“Suggestive CLINICAL SIGNS OF MENINGITIS IN CHILDREN: A SYSTEMATIC REVIEW OF
Sarah Curtis , MD, FRCPC
Kent Stobart , MD, MSc, FRCPC
Ben Vandermeer , BSc, MSc
David L. Simel , MD, MHS
Terry Klassen , MD, FRCPC, MSc
BACKGROUND: The clinical diagnosis of childhood meningitis is essential, therefore, familiarity
with evidence underscoring clinical features of meningitis is very important.
OBJECTIVE: To find evidence supporting the accuracy of the clinical features of pediatric bacterial
METHODS: A review of Medline, Embase, the Cumulative Index to Nursing and Allied Health
Literature, Web of Science, and PubMed was conducted for all articles of relevance. Articles
containing prospective data on the clinical characteristics of laboratory-confirmed bacterial
meningitis and comparison groups of those without children. Two authors independently assessed
trial quality and extracted data to calculate the precision data of clinical features.
RESULTS: Of the 14 145 references initially identified , 10 met the inclusion criteria. In the story, a
report from the bulging fontanelle (odds ratio [LR ] : 8.00 [ confidence interval (95% CI) : 2.4 to 26 ]
) , neck stiffness (7.70 [3, 2-19 ] ), seizures ( outside the age range of febrile seizures ) (4.40 [ 3.0 to
6.4 ] ) or reduced food (2.00 [ 1.2-3.4 ] ) expressed concern about the presence of meningitis. On
examination, jaundice (OR 5.90 [95 % CI 1.8 to 19] ), being toxic or moribund (5.80 [ 3.0 to 11 ] ) ,
meningeal signs (4.50 [2 ,4- 8, 3 ] ) , neck stiffness (4.00 [ 2.6 to 6.3 ] ) , bulging fontanelle (3.50 [
2.0-6.0 ] ) , Kernig sign (3.50 [ 2.1 to 5.7 ] ) , tone (3.20 [ 2.2 to 4.5 ] ) , fever > 40 ° C (2.90 [ 1.6-5.5 ]
) , and Brudzinski's sign (2.50 [ 1.8-3.6 ] ) independently increased the risk of meningitis. The
absence of meningeal signs (LR: 0.41 [95% CI : 0.30 to 0.57 ] ) and abnormal crying (0.30 [ 0.16 to
0.57 ] ) independently low probability meningitis . The absence of fever does not rule out
meningitis (OR, 0.70 [ 95% CI 0.53-0.92 ] ) .
CONCLUSIONS: Clinical evidence of several useful functions that influence the likelihood of
pediatric meningitis exists. No isolated clinical feature is diagnostic , and the combination of more
accurate diagnosis is unclear.
Meningitis can be difficult to diagnose clinically , particularly in young children who do not seem to
reliably show the classic features of the disease. Cerebrospinal fluid (CSF ) analysis through a
lumbar puncture (LP ) is the most important laboratory test for diagnosis. However, LP is an
invasive and painful procedure , and may be difficult and anxiety-provoking for caregivers. It has
been commonly associated with adverse events such as headache and back pain and rarely
associated with infection , cerebral herniation and spinal epidural and subdural hemorrhage . In
addition , CSF analysis is not readily available in many regions of the world. Therefore, it may not
be desirable or feasible to carry out an LP in every child presenting with nonspecific symptoms
that may be attributable to bacterial meningitis, but are much more commonly associated with
less serious conditions.
The delay in diagnosis or failure of meningitis is reflected in reviews of medical malpractice in the
pediatric setting . Undiagnosed Meningitis is the most common diagnosis in children involved in
malpractice claims emergency and has been associated with higher compensation payments and
median payments for pediatricians defense . Negligence cases involving children under 2 years
where the dead child were most often associated with the diagnosis of meningitis. Because
incidence rates decrease with vaccination, the opportunity for recognition and familiarity with the
clinical features of this disease for practicing physicians and students is becoming increasingly rare
. However, this devastating disease has a permanent potential for resurgence with occasional
outbreaks of known or new organisms.
Ideally, the primary clinical assessment should provide an estimate of the probability of the
disease and help determine if diagnostic tests are needed. Identification and use of features that
increase the pretest probability of disease as opposed to those who should not improve the
efficiency and accuracy of clinical evaluation. To our knowledge, a systematic synthesis of
prospective data on the clinical features of meningitis has not yet been made despite the
importance of this disease in training and clinical practice.
Literature Search and Selection: Using a strategy of structured search, a review of Medline,
Embase, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Web of Science,
PubMed, and Cochrane databases was conducted in June 2009, without time constraints, for all
items of interest. a meningitis, diagnostic accuracy, and a chain of pediatrics search terms were
used. The included studies were relevant to describe the historical and physical characteristics of
children with bacterial meningitis confirmed LP and prospectively collected data for calculating the
likely precision of the estimates. Similar data from an LP-negative comparison group had to also be
Quality Assessment : Two authors assessed quality using the quality assessment of diagnostic
accuracy studies ( QUADAS ) checklist and guidelines for assigning levels of quality of evidence .
The QUADAS checklist was developed for quality assessment in systematic reviews of studies of
diagnostic test accuracy . This is a list of 14 items with a "yes ", "no " or " unclear options " and
analyzes population inclusion criteria, and descriptions , time , independence and blinding indices
and benchmarks . The studies were also assessed for the execution of tests , the constant use of a
standard single reference well (LP ), the availability of the results for all patients, and details of the
CSF analysis .
Data Extraction : For both signs and symptoms, if the same word is used to describe a clinical
finding in multiple studies , it was assumed that the test was similar enough to combine
numerically. The decision to combine terms are reached by consensus after examining the terms
that reasonably can be combined without losing its essential meaning.
Data Analysis : The reasons of sensitivity, specificity , and likelihood (LRS ) with confidence
intervals (95% CIs) were calculated for symptoms and signs. When the data was considered
clinically and methodologically similar enough to warrant meta-analysis , Review Manager (
RevMan ) was used to calculate summary measures using the generic inverse variance function.
Heterogeneity was assessed using the statistics stay , which measures the amount of variance
attributable to between-study variation compared to the variation within the study .
sample flow and process selection study . One of the selected 14 145 titles and abstracts, which
resulted in 760 potentially relevant articles ultimately 10 articles met the inclusion criteria (Table 1
) authors. All studies had a quality score of 1 or 2 tests (level 1 : n = 4, level 2: n = 6) and scored ≥
10 on the QUADAS checklist .
FIGURE 1: Study flow diagram .
TABLE 1: Studies that met the inclusion criteria for the accuracy of clinical signs suggestive of
bacterial meningitis in children
CSF analysis was the gold standard for defining the presence of meningitis. The definition of CSF
meningitis vary in detail , but includes a combination of culture positivity CSF or CSF pleocytosis
along culture positivity , either blood or CSF positivity latex agglutination (Table 2 ) . Results of
normal CSF tests and negative results of microbiological studies excluded bacterial meningitis.
TABLE 2: LP (gold standard) Definitions used in each study
Eighteen descriptors of symptoms and signs 48 descriptors was found and extracted for metaanalysis . Of these descriptors , only 5 symptoms and signs 21 resulted in significant data (Table 3).
No significant results for the positive and negative RP shown in Table 4 .
TABLA 3: La precisión de las características clínicas
TABLA 4: Características no admitidas de meningitis pediátrica: Características clínicas de
los estudios prospectivos con resultados estadísticamente insignificantes
Características fueron consideradas como signos designados por el médico o los síntomas
si se informa sobre la historia de los cuidadores. Ningún estudio evaluó la precisión de los
hallazgos clínicos, por lo que el enfoque de la revisión fue el diagnóstico de la
enfermedad. Sólo 2 artículos que describen las combinaciones de resultados.
La prevalencia de la meningitis
El estudio (punto), la prevalencia de la meningitis varían ampliamente de región a región
( Tabla 1 ). La alta prevalencia de la meningitis refleja la naturaleza de la seleccionada el
tipo de paciente estudiado o brotes estacionales de patógenos específicos en las distintas
regiones del mundo. Los criterios de inclusión representa 2 categorías de niños: (1) los
niños con convulsiones y fiebre (2) niños con sospecha clínica de enfermedad bacteriana
invasiva o meningitis. De este modo, la LRS de los siguientes síntomas y los signos se debe
aplicar sólo a estas poblaciones infantiles.
Precisión de las funciones de la historia clínica sugestiva de meningitis
Cuando un médico informó de que su hijo o hija tenía una fontanela abultada, o rigidez en
el cuello, la probabilidad de meningitis aumentó casi ocho veces ( Tabla 3 ). Si un niño
había sufrido un ataque, pero la edad del niño estaba fuera de la franja de edad típica
para las convulsiones febriles, la probabilidad de meningitis se multiplicó por cuatro. La
falta de irritabilidad redujo las probabilidades de la enfermedad a la mitad, pero la
presencia de la irritabilidad no muy significar la presencia de meningitis. Un niño con una
historia reducción de los alimentos tenían una probabilidad algo mayor de la meningitis.
Precisión de las funciones de la exploración física de meningitis.
The presence of complex attacks doubled the risk of meningitis (Table 3). When the type of attack
was classified as "non-specific" when multiple seizures described, most likely was weaker.
Additional descriptors of seizures have been reported in primary studies, but the data were not
The meningeal signs
The definition of " meningeal signs ' varied (eg, irritation, tightness or stiffness or meningeal or
Brudzinski and Kernig sign ) and the presence of any one of them had a summary LR of 4.50 . The
absence of signs of meningitis was more consistent and decreased likelihood of meningitis . When
meningeal signs they were defined only as " stiff neck " were the most heterogeneous results, but
the TP were comparable with the more general term . As Walsh -Kelly et al evaluated the Kernig
and Brudzinski signs in isolation. The presence of any signs increased the likelihood of meningitis,
while the absence of any sign low probability. The presence of a bulging fontanelle increased the
risk of meningitis in an infant of 3.5 times , but when it is absent , the risk of meningitis declined
only slightly .
The mental state or Appearance
The descriptors of a "change in mental status ," " restless or irritable or agitated ," " lethargy or
sleep ," or being " unconscious or comatose RP had comparatively weak short ranging from 1.40 to
1.90 . a " toxic or moribund " appearance had a high LR of 5.80 , the absence of which half the risk
of meningitis. presence of an "abnormal cry " increased the likelihood of meningitis, but his
absence had a greater impact on the likelihood of meningitis (OR 0.30) .
Various other signs
The presence of high fever ( ≥ 40 ° C ) was helpful with a summary LR of 2.90 , but the LR for
temperatures <40 ° C (or otherwise specified ) had a CI that included 1.00. It should be noted that
the absence of fever does not rule out meningitis.
Several other signs were evaluated in each of only one study, and results require validation LR .
Among 341 patients with a 19% prevalence of meningitis, only patients with petechiae (n = 4 ) had
meningitis. Similarly, the presence of jaundice was also notable , as a sign of meningitis ( positive
LR of 5.90) , but was less useful to rule out the disease.
" Tone " had an LR of 3.20 clinical utility . The absence of high pitch reduces the likelihood of half
meningitis . The characteristic of having "open eyes " had a RP of 2.40 , the absence of which only
decreased the likelihood of the disease by a third, "can not or will not eat " seems to be clinically
useful , with a RP 2 10 , while the normal supply reduces the likelihood of meningitis something.
The information on the efficient use of clinical findings is important for physicians. Useful
functions for estimating the probability of meningitis are the features that show the strongest RL
for the presence or absence of disease. The LR of a clinical feature is the probability of finding in
patients with disease divided by the probability that the same feature in patients without the
disease ( LR range from 0 to infinity). Features LR equal to 1.00 have no diagnostic value , as it has
the same probability of finding the function in patients with the disease and in those without the
disease. Features LR > 1.00 support the diagnosis of interest in the magnitude of the increase in
numerical value. The characteristics with TP between 0 and 1.00 , the smallest of the LR, the less
likely the disease.
The useful features found in this review a report of caregivers of neck stiffness , bulging fontanelle
, seizures are listed in Table 3 . Throughout history, in descending order of magnitude, ( outside
the range of febrile seizures ) or reduce food raise concerns about the presence of meningitis. On
physical examination , in order of decreasing magnitude , the presence of jaundice, signs that are
toxic or dying , or have meningeal , neck stiffness , bulging fontanel , Kernig sign , tone , fever > 40
° C , or Brudzinski sign all increase the likelihood of meningitis in varying degrees in the patient.
Several other clinical features with RP between 1.30 and 2.40 , are less strong , but warrants
further study. Note that petechiae sign is strong with a TP of 37.00 , but was surprisingly only
examined in one small prospective study , only 4 patients showed the characteristic. Therefore,
the relevance of this well-known feature is currently uncertain, and systematic prospective
evaluations that among a large number of patients to establish with clarity.
As an example of applicability , assuming statistical independence, a pretest probability of disease
of 10% , and using the LR nomogram , 20 a combination of the presence of meningeal signs (LR:
4.50) , a bulging fontanelle ( CP 3 , 50) , and high fever (OR 2.90) (hence , a combination of LR
45.60 ) increase the likelihood of a child from meningitis to 84%. Although the presence or
absence of these findings , in combination or separately , just confirm or refute the diagnosis of
meningitis , which sufficiently increases the probability that an LP must be performed .
Each routine doctor incorporates a sense of the likelihood of disease through careful consideration
of the clinical evaluation , experience, and estimates of disease prevalence in the population. All
studies included patients with suspected meningitis or serious illness. The point prevalence of
meningitis ranged from 4.2 % to 19 % across these studies , the prevalence of each reflects the
clinical impression of possible meningitis (through initial inclusion in each study) . Prevalence
summary of these studies is 10% . Summary This prevalence could be seen as the posttest
probability of clinical examination , since all the children were judged by disease to undergo a
definitive test for meningitis . Assuming a disease prevalence of 1% for RA clinical impression of
meningitis as their own independent "test" would be 11.00 . Therefore clinical suspicion of the
disease that a provider of health care is derived from the clinical history and examination can , in
itself, be a useful test to monitor orders for diagnostic tests . However, although necessary for the
complete rapid synthesis of complex clinical information, much is known about the process of
clinical judgment and decision making. Clinical impressions are prone to errors, and efforts to
minimize the error by maximizing the pretest probability through an accurate prediction or clinical
decision rules will provide better care to the patient.
Seems clinically sensible combinations of some of the findings in Table 3 would have a greater
impact on the likelihood of meningitis than individual results. Only two studies examined
combinations of results. It is unfortunate that the original data concerned from statistical models
used in these studies were not available , so TP could not be calculated . However, Weber and
Berkley had built logistic regression models of different combinations of features, in an attempt to
obtain sets of predictor variables with an optimal balance of sensitivity and specificity. The best
combination model in the study of Weber et al, combining a history of seizures, being lethargic or
unconscious , or have a stiff neck , had a sensitivity of 98 % and a specificity of 70 %. This
combination of features is a simplified version Integrated criteria childhood disease reference , a
set of guidelines originally developed by the World Health Organization to identify sick children in
need of referral.
However, Berkley and others later in the model tested and found to be only 85 % sensitive and
59% specific . Other models included a Berkley with a high sensitivity of 97 % but low specificity of
44% and no malarial fever combined with any one of the following : bulging fontanelle , neck
stiffness , cyanosis , convulsions ( outside the age range febrile ) seizures, partial seizures, and
altered consciousness . Another model combined with any alteration of consciousness one of the
following : bulging fontanelle , nuchal rigidity , partial seizures , cyanosis , convulsions ( outside the
range of feverish age , seizures), which was found to be less sensitive (79% ) but more specific ( 80
% ) . With a threatening condition highly morbid life models that maximize diagnostic sensitivity
are essential . However, on evaluation , resulting from the application of models of low specificity ,
population is also of concern , especially for regions where distance restrictions or resources limit
access to care even more . Therefore, the ideal clinical pediatric meningitis model is still uncertain,
and prospective evaluation and validation of predictive models of known and novel in different
populations are essential.
Although many of the symptoms and signs with the available data demonstrate the unreliability
(Table 4 ) , these findings have not otherwise been studied in combination . Also, many other
features that are widely described , otherwise reported in textbooks or review articles, have not
been examined for validity in prospective studies. These clinical features commonly described
require a prospective further examination to confirm the soundness of its continued use in the
context of meningitis.
In considering the results of this systematic review , clinicians should be cautious in relation to
decision making for young children especially should not rely on the absence of archetypal
elements to guarantee the absence of disease. Several researchers noted that the studies included
infants with meningitis the classic features of few or no symptoms of the disease is. It is well
accepted clinically that infants with nonspecific features however , relative , such as fever ,
lethargy, lack of appetite, irritability, among others, should be approached with a high index of
suspicion , no matter how good it looks, because the incidence of serious bacterial infection in this
age group is much higher than in older.
This review was limited by heterogeneity in study settings , patient age , comorbidities , inclusion
criteria , gold standard , and the definitions of the index tests . However, the weight of each of
these characteristics in the clinical heterogeneity is variable and uncertain. All studies were similar
in the children examined in good health initially in outpatients at the emergency departments of
hospitals or clinics for acute care hospitals . All children had a spectrum of diseases that raised the
suspicion of meningitis were not previously treated with antibiotics, and performed all LPs had.
However, the degree of tolerance to the increasing heterogeneity must be balanced against the
potential decrease of accuracy measures of general summary. The results of this meta- analysis
should be applied with careful consideration of its limitations and patient populations that
resemble those of the included studies ( Table 1).
Ideally, meta- análisiss of clinical features in pediatric accurate summary reports of the clinical
utility of clinically relevant age groups reflective characteristics change pediatric physiology is. It is
unfortunate that this meta- analysis can only provide individual data summary for the child ( the
undefined age) due to age categorization accurate results or were not different . This leaves
uncertainty, for example, as to when the review of an older child begins to mirror that of an adult
or how the examination of a neonate differs from that of an older child.
Other limitations are notable gaps in a priori definition of individual clinical findings. When viewed
as separate diagnostic " tests " of each clinical feature, in any study of diagnostic accuracy requires
precise definitions to ensure standardized reproducibility and interpretability . For example , neck
stiffness may have varied from a little stiff or tender of a rigid set of researchers to other
researchers . Tone can mean increased muscle tone or hypertonia, but not specifically defined in
this article . Although fever was no description of the variables, and the finding was not useful
when not quantified by the actual temperature. For future research , attention to clear definitions
and classifications accuracy of clinical findings to standardize the performance of the physical
examination and must be paid to ensure reproducibility .
Several clinical useful features that are more likely to be present in children with meningitis
compared with those without the disease have been identified and are compatible with the
limitations , the data collected prospectively . Many other features described meningitis are
currently supported by the available data and ensures a more definitive examination. No clinical
feature is diagnostic in isolation, and the most accurate combination of clinical features to raise or
lower the suspicion of meningitis is still unclear.
He did not obtain external funding for the design or conduct of the study , collection ,
management, analysis or interpretation of data , or the preparation, review or approval of the
manuscript. Dr Curtis had full access to all data in the study and takes responsibility for the
integrity of the data and the accuracy of the data analysis .
We thank Lisa Tjosvold , BA , MLIS ( Alberta Research Centre for Health Evidence ) for assistance
with the literature search ; Belinda Allan and Lisa Chambers (Division of Pediatric Emergency
Medicine , Department of Pediatrics, University of Alberta ) for help with the retrieval of relevant
articles, and Clay Bordley , MD , MPH (Division of Hospital and Emergency Medicine, Department
of Pediatrics, Faculty of Medicine, Duke University Medical Center, Durham , North Carolina),
Dennis a . Clements , MD , PhD , MPH ( Duke Primary Care for Children, Duke Global Health
Institute 's Center for Latin American and Caribbean Studies , Duke University ) , and Rose Hatala ,
MD ( Department of Medicine, Hospital of San Pablo , University of British Columbia, Vancouver ,
British Columbia, Canada) for their valuable advice on earlier versions of the manuscript. None of
the recognized individuals received compensation for their contributions.