INTRODUCTION

Top Previous Next References

"He who knows syphilis, knows medicine"

Sir William Osler

The origins of syphilis have been discussed for many centuries (68, 251, 266, 341). Two main theories have been proposedthe New World or Columbian theory and the Old World or pre-Columbian theory. The former holds that syphilis was endemic in the part of the world now known as Haiti and was then acquired and carried to Europe by Columbus in the 1400s. The pre-Columbian theory purports that syphilis originated in central Africa and was introduced to Europe prior to the voyage by Columbus. A third theory, the Unitarian theory, could be made to fit the pre-Columbian theory (251). This theory proposed that syphilis and the nonvenereal treponematoses were all manifestations of the same infection, with the observed clinical differences being due mainly to environmental factors, especially temperature. However, recent bacteriological work has demonstrated genetic differences between these organisms (44). Regardless of the origins, however, it remains clear that by 1495 a widespread syphilis epidemic had spread throughout Europe (251). From there the disease spread to India in 1498 and China in 1505 (251).

Early names for syphilis included the Great Pox, lues venereum (venereal disease), morbus gallicus (French disease), and the Italian, Spanish, German, or Polish disease, but the name that was to become part of the everyday language was syphilis (251). Hieronymus Fracastorius in 1530 is believed to be the first to coin the term "syphilis," derived from a mythical shepherd, Syphilus, described in his poem Syphilis Sive Morbus Gallicus, which means "Syphilis or the French Disease" (251).

Many early investigators contributed to our knowledge of syphilis, and the following highlights but a few. John Hunter (1728 to 1793) believed the Unity or Monist Theory, which holds that syphilis and gonorrhea were the same disease (68, 251). This theory was supported by his well-known experiment of 1767 in which he inoculated matter from a patient whom he believed to have gonorrhea onto the prepuce and glans of a recipient, who traditionally is believed to be himself. However, since the experiment was reported in the third person, it is now believed that the recipient was someone else (67). Ten days after the inoculation, a chancre appeared, followed by signs of secondary syphilis (251). It is now believed that the donor had both syphilis and gonorrhea, but Hunter was convinced that he had induced syphilis by inoculation of gonorrheal pus. It was not until 1838 that Philippe Ricord demonstrated conclusively that syphilis and gonorrhea were separate diseases based on over 2,500 human inoculations (251, 341). Ricord was also the first to propose a scheme for the categorization of syphilis into primary, secondary, and tertiary stages, which is still used today (251). Giovanni Lancisi (1654 to 1720) and Herman Boerhaave (1668 to 1738) implicated syphilis as a cause of cardiovascular disease. Alfred Fournier (1832 to 1914) confirmed the syphilitic origins of neurosyphilis (294). Paul Diday (1812 to 1894) and Jonathan Hutchinson (1828 to 1913) contributed greatly to our knowledge of congenital infections (251).

In 1905, the association of Treponema pallidum with syphilis was described by Schaudinn and Hoffman, who demonstrated spirochetes in Giemsa-stained smears of fluid from secondary syphilitic lesions (295). August von Wassermann devised a serum reaction test for syphilis in 1906, and serologic tests for syphilis were born (349).

Treatments for syphilis included mercury, organic arsenical compounds, and bismuth until the advent of penicillin (294). In 1943, Mahoney et al. successfully treated the first four cases of syphilis with penicillin, and more than half a century later penicillin remains the drug of choice (204).

Although given various names following its discovery, the causative organism of syphilis was finally named Treponema because of its resemblance to a twisted thread and pallidum because of its pale color (266).

T. pallidum is a member of the order Spirochaetales, family Spirochaetaceae, and genus Treponema, which includes four human pathogens and at least six human nonpathogens (245). The pathogenic species are T. pallidum subsp. pallidum which causes venereal syphilis, T. pallidum subsp. endemicum, which causes endemic syphilis (bejel), T. pallidum subsp. pertenue, which causes yaws, and T. carateum, which is the etiologic agent of pinta. Initial studies indicated that these four agents were morphologically indistinguishable, with >95% DNA homology (223, 241). Recently, however, a genetic signature was defined in the 5'-flanking region of the 15-kDa lipoprotein gene (tpp15) that distinguishes T. pallidum subsp. pallidum from T. pallidum subsp. pertenue and endemicum (44).

Most of our knowledge of the physiology, metabolism, and antigenic structure of T. pallidum is derived from the Nichols strain, which has been maintained in rabbits since 1912 (240). T. pallidum is a spirochete varying from 0.10 to 0.18 µm in diameter and from 6 to 20 µm in length, making it invisible by light microscopy (245, 360). Dark-field microscopy is generally used in clinical practice for visualization (60). The average number of windings is 6 to 14, and the organism has pointed ends and lacks the hook shape seen in some commensal human spirochetes (245). The bacterium exhibits characteristic corkscrew motility due to endoflagella, with rapid rotation about the longitudinal axis and flexing, bending, and snapping about the full length (245).

The genome of T. pallidum subsp. pallidum has recently been sequenced by the whole pulse genome random sequencing method (106). The genome is a circular chromosome of 1,138,006 bp and contains 1,041 open reading frames (ORFs). Predicted biological roles were assigned to 55% of ORFs, while 17% match those encoding hypothetical proteins from other species and 28% represent novel genes. Physiologic studies have previously shown that the organism has limited biosynthetic capabilities, requiring multiple nutrients from the host (106). The T. pallidum genome confirms this by demonstrating the inability of the organism to synthesize enzyme cofactors, fatty acids, and nucleotides de novo and by encoding for a pathway for the conversion of phosphoenolpyruvate or pyruvate to aspartate. Given its limited biosynthetic properties, it is assumed that T. pallidum must have good transport proteins, and, indeed, the T. pallidum genome contains 57 ORFs (5% of the total) that encode 18 distinct transporters specific for amino acids, carbohydrates, and cations. Metabolio-pathway analysis shows that genes encoding all of the enzymes of the glycolytic pathway are present in T. pallidum, suggesting that it uses several carbohydrates as energy sources. The organism has previously been demonstrated to survive better in very low concentrations of oxygen and is therefore considered microaerophilic (89). This is confirmed by the lack of genes encoding superoxide dismutase, catalase, or peroxidase, which protect against oxygen toxicity. Motility-associated genes are highly conserved in T. pallidum, consistent with the importance of this activity; 36 genes encode proteins in the flagellar structure. Freeze fracture studies of the outer membrane of T. pallidum show that it contains a small number of integral membrane proteins which may allow it to evade the host immune response (59, 271, 348); genomic analysis indicates 22 putative lipoproteins consistent with these studies. Potential virulence factors are hinted at by the presence of a large number of duplicated genes (tprA through tprL) that encode putative membrane proteins that may function as porins and adhesins. Multiple copies of the tpr genes may represent a mechanism for antigenic variation in T. pallidum and in addition may provide new targets for vaccine development. Previous studies have indicated that T. pallidum does not produce lipopolysaccharide or potent exotoxins, although cytotoxic activity against neuroblasts and other cell types has been observed (94, 363). Genome analysis shows five genes encoding proteins similar to bacterial hemolysins. Comparison of the T. pallidum genome with that of another spirochete, Borrelia burgdorferi, showed that 46% of the T. pallidum ORFs have orthologs in B. burgdorferi, of which 76% have predicted biological function. A total of 115 ORFs shared by T. pallidum and B. burgdorferi encode proteins of unknown biological function, of which 50% appear unique to the spirochete group.

EPIDEMIOLOGY

Top Previous Next References

Transmission of Disease

The primary mode of transmission is by sexual contact, and the next most common is transfer across the placenta (324). Kissing, blood transfusion, and accidental inoculation have also been reported as routes of transmission but are of minor importance today (324). Partner notification studies have estimated a transmission rate of primary, secondary, and early latent syphilis of 18 to 80% (54, 187, 299, 302, 312, 347), while prospective trials of prophylactic treatment have estimated a transmission rate of 9 to 63% (3, 228, 304). The limitations of these studies have recently been discussed, and the authors conclude that the transmission probability per partner is around 60% (108).

The majority of infants with congenital syphilis are infected in utero, but the newborn can also be infected by contact with an active genital lesion at the time of delivery (95, 140). Presumed nonsexual transmission resulting in lesions of the finger and hand has been described in health care workers before the routine use of gloves and in young children sharing a bed with an infected person (324, 340).

The risk of transmission through blood is negligible due to improved donor selection, uniform serologic testing of all blood donors, and a shift from transfusion of fresh blood to transfusion of refrigerated blood components (9, 360). Transmission via blood products is nonetheless theoretically possible since organisms may survive for up to 5 days in refrigerated blood (342, 343). Needle sharing probably does not play a significant role in syphilis transmission, but this remains unclear (175).

(i) United States. (a) Primary and secondary syphilis. In 1947, prior to the penicillin era, the incidence of primary and secondary syphilis was reported at 66.4 cases per 100,000 persons (236). Rates declined to 3.9 cases per 100,000 persons by 1956 due to the availability of penicillin, changes in sexual behavior, and public health measures (236). Studies published by the U.S. Public Health Service suggest that antibiotic therapy, whether intentional or not, contributed to this decline. Since the nadir in 1956, syphilis rates have peaked and troughed in approximately 10-year cycles, with the overall trend being toward increasing rates. The most recent epidemic was noted in 1990, with reported rates for primary and secondary syphilis at 20 per 100,000 persons, a 59% increase since 1985 (41, 111, 351). Although no single factor can explain this trend, an important contributing factor is crack cocaine use and the exchange of illegal drugs for sex (7, 37, 39, 212, 225, 281). In 1991, there was a reversal of the 5-year increasing trend (72). The reasons for the decline are unclear but may be related to renewed priority and increased resources given to syphilis control programs (350). By 1997, a rate of 3.2 cases per 100,000 persons was reported, the lowest incidence ever reported (320).

(ii) Other developed countries. The rate of primary and secondary syphilis in Canada peaked at a rate of 5.5 per 100,000 population in 1984 and then declined (73). Except for a small rise in 1992, the rate has remained low, with a reported rate of 0.4 per 100,000 in 1995 (73). Overall rates of latent and congenital syphilis have also fallen in the last decade, and it is noteworthy that only five cases of congenital syphilis have been reported in Canada since 1990 (73). The decline in latent syphilis may be related to improved screening and treatment of early symptomatic disease.

Developing countries. Although accurate figures on the incidence of syphilis are not available for most developing countries, population-based studies have shown the seroprevalence to vary widely, from 0.9 to 94% depending on the group tested (12, 355). The World Health Organization estimated that there were approximately 12 million new cases of syphilis in adults worldwide in 1995 (355). The greatest number of cases was estimated to have occurred in South and Southeast Asia, with 5.8 million cases, while a further 3.5 million cases occurred in sub-Saharan Africa (355). Factors associated with a high prevalence of syphilis include certain occupations such as long-distance truck driving and commercial sex work, presence of other sexually transmitted diseases (STDs), lack of male circumcision, and level of education (239, 355). African long-distance drivers have been reported to have rates of syphilis as high as 15%, while commercial sex workers have rates varying from 23 to 47% in North Africa and the Middle East (355).

PATHOGENESIS AND PATHOLOGIC FINDINGS

Top Previous Next References

Pathogenesis

Most of the information on the pathogenesis of syphilis is derived from animal models because of the limited information available from human studies (91, 233). T. pallidum is presumed to penetrate through small breaks in the skin (91, 205, 206). Magnuson et al. were able to demonstrate that two organisms inoculated intracutaneously in rabbits produced a dark-field positive lesion in 47% of cases (202). This increased to 71 and 100% when 20 and 200,000 organisms were inoculated, respectively. On intracutaneous inoculation, the incubation period varied with the size of the inoculum, so that with a large inoculum, e.g., 10⁷ organisms, a chancre appeared in 5 to 7 days (202). An inoculation study in human volunteers, which would be considered unethical by today's standards, showed similar findings (203). Animal studies have shown that the organisms appear within minutes in lymph nodes and disseminate widely within hours (61, 272, 324). Although the exact mechanisms by which T. pallidum enters cells is not known, it has been shown to attach to mammalian cells in vitro (93, 143). Attachment may occur by specific attachment ligands (17, 331, 332). Invasion appears to be a critical virulence factor for T. pallidum, as demonstrated by its ability to penetrate endothelial cell monolayers and intact membranes (277, 333).

The pathologic findings of all stages of syphilis are characterized by vascular involvement with endarteritis and periarteritis and in the gummatous stage by granulomatous inflammation. In most primary lesions, the epidermis demonstrates hyperplasia with widening and elongation of the rete ridges (81). The ulcer surface is covered with an exudate consisting of fibrin, necrotic tissue fragments, and polymorphonuclear leukocytes. A dense inflammatory infiltrate is seen in the adjacent dermis, with predominantly lymphocytes and plasma cells, but histiocytes and polymorphonuclear cells are also seen. The perivascular area is particularly involved and associated with endothelial-cell swelling. Silver staining invariably demonstrates the presence of spirochetes, mainly in the dermal-epidermal junction in the perivascular area.

In secondary syphilis, a wide variety of histologic changes are seen, with lymphocytes and plasma cells present in the dermis in 75 to 100% of patients (1, 81). Silver staining reveals the presence of treponemes in around 70% of patients (81). Small blood vessels often show endothelial-cell swelling (1). The epidermis is frequently involved, with exocytosis, spongiosis, parakeratosis, and acanthosis being the most frequent changes (1). In late secondary lesions, the infiltrate may become granulomatous (1).

Histopathologic examination of syphilitic aneurysms demonstrates invasion of the aorta by spirochetes (156). An inflammatory exudate of lymphocytes and plasma cells forms about the vasa vasorum of the adventitia and is followed by obliterative endarteritis of nutrient vessels. In later cases, gumma and scar tissue develops.

In meningitis, the meninges are infiltrated by lymphocytes, plasma cells, and occasionally polymorphonuclear leukocytes with or without perivascular infiltration and vessel thrombosis due to endarteritis (222). In meningovascular syphilis, there is diffuse thickening and lymphocytic infiltration of the meninges and the perivascular spaces (148). The large and medium-sized vessels show fibroblastic and collagenous thickening of the intima, thinning of the media, and fibrous change of the adventitia, together with lymphocytic and plasma cell infiltration (148). Luminal narrowing leads to thrombosis and ischemic infarction of adjacent brain (148). Rarely, aneurysmal dilatation of the affected vessel may occur (148). When a gumma of the central nervous system forms, the histologic tests show a chronic inflammatory infiltrate with primarily lymphocytes, plasma cells, and multinucleate giant cells with or without necrosis (104, 179).

The manifestations of syphilis have been recognized for centuries. Juan de Vigo described genital pustulae in his account of the "French sickness" in 1514 (266). Many others subsequently described the primary chancre followed by a reddish rash whose superficial resemblance to smallpox was responsible for the origin of the term "pox." Protean late manifestations involving all organ systems have been described, heralding syphilis as the "great imitator." The key to diagnosis remains a high index of suspicion.

Boeck performed a prospective natural history study of 1,978 patients with early syphilis in 1891 (56). His observations over 20 years were then continued by Brusgaard, and this information was later termed the Oslo Study (56). Between 1949 and 1951, Gjestland undertook a monumental follow up of 1,404 of the original 1,978 patients (115). The data were reviewed in 1955 and then reanalyzed in 1964 by Clark and Danbolt (55, 56). Although the study is important in terms of defining natural history because of its large sample size, long follow-up, and almost complete absence of treatment, it does have some major design flaws. Diagnoses were made clinically since neither serologic testing nor microscopy was available when the study began, more women than men were admitted because more female beds were available, there were no controls, and autopsies were carried out on only 24% of patients. The results indicate that approximately one-third of the patients developed tertiary manifestations of neurologic, cardiovascular, and gummatous (late benign) syphilis and that the probability of dying due to untreated syphilis was 17% in males and 8% in females (55). Late benign syphilis was the most frequent manifestation, occurring in 14% of males and 17% of females, 1 to 46 years after healing of the secondary lesions. The incidence of clinically apparent cardiovascular syphilis was 13.6% in males and 7.6% in females, but the true incidence may have been higher had more autopsies been carried out. Symptomatic neurosyphilis developed in approximately 9.4% of males and 5.0% of females.

In 1932, the U.S. Public Health Service initiated the Tuskegee study observing African American men with untreated syphilis for 40 years (279, 335). The study has been heavily criticized on ethical grounds, primarily because written informed consent was not obtained from patients and also because treatment was withheld (335). A total of 412 men with untreated latent syphilis and 204 uninfected matched controls were monitored prospectively. Vonderlehr reviewed autopsy material from the first years of the study and noted that only one-fourth of the untreated patients were without evidence of tertiary syphilis after 15 years of infection (346). Cardiovascular involvement was the most frequently detected abnormality, with 50% of the patients infected for 10 years demonstrating cardiovascular involvement (259). Of the 41% survivors at 30 years follow-up, 12% had evidence of late, predominantly cardiovascular syphilis (279).

Rosahn retrospectively reviewed all autopsies at the Yale University School of Medicine from 1917 to 1941 (289). Of 4,000 autopsies of persons aged over 20 years at death, 9.7% had clinical, laboratory, or autopsy evidence of syphilis. About half had received no therapy. Among clinically diagnosed patients, 39% had late anatomic lesions. Of 77 patients with untreated syphilis and late anatomic lesions at autopsy, 83% had cardiovascular complications, 9% had late benign lesions, and 8% had neurosyphilis (289).

In summary, the three studies demonstrate that 15 to 40% of untreated patients with syphilis develop recognizable late complications. A higher mortality rate was noted in populations with syphilis, men were twice as likely to develop late complications, and it was suggested that African Americans were more likely to develop cardiovascular syphilis whereas whites were more likely to develop neurosyphilis (56, 289).

Effects of untreated syphilis on pregnancy outcome in 1951 demonstrated marked differences in risk according to the stage of maternal infection (162). If maternal infection occurred early during pregnancy, the incidence of stillbirth was 25%, that of neonatal death was 14%, that of infected infants was 41%, and that of noninfected infants was 20%. This is in contrast to rates of 12% for stillbirth, 9% for neonatal death, 2% for infected infants, and 77% for noninfected infants when maternal infection occurred late during pregnancy.

The clinical manifestations of the various stages of syphilis are summarized in Table 1.

Primary syphilis. The classic lesion of primary syphilis, the chancre, is a single, painless, indurated ulcer with a clean base (324). However, there is significant variability in morphologic presentation, making clinical diagnosis unreliable (46, 48, 70). The classic combination of a painless, ulcerated lesion with a clean base has a sensitivity of only 31% but a specificity of 98% (70). Induration is the single most specific sign, occurring in 47 to 92% of patients (46, 70). Purulence involving less than 30% of the base has been reported to be the most sensitive finding (70). The size of the chancre varies from 0.3 to 3.0 cm, and occasionally there are multiple lesions (46, 70). There is usually a sharply marginated border, although lesions with undetermined edges have been described (46, 70). Painless regional lymphadenopathy occurs in up to 80% of patients and is more often associated with genital lesions (46, 70, 224). Although pain is uncommon, as many as 36% of patients experience tenderness on palpation (70). The incubation period varies from 3 to 90 days.

Secondary syphilis. There may be no sharp demarcation between primary and secondary syphilis. A primary chancre is still present in as many as one-third of patients with secondary syphilis (47, 224). Alternatively, the primary lesion may have healed for, typically, 8 weeks before any cutaneous or constitutional signs appear. As many as 60% of patients found to have syphilis do not recall lesions of any sort. Secondary syphilis is often a subtle disease; the skin lesions may be easily overlooked and may mimic other dermatologic diseases (47, 324). The appearance of the skin rash ranges from macular to maculopapular, follicular, and occasionally pustular (47, 324). Classically, the lesions are described as "raw ham" or copper colored. They tend to be universally distributed, and the palms and soles are commonly involved (47, 324). Although the lesions are generally described as nonpruritic, patients may experience various degrees of pruritus (47). In untreated patients, the lesions resolve over several weeks and may heal with scarring or hyper- or hypopigmentation, although the majority heal without any scarring. The classic alopecia of secondary syphilis is patchy with a "moth-eaten appearance" and has been reported in up to 7% of patients (47, 224, 324).

Latent (early and late) syphilis. The latent or asymptomatic stage of syphilis is defined as the period from disappearance of the secondary manifestations until therapeutic cure occurs or tertiary manifestations develop (156). Latent disease is arbitrarily divided into early and late syphilis based on the time to spontaneous mucocutaneous (infectious) relapse of untreated patients. About 90% of first relapses occur within 1 year, 94% occur within 2 years, and the rest occur over 4 years (115). Early latent syphilis is therefore defined as occurring within 1 year of infection, and late latent syphilis is defined as occurring after 1 year. A patient with early latent syphilis is considered to be infectious due to the 25% risk of relapse to secondary syphilis (115).

(i) Cardiovascular syphilis. Although in the prepenicillin era, it was estimated that cardiovascular syphilis accounted for 10 to 15% of all clinical cardiovascular disease (324) and was demonstrable in 55 to 86% of all patients with syphilis at autopsy (156), it is now considered rare (144, 166). Concomitant involvement of the nervous system has been reported in up to 43% of patients (57). Cardiovascular involvement usually occurs between 10 and 30 years after the initial infection (57).

(ii) Neurosyphilis (asymptomatic, meningeal, meningovascular, parenchymatous, and gumma). Although many patients with syphilis undergo cerebrospinalfluid (CSF) invasion by spirochetes, not all will develop CSF abnormalities or neurosyphilis (51, 200, 358). After initial invasion of the CSF by spirochetes, untreated or inadequately treated infection may follow one of several courses: spontaneous resolution, asymptomatic meningitis, or acute syphilitic meningitis (221). After this, the disease may either remain asymptomatic or progress to meningovascular syphilis, tabes dorsalis, or paresis (221). Although neurosyphilis has been divided into five major categories, i.e., asymptomatic, meningeal, meningovascular, parenchymatous, and gummatous, these entities represent a continuum and frequently overlap (221). More recently, reports of unusual presentations and rapid progression of syphilis in patients with concurrent HIV infection has led to the hypothesis that infection with HIV may alter the natural history of syphilis (21, 172, 182, 200, 235).

(iii) Late benign syphilis. The essential lesion of late benign syphilis is the gumma (246). The term "benign" implies that these lesions rarely cause total physical incapacity or death, but when the lesions occur in organs such as the brain or heart, serious complications may occur (156, 178, 220, 314, 324). The most common sites of involvement are the skin, bone, and liver (177). The gumma develops from 1 to 46 years after healing of secondary lesions, with the majority developing by the end of the 15th year (56). Skin lesions may appear as nodular, noduloulcerative, or ulcerative lesions (53, 246). Bone lesions are marked by periostitis involving the cranial bones, tibia, and clavicle, with the cardinal features of nocturnal pain and local swelling (177).

Congenital syphilis. Untreated syphilis can profoundly affect pregnancy outcome, resulting in spontaneous abortion, stillbirth, premature delivery, or perinatal death (103, 162, 352, 353). Prematurity and low birth weight have been reported for 10 to 40% of infants born to untreated mothers (184, 213). Congenital syphilis is also a potentially treatable cause of nonimmune hydrops fetalis (16, 79).

The diagnosis of syphilis depends on clinical findings, examination of lesion material for treponemes, and/or serologic tests for syphilis (366). The laboratory diagnosis of syphilis has been extensively reviewed in a previous issue of this journal (192).

Dark-field microscopy is the main diagnostic method for primary syphilis (324). Direct fluorescent-antibody testing for T. pallidum (DFA-TP) has also been described (163, 164, 283). However, neither dark-field microscopy nor the DFA-TP can distinguish T. pallidum from the other pathogenic species of Treponema (192).

Efforts to cultivate the organism in vitro have been largely unsuccessful, but there has been one report in the literature of successful culture of T. pallidum on artificial medium (58). T. pallidum is readily cultivated by animal inoculation (313, 360). However, rabbit infectivity tests (RIT) are expensive and require 3 to 6 months to complete (192). This method is impractical for routine clinical use but is very sensitive, with the theoretical ability to identify one viable organism (192).

Serologic testing remains the mainstay of laboratory diagnosis for secondary, latent, and tertiary syphilis (192). Serologic tests are divided into nontreponemal and treponemal, tests and neither alone is sufficient for diagnosis. Nontreponemal tests are useful for screening, while treponemal tests are used to confirm the diagnosis.

Nontreponemal tests include the Venereal Disease Research Laboratory (VDRL) and the Rapid Plasma Reagin (RPR) card test, both modifications of the original Wasserman reaction (134, 265). Nontreponemal tests use an antigen comprising lecithin, cholesterol, and purified cardiolipin (a component of mammalian cell membranes) to detect an antibody against cardiolipin that is present in the sera of many patients with syphilis. Although these tests are widely available and relatively inexpensive, they are limited by their lack of sensitivity in early and late syphilis and by false-positive reactions. False-positive reactions are associated with increased age, pregnancy, drug addition, malignancy, and autoimmune diseases, such as systemic lupus erythematosus, as well as with viral (particularly Epstein-Barr virus and hepatitis virus), protozoal, or mycoplasmal infection (152, 192).

Treponemal tests include the serum fluorescent treponemal antibody absorption test (FTA-ABS) and the microhemagglutination test for T. pallidum (MHA-TP), which use lyophilized T. pallidum or a lysate of pathogenic T. pallidum, respectively (66, 133, 158, 291). These tests have higher sensitivity and specificity than the nontreponemal tests and are used as confirmatory tests for syphilis after a reactive nontreponemal test has been reported. False-positive results are rare but have been found in association with mixed connective tissue and autoimmune disease, viral infections, and pregnancy (316).

DNA PCR has been used to identify T. pallidum in clinical specimens but is not readily available for routine clinical use (31, 124, 142, 242). While DNA PCR methods have not been shown to be as sensitive as RIT, a highly sensitive reverse transcriptase PCR, which is able to detect a single treponeme, has been described (45). In addition, multiplex PCRs have been described for the simultaneous detection of the three major causes of genital ulcer disease: herpes simplex virus types 1 and 2, Haemophilus ducreyi, and T. pallidum (252).

There is no "gold" standard for the diagnosis of neurosyphilis, and it is usually based on a combination of reactive serologic tests, abnormalities of CSF cell count and protein levels, or a reactive CSF VDRL (43). Two or more CSF abnormalities are generally considered to be compatible with a diagnosis of neurosyphilis. The CSF VDRL has high diagnostic specificity but low diagnostic sensitivity (30 to 78%) in neurosyphilis (30, 154, 169, 316). A false-positive VDRL titer should occur only if the CSF is contaminated with blood (65, 165). The presence of treponemal antibody in the CSF is not diagnostic for neurosyphilis since it may represent passive diffusion of treponemal antibody from the blood into the CSF rather than active central nervous system infection (169). A negative CSF FTA-ABS however, may be useful to rule out neurosyphilis (170, 192, 209, 316). Some investigators have suggested that the CSF FTA-ABS, T. pallidum hemagglutination test, and MHA-TP are more sensitive than the VDRL in the diagnosis of neurosyphilis, but this is not uniformly accepted (64, 138, 170). Other CSF abnormalities include an elevated leukocyte count with predominantly lymphocytes and elevated protein levels (30). The indications for performing a lumbar puncture in patients with syphilis include neurologic or ophthalmologic symptoms or signs, tertiary syphilis (e.g., aortitis, gumma, and iritis), treatment failure, HIV infection if the RPR is >1:32 or if there are neurologic symptoms or signs, and congenital syphilis (43). Some experts recommend that all patients with HIV infection undergo lumbar puncture, but this recommendation is not strongly supported by available evidence.

The diagnosis of syphilis in pregnancy is most often made by serologic screening done at the first antenatal visit. In the United States, it is currently recommended that all pregnant women be screened by a nontreponemal test early in pregnancy (43). In areas of high prevalence and in patients considered to be at high risk, testing should also be done twice during the third trimester (43). If the nontreponemal test is positive, confirmatory testing by treponemal tests should be carried out. In pregnant women with a history of previous syphilis, nontreponemal titers may increase nonspecifically; this increase may be confused with reinfection or relapse (36, 192). The increase in titer may be attributed to pregnancy if there is clear documentation of previous treatment in the absence of lesions suggestive of syphilis, a fourfold increase in titer, or history of sexual contact with a person with infectious syphilis (192).

All infants born to mothers with positive serologic tests for syphilis should be evaluated by a quantitative nontreponemal serologic test within the first month of life (43). Serum from the neonate is the preferred specimen, since cord blood may produce false-positive results (52). The diagnosis of congenital syphilis is complicated by passive transfer of antibodies from mother to infant, but most of these antibodies should be catabolized and undetectable in noninfected infants by age 6 to 12 months (36, 43, 325). In addition, during the first month of life, infants born to mothers with positive serologic tests should be examined thoroughly for physical signs of congenital syphilis (43, 325). Maternal serologic test results and records should also be reviewed (43). Pathologic examination of the placenta or umbilical cord by using DFA-TP testing is also suggested (43). Dark-field microscopy with DFA-TP testing should be carried out on any suspicious lesions or body fluids, e.g., nasal discharge (43). Additional investigations, such as long-bone radiographs or chest X ray, are dependent on abnormalities found on physical examination. CSF analysis for cell count, protein, and VDRL is indicated if the infant has an abnormal physical examination consistent with congenital syphilis, a quantitative nontreponemal serologic titer in serum that is fourfold greater than the mother's titer, or a positive dark-field or DFA-TP result for tests of body fluids (43). Any prepubertal child with a reactive syphilis serologic test should undergo a full evaluation, and maternal serologic test results and records should also be reviewed (43).

The diagnosis of congenital syphilis depends on a combination of physical, radiographic, serologic, and direct microscopic criteria (36, 325, 369). A confirmed case is defined as occurring in an infant in whom T. pallidum is identified in specimens from lesions, placenta, umbilical cord, or autopsy material (36). A presumptive case is one in any infant or child whose mother was untreated or inadequately treated at delivery or who has a reactive treponemal test for syphilis and one of the following: evidence of congenital syphilis on physical examination or on long-bone X ray, elevated cell count or protein level (without other cause) in CSF, or reactive FTA-ABS immunoglobulin M antibody (36). A syphilitic stillbirth is defined as a fetal death occurring after 20 weeks gestation or weighing more than 500 g in which the mother had untreated or inadequate treatment for syphilis at delivery (36).

TREATMENT

Top Previous Next References

History

Mercury was one of the first treatments used for syphilis, coming into use in 1497 (294, 341). In 1909, Ehrlich introduced arsphenamine (Salvarsan), which was quickly shown to be superior to mercury and became the drug of choice (294). Clinical cure, however, required repeated injections over 18 months. Other preparations of heavy metals, including bismuth compounds, were found to be efficacious and less toxic. In 1917, Wagner von Jauregg introduced the technique of malaria inoculation for the treatment of paretic cases (294). The real breakthrough in the treatment of syphilis occurred in 1943, when Mahoney et al. successfully used penicillin to treat four patients with primary syphilis (204).

Current recommendations for the treatment of syphilis are summarized in Table 2. Although many guidelines have been published for the treatment of syphilis, no prospective controlled studies have addressed the optimum dosage or duration of therapy for any stage of syphilis (43, 191). Since the introduction of penicillin, its use in the treatment of syphilis has been based on clinical experience and on observational uncontrolled studies.

Although T. pallidum has not developed any measurable penicillin resistance, there have been anecdotal reports of persistence of the organism in aqueous humor, CSF, and inner ear despite treatment with recommended regimens (183, 200, 201, 339). In vitro resistance to erythromycin has been described in a single clinical isolate of T. pallidum from a penicillin-allergic patient who failed treatment with erythromycin (319). Norgard and Miller (244) have reported the identification of a plasmid in T. pallidum, which also suggests that the organism has the genetic potential to develop antibiotic resistance, but these findings have not been confirmed or reproduced.

The antibiotics used, as well as the dosage and length of treatment, are dependent on the stage of syphilis (43). Oral penicillin is not used due to inadequate levels in blood. Although other antibiotics have been used, they are alternate choices because of less spirocheticidal action, variable absorption, concerns with patient compliance, and diminished efficacy, especially in pregnant women (341).

The cure of syphilis depends not only on the antibiotic effect but also on the integrity of the immune capacity of the host. Although this has been recognized for many years, HIV infection has made it a very important and controversial issue.

Penicillins and cephalosporins. Benzathine penicillin remains the recommended first-line therapy for early syphilis (43). A retreatment rate of approximately 10% is reported for treatment failures (303). It is worth noting, however, that the retreatment rates in this and many of the treatment studies that follow did not differentiate between treatment failure and reinfection. Some reports of penicillin failure have been attributed to poor penetration of benzathine penicillin into the CSF and persistence of T. pallidum in CSF in early syphilis (21, 172, 200, 339).

Erythromycin. Current guidelines recommend erythromycin base in the treatment of penicillin-allergic patients with early syphilis, at doses of 40 mg/kg/day orally (p.o.) (maximum 500 mg per dose) for 14 days (43, 191). The initial studies with erythromycin estolate in syphilis recommended total doses of 15 to 20 g over 10 days (27). However, other forms of erythromycin were often substituted due to the complications of cholestatic jaundice (24, 303). Schroeter et al. recommended a minimum total dose of 30 g of erythromycin base after their study demonstrated a 25% retreatment rate with lower dosages (303). However, failure of erythromycin therapy, even with the recommended total dosage, has been reported in the treatment of primary syphilis (141).

Tetracycline. Tetracycline, 3 g per day p.o. for 10 days, results in 9.2% retreatment rates at 12 months compared to 3 to 5% with penicillin regimens (303). Only very limited studies of doxycycline have been done, but it has been added to treatment guidelines due to improved compliance and presumed efficacy equal to tetracycline (135, 250).

Chloramphenicol. The early literature concluded that chloramphenicol was as effective as tetracycline in all stages of syphilis, with the advantage of few or no gastrointestinal side effects (247). However, the blood dyscrasias later reported led to its removal as an alternate therapy for syphilis (247).

A longer duration of therapy has been suggested for late latent syphilis because treponemes divide more slowly in late infection and presumably require longer exposure to penicillin (275). Reviews of the limited studies in the treatment of late latent syphilis have deemed them inadequate to determine an optimal treatment regimen (168, 370). The main concern at this stage is to prevent the development of late complications of syphilis (280). The available data assessing the use of penicillin in preventing progression to late disease was accumulated in studies of patients with asymptomatic neurosyphilis rather than latent syphilis. However, it is likely that the findings from these studies can be extrapolated to include latent syphilis (280). In 1956 Smith et al. reported a cumulative relapse rate of 21% at 18 months in 47 patients treated with 2.4 to 2.5 mU of benzathine penicillin G (311). They also reported a 10% relapse rate in 53 patients treated with 4 to 5.9 mU of other penicillin preparations. Even given the obvious concerns about the nonrandomization of patients in this study and the use of different penicillin preparations, the data suggests that 2.4 mU of benzathine penicillin G is inadequate for the treatment of asymptomatic neurosyphilis. The currently recommended three-dose regimen for late latent syphilis is largely empirical but is likely to be effective (280).

Cardiovascular syphilis. The effect of treatment on cardiovascular syphilis remains unclear (177). It is known that the acute aortitis heals with scar formation and that treatment may hasten this healing, leading to the development of rapidly progressing aortic regurgitation during treatment (177). The currently recommended regimen is three doses of benzathine penicillin, 2.4 mU, at weekly intervals (43).

Neurosyphilis. The first question in addressing the treatment of neurosyphilis is whether the disease is present (63). The presence of a reactive VDRL in CSF associated with increased cell counts and increased total protein levels in the CSF is strong evidence of active infection (63). A decision analysis study by Wiesel et al. has suggested that CSF examination offers little additional benefit to empirical treatment with benzathine penicillin for most patients with late latent syphilis (356). In this study of asymptomatic patients with untreated syphilis for more than 1 year, they compared treatment involving 7.2 mU of benzathine penicillin G with performing a lumbar puncture to test for asymptomatic neurosyphilis and management based on CSF analysis. Both strategies resulted in a cure rate of at least 99.7%, and although using lumbar puncture resulted in a 0.2% higher cure rate, a 0.3% complication rate was also reported.

The aim of treatment in pregnancy is to cure the mother and to prevent fetal infection (280). Even before the penicillin era, it was clear that maternal treatment early in pregnancy resulted in a reduction of the number of infected fetuses (255).

All pregnant women with syphilis who have not been previously treated should receive penicillin appropriate to their stage of disease. Unless there is clinical or serologic evidence of new infection or a history of recent sexual contact with a person with early syphilis, retreatment during pregnancy is not necessary. If penicillin allergy is reported, desensitization is strongly recommended. If this is not possible (for example, if the patient refuses to undergo penicillin desensitization), erythromycin may be used, but the infant should then be managed as if born to an untreated mother. Erythromycin does not readily cross the placenta, and the fetus may still develop congenital infection (87, 315). Unpredictable maternal levels of erythromycin in serum are also a concern (261). Tetracycline should not be used because it may cause yellow-brown discoloration of the teeth and have adverse effects on long-bone growth in the newborn (78).

Infants should be treated for congenital syphilis if they were born to mothers with (i) untreated syphilis at delivery, (ii) serologic evidence of relapse or reinfection, (iii) treatment with nonpenicillin regimens for syphilis during pregnancy, (iv) treatment for syphilis within 1 month of delivery, (v) poor documentation of treatment for syphilis, or (vi) insufficient serologic follow-up during pregnancy (43). In addition, any infant with physical or laboratory findings consistent with the diagnosis of congenital syphilis should receive treatment (22, 43).

Single-dose benzathine penicillin has not been evaluated in prospective, randomized controlled trials but was used for many years to treat congenital syphilis. However, limited use of this regimen is currently recommended because benzathine penicillin may not achieve or maintain treponemicidal levels in the CSF of neonates (180, 317). In addition, a small number of treatment failures have been reported with single-dose benzathine penicillin therapy (18, 132, 364).

Both crystalline penicillin G and procaine penicillin have been studied in patients with congenital syphilis (216, 263). The regimens described in Table 2 are those currently recommended for congenital syphilis (43, 191).

Some experts believe that therapy with other penicillins may suffice in the treatment of congenital syphilis (234). Small case series have shown that oral amoxicillin, 6 g daily, plus probenecid twice daily achieved treponemicidal levels in CSF (85, 231). However, these limited data should be interpreted with caution until further studies are done.

Given the reported failures of penicillin therapy in some patients and the limited treatment options in penicillin-allergic patients, there is a need for newer agents. Limited animal studies have shown the newer macrolides clarithromycin, roxithromycin (RU965), and azithromycin to be effective treatments for experimental syphilis (2, 198, 199). A pilot study of azithromycin documented cure in 11 of 13 patients with primary or secondary syphilis (345). A nonrandomized study of azithromycin, 500 mg daily for 10 days or 500 mg on alternate days for 11 days, in 100 patients with early syphilis reported complete resolution of clinical signs and 90% reversion to a negative serum Wasserman reaction (214). T. pallidum immobilization (TPI) and FTA-ABS became negative 12 months after treatment in 40% of patients. Time to resolution of clinical signs was also shorter in the azithromycin group than in the penicillin and erythromycin groups. The newer cephalosporins, i.e., cefmetazole, ceftizoxime, and cefetamet (Ro 15-8074), also show promise in animal models (14, 92, 188). Although useful in the treatment of other sexually transmitted diseases, quinolones have shown poor efficacy in animal models of syphilis (6, 327, 344).

The Jarisch-Herxheimer (J-H) reaction was first observed by Jarisch in 1895 and Herxheimer in 1902 after the use of mercury but has since been described with other treatments (25, 156, 324, 341). One large study showed that the J-H reaction occurred more often after treatment with penicillin than with erythromycin or tetracycline (5). It has been reported to occur in 30 to 70% of patients with early syphilis, 43 to 55% of infants with congenital syphilis, 45% of pregnant patients with syphilis, and 2% of patients with neurosyphilis (25, 86, 186). The cause is not completely understood, but the release of T. pallidum lipoproteins with inflammatory activities from dead or dying organisms is the likely inducer of this clinical phenomenon (243,