Background: Accurate, updated estimates of the incidence and prevalence of sexually transmitted infections in the United States remain elusive. The most widely quoted number of new sexually transmitted disease (STD) cases each year is 12 million. However, this figure has not changed in more than a decade, despite improvements in detection methods and the effects of STD control programs.
Goals: To propose a system for weighing the strength of STD surveillance data and to estimate the incidence and prevalence of STDs in the United States for 1996, using the available published data.
Results: We estimate that more than 15 million STD infections occurred in the United States in 1996. This number exceeds the earlier estimate primarily because improved detection techniques have allowed an assessment of previously undiagnosed infections.
Conclusions: Large numbers of new STDs continue to occur each year in the United States, with serious health and economic consequences. More than two thirds of our current estimate of 15 million STDs annually consists of two infections-trichomoniasis and human papillomavirus-for which we have only level III surveillance data. As the quality of our surveillance data improves, we can further refine the precision of our national estimates.
From Family Health International, Research Triangle Park, North Carolina
This supplement has been supported by the Kaiser Family Foundation, Menlo Park, CA. The information contained within the supplement was presented at a sponsored forum on Sexually Transmitted Diseases: Conspiracy to Innovate, Menlo Park, CA, July 10, 1998. The article has not undergone peer review and is not necessarily attributable to the publisher or sponsor.
The authors thank Cynthia Kay for her assistance in preparing the manuscript.
This paper is a revised version of a paper published in "Sexually Transmitted Diseases in America: How Many and at What Cost?" Kaiser Family Foundation, Menlo Park, CA, and the American Social Health Association, Research Triangle Park, NC, Dec. 1998.
ASHA Panel to Estimate STD Incidence and Cost: Gail A. Bolan, MD, California Department of Health Services; Virginia Caine, MD, Indiana University School of Medicine; Willard Cates, Jr., MD, MPH (Chair), Family Health International; Jacqueline Darroch, PhD, The Alan Guttmacher Institute; Edward W. Hook, III, MD, University of Alabama at Birmingham; James G. Kahn, MD, MPH, Institute for Health Policy Studies; William J. Kassler, MD, MPH, Centers for Disease Control and Prevention; Stephen Morse, PhD, Centers for Disease Control and Prevention; Michael E. St. Louis, MD, Centers for Disease Control and Prevention; James Trussell, PhD, Princeton University; and Anna Wald, MD, MPH, University of Washington.
American Social Health Association: Linda L. Alexander, PhD, FAAN, President; Joan R. Cates, Vice President of Development and Policy; Jennifer F. Ratcliffe, PhD, MSC, Epidemiologist; Nancy Herndon, Development Programs Manager.
Correspondence: Dr. Willard Cates, Jr., Family Health International, PO Box 13950, Research Triangle Park, NC 27709-3905. E-mail: email@example.com
Sexually Transmitted Diseases (STDs) remain a major public health problem in the United States.1 However, accurate estimates of the overall incidence and prevalence of STDs remain elusive. For the past decade, the most widely quoted figure has been 12 million sexually transmitted infections occurring annually.2-4 The scientific basis for this number has not been closely examined, and variants have been proposed recently. For example, the sum of specific STDs listed in the Institute of Medicine's 1994 report was 10.5 million cases,1 whereas a 1998 National Institutes of Health Program Announcement cited 14 million STD cases annually.5
Are 12 million cases of STDs still an accurate estimate for the United States today? Several developments suggest a readjustment may be in order. Because control programs directed against gonorrhea and syphilis have been successful recently, the incidence and prevalence of these infections has declined.6 Chlamydia control programs, which have emphasized increased screening, have led to a paradoxic (though predictable) situation where reported cases are increasing despite decreasing incidence and prevalence.6 Also, improved detection techniques have made us aware of the unrecognized extent of genital herpes, human papillomavirus, and trichomoniasis. This review examines the available published evidence to provide an updated point estimate and range for the incidence and prevalence of selected STDs in the United States.
Measures of STDs
Estimates of the incidence and prevalence of STDs in the United States vary according to the source of data and the methods used to detect infections.1,7 Sources generally include: 1) reportable infections (such as gonorrhea, syphilis, and chlamydia), 2) diagnoses made during visits to office-based practices, 3) national surveys of representative populations, 4) prevalence data on individuals attending specialized health facilities (such as STD clinics and family planning clinics), and 5) data from mathematical models of STD natural history.
Case reports for notifiable STDs for the United States are collected by the Centers for Disease Control and Prevention through formal surveillance systems based in the states.6 Reported data may vary in accuracy, depending upon the surveillance priorities of STD control programs. These data on notifiable infections tend to be more accurate in states that have laws requiring positive STD tests to be reported. Other data on infections can be derived from five key sources: 1) the National Health and Nutrition Examination Survey (NHANES), conducted by the National Center for Health Statistics (NCHS), which collects clinical and biologic data on a random sample of Americans; 2) the Hospital Discharge Survey of NCHS, which includes 7,500 randomly selected hospitals from throughout the United States; 3) the National Ambulatory Medical Care Survey of NCHS, which is a probability sampling of the diagnoses of 1,900 physicians; 4) the National Hospital Ambulatory Medical Care Survey of NCHS, which is a probability sampling of visits to hospital emergency and outpatient departments; and 5) the National Disease and Therapeutic Index (NDTI), which is a private survey of a random sample of office visits to American physicians in office-based practices.
Unfortunately, each of these sources has limitations. Data on reported infections are affected by differences in the completeness of reporting between public and private health care sources and by interstate variation. Because infections diagnosed in public facilities are reported more frequently, these data are susceptible to biases related to the characteristics of individuals who use public clinics. A cascading set of circumstances must occur for STDs to be measured accurately by public health authorities. For symptomatic infections, the symptoms must be initially perceived as abnormal by the individual and must be severe enough to cause the person to seek health care. The STD must then be diagnosed and, for reportable infections, it must be communicated to appropriate authorities by clinicians. For asymptomatic infections, screening programs must be available from health services routinely used by infected persons. Likewise, data from private clinicians' practices are often affected by the absence of diagnostic validation.
States differ markedly in the quality of their surveillance data on specific STDs. Most have ongoing systems to collect information on syphilis and gonorrhea; however, cutbacks on testing for the latter may have affected the consistency of reporting.8 Recently, the reported number of genital chlamydial infections has been increasing, in large part because a growing array of states has made genital chlamydia a reportable infection, and wider screening has identified more cases.9 Finally, data from specific health facilities are subject to patient selection bias and local geographic variation.
National surveys are limited by their size and sporadicity and by the superficial nature of their analytic variables. The national databases include relatively small numbers of STDs in the samples, which leads to wide confidence intervals in subpopulations. Most national surveys are conducted years apart, which makes timely interpretation of trends difficult. Several rely on self-reports for identifying a history of STDs, which is limited by a respondent's inability to recognize subclinical infection and reluctance to admit a stigmatic condition. Growing use of biomarkers as measures of current (urine LCR) or past (antibodies) STDs will reduce the problems of self-reported data.
Data on specific STDs also vary by the type of infection,6 depending on whether current or cumulative infection is being measured. Symptomatic viral infections (measured by physician visits) occur less frequently than serologic or cytologic indicators of the cumulative number of infected persons. Therefore, comparisons among the different estimates of STDs must be made carefully, and differences between incident and prevalent infections should be kept in mind.
Finally, the World Health Organization (WHO) has used a simple prevalence model to estimate the magnitude of curable STDs worldwide.10 First, the available information on STD prevalence in developed and developing countries was summarized. Then, the prevalence of gonorrhea, chlamydial infection, syphilis, and trichomoniasis was estimated by gender and by United Nations region. The 1995 regional "denominator" was calculated using mid-year population estimates of adults 15 to 49 years of age. Next, the duration of each curable infection was estimated and classified by gender and by region. These duration estimates were based on the probability that a symptomatic or asymptomatic person received treatment for his or her STD. Regional STD incidence in adults was then calculated by dividing the estimated prevalence by the estimated duration of each disease. Although based on broad assumptions, the WHO approach provides a standardized mechanism for global public health estimates. It also is the model we used to estimate for trichomoniasis in the United States.
Strength of Surveillance Evidence
Several approaches have been suggested to assess the quality and reliability of the estimates for specific STDs within the United States. One is based on a characterization of the quality, generalizability, and precision of the available data (Table 1); a second has listed estimation methods by source of data.7 We have chosen to use the first approach. Even though it relies heavily on qualitative assessments of the data sources, the concept can be used to give readers an idea of the level of confidence the panel has in its estimates of the incidence and prevalence of specific STD. We have categorized the strength of evidence and surveillance systems of STDs as "good," "fair," and "poor," or levels I, II, and III, respectively (Table 1).
Level I surveillance data, such as those for AIDS, come from representative national surveys such as NHANES or national reporting systems with nearly complete counts. Level II surveillance data, such as those for chlamydia and gonorrhea, are derived from composite prevalence figures obtained from multiple populations over time or from less complete national reporting systems. Finally, level III surveillance data, such as those for HPV and trichomoniasis, are based on even weaker evidence and rough extrapolations.
Epidemiology of STDs in the United States
In the United States, the incidence of reported genital chlamydial infections and viral STDs steadily increased in recent years, whereas the incidence of gonorrhea generally declined during the same interval. However, the actual number of chlamydial infections probably fell as control programs expanded. Levels of syphilis varied among different population subgroups, but they have reached record lows since 1995. Vaginal infections such as trichomoniasis and bacterial vaginosis have seemingly remained high, but surveillance for these conditions is rudimentary.
Genital chlamydial infections became the most prevalent bacterial STD in the United States during the 1980s, at the time gonorrhea levels began to decline. In 1996, nearly 500,000 cases of genital chlamydia were reported to the Centers for Disease Control and Prevention (CDC)-exceeding all other notifiable disease in the United States.11 Reported chlamydial infections in women greatly exceed those in men, primarily because screening programs have focused on women. In addition, chlamydial prevalence is strongly correlated with younger age and heterosexual behavior. A previous estimate of 4 million new chlamydial infections annually in the United States was made more than a decade ago, using a prevalence ratio.12 Because the expansion of chlamydia control programs has probably led to declining chlamydial prevalence in the interim,6 this estimate has been updated. In 1997, between 2.6 and 3.2 million new cases of genital chlamydia were estimated to have occurred in persons aged 10-44 years.13 We made a point estimate of 3 million new chlamydia infections having occurred in 1996 (Table 2).
Gonorrhea trends have been quite consistent since 1975. The number of reported gonorrhea cases has generally declined, starting in the mid-1970's with the introduction of the national gonorrhea control program. A disproportionate share of the decline occurred among older, white populations, with infection rates remaining relatively high among minority races and adolescents.8,14,15 In addition, reported gonorrhea is associated with a younger mean age among all gender and race categories than is syphilis. In 1996, CDC reported 325,900 new cases of gonorrhea.11 Because previous investigations have shown that about half of all diagnosed gonorrhea cases are reported to public health authorities, we estimated that a total of 650,000 new gonorrhea infections occurred in 1996 (Table 2).
Syphilis trends have fluctuated significantly during the last half-century. Its incidence rose during World War II and fell thereafter, coinciding with the introduction of penicillin. The lowest levels were observed at the end of the 1950s.16 However, a rapidly rising number of infected males coincided with the spread of the disease throughout the 1970s among men having sex with men. However, in the 1980s, indicative of the safer sexual behavior stimulated by HIV prevention messages, syphilis cases in gay males declined precipitously.17 This encouraging trend was directly countered by the climbing number of syphilis cases during the late 1980s among heterosexuals of minority races, in large part caused by the crack cocaine epidemic. Nonetheless, during the 1990s, syphilis levels again fell to numbers seen three decades earlier, leading public health authorities to entertain notions of syphilis elimination.18,19 In 1996, CDC reported 11,400 new cases of primary and secondary syphilis and 53,000 new cases of all stages of syphilis.11 Accounting for an estimated 20% underreporting, approximately 70,000 total syphilis infections in 1996 were estimated to have been diagnosed (Table 2).
The numbers of symptomatic genital herpes cases increased eleven fold during the 1970s and 1980s.6 Genital herpes causes at least ten times more genital ulcer cases than does syphilis. A comprehensive analysis of existing national databases estimated nearly 150,000 clinical visits for genital herpes in 1992.20 In addition, recent investigations have shown that symptomatic infections caused by herpes simplex viruses (HSV) tell only part of the story.21 Infection with HSV-2 has occurred in an estimated 45 million Americans, even though less than one quarter perceive themselves to have genital herpes. Based on differences between HSV-2 levels measured cross-sectionally in the late 1970s and the late 1980s, up to 1 million new HSV-2 infections may be transmitted each year in the United States (Table 2). This number ignores the sizable percentage of genital herpes contributed by HSV-1, and might therefore be considered a minimum estimate.
Likewise, the diagnosis of symptomatic genital warts caused by the human papilloma viruses (HPV) has skyrocketed during the last two decades.6 Its asymptomatic counterparts, HPV infections of the cervix and vagina, have emerged as the most common STD among young sexually active populations. An estimated 75% of the reproductive-age population have been infected with genital strains of HPV.22 The cumulative three-year incidence of HPV infection among college-aged students was 43 percent, with a duration of eight months.23 Making conservative assumptions and extrapolating these data to the general population in the United States, one obtains an annual estimate of at least 5.5 million new HPV infections (Table 2).24 Likewise, a conservative estimate of the prevalence of productive HPV (persons with active shedding of HPV DNA) is approximately 20 million (Table 2).22
The Hepatitis B virus (HBV), despite the availability of a preventive vaccine, still remains a prominent sexually transmitted viral infection. Based on data from CDC's Sentinel Counties Study of Viral Hepatitis, 58% of hepatitis B cases are spread sexually.25 In the early 1990s, CDC estimated that 335,000 HBV infections occurred annually in the United States.26 However, the HBV incidence has declined in the interim, probably due to wider availability and use of HBV vaccine.25 Based on CDC's estimates of 200,000 overall HBV infections in 1996, we estimate 120,000 incident sexually transmitted hepatitis B cases (Table 2). Based on serologic measures from NHANES-III, an estimated total of 1,250,000 prevalent cases of chronic hepatitis B exist in the United States.27 We therefore estimate a prevalence of approximately 750,000 currently infectious persons with sexually acquired HBV (Table 2).
Vaginal infections caused by Trichomonas vaginalis are among the most common conditions found in women attending reproductive health facilities. Between 3% and 48% of sexually active young women requesting routine care at prenatal, family planning, or college health clinics were diagnosed with trichomoniasis.28 The WHO estimated that this STD accounted for nearly half of all curable infections in the world. Based on WHO estimates for North America, we extrapolate 5 million cases of T. vaginalis infection annually in America (Table 2).
Bacterial vaginosis (BV), a sexually associated condition, is the most frequent cause of vaginitis in sexually active women of reproductive age.29 Depending on the population studied, the prevalence of BV in the United States varies from 17% in family planning settings to 37% among selected groups of pregnant women. In developing countries, BV is even more common in the general population, affecting approximately half of all women, including some with only one reported lifetime sex partner.30 The natural history of untreated BV remains a controversial research topic, although the presence of this condition has been linked to pelvic inflammatory disease and HIV acquisition. In addition, in high prevalence areas, BV tends to recur even after mass treatment of asymptomatic women and their partners.30 Because no established surveillance system exists for BV and no previous estimates of its incidence or prevalence have been made, we elected not to include this condition in the aggregate estimates we have derived.
Human Immunodeficiency Virus and AIDS
Trends in HIV infection in the United States have been evolving. Beginning in the mid-1970s, HIV was transmitted primarily among homosexual and bisexual men, and AIDS was first diagnosed in this group by the early to mid-1980's. The virus entered the injection drug-using (IDU) populations in the early 1980s and rapidly spread during the decade. Limited heterosexual transmission occurred until the late 1980s. However, since 1989, the greatest proportionate increase of reported AIDS cases has been among heterosexuals, and this trend is expected to continue.31 In 1993, an estimated 750,000 persons in the United States were infected with HIV, with approximately 40,000 new infections occurring each year.31 By 1996, another approach to estimating HIV incidence and prevalence yielded an estimate of 41,000 new HIV infections annually, with between 700,000 to 800,000 prevalent HIV infections (Table 2).32 The introduction of protease inhibitors may increase the number of prevalent infections by extending the life of HIV-infected people. Approximately half of the incident infections and three-quarters of prevalent infections were considered sexually transmitted. Therefore, it appears as if the incidence of HIV has been relatively stable over the past several years. Globally, the incidence of HIV is much higher, with an estimated 6 million new HIV infections annually and more than 33 million persons currently living with HIV.33 More than 90% of global cases have been spread sexually.33
Our updated estimate of the number of annual STD infections is somewhat higher than the earlier estimate of 12 million. The cumulative number of incident infections spread sexually is more than 15 million cases per year. However, nearly 70% of that total are contributed by two infections-trichomoniasis and HPV-for which we have only level III surveillance evidence. Therefore, this point estimate is not precise; the true number could be as low as 10 million or as high as 20 million STDs annually.
A variety of explanations can account for the larger number of STDs. As diagnostic sensitivity improved through use of amplification techniques, we gained a greater understanding of the magnitude of asymptomatic HPV infections. Our update of new trichomonal infections, albeit extrapolated from WHO methods, also raised the total. Finally, use of serologic data from NHANES allowed better estimates of asymptomatic acquisition of HSV and HBV. These increases were only partially offset by decreases in the number of chlamydia, gonorrhea, and syphilis infections resulting from recent successes in STD control programs.
These estimates of STD incidence and prevalence represent a snapshot in time, based on the available published data. The CDC has a goal of strengthening the evidence supporting its surveillance of each STD.7 For example, a pilot study of genital chlamydia from urine samples obtained from NHANES participants suggested that reliable national prevalence estimates could be obtained.34 As the methods to detect and measure the magnitude of STDs improve, so will the precision and accuracy of these estimates. Nonetheless, as a barometer of the STD burden in the late 1990s, and as a number upon which public policy can be based, the national estimate of 15 million new STD cases annually is a useful tool.
1. Eng TR, Butler WT, eds. The Hidden Epidemic: Confronting Sexually Transmitted Diseases. Washington, DC, National Academy Press, 1997.
2. American Social Health Association. STD News. Fall 1995.
3. Gunn RA, Rolfs RT, Greenspan JR, Seidman RL, Wasserheit JN. The changing paradigm of sexually transmitted disease control in the era of managed health care. JAMA 1998; 279:680-684.
4. Division of STD/HIV Prevention. Annual Report, 1987. Atlanta, GA: Centers for Disease Control, April 1988.
5. National Institute of Allergy and Infectious Diseases. Sexually Transmitted Diseases Cooperative Research Centers (RFA: AI-98-007). Bethesda, Maryland; April 1, 1998.
6. Division of STD Prevention. Sexually transmitted disease surveillance, 1997. Atlanta, Georgia: Centers for Disease Control and Prevention, September, 1998.
7. St. Louis ME, Koumans E, Garnett G. Methods for estimating the incidence and prevalence of STDs. In: Proceedings of the ICSTD, Seville, Spain, October 20, 1997. Abstract 0152.
8. Gershman KA, Rolfs RT. Diverging gonorrhea and syphilis trends in the 1980s: Are they real? Am J Public Health 1991; 81:1263-1267.
9. Centers for Disease Control and Prevention. Chlamydia trachomatis
genital infections-United States, 1995. Morbid Mortal Wkly Rep 1997; 46:193-198.
10. Gerbase AC, Rowley JT, Heymann DHL, Berkley SFB, Piot P. Global prevalence and incidence estimates of selected curable STDs. Sex Transm Infect 1998; 74(suppl 1):512-517.
11. Division of STD Prevention. Sexually transmitted disease surveillance, 1996. Atlanta, Georgia: Centers for Disease Control and Prevention, September, 1997.
12. Washington AE, Johnson RE, Sanders LL, Barnes RC, Alexander ER. Incidence of Chlamydia trachomatis infections in the United States: Using reported Neisseria gonorrhoeae as a surrogate. In: Oriel D, Ridgway G, Schachter J, et al., eds. Chlamydial infections. Cambridge, England: Cambridge University Press, 1986:487-490.
13. Groseclose SL, Zaidi AA, De Lisle SJ, Levine WC, St. Louis ME. An approach to estimation of chlamydia incidence and prevalence in the United States. In: Proceedings of the ICSTD, 1997 Seville, Spain, October 20, 1997. Abstract 0158.
14. Fox KK, Whittington WL, Levine WC, Moran JS, Zaidi AA, Nakashima AK. Gonorrhea in the United States, 1981-1996: Demographic and geographic trends. Sex Trans Dis 1998; 25:386-393.
15. Webster LA, Berman SM, Greenspan JR. Surveillance for gonorrhea and primary and secondary syphilis among adolescents, United States-1981-1991. MMWR 1993; 42(suppl 3):1-11.
16. Nakashima AK, Rolfs RT, Flock ML, Kilmarx P, Greenspan JR. Epidemiology of syphilis in the United States, 1941 through 1993. Sex Transm Dis 1996; 23:16-23.
17. Rolfs RT, Nakashima AK. Epidemiology of primary/secondary syphilis in the United States, 1988 JAMA 1990; 264:1432-1437.
18. Centers for Disease Control and Prevention. Primary and secondary syphilis-United States, 1997. Morb Mort Wkly Rpt 1998; 47:493-497.
19. St. Louis ME, Wasserheit JN. Elimination of syphilis in the United States. Science 1998; 281:353-354.
20. Tao G, Kassler WJ, Rein DB. Medical care expenditures for genital herpes in the United States. Unpublished data.
21. Fleming DT, McQuillan GM, Johnson RE, et al. Herpes simplex virus type 2 in the United States, 1976 to 1994. N Engl J Med 1997; 337:1105-1111.
22. Koutsky L. Epidemiology of genital human papilloma virus infection. Am J Med 1997; 102(suppl 5A):3-8.
23. Ho GYF, Bierman R, Beardsley L, Chang CJ, Burk RD. Natural history of cervicovaginal papilloma virus infection in young women. N Engl J Med 1998; 338:423-428.
24. Ratcliffe JM. Estimation of HPV incidence in the US population. Report for ASHA, May 23, 1998.
25. Mast EE, Mahoney FJ, Alter MJ, Margolis HS. Progress toward elimination of hepatitis B virus transmission in the United States. Vaccine 1998; (suppl):S48-S51.
26. Coleman PJ, McQuillan GM, Moyer LA, Lambert SB, Margolis HS. Incidence of hepatitis B infection in the United States, 1976-1994: Estimates from the National Health and Nutrition Examination Surveys. J Infect Dis 1998; 178:954-959.
27. Sabin K. Surveillance of hepatitis B in the US. [Centers for Disease Control Website]. Available at: http://www.cdc.gov/ncidad/diseases/hepatitis. Accessed June 3, 1998.
28. Cotch MF, Pastorek JG, Nugent RG, et al. Trichomonas vaginalis associated with low birth weight and preterm delivery. Sex Transm Dis 1997; 24:353-360.
29. Sobel JD. Vaginitis. N Engl J Med 1997; 337:1896-1903.
30. Wawer MJ, Sewankambo NK, Serwadda D, et al. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomized community trial. Lancet 1999; 353:525-535.
31. Rosenberg PS. Scope of the AIDS epidemic in the United States. Science 1995; 270:1372-1375.
32. Holmberg SD. The estimated prevalence and incidence of HIV in 96 large U.S. metropolitan areas. Am J Public Health 1996; 86:642-654.
33. UNAIDS. AIDS Epidemic Update: December 1998. Geneva, Switzerland: UNAIDS, 1998.
34. Mertz KJ, McQuillan GM, Levine WC, et. al. A pilot study of the prevalence of chlamydial infection in a national household survey. Sex Transm Dis 1998; 25:225-228.
35. Zaidi AA. Reliability of estimates for STD surveillance. Presented at Meeting on Estimates of STD Incidence and Prevalence, Atlanta, Georgia, October 30, 1996.