.62. Leblebicioglu H, Esen S, Bedir A, et al. The valid-ity of Spanos’ and Hoen’s models for differential di-agnosis of meningitis. Eur J Clin Microbiol Infect Dis.1996;15:252-254.63. McKinney WP, Heudebert GR, Harper SA, et al.Validation of a clinical prediction rule for the differ-
ential diagnosis of acute meningitis. J Gen Intern Med.1994;9:8-12.64. Baty V, Viel JF, Schuhmacher J, et al. Prospectivevalidation of a diagnosis model as an aid to therapeu-tic decision-making in acute meningitis. Eur J Clin Mi-crobiol Infect Dis. 2000;19:422-426.65. McGinn TG, Guyatt GH, Wyer PC, et al; Evidence-Based Medicine Working Group. Users’ guides to themedical literature. XXII: how to use articles about clini-cal decision rules. JAMA. 2000;284:79-84.66. Brivet FG, Ducuing S, Jacobs F, et al. Accuracy ofclinical presentation for differentiating bacterial fromviral meningitis in adults: a multivariate approach.Intensive Care Med. 2005;31:1654-1660.67. Boon JM, Abrahams PH, Meiring JH, Welch T.Lumbar punctures: anatomical review of a clinical skill.Clin Anat. 2004;17:544-553.68. Eskey CJ, Ogilvy CS. Fluoroscopy-guided lum-bar puncture: decreased frequency of traumatic tapand implications for the assessment of CT-negativeacute subarachnoid hemorrhage. AJNR Am JNeuroradiol. 2001;22:571-576.69. Loman J, Myerson A, Goldman D. Effects of al-terations in posture on the cerebrospinal fluid pressure.Arch Neurol Psych. 1934:1279-1295.70. Serpell MG, Haldane GJ, Jamieson RS, Carson D.Prevention of headache after lumbar puncture: ques-tionnaire survey of neurologists and neurosurgeonsin United Kingdom. BMJ. 1998;316:1709-1710.71. Birnbach DJ, Kuroda MM, Sternman D, Thys DM.Use of atraumatic spinal needles among neurologistsin the United States. Headache. 2001;41:385-390.72. Hicks CM, Gonzales R, Morton MT, et al. Proce-dural experience and comfort level in internalmedicine trainees. J Gen Intern Med. 2000;15:716-722.73. Kiefe CI, Weissman NW, Allison JJ, et al. Identi-fying achievable benchmarks of care: concepts andmethodology. Int J Qual Health Care. 1998;10:443-447.74. Gorman P, Krummel T, Webster R, et al. A pro-totype haptic lumbar puncture simulator. Stud HealthTechnol Inform. 2000;70:106-109.75. Moorthy K, Jiwanji M, Shah J, et al. Validationof a web-based training tool for lumbar puncture. StudHealth Technol Inform. 2003;94:219-225.
DIAGNOSING BACTERIAL MENINGITIS BY LUMBAR PUNCTURE
