Syphilis: Advances in Diagnosis, Treatment, and Public Health

Syphilis cases in the United States roughly doubled between 2018 and 2022, and the trend is not uniquely American. This resurgence has pushed researchers to rethink how the disease is diagnosed, how treatment supply chains are managed, and how public health systems can intervene before infections spread. The science around syphilis has moved quickly in several directions at once, from rapid point-of-care tests and revised screening algorithms to preventive antibiotic strategies and early-stage vaccine research.

Why Syphilis Is Surging

Reported syphilis cases in the United States climbed from roughly 114,000 in 2018 to over 203,000 in 2022, with primary and secondary cases alone reaching about 59,000 that year.¹ The drivers are layered: reduced condom use, changing sexual behaviors, gaps in sexual health infrastructure, and disruptions to screening during the COVID-19 pandemic all played a role. Men who have sex with men remain disproportionately affected, but heterosexual transmission has climbed sharply as well, which is one reason congenital syphilis numbers have exploded in tandem. The disease had been declining for decades before this reversal, and many clinicians who trained during the low-prevalence era had little firsthand experience recognizing it. That knowledge gap compounds the problem.

The Organism Behind the Disease

Syphilis is caused by Treponema pallidum, a corkscrew-shaped bacterium that remains one of the most elusive human pathogens. It cannot be grown in standard lab cultures, which has hampered research for over a century. Its genome is small, making it heavily dependent on the human host for nutrients and metabolic building blocks.² Despite that stripped-down genome, T. pallidum is remarkably good at evading the immune system. It keeps very few proteins exposed on its outer surface, which limits what the immune system can target. One key surface protein, TprK, undergoes constant variation, essentially reshuffling its identity so that antibodies generated during an earlier encounter no longer recognize it. The bacterium also hides its flagella in the space between its inner and outer membranes, keeping those structures out of immune reach. These traits allow T. pallidum to disseminate rapidly through the bloodstream, cross into the brain and eyes, and establish latent infections that can last decades.

How Screening Has Changed

Syphilis testing has traditionally started with a nontreponemal test, such as the RPR or VDRL, that detects antibodies to tissue damage rather than to the bacterium itself. If that screen came back positive, a more specific treponemal test confirmed the result. This “traditional algorithm” is cheap, widely available, and works well for catching active infections. But it misses people whose nontreponemal antibodies have faded, including those with late latent syphilis or previously treated infections.

The “reverse algorithm” flips the order: a treponemal test screens first, then a nontreponemal test confirms. Many labs have adopted this approach because treponemal tests are easier to automate. In a head-to-head comparison at an Indian referral lab, the reverse algorithm detected about 62% of patients as seropositive compared to about 41% with the traditional algorithm, catching an additional 99 cases the traditional approach missed entirely. The traditional algorithm’s false-positive rate from its initial screen was around 7%, versus under 1% for the reverse algorithm’s initial screen.³

In pregnancy, however, the picture is more nuanced. A study comparing the two algorithms in pregnant women found that while the reverse algorithm flagged more women as screen-positive (about 1.6% versus 0.6%), the proportion ultimately diagnosed with confirmed infection was similar in both groups. Among those who screened positive and had follow-up testing, roughly half to 60% in both algorithms turned out to be false positives after confirmatory testing.⁴ In low-prevalence populations more broadly, reverse screening identified a handful of patients with possible latent syphilis that traditional screening missed, but also produced more false-reactive results.⁵ There is no universally superior algorithm. The choice depends on whether a clinic’s priority is catching every possible latent case or avoiding the follow-up burden of false positives.

Point-of-Care and Molecular Tests

Lab-based screening, whether traditional or reverse, requires a blood draw, a functioning lab, and a follow-up visit to deliver results. In resource-limited settings or outreach clinics, patients may never return for those results. That gap has driven the development of rapid point-of-care tests that can be done with a finger prick and read in minutes.

One of the more promising designs is a dual-pathway rapid test that detects both treponemal and nontreponemal antibodies at the same time, essentially performing both halves of the screening algorithm in a single strip. In a large multisite evaluation using over 3,000 specimens in China, the treponemal line showed sensitivities above 94% and specificities above 99% across three specimen types. The nontreponemal line was somewhat less sensitive for low-titer infections but reached essentially 100% sensitivity when antibody levels were at or above a moderate threshold.⁶ An earlier evaluation of the same dual test concept, using over 1,600 banked serum samples, reported similar concordance: above 98% agreement with RPR at meaningful titers and above 95% agreement with the standard treponemal assay.⁷ These rapid dual tests are especially valuable in antenatal clinics in low-income countries, where a same-visit diagnosis can mean the difference between treatment and a baby born with congenital syphilis.

On the molecular side, PCR-based tests that detect T. pallidum DNA directly from swabs of sores or other clinical specimens have been in development for decades. They are particularly useful in primary syphilis, when a chancre is present but antibodies may not have developed yet. Across many evaluations, these tests show excellent specificity, generally 97% to 100%, though sensitivity varies depending on the specimen type and disease stage.⁸ One real-time PCR assay targeting the polA gene achieved sensitivities and specificities between 94% and 100% in clinical validation.⁹ Despite this performance, no commercially approved molecular diagnostic test for syphilis exists yet. Every assay in use is lab-developed. Getting a standardized, commercially available test through regulatory approval would be a meaningful step forward, especially for early diagnosis when serology is unreliable.

Treatment and the Penicillin Problem

Benzathine penicillin G (BPG) has been the standard treatment for syphilis since the mid-twentieth century, and remarkably, T. pallidum has never developed resistance to it. A single intramuscular injection treats early syphilis effectively, with two-year follow-up data confirming results comparable to or better than more complex penicillin regimens.¹⁰ Late latent syphilis requires three weekly injections. The drug is inexpensive, well-tolerated, and uniquely effective because of T. pallidum’s slow replication cycle and minimal surface armor.

The problem is supply. A multicountry survey found that 41% of the 95 countries and territories assessed reported shortages of BPG between 2014 and 2016. Some countries reported using alternative antibiotics for maternal syphilis during shortages, which is concerning because alternatives are less proven in pregnancy.¹¹ BPG is a low-margin product, and relatively few manufacturers produce it, making the supply chain fragile. These shortages have real consequences: every missed dose in a pregnant person is a potential case of congenital syphilis.

When penicillin is unavailable or a patient is allergic, doxycycline and ceftriaxone are the main alternatives. Azithromycin was once considered a viable option, but resistance has become essentially universal. A survey of over 600 T. pallidum strains collected across 13 US states and two Canadian provinces between 2017 and 2023 found that over 99% carried genetic mutations conferring azithromycin resistance.¹² Azithromycin is no longer recommended for syphilis treatment in most guidelines.

Congenital Syphilis

When a pregnant person has untreated syphilis, the bacterium can cross the placenta and infect the fetus. The risk of adverse outcomes is severe: untreated maternal syphilis is associated with up to an 80% chance of complications including stillbirth, preterm birth, low birth weight, and congenital infection.¹³ Infants born with congenital syphilis who go undiagnosed and untreated can develop serious long-term damage, including hearing loss, skeletal deformities, and neurological impairment.¹⁴

All of this is preventable. Screening in early pregnancy followed by a single injection of BPG can prevent vertical transmission. The failures are almost entirely systemic: late or absent prenatal care, screening gaps, delays in treatment after diagnosis, and the BPG shortages described above. In the United States, congenital syphilis cases have risen in lockstep with adult cases, and the Centers for Disease Control and Prevention has classified the situation as a public health crisis. Many of the affected pregnancies involve people who did receive some prenatal care but were either tested too late or not treated promptly enough after diagnosis.

Neurosyphilis and Why It Gets Missed

T. pallidum can invade the central nervous system at any stage of infection, including early on. When it does, the resulting condition, neurosyphilis, can produce a wide range of symptoms: headaches, visual problems, hearing loss, cognitive changes, personality shifts, balance difficulties, and eventually severe neurological deterioration if left untreated. The classic descriptions of “general paresis of the insane” that filled psychiatric hospitals in the pre-antibiotic era were often neurosyphilis.

Diagnosing neurosyphilis remains genuinely difficult. There is no single definitive test. The standard approach involves a lumbar puncture to examine cerebrospinal fluid for signs of infection and inflammation, but the results are often ambiguous. In one retrospective analysis, only about 11% of patients evaluated for neurosyphilis met criteria for a confirmed diagnosis. Among those treated with neurosyphilis-directed antibiotics, roughly 17% had no cerebrospinal fluid findings actually supporting the diagnosis. Conversely, nearly half of patients who were not treated for neurosyphilis had cerebrospinal fluid results that would have supported the diagnosis.¹⁵ In other words, there is a substantial gap between what the fluid samples suggest and what clinicians actually do. Overtreatment and undertreatment both occur frequently.

Researchers have started looking for better biomarkers. A large-scale proteomic analysis of over 200 cerebrospinal fluid samples identified three protein biomarkers, SEMA7A, SERPINA3, and ITIH4, that showed promise in distinguishing neurosyphilis from other conditions when combined with machine learning models. These were validated in a separate set of 115 additional samples from multiple centers.¹⁶ This is still early-stage work, but it signals a real effort to replace the current diagnostic guesswork with something more reliable.

Ocular and Otic Syphilis

Syphilis has a particular affinity for two sensory organs: the eyes and the ears. Ocular syphilis can affect virtually every structure in the eye, producing inflammation of the iris, retina, optic nerve, or choroid. A review of the clinical spectrum noted presentations ranging from mild anterior inflammation to severe posterior uveitis, retinal detachment, and optic nerve dysfunction.¹⁷ In a clinical series of 127 patients with ocular syphilis, inflammation was bilateral in about two-thirds, and posterior uveitis was the most common pattern. Roughly 39% of patients presented with vision of 20/200 or worse, meaning significant impairment at the time of diagnosis.¹⁸ The clinical features varied somewhat between people with and without HIV: posterior and pan-uveitis patterns were more frequent in HIV-positive individuals, while anterior and intermediate patterns were more common in HIV-negative patients.¹⁸

Otosyphilis, syphilis affecting the ear, tends to be even more underdiagnosed. In a retrospective analysis of HIV-positive syphilis patients, none of those ultimately found to have otosyphilis reported hearing loss on their own; symptoms were only identified through systematic questioning or audiometry.¹⁹ All patients in that series who received intravenous penicillin showed clinically meaningful hearing improvement. The takeaway for clinicians: when syphilis is diagnosed, actively asking about visual and auditory symptoms is essential. Patients will not always volunteer them.

Doxycycline as Prevention

One of the most debated developments in sexual health over the past few years is doxycycline post-exposure prophylaxis, commonly called doxyPEP. The concept is simple: taking a dose of doxycycline within 72 hours of a sexual encounter to prevent bacterial sexually transmitted infections. A meta-analysis of randomized trials in men who have sex with men and transgender women found that doxyPEP reduced syphilis incidence by about 77%, chlamydia by about 81%, and gonorrhea by about 45%.²⁰ These are striking numbers, and the evidence for syphilis prevention specifically has been classified as high-certainty.

The obvious concern is antimicrobial resistance. Widespread doxycycline use could, in theory, breed resistant strains not just of T. pallidum but of other bacteria, including gonorrhea, which already has a worrying resistance profile. Some sexual health organizations have recommended caution for this reason. However, for T. pallidum specifically, the resistance concern appears to be mostly theoretical. No clinically confirmed cases of doxycycline-resistant T. pallidum have been documented. Surveillance efforts have not found relevant mutations in the gene that doxycycline targets. Laboratory experiments exposing T. pallidum to sublethal doxycycline concentrations for up to seven months, both continuously and intermittently, produced no changes in susceptibility or resistance-associated mutations.²¹ The bacterium’s minimal genome and lack of genetic exchange mechanisms appear to make it an unlikely candidate for developing resistance quickly. That said, ongoing surveillance remains important, and the resistance picture for gonorrhea is a separate and more complicated story.²²

The Search for a Vaccine

Given the rising case counts and the fragile treatment supply chain, a vaccine against syphilis would be transformative. But T. pallidum makes vaccine development exceptionally difficult. The bacterium’s outer surface has very few exposed proteins for a vaccine to target, and the proteins that are exposed, like TprK, change their structure constantly through antigenic variation. Researchers also lack a clear picture of what kind of immune response would actually protect someone from infection, a problem known as not having defined “correlates of protection.”²³

Research has focused on chimeric antigens, engineered proteins that combine pieces of multiple T. pallidum surface molecules into a single target, to try to generate broader immune responses. Progress has been made in mapping out T. pallidum genomes from populations around the world to assess whether vaccine candidates would cover the genetic diversity of circulating strains.²⁴ No candidate is near human clinical trials yet. The field is still in the antigen-discovery and platform-selection phase. Rabbit models have historically been used for syphilis vaccine testing, but the inability to culture T. pallidum in the lab means that growing and purifying the organism for study remains painstaking. A functional vaccine is probably still many years out, but the investment in research has increased noticeably since the resurgence began.

Self-Testing and Digital Health

Beyond the clinical lab, there is growing interest in making syphilis testing more accessible through self-testing kits and digital health platforms. A scoping review of screening innovations among men who have sex with men highlighted self-testing and digital health tools, alongside doxyPEP and point-of-care testing, as strategies with genuine potential to reach people who face barriers to traditional clinic-based care, whether those barriers involve geography, stigma, or scheduling.²⁵ HIV self-testing kits have already demonstrated that people will test at home when given the option, and similar models are being explored for syphilis. The challenge is ensuring that a positive result leads to actual treatment. A home test is only useful if the person who tests positive has a clear and accessible pathway to care.

The Origins Debate

The origin of syphilis has been argued about for centuries. The “Columbian hypothesis” holds that European explorers brought treponemal disease back from the Americas in the late 1400s. The “pre-Columbian hypothesis” argues that treponematoses were already present in the Old World. Ancient DNA evidence has recently shifted this debate significantly. Researchers recovered treponemal genomes from nearly 2,000-year-old human remains in Brazil, reconstructing four ancient genomes most closely related to the subspecies that causes bejel (endemic syphilis). These genomes helped refine molecular clock estimates, placing the divergence of modern T. pallidum subspecies firmly in pre-Columbian times.²⁶

Separately, ancient T. pallidum DNA recovered from pre-Columbian child sacrifices in Tlatelolco, Mexico defined a novel ancient treponemal population. Phylogenetic analysis showed this population displayed ancestral features and carried the genetic building blocks of the subspecies that cause disease today, supporting the Americas as the source of worldwide treponemal spread.²⁷ Interestingly, the Brazilian genomes most closely resembled the bejel agent rather than the venereal syphilis agent, which complicates the idea that the subspecies causing different treponemal diseases occupied neat, geographically determined niches. The treponemal family, which includes yaws, bejel, and pinta in addition to venereal syphilis, cannot currently be distinguished by standard serological tests, only by the clinical and epidemiological context.²⁸ These ancient DNA studies are beginning to reveal that the boundaries between subspecies were more fluid than previously assumed, and that treponematoses have been shaping human health far longer than the Columbus timeline would suggest.