Rabies encephalitis is a feared, virtually uniformly fatal form of central nervous system infection. The incidence of rabies encephalitis in the United States is almost certainly underestimated because of the predominance of bat-borne rabies, which can be spread without traumatic exposure. Because of its rarity in developed countries, rabies encephalitis has been seldom studied with modern imaging techniques.
University-based teaching hospital.
A case of pathologically confirmed rabies encephalitis is presented. Diagnosis of rabies was made by seroconversion testing while the patient was alive and was confirmed postmortem by the presence of rabies antigens and Negri bodies in the brain. The patient had 2 magnetic resonance studies done that showed dramatic abnormalities in the medulla and pons that correlated with features of the neurologic examination and hypothalamic-pituitary abnormalities.
The patient had a fulminant encephalitic course that ended in death.
Rabies is an uncommon cause of fatal encephalitis. Anatomic imaging studies such as computed tomographic and magnetic resonance scans have generally been negative in confirmed cases of rabies. We report a case of confirmed rabies with extensive brainstem and hypothalamic-pituitary abnormalities on magnetic resonance imaging. Although these findings are nonspecific, they should raise the clinical suspicion of rabies in the setting of aggressive encephalitis of unclear cause, and appropriate diagnostic tests should be performed.
RABIES HAS long been one of the most feared illnesses encountered by humans. It is uniformly fatal once clinical symptoms are evident unless the patient has been previously vaccinated.1 The incidence of rabies in the United States is very low, and rabies infections are uncommon in humans. There are, however, significant reservoirs of infection in small mammals in the wild,2 and rabies infection is still an important risk for people who are exposed to bats or other small mammals in the wild. Rabies infections are much more common in certain countries, such as India, where the rabies virus has not been eradicated from the canine population.1 Infection may occur by animal bite, scratch, or inhalation of virions.1,2 Once inoculated, the virus spreads centripetally via peripheral axons to the central nervous system (CNS), and once in the CNS, it is capable of transsynaptic spread within the CNS and back to the periphery.1,3
Although many clinical descriptions and reviews of rabies infections have been published, few reports include a detailed correlation of the progressive clinical findings and the accompanying neuroradiologic findings.4 We describe a case of rabies in a migrant farm worker presenting with fulminant encephalitis associated with prominent brainstem signs. During the course of the illness, the patient seroconverted and at autopsy had findings indicative of rabies. A summary of this patient's clinical course has been reported by the Centers for Disease Control and Prevention.5
The patient was a 27-year-old male migrant worker. He developed headaches on September 8, 1995, that worsened over the next 2 days. On September 9, he was admitted to a hospital in California's Central Valley with headache, temperature of 39.5°C, chills, and vomiting. He had a lumbar puncture with no recorded opening pressure, 20 red blood cells, 0 white blood cells, a protein level of 680 g/L, and a normal glucose level. Treatment was started with antibacterial agents and he was transferred to the University of California, San Francisco, on September 11 because of deteriorating mental status. Another lumbar puncture showed an opening pressure of 280 mm of water, clear fluid, 6 white blood cells (differential cell count: lymphocytes, 0.84), protein level of 1240 g/L, and a glucose level of 5.1 mmol/L (92 mg/dL). We felt a nonfocal neurologic examination would be useful, but he was very agitated and combative, and was sedated and intubated. The remainder of his examination was notable for a temperature of 40.3° C, and acyclovir therapy was added to broad-spectrum antibacterial agents. Detailed neurologic examination was not possible, but the neurology consultant noted that the patient had minimally responsive pupils both directly and consensually, and an absent corneal reflex on the left side. He had spontaneous movements of all 4 extremities and withdrew from painful tactile stimuli. He also had multifocal myoclonus. Reflexes and flexor plantar responses were normal.
Magnetic resonance imaging (MRI) (GE 1.5T Signa system; General Electric; Milwaukee, Wis) of the brain was performed on September 12 and showed an abnormally high signal on T2-weighted images in the pontine tegmentum and area postrema of the medulla, with only minimal postgadolinium enhancement in these areas (Figure 1). Subtle increased enhancement in the infundibulum and suprasellar regions was noted. At this time, these findings were thought to be nonspecific and consistent with encephalitis, vasculitis, or a demyelinating process. On September 13, the serum creatine kinase level was greater than 8000 U/L, and an electroencephalogram showed diffuse slowing with no abnormal spike activity. The serum testosterone level was 1.1 nmol/L (32 ng/dL; normal, 7.8-31.2 nmol/L [225-900 ng/dL]), and the luteinizing hormone level was less than 2.5 IU/L (normal, 4-18 IU/L), suggesting dysfunction of the hypothalamic-pituitary axis.
Images from the magnetic resonance examination of September 12 (GE 1.5T Signa system; General Electric). For T1-weighted images, repetition time (TR) was 500 milliseconds; echo time (TE), 11 milliseconds; for T2-weighted images, TR, 2500 milliseconds; TE, 80 milliseconds; for FLAIR (fluid-attenuated inversion recovery) sequences, TR, 10,000 milliseconds; inversion time, TI, 2200 milliseconds; TE, 140 milliseconds. The slice thickness was 5 mm. A, An axial T2-weighted image through the medulla demonstrates focal increased signal in the dorsal midline (arrow). B, A more cephalad T2-weighted image demonstrates focal increased signal in the dorsal pons (arrow). C, A sagittal fast spin-echo T2-weighted image shows the areas of high signal in the dorsal medulla and pons. There is also high signal in the region of the hypothalamus (arrows) and subtle high signal in the splenium of the corpus callosum (arrowhead). D, A sagittal postgadolinium T1-weighted image shows very faint enhancement in the dorsal medulla and pons, as well as questionable faint enhancement in the inferior colliculus (arrow).
On September 14, the patient's fever dropped and he was somewhat more alert at times and occasionally responsive to commands, but significant agitation necessitated continued sedation. On September 15, he was noted to have persistent upward gaze deviation, absent oculocephalic reflexes, and absent corneal reflexes bilaterally, and was completely unresponsive to verbal or tactile stimuli. His pupils were small but minimally reactive. His stretch and plantar reflexes were unchanged. An electroencephalogram, repeated to rule out status epilepticus as the cause of tonic eye deviation, showed diffuse slowing with no epileptiform activity. Magnetic resonance imaging was repeated to check for any supratentorial lesions that could be biopsied and showed an interval increase in the extent and severity of the T2 signal abnormalities in the pons and medulla, but no increase in enhancement of the brainstem lesions (Figure 2). On September 18, the patient was noted to have no remaining brainstem function. The invasiveness of his level of care was diminished, and support was withdrawn on September 21 after consultation with his family in Mexico.
Multiple images from the magnetic resonance examination of September 15. A, An axial T2-weighted image shows marked progression of signal abnormality, now involving the dorsal half of the medulla. B, A more cephalad T2-weighted image similarly shows significant progression of signal abnormality in the dorsal pons. There is a suggestion that faintly increased signal may be tracking along the course of the facial nerve as it travels from its nucleus to the root exit zone (arrowheads), possibly related to antegrade spread of virus. Inflammatory changes in the paranasal sinuses and mastoid air cells are presumably secondary to intubation. C, A sagittal FLAIR (fluid-attenuated inversion recovery) sequence demonstrates to good advantage the progression of signal abnormality in the dorsal brainstem, hypothalamic region, and splenium. There is also a suggestion of abnormal high signal in the inferior and superior colliculi.
The diagnosis of rabies was first considered on September 12 because the patient had a history of multiple exposures to bats and their roosting areas, including an exposure 1 or 2 years earlier, when a bat reportedly landed on his chest without biting him. A cerebrospinal fluid sample from September 12 and serum sample from September 13 were sent for rabies testing; both results were negative, with titers of less than 1:8. A nuchal biopsy and corneal impressions performed on September 13 did not show any rabies antigens using immunofluorescent stains. Repeated rabies titers on September 19 and 21 were positive at 1:256 and greater than 1:2048, respectively, however, confirming the clinical diagnosis of rabies.
An autopsy, including skin biopsy, corneal imprints, and brain examination (with immunofluorescent examination of brain for rabies antigens), was performed. On gross examination, the brain was edematous with marked congestion of the leptomeninges. The midbrain was soft, and the cerebellar vermis and tonsils were soft and necrotic. There were no Duret hemorrhages noted. Microscopic examination revealed microglial nodules in the basis pontis and pontine tegmentum with associated lymphocytic infiltrates, and electron microscopy revealed cytoplasmic rabies viral particles. The cerebellum had a diffuse loss of Purkinje cells, and a single Negri body was noted in a Purkinje cell. The basal ganglia, cortex, and hippocampus had extensive perivascular and parenchymal chronic inflammatory cells with associated microglial nodules. The pituitary gland showed extensive coagulative necrosis and evidence of a chronic inflammatory infiltrate. Direct immunofluorescence showed rabies antigens diffusely in the pons, cerebral cortex, and hippocampus, but not in the cerebellum, corneal imprints, or nuchal skin biopsy specimen. Polymerase chain reaction analysis of the tissue at the Centers for Disease Control and Prevention implicated a strain of the rabies virus usually associated with the Mexican free-tailed bat.5
There have been 35 confirmed cases of rabies in the United States since 1980, all ending in death,2,6,7 and it is estimated that there are about 35,000 deaths worldwide from rabies annually. The incidence of rabies may be underestimated because of the decreasing number of autopsies performed and the difficulty of making the diagnosis antemortem, particularly if appropriate clinical suspicion is not raised. The diagnosis of rabies is quite challenging, and only 20 of 35 cases in the United States since 1980 were diagnosed while the patients were alive.2,6,7 The clinical presentation can vary greatly, ranging from fulminant encephalitis to a relatively subacute presentation that may overlap with rapidly progressive dementias such as Creutzfeld-Jakob disease.1,6 The incubation time is usually weeks to months, with a modal length of 40 days,8 but well-documented cases of very long incubation periods have been reported (for a very well-documented case with a 6-year incubation, see Grattan-Smith et al9). In most cases in the United States, there is no clear-cut exposure to a rabid animal; instead, the infection can usually be traced serologically to a viral strain carried by bats.2,6,7 To make a laboratory-confirmed diagnosis of rabies, the clinician must have a high level of suspicion because the laboratory diagnosis relies on the physician ordering specific tests.1,3 Because of this problem, some have felt that a diagnosis of rabies should be considered in all patients presenting with encephalitis of unclear cause.1,2
Our patient presented with fulminant encephalitis and early cranial nerve signs. In this case, MRI revealed the presence of signal abnormalities in the medulla and pons that correlated with the patient's cranial nerve abnormalities. In other cases, there have been no such abnormalities noted, but only a few patients with confirmed rabies have been reported to have undergone MRI.4- 6,10- 12 A greater number of patients with rabies have been reported who underwent computed tomographic scanning9,10,12 with no findings. Computed tomographic scanning, however, would be almost certain to miss abnormalities such as those reported here because of its insensitivity to posterior fossa abnormalities. Since the findings in our patient were initially somewhat subtle and only became obvious on repeated imaging, it is unclear whether the other patients with normal MRI scans would have had the same or similar abnormalities if they had had scans at a different point in the course of their disease. Additionally, as the scans were not shown in some cases,12 they could not be confirmed to be technically adequate. Only prospective analysis of future rabies cases with technically adequate contrast-enhanced MRI scans will determine if the findings seen in this patient are present in others.
Rabies is caused by a group of several closely related viruses.1 The identity of the particular serotype can be helpful in determining the source of exposure since individual types are associated with specific animal vectors. In most of the world, rabies is spread by dog bites, while in the Americas it is overwhelmingly spread by bats.1,2 Because of the ability of the rabies variants spread by bats to replicate effectively in nonneuronal cells, they may be spread easily by nonbite exposure.1,2
Interestingly, the clinical presentations of rabies from canine and bat sources are quite different.1 A characteristic presentation of rabies following dog bite is a local sensory prodrome at the site of the bite, which is thought to reflect a local ganglionitis.1 It is common for these symptoms to then spread centrally.1,3 Canine rabies presents as either a furious (encephalitic) form or a dumb (paralytic) form.1 Patients with furious rabies usually die within a week of symptom onset, while patients with the dumb form usually survive for 2 weeks.1 High fevers, hyperactivity, and autonomic dysfunction (including excess salivation), piloerection, and pupillary abnormalities usually accompany furious rabies. Dumb rabies begins with progressive weakness in the bitten limb that spreads inexorably, causing generalized craniospinal weakness.1 Encephalitic features are usually late symptoms in dumb rabies.
Bat rabies is reported to have clinical features substantially different from canine rabies.1 In patients with bat rabies, there is a high incidence of focal brainstem signs and myoclonus, as in our patient. Other patients with rabies caused by bat exposure have been described as having hemipareses or hemisensory deficits, ataxia, chorea, or Horner syndrome. In canine-bite rabies, focal weakness is described only in those with dumb rabies, and this invariably begins in the bitten limb.1
The pathophysiology of rabies infection is fascinating and quite distinct from other viral infections of the nervous system. When an individual is inoculated by dog bite, the likelihood of rabies infection depends on whether the virus makes contact with muscle. The virus replicates in muscle and infects the motor neurons that innervate the muscle, and then spreads centripetally by axonal transport to the CNS at a speed of 1 to 2 cm/d.1,3 This process is thought to explain the usual incubation time, and there is no satisfactory explanation for cases with longer incubation periods. Bat rabies replicates in skin and connective tissue, which allows spreading without direct inoculation of muscle tissue.1,3 Significant dysfunction of the hypothalamic-pituitary axis has been described in many patients with rabies1; in our patient, it was associated with radiologic and pathologic abnormalities of the pituitary stalk and of the gland itself. No viremia develops at any stage, and an immune response is not mounted until the infection reaches the level of the CNS.8
Notable brainstem abnormalities on neuroimaging in the clinical setting of rabies have not been reported previously. However, given the propensity of rabies spread by bat exposure to cause brainstem findings clinically and pathologically, we suspect that other patients would show similar abnormalities if studied at the appropriate stage of disease. Since imaging findings such as these have been described in cases of rhombencephalitis caused by other organisms or viruses,13- 15 these findings are not diagnostic of rabies. It is of interest to note that enhancement was minimal in our case. Pathologically, inflammation may be quite inconspicuous in cases of confirmed rabies,8 and if the blood-brain barrier is intact, abnormal gadolinium enhancement will not occur.
Accepted for publication July 28, 2000.
Dr Pleasure was supported by a Howard Hughes Medical Institute postdoctoral fellowship for physicians.
Corresponding author: Samuel Pleasure, MD, PhD, Department of Neurology, Room S262, 513 Parnassus Ave, University of California, San Francisco, CA 94143 (e-mail: firstname.lastname@example.org).
Pleasure SJ, Fischbein NJ. Correlation of Clinical and Neuroimaging Findings in a Case of Rabies Encephalitis. Arch Neurol. 2000;57(12):1765-1769. doi:10.1001/archneur.57.12.1765