This chapter is concerned mainly with bacterial infections of the central nervous system (CNS), particularly bacterial meningitis, septic thrombophlebitis, brain abscess, epidural abscess, and subdural empyema. The granulomatous infections of the CNS, notably tuberculosis, syphilis and other spirochetal infections, and certain fungal infections are also discussed in some detail. In addition, consideration is given to sarcoidosis, a granulomatous disease of uncertain etiology, and to the CNS infections and infestations caused by certain rickettsias, protozoa, worms, and ticks.

A number of other important infectious diseases of the nervous system are discussed elsewhere in this book. Viral infections, because of their frequency and importance, are allotted a chapter of their own (see Chap. 33). Diseases caused by bacterial exotoxins—diphtheria, tetanus, botulism—are considered with other toxins that affect the nervous system (see Chap. 43). Leprosy, which is essentially a disease of the peripheral nerves, is described in Chap. 46, and trichinosis, mainly a disease of muscle, in Chap. 48.

These infections reach the intracranial structures by one of two pathways, either by hematogenous spread (emboli of bacteria or infected thrombi) or by extension from cranial structures adjacent to the brain (ears, paranasal sinuses, osteomyelitic foci in the skull, penetrating cranial or congenital sinus tracts). In a number of cases, infection is iatrogenic, being introduced in the course of cerebral or spinal surgery, the placement of a ventriculoperitoneal shunt or, rarely, by a lumbar puncture needle. Increasingly, craniospinal infections are nosocomial, i.e., acquired in-hospital; in urban hospitals, nosocomial meningitis is now as frequent as the non–hospital-acquired variety (Durand et al).

Surprisingly little is known about the mechanisms of hematogenous spread and animal experiments involving the injection of virulent bacteria into the bloodstream have yielded somewhat contradictory results. In most instances of bacteremia or septicemia, the nervous system seems not to be infected; yet sometimes a bacteremia caused by pneumonia or endocarditis is the only apparent predecessor to meningitis. With respect to the formation of brain abscess, cerebral tissue has a notable resistance to infection. Direct injection of virulent bacteria into the brain of an animal seldom results in abscess formation. In fact, this condition has been produced consistently only by injecting culture medium along with the bacteria or by causing necrosis of the tissue at the time bacteria are inoculated. In humans, infarction of brain tissue because of arterial occlusion (thrombosis or embolism) or venous occlusion (thrombophlebitis) appears to be a common and perhaps necessary antecedent by way of causing of a necrotic nidus.

The mechanism of meningitis and brain abscess from infection of the middle ear and paranasal sinuses is easier to understand. The cranial epidural and subdural spaces are practically never the sites of blood-borne infections, in contrast to the spinal epidural space, where such infections are either hemtogenous spread but can be from contiguous osteomyelitis. Furthermore, the cranial bones and the dura mater (which essentially constitutes the inner periosteum of the skull) protect the cranial cavity against the ingress of bacteria. This protective mechanism may fail if suppuration occurs in the middle ear, mastoid cells, or frontal, ethmoid, and sphenoid sinuses. Two pathways from these sources have been demonstrated: (1) infected thrombi may form in diploic veins and spread along these vessels into the dural sinuses (into which the diploic veins flow), and from there, in retrograde fashion, along the meningeal veins into the brain, and (2) an osteomyelitic focus may erode the inner table of bone and invade of the dura, subdural space, pia-arachnoid, and even brain. Each of these pathways has been observed by the authors in some fatal cases of epidural abscess, subdural empyema, meningitis, cranial venous sinusitis and meningeal thrombophlebitis, and brain abscess. However, in many cases coming to autopsy, the pathway of infection cannot be determined.

With a hematogenous infection in the course of a bacteremia, usually a single type of virulent bacterium gains entry to the cranial cavity. In the adult the most common pathogenic organisms are pneumococcus (Streptococcus pneumoniae), meningococcus (Neisseria meningitidis), Haemophilus influenzae in unvaccinated children, Listeria monocytogenes, and staphylococcus; in the neonate, Escherichia coli and group B streptococcus; in the infant and unvaccinated child, H. influenzae. By contrast, when septic material embolizes from infected lungs, pulmonary arteriovenous fistulas, or congenital heart lesions, or extends directly from ears or sinuses, more than one type of bacterial flora common to these sources may be transmitted. Such “mixed infections” pose difficult problems in therapy. Occasionally in these latter conditions, the demonstration of the causative organisms may be unsuccessful, even from the pus of an abscess (mainly because of inadequate culturing techniques for anaerobic organisms and the prior use of antibiotics). Infections that follow neurosurgery or the insertion of a cranial appliance are usually staphylococcal; a small number are a result of mixed flora, including anaerobic ones, or one of the enteric organisms. In determining the most likely invading organism, the age of the patient, the clinical setting of the infection (community-acquired, postsurgical, or nosocomial), the immune status of the patient, and evidence of systemic and local cranial disease all must be taken into account.

The Biology of Bacterial Meningitis

The immediate effect of bacteria or other microorganisms in the subarachnoid space is to cause an inflammatory reaction in the pia and arachnoid as well as in the cerebrospinal fluid (CSF). Because the subarachnoid space is continuous around the brain, spinal cord, and optic nerves, an infective agent gaining entry to any one part of the space allows it to spread rapidly to all of it, even its most remote recesses; in other words, meningitis is always cerebrospinal. Infection also reaches the ventricles, either directly from the choroid plexuses or by reflux through the foramina of Magendie and Luschka.

The first reaction to bacteria or their toxins is hyperemia of the meningeal venules and capillaries and an increased permeability of these vessels, followed shortly by exudation of protein and the migration of neutrophils into the pia and subarachnoid space. The subarachnoid exudate increases rapidly, particularly over the base of the brain; it extends into the sheaths of cranial and spinal nerves and, for a very short distance, into the perivascular spaces of the cortex. During the first few days, mature and immature neutrophils, many of them containing phagocytosed bacteria, are the predominant cells. Within a few days, lymphocytes and histiocytes increase gradually in numbers. During this time, there is exudation of fibrinogen, which is converted to fibrin after a few days. In the latter part of the second week, plasma cells appear and subsequently increase in number. At about the same time the cellular exudate becomes organized into two layers—an outer one, just beneath the arachnoid membrane, made up of neutrophils and fibrin, and an inner one, next to the pia, composed largely of lymphocytes, plasma cells, and mononuclear cells or macrophages. Although fibroblasts of the meninges begin to proliferate early, they are not conspicuous until later, when they take part in the organization of the exudate, resulting in fibrosis of the arachnoid and loculation of pockets of exudate.

During the process of resolution, the inflammatory cells disappear in almost the reverse order as they had appeared. Neutrophils begin to disintegrate by the fourth to fifth day, and soon thereafter, with treatment, no new ones appear. Lymphocytes, plasma cells, and macrophages disappear more slowly, and a few lymphocytes and mononuclear cells may remain in small numbers for several months. The completeness of resolution depends on the stage at which the infection is arrested. If it is controlled in the very early stages, there may not be any residual change in the arachnoid; following an infection of several weeks’ duration, there is a permanent fibrous overgrowth of the meninges, resulting in a thickened, cloudy, or opaque arachnoid and often in adhesions between the pia and arachnoid and even between the arachnoid and dura.

From the earliest stages of meningitis, changes are also found in the small and medium-sized subarachnoid arteries. The endothelial cells swell, multiply, and crowd into the lumen. This reaction appears within 48 to 72 h and increases in the days that follow. The adventitial connective tissue sheath becomes infiltrated by neutrophils. Foci of necrosis of the arterial wall sometimes occur. Neutrophils and lymphocytes migrate from the adventitia to the subintimal region, often forming a conspicuous layer. Later there is subintimal fibrosis. This is a striking feature of nearly all types of subacute and chronic infections of the meninges but most notably of tuberculous and syphilitic meningitis (Heubner arteritis). In the veins, swelling of the endothelial cells and infiltration of the adventitia also occur. Subintimal layering, as occurs in arterioles, is not observed, but there may be a diffuse infiltration of the entire wall of the vessel. It is in veins so affected that focal necrosis of the vessel wall and mural thrombi are most often found. Cortical thrombophlebitis of the larger veins does not usually develop before the end of the second week of the infection.

The unusual prominence of the vascular changes may be related to their anatomic peculiarities. The adventitia of the subarachnoid vessels, both of arterioles and venules, is actually formed by an investment of the arachnoid membrane, which is invariably involved by the infectious process. Thus, in a sense, the outer vessel wall is affected from the beginning by the inflammatory process—an infectious vasculitis. The much more frequent occurrence of thrombosis in veins than in arteries is probably accounted for by the thinner walls and the slower current of blood flow in the former.

Although the spinal and cranial nerves are surrounded by purulent exudate from the beginning of the infection, the perineurial sheaths become infiltrated by inflammatory cells only after several days. Exceptionally, there is infiltration of the endoneurium and degeneration of myelinated fibers, leading to the appearance of fatty macrophages and proliferation of Schwann cells and fibroblasts. More often, there is little or no damage to nerve fibers. Occasionally cellular infiltrations may be found in the optic nerves or olfactory bulbs.

The arachnoid membrane tends to serve as an effective barrier to the spread of infection into the adjacent subdural compartment, but some secondary reaction in this space may occur nevertheless (subdural effusion). This happens far more often in infants than in adults; according to Snedeker and coworkers, approximately 40 percent of infants with meningitis younger than 18 months of age develop subdural effusions. As a rule, there is no subdural pus, only a sterile yellowish exudate. In an even higher percentage of cases, small amounts of fibrinous exudate are found in microscopic sections that include the spinal dura.

In the early stages of meningitis, very little change in the substance of the brain can be detected. Neutrophils appear in the Virchow-Robin perivascular spaces but enter the brain only if there is necrosis. After several days, microglia and astrocytes increase in number, at first in the outer zone and later in all layers of the cortex. The associated nerve cell changes may be very slight. Obviously some disorder of the cortical neurons must take place from the beginning of the infection to account for the stupor, coma, and convulsions that are sometimes observed, but several days must elapse before any change can be demonstrated microscopically. It is uncertain whether these cortical changes are a result of the diffusion of toxins from the meninges, of a circulatory disturbance, or of some other factor, such as increased intracranial pressure or cortical venous thrombosis. The aforementioned changes are not because of invasion of brain substance by bacteria and should therefore be regarded as a noninfectious encephalopathy. When macrophages are exposed to endotoxins, they synthesize and release cytokines, among which are interleukin-1 and tumor necrosis factor. These cytokines are believed to stimulate and modulate the local immune response but may also affect neurons.

There is also little change initially in the ependyma and the subependymal tissues; but in later stages of meningitis, conspicuous changes are invariably found. The most prominent finding is infiltration of the subependymal perivascular spaces and often of the adjacent brain tissue with neutrophilic leukocytes and later with lymphocytes and plasma cells. Microglia and astrocytes proliferate, the latter sometimes overgrowing and burying remnants of the ependymal lining. The bacteria may pass through the ependymal lining and set up this inflammatory reaction in part because this sequence of events is favored by a developing hydrocephalus, which stretches and breaks the ependymal lining. Collections of subependymal astrocytes then begin to protrude into the ventricle, giving rise to a granular ependymitis, which, if prominent, may narrow and obstruct the aqueduct of Sylvius. As any meningitis becomes more chronic, the pia-arachnoid exudate tends to accumulate around the base of the brain (basilar meningitis), obstructing the flow of CSF and giving rise to hydrocephalus. In a survey of community-acquired bacterial meningitis, hydrocephalus occurred in only 5 percent, but it was associated with poor outcome (Kasanmoentalib et al).

The reader may question this digression into matters that are more pathologic than clinical, but knowledge of the morphologic features of meningitis enables one to understand the clinical state and its sequelae. The meningeal and ependymal reactions to bacterial infection and the clinical correlates of these reactions are summarized in Table 32-1.

Types of Bacterial Meningitis

Almost any bacterium gaining entrance to the body may produce meningitis but by far the most common are H. influenzae, N. meningitidis, and S. pneumoniae, which account for approximately 75 percent of sporadic cases. Infection with L. monocytogenes is now the fourth most common type of nonsurgical bacterial meningitis in adults. The following are less frequent causes: Staphylococcus aureus and group A (Streptococcus pyogenes) and group D streptococci, usually in association with brain abscess, epidural abscess, head trauma, neurosurgical procedures, or cranial thrombophlebitis; E. coli and group B streptococci in newborns; Pseudomonas and the Enterobacteriaceae, such as Klebsiella, Proteus, which are usually a consequence of lumbar puncture, spinal anesthesia, or shunting procedures to relieve hydrocephalus. Less-common meningeal pathogens include Salmonella, Shigella, Clostridium, Neisseria gonorrhoeae, and Acinetobacter calcoaceticus. In endemic areas, mycobacterial infections (to be considered further on) are as frequent as those caused by other bacterial organisms. They now assume greater importance in developed countries as the number of immunosuppressed persons increases.

Epidemiology

Pneumococcal, influenzal (H. influenzae), and meningococcal forms of meningitis have a worldwide distribution, occurring mainly during the winter and early spring and, in the case of the first two, also in the fall, and predominating slightly in males. Each has a relatively constant incidence, although epidemics of meningococcal meningitis seem to occur roughly in 10-year cycles. Drug-resistant strains appear with varying frequency, and such information, gleaned from surveillance reports issued by the Centers for Disease Control and Prevention and from reports of local health agencies and hospital infection surveillance, are of great practical importance.

H. influenzae meningitis, formerly encountered mainly in infants and young children, has been nearly eliminated in this age group as a result of vaccination programs in developed countries. It continues to be common in less-developed nations and is now occurring with increasing frequency in adults. Meningococcal meningitis occurs most often in children and adolescents but is also encountered throughout much of adult life, with a sharp decline in incidence after the age of 50 years. Pneumococcal meningitis predominates in the very young and in older adults. Perhaps the greatest change in the epidemiology of bacterial meningitis, aside from the one related to H. influenzae vaccination, has been the increasing incidence of nosocomial infections, accounting for 40 percent of cases in large urban hospitals (Durand et al); staphylococcus and gram-negative bacilli account for a large proportion of these. Noteworthy is the report of Schuchat and colleagues, who found that in 1995, some 5 years after the introduction of the conjugate H. influenzae vaccine, the incidence of bacterial meningitis in the United States had been halved. The yearly incidence rate (per 100,000 population) of the responsible pathogens is approximately as follows: S. pneumoniae, 1.1; N. meningitidis, 0.6; group B streptococcus (newborns), 0.3; L. monocytogenes, 0.2; and H. influenzae, 0.2. In an informative epidemiologic survey of bacterial meningitis in the United States from 1998 to 2007, Thigpen and colleagues found the relative order of the frequency of various organisms to be much the same and again emphasized the decrease in incidence of the disease due mainly to the H. influenza vaccination program. They estimated the recent overall incidence to be 4,100 cases annually, resulting in 500 deaths. Their article is recommended for its detailed analysis of age, race, and underlying medical condition.

Pathogenesis

The most common meningeal pathogens are all normal inhabitants of the nasopharynx in a significant part of the population and depend on antiphagocytic capsular or surface antigens for survival in the tissues of the infected host. To a large extent they express their pathogenicity by extracellular proliferation. It is evident from the frequency with which the carrier state is detected that nasal colonization is not a sufficient explanation of infection of the meninges. Factors that predispose the colonized patient to invasion of the bloodstream, which is the usual route by which these bacteria reach the meninges, are obscure but include antecedent viral infections of the upper respiratory passages or, in the case of S. pneumoniae, infections of the lung. Once blood-borne, it is evident that pneumococci, H. influenzae, and meningococci possess a predilection for the meninges, although the precise factors that determine this tropism are not known. Whether the organisms enter the CSF via the choroid plexus or meningeal vessels is also unknown. It has been variously postulated that the entry of bacteria into the subarachnoid space is facilitated by disruption of the blood–CSF barrier by trauma, circulating endotoxins, or an initial viral infection of the meninges. These organisms, being commensal in most persons, create immunity, but bacteria may nonetheless penetrate the mucosa. Certain features of the organisms enhance their ability to cause infection; this is particularly true of the meningococcus (Rosenstein et al).

Avenues other than the bloodstream by which bacteria can gain access to the meninges include congenital neuroectodermal defects; craniotomy sites; diseases of the middle ear and paranasal sinuses, particularly perilymphatic fistulas; skull fractures; and, in cases of recurrent infection, dural tears from remote minor or major trauma. Occasionally, a brain abscess may rupture into the subarachnoid space or ventricles, thus infecting the meninges. The isolation of anaerobic streptococci, Bacteroides, Actinomyces, or a mixture of microorganisms from the CSF should suggest the possibility of a brain abscess with an associated meningitis.

Clinical Features

Adults and Children

The early clinical effects of acute bacterial meningitis are fever, headache, usually severe, and stiffness of the neck (resistance to passive movement on forward bending), and less often initially, generalized convulsions and a disorder of consciousness (i.e., confusion, drowsiness, stupor, and coma). Flexion at the hip and knee in response to forward flexion of the neck (Brudzinski sign) and inability to completely extend the legs with the hips flexed (Kernig sign) have the same significance as stiff neck but are less-consistently elicitable. Basically, all of these signs are part of a flexor protective reflex (one of the “nocifensive” responses in Fulton’s terms). Stiffness of the neck that is part of paratonic or extrapyramidal rigidity should not be mistaken for that of meningeal irritation. The former is more or less equal in all directions of movement, in distinction to that of meningitis, which is present only or predominantly on forward flexion. Whether it is stiffness in the initial few degrees of flexion of the neck or in the subsequent part of the movement that is more specific for meningitis has been debated; our experience has been that the latter is more sensitive but also proves to be mistaken for other disorders; thus the first may be more specific for meningitis.

Diagnosis of meningitis may be difficult when the initial manifestations consist only of fever and headache, when stiffness of the neck has not yet developed, or when there is only pain in the neck or abdomen or a febrile confusional state or delirium. Also, stiffness of the neck may not be apparent in the deeply stuporous or comatose patient or in the infant or the elderly, as indicated further on.

The symptoms comprised by the meningitic syndrome are common to the three main types of bacterial meningitis, but certain clinical features and the setting in which each of them occurs correlate more closely with one type than another.

Meningococcal meningitis should be suspected when the evolution is extremely rapid (delirium and stupor may supervene in a matter of hours), when the onset is attended by a petechial or purpuric rash or by large ecchymoses and lividity of the skin of the lower parts of the body, when there is circulatory shock, and especially during local outbreaks of meningitis. Because a petechial rash accompanies approximately 50 percent of meningococcal infections, its presence dictates immediate institution of antibiotic therapy, even though a similar rash may be observed with certain viral (echovirus serotype 9 and some other enteroviruses), as well as S. aureus infections, and, rarely, with other bacterial meningitides.

Pneumococcal meningitis is often preceded by an infection in the lungs, ears, sinuses, or heart valves. In addition, a pneumococcal etiology should be suspected in alcoholics, in splenectomized patients, in the very elderly, and in those with recurrent bacterial meningitis, dermal sinus tracts, sickle cell anemia (“autosplenectomized”), and basilar skull fracture. On the other hand, H. influenzae meningitis usually follows upper respiratory and ear infections in the uninocualted child.

Other specific bacterial etiologies are suggested by particular clinical settings. Meningitis in the presence furunculosis or following a neurosurgical procedure directs attention to the possibility of a coagulase-positive staphylococcal infection. Ventricular shunts or drains inserted for the relief of hydrocephalus are particularly prone to infection with coagulase-negative staphylococci and Proprionobacerium acnes and diphteroids. HIV infection, myeloproliferative or lymphoproliferative disorders, defects in cranial bones (tumor, osteomyelitis), collagen diseases, metastatic cancer, and therapy with immunosuppressive agents are clinical conditions that favor invasion by such pathogens as Enterobacteriaceae, L. monocytogenes, A. calcoaceticus, Pseudomonas, and occasionally by parasites.

Focal cerebral signs in the early stages of the disease, although seldom prominent, are most frequent in pneumococcal and H. influenzae meningitides. Some of the transitory focal cerebral signs may represent postictal phenomena (Todd paralysis); others may be related to an unusually intense focal meningitis, for example, purulent material collected in one sylvian fissure. Seizures are encountered most often with H. influenzae meningitis. Although seizures are most common in infants and children, it is difficult to judge the significance, because young children may convulse with fever of any cause. Persistent focal cerebral lesions or intractable seizures usually develop in the second week of the meningeal infection and are caused by an infectious vasculitis, as described earlier, usually with occlusion of surface cerebral veins and consequent infarction of cerebral tissue. Cranial nerve abnormalities are particularly frequent with pneumococcal meningitis, the result of invasion of the nerve by purulent exudate and possibly ischemic damage as the nerve traverses the subarachnoid space.

Infants and Newborns

Acute bacterial meningitis during the first month of life is said to be more frequent than in any subsequent 30-day period of life. It poses a number of special problems. Infants, of course, cannot complain of headache, stiff neck may be absent, and one has only the nonspecific signs of a systemic illness: fever, irritability, drowsiness, vomiting, convulsions, and a bulging fontanel to suggest the presence of meningeal infection. Signs of meningeal irritation do occur, but only late in the course of the illness. A high index of suspicion and liberal use of the lumbar puncture needle are the keys to early diagnosis. Lumbar puncture is ideally performed before any antibiotics are administered for other neonatal infections. An antibiotic regimen sufficient to control a septicemia may allow a meningeal infection to smolder and to flare up after antibiotic therapy for the systemic infection has been discontinued.

A number of other facts about the natural history of neonatal meningitis are noteworthy. It is more common in males than in females, in a ratio of about 3:1. Obstetric abnormalities in the third trimester (premature birth, prolonged labor, premature rupture of fetal membranes) occur frequently in mothers of infants who develop meningitis in the first weeks of life. The most significant factor in the pathogenesis of the meningitis is maternal infection (usually a urinary tract infection or puerperal fever of unknown cause). The infection in both mother and infant is most often caused by gram-negative enterobacteria, particularly E. coli, and group B streptococci, and less often to Pseudomonas, Listeria, S. aureus or epidermidis (formerly albus), and group A streptococci. Analysis of postmortem material indicates that in most cases infection occurs at or near the time of birth, although clinical signs of meningitis may not become evident until several days or a week later.

In infants with meningitis, one should be prepared to find a unilateral or bilateral sympathetic subdural effusion regardless of bacterial type. Young age, rapid evolution of the illness, low polymorphonuclear cell count, and markedly elevated protein in the CSF correlate to some extent with the formation of effusions, according to Snedeker and coworkers. Also, these attributes greatly increase the likelihood of the meningitis being associated with neurologic signs. Transillumination of the skull is the simplest method of demonstrating the presence of an effusion, but CT and MRI are the definitive diagnostic tests. When aspirated, most of the effusions prove to be sterile. If recovery is delayed and neurologic signs persist, a succession of aspirations is required. Children in whom meningitis is complicated by subdural effusions are no more likely, according to authorative sources, to have residual neurologic signs and seizures than are those without effusions.

Spinal Fluid Examination

As already indicated, the lumbar puncture is an indispensable part of the examination of patients with the symptoms and signs of meningitis or of any patient in whom this diagnosis is suspected. Bacteremia is not a contraindication to lumbar puncture. The dilemma concerning the risk of promoting transtentorial or cerebellar herniation by lumbar puncture, even without a cerebral mass, as indicated in Chaps. 2 and 17, has been settled in favor of performing the tap if there is a reasonable suspicion of meningitis. The highest estimates of risk come from studies such as those of Rennick, who reported a 4 percent incidence of clinical worsening among 445 children undergoing lumbar puncture for the diagnosis of acute meningitis; most other series give a lower number. It must be pointed out that a cerebellar pressure cone (tonsillar herniation) may occur in fulminant meningitis independent of lumbar puncture; therefore the risk of the procedure is probably even less than usually stated.

If there is clinical evidence of a focal lesion with increased intracranial pressure, then CT or MRI scanning of the head, looking for a mass lesion, is a prudent first step, but in most cases this is not necessary and should not delay the administration of antibiotics. In an attempt to determine the utility of the CT scan performed prior to a lumbar puncture, Hasbun and colleagues were able to identify several clinical characteristics that were likely to be associated with an abnormality on the scan in patients with suspected meningitis; these included a recent seizure, coma or confusion, gaze palsy, and others. The more salient finding from this study in our opinion was that only 2 percent of 235 patients had a focal mass lesion that was judged a risk for lumbar puncture; many others had CT findings of interest, including some with diffuse mass effect. This study does not entirely clarify the issue of the safety of lumbar puncture but it emphasizes that patients who lack major neurologic findings are unlikely to have findings on the scan that will preclude lumbar puncture.

Only a sizable brain abscess or substantial brain swelling entirely interdicts a lumbar puncture in suspected bacterial meningitis. Furthermore, the fact that death results from cerebral herniation in many fatal cases of bacterial meningitis does not, of course, mean that lumbar puncture precipitated the demise. When there are signs of impending herniation or indications of a dangerous configuration on cerebral images, one may wish to draw blood cultures and and institute empiric treatment rather than take the small risk of precipitation herniation with a lumbar puncture. Any coagulopathy that is deemed a risk for hemorrhagic complication of lumbar puncture should be rapidly reversed if possible.

The spinal fluid pressure is so consistently elevated (above 180 mm H2O) that a normal pressure on the initial lumbar puncture in a patient with suspected bacterial meningitis suggests another diagnosis or raises the possibility that the needle is partially occluded or the spinal subarachnoid space is blocked. Pressures over approximately 350 mm H2O suggest the presence of brain swelling and the potential for cerebellar herniation. Many neurologists favor the administration of intravenous mannitol if the pressure is this high, but this practice does not provide assurance that herniation will be avoided.

A pleocytosis in the spinal fluid is diagnostic. The number of leukocytes ranges from 250 to 100,000/mm3, but the usual number is from 1,000 to 10,000. Occasionally, in pneumococcal and influenzal meningitis, the CSF may contain a large number of bacteria but few, if any, neutrophils for the first few hours. Cell counts of more than 50,000/mm3 raise the possibility of a brain abscess having ruptured into a ventricle. Neutrophils predominate (85 to 95 percent of the total), but an increasing proportion of mononuclear cells is found as the infection continues for days, and especially in partially treated meningitis. In the early stages, careful cytologic examination may disclose that some of the mononuclear cells are myelocytes or young neutrophils. Later, as treatment takes effect, the proportions of lymphocytes, plasma cells, and histiocytes steadily increase.

Substantial hemorrhage or substantial numbers of red cells in the CSF are uncommon in meningitis, the exceptions being anthrax meningitis (see Lanska) as well as certain rare viral infections (Hantavirus, dengue fever, Ebola virus, etc.) and some cases of amebic meningoencephalitis.

The protein content is higher than 45 mg/dL in more than 90 percent of the cases; in most cases, it falls in the range of 100 to 500 mg/dL. The glucose content is diminished, usually to a concentration below 40 mg/dL, or less than 40 percent of the blood glucose concentration (measured concomitantly or within the previous hour), provided that the latter is less than 250 mg/dL. However, in atypical or culture-negative cases, other conditions associated with a reduced CSF glucose should be considered. These include hypoglycemia from any cause; sarcoidosis of the CNS; fungal or tuberculous meningitis; and some cases of subarachnoid hemorrhage, meningeal carcinomatosis, chemically induced inflammation from craniopharyngioma or teratoma, and meningeal gliomatosis. The factors that alter CSF glucose concentration, especially at the extremes of blood glucose, are discussed in Chap. 2.

A special problem pertains to identifying patients with a meningitic syndrome and CSF pleocytosis who do not, in fact, have bacterial meningitis but likely have a viral or other cause for their syndrome. This is driven by a desire to avoid exposure to high-potency intravenous antibiotics that are expensive and potentially dangerous. To address this problem, Nigrovic and colleagues have developed a clinical prediction rule that classifies patients at very low risk for bacterial meningitis if they lack all of the following criteria: positive CSF Gram stain, CSF absolute neutrophil count of at least 1,000 cells/mL, CSF protein of at least 80 mg/dL, peripheral absolute neutrophil count of at least 10,000 cells/mL, and a history of a seizure at or after the time of presentation. This rule was validated in a multicenter retrospective cohort study that encompassed 3,295 patients. Of those who were categorized at very low risk, only 2 had bacterial meningitis. Whether this low rate justifies withholding antibiotics is, of course, a clinical judgement made at the bedside.

The Gram stain of the spinal fluid sediment permits identification of the causative agent in most cases of bacterial meningitis; pneumococci and H. influenzae are identified more readily than meningococci. Small numbers of gram-negative diplococci in leukocytes may be indistinguishable from fragmented nuclear material, which may also be gram-negative and of the same shape as bacteria. In such cases, a thin film of uncentrifuged CSF may lend itself more readily to morphologic interpretation than a smear of the sediment. The most common error in reading Gram-stained smears of CSF is the misinterpretation of precipitated dye or debris as gram-positive cocci or the confusion of pneumococci with H. influenzae. The latter organism may stain heavily at the poles, so that they resemble gram-positive diplococci, and older or rapidly growing pneumococci often lose their capacity to take a gram-positive stain.

Cultures of the spinal fluid, which prove to be positive in 70 to 90 percent of cases of bacterial meningitis, are best obtained by collecting the fluid in a sterile tube and immediately inoculating plates of blood, chocolate, and MacConkey agar; tubes of thioglycolate (for anaerobes); and at least one other broth. The advantage of using broth media is that large amounts of CSF can be cultured. The importance of obtaining blood cultures is mentioned below.

The problem of identifying causative organisms that cannot be cultured, particularly in patients who have received antibiotics, may be overcome by the application of special laboratory techniques. One of these is counterimmunoelectrophoresis (CIE), a sensitive test that permits the detection of bacterial antigens in the CSF in a matter of 30 to 60 min. It is particularly useful in patients with partially treated meningitis, in whom the CSF still contains bacterial antigens but no organisms on a smear or grown in culture.

Several more recently developed serologic methods, radioimmunoassay (RIA) and latex-particle agglutination (LPA), as well as an enzyme-linked immunosorbent assay (ELISA), may be even more sensitive than CIE. An argument has been made that these procedures are not cost-effective, as—in virtually all instances in which the bacterial antigen can be detected—Gram stain also shows the organism. Our sense is that the more expensive tests are of some assistance if Gram stain is difficult to interpret and one or more doses of antibiotics render the cultures negative. Gene amplification by the polymerase chain reaction (PCR) is the most recently developed and most sensitive technique. As it has become more widely available in clinical laboratories, rapid diagnosis has been facilitated (Desforges; Naber), but the use of carefully Gram-stained preparations still needs to be encouraged.

It is of interest that chloride concentrations in the CSF are usually found to be low, possibly reflecting dehydration and low serum chloride levels. In contrast, CSF lactate dehydrogenase (LDH), although also infrequently measured, can be of diagnostic and prognostic value. A rise in total LDH activity is consistently observed in patients with bacterial meningitis; most of this is because of fractions 4 and 5, which are derived from granulocytes. Fractions 1 and 2 of LDH, which are presumably derived from brain tissue, are only slightly elevated in bacterial meningitis but rise sharply in patients who develop neurologic sequelae or later die. Various enzymes in the CSF, derived from leukocytes, meningeal cells, or plasma, may also be increased in meningitis, but the clinical significance of this observation is unknown. Levels of lactic acid in the CSF (determined by either gas chromatography or enzymatic analysis) are also elevated in both bacterial and fungal meningitides (greater than 35 mg/dL) and may be helpful in distinguishing these disorders from viral meningitides, in which lactic acid levels remain normal; however, these ancillary tests are infrequently performed.

Other Laboratory Findings

In addition to CSF cultures, blood cultures should be obtained if possible because they are positive in 40 to 60 percent of patients with H. influenzae, meningococcal, and pneumococcal meningitis, and may provide the only definite clue as to the causative agent. Routine cultures of the oropharynx are as often misleading as helpful, because pneumococci, H. influenzae, and meningococci are common in the throats of healthy persons. In contrast, cultures of the nasopharynx may aid in diagnosis, although often not in a timely way; the finding of encapsulated H. influenzae or groupable meningococci may provide the clue to the etiology of the meningeal infection. Conversely, the absence of such a finding prior to antibiotic treatment makes an H. influenzae and meningococcal etiology unlikely. The leukocyte count in the blood is generally elevated, and immature forms are usually present. Meningitis may be complicated after several days by severe hyponatremia, the result of inappropriate secretion of antidiuretic hormone (ADH).

Imaging Studies

In patients with bacterial meningitis, chest films are essential because they may disclose an area of pneumonia or abscess. Sinus and skull films may provide clues to the presence of cranial osteomyelitis, paranasal sinusitis, mastoiditis, or cranial osteomyelitis, but these structures are better visualized on CT scans, which have supplanted conventional films in most cases. The CT scan is particularly useful in detecting lesions that erode the skull or spine and provide a route for bacterial invasion, such as tumors or sinus wall defects, as well as in demonstrating a brain abscess or subdural empyema. MRI with gadolinium enhancement may display the meningeal exudate and cortical reaction, and both types of imaging, with appropriate techniques, will demonstrate venous occlusions and adjacent infarctions. The issues pertaining to an abscess and to brain swelling in meningitis have already been noted and are disucussed further on as well.

Recurrent Bacterial Meningitis

This is observed most frequently in patients who have had some type of shunting procedure for the treatment of hydrocephalus or who have an incompletely closed dural opening after cranial or spinal surgery. When the origin of the recurrence is inapparent, one should suspect a congenital neuroectodermal sinus or a fistulous connection between the nasal sinuses and the subarachnoid space. The fistula in these latter cases is more often traumatic than congenital in origin (e.g., a previous basilar skull fracture), although the interval between injury and the initial bout of meningitis may be several years. The site of trauma is in the frontal or ethmoid sinuses or the cribriform plate, and S. pneumoniae is the usual pathogen. Often it reflects the predominance of such strains in nasal carriers. These cases usually have a good prognosis; mortality is much lower than in ordinary cases of pneumococcal meningitis.

CSF rhinorrhea is present in most cases of posttraumatic meningitis, but it may be transient and difficult to find. Suspicion of its presence is raised by the recent onset of anosmia or by the occurrence of a watery nasal discharge that is salty to the taste and increases in volume when the head is dependent. One way of confirming the presence of a CSF leak is to measure the glucose concentration of nasal secretions; ordinarily they contain little glucose, but in CSF rhinorrhea the amount of glucose approximates that obtained by lumbar puncture (two-thirds of the serum value). A “dipstick” used for urine testing is sometimes adequate but these are regrettably decresingly available on general hospital wards. Another bedside test for CSF rhinorrhea or otorrhea is to estimate the amount of protein in the fluid. A high protein, sufficient to make a handkerchief stiff on drying, suggests it is of nasal mucosal origin. If the fluid fails to cause a handkerchief to stiffen on drying, a spinal fluid leak is suspected. The most specific and sensitive test for CSF otorrhea and rhinorrhea is the finding of beta2-transferrin (tau), not found in fluids other than CSF.

The site of a CSF leak can sometimes be demonstrated by injecting a dye, radioactive albumin, or water-soluble contrast material into the spinal subarachnoid space and detecting its appearance in nasal secretions or its site of exit by CT scanning. This testing is best performed after the acute infection has subsided. Persistence of CSF rhinorrhea or a spinal CSF leak usually requires surgical repair.

Differential Diagnosis

The diagnosis of bacterial meningitis is usually not difficult in an immunocompetent individual. Febrile patients with lethargy, headache, stiff neck, or confusion of sudden onset—even those with low-grade fever—should generally undergo lumbar puncture if no alternative explanation for the state is evident. It is particularly important to recall the possibility of meningitis in drowsy, febrile, and septic patients in an intensive care unit when no obvious source of fever is apparent. Overwhelming sepsis itself, or the multiorgan failure that it engenders, may cause an encephalopathy; but if there is meningitis, it is imperative, in deciding on the choice of antibiotics, to identify it early. The same can be said for the confused alcoholic patient. Too often, the symptoms are ascribed to alcohol intoxication or withdrawal, or to hepatic encephalopathy, until examination of the CSF reveals meningitis. Although this approach undoubtedly results in many negative spinal fluid examinations, it is preferable to the consequence of overlooking bacterial meningitis. Viral meningitis (which is far more common than bacterial meningitis), subarachnoid hemorrhage, chemical meningitis (following lumbar puncture, spinal anesthesia, or myelography), and tuberculous, leptospiral, sarcoid, and fungal meningoencephalitis, and allergic-immune reactions enter into the differential diagnosis as well, as discussed in later sections.

A number of nonbacterial meningitides must be considered in the differential diagnosis when the meningitis recurs repeatedly and all cultures are negative. Included in this group are Epstein-Barr virus (EBV) infections; Behçet disease, which is characterized by recurrent oropharyngeal mucosal ulceration, uveitis, orchitis, and meningitis; Mollaret meningitis, which consists of recurrent episodes of fever and headache in addition to signs of meningeal irritation (in many cases caused by herpes simplex, as discussed in Chap. 33); and the Vogt-Koyanagi-Harada syndrome, in which recurrent meningitis is associated with iridocyclitis and depigmentation of the hair and skin (poliosis and vitiligo). The CSF in these recurrent types may contain large numbers of lymphocytes or polymorphonuclear leukocytes but no bacteria, and the glucose content is not reduced (see discussion of Chronic and Recurrent Meningitis Chap. 33). These recurrent syndromes rarely present in the fulminant manner of acute bacterial meningitis but sometimes they do, and the CSF formulas can be similar, including a reduction in glucose concentration. Rarely, a fulminant case of cerebral angiitis or intravascular lymphoma will present with headache, fever, and confusion in conjunction with a meningeal inflammatory reaction.

The other intracranial purulent diseases and their differentiation from bacterial meningitis are considered further on in this chapter.

Treatment

Bacterial meningitis is a medical emergency. The first therapeutic measures are directed to sustaining blood pressure and treating septic shock (volume replacement, pressor therapy). A premium is then placed on choosing an antibiotic that is known both to be bactericidal for the suspected organism and is able to enter the CSF in effective amounts. Treatment should begin while awaiting the results of diagnostic tests and may be altered later in accordance with the laboratory findings. Whereas penicillin formerly sufficed to treat almost all meningitides acquired outside the hospital, the initial choice of antibiotic has become increasingly complicated as resistant strains of meningitic bacteria have emerged. The selection of drugs to treat nosocomial infections also presents special difficulties.

In recent years, many reports have documented an increasing incidence of pneumococcal isolates that have a relatively high resistance to penicillin, reaching 50 percent in some European countries. Current estimates are that, in some areas of the United States, 15 percent of these isolates are penicillin-resistant to some degree (most have a relatively low level of resistance). In the 1970s, H. influenzae type B strains producing beta- lactamase, which are resistant to ampicillin and penicillin, were recognized. Currently, 30 percent of H. influenzae isolates produce the beta-lactamase enzyme, but almost all remain sensitive to third-generation cephalosporins (e.g., cefotaxime, ceftizoxime, ceftriaxone).

Recommendations for the institution of empiric treatment of meningitis have been reviewed by van de Beek and colleagues (2006) and by Tunkel and colleagues, often updated, and are summarized in modified form in Table 32-2. The choice of agents varies every few years based on epidemiology and geographic region, but these ones given here are a good approximation to current practice in developed countries.

In children and adults, third-generation cephalosporins such as ceftriaxone, combined with vancomycin is probably the best initial therapy for the three major types of community-acquired meningitides. In areas with low numbers of high-level penicillin-resistant pneumococci, it is possible to avoid adding vancomycin or rifampin. Ampicillin should be added to the regimen in cases of suspected Listeria meningitis, particularly in an immunocompromised patient. Intravenous drug abusers have high rates of meningitis due to S. Aureus and should receive cefepime or ceftazidime with vancomycin. When serious allergy to penicillin and cephalosporins precludes their use, chloramphenicol may be a suitable alternative in some regions, but not for Listeria.

Isolation from the blood or CSF of a resistant organism requires the use of ceftriaxone with the addition of vancomycin and rifampin. N. meningitides, at least in the United States, remains highly susceptible to penicillin and ampicillin. Regional variations and ongoing antibiotic-induced changes in the infecting microorganisms underscore the need for constant awareness of drug resistance in the physician’s local area, especially in the case of pneumococcal infections. Throughout the course of treatment, it is necessary to have access to a laboratory that can carry out rapid and detailed drug-resistance testing.

Nosocomial Meningitis

In cases of meningitis caused by coagulase-positive S. aureus, including those that occur after neurosurgery or major head injury, administration of vancomycin plus a third-generation cephalosporin (e.g., cefepime, ceftazadime, or meropenem) is a reasonable first approach. If Pseudomonas is considered possible, such as after neurosurgery, an antipseudomonal cephalosporin such as ceftazidime or cefapime should be added. Once the sensitivity of the organism has been determined, therapy may have to be altered or may be simplified by using vancomycin or nafcillin alone. These approaches have been reviewed by van de Beek and colleagues (2010). They note that the CSF cell count may be low in cases of ventricular catheter-associated meningitis. They also provide recommendations on the use of prophylactic antibiotics after a basilar skull fracture, a controversial problem that is reviewed in Chap. 35.

Table 32-3 lists the approximate dosages of the most used antibiotics, and Table 32-4 gives reasonable choices of antibiotic for the treatment of specific bacterial isolates.

Duration of Therapy

Most cases of bacterial meningitis should be treated for a period of 10 to 14 days except when there is a persistent parameningeal focus of infection (otitic or sinus origin), in which cases longer treatment may be needed. Antibiotics should be administered in full doses parenterally (preferably intravenously) throughout the period of treatment. Treatment failures with certain drugs, notably ampicillin, may be attributable to oral or intramuscular administration, resulting in inadequate concentrations in the CSF. Repeated lumbar punctures are not necessary to assess the effects of therapy as long as there is progressive clinical improvement. The CSF glucose may remain low for many days after other signs of infection have subsided and should occasion concern only if bacteria are present in the fluid and the patient remains febrile and ill.

Persistence of fever or the late appearance of drowsiness, hemiparesis, or seizures should raise the suspicion of subdural effusion, mastoiditis, venous sinus thrombosis, cortical vein or jugular phlebitis, or brain abscess; all require that therapy be continued for a longer period. Bacteriologic relapse after treatment is discontinued requires reinstitution of therapy and exploration for a persistent parameningeal focus of infection, such as in the spinal column.

Corticosteroids

Controlled studies several decades ago were unable to demonstrate beneficial effects of corticosteroids in the treatment of bacterial meningitis. More recent studies have given another perspective of the therapeutic value of dexamethasone in children and adults with meningitis. In children, although mortality was not affected in the main study conducted by Lebel and colleagues, fever subsided more rapidly and the incidence of sensorineural deafness and other neurologic sequelae was reduced, particularly in those children with H. influenzae meningitis. On these grounds, it has been recommended that the treatment of childhood meningitis include dexamethasone in high doses (0.15 mg/kg qid for 4 days), instituted as soon as possible.

Despite similarly conflicting results from earlier studies of corticosteroids in adults, the trial by deGans and van de Beck has demonstrated a reduction in mortality and improved overall outcome if dexamethasone 10 mg is given just before the first dose of antibiotics and then repeated q6h for 4 days. The improvement was largely in patients who were infected with pneumococcus. Seizures and coma were reduced in incidence as a result of the administration of corticosteroids, but neurologic sequelae, such as hearing loss, were not affected. Based on a number of smaller studies, some authorities in the field of bacterial meningitis have endorsed the administration of dexamethasone in the doses mentioned above, but only if they can be started before antibiotics, and only in those with presumed pneumococcal infection (see Tunkel and Scheld). They also advise against the use of the drug if there is septic shock. In developing countries, especially those with high rates of AIDS, the benefits of adjuvant dexamethasone have not been clear. Improved survival was limited to those who ultimately had bacteria isolated from the CSF, in contrast to those with suspected meningitis but negative cultures.

Nonetheless, the incidence of deafness was reduced (Nguyen et al; Scarborough et al). The use of corticosteroids is therefore suggested in cases with overwhelming infection at any age (very high CSF pressure or signs of herniation, high CSF bacterial count with minimal pleocytosis, and signs of acute adrenal insufficiency, i.e., the Waterhouse-Friderichsen syndrome). It is not always possible to determine with certainty at the first presentation those cases that will be culture positive but one is referred back to the prediction “rules” validated by Nigrovic and colleagues.

Other Forms of Therapy

There is no evidence that repeated drainage of CSF is therapeutically effective. In fact, increased CSF pressure in the acute phase of bacterial meningitis is largely a consequence of cerebral edema, in which case the lumbar puncture may predispose to cerebellar herniation. As already mentioned, a second lumbar puncture to gauge the effectiveness of treatment is generally not necessary, but it may be of value if the patient is worsening without explanation. Mannitol and urea have been employed with apparent success in some cases of severe brain swelling with unusually high initial CSF pressures (400 mm H2O). Acting as osmotic diuretics, these agents enter cerebral tissue slowly, and their net effect is to decrease brain water. However, neither mannitol nor urea has been studied in controlled fashion in the management of meningitis. An adequate but not excessive amount of intravenous normal saline (and avoiding fluids with free water) should be given. Particular care should be taken with children to avoid hyponatremia and water intoxication—potential causes of brain swelling.

Antiepileptic drugs need not be administered routinely but should be given if a seizure has occurred or there is evidence of cortical vein thrombosis.

Prophylaxis

Household contacts of patients with meningococcal meningitis should be protected with antibiotic treatment. The risk of secondary cases is small for adolescents and adults, but ranges from 2 to 4 percent for those younger than 5 years of age and is probably higher in the elderly. A single dose of ciprofloxacin is effective. An alternative is a daily oral dose of rifampin—600 mg q12h in adults and 10 mg/kg q12h in children—for 2 days. If 2 weeks or more have elapsed since the index case was found, no prophylaxis is needed.

As mentioned, immunization against H. influenzae is steadily reducing the incidence of meningitis from this organism. Also, many institutions housing young adults, such as colleges and the military, have instituted programs of immunization against N. meningitidis.

Prognosis and Sequelae of Meningitis

Untreated, bacterial meningitis is usually fatal. The overall mortality rate of uncomplicated but treated H. influenzae and meningococcal meningitis has remained at approximately 5 percent for many years; in pneumococcal meningitis, the rate is considerably higher (approximately 15 percent), perhaps related to the older and sicker population that is affected. Fulminant meningococcemia, with or without meningitis, also has a high mortality rate because of the shock associated with adrenocortical hemorrhages (Waterhouse-Friderichsen syndrome). A disproportionate number of deaths from meningitis occur in infants and in the aged. The mortality rate is highest in neonates, from 40 to 75 percent in several reported series, and at least half of those who recover show serious neurologic sequelae. In adults, the presence of bacteremia, coma, seizures, and a variety of concomitant diseases—including alcoholism, diabetes mellitus, multiple myeloma, and head trauma—all worsen the prognosis. The triad of pneumococcal meningitis, pneumonia, and endocarditis (Osler triad) has a particularly high fatality rate.

Suprisingly often, it is impossible to explain the death of a patient with meningitis or at least to trace it to a single specific mechanism. The effects of overwhelming infection, with bacteremia and hypotension, or brain swelling and cerebellar herniation, are clearly implicated in some patients during the initial 48 h. These events may occur in bacterial meningitis of any etiology; however, they are far more frequent in meningococcal and pneumococcal infection. Some of the deaths occurring later in the course of the illness are attributable to respiratory failure, often consequent to aspiration pneumonia.

It has been stated that relatively few adult patients who recover from meningococcal meningitis show residual neurologic defects, whereas such defects are encountered in at least 25 percent of children with H. influenzae meningitis and up to 30 percent of child and adult patients with pneumococcal meningitis. Kastenbauer and Pfister, reporting on adults with pneumococcal meningitis, have emphasized that the mortality remains quite high and that cerebral venous or arterial thrombosis occurred in almost a third of cases, as discussed further on. They also had two patients with an associated myelitis. We have seen several instances of upper cervical cord and lower medullary infarction in bacterial meningitis; quadriparesis and respiratory failure were the result of compression from descent of the cerebellar tonsils (Ropper and Kanis). As already discussed, the role of lumbar puncture in promoting this complication of cerebellar herniation has not been clarified.

Among infants who survive H. influenzae meningitis, Ferry and coworkers, in a prospective study of 50 cases, found that about half were normal, whereas 9 percent had behavioral problems and about 30 percent had neurologic deficits (seizures or impairment of hearing, language, mentation, and motor function). In a report of a series of 185 children recovering from bacterial meningitis, Pomeroy and associates found that 69 were not normal neurologically at the end of a month; however, at the end of a year, only 18 were left with a hearing deficit, 13 with late afebrile seizures, and 8 with multiple deficits. The presence of a persistent neurologic deficit was the only independent predictor of later seizures. Dodge and colleagues in past decades found that 31 percent of children with pneumococcal meningitis were left with persistent sensorineural hearing loss; for meningococcal and H. influenzae meningitis, the figures were 10.5 and 6 percent, respectively. These events are seemingly less frequent now, specifically in developed countries, but still reflect the seriousness of these sequealae in less advantaged regions of the world.

Cranial nerve palsies other than deafness, if they occur, tend to disappear after a few weeks or months. Deafness in these infections is a result of suppurative cochlear destruction or, less often now, of the ototoxic effects of aminoglycoside antibiotics. Bacteria reach the cochlea mainly via the cochlear aqueduct, which connects the subarachnoid space to the scala tympani. This occurs quite early in the course of infection, hearing loss being evident within a day of onset of the meningitis; in about half or most of such cases, the acute deafness resolves. Hydrocephalus is an infrequent complication that may become manifest months after treatment and then requires shunting if gait or mentation is affected. It may be difficult to determine on clinical grounds whether a residual state of imbalance is the result of hydrocephalus or of eighth nerve damage. The acute complications of bacterial meningitis, the intermediate and late neurologic sequelae, and the pathologic basis of these effects are summarized in Table 32-1.

Quite apart from acute and subacute bacterial endocarditis, which may give rise to cerebral embolism and characteristic inflammatory reactions in the brain (see further on), there are several systemic bacterial infections that are complicated by a special type of encephalitis or meningoencephalitis. Three common ones are Mycoplasma pneumoniae infections, L. monocytogenes meningoencephalitis, and Legionnaire disease. Probably Lyme borreliosis should be included in this category but it is more chronic and is described further on in this chapter with the spirochetal infections. The rickettsial encephalitides (particularly Q fever), which mimic bacterial meningoencephalitis, are also addressed later in the chapter. Catscratch disease is another rare cause of bacterial meningoencephalitis. Meningoencephalitis caused by brucellosis occurs very rarely in the United States. Whipple disease, discussed later, which appears to be a focal invasion of the brain by an unusual intracellular bacterium, is an oddity but also belongs in this category.

Mycoplasma pneumoniae

This organism, which causes 10 to 20 percent of all pneumonias, is associated with a number of neurologic syndromes. Guillain-Barré polyneuritis, cranial neuritis, acute myositis, aseptic meningitis, transverse myelitis, global encephalitis, seizures, cerebellitis, acute disseminated (postinfectious) encephalomyelitis, and acute hemorrhagic leukoencephalitis (Hurst disease) have all been reported in association with mycoplasmal pneumonias or with serologic evidence of a recent infection (Westenfelder et al; Fisher et al; Rothstein and Kenny). We have observed several patients with striking cerebral, cerebellar, brainstem, or spinal syndromes incurred during or soon after a mycoplasmal pneumonia or tracheobronchitis. In addition to the cerebellitis, which is clinically similar to the illness that follows varicella, unusual encephalitic syndromes of choreoathetosis, seizures, delirium, hemiparesis, and acute brain swelling (Reye syndrome) have each been reported in a few cases. The incidence of these complications has been estimated as 1 in 1,000 mycoplasmal infections, but it may approach 5 percent when more careful surveillance is carried out during epidemics. A severe prodromal headache has occurred in most of our cases. At the time of onset of the neurologic symptoms, there may be scant signs of pneumonia, and in some patients, only an upper respiratory syndrome occurs.

The mechanism of cerebral damage that complicates mycoplasmal infections has not been established, but recent evidence suggests that the organism is present in the CNS during the acute illness. To our knowledge, the organism has been cultured from the brain in only one fatal case, but PCR techniques have detected fragments of mycoplasmal DNA in the spinal fluid from several patients (Narita et al). In other instances, the nature of the neurologic complications and their temporal relationship to the mycoplasmal infection clearly suggest that secondary autoimmune factors are operative, i.e., that these are instances of postinfectious encephalomyelitis (a type of acute disseminated encephalomyelitis described in Chap. 36). This is almost certainly the mechanism of postmycoplasmal Guillain-Barrè syndrome. Most of the patients with the infectious variety have recovered with few or no sequelae, but rare fatalities are reported.

The CSF usually contains small numbers of lymphocytes and other mononuclear cells and an increased protein content. The diagnosis can be established by culture of the organism from the respiratory tract (which is difficult), by rising serum titers of complement-fixing immunoglobulin IgG and IgM antibodies and cold agglutinin antibodies in the blood and CSF, or by DNA detection techniques from the CSF.

Treatment

Macrolide antibiotics such as azithromicin and clarithromicin but also erythromycin and tetracycline derivatives reduce morbidity, mainly by eradicating the pulmonary infection, but the effects of antibiotics on the nervous system complications are not known.

Listeria monocytogenes

Meningoencephalitis from this organism is most likely to occur in immunosuppressed and debilitated individuals and is a well-known and occasionally fatal cause of meningitis in the newborn. Meningitis is the usual neurologic manifestation, but there are numerous recorded instances of isolated focal bacterial infectious encephalitis, rarely with a normal CSF, most cases showing a pleocytosis that may be initially polymorphonuclear. Between 1929, when the organism was discovered, and 1962, when Gray and Killinger collected all the reported cases, it was noted that 35 percent of patients had either meningitis or meningoencephalitis as the primary manifestation.

The infection may take the form of a brainstem encephalitis, or “rhombencephalitis,” specifically with several days of headache, fever, nausea, and vomiting followed by asymmetrical cranial-nerve palsies, signs of cerebellar dysfunction, hemiparesis, quadriparesis, or sensory loss. Respiratory failure has been reported. Of 62 cases of Listeria brainstem encephalitis reported by Armstrong and Fung, 8 percent were in immunosuppressed patients, meningeal signs were present in only half the patients, and the spinal fluid often showed misleadingly mild abnormalities. CSF cultures yielded Listeria in only 40 percent of cases (blood cultures were even more often normal). Consistent with our experience, the early CT scan was often normal; MRI, however, has revealed abnormal signals in the parenchyma of the brainstem.

The monocytosis, which gives the organism its name, refers to the reaction in the peripheral blood in rabbits but these cells have not been prominent in the blood or CSF of patients. One patient described by Lechtenberg and coworkers had a proven brain abscess; other patients have had multiple small abscesses (Uldry et al) but it is not clear if this is a uniform feature of the illness that explains the rhombencephalitis. Judging from the clinical signs in some cases, the infection appears to affect both the brainstem parenchyma and the extraaxial portion of the lower cranial nerves.

Treatment

The treatment is ampicillin (2 g intravenously q4h) in combination with gentamicin (5 mg/kg intravenously in 3 divided doses daily). If the condition of the host is compromised, the outcome is often fatal, but most of our patients without serious medical disease have made a full and prompt recovery with treatment.

Meliodosis

In India and Southeast Asia, particularly Cambodia and Thailand, a brainstem, cerebellar, and meningitic illness, similar to that caused by Listeria, results instead from meliodosis (Burkholderia pseudomallei). It should be suspected in returning travelers from that region but the disease is, of course, well known to physicians in areas endemic for the organism. Diabetics are particularly prone to this infection. The CSF shows one to several dozen white blood cells and raised protein but glucose may be normal. There is usually an associated pulmonary infection but this may be minor and the degree of temperature elevation varies. The diagnosis can be made by the culture of the organism from any body site, CSF, pharynx, blood, urine or sputum, as it is not a normal commensal bacterium but both blood agar and special Ashdown’s medium containg gentamicin are required. There is a commercial serologic test but there are high background rates of positivity in endemic regions.

Treatment

This is in two phases, an intesive eradication component with high intravenous doses of ceftazidime (or several equivalent regimens) for 10 to 14 days, followed by an eradication phase that is necessary to prevent relapse, using co-trimoxazole alone or accompanied by doxycycline.

Legionella

This potentially fatal respiratory disease caused by the gram-negative bacillus Legionella pneumophila, first came to medical notice in July 1976, when a large number of members of the American Legion fell ill at their annual convention in Philadelphia. The fatality rate was high. In addition to the obvious pulmonary infection, manifestations referable to the CNS and other organs were observed regularly. Lees and Tyrrell described patients with severe and diffuse cerebral involvement, and Baker and associates and Shetty and colleagues described others with cerebellar and brainstem syndromes. The clinical details have varied. One constellation consisted of headache, obtundation, acute confusion or delirium with high fever, and evidence of pulmonary distress; another took the form of tremor, nystagmus, cerebellar ataxia, extraocular muscle and gaze palsies, and dysarthria.

Other neurologic abnormalities have been observed, such as inappropriate ADH secretion, or a syndrome of more diffuse encephalomyelitis or transverse myelitis, similar to that observed with Mycoplasma infections. The CSF is usually normal and CT scans of the brain are negative, a circumstance that makes diagnosis difficult. The neuropathologic abnormalities have not been studied. Suspicion of the disease, based on exposure or on the presence of an atypical pneumonia, should prompt urine antigen and culture of blood and CSF. Serologic tests are available but require paired sera and have little impact on clinical decision making. In most patients the signs of CNS disorder resolve rapidly and completely, although residual impairment of memory and a cerebellar ataxia have been recorded. To date, the Legionella bacillus has not been isolated from the brain or spinal fluid.

Treatment

Treatment in adults has consisted of one of levofloxacin, moxifloxacin, or azithromycin; rifampin is sometimes used. In the past, erythromycin, 0.5 to 1.0 g was used intravenously q6h for a 7 to 10 days.

Catscratch Fever (Bartonella henselae)

Reports of over 100 cases of encephalitis from catscratch disease have appeared in the medical literature and several have occurred on our services over the years, for which reason we do not consider it rare. The causative organism is a gram-negative bacillus now called Bartonella henselae (formerly Rochalimaea henselae). The illness begins as unilateral axillary or cervical adenopathy occurring after a seemingly innocuous scratch (rarely a bite) from an infected cat. The cases with which we are familiar began with an encephalopathy and high fever (higher temperature than with most of the other organisms that are capable of causing a bacterial encephalitis), followed by seizures or status epilepticus. The organism has also been implicated in causing a focal cerebral vasculitis in AIDS patients as well as neuroretinitis in both immunocompromised and immunocompetent patients. Demonstration of elevated complement-fixing titers and detection of the organism by PCR or by silver staining from an excised lymph node are diagnostic. A single high antibody titer is probably inadequate for this purpose.

Treatment

First-line treatment is with azithromycin or doxycycline, sometimes with rifampin in recalcitrant cases. Erythromycin is used less frequently. Most patients recover completely, but one of our patients and a few reported by others have died.

Anthrax

This rare form of meningoencephalitis is included here because of the current interest in Bacillus anthracis as a bioweapon. Lanska was able to collect from the literature 70 patients with meningeal infection, most of whom were encephalopathic. He has estimated that fewer than 5 percent of infected individuals will acquire a meningoencephalitis; in a 2001 U.S. outbreak, only 1 of 11 cases with anthrax pneumonitis developed this complication. Reflecting the main site of natural infection, the majority of cases originated in cutaneous anthrax. In addition to a typically fulminating course after a prodrome of one or several days, the exceptional feature was a hemorrhagic and inflammatory spinal fluid formula. Subarachnoid hemorrhage was prominent in autopsy material, presumably reflecting necrosis of the vessel walls as a toxic effect of B. anthracis.

Treatment

Although natural isolates are sensitive to penicillin, bioengineered strains are resistant; therefore, combined treatment with ciprofloxacin with clindamycin, rifampin, or meropenem has been recommended initially. The benefit of specific antitoxin is uncertain once meningoencephalitis has occurred.

Recently, very similar overwhelming cases with meningitis and subarachnoid hemorrhage caused by Bacillus cereus have appeared in immunosuppressed patients.

Brucellosis

This worldwide disease of domesticated livestock is frequently transmitted to humans in areas where the infection is enzootic. In the United States, it is distinctly rare, with 200 cases or less being reported annually since 1980, some in abattoir workers. During the 1950s it was a fashionable explanation for chronic fatigue. In the Middle East, infection with Brucella is still frequent, attributable to the ingestion of raw milk. In Saudi Arabia, for example, al Deeb and coworkers reported on a series of 400 cases of brucellosis, of which 13 presented with brain involvement (acute meningoencephalitis, papilledema and increased intracranial pressure, and meningovascular manifestations). The CSF showed a lymphocytic pleocytosis and increased protein content. Blood and CSF antibody titers to the organism were greater than 1:640 and 1:128, respectively.

Treatment

Prolonged treatment with doxycycline with streptomycin or gentamicin; an alternative is doxycycline plus rifampin to suppress the infection.

Whipple Disease

This is a rare but often-discussed disorder, predominantly of middle-aged men. Weight loss, fever, anemia, steatorrhea, abdominal pain and distention, arthralgia, lymphadenopathy, and hyperpigmentation are the usual systemic manifestations. Less often, infection is associated with a number of neurologic syndromes. It is caused by a gram-positive bacillus, Tropheryma whipplei, which resides predominantly in the gut. Biopsy of the jejunal mucosa, which discloses macrophages filled with the periodic acid-Schiff (PAS)-positive organisms, is diagnostic. PAS-positive histiocytes have also been identified in the CSF, as well as in periventricular regions, in the hypothalamic and tuberal nuclei, and diffusely scattered in the brain.

The neurologic manifestations most often take the form of a slowly progressive memory loss or dementia of subacute or early chronic evolution. Supranuclear ophthalmoplegia, ataxia, seizures, myoclonus, nystagmus, and a highly characteristic oculomasticatory movement described as myorhythmia (which looks to us like rhythmic myoclonus) have been noted less often than the dementing syndrome. The rhythmic myoclonus or spasm occurs in synchronous bursts involving several adjacent regions, mainly the eyes, jaw, and face. This movement disorder is fairly specific but insensitive for Whipple disease, occurring in only approximately 10 percent of patients. As pointed out by Matthews and colleagues, cerebellar ataxia, although obviously much less specific for Whipple disease of the brain, is more frequent, occurring in about half the documented cases. Almost always, the myorhythmias are accompanied by a supranuclear vertical gaze paresis that sometimes affects horizontal eye movements as well. Presumably, the neurologic complications are the result of infiltration of the brain by the organism, but this has not been satisfactorily established.

Approximately half of the patients have a mild pleocytosis and some of these have PAS-positive material in the CSF. A variety of brain imaging abnormalities have been recorded, none characteristic, but either enhancing focal lesions or a normal scan may be found. The diagnosis is made mainly from PAS staining of an intestinal (jejunal) biopsy, as already mentioned, supplemented by PCR testing of the bowel tissue or biopsy material from brain or lymph node. In cases of subacute progressive limb and gait ataxia occurring in middle-aged or older men in whom no cause is uncovered by less-invasive means, it is justifiable to perform these tests (see Chap. 5). Rarely, the neurologic symptoms may occur in the absence of gastrointestinal disease (Adams et al, 1987). In the review of 84 cases of cerebral Whipple disease by Louis and colleagues, 71 percent had cognitive changes, half with psychiatric features; 31 percent had myoclonus; 18 percent had ataxia; and 20 percent had the oculomasticatory and skeletal myorhythmias (Schwartz et al).

Treatment

A course of induction by penicillin or ceftriaxone for 2 weeks followed by trimethoprim-sulfamethoxazole or doxycycline continued for 1 year are the currently recommended regimens. An alternative approach is 2 weeks of ceftriaxone followed by treatment with trimethoprim-sulfamethoxazole or a tetracycline for a year. Antibiotic-resistant cases and instances of relapse after antibiotic treatment are known. The review by Anderson may be consulted for details.

“Acute Toxic Encephalopathy”

We are uncertain of its status but have been shown putative cases of this disorder described by Lyon and colleagues as “acute encephalopathy of obscure origin in children,” a febrile and sometimes fatal illness that could not be ascribed to direct infection of the nervous system. During the height of a systemic bacterial or sometimes viral infection, the child sinks into coma, seizures are infrequent, the neck is supple, and the spinal fluid shows no changes or only a few cells. This is undoubtedly an illness of diverse causes, common among them being fluid overload and electrolyte imbalance, Reye syndrome (see Chap. 30) and, possibly most commonly, the immune condition of postinfectious encephalitis (see Chap. 36). Nonetheless, cases continue to be reported, such as those of Thi and colleagues, which can only be classified as a noninfectious bacterial encephalopathy or encephalitis. A relationship to the “septic encephalopathy” of adults, which has been emphasized by the group from London, Ontario, is possible but unproved. The term acute toxic encephalopathy still has some utility in cases of obscure cause, but a careful search for better-characterized causes of febrile coma must be undertaken. The acute necrotizing encephalopathy that has been reported, particularly in Asian children after influenza, belongs to this category and consists of a number of diseases as discussed by Mizuguchi and coworkers.

Subdural empyema is an intracranial (rarely intraspinal) purulent process between the inner surface of the dura and the outer surface of the arachnoid that occurs mainly in children. The term subdural abscess, among others, had been applied to this condition but the proper name is empyema

Related posts:

Disorders of the Special Senses Chapter 20. Delirium and Other Acute Confusional States Chapter 24. Fatigue, Asthenia, Anxiety, and Depression Chapter 28. Normal Development and Deviations in Development of the Nervous System Chapter 39. Degenerative Diseases of the Nervous System Chapter 38. Developmental Diseases of the Nervous System

Stay updated, free articles. Join our Telegram channel

Join