Clinical Features of Meningitis in Children

Abstract & Commentary

By Hal B. Jenson, MD, FAAP, Professor of Pediatrics, Tufts University School of Medicine, and Chief Academic Officer, Baystate Medical Center, Springfield, MA, is Associate Editor for Infectious Disease Alert.

Dr. Jenson reports no financial relationships relevant to this field of study.

Synopsis: A structured meta-analysis of signs and symptoms of meningitis in children quantitated the diagnostic value of individual features. Gaps — and opportunities — exist in our clinical diagnostic approach for pediatric meningitis.

Source: Curtis S, et al. Clinical features suggestive of meningitis in children: A systematic review of prospective data. Pediatrics. 2010;126:952-960.

A structured search and analysis was conducted of all published articles describing historical and physical features of children with culture-confirmed bacterial meningitis. Study inclusion criteria included studies of either: 1) children with seizure and fever; or 2) children with a clinical suspicion of invasive bacterial disease or meningitis. There were 14,145 titles and abstracts that were screened, with 760 potentially relevant articles, and ultimately 10 articles that met all inclusion criteria (appropriate design, clear reporting of data, and a comparison group). The analysis used the Quality Assessment for Diagnostic Accuracy Studies (QUADAS) 14-item checklist, which was developed for quality assessment and systematic reviews of diagnostic test-accuracy studies, and also generally accepted guidelines for assigning quality levels of evidence to studies. The 10 studies had a quality level of evidence of 1 (n = 4) or 2 (n = 6).

There were 18 symptom descriptors (reported on history by caregivers) and 48 sign descriptors (described by the physician) that were extracted for a meta-analysis. Only 5 symptoms and 21 signs had a significant likelihood ratio (LR) that increased the likelihood of meningitis. The presence of complex seizures (focal, multiple, or > 15 minute in duration) doubled the risk of meningitis. The 5 significant symptoms were:

  • Bulging fontanel (LR 8.0, 14% sensitivity, 98% specificity)
  • Neck stiffness or bulging fontanel (LR 7.7, 20% sensitivity, 98% specificity)
  • Seizures (not febrile seizures) (LR 4.4, 32% sensitivity, 93% specificity)
  • Reduced feeds (LR 2.0, 52% sensitivity, 70% specificity)
  • Irritability (LR 1.3, 82% sensitivity, 34% specificity)

Of the 21 signs, petechiae (LR 37) and jaundice (LR 5.9) had the highest LRs, but very low sensitivity (6% each). Some of the common key signs included:

  • Toxic or moribund (LR 5.8, 49% sensitivity, 92% specificity)
  • Neck stiffness or rigidity, meningeal irritation, Brudzinski or Kernig sign (LR 4.5, 64% sensitivity, 89% specificity)
  • Bulging fontanel (LR 3.5, 36% sensitivity, 90% specificity)
  • Increased tone (LR 3.2, 59% sensitivity, 82% specificity)
  • Fever > 40°C (LR 2.9, 19% sensitivity, 93% specificity)
  • Complex seizures (LR 2.0, 27% sensitivity, 82% specificity)
  • Lethargic or drowsy (LR 1.8, 40% sensitivity, 79% specificity)
  • Abnormal cry (LR 1.8, 84% sensitivity, 52% specificity)

An additional 13 symptoms and 28 sign pediatric meningitis had statistically insignificant results. Notably, the absence of fever did not exclude meningitis (LR: 0.70 [95% CI: 0.53–0.92]). Non-significant symptoms included lethargy, vomiting, and duration of fever. Non-significant signs included simple seizures, fever < 40°C, tachypnea, dehydration, pallor, cyanosis, and delayed capillary refill.


The LR of a clinical feature is the probability of that finding in patients with the disease divided by the probability of that feature in patients without the disease. LRs range from zero to infinity. The smaller the LR between 0 and 1.0, the less likely the disease with that feature. Features with an LR of 1.0 have no diagnostic value because they are equally likely to be present in those with the disease as in those without the disease. Features with LRs > 1.0 increase the likelihood of the diagnosis, with the diagnostic value increasing as the magnitude of the LR increases.

The clinical assessment of meningitis in children remains a combination of the art and science of medicine. For example, "irritability" is easy to state in a textbook as a key clinical feature of meningitis, but recognizing this sign accurately is a highly refined skill. No single clinical feature is diagnostic of meningitis in children. A different constellation of symptoms and signs must be used to assess the likelihood of meningitis at different ages. This analysis reveals that several symptoms and signs do have quantifiable clinical utility for the diagnosis of pediatric meningitis. However, the most accurate diagnostic combination that should be used remains suboptimally defined. This analysis also shows that the available data for many of the clinical features that are often used by clinicians and tend to be cited in textbooks — 13 symptoms and 28 signs — do not have demonstrated accuracy for diagnosis. Many of these clinical features have not been clearly defined for accuracy and precision, and have not been adequately studied for validity by prospective studies to confirm the soundness of using these features for the diagnosis of meningitis.