Accepted Manuscript

Title: Molecular Diagnosis of Rickettsia Infection in Patients

from Tunisia

Author: Fatma Khrouf Hanene Sellami Emna Elleuch

Zouhour Hattab Lamia Ammari Moncef Khalfaoui Jihed
Souissi Hejer Harrabi Youmna M’ghirbi Hanene Tiouiri
Mounir ben jemaa Adnene Hammami Amel Letaief Ali
Bouattour Abir Znazen

PII: S1877-959X(16)30024-3
DOI: http://dx.doi.org/doi:10.1016/j.ttbdis.2016.02.010
Reference: TTBDIS 616

To appear in:

Received date: 24-7-2015

Revised date: 5-2-2016
Accepted date: 6-2-2016

Please cite this article as: Khrouf, F., Sellami, H., Elleuch, E., Hattab, Z., Ammari,
L., Khalfaoui, M., Souissi, J., Harrabi, H., M’ghirbi, Y., Tiouiri, H., jemaa, M.,
Hammami, A., Letaief, A., Bouattour, A., Znazen, A.,Molecular Diagnosis of
Rickettsia Infection in Patients from Tunisia, Ticks and Tick-borne Diseases (2016),
http://dx.doi.org/10.1016/j.ttbdis.2016.02.010

This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.
Molecular Diagnosis of Rickettsia Infection in Patients from Tunisia

Author names and affiliations:

Fatma Khroufa, Hanene Sellamib, Emna Elleuchc, Zouhour Hattabd, Lamia Ammarie, Moncef

Khalfaouif, Jihed Souissid, Hejer Harrabie, Youmna M’ghirbia, Hanene Tiouirie, Mounir ben

t
jemaac, Adnene Hammamib, Amel Letaiefd, Ali Bouattoura# and Abir Znazenb.

ip
a: Laboratory of entomology, Pasteur Institute of Tunis, Tunis, Tunisia.

cr
b: Laboratory of Microbiology, Research Laboratory “MPH”, Habib Bourguiba University

Hospital of Sfax, Sfax University, Sfax, Tunisia.

us
c: Infectious diseases department, Hedi Chaker University Hospital of Sfax, Sfax University,

an
Sfax, Tunisia.

d: Infectious diseases department, Farhat Hached University Hospital of Sousse, Sousse,

M
Tunisia,, Tunisia.

e: Infectious diseases department, La Rabta University Hospital of Tunis, Tunis University,

d

Tunisia.
te

f: Infectious diseases department Menzel Bourguiba regional hospital, Bizerte, Tunisia.

# Corresponding author: Ali Bouattour

p

ali.bouattour@pasteur.rns.tn
ce

Institut Pasteur de Tunis, BP74, 1002 Tunis Bélvèdère, TUNISIA

Ac

Phone: +216 71 893 340

Fax : +216 71 791 833

1 
 
Page 1 of 13
Abstract

Diagnosis of rickettsioses had largely benefited from the development of molecular

techniques. Unfortunately, in Tunisia, despite the large number of rickettsial cases registered

t
ip
every year, the Rickettsia species remain unidentified. In this study, we aimed to detect the

cr
Rickettsia species in clinical samples using molecular tests. A study was established to

analyze skin biopsies, cutaneous swabs, and cerebrospinal fluid samples taken from clinically

us
suspected patients to have rickettsial infection. Two molecular techniques were used to detect

Rickettsia DNA: Quantitative Real time PCR (qPCR) and Reverse Line Blot test (RLB). An

an
analysis of the RLB hybridization assay results revealed the presence of Rickettsia DNA in

skin biopsies (40.6%) and swabs (46.7%). R. conorii was the most prevalent identified species
M
among tested samples. Other species of interest include R. typhi and R. massiliae. Using

qPCR positivity rates in skin biopsies was 63.7% against 80% in swabs. R. conorii was the
d

most frequently detected species, followed by R. typhi. The agreement between the two
te

techniques was 68.6% (kappa = 0.33). Molecular tests, especially using specific probes qPCR,
p

allow for a rapid, better and confident diagnosis in clinical practice. They improve the survey
ce

of Mediterranean spotted fever which is considered to be the most important rickettsial

infection in humans in Tunisia.

Ac

KEY WORDS: Rickettsia, skin biopsy, swab, cerebrospinal fluid RLB, qPCR

2 
 
Page 2 of 13
Introduction

Rickettsioses are emerging infectious diseases caused by Rickettsia and transmitted by the

bites of arthropods (ticks and fleas in particular). Although classified as neglected diseases,

t
ip
they continue to cause severe illnesses and death in adults and children worldwide (Parola et

al., 2013). A delayed diagnosis can lead to serious complications -- acute renal failure,

cr
meningoencephalitis, gastrointestinal bleeding, and multiple organ failure -- and death (Boillat

us
et al., 2008; Charra et al., 2005; Jensenius et al., 2004). An early, rapid diagnosis is the

greatest clinical challenge for a successful treatment. A serology test is retrospective and

an
infections are underestimated because control serums are rarely taken. Real-time quantitative

PCR (qPCR) and Reverse Line Blot (RLB) targeting a specific gene are currently widespread
M
techniques with good sensitivity and specificity for detecting rickettsiosis (Jado et al., 2006;

Renvoisé et al., 2012), but, few studies have compared them. We conducted a multicenter
d

study using these two molecular techniques to identify rickettsial species in human samples
te

and compared their diagnostic accuracy to develop a strategy to more rapidly and precisely

identifying the responsible species.

p
ce

Materials and Methods

Human sampling
Ac

A study was conducted in 2012-2014 from June to October, when higher exposure to Ticks

and /or fleas produces the greatest number of rickettsiosis cases. A total of 121 samples from

101 symptomatic adult patients (excluding children) with acute fever and cutaneous rashes

and/or eschars in 5 hospitals (Fig. 1) was collected. The samples included 69 skin biopsies

(from 35 men and 34 women), 15 cutaneous swabs (from the eschar site (9 men and 6

women)) and 37 cerebrospinal fluid (CSF) samples (23 men and 14 women). Skin biopsies

3 
 
Page 3 of 13
were taken using a punch (3.5 mm of diameters) from eschar, when they existed, or from the

rash. The distribution of patients was mapped with ArcGIS 9 (version 9.3.1) software.

DNA extraction and QPCR amplification

DNA was extracted from all samples using the QIAamp DNA tissue extraction kit (Qiagen,

t
Hilden, Germany) as per instructions. Biopsied tissues weighed between 18-20 mg. DNA was

ip
eluted in100 μl and extracts stored at -20°C.

cr
All samples were screened for Rickettsia spp. DNA using qPCR amplification (Rickettsia spp-

qPCR) (Renvoisé et al., 2012). The positive screening was followed by runs of qPCRs with

us
specific probes for R. conorii, R. typhi, R. felis and R. africae, respectively (Giulieri et al.,

an
2012; Renvoisé et al., 2012). In ambiguous cases, the Rickettsia infection was checked using

primers and probes for the Rickettsia Spotted Fever Group (SFG) (RC00338) and for the
M
Typhus group, Rickettsia (RP278) described by Renvoisé et al., (2012). DNA positive

controls and amplification programs were used as described (Znazen et al., 2015). QPCR
d

amplifications and product detections were carried out in the CFX96 Touch Real-Time PCR
te

Detection System (Biorad, US). To determine the quantity of DNA in each sample, five

standards (104, 103, 102, 10, 1 copies/μl) were tested per run and the quantity subsequently
p

measured using Bio-Rad CFX Manager software. A sample was considered positive when the
ce

Ct (cycle number at threshold of log-based fluorescence) was below 35 (≈18-35 copies of

Rickettsia gene/PCR reaction).

Ac

PCR and Reverse Line Blot (RLB) hybridization assay

All DNA samples were tested twice using the RLB hybridization targeting 23S-5S rRNA

gene region assay as described by Khrouf et al., (2014). To monitor cross contamination and

false-positive results, negative controls from DNA extraction and PCR amplification were

included in each batch.

Statistical analysis

4 
 
Page 4 of 13
The agreement between the two tests was calculated with the widely-used kappa value using

the MedCalc program which also used to determine RLB specificity and sensitivity. A kappa

of 0 indicates no agreement beyond chance; a kappa of 1 indicates perfect agreement. For data

comparison, the Student’s t-test or Fischer test was performed using EpiInfo 6.0 software.

t
Statistical significance was defined as p<0.05.

ip
Results

cr
Prevalence and species detection using qPCR and RLB

Rickettsia spp- DNA was detected by qPCR in 56 samples: 44 skin biopsies (63.7%) and 12

us
swabs (80%). RLB revealed Rickettsia DNA in 35 samples: 28 skin biopsies (40.6%) and 7

an
swab samples (43.7%). Neither test revealed any CSF positive samples. Among 58 patients

(57%) infected by Rickettsia spp. by either qPCR or RLB, R. conorii was identified in 48
M
cases (80.8%) and R. typhi in 5 (10.6%). The remaining 5 cases were negative in either R. felis

or R. africae species specific qPCR assays. However, a skin biopsy sample was positive for
d

SFG-qPCR; and showed hybridization by RLB assay with a R. massiliae probe.

te

Using the two techniques, the differences of Rickettsia prevalence rates between the 5 sites

were not statistically significant (p=0.13) (Fig.1). The Rickettsial infection rate was 66% and
p

68% for men and women respectively are not statistically significant (p=0.1).
ce

Tests Comparison

When compared to qPCR, RLB had a sensitivity of 46.4% (95% CI, 32.9 to 60.2), a
Ac

specificity of 86.1% (95% CI, 75.3 to 93.5), and an AUC of 0.66 (95% CI, 0.57 to 0.75); test

agreement was 68.6%. Comparing RLB and qPCR, the Kappa value was 0.33 with a standard

error =0.081 and 95%CI [0.0175 to 0.494]. Both techniques were positive in 26 samples. RLB

was negative in 30 samples that proved positive by Rickettsia spp-qPCR. By RLB, 9 samples

were positive and Rickettsia spp. qPCR negative (Table 1). R. conorii was detected in 42

samples with qPCR and 33 samples by RLB. R. typhi was identified in 5 rash skin biopsies by

5 
 
Page 5 of 13
PCR and in one case by RLB. By qPCR, the Rickettsia infection remains unidentified in 9

samples (6 skin biopsies; 3 swabs). However, DNA showed hybridization by RLB assay with

R. massiliae in one skin biopsy and with R. conorii probe in four samples (3 skin biopsies;

one swab). Indeed, 4 patients had unidentified rickettsial DNA.

t
Comparison of the sample type using qPCR

ip
For 10 patients, both samples (eschar swab, skin biopsy) were collected. Positive rates using

cr
qPCR were 70% (7/10) and 80% (8/10), respectively. Concordant results were found in 9

cases by qPCR: 7 cases were positive and 2 were negative in both samples. There was a single

us
discordant result showing positive qPCR on the skin biopsy only. Comparing mean DNA

an
quantities detected in skin biopsies and swabs showed no statistical difference for the

Rickettsia spp-qPCR (p=0.28). Both samples yielded the same results for species
M
identification.

Discussion
d

Our results provide new insights into molecular identification and evidence of Rickettsia
te

infection in humans in Tunisia. Both QPCR and RLB tests indicated a high prevalence of R.

conorii exposure in skin biopsies and cutaneous swab samples. QPCR was rapid and more
p

efficient than RLB for detecting Rickettsia, due perhaps to weak hybridization and/or cross
ce

reactivity of probes in RLB. The qPCR used in this study had been evaluated in our laboratory

as having good sensitivity (Znazen et al., 2015). Indeed, qPCR assays are widely used for
Ac

rapid detection and identification of fastidious organisms such as Rickettsiae in clinical

samples (Eygelaar et al., 2015; Kidd et al., 2008). RLB is a flexible technique permitting the

simultaneous detection of different targets especially in coinfections, and adequate for both

epidemiological studies and clinical diagnosis (Khrouf et al., 2014).

In this study, qPCR (66.7%) was more sensitive than RLB (41.7%) in detecting Rickettsia in

suspected infected samples. Both techniques had weaknesses: 30 samples were positive using

6 
 
Page 6 of 13
qPCR and were negative using RLB. In these samples, we detected a low copy number of

Rickettsia in biopsies and swabs -- about 18-90 Rickettsia gene/PCR reaction. Nine samples (8

skin biopsies; 1 swab) were positive by RLB and negative by Rickettsia spp-qPCR. When

tested by R. conorii qPCR, only 2 of these were weakly positive (a ct value of more than 35

t
cycles). Regarding the 4 negative Rickettsia conorii qPCR and R. typhi qPCR, DNA samples

ip
were caught using the R. conorii RLB probe. The R. conorii qPCR test carried out for R.

cr
conorii israelensis was negative. Since, two R. conorii subspecies (R. conorii conorii and R.

conorii israelensis were previously described in Tunisia, we suggested that these patients

us
were probably infected by R. conorii israelensis (Znazen et al., 2011; Khrouf et al., 2014).

an
RLB and qPCR tests all indicated a high prevalence of Rickettsia exposure in the two types of

clinical samples, confirming that swab sampling is as efficient as skin biopsies for diagnosing
M
Rickettsia (Bechah et al., 2011; Wang et al., 2009). Using qPCR, cutaneous swabs, all

performed in the eschar, showed a positive rate of 80% (12/15) compared to 63.7% (44/69)
d

for skin biopsies. The difference that may be due to the fact that skin biopsies were performed
te

on both eschar (n= 36) and cutaneous rash (n= 33). For Rickettsia DNA quantification, a

comparison of mean DNA quantities detected in skin biopsies from eschar and from
p

cutaneous rash showed no statistical difference for the Rickettsia spp-qPCR (p=0.36). For
ce

both sample sources, the DNA quantity was 10 copies to 103 copies of Rickettsia gene/PCR

reaction. Our results showed a high rate of infection by R. conorii, the agent of Mediterranean
Ac

Spotted Fever endemic in the Mediterranean, and particularly in Tunisia. The finding has been

confirmed by various serological and molecular previous studies (Sfar et al., 2009; Kaabia et

al., 2006; Kaabia and Letaief, 2009). A recent serological study in Tunisia confirmed and

highlighted the presence of R. conorii, R. typhi and R. felis in 197 out of 380 suspected cases

(Hattab et al., 2014). In our study, R. massiliae was identified by RLB in a skin biopsy of one

patient, which is unsurprising given its detection in ticks (Khrouf et al., 2014), and also

7 
 
Page 7 of 13
widely described as causing human and vector infections in the Mediterranean region (Harrus

et al., 2011; Boudebouch et al., 2009; Vitale et al., 2006). Additional diagnostic methods

should be used to confirm the role of this species in human cases.

We mentioned the detection in 5 skin rash biopsies of R. typhi, the agent of Murine typhus,

t
confirming other studies in Tunisia (Znazen et al., 2013; Letaïef et al., 2005). The result was

ip
confirmed by sequencing the partial sequence of gltA gene of one sample (data not published).

cr
Although Mediterranean Spotted Fever caused severe complications, including neurological

syndromes in 10% of the cases (Ezpeleta et al., 1999), none of the tested CSF was positive for

us
Rickettsia qPCR or RLB. However, some of these CSF were collected from patients

an
presenting positive skin biopsies for R. conorii qPCR. Rickettsia penetrates the CNS barrier

and causes neurological complications only in late untreated cases; none of the cases studied
M
seem to meet this criterion.

The negative results using molecular assays do not exclude Rickettsia infection. To increase
d

the sensitivity of qPCR, skin biopsies should be sampled before treatment early in the course
te

of the disease.

Conflict of interest
p

The authors declare that the research was conducted in the absence of any commercial and
ce

financial relationships that could be construed as potential conflict of interest.

Ethics statement
Ac

This study was approved by our institutional review board the “Habib Bourguiba University

hospital ethics committee” with the given number 11–13. All test subjects (all were adults, no

children were included) provided an informed written consent.

Acknowledgments

We thank Prof. Didier Raoult for providing antigens used in the serology and all Rickettsia

DNA used to construct positive recombinant plasmid controls. We also thank Dalinda

8 
 
Page 8 of 13
Charfeddine, Radhia Turki and Lamia Charfi for their technical help. We thank Deborah

Glassman for her comments and corrections on the early drafts of the manuscript.

This work was financed by the IEVP coopération transfrontalière Italie-Tunisie 2007-2013

PROJET 2 PS1.3.023-RESTUS.

t
ip
References

cr
Bechah, Y., Socolovschi, C., Raoult, D., 2011. Identification of Rickettsial Infections by
Using Cutaneous Swab Specimens and PCR. Emerg. Infect. Dis. 17, 83–86.
doi:10.3201/eid1701.100854

us
Boillat, N., Genton, B., D’Acremont, V., Raoult, D., Greub, G., 2008. Fatal case of Israeli
spotted fever after Mediterranean cruise. Emerg. Infect. Dis. 14, 1944–1946.
doi:10.3201/eid1412.070641
Boudebouch, N., Sarih, M., Socolovschi, C., Amarouch, H., Hassar, M., Raoult, D., Parola,

an
P., 2009. Molecular survey for spotted fever group rickettsiae in ticks from Morocco.
Clin. Microbiol. Infect. 15, 259–260. doi:10.1111/j.1469-0691.2008.02226.x
Charra, B., Berrada, J., Hachimi, A., Judate, I., Nejmi, H., Motaouakkil, S., 2005. [A fatal
case of Mediterranean spotted fever]. Médecine Mal. Infect. 35, 374–375.
M
doi:10.1016/j.medmal.2005.03.012
Eygelaar, D., Jori, F., Mokopasetso, M., Sibeko, K.P., Collins, N.E., Vorster, I., Troskie, M.,
Oosthuizen, M.C., 2015. Tick-borne haemoparasites in African buffalo (Syncerus
caffer) from two wildlife areas in Northern Botswana. Parasit. Vectors 8.
d

doi:10.1186/s13071-014-0627-y
Ezpeleta, D., Muñoz-Blanco, J.L., Tabernero, C., Giménez-Roldán, S., 1999. [Neurological
te

complications of Mediterranean boutonneuse fever. Presentation of a case of acute

encephalomeningomyelitis and review of the literature]. Neurol. Barc. Spain 14, 38–
42.
p

Giulieri, S., Jaton, K., Cometta, A., Trellu, L.T., Greub, G., 2012. Development of a duplex
real-time PCR for the detection of Rickettsia spp. and typhus group rickettsia in
ce

clinical samples. FEMS Immunol. Med. Microbiol. 64, 92–97. doi:10.1111/j.1574-

695X.2011.00910.x
Harrus, S., Perlman-Avrahami, A., Mumcuoglu, K.Y., Morick, D., Baneth, G., 2011.
Molecular detection of Rickettsia massiliae, Rickettsia sibirica mongolitimonae and
Ac

Rickettsia conorii israelensis in ticks from Israel. Clin. Microbiol. Infect. 17, 176–180.
doi:10.1111/j.1469-0691.2010.03224.x
Jado, I., Escudero, R., Gil, H., Jimenez-Alonso, M.I., Sousa, R., Garcia-Perez, A.L.,
Rodriguez-Vargas, M., Lobo, B., Anda, P., 2006. Molecular Method for Identification
of Rickettsia Species in Clinical and Environmental Samples. J. Clin. Microbiol. 44,
4572–4576. doi:10.1128/JCM.01227-06
Jensenius, M., Fournier, P.-E., Raoult, D., 2004. Rickettsioses and the international traveler.
Clin. Infect. Dis. Off. Publ. Infect. Dis. Soc. Am. 39, 1493–1499. doi:10.1086/425365
Kaabia, N., Letaief, A., 2009. Characterization of Rickettsial Diseases in a Hospital-Based
Population in Central Tunisia. Ann. N. Y. Acad. Sci. 1166, 167–171.
doi:10.1111/j.1749-6632.2009.04521.x

9 
 
Page 9 of 13
Kaabia, N., Rolain, J.., Khalifa, M., Jazia, E.B., Bahri, F., Raoult, D., Letaief, A., 2006.
Serologic Study of Rickettsioses among Acute Febrile Patients in Central Tunisia.
Ann. N. Y. Acad. Sci. 1078, 176–179. doi:10.1196/annals.1374.126
Khrouf, F., M’Ghirbi, Y., Znazen, A., Ben Jemaa, M., Hammami, A., Bouattour, A., 2014.
Detection of Rickettsia in Rhipicephalus sanguineus Ticks and Ctenocephalides felis
Fleas from Southeastern Tunisia by Reverse Line Blot Assay. J. Clin. Microbiol. 52,
268–274. doi:10.1128/JCM.01925-13
Kidd, L., Maggi, R., Diniz, P.P.V.P., Hegarty, B., Tucker, M., Breitschwerdt, E., 2008.

t
Evaluation of conventional and real-time PCR assays for detection and differentiation

ip
of Spotted Fever Group Rickettsia in dog blood. Vet. Microbiol. 129, 294–303.
doi:10.1016/j.vetmic.2007.11.035
Letaïef, A.O., Kaabia, N., Chakroun, M., Khalifa, M., Bouzouaia, N., Jemni, L., 2005.

cr
Clinical and laboratory features of murine typhus in central Tunisia: a report of seven
cases. Int. J. Infect. Dis. 9, 331–334. doi:10.1016/j.ijid.2004.09.011
Parola, P., Paddock, C.D., Socolovschi, C., Labruna, M.B., Mediannikov, O., Kernif, T.,

us
Abdad, M.Y., Stenos, J., Bitam, I., Fournier, P.-E., Raoult, D., 2013. Update on Tick-
Borne Rickettsioses around the World: a Geographic Approach. Clin. Microbiol. Rev.
26, 657–702. doi:10.1128/CMR.00032-13

an
Renvoisé, A., Rolain, J.-M., Socolovschi, C., Raoult, D., 2012. Widespread use of real-time
PCR for rickettsial diagnosis. FEMS Immunol. Med. Microbiol. 64, 126–129.
doi:10.1111/j.1574-695X.2011.00899.x
Sfar, N., Kaabia, N., Letaief, A., Rolain, J.M., Parola, P., Bouattour, A., Raoult, D., 2009.
M
First molecular detection of R. conorii subsp. conorii 99 years after the Conor
description of Mediterranean spotted fever, in Tunisia. Clin. Microbiol. Infect. 15,
309–310. doi:10.1111/j.1469-0691.2008.02253.x
Vitale, G., Mansueto, S., Rolain, J.-M., Raoult, D., 2006. Rickettsia massiliae Human
d

Isolation. Emerg. Infect. Dis. 12, 174–175. doi:10.3201/eid1201.050850

Wang, J.-M., Hudson, B.J., Watts, M.R., Karagiannis, T., Fisher, N.J., Anderson, C., Roffey,
te

P., 2009. Diagnosis of Queensland Tick Typhus and African Tick Bite Fever by PCR
of Lesion Swabs. Emerg. Infect. Dis. 15, 963–965. doi:10.3201/eid1506.080855
Znazen, A., Hammami, B., Mustapha, A.B., Chaari, S., Lahiani, D., Maaloul, I., Jemaa, M.B.,
p

Hammami, A., 2013. Murine typhus in Tunisia: A neglected cause of fever as a single
symptom. Med. Mal. Infect. doi:10.1016/j.medmal.2013.02.007
ce

Znazen, A., Hammami, Boussayma, Lahiani, Dorra, Ben Jemaa, Mounir, Hammami, Adnene,
2011. Israeli Spotted Fever, Tunisia. Emerg. Infect. Dis. 17, 1328–1330.
doi:10.3201/eid1707.101648
Znazen, A., Sellami, H., Elleuch, E., Hattab, Z., Ben Sassi, L., Khrouf, F., Dammak, H.,
Ac

Letaief, A., Ben Jemaa, M., Hammami, A., others, 2015. Comparison of two
quantitative real time PCR assays for Rickettsia detection in patients from Tunisia.
PLoS Negl Trop Dis 9, e0003487.

10 
 
Page 10 of 13
Table 1. Comparison of qPCR and RLB positivity rate for Rickettsia infection in collected

samples.

Figure legend

t
Figure 1. Distribution of included patients and species detection in confirmed infected patients

ip
by qPCR and/or RLB.

cr
us
an
M
d
p te
ce
Ac

11 
 
Page 11 of 13
Table 1. Comparison of qPCR and RLB positivity rate for Rickettsia infection in collected

samples.

qPCR

Skin biopsies + -

t
(n=69)

ip
+ 21 7

cr
- 23 18

+ -

us
Cutaneous swabs

RLB (n=15)

an
+ 6 2

- 6 1
M
CSF (n=37) + -

+ 0 0
d

- 0 37
te

n= number of samples

CFS= cerebrospinal fluid

p
ce
Ac

12 
 
Page 12 of 13
 Ac
ce
pt
ed
M
an
us
cr
ip
t

Page 13 of 13