Diagnostic accuracy of CSF cell index and corrected CSF white blood cell count in healthcare-associated ventriculitis and meningitis after intracranial hemorrhage

Abstract

Background
The diagnosis of healthcare-associated meningitis and ventriculitis (HCAMV) in patients with intracranial hemorrhage (ICH) is challenging. The purpose of this study was to evaluate the diagnostic accuracy of routine cerebrospinal fluid (CSF) studies including a cell index and a corrected white blood cell (WBC) count.

Methods
Case control study of adult patients with the diagnosis of ICH and HCAMV at a large tertiary care hospital in Houston, Texas from 2003 to 2016. Cases were defined as patients with ICH and HCAMV as documented by a positive CSF culture. Controls were selected as patients with ICH without evidence of HCAMV, no previous antibiotic therapy and a negative CSF culture. Cases and controls were matched 1:2 by age, Glasgow Coma Scale (GCS) and Apache II scores. Cell index was calculated using the following formula: (CSF leukocytes / CSF erythrocytes) / (blood leukocytes / blood erythrocytes). Corrected WBC count was calculated using the following formula: CSF leukocytes - (CSF erythrocytes/1,000). Area under the curve of receiver operating characteristic (AUC-ROC) and 95% confidence interval (CI) for CSF cell index greater than or equal to absolute value of 1, corrected CSF WBC count greater than 5 K/uL, CSF lactate greater than 4 mmol/L, and CSF glucose less than 40 mmol/L, respectively, were calculated in order to determine the accuracy of these studies.

Results
A total of 120 patients with ICH were included in this study; 40 patients had proven HCAMV whereas 80 patients had ICH with no evidence of HCAMV. Matching of cases and controls by age, GCS, and Apache II score was appropriate (p>0.05). The AUC-ROC values for CSF cell index, corrected CSF WBC count, CSF lactate, and CSF glucose were all low at 0.609 (95% CI = 0.449–0.768), 0.731 (95% CI = 0.589–0.872), 0.719 (95% CI = 0.573–0.864), and 0.609 (95% CI = 0.449–0.768), respectively.

Conclusion
This study demonstrated poor accuracy of CSF cell index, corrected CSF WBC count, CSF lactate, and CSF glucose in diagnosis of HCAMV after ICH.

Disclosures
R. Hasbun, Biomeriaux: Consultant, Consulting fee. Biofire: Speaker’s Bureau, Speaker honorarium. Merck: Speaker’s Bureau, Speaker honorarium. Pfizer: Speaker’s Bureau, Speaker honorarium. Medicine’s Co: Speaker’s Bureau, Speaker honorarium.

Full text

S304 • OFID 2017:4 (Suppl 1) • Poster Abstracts
Disclosures. Al l authors: No reported disclosures.
1006. Diagnostic accuracy of CSF cell index and corrected CSF white blood cell
count in healthcare-associated ventriculitis and meningitis after intracranial
hemorrhage
HelenaJenkinson, MD1; OnaizahHabib, MD2; LucreciaSalazar, MD2 and
RodrigoHasbun, MD, MPH3; 1University of Texas McGovern Medical School at
Houston, Houston, Texas, 2Division of Infectious Diseases, University of Texas
McGovern Medical School at Houston, Houston, Texas, 3Division of Infectious
Diseases, University of Texas Health Science Center at Houston, McGovern Medical
School, Houston, Texas
Session: 137. Adult CNS Infection
Friday, October 6, 2017: 12:30 PM
Background. e diagnosis of healthcare-associated meningitis and ventricu-
litis (HCAMV) in patients with intracranial hemorrhage (ICH) is challenging. e
purpose of this study was to evaluate the diagnostic accuracy of routine cerebro-
spinal uid (CSF) studies including a cell index and a corrected white blood cell
(WBC)count.
Methods. Case control study of adult patients with the diagnosis of ICH and
HCAMV at a large tertiary care hospital in Houston, Texas from 2003 to 2016. Cases
were dened as patients with ICH and HCAMV as documented by a positive CSF
culture. Controls were selected as patients with ICH without evidence of HCAMV,
no previous antibiotic therapy and a negative CSF culture. Cases and controls were
matched 1:2 by age, Glasgow Coma Scale (GCS) and Apache II scores. Cell index was
calculated using the following formula: (CSF leukocytes / CSF erythrocytes) / (blood
leukocytes / blood erythrocytes). Corrected WBC count was calculated using the
following formula: CSF leukocytes - (CSF erythrocytes/1,000). Area under the curve
of receiver operating characteristic (AUC-ROC) and 95% condence interval (CI)
for CSF cell index greater than or equal to absolute value of 1, corrected CSF WBC
count greater than 5 K/uL, CSF lactate greater than 4mmol/L, and CSF glucose less
than 40mmol/L, respectively, were calculated in order to determine the accuracy of
these studies.
Results. A total of 120 patients with ICH were included in this study; 40 patients
had proven HCAMV whereas 80 patients had ICH with no evidence of HCAMV.
Matching of cases and controls by age, GCS, and Apache II score was appropriate
(p>0.05). e AUC-ROC values for CSF cell index, corrected CSF WBC count, CSF
lactate, and CSF glucose were all low at 0.609 (95% CI= 0.449–0.768), 0.731 (95%
CI=0.589–0.872), 0.719 (95% CI= 0.573–0.864), and 0.609 (95% CI=0.449–0.768),
respectively.
Conclusion. is study demonstrated poor accuracy of CSF cell index, corrected
CSF WBC count, CSF lactate, and CSF glucose in diagnosis of HCAMV aerICH.
Disclosures. R. Hasbun, Biomeriaux: Consultant, Consulting fee. Biore:
Speaker’s Bureau, Speaker honorarium. Merck: Speaker’s Bureau, Speaker honorar-
ium. Pzer: Speaker’s Bureau, Speaker honorarium. Medicine’s Co: Speaker’s Bureau,
Speaker honorarium.
1007. Achieving Optimal Specialty Cerebrospinal Fluid (CSF) Testing: Are
Electronic Medical Record Order Sets Helpful?
Jessica I.Abrantes-Figueiredo, MD1; Virginia M.Bieluch, MD, FIDSA2 and Michael
D.Nailor, PharmD, BCPS (AQ-ID)3; 1University of Connecticut Health Center,
Farmington, Connecticut, 2Medicine, Hospital of Central Connecticut, New Britain,
Connecticut, 3University of Connecticut School of Pharmacy, Storrs, Connecticut
Session: 137. Adult CNS Infection
Friday, October 6, 2017: 12:30 PM
Background. Specialty PCR testing has become available for lumbar puncture to
determine the cause of infectious meningitis and encephalitis. Testing with low pre-test
probability may increase antimicrobial therapy while results are pending and create
increased direct costs. We aim to describe the appropriateness of testing before and
aer the implementation of electronic medical record (EMR) order sets designed to
reduce excessive testing of CSF by creating two lists of tests: (1) a routine panel for
all patients and (2) a list of optional specialty tests designed to be utilized aer the
nucleated cells are resulted.
Methods. Retrospective study of adult patients undergoing lumbar puncture with
suspicion for CNS infection pre-and post-implementation of EMR order sets from
January 2016–March 2017. Consecutive patients with complete charts were reviewed
from a tertiary care center. Data collected included demographics, co-morbid con-
ditions, clinical presentation, and lumbar puncture results. e primary outcome of
interest was the frequency of CSF specialty testing in patients with ≤ 10 nucleated cells/
µL in theCSF.
Results. Two hundred patients had ≤ 10 nucleated cells/µL in the CSF (n = 108
in pre-EMR group; n = 92 in post-EMR group). Of these patients 74% and 48.9% had
Herpes Simplex Virus (HSV) PCR testing done pre and post EMR changes (P<0.05).
Enterovirus PCR testing remained similar among both groups (37% pre-EMR order
sets vs. 36.9% post-EMR order sets, P= 0.99). Lyme PCR testing decreased between
pre- and post-groups (26.8% vs. 9.7%, P<0.05). CSF Epstein-Barr virus PCR testing
also dropped signicantly from 26.9% to 7.6% (P< 0.05). All specialty PCR testing
that was performed on patients with ≤ 10 nucleated cells/µL in the CSF were negative.
Paradoxically, HSV antibody testing increased post-implementation of EMR order sets
(21.7% vs. 0%, P <0.05). Total costs of tests on average decreased by $70.71 per patient
post EMR changes.
Conclusion. In this cohort, CSF specialty testing was common but decreased aer
EMR changes. Laboratory stewardship can be improved with EMR changes but further
education is needed to prevent unnecessary tests. Unwanted tests (HSV antibodies)
may be increased as prescribers are unable to locate familiartests.
Disclosures. M. D. Nailor, Astellas: Scientic Advisor and Speaker’s Bureau,
Consulting fee and Speaker honorarium. Merck: Grant Investigator, Research grant.
1008. Brain Abscess Risk Associated with Genotypic Polymorphism of the Matrix
Metalloproteinase-1, -2, -3, and -9 in North Indian Population
PriyankaMishra, MSc1; Kashi NathPrasad, MD2; KaminiSingh, PhD3; Rabi
NarayanSahu, MS4; SwayamPrakash, PhD5 and Bal KOjha, MS6; 1Microbiology,
Sanjay Gandhi Post Graduate Institute of Medical Science, lucknow, India,
2Microbiology, Sanjay Gandhi Post Graduate of Medical Science, lucknow, India,
3Microbiology, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow,
India., lucknow, India, 4Neurosurgery, Sanjay Gandhi Postgraduate Institute of
Medical Sciences, Lucknow, India, lucknow, India, 5Sanjay Gandhi Postgraduate
Institute of Medical Sciences, Lucknow, India, 6Sanjay Gandhi Postgraduate Institute
of Medical Sciences, lucknow, India
Session: 137. Adult CNS Infection
Friday, October 6, 2017: 12:30 PM
Background. Brain abscess develops in response to a parenchymal infection
due to pyogenic bacteria. MMPs (matrix metalloproteinases) play vital role in many
infectious and central nervous system (CNS) diseases. e present study evaluated
the association of specic alleles/ genotypes of MMP-1, -2, -3, and -9 with brain
abscess.
Methods. A total of 100 brain abscess patients and 100 healthy controls were
included in the study. Predisposing factors were identied in 70 brain abscess patients.
Out of 100 brain abscess samples, 66 were culture positive. MMP-1-1607 1G/2G,
MMP-2- C-1306-T, MMP-3 -1171 5A/6A, and MMP-9 C-1562Tgenotypes were
detected by PCR-RFLP. Levels of these MMPs were determined in patients’ sera by
ELISA and correlated with dierent genotypes.
Results. e genotypic distributions of MMP-1-1607 1G/2G, MMP-2-
C-1306-T, MMP-3 -1171 5A/6A, and MMP-9 C-1562T were signicantly dierent
between patients and controls. Homozygous genotype of MMP-1, -3, and -9 (P <
0.001, P = 0.04, and P = 0.03, respectively) and heterozygous genotype of MMP-2 (P
< 0.001) showed signicant association with brain abscess. Individuals with mutant
genotypes had elevated levels of these MMPs. Furthermore, heterozygous (5A/6A
and C/T, respectively) genotypes of MMP-3 and -9 also showed signicant associ-
ation with brain abscess patients having predisposing factors. When comparison
was made between culture positive and culture negative results, of MMP-1 2G/2G
and MMP-9 T/T, C/T genotype showed signicant association with culture positive
patients
Conclusion. Polymorphism of MMP-1-1607 1G/2G, MMP-2- C-1306-T, MMP-3
-1171 5A/6A, and MMP-9 C-1562T polymorphisms lead to increased production of
these molecules, which appear to be a risk for the development of brain abscess in
North Indian population.
Disclosures. Al l authors: No reported disclosures.
1009. The Etiologies and Clinical Characteristics of Patients Hospitalized with an
Acute Febrile Illness and Central Nervous System Sydromes in Indonesia
Abu olibAman, MSc, PhD, SpMK (K)1; Muhammad HusseinGasem, PhD,
SpPD-KPTI2; EmilianaTjitra, MSc, PhD3; BachtiAlisjahbana, PhD, SpPD-KPTI4;
HermanKosasih, PhD5; Ketut TutiMerati, SpPD-KPTI6; MansyurArif, SpPK7;
MuhammadKaryana, M.Kes8; PratiwiSudarmono, PhD, SpMK(K)9; SuhartoSuharto,
MPdk, DTMH, SpPD-KPTI10; ViviLisdawati, M.SI, Apt11; AaronNeal, PhD12 and
SophiaSiddiqui, MPH12 1Gadjah Mada University, Yogyakarta, Indonesia, 2Faculty
of Medicine, Diponegoro University, Semarang, Indonesia, 3National Institute of
Health Research and Development, Ministry of Health, Jakarta, Indonesia, Jakarta,
Indonesia, 4Faculty of Medicine, Hasan Sadikin General Hospital, Bandung,
Indonesia, 5INA-RESPOND, Jakarta, Indonesia, 6Faculty of Medicine Udayana
University & Sanglah Hospital, Bali, Indonesia, 7Dr. Wahidin Soedirohusodo Hospital
& Hasanuddin University, Makassar, Indonesia, 8National Institute of Health Research
and Development, Ministry of Health, Jakarta, IndonesiaD, Jakarta, Indonesia, 9Cipto
Mangunkusumo Hospital, Jakarta, Indonesia, 10Faculty of Medicine, Airlangga
University & Dr. Soetomo Hospital, Surabaya, Indonesia, 11Sulianti Saroso Hospital,
Jakarta, Indonesia, 12Clinical Research Center, NIAID, National Institutes of Health,
Bethesda, United States, Bethesda, Maryland
Session: 137. Adult CNS Infection
Friday, October 6, 2017: 12:30 PM
Background. Acute febrile illness is a common reason for hospitalization in many
developing countries, including Indonesia. While patients can oen be categorized
and managed based on clinical presentations, diagnostic capacity in these countries
remains limited, leading to poor patient outcomes. For patients with central nervous
system (CNS) infections, identifying the underlying etiologies is particularly impor-
tant to prevent lifelong neurological complications anddeath.
Methods. As part of a study conducted at 8 top-referral hospitals across
Indonesia from 2013 to 2016, 114 of 1,486 enrolled subjects presented with an
acute fever and a CNS syndrome. To identify the etiologies and clinical manifesta-
tions of these infections, as well as the management of febrile patients at the hospi-
tals, demographic and clinical data were collected at enrollment, and blood samples
Downloaded from https://academic.oup.com/ofid/article-abstract/4/suppl_1/S304/4294280/Diagnostic-accuracy-of-CSF-cell-index-and
by guest
on 04 October 2017