In the prepenicillin era, infection with Treponema pallidum was common, and syphilis was a major cause of disability and death. The annual incidence of infection was 1.8% in US Army soldiers stationed in the United States in 1914.1 Over 575,000 cases (of all stages) were reported in the United States in 1943.2 Long-term complications from syphilis occurred in 28% of the Oslo cohort of patients with untreated syphilis diagnosed in 1891 to 1910.3 Cardiovascular syphilis was diagnosed in 10.4% of patients, and 6.5% developed neurosyphilis.3 The case fatality rate was 10.8% (15.1% in men, 8.3% in women) with deaths occurring up to 40 years after infection.3 Syphilis was the 10th leading cause of death in the United States in 1923 to 1925, averaging 17,515 deaths per year.4
When penicillin became available in the 1940s, there were major decreases in incidence of syphilis. Reported cases of primary and secondary syphilis fell from 94,957 in 1946 to 6454 in 1955.2 Serious complications of syphilis also decreased. Between 1927 and 1937, 6.9% (1040) of the 15,000 autopsies at Philadelphia General Hospital had evidence of cardiovascular syphilis.5 By 1950 to 1960, the prevalence of syphilitic aortitis had fallen to 0.8% (100) of 13,082 autopsies at Kings County Hospital in Brooklyn, NY.6 Since then, case series of cardiovascular syphilis have been reported,7 but it is difficult to say what the rates of disease due to cardiovascular syphilis have been. Similarly, syphilis was responsible for 8.6% of admissions to the Boston Psychopathic Hospital between 1912 and 1934.8 However, by 2001, syphilis was so rarely associated with dementia that the American Academy of Neurology dropped syphilis testing from the routine work-up.9
Increases in syphilis in the late 1980s,2 high prevalence of coinfection with human immunodeficiency virus,2 and increased recognition of ocular syphilis10 have raised concerns about potential increases in syphilis-related morbidity among persons whose infections have remained undiagnosed. Syphilis surveillance focuses on infection and does not monitor complications of syphilis. Trends in tertiary syphilis are not available.11 However, death certificates offer an opportunity to search for trends in severe complications of syphilis.12 We studied trends in syphilis as an underlying cause of death as a surrogate for trends in severe complications of syphilis and to help assess the frequency of severe complications.
We used CDC Wonder13 to access US mortality data which were available from 1968 to 2015. Three different International Classification of Diseases (ICD) codes were used during this time, ICD8 (1968–1979), ICD9 (1980–1998), and ICD10 (1999–2015). Categories for classifying syphilis changed minimally with these different code sets. The main diagnostic categories for syphilis were consistent over time: congenital, early, cardiovascular, neuro, and late unspecified. ICD9 contained a few more subcategories for early, cardiovascular, and neurosyphilis than ICD8 or ICD10. For example, ICD8 cases of “other cardiovascular syphilis” could be specified as “syphilitic aortitis,” “syphilitic endocarditis,” or “other specified cardiovascular syphilis” in ICD9.
We searched for mortality records with the reported underlying cause of death attributed to syphilis (ICD codes 090-097). Deaths from subcategories were collapsed and counted according to the main diagnostic categories: congenital, early, cardiovascular, neuro, and late unspecified. Numbers of deaths were examined after stratifying by age, race, and gender.
Reported cases of primary and secondary syphilis, by gender, were obtained from the National Electronic Telecommunications System for Surveillance, the system through which CDC receives notifiable STD data from all 50 states and the District of Columbia. Trends in syphilis deaths were plotted to compare with trends in primary and secondary syphilis among men and women. Trends in congenital syphilis deaths were compared with trends in primary and secondary syphilis among women.
During 1968 to 2015, syphilis was listed as the underlying cause for 6498 deaths. The annual number of deaths decreased rapidly from 586 in 1968 to 80 in 1985. Since 1998, deaths due to syphilis have fluctuated between 28 and 47 per year (Fig. 1). Yearly reports of primary and secondary syphilis declined rapidly from 93,545 in 1947 to 10,449 in 1952, increased to over 20,000 in the 1960s and then peaked at 50,578 in 1990 (Fig. 1). There was no clear increase in deaths due to syphilis among adults after the increase in syphilis cases in the late 1980s.
Congenital syphilis was the underlying cause for 568 of the 6498 syphilis deaths from 1968 through 2015 (this does not include stillbirths). The number of deaths decreased from 28 in 1968 to 9 in 1972, then remained relatively stable until they increased from 9 in 1986 to 35 in 1990 (Fig. 2). Deaths due to congenital syphilis tracked closely with reported cases of primary and secondary syphilis among women which increased 4-fold between 1978 and 1990.
The total number of deaths for the major categories were as follows: congenital, 568; cardiovascular syphilis, 2702; neurosyphilis, 2533; late unspecified syphilis, 668; and early syphilis, 27. All categories of syphilis deaths decreased over time; however, cardiovascular syphilis deaths decreased more than neurosyphilis deaths (Fig. 3). Cardiovascular syphilis accounted for 338 deaths in 1968 but only between 1 and 6 per year since 1998. Neurosyphilis accounted for 191 deaths in 1968 and between 16 and 34 per year since 1998.
Overall, 63.6% of syphilis deaths were in persons aged 65 years or older. The age at death increased slightly over time. The most common age group was 65 to 74 years during 1968 to 1978 and 75 to 84 years during 1979 to 2015. By race, there were more deaths among whites (151 per year) than blacks (111 per year) during 1968 to 1985; thereafter, numbers of deaths were similar (28 per year for each).
More deaths due to syphilis occurred among men (4149) than women (2349). There were 2.2 times more male deaths than female deaths during 1968 to 1975 (Fig. 4). Since that time, the male/female ratio of deaths has fluctuated, but remained at about 1:6. There was a 4-fold increase in reported primary and secondary syphilis among women between 1978 (5,331 cases) and 1990 (22,088 cases). The number of syphilis deaths among women (excluding congenital cases) did not increase after this epidemic; it has continued to fall from an average of 23.6 per year (1991–1995) to 14.4 per year (2011–2015) (Fig. 4). Similarly, deaths among men have not increased despite increases in reported cases from 1968 to 1990.
Death certificates are a crude approach to tracking severity of disease over time,14,15 but for syphilis, there are few alternatives. The trends in deaths caused by syphilis showed little year-to-year variation, suggesting that they might be valid reflections of trends in disease. The rapid decline in syphilis deaths among adults that occurred through 1980 likely reflects the dramatic decreases in incidence that occurred 30 to 40 years earlier and the ability of penicillin to prevent progression of prevalent infections to tertiary disease. For adults, deaths due to untreated syphilis can occur 10 to 40 years after infection.3 This broad interval blurs the correlation between infection and death. In contrast, deaths due to congenital syphilis usually occur within a year of birth, so trends in congenital syphilis deaths would be expected to mirror trends in congenital syphilis and syphilis among pregnant women. Indeed, the increase in deaths due to congenital syphilis in the late 1980s very closely matches the increases in syphilis that occurred among women at that time.
The absence of any increase in adult deaths 25 years after the major increases in syphilis in the late 1980s suggests that the attack rate for late manifestations of syphilis is much lower than it was in the past. This finding is consistent with observations in the 1970s that a predicted rise in neurosyphilis due to untreated infections acquired during World War II never materialized.16 Perhaps deaths have decreased because nearly all infections are detected and treated before late complications can occur. Alternatively, many infections may be inadvertently cured when antibiotics are taken for other reasons.16 It is also possible that the organism has evolved to be less pathogenic.17 Even in the 1930s, before penicillin, it was noted that “an often fatal infection has evolved into one of exquisite chronicity.”1
In contrast to the decrease in likelihood of progression to tertiary syphilis in adults, the likelihood of progression to serious disease after congenital infection has not changed; 66.5% of women with untreated syphilis will have an adverse outcome of pregnancy, including 25.6% who will have fetal loss or a stillbirth.18 This change for adults but not for babies may be due to differences in pathogenesis or to the differences in opportunities for treatment (including inadvertent treatment). The time between infection and serious complications is often measured in decades for adults, but can be a matter of weeks for the developing fetus.
Our findings raise a few questions about complications of syphilis. Why did cardiovascular syphilis deaths decrease faster than neurosyphilis deaths? This trend appears to be consistent with changes in nonfatal tertiary syphilis, though data are sparse.19,20 These changes could be influenced by changes in clinical suspicion, ability to diagnose and treat complications, or changes in autopsy rates. It is possible that the blood brain barrier limits the impact of antibiotics taken for other reasons, so inadvertent treatment reduces cardiovascular syphilis more than neurosyphilis.21,22 Maybe the pathogenesis of neurosyphilis begins earlier in the course of infection or is less likely to be stopped by treatment22 than cardiovascular syphilis,21 providing a shorter opportunity for cure.
Why are men more likely to develop complications and die from syphilis than women?3,5,6,23 Men have been more likely to acquire infection than women,2 and there may also be differences in how often infections are detected and treated. However, difference in complication rates was consistently seen before penicillin, even when confining analyses to infected persons.3,23 In the Oslo cohort of untreated syphilis, infected men were more likely than infected women to develop neurosyphilis (9.2% vs. 5.1%), cardiovascular syphilis (14.9% vs. 8.0%), or to die from syphilis (15.1% vs. 8.3%).3 Previous hypotheses to explain these differences include the possibility that physical exertion increased the risk of cardiovascular syphilis.24 Differences in immune systems are now thought to contribute to gender differences in response to many other infections.25 Advances in understanding of physiology and new laboratory techniques may someday solve this mystery.
These changes in disease caused by syphilis have implications for prevention. Reduction in adverse effects of syphilis among adults means that there is less urgency to find and cure adults with syphilis than in the past. Continuing high risk for complications from congenital syphilis means congenital syphilis should be considered a higher priority. In 2016, deaths from congenital syphilis (including stillbirths) exceeded deaths from syphilis among adults even though adult cases exceeded congenital cases by a factor of 100.2 Most congenital infections can be prevented by screening and rescreening pregnant women.26 However, preventing all congenital infections requires preventing syphilis among women. Low rates of infection among women and recent successes in stopping outbreaks among heterosexuals suggest that this is possible.27
Our study has important limitations. We measured deaths, not nonfatal outcomes which can also be serious. Deaths were determined by a wide variety of providers and coders which likely leads to inconsistency among observers and could cause trends over time. We did not examine medical records to validate the cause of death. We had little information on each death beyond basic demographics. We did not have information on sexual orientation, so we could not look specifically at trends among men who have sex with men. We compared trends in deaths to trends in reported cases. Trends in deaths are more likely related to changes in the number of untreated infections, which are not known.
In conclusion, the reduction in deaths due to syphilis is miraculous and continues despite the inability to prevent infection among adults. Some characteristics of syphilis remain medical mysteries, including the greater decreases in cardiovascular syphilis compared to neurosyphilis, and the higher case fatality rate for men compared with women. Changes in prognosis after an infection call for changes in priorities to prevent serious disease with an emphasis on preventing syphilis among women to prevent congenital syphilis.
1. Moore JE. An evaluation of public-health measures for the control of syphilis: An epidemiological study. Lancet 1951; 1(6657):699–711.
2. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2016. Atlanta: U.S: Department of Health and Human Services, 2017.
3. Gjestland T. The Oslo study of untreated syphilis; an epidemiologic investigation of the natural course of the syphilitic infection based upon a re-study of the Boeck-Bruusgaard material. Acta Derm Venereol 1955; 35(Supp 34):364.
5. Welty JW. A necropsy survey of cardiovascular syphilis with particular reference to its decreasing incidence. Am J Med Sci 1939; 197:782–793.
6. Heggtveit HA. Syphilitic aortitis. A clinicopathologic autopsy study of 100 cases, 1950 to 1960. Circulation 1964; 29:346–355.
7. Roberts WC, Barbin CM, Weissenborn MR, et al. Syphilis as a cause of thoracic aortic aneurysm. Am J Cardiol 2015; 116:1298–1303.
8. Moore M, Merritt HH. Role of syphilis of the nervous system in the production of mental disease. JAMA 1936; 107:1292–1293.
9. Knopman DS, DeKosky ST, Cummings JL, et al. Practice parameter: Diagnosis of dementia (an evidence-based review). Report of the quality standards subcommittee of the American Academy of Neurology. Neurology 2001; 56:1143–1153.
10. Oliver SE, Cope AB, Rinsky JL, et al. Increases in ocular syphilis—North Carolina, 2014–2015. Clin Infect Dis. In Press.
11. Tuddenham S, Ghanem KG. Neurosyphilis: Knowledge gaps and controversies. Sex Transm Dis 2018; 45:147–151.
12. Barragan NC, Moschetti K, Smith LV, et al. Differential declines in syphilis-related mortality in the United States, 2000–2014. Am J Infect Control 2017; 45:417–420.
14. Johansson LA, Westerling R, Rosenberg HM. Methodology of studies evaluating death certificate accuracy were flawed. J Clin Epidemiol 2006; 59:125–131.
15. McGivern L, Shulman L, Carney JK, et al. Death certification errors and the effect on mortality statistics. Public Health Rep 2017; 132:669–675.
16. Catterall RD. Neurosyphilis. Br J Hosp Med 1977; 17:585–604.
17. Smajs D, Norris SJ, Weinstock GM. Genetic diversity in Treponema pallidum
: Implications for pathogenesis, evolution and molecular diagnostics of syphilis and yaws. Infect Genet Evol 2012; 12:191–202.
18. Gomez GB, Kamb ML, Newman LM, et al. Untreated maternal syphilis and adverse outcomes of pregnancy: A systematic review and meta-analysis. Bull World Health Organ 2013; 91:217–226.
19. Sparling PF, Swartz MN, Musher DM, et al. Clinical Manifestations of Syphilis. Chapter 37 in Holmes KK, Sparling PF, Stamm WE, et al. Eds. Sexually Transmitted Diseases Fourth Edition. 2008 McGraw-Hill, New York.
20. Hook EW 3rd. Syphilis. Lancet 2017; 389:1550–1557.
21. Thompson WP, Comeau WJ, White PD. The role of the treatment of syphilis in the prevention of cardiovascular involvement. Am Heart J 1939; 17:286–293.
22. Drago F, Javor S, Parodi A. Neurosyphilis: From infection to autoinflammation? Int J STD AIDS 2016; 27:327–328.
23. Singh AE, Romanowski B. Syphilis: Review with emphasis on clinical, epidemiologic, and some biologic features. Clin Micro Rev 1999; 12:187–209.
24. NICOL CS. Cardiovascular syphilis. Br J Vener Dis 1950; 26:109–125.
25. Giefing-Kröll C, Berger P, Lepperdinger G, et al. How sex and age affect immune responses, susceptibility to infections, and response to vaccination. Aging Cell 2015; 14:309–321.
26. Matthias JM, Rahman MM, Newman DR, et al. Effectiveness of prenatal screening and treatment to prevent congenital syphilis, Louisiana and Florida, 2013–2014. Sex Transm Dis 2017; 44:498–502.
27. Peterman TA, Su J, Bernstein KT, et al. Syphilis in the United States: On the rise? Expert Rev Anti Infect Ther 2015; 13:161–168.