The assumed disease progression of a syphilis-infected individual is described in the supplementary material and is shown in Figure S1A (available online only, Supplemental Digital Content 1, avialable at: http://links.lww.com/OLQ/A6). Individuals are designated to be infectious if they are in the exposed, primary, secondary, early latent, or recurrent infectious stages of syphilis with the transmission probability per sexual act depending on the infection stage (Table 1). Individuals treated in the early stages are assumed to be immediately susceptible to reinfection, while those treated in the later stages of syphilis are immune to reinfection for an average of 5 years.18,25 Before the introduction of interventions 55% to 70% of the population is tested for syphilis at least once every year21 depending on their sexual activity and HIV status. We also include a proportion of the population who never get tested for syphilis.25,44
The model was implemented using Matlab with each simulation tracking the dynamic sexual network, syphilis transmission, and disease progression of syphilis-infected individuals. Syphilis transmission was tracked over the years 1998 to 2007. The model was specifically calibrated to match the estimated infectious syphilis diagnoses among gay men in Victoria, Australia. The median trajectory of 50 model simulations, using realistic parameter values (Table 1), accurately reflected surveillance data for Victoria (Fig. S1B in supplementary material, available online only, Supplemental Digital Content 1, available at: http://links.lww.com/OLQ/A6). The 10 simulations that best fit the epidemic data were selected (using a Pearson chi-squared test) to forecast epidemic trajectories over the next 10 years under various interventions. Although the model used Australian behavioral data and was calibrated to reflect the epidemic in the state of Victoria, the relative impact of each intervention should be generally applicable to other settings.
To compare different interventions, the median prevalence and syphilis diagnoses (taken as the number of treatments) per year were recorded. The following interventions were simulated:
1. No change in screening;
2. Increase coverage to screen all men previously tested (85% of the population) once per year; or screen all men once per year (including men not previously tested);
3. Maintain the current testing coverage but increase the frequency to 2 or 4 times per year (with the testing frequency of HIV-positive individuals on antiretroviral therapy (ART) increased from 3 to 6 times per year in both cases); and
4. Contact tracing and testing 50% and 75% of regular partners over the previous 3 months with 5% and 10% of casual partners over the previous month. The ideal scenario where all partners over these time periods are tested was also investigated. It is important to note that “contact tracing” is synonymous with “partner notification” in some settings. Contact tracing usually refers to that carried out by the Health Department's Disease Intervention Specialists to notify and test partners while partner notification is when the patient informs their sexual partners of the infection and advises the partner to seek testing. The coverage rates for our simulations are the actual levels of partner testing that is achieved independent of the means by which the partners are reached. For the remainder of this article we use the term contact tracing to describe this type of intervention.
These scenarios were independently applied to: HIV-positive gay men on ART, highly sexually active gay men, and gay men who engage in group sex, with the remaining population screened at baseline levels. The number of tests carried out and the number of infections averted (compared to the baseline case) was also recorded to measure the efficiency of each intervention scenario.
Our model projected future epidemic trajectories under conditions that all parameters and screening rates remain unchanged. The median prevalence and number of diagnoses for these simulations was used as a baseline from which the relative impact of different screening interventions could be measured.
Increasing Testing Coverage
Increasing the testing coverage (from 55% to 70% per year) to screening all gay men previously tested (85% of the population) results in a relative decrease in the peak prevalence and number of diagnoses of ∼38% and ∼25% (to ∼6% and ∼1011 cases), respectively (Figs. 1A, B). However, there is an increase in diagnoses above the baseline case immediately after this intervention is introduced (Fig. 1B), reflecting the increased number of syphilis infections discovered. Increasing the testing coverage to all gay men, including men not previously tested, results in an expected reduction in prevalence and diagnoses (with prevalence of ∼1.6% and 504 diagnosed cases after 10 years). However, the rate of decrease in diagnoses in this case is similar to the other coverage scenario after a large initial spike due to the group of men not previously tested being a pool for late stage (noninfectious) syphilis. It is unlikely that 100% screening coverage will be achieved. Hence, increasing the coverage within the gay male population will likely have a relatively small impact on syphilis epidemics.
Increasing Testing Frequency
Increasing the frequency of testing can have a large impact on syphilis epidemics (Figs. 1C, D). Maintaining current coverage levels but increasing the frequency of testing to 2 times per year for HIV-negative gay men and HIV-positive gay men not receiving ART and to 6 times per year for HIV-positive men on ART results in a large relative decrease in the peak prevalence of ∼ 61% (to ∼3.8%) (Fig. 1C). However, when the testing frequency for gay men not on ART is increased to 4 times per year the syphilis prevalence immediately declines and slowly decays to a prevalence of ∼1.47% by 2017 (Fig. 1C); that is, a relative decrease (from peak) of ∼84%. For these interventions, the expected number of syphilis diagnoses immediately increases to a sharp peak during the first year after initiating the intervention before decreasing (Fig. 1D). When testing is carried out every 3 months, the number of diagnoses decays rapidly to approximately zero after 10 years. Thus, increasing the frequency of syphilis testing could substantially mitigate an epidemic. However, the prevalence of syphilis in the population rapidly decreases to ∼1.5% and then only very slowly decreases from this value despite diagnoses dropping to zero. The remaining prevalence is predominantly due to men who have never tested for syphilis progressing to late-stage (noninfectious) syphilis. This further emphasizes the importance of encouraging all men to get tested for syphilis.
Tracing and testing a proportion of diagnosed individual's partners, and treating infected cases can be expected to reduce the level of syphilis prevalence (Fig. 2A). Treating 50% of regular sexual partners (from the previous 3 months) and 5% of casual partners (from the previous month) of diagnosed syphilis cases leads to a relative decrease in peak prevalence of ∼29% (from ∼9.6% to ∼6.8%). Increasing the proportion of regular partners treated to 75% reduces the peak prevalence further, to ∼6%. However, doubling the proportion of casual partnerships treated to 10% has only a minimal additional effect (Fig. 2A). In the unrealistic scenario of 100% of all sexual partners being traced and treated, the syphilis prevalence would immediately flatten out to a peak level of ∼3.7% (∼60% relative decrease) and then decrease slightly over 10 years to ∼2.7%. These reductions are not sufficient to reverse trends in syphilis epidemics, even at unrealistically high rates of effective contact tracing.
Synchronized Testing and Follow-up Screening
We also investigated screening strategies where current testing rates are maintained but all syphilis testing is synchronized to occur during a 1-month period and screening strategies that follow-up men who have previously been treated for syphilis. However, we determined that these interventions had minimal impact on syphilis epidemics (results not shown).
Targeting Specific Subpopulations
Increased syphilis testing in specific at-risk subpopulations that may be reached by public health services and community-based campaigns was also investigated. We simulated increased syphilis testing in gay men who engage in group sex, men who have greater than 10 partners per year, and HIV-infected men who are on ART. We found that if all men who engage in group sex are tested 2 or 4 times per year then the peak syphilis prevalence can be reduced by ∼27% and ∼50% (to ∼7.0% and ∼4.8%), respectively (Fig. 2B). Although this targeted strategy has some impact, the effect may be relatively small since it is estimated that only 17% of HIV-negative gay men and 30% of HIV-positive gay men regularly engage in group sex (Table 1).
Targeting men who have more than 10 partners per year are predicted to have a substantial impact on syphilis epidemics. Testing highly active gay men and men who engage in group sex an average of 2 times per year has a larger effect on prevalence than only testing gay men who engage in group sex on a 3-monthly basis. Testing highly sexually active gay men and men who engage in group sex every 3 months has almost the same impact on syphilis epidemics as testing the entire population every 3 months (with HIV-positive men on ART tested 6 times per year) with a similar immediate decay in prevalence and only a slightly higher prevalence (∼1.51% vs. ∼1.47%) by 2017 (Figs. 1C, 2B). This suggests that men who have less than 10 partners per year do not contribute significantly to syphilis epidemics and targeting men with more partners is most efficient.
Efficiency of Interventions
The efficiency of each intervention was measured by calculating the number of infections averted compared with the baseline case versus the total number of tests performed for the intervention over 10 years (Table 2). We found that for interventions not involving contact tracing there is a relatively linear relationship between the number of infections averted and the number of additional tests carried out. This relationship breaks down when the average testing rate is 4 tests/year for each individual not on ART and 6 tests/year for men on ART. Almost twice as many tests are performed over 10 years (∼550,183 vs. ∼277,070) with almost no increase in the number of infections averted (from ∼9929 to ∼9970). Therefore, increasing the testing rate in gay men with few partners is highly inefficient and gives little additional public health benefit. In contrast, contact tracing and testing the partners of an infected individual is highly efficient (Table 2) despite having less effect on the epidemic (Fig. 2A).
Using a simulation model that incorporates detailed sexual-activity and testing data, realistic biological properties and heterogeneous subpopulations, we have shown that syphilis epidemics can occur in highly tested populations of gay men. Our model was used to predict the epidemiological outcomes of changes in testing strategies. Our results suggest that increasing the coverage of syphilis testing within the population of men who are already willing to get tested has minimal impact; coverage is already relatively high and only moderate improvements in coverage can be achieved. However, encouraging syphilis testing among men who have not previously been tested can have a large impact on syphilis epidemics. This is particularly important as this subpopulation of men is at high risk of becoming infected and progressing to tertiary syphilis.
What is required is greater frequency of testing; if testing occurred every 3 months, on average, then syphilis incidence could be reduced substantially. However, targeting men who have low numbers of partners (<10 per year) is ineffective. Increasing testing rates of HIV-positive men on ART is also ineffective since they are a relatively small subpopulation and they already have relatively high testing rates (an average of 3 tests per year). However, epidemiological data suggest that a disproportionate number of syphilis diagnoses are in HIV-positive men,45,46 therefore targeting this population may have significant benefits not detected by our model.
According to our model the key subpopulations that should be targeted for increased syphilis screening are gay men who engage in group sex and those with large numbers of partners (>10 per year). Testing “higher activity” men every 3 months on average should reverse current trends in syphilis incidence and is an efficient intervention strategy.
We also explored the likely impact of tracing partners of cases who are treated for syphilis. We found that this would be a very effective and efficient strategy and should be carried out wherever possible. However, contact tracing will only have a substantial effect on a syphilis epidemic if a very large proportion of regular and casual partners are notified and tested. The intensive labor required to carry out contact tracing is unlikely to achieve the very high rates required to reverse syphilis epidemics. Therefore, we recommend that contact tracing should be a secondary objective for controlling syphilis epidemics among gay men. It should be a supporting priority coupled with the primary objective of increasing the frequency of testing in the population. Although this analysis is limited to gay men our conclusions may be valid in some heterosexual settings, such as those in which large proportions of sexual partners are unlikely to be accessed. While partner notification and contact tracing are very important, increasing population-level testing frequencies is recommended as the most effective way to reduce the incidence and prevalence of syphilis.
Our results were obtained by calibrating our model to gay men in Victoria, Australia. These results are likely to be applicable to other populations worldwide. While quantitative results will vary between settings due to differences in sexual mixing behavior, accessibility to health services, different epidemic stages, and other social variables our qualitative findings should be generalizable to most settings. The need to substantially increase the frequency of syphilis testing (particularly among “high-activity” subpopulations) and to promote contact tracing to reduce syphilis epidemics can be considered in any jurisdiction as interventions most likely to be effective and efficient.
Our model, although informed by the best data available to us, has a number of limitations. It held all parameters constant and thus did not capture changes in behavior over the last 10 years. Numerous biological and clinical aspects of syphilis infection are unknown. For example, it is difficult to model the infectiousness of individuals in the primary and secondary stages of syphilis: infected individuals could be highly infectious over a shorter time period. The degree of immunity following treatment at different stages is also uncertain. For these factors we reviewed the available literature and made the best assumptions possible (Table 1). Furthermore, we also neglected age- or group-specific assortative sexual mixing patterns. However, we believe that our model well characterizes syphilis epidemics (Fig. S1B, available online only, Supplemental Digital Content 1, available at: http://links.lww.com/OLQ/A6) and has been useful in evaluating potential interventions, gaining insight into those that are not likely to succeed, and determining which strategies should be adopted as goals.
The feasibility of increasing screening should be evaluated in each setting. Serologic screening is sensitive and inexpensive with infectious syphilis easily and cheaply treated using intramuscular penicillin G. The cost-effectiveness of syphilis testing depends on the efficiency of the screening strategy. We have provided estimates of the number needed to test in order to avert an infection. While these numbers will differ between locations and settings our estimates can be used as a guide in the decision-making of public health policies and campaigns. Of course, ease of access to testing facilities as well as the likely behavior and attitudes of people to whom strategies are targeted should be considered. There is likely to be a maximum frequency of screening and some people will not be screened regardless of the extent of educational messages. However, if interventions successfully scale-up the frequency of syphilis testing then it is possible to reverse epidemiological trends of this important condition.
1. Golden MR, Marra CM, Holmes KK. Update on syphilis: Resurgence of an old problem. JAMA 2003; 290:1510–1514.
2. Fenton KA. A multilevel approach to understanding the resurgence and evolution of infectious syphilis in Western Europe. Euro Surveill 2004; 9:3–4.
3. CDC. Sexually Transmitted Disease Surveillance 2004 Supplement: Syphilis Surveillance Report. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, 2005.
4. 2004 Canadian sexually transmitted infections surveillance report. Can Commun Dis Rep 2007; 33(suppl 1):1–69.
5. Botham SJ, Ressler KA, Bourne C, et al. Epidemic infectious syphilis in inner Sydney–strengthening enhanced surveillance. Aust N Z J Public Health 2006; 30:529–533.
6. Fenton KA, Imrie J. Increasing rates of sexually transmitted diseases in homosexual men in Western Europe and the United States: Why? Infect Dis Clin North Am 2005; 19:311–331.
7. Chen SY, Gibson S, Katz MH, et al. Continuing increases in sexual risk behavior and sexually transmitted diseases among men who have sex with men: San Francisco, California, 1999–2001, USA. Am J Public Health 2002; 92:1387–1388.
8. Lee DM, Chen MY. The re-emergence of syphilis among homosexually active men in Melbourne. Aust N Z J Public Health 2005; 29:390–391.
9. Branger J, van der Meer JT, van Ketel RJ, et al. High incidence of asymptomatic syphilis in HIV-infected MSM justifies routine screening. Sex Transm Dis 2009; 36:84–85.
10. Thomas SB, Quinn SC. The Tuskegee Syphilis Study, 1932 to 1972: Implications for HIV education and AIDS risk education programs in the black community. Am J Public Health 1991; 81:1498–1505.
11. Clark EG, Danbolt N. The Oslo study of the natural course of untreated syphilis. An epidemiologic investigation based on a re-study of the Boeck-Bruusgaard material. Med Clin North Am 1964; 48:613–623.
12. Sellati TJ, Wilkinson DA, Sheffield JS, et al. Virulent Treponema pallidum, lipoprotein, and synthetic lipopeptides induce CCR5 on human monocytes and enhance their susceptibility to infection by human immunodeficiency virus type 1. J Infect Dis 2000; 181:283–293.
13. Fisher M, Pao D, Murphy G, et al. Serological testing algorithm shows rising HIV incidence in a UK cohort of men who have sex with men: 10 years application. AIDS 2007; 21:2309–2314.
14. Grulich AE, Kaldor J. Trends in HIV incidence in homosexual men in developed countries. Sex Health 2008; 5:113–118.
15. Anderson RM, May RM. Infectious Diseases of Humans: Dynamics and Control. NY: Oxford University Press, 1991.
16. Grakovich RI, Milich MV, Kulikova NP. Experience in using mathematical analysis for predicting morbidity in infectious forms of syphilis. Vestn Dermatol Venerol 1987; 8:36–41.
17. Tesalova OT, Novikova NF, Minaev VA, et al. Practical results of modeling the dynamics of syphilis morbidity. Vestn Dermatol Venerol 1985; 1:41–45.
18. Garnett GP, Aral SO, Hoyle DV, et al. The natural history of syphilis. Implications for the transmission dynamics and control of infection. Sex Transm Dis 1997; 24:185–200.
19. Oxman GL, Smolkowski K, Noell J. Mathematical modeling of epidemic syphilis transmission. Implications for syphilis control programs. Sex Transm Dis 1996; 23:30–39.
20. Pourbohloul B, Rekart ML, Brunham RC. Impact of mass treatment on syphilis transmission: A mathematical modeling approach. Sex Transm Dis 2003; 30:297–305.
21. NSW, VIC and QLD Gay Periodic Surveys, 1998–2006.
22. Tramont EC. Treponema pallidum. In: Mandell GL, Bennett JE, Dolm R, eds. Principles and Practice of Infections Disease. 4th ed. New York: Churchill Livingstone, 1995: 2117–2132.
23. Schrijvers D, Josse R, Trebucq A, et al. Transmission of syphilis between sexual partners in Gabon. Genitourin Med 1989; 65:84–85.
24. Stamm LV. Biology of Treponema pallidum. In: Holmes KK, Sparling PR, Mardh PA, eds. Sexually Transmitted Diseases. 3rd ed. New York: McGraw-Hill, 1998: 467–472.
25. Prestage G, Hudson J, Bradley J, et al. TOMS: Three or More Study. Sydney, Australia: National Centre in HIV Epidemiology and Clinical Research, University of New South Wales, 2008.
26. Fogarty A, Mao L, Zablotska I, et al. The Health in Men and Positive Health cohorts: A Comparison of Trends in the Health and Sexual Behavior of HIV-Negative and HIV-Positive Gay Men, 2002–2005, National Centre in HIV Social Research Annual Report of Trends in Behavior. Sydney: University of New South Wales, 2006.
27. Mao L, Crawford JM, Hospers HJ, et al. “Serosorting” in casual anal sex of HIV-negative gay men is noteworthy and is increasing in Sydney, Australia. AIDS 2006; 20:1204–1206.
28. Fitch TJ, Stine C, Hagar DW, et al. Condom effectiveness: Factors that influence risk reduction. Sex Transm Dis 2002; 29:811–817.
29. Martin IE, Lau A, Sawatzky P, et al. Serological diagnosis of syphilis: Enzyme-linked immunosorbent assay to measure antibodies to individual recombinant Treponema pallidum antigens. J Immunoassay Immunochem 2008; 29:143–151.
30. Nessa K, Alam A, Chawdhury FA, et al. Field evaluation of simple rapid tests in the diagnosis of syphilis. Int J STD AIDS 2008; 19:316–320.
31. Prestage GP, Van de Ven PG, Knox S, et al. The Sydney Men and Sexual Health Study: Changes in Behavior Over Time. Sydney: Monograph, HIV AIDS and Society Publications, 2000.
32. Prestage G, Ferris J, Grierson J, et al. Homosexual men in Australia: Population, distribution and HIV prevalence. Sex Health 2008; 5:97–102.
33. Brunham RC, Plummer FA. A general model of sexually transmitted disease epidemiology and its implications for control. Med Clin North Am 1990; 74:1339–1352.
34. Lossick JG, Kraus SJ. Syphilis. In: Evans AS, ed. Bacterial Infections of Humans: Epidemiology and Control. New York: Plenum, 1991: 675–695.
35. Schober PC, Gabriel G, White P, et al. How infectious is syphilis? Br J Vener Dis 1983; 59:217–219.
36. Moore MB, Price EV, Knox JM, et al. Epidemiologic treatment of contacts to infectious syphilis. Public Health Rep 1963; 78:966–970.
37. Schroeter AL, Turner R, Lucas JB, et al. Therapy for incubation syphilis. JAMA 1971; 218:711–713.
38. Rottingen JA, Garnett GP. The epidemiological and control implications of HIV transmission probabilities within partnerships. Sex Transm Dis 2002; 29:818–827.
39. Magnuson HJ, Thomas EW, Olansky S, et al. Inoculation syphilis in human volunteers. Medicine (Baltimore) 1956; 35:33–82.
40. Driver JR. Reinfection in syphilis. JAMA 1924; 83:1728–1733.
41. Salazar JC, Hazlett KR, Radolf JD. The immune response to infection with Treponema pallidum
, the stealth pathogen. Microbes Infect 2002; 4:1133–1140.
42. Miller JN. Immunity in experimental syphilis. VI. Successful vaccination of rabbits with Treponema pallidum
, Nichols strain, attenuated by -irradiation. J Immunol 1973; 110:1206–1215.
43. Prestage GP, Hudson J, Down I, et al. Gay men who engage in group sex are at increased risk of HIV infection and onward transmission. AIDS Behav 2009; 13:724–730.
44. Zablotska IB, Imrie J, Bourne C, et al. Improvements in sexual health testing among gay men in Sydney, Australia, 2003–2007. Int J STD AIDS 2008; 19:758–760.
45. Dougan S, Evans BG, Elford J. Sexually transmitted infections in Western Europe among HIV-positive men who have sex with men. Sex Transm Dis 2007; 34:783–790.
46. Jin F, Prestage GP, Kippax SC, et al. Epidemic syphilis among homosexually active men in Sydney. Med J Aust 2005; 183:179–183.
Supplemental Digital Content
© Copyright 2010 American Sexually Transmitted Diseases Association