G Model

JIPH-770; No. of Pages 5 ARTICLE IN PRESS

Journal of Infection and Public Health xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Journal of Infection and Public Health

journal homepage: http://www.elsevier.com/locate/jiph

Risk factors for poor prognosis in children with refractory purulent

meningitis and the discharge criteria
Hai-Lun Peng a,b,c,d, Yue Hu a,b,c,d,∗, Hong-Jia Chen a,b,c,d, Pan-Pan Song a,b,c,d, Li Jiang a,b,c,d
a
Department of Neurology, Children’s Hospital of Chongqing Medical University, China
b
Ministry of Education Key Laboratory of Child Development and Disorders, China
c
Key Laboratory of Pediatrics in Chongqing, China
d
Chongqing International Science and Technology Cooperation Center for Child Development and Disorders, China

a r t i c l e i n f o a b s t r a c t

Article history: This study was undertaken to investigate the relationship between cerebrospinal fluid abnormalities
Received 1 November 2016 and prognosis in pediatric refractory purulent meningitis. Ninety cases of pediatric refractory purulent
Received in revised form 14 June 2017 meningitis were stratified into “good” (n = 33) or “poor” (n = 57) prognosis groups according to the Glas-
Accepted 9 July 2017
gow clinical outcome scores. The symptoms, laboratory results, and prognosis were compared by using
univariate and multivariate logistic regression analyses. Univariate analysis showed that poor prognosis
Keywords:
was associated with: unequal pupil size in both eyes; positive Babinski sign; CSF-WBC >500 × 106 /L, CSF
Children
protein concentration >1.0 g/L, CSF glucose content <1.5 mmol/L; initial procalcitonin result >0.1 ng/dL
Refractory purulent meningitis
Prognosis
on admission; hemoglobin <90 g/L during hospitalization; abnormal head imaging, and abnormal elec-
Discharge criteria troencephalogram. On multivariate analysis only unequal pupil size in both eyes and CSF glucose content
Antibiotic course <1.5 mmol/L remained significant. The CSF protein concentration was significantly different between
groups at discharge. The cutoff value was 0.68 g/L. We recommend that discharged patients meet the
following criteria: full antibiotic course and over 1 week of defervesce, disappearance of acute phase
symptoms, CSF-WBC ≤28 × 106 /L, CSF glucose >1.75 mmol/L, and protein <0.68 g/L. The patient may be
discharged for follow-up if no relapse occurs during 3–5 days of observation after drug withdrawal.
© 2017 Published by Elsevier Limited on behalf of King Saud Bin Abdulaziz University for Health
Sciences. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/).

Introduction
Purulent meningitis, also known as acute bacterial meningitis, is
a common infectious disease that affects the central nervous system
in children. The main clinical features of the disease include fever,
increased intracranial pressure, meningeal irritation, and purulent
changes to the cerebrospinal fluid (CSF) [1]. Purulent meningitis
is an important cause of morbidity and mortality in neonates and
children [2–9]. Mortality can reach 100% in pediatric patients with-
out any treatment [10]. The onset of refractory purulent meningitis
Abbreviations: CSF, cerebrospinal fluid; EEG, electroencephalography; CNS,
central nervous system; GOS, Glasgow Outcome Score criteria; ROC, operating
characteristic curve; WBC, white blood cell; CMA, China Medical Association; PCT,
procalcitonin.
∗ Corresponding author at: Department of Neurology, Children’s Hospital of
Chongqing Medical University, No.136 Zhongshan 2nd Road, Yu Zhong District,
Chongqing 400014, China. Fax: +86 23 67460022.
E-mail addresses: 158156396@qq.com (H.-L. Peng), huyue915@163.com (Y. Hu),
topmax@163.com (H.-J. Chen), 34186586@qq.com (P.-P. Song),
jiangli19640718@163.com (L. Jiang).
is acute, it develops rapidly, and mortality and morbidity are higher
than those of common purulent meningitis. Some pediatric patients
have poor therapeutic outcomes such as minimal effectiveness of
conventional antibiotic therapy, relapse rapidly after showing signs
of improvement on treatment, CSF remains abnormal, or in some
cases there are serious sequelae including death. It is unclear why
the disease is more refractory in some patients than others and
why some patients have a better prognosis than others with a sim-
ilar disease course. Therefore, early recognition of the symptoms of
severe disease and understanding the risk factors resulting in poor
prognosis for refractory purulent meningitis is critical to reducing
the mortality and morbidity of purulent meningitis [11].
This study was undertaken with multiple objectives: (1) to
understand the risk factors for poor prognosis in pediatric patients
with refractory purulent meningitis; (2) characterize the relation-
ship between CSF recovery and the prognosis of refractory purulent
meningitis; and (3) provide evidence based recommendations for
an appropriate antibiotic treatment course, and discharge criteria.
The study is a retrospective analysis of clinical data and outcomes
for 90 children with refractory purulent meningitis in the Children’s

http://dx.doi.org/10.1016/j.jiph.2017.07.007
1876-0341/© 2017 Published by Elsevier Limited on behalf of King Saud Bin Abdulaziz University for Health Sciences. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: Peng H-L, et al. Risk factors for poor prognosis in children with refractory purulent meningitis and
the discharge criteria. J Infect Public Health (2017), http://dx.doi.org/10.1016/j.jiph.2017.07.007
G Model
JIPH-770; No. of Pages 5 ARTICLE IN PRESS
2 H.-L. Peng et al. / Journal of Infection and Public Health xxx (2017) xxx–xxx

Hospital of Chongqing Medical University from February 2011 to
July 2014.
Material and methods
Subject inclusion and exclusion criteria
For inclusion in the study, subjects had to meet the following
criteria: (a) age ≥1 month; (b) symptoms and findings consistent
with the diagnostic criteria for purulent meningitis [2]; and (c)
pediatric patients meeting one or more of the following criteria
were classified as having refractory purulent meningitis [12]: (i)
severe disturbance of consciousness (prolonged coma); (ii) cere-
bral hernia; (iii) repeated or prolonged convulsions; (iv) abnormal
plate-like brain parenchymal loci on head CT or MRI; (v) compli-
cations persisting during late follow-up, such as subdural effusion,
ependymitis and hydrocephalus; (vi) death, or sequelae during late
follow-up such as secondary epilepsy, cranial nerve damage, and
psychomotor retardation; (vii) treatment for >3 weeks according
to a regular treatment program (except for a prolonged course
of treatment caused by nosocomial infections or underlying dis-
ease); and (viii) unexplained recurrent suppurative intracranial
infections.
This study was approved by the Ethics Committee of the Chil-
dren’s Hospital affiliated with Chongqing University of Medical
Sciences. Informed consent was obtained from the parents or legal
guardians of the subjects via signed consent forms.
Subjects were excluded from the study if they met the following
criteria: (i) cases complicated with viral encephalitis, tubercular
meningitis or cryptococcus meningitis and (ii) cases in which other
central nervous system (CNS) disorders could not be excluded.
chi-square test. Multivariate logistic regression analysis was used
for the indicators with statistical significance in the univariate sta-
tistical analysis. A receiver operating characteristic (ROC) curve was
used to determine the best cut-off value for diagnosis. Statistical
significance was established at P < 0.05.
Results
For the ninety cases, the onset age range was from 1-month-
old to 14-years-old. The average duration of antibiotic usage in this
study was 37 days, and the range was 6–76 days. Subjects were
followed for a minimum of 8 months and a maximum of 4 years.
Of the 90 enrolled subjects, 33 (33/90; 36.67%) were included in
the “good prognosis” group. Of these, 9 cases were cured and 24
cases showed improvement. The “poor prognosis” group included
the remaining 57 subjects (57/90; 63.33%). Of these, 12 cases died
(12/90; 13.33%), and none of the subjects underwent an autopsy.
The incidences of sequelae in the “poor prognosis” group were: 35
cases of mental retardation (35/90; 38.89%), 9 cases of secondary
epilepsy (9/90; 10%), 6 cases of hydrocephalus (6/90; 6.67%), 4
cases of hearing impairment (4/90; 4.44%), and 3 cases of move-
ment disorders (3/90; 3.33%). Eleven (11) cases (11/90; 12.22%)
were complicated with two or more sequelae, including 4 cases of
mental retardation complicated with secondary epilepsy, 3 cases
of mental retardation complicated with hydrocephalus, 2 cases of
mental retardation complicated with hearing impairment, 1 case
of mental retardation complicated with movement disorders, and
1 case of mental retardation complicated with hydrocephalus and
movement disorders.
Etiology and treatment

Methods
CSF culture was performed in 90 patients. CSF smears were pos-
itive in 4 patients (Gram-positive cocci), while CSF cultures were
Ninety (n = 90) cases met the inclusion criteria in the neurology
positive in 19 patients (total: 19/90, 21.11%). Blood cultures were
ward in the Children’s Hospital of Chongqing Medical University
performed in 90 patients. Blood smear was positive in only 1 patient
between February 2011 and July 2014. The cases were divided into
(Gram-positive cocci), while blood cultures were positive in 35
the “good prognosis” group and “poor prognosis” group based on
patients (total: 35/90, 38.89%) (Table 2). The isolates from CSF and
the clinical Glasgow Outcome Score criteria (GOS; Table 1) [13]. The
blood matched.
cases scoring a 5 on the GOS were stratified to the “good prognosis”
After admission, antibiotics were administered to all 90 patients
group and those scoring 1–4 points were stratified to the “poor
(100%). The empirical antimicrobial therapy for purulent meningi-
prognosis” group.
tis was based on patient age and specific predisposing condition.
Targeted antimicrobial therapy was based on positive pathogen
Statistical analysis identification by CSF or blood cultivation. Surgery was performed
in 25 patients (25/90, 27.78%) (Table 3).
Statistical analysis was performed using SPSS22.0 statistical
software package. Measurement data were analyzed with the
t test and rank sum test; and count data were analyzed with Table 2
the chi-square test. If measurement data were not normally dis- The etiology of CSF and blood cultures.

tributed, then they were converted to count data to be analyzed by Item Etiology N

CSF Streptococcus pneumoniae 9

Table 1 cultures Escherichia coli 6
Glasgow outcome score criteria. Enterobacter cloacae 1
Staphylococcus aureus + Streptococcus pneumoniae 1
Score Criteria Klebsiella pneumoniae subsp. pneumoniae + Acinetobacter 1
baumannii
5 Cure: purulent meningitis symptoms and the signs have disappeared
Klebsiella pneumoniae subsp. pneumonia + Staphylococcus 1
completely, two consecutive CSF values are normal, peripheral blood is
aureus
normal, no complications or sequelae are present. Significant
improvement: purulent meningitis symptoms and signs are relieved Blood Streptococcus 15
significantly, CSF results are basically normal, various complications cultures Staphylococcus 9
are improved, no sequelae. Escherichia coli 4
4 Frequent seizures or movement disorders occur due to central nervous Klebsiella pneumoniae 2
system dysfunction, CSF values show no improvement, no Listeria monocytogenes 1
complications mitigation is available Pasteurella pneumotropica, 1
3 Severe central nervous system disorders Staphylococcus aureus + Streptococcus pneumoniae 1
2 Vegetative state Staphylococcus hominis subspecies + Staphylococcus epidermidis 1
1 Death Staphylococcus haemolyticus + Staphylococcus capitis 1

Please cite this article in press as: Peng H-L, et al. Risk factors for poor prognosis in children with refractory purulent meningitis and
the discharge criteria. J Infect Public Health (2017), http://dx.doi.org/10.1016/j.jiph.2017.07.007
G Model
JIPH-770; No. of Pages 5 ARTICLE IN PRESS
H.-L. Peng et al. / Journal of Infection and Public Health xxx (2017) xxx–xxx 3

Table 3 refractory meningitis was the dependent variable. Two factors

The treatment for antibiotic and surgery.
remained significantly associated with the risk for poor prognosis
Treatment N in subjects with severe meningitis, unequal pupil size in both eyes
Meropenem 4 (regression coefficient 1.662, OR 5.268) and CSF glucose content
Third-generation cephalosporin 4 <1.5 mmol/L (regression coefficient 1.500, OR 4.483).
Third-generation cephalosporin + vancomycin 32
Vancomycin + carbapenem 17
Carbapenem + third-generation cephalosporin 12 Relationship between the last CSF before discharge and subject
Penicillin + third-generation cephalosporin 9 prognosis
Chloramphenicol + third-generation cephalosporin 7
Vancomycin + ceftriaxone + rifampicin 1
The CSF of pediatric subjects with purulent meningitis was
Vancomycin + meropenem + rifampicin 3
Vancomycin + meropenem + ceftizoxime 1
described using three parameters: WBC count (number of
Subdural effusion drainage 13 cells × 106 /L), glucose concentration (mmol/L), and protein con-
Ommaya reservoir implantation into encephalocoele 9 centration (g/L). The data were not normally distributed according
Epidural hematoma drainage 1 to the chi-square test, therefore, the subjects were stratified into
Partial small bowel resection due to “meconium peritonitis” 1
categories, and the chi-square test was used to compare between
Resection of pilonidal sinus in the occipital region and Ommaya 1
reservoir implantation into encephalocoele groups. The one-sided 95% CIs were CSF-WBC count <28 × 106 /L
and CSF glucose >1.75 mmol/L respectively. The difference in CSF
protein concentration was statistically significant (P = 0.009). The
Univariate analysis for risk factors of poor prognosis in subjects cut-off point for the two groups was determined to be <0.68 g/L
with severe purulent meningitis using the ROC (Table 5).

To determine which clinical factors might be associated with
poor prognosis, we conducted a univariate analysis using a num-
ber of common clinical tests that monitor subjects with purulent
meningitis (Table 4). Of these, poor prognosis was associated with
the following 9 clinical indicators: unequal pupil size in both
eyes, positive Babinski sign, CSF-white blood cell (WBC) count
>500 × 106 cells/L, CSF protein values >1.0 g/L, CSF glucose con-
tent of <1.5 mmol/L, initial procalcitonin (PCT) results >0.1 ng/dL
on admission, hemoglobin <90 g/L during hospitalization, abnormal
head imaging, and abnormal video EEG results.
Multivariate analysis for risk factors of poor prognosis of
refractory purulent meningitis
We next used the nine significant clinical parameters identified
as predicting poor prognosis in the univariate analysis to perform
a multivariate analysis. The nine associated factors were included
as independent variables and the prognosis of children with
Duration of CSF abnormalities in pediatric subjects with severe
purulent meningitis
CSF-WBC counts, protein concentration, and glucose concentra-
tion at discharge or during follow-up normalized, or were nearly
normal, in 28 of 90 cases (28/90; 31.11%).
In the “good prognosis” group, the CSF completely normalized in
16 cases (16/33; 48.48%). Of these 16 cases, the CSF had normalized
by discharge in 13 cases, and normalized later during follow-up
in 3 cases. The minimum period from disease onset to recovery
in cases where the CSF normalized was 2 weeks (2 cases). Eleven
(11) cases took approximately 1–2 months to recover, 1 case recov-
ered after 3 months, 1 case recovered after 7 months, and 1 case
caught unexplained recurrent meningitis. In this subject the CSF
had normalized at discharge, and then the subject could take care
of himself/herself in daily life and learning. The CSF did not nor-
malize in the remaining 17 cases (17/33; 51.52%). Of these, 12 cases
did not have CSF parameters followed after discharge. In 9 of the 17

Table 4
Univariate analyses for risk factors predicting poor prognosis in pediatric subjects with refractory purulent meningitis.

Influence factor Good prognosis group (n = 33) Poor prognosis group (n = 57) Statistical quantity P (<0.05)

Age ≤1 years 24 45 2 = 0.452 0.607

Gender (male) 19 31 2 = 0.086 0.501
Headache 7 4 2 = 3.925 0.091
Vomiting 21 36 2 = 0.002 0.964
Convulsion 17 38 2 = 2.019 0.155
Listlessness, restlessness, irritability 12 24 2 = 0.287 0.592
Disturbance of consciousness 8 22 2 = 1.938 0.164
Bulging anterior fontanelle 9 15 2 = 0.010 0.921
Neck hindered positive 14 17 2 = 1.469 0.225
Positive Brudzinski signs 0 3 2 = 1.797 0.296
Positive Kerning signs 3 4 2 = 0.125 0.704
Slow/loss of pupillary light reflex 8 13 2 = 0.024 0.877
Unequal pupil size of both eyes 0 12 2 = 8.016 0.003
Dystonia 8 19 2 = 0.825 0.304
Babinski sign positive 4 21 2 = 6.367 0.014
Abnormal patellar reflex 0 5 2 = 3.065 0.154
Peripheral WBC values (<4*109 /L/>12*109 /L) 14 16 2 = 1.938 0.164
CSF-WBC (>500*106 /L) 6 25 2 = 6.103 0.013
CSF-protein (>1.0 g/L) 16 44 2 = 7.751 0.005
CSF glucose (<1.5 mmol/L) 6 24 2 = 5.383 0.020
PCT >0.1 ng/dL 9 27 2 = 7.748 0.005
Ultra-sensitive CRP >8 mg/L 19 39 2 = 0.454 0.500
Hb <90 g/L 11 33 2 = 5.046 0.025
Abnormal video EEG 9 25 2 = 5.046 0.025
Abnormal head imaging 21 49 2 = 5.967 0.015
Temperature 39.18 ± 0.79 39.31 ± 0.75 t = −0.750 0.455

Please cite this article in press as: Peng H-L, et al. Risk factors for poor prognosis in children with refractory purulent meningitis and
the discharge criteria. J Infect Public Health (2017), http://dx.doi.org/10.1016/j.jiph.2017.07.007
G Model
JIPH-770; No. of Pages 5 ARTICLE IN PRESS
4 H.-L. Peng et al. / Journal of Infection and Public Health xxx (2017) xxx–xxx

Table 5 content <1.5 mmol/L, initial PCT results >0.1 ng/dL on admission,
CSF parameters immediately before discharge in pediatric subjects with refractory
hemoglobin <90 g/L during hospitalization, abnormal head imag-
purulent meningitis who have a good or poor prognosis.
ing, and abnormal video-EEG findings. In multivariate analysis
Item Good prognosis group Poor prognosis group P only unequal pupil sizes in both eyes and CSF glucose content
(n = 33) (n = 57)
<1.5 mmol/L remained significant.
CSF-WBC (*106 /L) 0372 Clinically, unequal pupil size in both eyes suggests possible ocu-
≤15 20 29 lomotor palsy or brain herniation. Liu et al. [17] performed a similar
>15 13 28
retrospective study of 148 pediatric subjects with purulent menin-
Range 1–290 0–2740
The median number 10*106 /L 18*106 /L gitis and also found that in a multivariate analysis the strongest
Mean ± SD 30.30 ± 54.25*106 /L 182.07 ± 518.11*106 /L risk factor for poor prognosis was unequal pupil size in both eyes.
CSF glucose (mmol/L) 0.428
Casado-Flores et al. [18–20] reported that abnormal pupils were
≥2.2 23 35 associated with death in purulent meningitis caused by Streptococ-
<1.5 10 22 cus pneumoniae.
Range 1.16–3.75 0.05–5.69 The degree of reduction in CSF glucose may indirectly reflect the
The median number 2.25 mmol/L 2.27 mmol/L
fecundity of pathogenic bacteria and the extent of brain damage,
Mean ± SD 2.47 ± 0.64 mmol/L 1.96 ± 1.30 mmol/L
and thereby suggest a poor prognosis [21]. The studies [15,17,22]
CSF protein (g/L) 0.009 show that a reduced CSF glucose concentration (<1.5 mmol/L [17]
≤0.45 21 20
>0.45 12 37
or <10 mg/dL [22] or <40 mg/dL [15]) is an important risk factor for
Range 0.14–1.38 0.1–20.80 poor prognosis in pediatric subjects with purulent meningitis. In
The median number 0.48 g/L 0.8 g/L this study, we analyzed the initial CSF glucose concentrations after
Mean ± SD 0.57 ± 0.30 g/L 1.74 ± 3.14 g/L admission. Using 1.5 mmol/L as the cutoff point, 6 cases in the “good
prognosis” group and 24 cases in the “poor prognosis” group fell
below 1.5 mmol/L; a significant difference in a multivariate logistic
cases (9/17; 52.94%) one CSF value remained abnormal; in 4 cases regression analysis.
the CSF-WBC values remained abnormal, in 1 case the CSF glucose The antibiotics course and discharge criteria for severe purulent
values remained abnormal, in 4 cases the CSF protein concentra- meningitis recommended in published treatment guidelines are
tion was abnormal. In 4 of the 17 cases (4/17; 23.53%), two CSF not consistent in China and abroad. The Clinical Practice Guideline
values remained abnormal; in 3 cases the CSF protein and glucose of the China Medical Association (CMA) [23] recommends discon-
concentrations were abnormal and in 1 case the CSF protein con- tinuing antibiotics 5–7 days after the CSF returns to normal after
centration and CSF-WBC count were abnormal. Finally, in 4 of the clinical symptoms resolve. In contrast, the antibiotic course rec-
17 cases (4/17; 23.53%) all three of the CSF parameters remained ommended abroad for purulent meningitis are based on positive
abnormal. pathogen cultivation [24–27]. In this study, the average antibi-
In the “poor prognosis” group, the CSF was completely back to otic course was 37 days, which was longer than recommended
normal in 12 of 57 cases (12/57; 21.05%). The CSF was completely by the international guidelines [24–27]. However, 72.2% of the
normalized in 4 cases at discharge, and gradually normalized dur- children did not complete the standard of care antibiotic course
ing follow-up in 8 cases. The minimum period from the onset to recommended in the guidelines from the CMA [23]. An insuffi-
recovery was a month (n = 4 cases). Five (5) cases took 1–2 months cient antibiotic course may affect the therapeutic effect, making
to recover and 3 cases recovered in 3–5 months. The CSF param- it easier to develop recurrent meningitis or serious neurological
eters did not normalize in 45 (45/57; 78.95%) cases. Of these, 38 sequelae, including death. However, a prolonged antibiotic course
cases did not have CSF parameters followed after discharge. In 28 of may cause the emergence of a superbug, increase the risk of nosoco-
the cases (28/45; 62.22%) a single CSF parameter remained abnor- mial infection [28], and affect growth and development of children
mal; in 6 cases the CSF-WBC values were abnormal, in 6 cases the [29,30].
CSF glucose values were abnormal, and in 16 cases the CSF protein There is no unified conclusion on when to withdraw drug for
remained abnormal. In 8 cases (8/45; 17.78%) two CSF parameters refractory purulent meningitis. One meta-analysis [31] suggests
remained abnormal; in 4 cases the CSF-WBC and CSF protein con- there is no significant difference in the final efficacy, sequelae of
centration was abnormal, in 1 case the CSF-WBC and CSF glucose long-term neurological damage or hearing loss, secondary nosoco-
concentration was abnormal, and in 3 cases the CSF protein and glu- mial infections, and other adverse events between a short course
cose concentrations remained abnormal. In 9 cases (9/45; 20.00%) (≤7 days) and prolonged course (standard treatment) of antibi-
all three of the CSF parameters remained abnormal. In the pedi- otics. The appropriate treatment course for children with purulent
atric subjects who died during follow-up, the CSF parameters had meningitis cannot be established due to the limitations of the avail-
not normalized at the time of discharge or during follow-up. able data. Molyneux et al. [32] recommended 5 days of antibiotic
The proportion of pediatric subjects in whom the CSF parame- for children with purulent meningitis caused by S. pneumoniae,
ters had completely normalized at discharge was not significantly Haemophilus influenzae, or Neisseria meningitidis to avoid producing
different between the “good prognosis” and “poor prognosis” resistant bacteria due to the excessive use of antibiotics.
groups (P = 0.075). After the start of antibiotic treatment of purulent meningitis,
CSF recovery often follows a predictable order: (1) disappearance
Discussion of bacteria, (2) glucose normalizes, (3) cell number normalizes,
and (4) protein content normalizes. Changes in the CSF parameters
Predictors of death and sequelae have been previously described in children with purulent meningitis, including low CSF glucose,
[14–16]. However, there are relatively few studies assessing risk high protein, and high cell count, are the result of an inflammatory
factors for poor prognosis in pediatric subjects with refractory cascade and metabolic disorders. They do not necessarily reflect
purulent meningitis. Here we show that in pediatric subjects bacterial persistence and do not necessarily predict intracranial
with refractory purulent meningitis, univariate analysis indi- infection recurrence or relapse. Even if the pathogens have been
cated poor prognosis may be associated with the following killed by antibiotics, the CSF still needs time to return to nor-
factors: unequal pupil size of both eyes, positive Pakistan’s sign, mal. In this study, the CSF took up to seven months to return
CSF-WBC >500 × 106 /L, CSF protein values >1.0 g/L, CSF glucose to normal in the “good prognosis” group. There is no statistically

Please cite this article in press as: Peng H-L, et al. Risk factors for poor prognosis in children with refractory purulent meningitis and
the discharge criteria. J Infect Public Health (2017), http://dx.doi.org/10.1016/j.jiph.2017.07.007
G Model
JIPH-770; No. of Pages 5 ARTICLE IN PRESS
H.-L. Peng et al. / Journal of Infection and Public Health xxx (2017) xxx–xxx
significant difference in the proportion of subjects in whom the
CSF completely normalized at discharge between the “good prog-
nosis” and “poor prognosis” groups, suggesting that whether CSF is
completely back to normal is unlikely to have a significant effect on
the outcome. Thus, there is no scientific basis to evaluate whether
antibiotics should be withdrawn using CSF indicators. However,
there is a significant difference in last CSF protein concentration
before discharge between subjects in the “good prognosis” and
“poor prognosis” groups. The cut-off point for the two groups was
determined to be <0.68 g/L using a ROC. Therefore, we recommend
that children with refractory purulent meningitis may stop using
antibiotics and consider discharge under the following conditions:
(1) completion of a regular anti-infection course and more than 1
week of defervesce; (2) disappearance of acute phase symptoms;
(3) CSF-WBC ≤28 × 106 /L, CSF glucose concentration >1.75 mmol/L,
and CSF protein concentration <0.68 g/L. The patient may be dis-
charged for follow-up if he/she has no relapse during 3–5 days of
observation after drug withdrawal.
Conclusions
We identified two independent risk factors for poor prognosis
in refractory pediatric purulent meningitis: unequal pupil size in
both eyes and CSF glucose content <1.5 mmol/L. Whether CSF is
completely back to normal is unlikely to have a significant effect on
the outcome. Thus, there is no scientific basis to evaluate whether
antibiotics should be withdrawn using CSF indicators.
Conflict of interest
The authors declare that they have no conflict of interest.
Acknowledgments
We sincerely thank all of the children and their parents for their
commitment and their participation in this study.
References
[1] Uiterwijk A, Koehler PJ. A history of acute bacterial meningitis. J Hist Neurosci
2012;21(3):293–313.
[2] Kim KS. Acute bacterial meningitis in infants and children. Lancet Infect Dis
2010;10(1):32–42.
[3] Castelblanco RL, Lee M, Hasbun R. Epidemiology of bacterial meningitis in the
USA from 1997 to 2010: a population-based observational study. Lancet Infect
Dis 2014;14(9):813–9.
[4] Li Y, Yin Z, Shao Z, Li M, Liang X, Sandhu HS, et al. Population-based surveillance
for bacterial meningitis in China, September 2006–December 2009. Emerg
Infect Dis 2014;20(1):61–9.
[5] Curtis S, Stobart K, Vandermeer B, Simel DL, Klassen T. Clinical features sug-
gestive of meningitis in children: a systematic review of prospective data.
Pediatrics 2010;126(5):952–60.
[6] Baraff LJ, Lee SI, Schriger DL. Outcomes of bacterial meningitis in children: a
meta-analysis. Pediatr Infect Dis J 1993;12(5):389–94.
[7] Berg S, Trollfors B, Hugosson S, Fernell E, Svensson E. Long-term follow-up of
children with bacterial meningitis with emphasis on behavioural characteris-
tics. Eur J Pediatr 2002;161(6):330–6.
[8] Chandran A, Herbert H, Misurski D, Santosham M. Long-term sequelae of child-
hood bacterial meningitis: an underappreciated problem. Pediatr Infect Dis J
2011;30(1):3–6.
Please cite this article in press as: Peng H-L, et al. Risk factors for poor prognosis in children with refractory purulent meningitis and
the discharge criteria. J Infect Public Health (2017), http://dx.doi.org/10.1016/j.jiph.2017.07.007
5
[9] Agrawal S, Nadel S. Acute bacterial meningitis in infants and children: epidemi-
ology and management. Paediatr Drugs 2011;13(6):385–400.
[10] Richard GC, Lepe M. Meningitis in children: diagnosis and treatment for the
emergency clinician. Elsevier Inc 2013;14(2):146–56.
[11] Koomen I, Grobbee DE, Roord JJ, Jennekens-Schinkel A, van der Lei HD, Kraak
MA, et al. Prediction of academic and behavioural limitations in school-age
survivors of bacterial meningitis. Acta Paediatr 2004;93(10):1378–85.
[12] Pan YL, Lin ZD. Relationship between cerebrospinal fluid protein content and
severity and prognosis of children with purulent meningitis. J Wenzhou Med
Coll 2011;41(5):483–5.
[13] Kastenbauer S, Pfister HW. Pneumococcal meningitis in adults: spectrum of
complications and prognostic factors in a series of 87cases. Brain 2003;126(Pt.
5):1015–25.
[14] De Jonge RC, van Furth AM, Wassenaar M, Gemke RJ, Terwee CB. Predicting
sequelae and death after bacterial meningitis in childhood: a systematic review
of prognosticstudies. BMC Infect Dis 2010;10:232–46.
[15] Vasilopoulou VA, Karanika M, Theodoridou K, Katsioulis AT, Theodoridou MN,
Hadjichristodoulou CS. Prognostic factors related to sequelae in childhood
bacterial meningitis: data from a Greek meningitisregistry. BMC Infect Dis
2011;11:214.
[16] Namani S, Milenković Z, Koci B. A prospective study of risk factors
for neurological complications in childhood bacterial meningitis. J Pediatr
2013;89(3):256–62.
[17] Liu G, Zhang EQ, Chen HY, Li SY, Chi W, Jiang ZF. Risk factors for an adverse
outcome of childhood bacterial meningitis. J Clin Pediatr 2011;29(2):148–52.
[18] Casado-Flores J, Aristegui J, de Liria CR, Martinón JM, Fernández C. Clinical data
and factors associated with poor outcome in pneumococcal meningitis. Eur J
Pediatr 2006;165(5):285–9.
[19] Kutz JW, Simon LM, Chennupati SK, Giannoni CM, Manolidis S. Clinical pre-
dictors for hearing loss in children with bacterial meningitis. Arch Otolaryngol
Head Neck Surg 2006;132(9):941–5.
[20] Singhi P, Bansal A, Geeta P, Singhi S. Predictors of long term neurological out-
come in bacterial meningitis. Indian J Pediatr 2007;74(4):369–74.
[21] McMillan JA, DeAngellis CD, Feigin RD, Warshaw JB. Oski’s pediatrics:principles
and practice. 3rd ed Philadelphia: Lippincott Williams and Wilkins; 2006.
[22] Lovera D, Arbo A. Risk factors for mortality in Paraguayan children with pneu-
mococcal bacterial meningitis. Trop Med Int Health 2005;10:1235–41.
[23] Clinical Practice Guidelines of China Medical Association. Beijing: People’s
Medical Publishing House; 2005. p. 177–8 [Pediatrics Volume].
[24] Chaudhuri A, Martinez-Martin P, Kennedy PG, Andrew Seaton R, Portegies P,
Bojar M, et al. EFNS guideline on the management of community-acquired bac-
terial meningitis: report of an EFNS TaskForce on acute bacterial meningitis in
older children and adults. Eur J Neurol 2008;15(7):649–59.
[25] Le Saux Nicole. Guidelines for the management of suspected and confirmed
bacterial meningitis in Canadian children older than one month of age. Paediatr
Child Health 2014;19(3):141–52.
[26] Pickering LK, editor. Red Book: 2003 report of the committee on infectious
diseases. 26th ed. American Academy of Pediatrics; 2003. p. 293–493.
[27] Tunkel AR, Hartman BJ, Kaplan SL, Kaufman BA, Roos KL, Scheld WM, et al.
Practice guidelines for the management of bacterial meningitis. Clin Infect Dis
2004;39(9):1267–84.
[28] Appelgren P, Hellström I, Weitzberg E, Söderlund V, Bindslev L, Ransjö U. Risk
factors for nosocomial intensive care infection: a long-term prospective anal-
ysis. Acta Anaesthesiol Scand 2001;45(6):710–9.
[29] Shilling V, Edwards V, Rogers M, Morris C. The experience of disabled children
as inpatients: a structured review and synthesis of qualitative studies report-
ing the views of children, parents and professionals. Child Care Health Dev
2012;38(6):778–88.
[30] Ekra EM, Gjengedal E. Being hospitalized with a newly diagnosed chronic
illness—a phenomenological study of children’s lifeworld in the hospital. Int
J Qual Stud Health Well-Being 2012;7:10–8.
[31] Karageorgopoulos DE, Valkimadi PE, Kapaskelis A, Rafailidis PI, Falagas ME.
Short versus long duration of antibiotic therapy for bacterial meningitis:
a meta-analysis of randomised controlled trials in children. Arch Dis Child
2009;94(8):607–14.
[32] Molyneux E, Nizami SQ, Saha S, Huu KT, Azam M, Bhutta ZA, et al. 5 versus
10 days of treatment with ceftriaxone for bacterial meningitis in children: a
double-blind randomised equivalence study. Lancet 2011;377(9780):1837–45.