[Downloaded free from http://www.ijpmonline.org on Monday, May 27, 2019, IP: 14.139.154.

130]

Utility of urine reagent strips in cerebrospinal fluid

Original Article

analysis: An aid to bedside diagnosis of meningitis

Sujaya Mazumder, B. S. Ramya, Dayananda S. Biligi
Department of Pathology, Bangalore Medical College and Research Institute, Bengaluru, Karnataka, India

Address for correspondence:

Dr. Sujaya Mazumder, No. 4D, Jeno Sougoumaran Apartment, Ajjis Nagar, Reddiyarpalayam, Puducherry ‑ 605 010, India.
E‑mail: suji0206@gmail.com

ABSTRACT Access this article online

Website: www.ijpmonline.org
Context: The provision of initial treatment to a patient with suspected meningitis PMID: xxxxxxxxx (when available)
depends greatly on early recognition and rapid diagnostic evaluation of DOI: 10.4103/IJPM.IJPM_821_16
cerebrospinal fluid  (CSF) leukocytes, proteins, and glucose. The diagnosis Quick Response Code:
is time critical and timely intervention has an implication on the prognosis
and outcome. Reasonably, sound laboratorial setups are not available in our
country in the primary health‑care level and, even in the settings where they
are available, long waiting periods precede the availability of results. Aims: We
conducted this study to emphasize the role of urine reagent strip test as a rapid
diagnostic tool in CSF analysis. Settings and Design: This is a prospective
single‑blinded study on 100 consecutive CSF samples received with in 1h of tap.
Subjects and Methods: All the 100 samples were subjected to definitive test
being CSF microscopy and biochemical analysis of proteins and sugar and index
test being a semi quantitative analysis of CSF leukocytes, proteins, and sugar facilities may not be present in the rural
by urinary reagent strips. Statistical Analysis Used: The diagnostic accuracy setup and, even in the settings where they
of the reagent strip for different cutoff levels was estimated and tabulated in the are available, the turnaround time period
form of sensitivity, specificity, positive predictive value, negative predictive value, for laboratory‑based CSF analysis is long.[4]
and likelihood ratio. Results: 77% of cases were in the pediatric age group and
Although many authors have studied the
23% cases were adults. The sensitivity and specificity for leukocytes by the strip
method for  ≥15  cells/cumm were 89.28% and 98.61%, respectively, which role of urinary reagent strips (a well‑known
increased to 100% with an increase in the counts. The reagent strip test had a semi‑quantitative method used widely
sensitivity of 85.71% and a specificity of 95.65% for the protein levels >30 mg/dl for urine analysis) in the diagnosis of
which increased to 100% with an increase in protein levels. The reagent strip meningitis, this method has not yet
test for glucose was highly specific (100%) but less sensitive. Conclusions: The gained the popularity it deserves.[4‑9] The
results indicate that urine reagent strip is instrumental in bedside CSF analysis reagent strip test could prove to be a rapid
and has a future stand in the diagnosis of meningitis. bedside diagnostic test for CSF pleocytosis
and glucose and protein estimation until
KEY WORDS: Cerebrospinal fluid, meningitis, urine reagent strips further substantiation by laboratory results
occurs.[10]

INTRODUCTION The present study is undertaken to

evaluate and establish the role of urinary
Meningitis, the infection of the brain and spinal cord, is a medical emergency and its reagent strip method in the analysis of
timely identification and treatment is essential for preventing permanent neurological
This is an open access journal, and
deficits and death.[1,2] articles are distributed under the
terms of the Creative Commons
Neonatal meningitis continues to contribute substantially to neurological deficits Attribution‑NonCommercial‑ShareAlike 4.0
License, which allows others to remix, tweak,
worldwide. At the same time, conditions such as febrile seizures, especially in the and build upon the work non‑commercially, as
pediatric population, make allusions to meningitis, resulting in a diagnostic dilemma long as appropriate credit is given and the new
for the clinicians. creations are licensed under the identical terms.

For reprints contact: reprints@medknow.com

Cerebrospinal fluid (CSF) cell counts and protein and sugar estimation, along with How to cite this article: Mazumder S,
microbiological studies, are required to make the diagnosis of meningitis.[3] The presence Ramya BS, Biligi DS. Utility of urine reagent
strips in cerebrospinal fluid analysis: An aid to
of an experienced pathologist and the availability of a reasonably sound laboratorial setup bedside diagnosis of meningitis. Indian J Pathol
are indispensable for the estimation of cell count and protein and sugar estimation. These Microbiol 2018;61:356-9.

356 © 2018 Indian Journal of Pathology and Microbiology | Published by Wolters Kluwer - Medknow
[Downloaded free from http://www.ijpmonline.org on Monday, May 27, 2019, IP: 14.139.154.130]

Mazumder, et al.: Role of urine reagent strips in CSF analysis

CSF to assist the clinician in the bedside diagnosis or exclusion and specificity as a function of the strip color cutoff was plotted
of meningitis. and the area under the curve (AUC) along with its standard
error was calculated using the SPSS  statistical software
SUBJECTS AND METHODS package.

We conducted a prospective single‑blinded study, after ethical RESULTS

committee clearance on 100 consecutive CSF samples received
in our laboratory within an hour of tap. Hemorrhagic taps were The study was conducted on 100 consecutive CSF samples.
excluded from the study. Age and sex were recorded and all the There was a fairly equal distribution of males (52%) and
samples were subjected to both index and definitive tests. females (48%).77% of cases were in the pediatric age
group and 23% of cases were adults. 41% of the cases were
Index test neonates [Table 1].
Index test was done by one of the investigators using urinary
reagent strips which can detect ten parameters including Three cardinal parameters, leukocytes, proteins, and glucose tests
protein, glucose, and leukocyte esterase. With the help of a of the reagent strip, were compared with those of the reference
pipette, 2–3 drops of undiluted CSF were added to patches of standard. The sensitivity and specificity for leukocytes by the
proteins, glucose, and leukocyte esterase, and the color change strip method for ≥15 cells/cumm were 89.28% and 98.61%,
was recorded and interpreted using the manufacturer‑provided respectively. The accuracy increased to 100% with higher cutoffs
color grading. of cell count[Table 2].

The reagent strip is designed to detect the range of leukocytes The reagent strip test had a sensitivity of 85.71% and a specificity
from 15to 500 cells/mm3. It detects the leukocytes by estimation of 95.65% for the protein levels >30 mg/dl. For a cutoff level of
of leukocyte esterase. Depending on the color change, leukocytes the proteins of >300 mg/dl, both the sensitivity and specificity
are graded as negative for the cell count of  <15  cells/cumm, increased to 100%[Table 2].
1+ for 15–70 cells/cumm, 2+ for 70–125 cells/cumm, and 3+ for
125–500 cells/cumm. The reagent strip test for glucose at cutoff levels
of  <40  mg/dl and  <50  mg/dl was highly specific with a
For proteins, the interpretation of the colors on the reagent specificity of 100% but it was less sensitive (71.42% and 48.2%,
strips was as follows:<30 as negative, 30–100 was graded as respectively) [Table 2].
1+, 100–300 was graded as 2+, and 300–2000 was graded as
3+. We determined if the protein was <30(normal), >30mg/dl, For each stripcolor’s cutoff value, the positive and negative
>100mg/dl, or >300mg/dl based on the color change. likelihood ratioswerealso derived [Table 3]. The ROC curves with
AUC for leukocytes >15, protein >30, and sugar <50 are shown
For sugars, the readings are as follows: no color change for in Figures 1‑3, respectively.
sugars <50mg/dl, traces for 50–100mg/dl, 1+ for 100–250mg/dl,
2+ for count between 250 and 500, and 3+ for counts between Table 1: Age‑wise distribution of cases
500 and 1000. Using the strip, we determined if the values Age Males Females Total
were above or below 50mg/dl with the interpretation being 0 day‑1 month 18 23 41
no change in color as  <50mg/dl and any change in color 2 months‑1 year 10 5 15
as >50mg/dl. 2 years‑5 years 8 5 13
6 years‑15 years 5 3 8
Definitive test ≥16 years 11 12 23
Definitive tests were done by an independent blind observer. Total 52 48 100
The tests done were cell count by Neubauer’s chamber followed
by differential counting on two centrifuged smears: one stained Table 2: Diagnostic accuracy of urinary reagent strips
with hematoxylin and eosin stain and the other with Leishman Index testCut‑off TP FP FN TN SN SP PPV NPV Accuracy
stain along with protein and sugar estimation by an automated values (%) (%) (%) (%) (%)
analyzer. Leukocytes >15 25 1 3 71 89.28 98.61 96.61 95.94 96
>70 12 0 0 88 100 100 100 100 100
>125 9 0 0 91 100 100 100 100 100
Statistical analysis
Protein >30 66 1 11 22 85.71 95.65 98.50 66.66 88
The diagnostic accuracy of the reagent strip versus the
>100 23 1 1 75 95.83 98.68 95.83 98.68 98
standard tests was estimated using standard statistical tests
>300 7 0 0 93 100 100 100 100 100
and tabulated in the form of sensitivity, specificity, positive
Sugar <50 14 0 15 71 48.2 100 100 82.55 85
predictive value (PPV) and negative predictive value (NPV),
<40 10 0 04 86 71.42 100 100 95.5 96
and positive and negative likelihood ratios.A receiver operating TP: True positive; FP: False positive; FN: False negative; TN: True negative; SN: Sensitivity;
curve (ROC) curve showing the relationship between sensitivity SP: Specificity; PPV: Positive predictive value; NPV: Negative predictive value

Indian Journal of Pathology and Microbiology ¦ Volume 61 ¦ Issue 3 ¦ July-September 2018 357
[Downloaded free from http://www.ijpmonline.org on Monday, May 27, 2019, IP: 14.139.154.130]

Mazumder, et al.: Role of urine reagent strips in CSF analysis

Table 3: Likelihood ratio of the test

Index test Cutoff values LR+ LR−
Leukocytes >15 64.23 0.111
>70 ∞ 0
>125 ∞ 0
Protein >30 19.7 0.150
>100 68.45 0.050
>300 ∞ 0
Sugar <50 ∞ 0.518
<40 ∞ 0.285
LR+: Positive likelihood ratio; LR−: Negative likelihood ratio

DISCUSSION

Meningitis is defined as inflammation of the leptomeninges.

It is associated with high mortality and psychoneurological Figure  1: Receiver operating curve for cells  >15. Area under the
sequelae due to damage to the neurological structures.[10] Delay curve = 98.7%, standard error = 1%
in the treatment of meningitis in neonates and infants can lead
to significant problems in them in the future with respect to
language, motor function, cognition, vision, hearing, and even
behavioral problems.[11,12] It is imperative to initiate treatment
within 3 h to reduce the morbidity and mortality.[13] The diagnosis
greatly depends on the estimate of CSF pleocytosis and glucose
and protein levels of the CSF.[14]

Chikkannaiahet al. have used urinary strips for the semi‑quantitative

analysis of CSF cellularity and chemistry and have observed high
sensitivity (96.6%) and specificity (94.5%) in leukocyte estimation.
The strip showed a sensitivity of 96% and a specificity of 87.1%
for proteins at a cutoff ≥100 mg/dl, while the strip was less specific
at a cutoff of ≥30 mg/dl. With respect to glucose, the strip was
highly specific (100%) and less sensitive at both the cutoff levels
(<40mg/dl and <50mg/dl).[4] Joshi et al. using the reagent strip
conducted a similar study and observed a sensitivity of 85.2% and Figure  2: Receiver operating curve for proteins  >30 area under the
a specificity of 89.6% for leukocytes >10 cells/cumm. For proteins, curve = 92.9%, standard error = 3.1%
at a cutoff of ≥30 mg/dl, the sensitivity was high (98.1%) but the
specificity was low (57.1%);however, at a cutoff of ≥100 mg/
dl, both the sensitivity and specificity were acceptable.The
strip‑based analysis for glucose at a cutoff of ≤40 mg/dl had
high specificity (96.5%) than sensitivity (61.1%); and, even at a
cutoff level of ≤50 mg/dl, the test had high specificity (98%) than
sensitivity (46.2%).[5] Our study was comparable, in terms of all the
three parameters, with the above studies. We also observed that
with respect to proteins, we had 11 false‑negative cases(<30mg/dl
as per strip method),10 out of which actually had laboratory values
of proteins <45mg/dl, which is the standard biochemical cutoff
for normal CSF proteins. Thus, if a strip is designed for this cutoff
level, the accuracy can improve remarkably.

Romanelli et al. compared the results of reagent strips and those

of the standard cytological and biochemical assays and obtained
values for sensitivity, specificity, PPV, and NPV(90.7%, 98.1%, Figure  3: Receiver operating curve for sugar  <50 area under the
95.1%, and 96.4%, respectively) in the diagnosis of bacterial curve = 92.8 with standard error = 2.8%
meningitis.[15] Kumar et al. observed a positive correlation between
the strip and the laboratory values for the diagnosis of meningitis respectively, with P < 0.0001, which was statistically significant.[1]
with κ = 0.94, 0.819, and 0.819 for cells, protein, and glucose, Maclennan et al. in their study observed that additional information
358 Indian Journal of Pathology and Microbiology ¦ Volume 61 ¦ Issue 3 ¦ July-September 2018
[Downloaded free from http://www.ijpmonline.org on Monday, May 27, 2019, IP: 14.139.154.130]

Mazumder, et al.: Role of urine reagent strips in CSF analysis

maybe obtained with the nitrate patch test and they concluded Conflicts of interest
that the nitrate patch, which is also a component of 10‑parameter There are no conflicts of interest.
urinary reagent strip, will become positive in bacterial meningitis
in situations where the granulocytes are increased but have not REFERENCES
released the esterase enzyme.[16]
1. Kumar A, Debata PK, Ranjan A, Gaind R. The role and reliability of rapid
In our study, we additionally observed that the accuracy of the bedside diagnostic test in early diagnosis and treatment of bacterial
tests increased with increase in the values of cell counts and meningitis. Indian J Pediatr 2015;82:311‑4.
2. Scarborough M, Thwaites GE. The diagnosis and management of
proteins and decrease in sugar reaching 100% accuracy for very
acute bacterial meningitis in resource‑poor settings. Lancet Neurol
high test results. This is significant as bacterial meningitis shows 2008;7:637‑48.
very high CSF cell counts and protein levels with very low sugars 3. Valmari P, Peltola H, Ruuskanen O, Korvenranta H. Childhood bacterial
as compared to aseptic meningitis and thus the strip test proves meningitis: Initial symptoms and signs related to age, and reasons for
to be of great relevance in identifying these cases. consulting a physician. Eur J Pediatr 1987;146:515‑8.
4. Chikkannaiah P, Benachinmardi KK, Srinivasamurthy V. Semi‑quantitative
We also observed that the specificity of all the three tests was analysis of cerebrospinal fluid chemistry and cellularity using urinary
very high nearing 100% and is thus instrumental in ruling out reagent strip: An aid to rapid diagnosis of meningitis. Neurol India
meningitis in doubtful cases. In the era of antibiotic resistance, 2016;64:50‑5.
5. Joshi D, Kundana K, Puranik A, Joshi R. Diagnostic accuracy of urinary
this test can be very beneficial in terms of avoiding unnecessary
reagent strip to determine cerebrospinal fluid chemistry and cellularity.
overtreatment in children with febrile seizures or neonates J Neurosci Rural Pract 2013;4:140‑5.
with hypoxic seizures where again the differential diagnosis 6. Parmar RC, Warke S, Sira P, Kamat JR. Rapid diagnosis of meningitis
of meningitis is considered. This is significant in our setup as using reagent strips. Indian J Med Sci 2004;58:62‑6.
we receive the majority of CSF analysis in our laboratory from 7. Molyneux E. Where there is no laboratory, a urine patch test helps
patients in the pediatric age group. diagnose meningitis. J Neurosci Rural Pract 2013;4:117‑8.
8. Oses Salvador JM, Zarallo Cortés L, Cardesa García JJ. Usefulness
This procedure is easy to perform, does not require any technical of reactive strips in the diagnosis of suppurative meningitis, at the
patient’s bedside. An Esp Pediatr 1988;29:105‑8.
expertise, and is a boon to the primary health‑care centers and
9. Moosa AA, Quortum HA, Ibrahim MD. Rapid diagnosis of bacterial
rural setups where no laboratory facilities exist. The operators
meningitis with reagent strips. Lancet 1995;345:1290‑1.
can be trained very easily and competency of the operators can 10. Ray P, Badarou‑Acossi G, Viallon A, Boutoille D, Arthaud M, Trystram D,
be monitored prior to routine use. et al. Accuracy of the cerebrospinal fluid results to differentiate bacterial
from non bacterial meningitis, in case of negative gram‑stained smear.
The only shortcoming is the slight variation in the cutoff normal Am J Emerg Med 2007;25:179‑84.
values of protein and sugar in CSF as compared to urine which 11. de Louvois J, Halket S, Har vey D. Effect of meningitis in
has added to the few false negatives and false positives in our infancy on school‑leaving examination results. Arch Dis Child
study. This can be overcome by designing strips specific for CSF 2007;92:959‑62.
12. Krebs VL, Costa GA. Clinical outcome of neonatal bacterial
analysis, at the manufacturer’s level.
meningitis according to birth weight. Arq Neuropsiquiatr
2007;65:1149‑53.
CONCLUSION 13. Brancusi F, Farrar J, Heemskerk D. Tuberculous meningitis in adults:
A review of a decade of developments focusing on prognostic factors
Therefore, in light of our study and with reference to the for outcome. Future Microbiol 2012;7:1101‑16.
results obtained in similar studies by various observers, we can 14. Chaudhuri A, Martinez‑Martin P, Kennedy PG, Andrew Seaton R,
conclude that urinary reagent strips can be used to facilitate Portegies P, Bojar M, et al. EFNS guideline on the management of
therapeutic decisions in cases of suspected meningitis, especially community‑acquired bacterial meningitis: Report of an EFNS task force
in resource‑constrained settings. Further studies with larger on acute bacterial meningitis in older children and adults. Eur J Neurol
2008;15:649‑59.
numbers and comparison with different manufacturer strips
15. Romanelli RM, Thome EE, Duarte FM, Gomes RS, Camargos PA,
can be undertaken. If standardized, this method can revamp the Freire HB, et al. Diagnosis of meningitis with reagent strips.
entire diagnostic protocol of meningitis. J Pediatr (Rio J) 2001;77:203‑8.
16. Maclennan C, Banda E, Molyneux EM, Green DA. Rapid diagnosis
Financial support and sponsorship of bacterial meningitis using nitrite patch testing. Trop Doct
Nil. 2004;34:231‑2.

Indian Journal of Pathology and Microbiology ¦ Volume 61 ¦ Issue 3 ¦ July-September 2018 359