[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.
Nonculture Cerebrospinal Fluid (CSF) Infection Testing Frequency Over Time
Nonculture Cerebrospinal Fluid (CSF) Infection Testing Frequency Over Time

Testing frequency did vary by season but has been consistent annually for at least the past 9 to 10 years. LP indicates lumbar puncture.

Figure 2.
Flow Diagram
Flow Diagram

Lumbar puncture (LP) specimens with normal cell counts were common in immune-competent children of all ages but were rarely associated with a positive result from a nonculture cerebrospinal fluid infection test.

Table 1.  
Characteristics of All Children Who Underwent Lumbar Puncture Procedure
Characteristics of All Children Who Underwent Lumbar Puncture Procedure
Table 2.  
Frequency of Nonculture Cerebrospinal Fluid Infection Tests and Positive Yield
Frequency of Nonculture Cerebrospinal Fluid Infection Tests and Positive Yield
Table 3.  
Clinical and CSF Characteristics of Children With Normal Cell Counts and Nonculture CSF Infection Testing With a Positive Result
Clinical and CSF Characteristics of Children With Normal Cell Counts and Nonculture CSF Infection Testing With a Positive Result
1.
Arcara  K, Tschudy  M.  The Harriet Lane Handbook. 19th ed. Maryland Heights, MO: Mosby; 2011.
2.
Mitchell  PS, Espy  MJ, Smith  TF,  et al.  Laboratory diagnosis of central nervous system infections with herpes simplex virus by PCR performed with cerebrospinal fluid specimens.  J Clin Microbiol. 1997;35(11):2873-2877.PubMedGoogle Scholar
3.
Iturriza-Gómara  M, Megson  B, Gray  J.  Molecular detection and characterization of human enteroviruses directly from clinical samples using RT-PCR and DNA sequencing.  J Med Virol. 2006;78(2):243-253. doi:10.1002/jmv.20533PubMedGoogle ScholarCrossref
4.
Avery  RA, Frank  G, Eppes  SC.  Diagnostic utility of Borrelia burgdorferi cerebrospinal fluid polymerase chain reaction in children with Lyme meningitis.  Pediatr Infect Dis J. 2005;24(8):705-708. doi:10.1097/01.inf.0000172903.14077.4cPubMedGoogle ScholarCrossref
5.
Marton  KI, Gean  AD.  The spinal tap: a new look at an old test.  Ann Intern Med. 1986;104(6):840-848. doi:10.7326/0003-4819-104-6-840PubMedGoogle ScholarCrossref
6.
Tunkel  AR, Hartman  BJ, Kaplan  SL,  et al.  Practice guidelines for the management of bacterial meningitis.  Clin Infect Dis. 2004;39(9):1267-1284. doi:10.1086/425368PubMedGoogle ScholarCrossref
7.
Heim  A, Ebnet  C, Harste  G, Pring-Åkerblom  P.  Rapid and quantitative detection of human adenovirus DNA by real-time PCR.  J Med Virol. 2003;70(2):228-239. doi:10.1002/jmv.10382PubMedGoogle ScholarCrossref
8.
Sanchez  JL, Storch  GA.  Multiplex, quantitative, real-time PCR assay for cytomegalovirus and human DNA.  J Clin Microbiol. 2002;40(7):2381-2386. doi:10.1128/JCM.40.7.2381-2386.2002PubMedGoogle ScholarCrossref
9.
Niesters  HG, van Esser  J, Fries  E, Wolthers  KC, Cornelissen  J, Osterhaus  AD.  Development of a real-time quantitative assay for detection of Epstein-Barr virus.  J Clin Microbiol. 2000;38(2):712-715.PubMedGoogle Scholar
10.
Kares  S, Lönnrot  M, Vuorinen  P, Oikarinen  S, Taurianen  S, Hyöty  H.  Real-time PCR for rapid diagnosis of entero- and rhinovirus infections using LightCycler.  J Clin Virol. 2004;29(2):99-104. doi:10.1016/S1386-6532(03)00093-3PubMedGoogle ScholarCrossref
11.
Gautheret-Dejean  A, Manichanh  C, Thien-Ah-Koon  F,  et al.  Development of a real-time polymerase chain reaction assay for the diagnosis of human herpesvirus-6 infection and application to bone marrow transplant patients.  J Virol Methods. 2002;100(1-2):27-35. doi:10.1016/S0166-0934(01)00390-1PubMedGoogle ScholarCrossref
12.
Kessler  HH, Mühlbauer  G, Rinner  B,  et al.  Detection of herpes simplex virus DNA by real-time PCR.  J Clin Microbiol. 2000;38(7):2638-2642.PubMedGoogle Scholar
13.
Nix  WA, Maher  K, Johansson  ES,  et al.  Detection of all known parechoviruses by real-time PCR.  J Clin Microbiol. 2008;46(8):2519-2524. doi:10.1128/JCM.00277-08PubMedGoogle ScholarCrossref
14.
Maertzdorf  J, Wang  CK, Brown  JB,  et al.  Real-time reverse transcriptase PCR assay for detection of human metapneumoviruses from all known genetic lineages.  J Clin Microbiol. 2004;42(3):981-986. doi:10.1128/JCM.42.3.981-986.2004PubMedGoogle ScholarCrossref
15.
Hardegger  D, Nadal  D, Bossart  W, Altwegg  M, Dutly  F.  Rapid detection of Mycoplasma pneumoniae in clinical samples by real-time PCR.  J Microbiol Methods. 2000;41(1):45-51. doi:10.1016/S0167-7012(00)00135-4PubMedGoogle ScholarCrossref
16.
Hibbitts  S, Rahman  A, John  R, Westmoreland  D, Fox  JD.  Development and evaluation of NucliSens basic kit NASBA for diagnosis of parainfluenza virus infection with ‘end-point’ and ‘real-time’ detection.  J Virol Methods. 2003;108(2):145-155. doi:10.1016/S0166-0934(02)00268-9PubMedGoogle ScholarCrossref
17.
Watzinger  F, Suda  M, Preuner  S,  et al.  Real-time quantitative PCR assays for detection and monitoring of pathogenic human viruses in immunosuppressed pediatric patients.  J Clin Microbiol. 2004;42(11):5189-5198. doi:10.1128/JCM.42.11.5189-5198.2004PubMedGoogle ScholarCrossref
18.
Jordan  JA.  Identification of human parvovirus B19 infection in idiopathic nonimmune hydrops fetalis.  Am J Obstet Gynecol. 1996;174(1, pt 1):37-42. doi:10.1016/S0002-9378(96)70370-8PubMedGoogle ScholarCrossref
19.
Pevenstein  SR, Williams  RK, McChesney  D, Mont  EK, Smialek  JE, Straus  SE.  Quantitation of latent varicella-zoster virus and herpes simplex virus genomes in human trigeminal ganglia.  J Virol. 1999;73(12):10514-10518.PubMedGoogle Scholar
20.
Graf  EH, Farquharson  MV, Cárdenas  AM.  Comparative evaluation of the FilmArray meningitis/encephalitis molecular panel in a pediatric population.  Diagn Microbiol Infect Dis. 2017;87(1):92-94. doi:10.1016/j.diagmicrobio.2016.09.022PubMedGoogle ScholarCrossref
21.
Jafri  SK, Kumar  R, Ibrahim  SH.  Subacute sclerosing panencephalitis - current perspectives.  Pediatric Health Med Ther. 2018;9:67-71. doi:10.2147/PHMT.S126293PubMedGoogle ScholarCrossref
22.
Tang  YW, Hibbs  JR, Tau  KR,  et al.  Effective use of polymerase chain reaction for diagnosis of central nervous system infections.  Clin Infect Dis. 1999;29(4):803-806. doi:10.1086/520439PubMedGoogle ScholarCrossref
23.
Hiscock  H, Neely  RJ, Warren  H, Soon  J, Georgiou  A.  Reducing unnecessary imaging and pathology tests: a systematic review.  Pediatrics. 2018;141(2):e20172862. doi:10.1542/peds.2017-2862PubMedGoogle ScholarCrossref
24.
Gawande  AA, Colla  CH, Halpern  SD, Landon  BE.  Avoiding low-value care.  N Engl J Med. 2014;370(14):e21. doi:10.1056/NEJMp1401245PubMedGoogle ScholarCrossref
25.
Hanson  KE, Alexander  BD, Woods  C, Petti  C, Reller  LB.  Validation of laboratory screening criteria for herpes simplex virus testing of cerebrospinal fluid.  J Clin Microbiol. 2007;45(3):721-724. doi:10.1128/JCM.01950-06PubMedGoogle ScholarCrossref
26.
Hauser  RG, Brandt  CA, Martinello  RA.  Criteria to screen molecular tests for the diagnosis of herpes simplex virus in the central nervous system have no propensity to harm.  J Pathol Inform. 2017;8:4. doi:10.4103/2153-3539.201113PubMedGoogle ScholarCrossref
27.
Wilen  CB, Monaco  CL, Hoppe-Bauer  J, Jackups  R Jr, Bucelli  RC, Burnham  CA.  Criteria for reducing unnecessary testing for herpes simplex virus, varicella-zoster virus, cytomegalovirus, and enterovirus in cerebrospinal fluid samples from adults.  J Clin Microbiol. 2015;53(3):887-895. doi:10.1128/JCM.03161-14PubMedGoogle ScholarCrossref
28.
Tan  NWH, Lee  EY, Khoo  GMC, Tee  NWS, Krishnamoorthy  S, Choong  CT.  Cerebrospinal fluid white cell count: discriminatory or otherwise for enteroviral meningitis in infants and young children?  J Neurovirol. 2016;22(2):213-217. doi:10.1007/s13365-015-0387-2PubMedGoogle ScholarCrossref
29.
Yun  KW, Choi  EH, Cheon  DS,  et al.  Enteroviral meningitis without pleocytosis in children.  Arch Dis Child. 2012;97(10):874-878. doi:10.1136/archdischild-2012-301884PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 804
    Original Investigation
    Infectious Diseases
    July 19, 2019

    Assessment of Diagnostic Yield of Nonculture Infection Testing on Cerebrospinal Fluid in Immune-Competent Children

    Author Affiliations
    • 1Division of Neurology, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania
    • 2Department of Pediatrics, The Perelman School of Medicine at the University of Pennsylvania, Philadelphia
    • 3Department of Neurology, The Perelman School of Medicine at the University of Pennsylvania, Philadelphia
    • 4Division of Infectious Disease, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania
    • 5Division of Ophthalmology, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania
    • 6Department of Ophthalmology, The Perelman School of Medicine at the University of Pennsylvania, Philadelphia
    JAMA Netw Open. 2019;2(7):e197307. doi:10.1001/jamanetworkopen.2019.7307
    Key Points español 中文 (chinese)

    Question  What is the value of nonculture cerebrospinal fluid infection testing in immune-competent children with normal cerebrospinal fluid cell counts?

    Finding  In this cross-sectional study of 4811 lumbar puncture specimens from immune-competent children aged 0.5 to 18.9 years, only 2 specimens with positive results from nonculture cerebrospinal fluid infection testing required clinical intervention, among all specimens with normal cell counts. Both children who required clinical intervention had other clinical signs of infection.

    Meaning  Nonculture cerebrospinal fluid infection testing appeared to be common in immune-competent children with normal cerebrospinal fluid cell counts, but positive results seemed rare and unlikely to be independently associated with care.

    Abstract

    Importance  Nonculture infection tests of cerebrospinal fluid (CSF) samples using polymerase chain reaction and antigen or antibody assays are frequently ordered on lumbar puncture specimens concurrently with routine CSF cell counts, but the value of CSF infection testing in otherwise healthy children is unknown.

    Objective  To determine the value of nonculture CSF infection testing in immune-competent children with normal CSF cell counts.

    Design, Setting, and Participants  This cross-sectional study reviewed screening and diagnostic tests in the electronic medical record system of a large academic tertiary care children’s hospital. Records of children aged 0.5 to 18.9 years (n = 4083) who underwent lumbar puncture (n = 4811 procedures) in an inpatient or outpatient facility of Children's Hospital of Philadelphia between July 1, 2007, and December 31, 2016, were reviewed. Those with indwelling CSF shunts or catheters; those with active or past oncologic, immunologic, or rheumatologic conditions; or those taking immune-suppressing medications were excluded from analysis. This study was conducted from July 20, 2017, to March 13, 2019.

    Main Outcomes and Measures  Outcome variables included frequency of nonculture CSF infection testing and frequency of positive results in the entire cohort, and among those with normal cell counts. Normal cell counts were defined as CSF white blood cell counts lower than 5 cells/μL and CSF red blood cell counts lower than 500 cells/μL.

    Results  In total, 4811 lumbar puncture procedures were performed on 4083 unique children, with a median (range) age of 7.4 (0.5-18.9) years, 2537 boys (52.7%), and 3331 (69.2%) with normal CSF cell counts. At least 1 nonculture CSF infection test was performed on 1270 lumbar puncture specimens with normal cell counts (38.1%; 95% CI, 36%-40%), and more tests were performed in the summer months. Only 18 (1.4%; 95% CI, 0.9%-2.2%) of 1270 lumbar puncture specimens with normal cell counts had at least 1 nonculture infection test with a positive result; 2 of these 18 children required clinical intervention for their positive results, but each also had other clear clinical signs of infection.

    Conclusions and Relevance  Nonculture CSF infection testing appeared to be common in immune-competent children with normal CSF cell counts, but positive results were uncommon and were not independently associated with clinical care; delaying the decision to send nonculture infection tests until CSF cell counts are available could reduce unnecessary diagnostic testing and medical costs, which may improve value-based care.

    Introduction

    Acute central nervous system (CNS) infections, such as meningitis and encephalitis, require prompt recognition of treatable etiologies (eg, bacterial pathogens, herpes simplex virus [HSV]) to reduce potentially significant morbidity and mortality. Infections of the CNS are typically identified with lumbar puncture and evaluation of cerebrospinal fluid (CSF) cell counts, protein, glucose, Gram stain, bacterial culture, and nonculture infection tests. A CSF white blood cell (WBC) count lower than 5 cells/μL in immune-competent children older than 2 months is considered normal, whereas a count of 5 cells/μL or higher suggests infection and/or inflammation1 (to convert to ×109 per liter, multiply by 0.001). The presence of red blood cells (RBCs) in CSF samples represents either intrathecal blood (eg, subarachnoid or intraparenchymal hemorrhage) or introduction of peripheral blood into CSF during the performance of the lumbar puncture. Introduction of peripheral blood into otherwise sterile CSF may result in the potential contamination of a bloodstream pathogen, and it also makes the CSF WBC difficult to clinically interpret because an elevated count may reflect a true pleocytosis and/or peripheral blood WBCs.

    Polymerase chain reaction (PCR) of CSF and antigen or antibody assays are used to identify specific pathogens in suspected CNS infection; however, their overall diagnostic yield is variable. Nonculture CSF infection test sensitivity and specificity are high for some organisms (eg, HSV PCR,2 enterovirus PCR3) but are low for others (eg, Borrelia burgdorferi PCR and antibody assays4). The positive and negative predictive values, as well as the clinical advantages of these tests in populations with a low pretest probability for CNS infection (such as immune-competent children with normal CSF cell counts), are unknown.

    Laboratory testing with little clinical advantage but a disproportionately high cost is one example of low-value care. When a lumbar puncture is performed, frontline clinicians often order nonculture CSF infection tests concurrently with CSF cell counts, protein level, glucose level, and Gram stain to optimize administrative workflow and minimize the potential need later to collect more CSF. Even though lumbar puncture has a low complication rate,5,6 it often requires sedation for children, making a repeated procedure undesirable. To determine if ordering practices and laboratory workflow for nonculture CSF infection tests may be a target to reduce low-value care in pediatrics, we sought to identify the frequency of positive nonculture CSF infection tests and their association with the clinical care of immune-competent children with normal CSF cell counts.

    Methods

    We reviewed the electronic medical record (EMR) of Children’s Hospital of Philadelphia, a large academic tertiary care children’s hospital. The Children’s Hospital of Philadelphia Institutional Review Board approved this study and waived the informed consent requirement and Health Insurance Portability and Accountability Act authorization because the research and use of personal health information poses no greater than minimal risk and the study could not be performed without the waiver, which did not adversely affect the rights or welfare of the patients. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies. The study was conducted from July 20, 2017, to March 13, 2019.

    We queried the EMR for any CSF cell count performed on a child aged 0.5 to 18.9 years at an inpatient or outpatient Children’s Hospital of Philadelphia facility between July 1, 2007, and December 31, 2016. Some children underwent more than 1 lumbar puncture during this study period and contributed multiple independent data points over time to the analysis. We defined exclusion criteria according to International Classification of Diseases, Tenth Revision, codes in the EMR at or before the time of the lumbar puncture. We excluded children (1) with indwelling CSF shunts or catheters; (2) with malignant oncologic, immunologic, or rheumatologic diagnosis that may have interfered with the ability to mount an appropriate pleocytosis in response to the infection at the time of the lumbar puncture; (3) who underwent lumbar puncture in an oncology department, when the location of the procedure was available; and (4) with EMR documentation of a medication, taken at the time of the lumbar puncture, known to suppress the immune system.

    Patient data, including medical record number, demographic characteristics, epidemiologic information, and CSF test results were extracted from the EMR on children who met the inclusion and exclusion criteria. These data were manually reviewed in the EMR to verify suspected errors and missing data, confirm the accuracy of extracted data (all nonculture CSF infection tests were manually reviewed for lumbar puncture specimens with normal cell counts), and develop an abstract of clinical data from specimens with normal cell counts and a positive nonculture infection test result.

    Study Definitions

    Normal cell counts were defined as CSF WBC lower than 5 cells/μL and CSF RBC lower than 500 cells/μL; elevated cell counts were defined as CSF WBC 5 cells/μL or higher and/or CSF RBC 500 cells/μL or higher. When more than 1 set of cell counts were obtained for a given lumbar puncture specimen, we used the lowest value of each WBC and RBC count. We recorded CSF protein and glucose values less than a lower limit of detection (eg, CSF glucose <10 mg/dL [to convert to millimoles per liter, multiply by 0.0555]) as the lower limit of detection.

    The International Classification of Diseases, Tenth Revision, codes used to define exclusion criteria are summarized in the eMethods in the Supplement. Corticosteroids were not included as an exclusionary immune-suppressing medication given the varying immune outcomes by medication and dosing by body weight.

    Nonculture CSF infection tests we examined were PCR, antibody testing (IgG and/or IgM), and antigen testing. We included these tests in the analysis only if they were ordered on the same day that CSF cell counts were obtained; tests ordered on an existing specimen at a later date were excluded. Specific monoplex, real-time, qualitative CSF PCR tests we examined were those for the following: adenovirus (laboratory developed test [LDT], using published primer and probe sequences),7Borrelia burgdorferi (sent to Associated Regional and University Pathologists [ARUP] Laboratories), cytomegalovirus (LDT, using published primer and probe sequences),8 Epstein-Barr virus (LDT, using published primer and probe sequences),9 enterovirus (LDT, using published primer and probe sequences),10 human herpesvirus 6 (HHV-6; LDT, using published primer and probe sequences),11 herpes simplex virus (HSV; LDT, using published primer and probe sequences),12 human parechovirus (LDT, using published primer and probe sequences),13 influenza A and B virus (LDT, based on communication with Centers for Disease Control and Prevention), human metapneumovirus (LDT, using published primer and probe sequences),14 mumps virus (sent to the Pennsylvania state laboratory), Mycoplasma pneumoniae (LDT, using published primer and probe sequences),15 parainfluenza virus 1-3 (LDT, using published primer and probe sequences),16,17 parvovirus B19 (LDT, using published primer and probe sequences),18 varicella zoster virus (VZV; LDT, using published primer and probe sequences),19 and West Nile virus (sent to ARUP Laboratories).

    The Children’s Hospital of Philadelphia does not use commercial multiplex PCR tests for CSF samples because of performance concerns.20 Antigen or antibody CSF testing included arbovirus IgG or IgM panels (testing for California encephalitis, Eastern equine encephalitis, St Louis encephalitis, Western equine encephalitis, and West Nile virus; sent to ARUP Laboratories), cryptococcal antigen (ran in house [lateral flow assay; Immuno Mycologics]), Epstein-Barr virus IgG panels (ran in house [Zeus Scientific]), lymphocytic choriomeningitis virus IgG or IgM (sent to ARUP Laboratories), Lyme total antibody level (to define if someone was tested for Lyme; sent to ARUP Laboratories), Lyme Western blot IgG or IgM (to define a positive Lyme test result; sent to ARUP Laboratories), measles IgG or IgM (sent to Pennsylvania Department of Health), rubella IgG or IgM (IgG ran in-house [Captia assay; Trinity Biotech]; IgM sent to ARUP Laboratories), toxoplasmosis IgG or IgM (IgG ran in-house [Captia assay; Trinity Biotech]; IgM sent to ARUP Laboratories), VZV IgG or IgM (IgG ran in-house [Captia assay; Trinity Biotech]; IgM sent to ARUP Laboratories), and venereal diseases research laboratory assay (sent to ARUP Laboratories).

    Nonculture CSF infection test results were recorded as positive, negative, or not performed. Medical records with either IgG- or IgM-positive antibody CSF test results were manually reviewed to determine the clinical interpretation as positive or negative for the child’s clinical presentation. We manually reviewed all indeterminate or equivocal antibody testing results in the EMR, and if repeated confirmatory testing was not performed, we conservatively considered the results positive. When any component of a multitest panel (eg, West Nile virus IgM in an arbovirus panel) had a positive result, we considered the overall panel results positive.

    Statistical Analysis

    Data were analyzed using Stata, version 14.2 (StataCorp LLC). Nonparametric methods were used because not all continuous variables (such as age) were normally distributed and to minimize the effect of outliers with statistical associations. Continuous variables were described using medians and interquartile ranges (IQRs), and intergroup differences were compared using Wilcoxon rank sum tests. Categorical variables were described using counts and frequencies, and intergroup differences compared using the χ2 test. Statistical significance was determined a priori as a 2-tailed P < .05.

    After the overall cohort was analyzed, the subset of infants aged 6 to 12 months was also examined separately, given our a priori concern that their relatively immature immune system could potentially be a factor in the test results being different from those of older children.

    Results

    Cohort demographics and key laboratory characteristics of the overall cohort and subgroups by cell counts (normal and elevated) are summarized in Table 1. Overall, 4811 lumbar puncture procedures were performed on 4083 unique children, with a median (range) age of 7.4 (0.5-18.9) years and 2537 (52.7%) of whom were boys. The distribution of age was skewed to the left, with 1161 (24.1%) of 4811 procedures performed in children younger than 2 years. Children with elevated CSF cell counts were more often male (825 [55.7%] vs 1712 [51.4%]; P = .005) and African American (519 [35.1%] vs 975 [29.3%]; P = .001) compared with children with normal CSF cell counts (Table 1); median age did not differ between groups.

    Among the total 4811 lumbar puncture specimens, 2038 (42.4%; 95% CI, 41%-44%) had at least 1 nonculture CSF infection test sent: 270 specimens had both PCR and antibody or antigen testing, 1613 had only PCR testing, and 155 had just antibody or antigen testing. Among those specimens that had at least 1 nonculture CSF infection test, 1368 (67.1%; 95% CI, 65%-69%) had more than 1 test (median [range], 2 [1-11]). Testing was more frequent in the summer months compared with the remainder of the year (Figure 1A) but was consistent annually overall (Figure 1B). Only 227 (11.1%; 95% CI, 10%-13%) of the 2038 lumbar puncture specimens that had nonculture CSF infection testing, or 229 (4.5%; 95% CI, 3.8%-5.1%) of the 5054 total nonculture infection tests performed on these specimens, had a positive result. Enterovirus PCR (n = 186) accounted for 81% (95% CI, 76%-86%) of all nonculture CSF infection tests with a positive result, followed by Lyme Western blot (n = 12), HSV PCR (n = 10), HHV-6 PCR (n = 7), VZV PCR (n = 7), VZV IgM (n = 1), and others (Table 2).

    Of the 4811 lumbar puncture procedures performed, 3331 specimens (69.2%; 95% CI, 68%-71%) performed on 3033 unique children had normal CSF cell counts, of which 1270 (38.1%; 95% CI, 36%-40%) had at least 1 sent for nonculture CSF infection test. Among those 1270 specimens with normal cell counts sent for a nonculture CSF infection test, 827 (65.1%; 95% CI, 62%-68%) were sent for more than 1 test (median [range], 2 [1-11] tests). Only 18 (1.4%; 95% CI, 0.9%-2.2%) of 1270 specimens with normal cell counts tested, or 19 (0.6%; 95% CI, 0.4%-1%) of the 3021 total nonculture CSF infection tests performed on these specimens, had a positive result (Figure 2). Positive CSF test results in this subgroup included enterovirus PCR (n = 12), HHV-6 PCR (n = 3), HSV PCR (n = 1), measles IgG (n = 2), and toxoplasmosis IgG (n = 1). The 2 children who each had positive antibody testing also had recently received intravenous immunoglobulin, which may have confounded positive test results. Overall, positive results had active treatment implications for only 2 children. A 2-year-old with Hunter syndrome (mucopolysaccharidosis type II) had a positive CSF HSV PCR result and a magnetic resonance imaging scan highly suggestive of HSV encephalitis. An 11-year-old had positive results for CSF measles IgG and toxoplasmosis IgG; the latter was believed to be a false-positive associated with intravenous immunoglobulin. The child presented with progressive neurologic deterioration and a profoundly abnormal electroencephalogram and magnetic resonance imaging scan and ultimately received a diagnosis of subacute sclerosing panencephalitis (Table 3).

    To explore the role of CSF RBC values in these data, we performed a sensitivity analysis to include children with elevated CSF RBC in the original normal cell count subgroup. This inclusion added 191 lumbar puncture specimens to the sensitivity analysis, for a total of 3522 specimens. Among these 3522 specimens with normal CSF WBC alone, 1348 (38.3%; 95% CI, 37%-40%) were sent for at least 1 nonculture CSF infection test, of which 22 (1.6%; 95% CI, 1.1%-2.5%) had a positive result. However, none of the additional 4 procedures with positive results (0 HSV, 1 enterovirus PCR, and 3 HHV-6 PCR) had active treatment implications.

    Of the 4811 lumbar puncture procedures that met inclusion and exclusion criteria, 537 (11.2%) were performed on 460 unique infants between 6 and 12 months of age. Among the 537 procedures performed on these infants, 169 (31.5%; 95% CI, 28%-36%) had specimens sent for at least 1 nonculture CSF infection test, for a total of 352 nonculture infection tests. Only 13 (7.7%; 95% CI, 4.2%-12.8%) of the 169 lumbar puncture procedures performed on infants who had nonculture CSF infection testing, or 13 (3.7%; 95% CI, 2.0%-6.2%) of the 352 total nonculture infection tests performed on these procedures, had a positive result. These proportions are each slightly lower than seen in the overall cohort. Three hundred fifty-one (65.4%; 95% CI, 61%-69%) of 537 lumbar puncture specimens in infants had normal cell counts, a similar proportion to the overall cohort. One hundred fifteen (32.8%; 95% CI, 28%-38%) of 351 lumbar puncture specimens in infants with normal cell counts had at least 1 nonculture CSF infection test sent, 3 (2.6%; 95% CI, 0.5%-7.4%) of which had a positive result, a slightly higher proportion than the overall cohort (Table 3).

    From a value perspective, if nonculture CSF infection testing on lumbar puncture specimens with normal cell counts was not performed in this cohort, we would have eliminated testing on 1270 lumbar puncture specimens, a total of 3021 of the 5054 total tests sent (59.7%, 95% CI, 58%-61%) of all nonculture CSF infection tests sent for this cohort. However, we would have missed the 19 positive test results in 18 children outlined in Table 3.

    Discussion

    Lumbar puncture specimens with normal CSF cell counts were common in this cohort of immune-competent children but were rarely associated with nonculture CSF infection tests with a positive result and implications for clinical care. The proportion of these specimens sent for nonculture infection testing was lower compared with the proportion observed in those with elevated cell counts (38.1% vs 51.9%; P < .001), but this is only a 27% difference in magnitude, suggesting that cell counts alone are not a factor in the decision to send specimens for a nonculture CSF infection test in most children. Furthermore, once 1 nonculture infection test was sent for testing, both groups had a similar proportion of multiple tests sent (65.1% vs 67.1%), suggesting that once the decision is made to send for nonculture infection testing, CSF cell counts are not a factor in the number of tests ordered. Annual testing frequency has not substantially changed in the past 10 years at Children’s Hospital of Philadelphia. Seasonal testing frequency increases annually in July to September, which may be associated with seasonal enteroviral transmission (and subsequent increased testing practice) or with the relative inexperience of new house officers, as proposed in previous CSF PCR testing epidemiologic studies.22 Testing practices were similar in infants 6 to 12 months of age compared with testing in older children.

    Eight of the 18 children with normal cell counts who had testing with a positive result (7 CSF EV PCR, 1 CSF HHV-6 PCR) in the cohort were discharged before the results were available. One child underwent a lumbar puncture as an outpatient, and the positive nonculture infection test result (CSF measles IgG) was believed to be associated with prior intravenous immunoglobulin and not clinically significant. One child had a positive CSF enterovirus PCR result that prompted an acyclovir sodium discontinuation. The 2 children with positive test results that required clinical intervention (1 CSF HSV PCR, and 1 CSF measles IgG) each had other clear clinical and neuroradiographic signs of infection. The remaining 6 children had positive results that did not affect their medical treatment.

    Cell count results from CSF are typically available within hours, but other nonculture CSF infection tests can take days or weeks to complete. Implementing a testing strategy to delay the decision to send specimens for additional nonculture CSF infection testing until CSF cell counts were available could have reduced up to 60% of all nonculture CSF infection testing for this cohort. This type of change could reduce unnecessary diagnostic testing and medical costs, without substantially delaying the results in children with elevated cell counts or those with normal cell counts and a specific clinical concern based on other clinical data, thus improving value-based care.

    Low-value care (or clinical care that provides little or no patient advantage, may cause harm, or may yield marginal positive outcomes at a disproportionately high cost) is believed to constitute up to one-third of health care costs in the United States.23,24 Identifying sources of low-value care and minimizing it are critical: small changes applied across populations can be associated with meaningful reductions in domestic health care expenditures.23

    Several groups have previously examined various interventions to reduce unnecessary nonculture CSF infection testing in different populations. In a retrospective study of 732 adult CSF specimens submitted for testing at the Mayo Clinic, Tang et al22 demonstrated that restricting HHV PCR testing to specimens with CSF WBC count higher than 5 cells/μL and protein level greater than 45 mg/dL would have saved almost one-third of costs without reducing sensitivity. Hanson et al25 demonstrated that limiting HSV PCR testing to requests on specimens with CSF WBC count higher than 5 cells/μL or protein level greater than 50 mg/dL, specimens obtained from an immunocompromised individual (HIV positive or recipient of organ transplant) or from a child younger than 2 years, reduced testing by 22% (363 of 1659 specimens) and missed only 2 positive CSF HSV PCR results, which in retrospect did meet the criteria for testing. Hauser et al26 demonstrated that, among 6357 CSF HSV PCRs ordered on adults at a Veterans Affairs hospital over a 13-year period, no clinically significant positive results would have been missed by applying the Hanson et al25 criteria. In a split retrospective observational study with a screening and validation cohort, Wilen et al27 demonstrated that limiting the 8791 HSV, VZV, cytomegalovirus, and enterovirus PCRs ordered during the screening phase to immune-competent adults with CSF WBC count higher than 10 cells/μL would have reduced testing by about 55%, missing only 1 clinically significant result. In their validation cohort, 17 of 2126 tests were positive for HSV, VZV, cytomegalovirus, or enterovirus; applying the same criteria would not have missed any clinically significant test results.

    This current study adds to the growing body of literature on CSF testing and low-value care. It provides pediatric data on diagnostic yield of nonculture CSF infection testing across a broad battery of common tests. Unlike the adult studies, protein levels were largely normal among the children in this cohort with positive nonculture CSF infection testing with the exception of 1 child (elevated to 49 mg/dL; Table 3).

    Ultimately, the decision to send specimens for nonculture CSF infection testing requires a nuanced approach. It should incorporate community prevalence of infection, known pathogen risk factors, and patient-specific clinical judgment.

    Limitations

    This study has several limitations. First, although we suspected that CSF cell counts were not available at the time the decision was made to perform nonculture CSF infection testing, we did not know the definite timing of clinical decision making for each lumbar puncture in this large cohort. For this reason, we excluded nonculture infection tests ordered on days after the lumbar puncture, to exclude tests that were ordered on the basis of preliminary CSF data. Second, given the large size of the cohort, we based the exclusion criteria on International Classification of Diseases, Tenth Revision, codes. During the manual EMR review for quality assurance, we discovered that some children with lumbar puncture procedures ordered for noninfectious concerns (eg, demyelinating lesions, cytologic testing, and hemorrhage) and for neurosurgical concerns (eg, lumbar drains, hydrocephalus status post shunt, and others) remained in the cohort, likely because the pertinent diagnosis had not yet been added to their EMR. These findings may imply that our conclusions are more generalizable to broader populations, but further focused study is warranted.

    Third, the data we used were produced by EMR query, making it possible that some CSF tests ordered may have been missed. To minimize this risk, we manually reviewed all 3331 records of children with normal cell counts to ensure accurate documentation of nonculture CSF infection tests performed and their results. Fourth, we did not include infants younger than 6 months, limiting the study’s application to this population. This decision was a priori, given that the CSF WBC profiles of infants in the face of infection have substantial variability.28,29 Thus, the interpretation of CSF data in younger infants compared with older infants is a different and unique clinical challenge. Fifth, given the variation in CNS pathogen prevalence in different source populations, these data may not be completely generalizable to all tertiary care pediatric hospitals. However, the potential to reduce nonculture CSF infection testing in children with normal CSF cell counts should be applicable in any setting.

    Conclusions

    This study’s findings suggest that lumbar puncture specimens with normal cell counts are common in immune-competent children older than 6 months but are rarely associated with nonculture CSF infection tests with positive results that are independently associated with clinical care. Implementing a hospitalwide intervention to delay the decision to order nonculture CSF infection testing, or to hold specimens in the laboratory until CSF cell counts are available, could dramatically reduce unnecessary testing and associated costs, improving value-based care. A nuanced approach is needed in the decision to send specimens for nonculture CSF infection testing. This approach should include community prevalence of infection, known pathogen risk factors, and patient-specific clinical judgment.

    Back to top
    Article Information

    Accepted for Publication: May 29, 2019.

    Published: July 19, 2019. doi:10.1001/jamanetworkopen.2019.7307

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 McGuire JL et al. JAMA Network Open.

    Corresponding Author: Jennifer L. McGuire, MD MSCE, Department of Pediatrics, Perelman School of Medicine at the University of Pennsylvania, 34th St and Civic Center Blvd, Philadelphia, PA 19104 (mcguirej@email.chop.edu).

    Author Contributions: Dr McGuire had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: McGuire, Swami, Avery.

    Acquisition, analysis, or interpretation of data: McGuire, Tuite, Avery.

    Drafting of the manuscript: McGuire, Avery.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: McGuire.

    Obtained funding: McGuire.

    Administrative, technical, or material support: All authors.

    Supervision: Avery.

    Conflict of Interest Disclosures: None reported.

    Funding/Support: Dr McGuire was supported by grant K23 NS094069 from the National Institutes of Health.

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    Disclaimer: The views expressed herein are those of the authors and do not reflect the official policy or position of the National Institutes of Health.

    Meeting Presentation: Portions of these data were presented in a poster at the Pediatric Academic Societies Annual Meeting; May 7, 2018; Toronto, Ontario, Canada.

    Additional Contributions: Ana María Cárdenas, PhD, D(ABMM), Infectious Disease Diagnostics Laboratory at the Children’s Hospital of Philadelphia, provided assistance with laboratory protocols used in this study. She did not receive compensation for her contribution.

    References
    1.
    Arcara  K, Tschudy  M.  The Harriet Lane Handbook. 19th ed. Maryland Heights, MO: Mosby; 2011.
    2.
    Mitchell  PS, Espy  MJ, Smith  TF,  et al.  Laboratory diagnosis of central nervous system infections with herpes simplex virus by PCR performed with cerebrospinal fluid specimens.  J Clin Microbiol. 1997;35(11):2873-2877.PubMedGoogle Scholar
    3.
    Iturriza-Gómara  M, Megson  B, Gray  J.  Molecular detection and characterization of human enteroviruses directly from clinical samples using RT-PCR and DNA sequencing.  J Med Virol. 2006;78(2):243-253. doi:10.1002/jmv.20533PubMedGoogle ScholarCrossref
    4.
    Avery  RA, Frank  G, Eppes  SC.  Diagnostic utility of Borrelia burgdorferi cerebrospinal fluid polymerase chain reaction in children with Lyme meningitis.  Pediatr Infect Dis J. 2005;24(8):705-708. doi:10.1097/01.inf.0000172903.14077.4cPubMedGoogle ScholarCrossref
    5.
    Marton  KI, Gean  AD.  The spinal tap: a new look at an old test.  Ann Intern Med. 1986;104(6):840-848. doi:10.7326/0003-4819-104-6-840PubMedGoogle ScholarCrossref
    6.
    Tunkel  AR, Hartman  BJ, Kaplan  SL,  et al.  Practice guidelines for the management of bacterial meningitis.  Clin Infect Dis. 2004;39(9):1267-1284. doi:10.1086/425368PubMedGoogle ScholarCrossref
    7.
    Heim  A, Ebnet  C, Harste  G, Pring-Åkerblom  P.  Rapid and quantitative detection of human adenovirus DNA by real-time PCR.  J Med Virol. 2003;70(2):228-239. doi:10.1002/jmv.10382PubMedGoogle ScholarCrossref
    8.
    Sanchez  JL, Storch  GA.  Multiplex, quantitative, real-time PCR assay for cytomegalovirus and human DNA.  J Clin Microbiol. 2002;40(7):2381-2386. doi:10.1128/JCM.40.7.2381-2386.2002PubMedGoogle ScholarCrossref
    9.
    Niesters  HG, van Esser  J, Fries  E, Wolthers  KC, Cornelissen  J, Osterhaus  AD.  Development of a real-time quantitative assay for detection of Epstein-Barr virus.  J Clin Microbiol. 2000;38(2):712-715.PubMedGoogle Scholar
    10.
    Kares  S, Lönnrot  M, Vuorinen  P, Oikarinen  S, Taurianen  S, Hyöty  H.  Real-time PCR for rapid diagnosis of entero- and rhinovirus infections using LightCycler.  J Clin Virol. 2004;29(2):99-104. doi:10.1016/S1386-6532(03)00093-3PubMedGoogle ScholarCrossref
    11.
    Gautheret-Dejean  A, Manichanh  C, Thien-Ah-Koon  F,  et al.  Development of a real-time polymerase chain reaction assay for the diagnosis of human herpesvirus-6 infection and application to bone marrow transplant patients.  J Virol Methods. 2002;100(1-2):27-35. doi:10.1016/S0166-0934(01)00390-1PubMedGoogle ScholarCrossref
    12.
    Kessler  HH, Mühlbauer  G, Rinner  B,  et al.  Detection of herpes simplex virus DNA by real-time PCR.  J Clin Microbiol. 2000;38(7):2638-2642.PubMedGoogle Scholar
    13.
    Nix  WA, Maher  K, Johansson  ES,  et al.  Detection of all known parechoviruses by real-time PCR.  J Clin Microbiol. 2008;46(8):2519-2524. doi:10.1128/JCM.00277-08PubMedGoogle ScholarCrossref
    14.
    Maertzdorf  J, Wang  CK, Brown  JB,  et al.  Real-time reverse transcriptase PCR assay for detection of human metapneumoviruses from all known genetic lineages.  J Clin Microbiol. 2004;42(3):981-986. doi:10.1128/JCM.42.3.981-986.2004PubMedGoogle ScholarCrossref
    15.
    Hardegger  D, Nadal  D, Bossart  W, Altwegg  M, Dutly  F.  Rapid detection of Mycoplasma pneumoniae in clinical samples by real-time PCR.  J Microbiol Methods. 2000;41(1):45-51. doi:10.1016/S0167-7012(00)00135-4PubMedGoogle ScholarCrossref
    16.
    Hibbitts  S, Rahman  A, John  R, Westmoreland  D, Fox  JD.  Development and evaluation of NucliSens basic kit NASBA for diagnosis of parainfluenza virus infection with ‘end-point’ and ‘real-time’ detection.  J Virol Methods. 2003;108(2):145-155. doi:10.1016/S0166-0934(02)00268-9PubMedGoogle ScholarCrossref
    17.
    Watzinger  F, Suda  M, Preuner  S,  et al.  Real-time quantitative PCR assays for detection and monitoring of pathogenic human viruses in immunosuppressed pediatric patients.  J Clin Microbiol. 2004;42(11):5189-5198. doi:10.1128/JCM.42.11.5189-5198.2004PubMedGoogle ScholarCrossref
    18.
    Jordan  JA.  Identification of human parvovirus B19 infection in idiopathic nonimmune hydrops fetalis.  Am J Obstet Gynecol. 1996;174(1, pt 1):37-42. doi:10.1016/S0002-9378(96)70370-8PubMedGoogle ScholarCrossref
    19.
    Pevenstein  SR, Williams  RK, McChesney  D, Mont  EK, Smialek  JE, Straus  SE.  Quantitation of latent varicella-zoster virus and herpes simplex virus genomes in human trigeminal ganglia.  J Virol. 1999;73(12):10514-10518.PubMedGoogle Scholar
    20.
    Graf  EH, Farquharson  MV, Cárdenas  AM.  Comparative evaluation of the FilmArray meningitis/encephalitis molecular panel in a pediatric population.  Diagn Microbiol Infect Dis. 2017;87(1):92-94. doi:10.1016/j.diagmicrobio.2016.09.022PubMedGoogle ScholarCrossref
    21.
    Jafri  SK, Kumar  R, Ibrahim  SH.  Subacute sclerosing panencephalitis - current perspectives.  Pediatric Health Med Ther. 2018;9:67-71. doi:10.2147/PHMT.S126293PubMedGoogle ScholarCrossref
    22.
    Tang  YW, Hibbs  JR, Tau  KR,  et al.  Effective use of polymerase chain reaction for diagnosis of central nervous system infections.  Clin Infect Dis. 1999;29(4):803-806. doi:10.1086/520439PubMedGoogle ScholarCrossref
    23.
    Hiscock  H, Neely  RJ, Warren  H, Soon  J, Georgiou  A.  Reducing unnecessary imaging and pathology tests: a systematic review.  Pediatrics. 2018;141(2):e20172862. doi:10.1542/peds.2017-2862PubMedGoogle ScholarCrossref
    24.
    Gawande  AA, Colla  CH, Halpern  SD, Landon  BE.  Avoiding low-value care.  N Engl J Med. 2014;370(14):e21. doi:10.1056/NEJMp1401245PubMedGoogle ScholarCrossref
    25.
    Hanson  KE, Alexander  BD, Woods  C, Petti  C, Reller  LB.  Validation of laboratory screening criteria for herpes simplex virus testing of cerebrospinal fluid.  J Clin Microbiol. 2007;45(3):721-724. doi:10.1128/JCM.01950-06PubMedGoogle ScholarCrossref
    26.
    Hauser  RG, Brandt  CA, Martinello  RA.  Criteria to screen molecular tests for the diagnosis of herpes simplex virus in the central nervous system have no propensity to harm.  J Pathol Inform. 2017;8:4. doi:10.4103/2153-3539.201113PubMedGoogle ScholarCrossref
    27.
    Wilen  CB, Monaco  CL, Hoppe-Bauer  J, Jackups  R Jr, Bucelli  RC, Burnham  CA.  Criteria for reducing unnecessary testing for herpes simplex virus, varicella-zoster virus, cytomegalovirus, and enterovirus in cerebrospinal fluid samples from adults.  J Clin Microbiol. 2015;53(3):887-895. doi:10.1128/JCM.03161-14PubMedGoogle ScholarCrossref
    28.
    Tan  NWH, Lee  EY, Khoo  GMC, Tee  NWS, Krishnamoorthy  S, Choong  CT.  Cerebrospinal fluid white cell count: discriminatory or otherwise for enteroviral meningitis in infants and young children?  J Neurovirol. 2016;22(2):213-217. doi:10.1007/s13365-015-0387-2PubMedGoogle ScholarCrossref
    29.
    Yun  KW, Choi  EH, Cheon  DS,  et al.  Enteroviral meningitis without pleocytosis in children.  Arch Dis Child. 2012;97(10):874-878. doi:10.1136/archdischild-2012-301884PubMedGoogle ScholarCrossref
    ×