eMedicine Specialties > Neurology > Neurological Infections

Haemophilus Meningitis: Differential Diagnoses & Workup

Author: Robert Rust Jr, MD, Thomas E Worrell Jr Professor of Epileptology and Neurology, Co-Director of FE Dreifuss Child Neurology and Epilepsy Clinics, University of Virginia School; Clinical and Residency Training, Child Neurology, University of Virginia Hospital and Clinics
Coauthor(s): Robert Cavaliere, MD, Assistant Professor of Neurology, Neurosurgery and Medicine, Ohio State University College of Medicine
Contributor Information and Disclosures

Updated: Sep 28, 2006

Differential Diagnoses

Other Problems to Be Considered

Brain abscess
Posterior fossa subdural or subdural empyema
Sepsis
Rocky Mountain spotted fever
Typhus
Cerebral malaria
Fungal meningitis
Carcinomatous meningitis

Workup

Laboratory Studies

• White blood cell count is elevated in the majority of cases at presentation, with a left shift. As with Hib epiglottitis, counts in excess of 20,000/µL may be found.
• Hib may be grown from blood cultures in at least 50-80% of cases if no prior treatment with antibiotics has been undertaken. Accurately diagnosing the agent responsible for meningitis may be more complicated in developing countries by the widespread use of antibiotics before blood or CSF cultures can be obtained. This approach is more understandable if one considers the delay that may be encountered in some nations in rapidly obtaining access to healthcare facilities capable of performing such studies.
• Chemistry, and in particular sodium, should be ascertained immediately and monitored at intervals throughout treatment. The syndrome of inappropriate antidiuretic hormone secretion (SIADH) develops in approximately half of all cases of Hib meningitis. It may cause stupor or seizure and may contribute to the elevation of intracranial pressure. Proper diagnosis requires demonstration of serum sodium less than 135 mEq/L, serum osmolality less than 270 mOsm/L, urine osmolality greater than twice serum osmolality, urine sodium greater than 30 mEq/L, and no evidence of confounding hypovolemia or dehydration. In some instances, low sodium is due to cerebral salt wasting, rather than SIADH. Unlike SIADH, serum salt wasting is associated with decline in patient mass, rather than increase in mass. In other instances, hyponatremia is produced by the excessive intravenous administration of hyposmolar fluids.

Imaging Studies

• Brain imaging studies may be of importance in patients with Hib meningitis. They are appropriately obtained in the acute setting to identify mass lesions that are in the differential diagnosis (eg, focal encephalitis, brain abscess, empyema, parasitism, subdural hemorrhages) not only for diagnostic purposes, but also to evaluate such risks as may be associated with lumbar puncture. Hence, evidence for focal neurologic dysfunction (ie, seizures, focal neurologic deficits) or of papilledema should prompt consideration of scanning. Other indications for scanning during the initial or subsequent phases of hospitalization include persistently depressed or unexplained deterioration in neurologic status and prolonged fever despite treatment.
• Scanning should never be performed before other critical management decisions have been made and acted upon. If lumbar puncture is deferred until after scanning, adequate intravenous access must be established, blood cultures must be drawn, and broad-spectrum antibiotic coverage pertinent to any suspected meningitic agent should be administered.
• If a scan is necessary and seizures have occurred or may continue to occur, full intravenous loading with an anticonvulsant should be considered. The authors regard phenobarbital as the treatment of choice in young children because its sedative properties may make other forms of sedation unnecessary. The wide therapeutic window of this agent permits multiple additional doses to be administered if seizures are resistant to treatment, and phenobarbital is easier to manage than phenytoin because of the nonlinear kinetics of phenytoin, if an anticonvulsant is judged necessary at discharge. Phenobarbital may also have beneficial effects in cases of increased intracranial pressure by reducing irritability and also cerebral metabolic demand.
• Brain imaging studies obtained at presentation are usually justified to identify an alternative diagnosis to meningitis (eg, brain abscess, subdural empyema) that may contraindicate a lumbar puncture. Results of imaging studies do not confirm the diagnosis of meningitis, which can only truly be confirmed by the performance of lumbar puncture. In instances where lumbar puncture is contraindicated, the presumption of meningitis may be made where the image findings or clinical circumstances and other testing do not disclose an alternative diagnosis.
• Either CT scanning or MRI may provide information concerning the usual space-occupying lesions or other complications that may result from Hib meningitis and either modality provides information concerning some alternative diagnoses. Generally, CT scanning is obtained because it is usually more readily available and requires less time. Patients must be monitored by qualified personnel during imaging because, during the scan interval, these patients may have or develop untreated seizures or critical elevation in intracranial pressure.
• The most common imaging findings in cases of Hib meningitis at or shortly after presentation are meningeal, ependymal, or choroidal enhancement due to meningitic inflammation. Inflammatory exudate may be demonstrable in the basilar cisterns, especially the foramen magnum. The accumulation of inflammatory exudate tends to widen the basilar cisterns and the cortical sulci (particularly over the convexities of the forebrain hemispheres) on CT scanning. Findings on CT scanning may be quite normal in the acute stage of Hib meningitis. MRI scanning, if performed, may reveal the abnormalities noted above with even greater sensitivity and definition than CT scanning.
• Abnormalities indicative of meningitic inflammation and exudate support the diagnosis of meningitis but are not very specific with regard to organism and usually do not modify therapy or prognosis. Thus, for example, the extent of meningeal enhancement is not indicative of prognosis. Rarely, adults may present with Hib meningoventriculitis manifesting ventricular debris, periventricular hyperintense signal, and periventricular ependymal enhancement (Nakayasu, 2005).
• Other important abnormalities that scans may detect tend to develop later in the course of illness and constitute complications of Hib meningitis. In many cases, these abnormalities are better defined by MRI than CT scanning. Special circumstances of the clinical or laboratory course may serve as indications for obtaining scans. Indications for obtaining such scans, whether CT or MRI, and the abnormalities that may be found in explanation of such findings in Hib meningitis include the following:
  • Persistence of fever after several days of appropriate intravenous antibiotic therapy have been provided (found in approximately 10% of all cases of Hib meningitis): Causes include nosocomial infection (eg, of the intravenous lines), subdural effusions (20-30% of cases), and comorbidity of pneumonia or arteritis (uncommon).
  • Return of fever after achievement of an afebrile state with appropriate intravenous antibiotic therapy (see discussion of prolonged fever above)
  • Evidence suggesting increased intracranial pressure (eg, bulging fontanelle, Cushing reflex, obtundation, meningismus, cranial nerve signs suggestive of herniation), hydrocephalus, large subdural effusion, or empyema
  • Focal neurologic deficits
  • Prolonged obtundation or coma
• On CT scanning performed because of these various indications, one or another of these various abnormalities is found in slightly more than 50% of all such scans. However, in most instances, these abnormalities do not require specific interventions and may not prove helpful in estimating prognosis. Discussions of the most common complications found on scans are discussed.
• Transependymal movement of CSF may be detected, especially in instances where noncommunicating hydrocephalus develops. Brain swelling may be found, and diffuse increased T2-weighted signal may be found, representing interstitial cerebral edema. This change may be diffuse. These various changes may indicate the need for management of increased intracranial pressure.
• Subdural effusions may develop in Hib meningitis. They are due to increased permeability of capillaries and veins of the inner dural surface. Only a small percentage of these are of clinical significance. They are usually sterile and seldom exert mass effect, although in a small number of cases they are infected and they may in rare cases produce mass effects. However, in general, they follow a self-limited course and become completely resorbed. Prolonged fever despite adequate antibiotic coverage may result from an effusion that has become secondarily infected. This is suggested by the presence of contrast enhancement. The development of new or progressive deficits, such as hemiparesis, during the course of illness may indicate that a subdural effusion has begun to exert mass effects.
• Hydrocephalus, either communicating or obstructive, may occur in Hib meningitis. Such a process should be suspected with progressive or prolonged altered consciousness despite appropriate antibiotic treatment. Communicating hydrocephalus probably develops because the inflammatory exudate across the vertices impairs the resorptive function of arachnoid granulations. Noncommunicating hydrocephalus usually develops because of exudative blockage of the foramina of Magendie and Luschka.
• Cerebral infarction as a consequence of meningitic vasculitis may be found. MRI is usually superior for the demonstration of these changes, particularly when sequences designed to demonstrate restricted diffusion are employed. These abnormalities tend to be found in subcortical white matter, cerebellum, and brainstem and resemble the changes that may be found in hypoxic-ischemic encephalopathy. Lesions such as these should be suspected when patients with Hib meningitis manifest focal deficits or seizures.
• Abscess formation may be detected in scans obtained because of the development of focal deficits or seizure, although this complication is uncommon in Hib meningitis.

Other Tests

• Electroencephalography: Electroencephalography is sometimes indicated to evaluate for seizure activity. This includes patients with persistent depressed mental status without obvious evidence of seizure activity. Nonconvulsive status epilepticus is common in this population, although altered mental status is more often caused by metabolic disarray.
• Brainstem auditory evoked response
  • Hearing impairment is a common complication in Hib meningitis. If present, it is usually permanent. Hearing may be difficult to assess clinically, and all children should have brainstem auditory evoked response (BAER) testing at some point during hospitalization or in the early period of posthospitalization recovery. However, the timing of BAER testing is important, and this form of testing does not influence acute management.
  • The value of BAER testing in predicting permanent sensorineural hearing loss from bacterial meningitis is limited if test results are abnormal before resolution of conductive loss (due to the presence of otitis media, which frequently precedes Hib meningitis) or other possible forms of acute inflammation of neural tissue. Thus, the test should be repeated some weeks or months later if results are initially abnormal. On the other hand, if the test results are normal during the acute phase of the disease, their predictive value for normal hearing is excellent. Unlike other focal complications of meningitis, sensorineural hearing loss is not a risk factor for epilepsy. However, sensorineural hearing loss is associated with language and learning delays. Thus, if present, children should be referred for further hearing and speech evaluations and therapies.

Procedures

• Lumbar puncture is critical in the evaluation of patients with suspected meningitis and should be performed unless some specific contraindication exists. In young febrile children, lumbar puncture should be performed if meningitis cannot be otherwise excluded (after appropriate consideration of such contraindications as asymmetrical space-occupying lesion). Lumbar puncture should also be strongly considered if another definite source of infection and fever cannot be found and outpatient antibiotic therapy is to be provided. Performing such a puncture avoids the diagnostic problems associated with partially treated meningitis in the event that the infant returns within the next few days with clinical worsening.
• Lumbar puncture results may confirm the diagnosis of meningitis or suggest an alternative diagnosis. In cases of bacterial meningitis, CSF Gram stain and culture may identify the organism causing meningitis, which is advantageous in that treatment and prognostication can be adjusted to the specific organism. Identification of increased pressure by lumbar puncture may also modify the therapy provided.
• Care must be taken not to perform lumbar punctures in patients who are at risk for herniation or are manifesting signs of impending herniation. Although the scientific underpinnings of the allegations of a relationship between lumbar puncture and herniation are in many cases weak, they may not appear to be so in the minds of nonmedical personnel called upon to review such an alleged relationship in retrospect in a courtroom.
• Findings that may indicate onset for herniation or impending herniation include focal brainstem signs, especially if present unilaterally (eg, dilation of pupil, diminution or loss of pupillary reactivity, diminution or loss of abducens function), head tilt, meningismus, deterioration in mental status, visual field defect, focal seizures, vomiting, increased tone in the lower extremities, Cushing reflex (ie, elevated blood pressure with slow heart rate), and hyperventilation or other disturbances of breathing rhythm consistent with brainstem regulatory failure. Papilledema is a very important sign, but it may not develop until several hours of increase in intracranial pressure have passed, and a large segment of the medical community cannot reliably determine the pertinent early funduscopic changes. Venous pulsation presence may be reassuring, but the absence of pulsations is of greatest value only in cases where they were known to be present prior to the current urgent presentation.
• In cases where concern is raised by any of these signs, deferring lumbar puncture until after brain imaging can be obtained is appropriate. However, in all such cases wherein the diagnosis of meningitis is entertained, obtaining a blood culture and initiating appropriate broad-spectrum antibiotic therapy immediately afterwards is crucial so that scanning the brain does not delay initiation of treatment. The authors re-emphasize the fact that the performance of brain imaging studies should never delay the initiation of treatment for increased intracranial pressure or seizures. Note that even in cases where intravenous antibiotics have been administered immediately prior to CT scanning, CSF from a lumbar puncture performed after the completion of the scan seldom has been sterilized by the antibiotics.
• In the absence of focal neurologic findings (such as those noted above), the risk of herniation in cases of Hib meningitis is low and one can safely proceed to lumbar puncture without imaging. In general, evidence for raised intracranial pressure is not considered to represent a categorical contraindication to lumbar puncture, as long as no signs suggesting focal space-occupying lesions are found. If evidence exists for increased intracranial pressure, a small needle (#22 gauge) should be employed by the most skilled available person and only as much CSF as is needed for essential tests should be collected.
• CSF abnormalities are found in approximately 16-20% of children who are evaluated by lumbar puncture for possible meningitis. Of the children with abnormalities, 60-68% are viral, 20-26% are found to be bacterial, and the cause remains unclear or unknown in 5-10%.
• Opening pressure should be recorded. In bacterial meningitis, it is frequently elevated and may have an impact on treatment. In small calm infants, pressures should be less than 160 mm H₂ 0. Older infants and children should have pressures less than 180 mm H₂ 0. Normal pressures of some obese individuals are as high as 250 mm H₂ 0. Pressure should only be recorded in individuals who are in the lateral recumbent position, while they are as relaxed and calm as possible. In fulminant cases of Hib meningitis, pressures as high as 300 mm H₂ 0 to more than 500 mm H₂ 0 may be recorded.
• CSF should be collected in a sterile manner in sufficient quantity and immediately submitted to the laboratory. If extra CSF is available, freeze it and store it for possible future evaluation. The appearance of the CSF should be noted. Normal CSF is clear. However, in bacterial meningitis, the presence of more than 200/mL white cells or more than 400/mL red cells or more than 100 mg/dL protein or more than 10⁵ colony-forming units (CFU) of bacteria may cause the CSF to appear cloudy. This change may be subtle and is best appreciated by flicking the bottom of the firmly held tube and observing for a shimmer of iridescence. In severe cases, the CSF may appear purulent. Protein values greater than 150 mg/dL cause the fluid to appear xanthochromic.
• CSF Gram stain may reveal the Hib pleomorphic gram-negative coccobacilli. If the CSF is cloudy, the stain should be performed on fresh uncentrifuged CSF. If the CSF is clear, it should be performed on the pellet of centrifuged CSF. The probability of visualizing bacteria depends on the concentration of bacteria in CSF. Bacteria are identifiable in 60-90% of all cases of acute bacterial meningitis, particularly in cases where more than 10⁵ CFU/mL are present. The specificity of a positive Gram stain for bacterial meningitis is approximately 95%. Gram staining does not provide a definitive identification of the bacteria and does not, of course, provide information concerning antibiotic sensitivities. Oral or intravenous pretreatment with antibiotics may disable the identification of organisms by Gram stain. Indeed, the presence of organisms on Gram stain 24 hours after intravenous treatment has been initiated may be an important indicator of treatment failure.
• CSF should be examined promptly for white cells because they tend to begin to disintegrate within about 90 minutes of the lumbar puncture. Greater than 10 WBC is usually considered abnormal, while the presence of even one polymorphonuclear (PMN) leukocyte is considered abnormal. WBC differential counts from cytocentrifuged CSF may falsely elevate the PMN leukocytes. The occurrence of a preceding convulsive seizure may elevate the white count, particularly the PMN leukocyte count. When modest CSF pleocytosis is due to seizure and not meningitis, opening pressure is usually normal, CSF is clear, fewer than 80 WBC/µL are found, and CSF glucose is normal.
• The typical finding in Hib meningitis is PMN leukocyte–predominant pleocytosis, as is the case with most other forms of bacterial meningitis. CSF WBC counts in Hib meningitis are greater than 100/µL in more than 90% of cases and greater than 1000 in 65-70% of cases. The mean CSF WBC counts for Hib meningitis approach 1100/µL.
• Note, however, that although most cases of the fully developed CSF pleocytosis of viral meningitis manifest lymphocytic predominance, PMN leukocytes may predominate in as many as 20-75% of lumbar puncture samples obtained in the early phases of viral encephalitis, and they may be found in 5-8% of viral encephalitides even after fully developed pleocytosis has been achieved. On the other hand, approximately 10-30% of bacterial meningitis cases are found to have early lymphocytic predominance, especially in cases where the CSF WBC count is less than 1000/µL.
• In at least half of all patients who receive appropriate antibiotic therapy for bacterial meningitis, the CSF WBC count remains elevated for at least one week after initiation, and in some cases, an elevated count persists for several weeks. However, falling CSF WBC counts on repeat lumbar punctures should be considered a reassuring indication of response to appropriate treatment.
• Relatively low CSF WBC counts in a very ill child with Hib meningitis may indicate a poor prognosis, especially if large numbers of nonengulfed Hib organisms are observed on the CSF Gram stain.
• CSF glucose concentrations lower than 40 mg/dL are found in approximately two thirds of all cases of acute bacterial meningitis. Comparison must always be made to serum glucose at the time of the lumbar puncture. The CSF-to-serum glucose ratios ought to be approximately 2:3 (ie, 0.6).
• CSF glucose within the reference range in the presence of elevated serum glucose may not actually be normal because the CSF value must be interpreted with respect to serum glucose concentration. Serum glucose values are often low or high in cases of acute bacterial meningitis. A CSF-to-serum glucose ratio of less then 0.31 is observed in 70% of patients with bacterial meningitis. Low CSF-to-serum glucose ratios are also found in fungal and carcinomatous meningitides.
• In as many as 80% of patients who receive appropriate intravenous antibiotic treatment for bacterial meningitis, CSF glucose concentration returns to the reference range by the third day of that treatment. However, even with appropriate treatment some patients continue to exhibit low CSF glucose for 7-10 days after the initiation of appropriate intravenous antibiotic treatment.
• As happens with any process that disturbs the BBB function, CSF protein concentrations increase in bacterial meningitis. In Hib meningitis, this value for lumbar CSF is typically greater than 50 mg/dL with a typical range of 100-500 mg/dL. In the event that ventricular CSF is available for analysis, note that abnormal values are those greater than 15 mg/dL. In the event of a traumatic tap, protein values may be grossly estimated by the subtraction of 1 mg/dL of protein for every 1000 RBC/µL.
• In the setting of bacterial meningitis, CSF lactate is frequently elevated. Values in excess of 3.5-3.8 mmol/L are sensitive indicators of acute bacterial meningitis, found in as many as 92% of cases. Specificity of this finding is comparatively low, although elevation of lactate to the concentrations noted above is more strongly indicative of bacterial than viral meningitis. However, elevation of lactate does not exclude the diagnosis of viral meningitis. Whether CSF lactate as a diagnostic test adds information that cannot be obtained from CSF cell counts, glucose, and protein is not clear. Moreover, elevation of CSF lactate may be due to other potential alternative diagnoses such as closed head injury, smothering and other causes of hypoxic-ischemic brain injury, neoplasia, or prolonged seizures from any of a wide variety of causes.
• Elevation of CSF lactate in Hib meningitis may be due to cerebral edema or changes in cell membranes or cellular energy metabolism leading to anaerobic glycolysis. CSF lactate may remain elevated for a fairly long interval after effective antimicrobial therapy has resulted in amelioration of brain edema and restoration of intracranial pressure to the reference range.
• Repeated lactate estimation (lumbar CSF analysis or MRI spectroscopically) may provide a method for estimating possible deleterious effects of fluid restriction in cases of Hib meningitis–induced brain swelling. Inadequate systemic volume may be deleterious in such cases because of the high intracranial pressure and pressure-passive nature of dysregulation of cerebral circulation in meningitis.
• Culture of the CSF yields the most specific information. However, CSF cultures are positive within 48 hours in approximately 75-80% of cases with sensitivity of 95% and specificity of 99%.
• A positive culture is the most valuable single test in confirming the diagnosis of bacterial meningitis. Although in some instances a false-positive CSF culture is obtained, these cultures tend to contain skin commensals such as Staphylococcus alba, and a false-positive CSF culture containing Hib is likely very rare. Moreover, ascribing such a positive culture to contamination is so risky that it prevents most physicians from any subsequent course other than completing the usual course of therapy for meningitis.
• The best method of confirming that Hib in culture is the cause of meningitis is to have found the organism on Gram stain of the initial CSF. This result is the other most sensitive method of confirming the diagnosis of meningitis in any given case where the diagnosis is considered. Although other tests, such as cell counts and CSF chemistries, are critical for the initial management, the results may (unless very abnormal) be less sensitive or specific than positive cultures and Gram stains. Evidence suggests that CSF glucose less than 34 mg/dL, a ratio of CSF to blood glucose of less than 0.23, CSF leukocyte counts greater than 2000/µL, or CSF neutrophil counts of greater than 1180/µL predict bacterial meningitis in cases with a clinically consistent picture with 99% certainty.
• Moreover, in cases where individuals have been treated with antibiotics within the week prior to the lumbar puncture, the less-irrefutable approaches to meningitis diagnosis by CSF cell counts and chemistries must nonetheless be relied upon for decisions concerning initiation of and persistence in therapy. Two careful prospective studies of the effect of pretreatment have been published and indicate that as many as one third of children with Hib meningitis receive such treatment, usually for suspected otitis media.
• Pretreatment with oral antibiotics may significantly reduce not only the yield of CSF culture and Gram stain, but also CSF protein concentration and neutrophil percentage. On the other hand, pretreatment was not shown to significantly decrease the yield of blood culture, total CSF white cell count, CSF glucose concentration, CSF-to-serum glucose ratio, or positivity of CSF counterimmunoelectrophoresis or latex agglutination studies for bacterial antigens. Attenuation of the significance of some of these indicators of bacterial meningitis was thought to be due to the fact that oral antibiotics had attenuated the severity of illness, although they had not prevented the development of meningitis. This concept is supported by the fact that pretreated children tended to present for lumbar puncture several days later than untreated children after an initial premonitory meningitic phase with otitis media or upper respiratory illness such as commonly precedes Hib meningitis.
• Studies of the effects of intravenous antibiotic administration on CSF characteristics of children who do have meningitis have shown that as many as several days of intravenous treatment with appropriate antibiotics does not significantly alter CSF protein, glucose, or white blood cell concentrations, although the yield of Gram stain and culture is lost.
• No results of CSF cell counts or chemistries can be employed to irrefutably rule out a meningitis diagnosis where the clinical indications of possible meningitis are present. This is particularly true because of the possibility of viral meningitis in such cases.
• A positive CSF culture provides the additional benefit of permitting the establishment of antimicrobial sensitivity in subcultures of recovered organisms, permitting the adjustment of treatment to these findings. However, several additional days after recovery of the organism may be necessary for such results to be available.
• Bacterial antigen tests such as counterimmunoelectrophoresis or latex agglutination immunologically detect the soluble antigens on many bacteria, including those of Hib. The tests are very rapid but detect only the most common forms of meningitis. The latex particle agglutination antigen tests for Hib have sensitivity of 97% and specificity of 95%.
• Polymerase chain reaction (PCR) is an emerging technique that may ultimately be useful in identifying the organism in patients when the Gram stain and culture results are negative. However, its application to bacterial meningitis has been limited by a significant number of false-positive results caused by amplification of contaminating DNA and mispriming.
• Some children have been pretreated with oral antibiotics prior to presentation. Treatment may impact the results of CSF analysis. CSF Gram stain and cultures are the most sensitive to pretreatment with antibiotics and may be rendered negative within 24 hours of treatment. However, bacterial antigens and PCR are not affected and therefore remain effective in identifying the organism, although not its antimicrobial sensitivities. CSF protein, glucose, and WBC counts are not significantly influenced by pretreatment.
• Various studies have been published concerning the utility of testing for fibrin degradation products, lactate dehydrogenase, creatine kinase, or other potential CSF constituents in evaluation of children who may have meningitis. As yet, no compelling evidence indicates that such testing is valuable.

Histologic Findings

No specific findings are noted on histologic studies of any tissues other than brain.

Staging

No system of staging exists.

• Search for CME/CE on This Topic »

• Brain and Nervous System Center
• Meningitis in Children Overview
• Meningitis in Children Symptoms
• Spinal Tap Introduction
• Sepsis Overview
• Cellulitis Overview
• Blood and Lymphatic System Center
• Meningitis in Children Causes
• Meningitis in Children Treatment
• Brain Infection Overview
• Epiglottitis Overview