Spirochetal Infections of the Nervous System

Background

Syphilis, caused by Treponema pallidum , became prominent in Europe shortly after the return of Columbus from the New World. Known at the time (in some circles) as “the French pox,” the mode of transmission was recognized early on. The Breviary of Health of 1547 states clearly that “specially it is taken when one pocky person doth synne in lechery the one with another.” At that time, as with other newly introduced infectious diseases, the morbidity and mortality associated with syphilis were high. Then, presumably as the most susceptible hosts died and the bacteria self-selected for strains that could persist without killing or incapacitating their hosts, the bacteria and humankind settled into their current symbiotic relationship. Currently, there are about 31,000 cases of primary and secondary syphilis each year in the United States, similar to the number of cases of Lyme disease confirmed by the Centers for Disease Control and Prevention (CDC).

Diagnosis

Diagnosis was originally clinical and rested on recognition of the characteristic chancre—a painless, ulcerated, indurated lesion at the site of primary infection that typically heals spontaneously. The lesion contains numerous spirochetes; one of the best methods of diagnosis is to scrape the lesion, place the scrapings in saline on a slide, and view the motile spirochetes with a dark-field microscope. In patients not seen at the time of the chancre, diagnosis initially rested on the recognition of one of the large number of clinical syndromes attributed to this disease.

More accurate diagnosis of later disease required advancements in laboratory techniques. Identification of the causative organism was an important first step, but diagnosis by culturing T. pallidum remains essentially impossible. The organism can be propagated by inoculation into susceptible animals but not cultured in vitro—hence the microbiologic approach traditionally used in clinical diagnosis of other bacterial infections is not useful.

The spirochete has now been very well characterized. Its length varies from 6 to 20 µm; its diameter is about 0.18 µm. It is shaped in a regular helix and exhibits corkscrew motility, propelled by its flagella. Its entire genome has now been sequenced. Despite this intimate knowledge of the spirochete, laboratory confirmation of infection, particularly after resolution of the chancre, remains dependent on serologic tests. These fall into one of two groups. Historically, screening has used nontreponemal tests, measuring “reaginic” antibodies—in essence, anticardiolipin antibodies. The reason that patients with this infection have high titers of anticardiolipin antibody remains unknown, but presumably reflects an interaction of the spirochete with host lipoproteins and other lipids. Tests such as the Wasserman reaction, rapid plasma reagin (RPR), Venereal Disease Research Laboratory (VDRL), and Hinton tests all fall into this group. These tests have been thought to have very high sensitivity but, as indirect measures of infection, lack specificity. Patients with positive nontreponemal tests would then be further tested for specific antitreponemal antibodies, using the FTA-Abs (fluorescent treponemal antibody, absorbed against nonpathogenic treponemata to decrease cross-reactivity), the MHA-TP test (microhemagglutination for T. pallidum ), or capture enzyme-linked immunosorbent (ELISA) or chemoluminescent assays (CIA).

Studies using CIAs or ELISAs for comparison suggest that reaginic tests are less sensitive than previously believed, leading to recommendations by some that direct tests be used for initial screening. Although the CDC still recommends the traditional sequence, it now also endorses the reverse sequence. In this approach the initial diagnosis starts with a specific treponemal test (ELISA or CIA), confirming positives with a nontreponemal test; discordant results are then further tested with a different specific treponemal test. To make matters even more confusing the European equivalent to the CDC recommends a different reverse approach, also starting with a specific treponemal test, but then confirming positives with a different specific test, testing discordant results with a nontreponemal test. Arguments in favor of each approach are based on assessments of positive and negative predictive values. Notably, a large study suggests that overall, both reverse approaches are equivalent, although increasing evidence suggests the traditional method, starting with nontreponemal tests, is less sensitive.

Response to therapy is generally assessed by measuring titers of reaginic antibody. Typically, the serum titer in a VDRL or similar test declines steadily after successful treatment and eventually becomes negative in most cases, unlike specific assays, which remain positive indefinitely.

When serologic testing was introduced early in the twentieth century, seroprevalence rates of 8 to 14 percent were reported in major cities. It was at that time that the medical literature blossomed, attributing all manner of disease to syphilis. It may well be that at least some of this was misattribution due to coincidentally positive serologic tests, much as occurs in Lyme disease–endemic areas today (where, coincidentally, seroprevalence rates of 5 to 15% are common).

Diagnosis of nervous system disease relies heavily on examination of the cerebrospinal fluid (CSF), which is recommended in the assessment of any patient with primary or secondary syphilis and clinical suspicion of central nervous system (CNS) involvement or with tertiary syphilis. Invasion and infection of the CNS typically result in increased protein concentration or a CSF pleocytosis, with an increased proportion of B cells. Rarely there is mild hypoglycorrhachia. Active disease typically results in the presence of reaginic antibodies (VDRL) in the CSF. Chronic infection and stimulation of the immune response within the CNS generally result in increased CSF IgG concentration as well as production of oligoclonal bands. Synthesis of specific anti- T. pallidum antibody can often be demonstrated in the CSF. Interestingly, active infection leads to increased CSF concentration of CXCL13, a B cell–attracting cytokine, just as in nervous system Lyme disease. Successful treatment usually leads to resolution of the CSF pleocytosis, with normalization of CSF protein, IgG concentration, CXCL13, and, usually, VDRL.

CSF-based diagnosis is usually straightforward but can sometimes be problematic, as reviewed elsewhere. The CSF VDRL is considered highly specific—a positive CSF VDRL result is regarded as diagnostic of CNS syphilis. (The CSF RPR test has far more false-positive results.) The VDRL’s specificity is probably related to the fact that the methodology is rather insensitive—a fairly high concentration of antibody is needed for the test to be judged positive. Thus, the presence of a positive VDRL in the CSF is unlikely to be an artifact of contamination of CSF with peripheral blood immunoglobulin. One potential diagnostic problem arises, however, in that although the CSF VDRL typically declines after successful treatment, it may remain positive for an extended period of time. In this circumstance, the return of the CSF white blood cell count and protein concentration to normal is usually taken as evidence of adequate treatment.

With the technically more sensitive methods used to detect antitreponemal antibodies (FTA-Abs, MHA-TP, ELISA, CIA), contamination is a more important issue. False-positive results can occur for one of at least three reasons. Antibodies against antigenically similar organisms (e.g., Borrelia burgdorferi ) can cross-react. More important, if the lumbar puncture is traumatic, peripheral blood can contaminate the specimen, resulting in artifactual elevations in antibody titers. Probably most important, however, is the fact that some peripheral blood immunoglobulin always filters into the CSF. In the absence of CNS inflammation, there is minimal if any intrinsic production of antibodies within the CNS. However, CSF IgG concentration is typically 2 to 3 mg/dL, approximately 0.2 percent of peripheral blood IgG concentration, representing the small amount of serum immunoglobulin that normally passes through the blood–brain barrier. Just as in testing CSF Lyme serologies, standard testing conditions (diluting CSF 1:5 for the FTA-Abs, for example) are selected to compensate for this leakage. However, artifacts may arise in two important circumstances. In patients with high serum FTA-Abs titers, CSF titers may be positive simply by virtue of the small amount of specific antibody that normally filters in. Second, in the presence either of CNS inflammation or of blood–brain barrier breakdown for other reasons, the amount of peripheral blood immunoglobulin leaking into the CNS may be greater than assumed, resulting in an apparent increase in the concentration of the tested antibody but actually simply reflecting an overall and nonspecific increase in all antibody levels. These considerations notwithstanding, the CSF FTA is a highly sensitive if not entirely specific marker of neurosyphilis.

The absence of an absolute “gold standard” test for the diagnosis of neurosyphilis has resulted in controversy, differing recommendations for diagnostic strategies, and varying estimates of the sensitivity and specificity of different CSF tests. Whereas the CSF VDRL test is often said to be 100 percent specific, estimates of its sensitivity vary widely, depending on the alternative criteria used to define neurosyphilis. Screening CSF samples using the FTA-Abs and considering patients to have probable neurosyphilis if they have (1) a positive CSF FTA-Abs result, (2) a consistent clinical syndrome, and (3) elevation in either the CSF cell count, protein concentration, or both, leads to an estimated sensitivity of the CSF VDRL of 27 percent. Performing a similar analysis, but using the CSF MHA-TP as the primary criterion and comparing the CSF MHA-TP, CSF FTA-Abs, and CSF VDRL, suggests that the sensitivity of the CSF VDRL is about 40 percent, but that of the CSF MHA-TP is substantially greater than that of the FTA-Abs. However, there is evidence that as many as one-third of human immunodeficiency virus (HIV)–seronegative individuals with positive serum VDRL results will have inflammatory changes in the CSF, but no serologic (CSF VDRL or FTA-Abs) evidence of neurosyphilis, calling into doubt the assumptions underlying the first two conclusions.

Theoretically, a more definitive approach would be to determine whether antibodies targeted against T. pallidum are actually being produced within the CNS. Although a variety of techniques can be used, conceptually this is accomplished by measuring the antibody concentration in CSF and serum, normalizing for the relative total immunoglobulin concentrations in both, and determining whether specific antibody is disproportionately represented in the CSF. This approach, used in many other infections, should have very high specificity, although, again, sensitivity is difficult to define. Despite the intuitive appeal of this approach, sensitivity estimates have varied widely, depending on underlying assumptions, making its applicability uncertain at this time.

Clinical Features

After initial inoculation of spirochetes, the organism disseminates rapidly. Within weeks of infection the chancre develops at the site of infection. Chancres contain large numbers of spirochetes and are highly infectious. Untreated, they typically subside spontaneously over weeks. This early, localized disease constitutes primary syphilis. Over the ensuing few months, spirochetes disseminate, setting the stage for secondary syphilis. In this phase, there is a high antigen load, often with a significant spirochetemia. Spirochetes can invade any organ of the body but have a particular predilection for the skin (with secondary lesions, particularly involving the palms and soles), lymph nodes (particularly the epitrochlear nodes), kidney (with an immune complex–mediated glomerulonephritis), and CNS (seeded in between 25 and 40% of infected individuals). As many as 40 percent of untreated patients develop a lymphocytic meningitis. As in other basilar meningitides, the cranial nerves, particularly II to VIII, can be compromised. Intracranial pressure may be increased.

The disease then typically enters a latent phase, during which infection—and the immune response to it—persist. In the majority, the host immune response clears the infection. In a subset, late or tertiary complications develop. In some, this consists of disseminated granulomata, known as gummas, that can occur anywhere in the body, causing focal symptoms depending on location. Patients may develop gummas in the brain, causing local structural damage and symptoms. Gummas tend to be relatively benign but can be confused with neoplasms or other processes. Some patients develop an endarteritis obliterans affecting the vasa vasorum of the ascending aorta, causing characteristic aneurysmal dilatation.

Although neurologic involvement is generally divided into early or late (secondary or tertiary) manifestations, there is some variability in how different pathophysiologic entities are grouped. The acute meningitis, with or without cranial nerve involvement, typically occurs during acute spirochetal dissemination and is considered part of secondary (acute disseminated) disease. All other manifestations occur later in the disease, in what might be termed late disseminated infection (“tertiary syphilis”), although some tend to occur earlier in this “late” phase, others later.

Symptomatic nervous system involvement ultimately occurs in 4 to 6 percent of untreated patients and tends to take one of several forms. One of the earlier occurring late manifestations is meningovascular syphilis, in which the characteristic obliterative endarteritis involves the intracranial vasculature, causing strokes. Although the resultant clinical phenomena are described as protean, they are no more protean than in other cerebrovascular disease—differences reflect the site of damage, not the mechanism.

Several other characteristic abnormalities may occur. Those not due to vasculitis are collectively referred to as parenchymal neurosyphilis. Involvement of the upper brainstem and surrounding basilar cisterns may lead to Argyll Robertson pupils, small irregular pupils that do not react to light but do to accommodation. Optic atrophy may develop slowly, presumably owing to both increased intracranial pressure caused by the meningitis and local infiltration of the optic nerves. Tabes dorsalis , consisting of damage to the dorsal roots, the dorsal root ganglia, and the posterior columns, results in marked loss of proprioception, as well as “lightning-like” shooting radicular pain. Gait is typical of that in individuals with proprioceptive loss—slapping, with poor awareness of the location of the feet in space. Loss of the dorsal root ganglion neurons and dorsal roots leads to loss of reflexes and diminished pain awareness, potentially leading to repeated injuries to lower extremity joints with severe cumulative damage (Charcot joints).

One of the most fascinating disorders associated with late neurosyphilis is the late dementia, referred to as general paresis of the insane , or GPI. Initial symptoms are described as personality changes with delusions and hallucinations; seizures and myoclonus may occur. Involvement tends to be frontotemporoparietal, with a slowly progressive dementia. Considered a consequence of long-standing meningitis, the pathology is rather subtle. Inflammatory changes are seen in the ventricular walls (granular ependymitis). There is brain atrophy, ventricular enlargement, and neuronal loss, with some reactive gliosis. Parenchymal inflammation is usually rather limited, without evidence of multiple infarcts—in brief, the degree of CNS dysfunction seems out of proportion to the evidence of disease activity. MRI studies have demonstrated fairly widespread abnormalities.

Since the early 1980s, the HIV epidemic has led to a resurgence of neurosyphilis, raising new concerns about both the biology of the infection and its treatment. In HIV-infected individuals, early disease and asymptomatic disease each account for about one-third of the cases of co-existing neurosyphilis. Of patients with early disease, one-half have meningitis and one-fourth have meningovascular syphilis. Of those with late-stage disease, the majority has general paresis, and a small subset has tabes.

Finally, congenital infection still occurs. This happens most commonly during early stages of maternal infection, when spirochete dissemination is still actively occurring. Treatment during the first 4 months of gestation usually protects the fetus; after that, death of the fetus or newborn, or congenital infection, can result. The common manifestations of newborn infections include “snuffles” (rhinitis due to mucocutaneous involvement) and a diffuse rash that is unusual in that it involves the palms and soles as well as skin at mucoepithelial junctions. Involvement of the liver is common, as are osteochondritis and perichondritis, most typically manifested as a “saddle nose” and “saber shins.” Unusual, centrally notched, peg-shaped upper teeth (“Hutchinson’s incisors”) occur, as does frontal bossing. IgM assays can be useful for diagnosis.