P&S Medical Review: Apr 1995, Vol.2, No.2
H. Houston Merritt and Leonidas Stefanis, M.D. and Lewis P. Rowland, M.D.
Columbia University College of Physicians and Surgeons, New York, N.Y.
Houston Merritt was the Chair of Neurology at Columbia-Presbyterian from 1948 to 1967. He was Dean of the College of Physicians and Surgeons from 1958 to 1969. His legacy as a teacher of neurology extended throughout the United States because 30 of his st udents headed departments at medical centers from coast to coast. At that time, one-third of all Neurology chairs had trained with Merritt and some members of that cohort have not yet retired. Merritt also wrote a Textbook of Neurology;1he was the sole author of that book for five editions. As the book increased in size with the accumulation of clinical knowledge, he reluctantly accepted the aid of others. The ninth edition of the text book was published in February 1995. Merritt’s influence as teacher is still strong.
Merritt attracted these future leaders partly because of his own research achievements. He was one of the first to appreciate the need for clinical research and the opportunities it afforded. His first papers appeared after 1930. At that time, research in neurology was essentially confined to clinico-pathological correlations. Electroencephalography came soon thereafter, but Merritt took another track. Modern clinical investigation may be said to have started with those famous investigators of salt-and-wa ter metabolism – James Gamble at Hopkins, Robert Loeb and Dana Atchley at P&S, and John Peters at Yale. Merritt, working at the Boston City Hospital at the time, focused on a different body fluid, the cerebrospinal fluid (CSF). His work established the no rms and deviations we still use in the diagnostic use of CSF analysis. With Frank Freemont-Smith, Merritt wrote a monograph2on CSF, setting down the results of all of his studies. One of his students, Robert Fishman, brought the review up to da te and Fishman’s book3is now in its second edition.
Merritt’s most famous contributions were the discovery of phenytoin and the development of methods to determine the efficacy of anticonvulsant drugs in animals.4Both the drug and modifications of the method are still in wide use. Merritt’s part ner in that enterprise at Harvard was Tracy Putnam, who was also his predecessor as Chair at P&S.
Merritt made many other contributions, including research on head injury, multiple sclerosis, and stroke. He was one of the world’s authorities on neurosyphilis, which was a major public health problem before World War II. Merritt’s first work on syphilis was published after he went to the Kaiser Wilhelm Institute in Munich in 1930.5,6Combining his clinical interests, analysis of CSF and a large population of patients with syphilis at the Boston City Hospital, Merritt published 20 papers on syp hilis, culminating in his monograph7on the subject, which was written with Raymond D. Adams (neurologist and neuropathologist) and Harry Solomon, psychiatrist. That book appeared in 1946, a year of transition because penicillin became widely av ailable at just about that time, drastically altering the incidence and prevalence of neurosyphilis. With the advent of AIDS and renewed interest in neurosyphilis, it is now worthwhile to contrast the view of neurosyphilis in the 1946 monograph and curren t concepts.
The monograph also ended Merritt’s publications on neurosyphilis. Of 90 papers he wrote between 1938 and 1945, 20 were on diverse topics related to neurosyphilis, establishing many concepts that still prevail and were summarized in the monograph. From 194 5 to 1963 (the year his available bibliography ends), he published another 118 papers for a total of 209. None of the later papers included “syphilis” in the title. Instead, the titles included headache, head injury, multiple sclerosis, and Parkinson’s di sease.
Then: In the 1930s, in Merritt’s time, syphilis was a common disease. He estimated that paresis, a late form of neurosyphilis, ultimately appeared in 5% of all patients with syphilis and that 5-10% of all adults in the US had syphilis. Therefore, t he prevalence of paresis would have been about 2.5-5.0/1,000. Paresis accounted for more than 20% of all patients in mental hospitals.8Available therapy at the time was limited in efficacy but the incidence declined in the 30s for several reaso ns, including the routine screening of blood serology, federally funded treatment programs, and preventive education. Even before the introduction of penicillin after World War II, the number of patients with paresis in mental hospitals had been reduced b y half.8
With an effective treatment readily available, the incidence declined from the more than 100,000 cases in 1946, to 29,000 cases in 1950, and 6,000 in the early 50s. The rate rose again to 20,000 cases a year in 1962 and leveled off at a rate of 19,000 to 26,000 cases annually until 1978. In the 70s, there was a popular movement to discontinue routine testing for syphilis in all patients admitted to hospitals. Even before the advent of AIDS in 1981, the incidence of syphilis began to increase (figure 1), e specially among homosexual men, and then the AIDS epidemic ended the status of syphilis “as a disease of primarily historical interest.”8,9,10,11,12
Now:From 1985 to 1990 the incidence rate increased by 75%. Urban centers in the Southeast had the highest rates and, for the first time, there was a great increase in the Midwest. Among the possible causes were poverty, limited access to health car e, the “crack cocaine” epidemic and reporting bias.8,11,12,13The epidemic peaked in 1990, when 50,233 new cases of early syphilis were reported. Since then, all subgroups have shown a progressive decline in the incidence of syphilis,14with fewer than 20,000 cases projected for 1994,15perhaps a result of a fear of AIDS and safer sexual practices. The epidemics of HIV and syphilis in the ’80s ran in parallel, appearing first in homosexual men, and later, disproportionately a ffecting minority groups in the inner cities.16,17The prevalence of neurosyphilis in HIV-infected subjects ranges from 1 to 5.9%.18,19
It is not known how often those with early syphilis later develop neurosyphilis. In 1976, when the incidence of neurosyphilis was last reported separately by the CDC, neurosyphilis appeared in about 4% of all people with syphilis,20compared to the 29% reported by Merritt for the years 1932-42. Presumably, penicillin therapy of early syphilis reduced the risk of neurosyphilis.
Serological Tests for Syphilis
Then: Much of the following discussion of manifestations and treatment of neurosyphilis depends on an understanding of the tests used to make the diagnosis. Physicians then depended on the detection of antibodies to what was called “reagin”, a comp lex of cardiolipin, lecithin, and cholesterol. That was the basis of the original Wassermann complement-fixation test. A more sensitive version of the same test was the Kolmer test. These are “non-treponema” tests, however, and in Merritt’s time there wer e no specific tests for treponema antigens.
Merritt relied heavily on the CSF Wassermann test, which was regarded as highly specific because false positive reactions were restricted to people with leprosy, frambesia and trypanosomiasis, none of which were commonly encountered in Boston. He recogniz ed that false positive reactions could result from contamination of CSF by Wassermann-positive blood.
Now: Modern non-treponema tests are the Venereal Disease Research Laboratory (VDRL) test and a variant called the rapid plasma reagin (RPR) test, which has become popular because it can be carried out rapidly as a flocculation that is macroscopical ly visible on a slide. These tests are direct descendants of the Wassermann and Kolmer tests, made more sensitive by purifying the components of the nonspecific antigen (cardiolipin-lecithin-cholesterol). The RPR is more sensitive and more liable to give false positive results.
When applied to the CSF, the VDRL shows high specificity because there are few false positive reactions. These false positive reactions might occur under the following circumstances: contamination of CSF by blood, high CSF protein content, or paraproteine mia.8It is said that the CSF-VDRL may be negative in some patients with neurosyphilis but it is then difficult to prove that the clinical syndrome is actually a form of neurosyphilis, especially if the manifestations are not absolutely typical. Nevertheless, the CSF-VDRL, like so many other laboratory tests, is more meaningful when it is positive than when it is negative. A positive CSF-VDRL establishes the diagnosis of neurosyphilis. A negative test does not completely rule it out.
Treponema antibody tests are also available. The most widely used is the fluorescent treponema antibody (FTA). The antigen is made from the spirochete. Absorption methods remove nonspecific and false reactions (FTA-ABS). This is now the most specific test for syphilis when applied to blood. However, it cannot be used on CSF because even the slightest contamination of CSF by blood can give a falsely positive reaction, which may also result from diffusion of serum immunoglobulins into the CSF, resulting in a false positive rate of 1.0-4.5%.21To increase the reliability of treponema antibody tests, the ratio of serum-to-CSF content of albumin or IgG has been used to assess intrathecal production of treponema antibodies, especially the treponema pa llidum hemagglutination assay (TPHA) index.22,23,24,25,26However, even with this test, it is difficult to be confident of the diagnosis of neurosyphilis if all other measures of the CSF are normal. Although the TPHA index may revert to normal a fter treatment,26the specific treponema antibody tests for CSF seem to be more reliable in excluding neurosyphilis than in confirming the diagnosis.
HIV co-infection might alter the serological response to the treponema, and neurosyphilis might develop in a person with negative serology in serum, CSF, or both.27However, studies of HIV-infected and noninfected subjects have shown no signific ant differences in the serological response to syphilis or only slightly higher titers in the HIV-infected group.28,29,30These slight differences, which have no apparent clinical significance, are attributed to the generalized increase of the i mmunoglobulin response in early HIV infection or to heavier spirochetemia, as a result of the subdued cellular immunity.8,29Nevertheless, in advanced stages of AIDS, the serological response might be attenuated.8,31HIV infection can cause a false-positive RPR.8
For these reasons, the VDRL is the preferred test for CSF diagnosis. The argument about limitations of the CSF-VDRL arises partly from modern technology because there is yet another test for treponema antigens, the polymerase chain reaction (PCR), which m ay be positive even if the CSF is completely normal in other respects, including VDRL, cell count, protein, g-globulin, and sugar. It has been claimed that the sensitivity of the test is about 50% and the specificity above 90%, with some false positive re sults.32,33However, in a study of ten HIV-positive patients with definite neurosyphilis, only three tested positive by PCR.34
Isolation of the Treponema
Then: Even in Merritt’s time, it was possible to isolate treponema from the CSF and other body fluids or tissues by inoculation of the material into the testicles of rabbits. This rabbit infectivity test (RIT) was available then but it was mentione d only in passing in his monograph, probably because it was cumbersome, time-consuming and expensive, factors that precluded its general application to populations. In 1924, Chesney and Kemp35found positive CSF-RIT in 15% of subjects with early syphilis and otherwise normal CSF studies.
Now: Because the CSF-RIT may be positive when the CSF-VDRL and other CSF characteristics are normal,35there is lingering uncertainty, especially when presumably adequate penicillin therapy has already been administered.8,28,36,37,38 It is still uncertain whether these patients are at increased risk of developing symptomatic neurosyphilis, which is implied by the presence of viable pathogenic treponema. The sensitivity of the test might be lower in late syphilis than in early sy philis.
Clinical Definition of Neurosyphilis
Then: In the monograph, Merritt did not state a formal definition of cases. He apparently accepted a case if there had been a history of syphilis and the neurological syndrome was compatible with concepts, or if the clinical syndrome was compatible and there was serological evidence of syphilis.
Doubts about the diagnosis of cases in the classic monograph have arisen. For instance, Merritt et al reported negative CSF Wassermann tests in 14% of those with meningitis, 19% of those with cerebrovascular syphilis, and 20% of those with tabes do rsalis. Without autopsy, how could it be proven that those seronegative cases were really due to syphilis? The CSF-VDRL test now in use is about 24 times more sensitive than the old Wassermann test;39as a result, most neurologists reject the ar gument that the CSF-VDRL may be non-reactive in established neurosyphilis. However, there has been at least one case of pathologically-proven paresis although the CSF-VDRL was repeatedly negative.40
Now:Most students agree with the criteria of Gordon et al34for the definite diagnosis of neurosyphilis: reactive treponema blood test (FTA-ABS), if symptomatic, a compatible clinical syndrome, and positive CSF-VDRL. All three must be present.
Clinical Varieties of Neurosyphilis
Then: Merritt and his colleagues collected cases of neurosyphilis primarily from the Boston City Hospital during the years 1932-42. The distribution of the several types of neurosyphilis is listed in Table 1, comparing figures from the monograph to those of the antibiotic and HIV-era.30,41
This was regarded as a rare form of neurosyphilis. Merritt et al found only 80 acceptable cases between 1920 and 1935 in three hospitals and only 37 cases among 2,263 people with syphilis. Their analysis of the syndrome was based on the 80 cases. T hey divided the series into three forms: syphilitic hydrocephalus, 26 cases, with a syndrome of acutely increased CSF pressure and papilledema; syphilitic meningitis of the vertex with seizures and focal cerebral symptoms, 20 cases; and syphilitic meningi tis of the base, 34 cases that were dominated by cranial nerve abnormalities.
Of the 80 patients, 57 had a chancre or rash. In 21 patients, meningitis was the first manifestation of syphilis. In 38, the interval between primary or secondary and neural lesions was less than one year. When the interval was longer, Merritt thought the patient had more chance to forget the early manifestations. Twenty-nine of the patients had not received treatment previously. Among the others, symptoms appeared in 48 patients one to seven months after completing treatment. The blood Wassermann was neg ative in 26 of the 80 patients (36%) and was not recorded in six others. However, 86% of patients had a positive CSF Wassermann, leaving 11 of 80 patients studied who had negative CSF serology. Presumably, those with negative serology had had a documented primary lesion. Merritt noted that many of the seronegative patients had been studied before the more sensitive tests became available. Also, four of the 26 patients with an initially negative CSF-Wassermann Test later had a positive CSF reaction when th ey were re-studied.
Rather than having luetic meningitis, some of those seronegative patients or even those with positive serology could have had viral infections that were not yet recognized as causes of meningeal infection. That possibility became a major problem years lat er, because HIV itself may cause meningitis and many patients are infected by both HIV and Treponema pallidum.
Nevertheless, the syndrome responded well to therapy. All but 2 of the 80 patients received intravenous arsenical treatment. Only three were known to have developed parenchymatous neurosyphilis.
Nine and one half percent of the Boston City Hospital patients (N=2,263) with syphilis had asymptomatic neurosyphilis. None had received adequate treatment by then-current standards; 82% were untreated and 18% inadequately treated. Diagnosis depended enti rely on serology and 98% had a positive blood Hinton test, a variant of the Wasserman Test. In the others the blood was negative but the CSF serology was positive.
Stroke is so common in the general population that Merritt et al had a problem in selecting cases for the diagnosis of luetic cerebrovascular disease. They reviewed 250 cases to exclude those with hypertensive or atherosclerotic changes, cerebral e mbolism, other non-syphilitic diseases, and paretic and acute meningeal syphilis. They excluded patients older than age 60. They were left with 42 cases for analysis and the syndromes affected virtually all major arterial territories of the cerebral hemis pheres, brain stem and cerebellum. They made the diagnosis when there were “symptoms and signs of a vascular lesion in a patient with a definite history or other evidence of a syphilitic infection.” All of their patients had a positive blood or CSF serolo gy; the blood was positive in 95% of cases and the CSF Wassermann in 81%. The authors noted that “probably a non-syphilitic etiology would have been found in most of the cases with negative CSF serology”, a view reiterated by Simon9and Fishman. 3Merritt explicitly noted that a person could have both a history of syphilis and a stroke due to atherosclerosis. Sixty six percent of their cases showed CSF pleocytosis, more likely in syphilis but not restricted to syphilis. They did not kno w about carotid vertebral artery disease as a cause of stroke, and they were unaware of other sources of cerebral emboli, so it is difficult to know how many of those 42 patients actually did have syphilitic arteritis.
Parenchymatous Neurosyphilis, Paresis and Tabes
Paretic neurosyphilis was a major cause of psychosis, dementia and corticospinal tract disease. It was attributed to the direct effects of the spirochete on the cerebral cortex. The diagnosis depended on the clinical manifestations and an abnormal CSF. Me rritt et al stated that “when the spinal fluid is normal, such a diagnosis is untenable.” The CSF changes included pleocytosis, increased protein content, an abnormal (but nonspecific) colloidal gold curve, and strongly positive Wassermann reaction s. The syndrome resisted arsenical therapy, and fever therapy (induced by injection of malarial organisms or by placing the patient in a hot box) was used in preparation for chemotherapy. The spinal fluid abnormalities might persist even if the patient im proved clinically but almost half the patients improved inadequately and remained in psychiatric hospitals.
Tabes dorsalis was a second form of parenchymatous neurosyphilis, causing a spinal cord syndrome in which degeneration of the posterior nerve roots and the posterior columns led to sensory ataxia due to loss of position sensation, with lancinating pains a nd urinary incontinence.
In the pre-penicillin era, paresis and tabes dorsalis were the most common forms of symptomatic neurosyphilis (Table 1). Hook and others believe that the organism gained entrance to the CNS early in the infection and, if treatment were inadequate, the spi rochetes could then multiply in a sanctuary protected from the host’s immune response. That could explain the prominence of parenchymatous neurosyphilis in those days.
Now: The prevalence of AIDS has made the diagnosis of syphilis more complicated but with or without AIDS, the relative frequency of the several clinical syndromes of neurosyphilis has changed from the pre-penicillin days (Table 1). Parenchymatous paretic or tabetic neurosyphilis is rare now. Gummas were always exceptional and still are. Modern neurosyphilis is virtually restricted to the asymptomatic form,23,42 meningitis,30cerebrovascular disease,43,44or gumma.45,46There is no evidence that “atypical” forms of neurosyphilis are more prevalent.9,21
Co-infection with HIV might alter the course of syphilis. Merritt et al recognized the importance of the host immune response in defining the clinical syndrome and the need for both therapy and the immune response to work together in eradicating the infection. In an immunodeficient host, neurosyphilis might appear earlier and in more severe form, especially in immunologic sanctuaries such as the eyes or CNS.47In experimental animals, cortisone is used to reactivate a syphilitic infection, presumably because of immunosuppression.38
Nevertheless, it is uncertain whether HIV infection affects the nature of concomitant syphilis.8,9,48There have been few prospective studies of HIV-positive and negative subjects with early syphilis and adequate follow-up to evaluate later neur osyphilis. In one small study28that lacked long-term follow-up, there were no clinical or serologic differences in HIV-positive and negative subjects, and no difference in the rate of isolation of treponema from CSF. In another approach to the problem,49the stage of syphilis was not related to HIV serology and there were no unusual or fulminant forms of syphilis among those who were HIV-positive; both groups responded to treatment of syphilis. In two studies from the same group of in vestigators,29,50the results were contradictory: in one study, HIV-infected patients were more likely to develop secondary syphilis, but this was not true in the second study. In a retrospective analysis of patients with neurosyphilis,30those who were HIV-positive more often had meningitis or ocular syphilis (Table 1).
Then: The history of therapy for syphilis was described in detail by Merritt et al. From the sixteenth century until 1907, syphilis was treated with mercury, although the mode of action was not known. The metal is treponemicidal in vitro< /I> but loses that property when given to animals. Mercury salves were effective in treating cutaneous syphilis but CNS syphilis did not respond to any form of mercury. In 1907, Ehrlich introduced salvarsan (arsphenamine), the first “magic bullet” that co uld kill an organism without destroying the host. It was also the beginning of true chemotherapy. Nevertheless, it was still ineffective in paretic neurosyphilis and only partially effective in tabes, presumably because it did not pass the blood-brain bar rier.
In 1921, bismuth therapy was introduced and was rapidly accepted because of greater efficacy and lower toxicity. Then, other forms of arsenical therapy were devised, some given by injection into the CSF in attempts to by-pass the blood-brain barrier. A ne w approach was that of Wagner-Jauregg who was impressed by a report that the remission rate of paresis was unusually high in a mental hospital where there had been epidemics of febrile diseases. He therefore sought to induce fever as a treatment for pares is. There seemed to be a relatively low incidence of neurosyphilis in areas where malaria was endemic. In 1917, Wagner-Jauregg reported improvement in four of nine patients with paresis after malaria-induced fever. Later studies found improvement in about 70% of cases, with marked improvement in 20-40%. Then attempts were made to find safer and less traumatic ways to induce fever, including hot boxes that could be adjusted to control body temperature.
By the time the Merritt monograph was written, acute syphilitic meningitis was being treated with a combination of an arsenical drug and bismuth. The patient was followed for two years of continuous treatment of weekly injections. If, at the end of that t ime, the CSF was still abnormal, the risk of paresis was deemed high and the patient was given fever therapy. Among 27 patients who were treated for at least 18 months, 21 were clinically and serologically normal. Two developed paresis and four had persis tent CSF abnormalities. Six patients received substandard therapy and were followed for two years. One became demented and one had tabes; only one was clinically and serologically normal. In another group of 61 patients, they found clinical recovery in 83 % of appropriately treated patients, compared to 35% of 31 patients who received poor treatment.
For asymptomatic neurosyphilis, 80-96% of adequately treated patients remained asymptomatic, with improvement or reversal of serology in 75-85%. Of those with a severe CSF abnormality, 20% developed paresis or tabes. Like other forms of stroke, those due to meningovascular syphilis also improved, but the CSF remained abnormal in 10-15% despite therapy; these patients were later given fever therapy. If the CSF became normal after two years, they attributed any progression of symptoms to some non-syphilitic cause.
For paresis, fever therapy resulted in improvement of 30% and arrest of progression in most of the others. Although the CSF cell count and protein content return to normal in about a year, the CSF serology often remained positive for a longer time. The CS F also improved after arsenical or fever treatment of tabes but symptoms persisted.
Examination of the CSF played a major role in the diagnosis and treatment of all forms of neurosyphilis. Among the principles established in the pre-antibiotic era were the following: 1) CSF gamma globulin content could not be measured directly but later studies showed that the “colloidal gold” test was really a measure of IgG content. 2) Treponema tests on CSF included the Wassermann, Kahn and Davies-Hinton tests. The transition to modern tests has been described above, but the value and limitations of C SF serology in following treatment were established early. 3) Experience documented the value of adequate treatment of early syphilis in preventing the later appearance of neurosyphilis. 4) The CSF abnormalities improved with clinical improvement, especia lly in meningeal and vascular neurosyphilis, but the response in paresis and tabes was slower or nonexistent. In all forms, the pleocytosis and protein content decreased more rapidly than the Wassermann and colloidal gold reactions. 5) Persistent pleocyto sis became an indication for re-treatment. All of these principles carried over to modern treatment.
Now: Penicillin was introduced during World War II, but was not available to civilians until 1945. It was rapidly found to be the most effective treatment of all forms of syphilis. The regimen recommended by the CDC and WHO for early syphilis is n ow intramuscular administration of benzathine penicillin G, 2.4 million units in one dose. For latent syphilis, the same regimen is given three times in three weeks. For neurosyphilis, asymptomatic or symptomatic, aqueous crystalline penicillin G is given intravenously, 2.4 million units every four hours for 10-14 days.8,51,52
Until 1986, the same regimen was recommended for both latent syphilis and neurosyphilis. The guidelines were modified because of treatment failures and isolation of treponema from the CSF after the “standard” regimen of benzathine penicillin; also, benzat hine penicillin fails to achieve treponemicidal levels in the CSF, despite adequate blood levels.53In contrast, intravenous administration of penicillin G achieves adequate CSF levels.54As an alternative therapy for neurosyphilis, th e WHO recommends intramuscular administration of procaine penicillin, 1.2 million units every day, in combination with oral probenecid, 500 mg, four times a day for 10-14 days.51This regimen almost always gives treponemicidal levels in the CSF. 55
Penicillin therapy has been effective but treatment failures were known from the start. Willcox stated in 1962 that “the very rarity (of cases of tertiary syphilis following penicillin therapy) makes them worthy of record,” but even then there were some s uch reported cases56and others came later.57,58,59In the pre-AIDS era, treatment failure and a need to re-treat, as defined by clinical and serological criteria, was seen in about 5% of patients with early syphilis, with reports rang ing from 1 to 14%, depending on the definition of “cure.”60,61,62,63Among 765 patients with asymptomatic neurosyphilis who had been treated with various regimens of penicillin, 3.31% developed symptomatic neurosyphilis within seven years. Twent y percent of the 765 patients were re-treated, most often because of persistent CSF pleocytosis.63However, those regimens did not include IV penicillin, the treatment of choice today.
Late parenchymatous neurosyphilis, especially general paresis, responds poorly to penicillin.64,65,66Tabes is now rare, but that syndrome in one patient66progressed clinically, with persistent CSF pleocytosis despite multiple treatme nt regimens, including IV penicillin. Meningovascular syphilis responds well to treatment. In some cases of late syphilis, disease activity may be related to the formation of scar tissue, so that any further treatment may be useless.21,67
Among people co-infected with HIV, there have been numerous reports of treatment failure and manifestations of neurosyphilis despite “adequate” treatment of early syphilis.46,68,69,70,71Also, neurosyphilis may progress despite recommended treat ment.34,72Some investigators therefore believe that subjects co-infected with HIV do not respond as well as immunocompetent subjects to standard regimens and that more aggressive treatment is indicated.47
However, these uncontrolled reports should be viewed with caution for several reasons: 1) Treatment failure is also seen in immunocompetent subjects. 2) Some of the reported cases may be due to reinfection rather than relapse. 3) The definition of “cure” of neurosyphilis frequently includes a fall in the CSF pleocytosis after treatment,21,61,67but HIV-positive people may show CSF pleocytosis even without syphilis.34,61
Lukehart et al28isolated treponema from the CSF after treatment of three HIV-infected subjects but not from any who were HIV-negative. However, treponema pallidum has also been isolated after “optimal” treatment38of immunocomp etent hosts. In one retrospective study of the serological response to treatment of early syphilis, there was a slower reduction of titers in the HIV-infected subjects, but it was unclear whether there was persistently active disease.8,73In ano ther retrospective study,49IV drug users at various stages of syphilis received different treatment regimens and there was no difference in clinical or serological responses in those who were HIV-positive or negative.49
Musher48was pessimistic about treating patients with both syphilis and HIV-infection, citing failure rates of 20-30% in patients treated with either IV ceftriaxone or IM benzathine penicillin.72He also cited progression of neurosyphi lis despite treatment with IV penicillin.34Musher advocated a return to the old regimen of IM benzathine penicillin for the treatment of neurosyphilis in HIV-infected people because the more intensive regimens seem no more effective. Even in im munocompetent subjects, no controlled comparative studies have been done to show that the more intensive IV regimens are more effective than the IM benzathine penicillin regimen for the treatment of neurosyphilis. Manu and Varade74advocated tre atment of asymptomatic neurosyphilis with daily IM procaine penicillin: even with a failure rate of 10%, that would be overall more cost-effective than in-hospital IV penicillin therapy.
Regardless of HIV status, patients with asymptomatic neurosyphilis can be treated as out-patients by using a WHO-recommended regimen of IM procaine penicillin G and probenecid,61,75provided that close follow-up can be assured. Patients with sym ptomatic neurosyphilis should receive IV penicillin. The reportedly different responses of syphilis in HIV-coinfected patients are anecdotal and unsubstantiated; the CDC does not recommend special treatment for those with both syphilis and HIV-infection.< SUP>51,52However, because of the persistent uncertainty, these patients should be followed closely. Further study is needed to determine whether patients need different treatment because of simultaneous infection with both HIV and syphilis.
This review follows publication of Merritt’s monograph on syphilis by 50 years. The interim was marked by more effective treatment in the form of penicillin and the advent of AIDS. Advances in our knowledge of the different causes of stroke and viral meni ngitis make it now seem likely that the monograph overestimated the prevalence of these syndromes as manifestations of neurosyphilis. Nevertheless, Merritt’s contribution has lasted. Today, his classification of neurosyphilis is still the basis for modern views, even though paresis and tabes have all but disappeared. His observations of the role of CSF examination in diagnosis and evaluation of therapy have also persisted. This lasting impact is testimony to the value of clinical scholarship and to Merritt’s personal approach.
The monograph had three authors and both of Merritt’s companions also had distinguished careers, Adams in clinical neurology and neuropathology, and Solomon in psychiatry. Their friendships lasted and Adams played an important role in Merritt’s terminal i llness, which was thought to be normal pressure hydrocephalus, a disorder first described by Adams and Hakim. Merritt, the pioneer of CSF studies, had been skeptical about the diagnosis (it is still difficult to identify which patients will improve by shu nting of CSF). Then, ironically, Merritt himself developed the triad of symptoms that characterize normal pressure hydrocephalus: gait disorder, sphincter problems, and early cognitive loss. Merritt was being treated by too many of his own neurological st udents in New York and they disagreed about the proper management. Merritt was finally referred as a patient to Adams and Miller Fisher, who were both immensely compassionate and kind. The shunting procedure was, however, a disaster, and Merritt never awo ke. Before the operation he had stated that he preferred death to the increasing restrictions of his illness; his wish was inadvertently granted.
From the Department of Neurology, Neurological Institute, Columbia-Presbyterian Medical Center, New York, NY.