Datta, S. Deblina MD*; Torrone, Elizabeth PhD, MSPH*,†; Kruszon-Moran, Deanna ScM‡; Berman, Stuart MD, ScM*; Johnson, Robert MD, MPH*; Satterwhite, Catherine L. PhD, MSPH, MPH*; Papp, John PhD*; Weinstock, Hillard MD, MPH*
From the *National Center for HIV/AIDS, Viral Hepatitis, STD and TB Prevention, Centers for Disease Control and Prevention, Atlanta, GA; †Epidemic Intelligence Service, Office of Scientific Education and Professional Development Program (proposed), Centers for Disease Control and Prevention, Atlanta, GA; and ‡National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, MD
S.D.D. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis; conception and design; acquisition of data; analysis and interpretation of data; drafting of manuscript; critical revision of manuscript for important intellectual content; statistical analysis; obtaining funding; administrative, technical, or material support; supervision; E.T.: analysis and interpretation of data; critical revision of manuscript for important intellectual content; statistical analysis; D.K.M: acquisition of data; analysis and interpretation of data; drafting of manuscript; critical revision of manuscript for important intellectual content; statistical analysis; administrative, technical, or material support; S.B.: conception and design; analysis and interpretation of data; drafting of manuscript; critical revision of manuscript for important intellectual content; obtaining funding; administrative, technical, or material support; supervision; R.J.: conception and design; acquisition of data; analysis and interpretation of data; critical revision of manuscript for important intellectual content; C.L.S.: analysis and interpretation of data; critical revision of manuscript for important intellectual content; J.P.: acquisition of data; analysis and interpretation of data; critical revision of manuscript for important intellectual content; supervision; H.W.: analysis and interpretation of data; critical revision of manuscript for important intellectual content; obtaining funding; supervision.
The authors have no financial conflicts of interest to declare: all are full-time government researchers.
Supported by the Centers for Disease Control and Prevention, Atlanta, GA. CDC researchers are responsible for design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
Received for publication April 5, 2011, and accepted August 4, 2011.
Correspondence: S. Deblina Datta, MD, 1600 Clifton Rd, MS E-05, Atlanta, GA 30333. E-mail: firstname.lastname@example.org.
To reduce Chlamydia trachomatis (C. trachomatis, chlamydia)-related morbidity and sequelae such as pelvic inflammatory disease, ectopic pregnancy, chronic pelvic pain, and infertility, the Centers for Disease Control and Prevention (CDC) has issued recommendations for routine chlamydia screening since 1993.1,2 Current CDC guidelines call for routine screening of all sexually active females aged <26 years and screening of women aged ≥26 years with risk factors.3 Other organizations also recommend routine chlamydia screening of women (US Preventive Services Task Force, American Medical Association, American Academy of Pediatrics, American Association of Family Physicians, American College of Obstetricians and Gynecologists, American College of Preventive Medicine), and the proportion of sexually active females aged 16 to 25 years receiving annual chlamydia screening in managed care organizations has been a HEDIS (Healthcare Effectiveness Data and Information Set) measure for managed care providers since 2000.4
The impact of prevention strategies, such as routine screening and safe sex education initiatives, on national trends in chlamydia is difficult to assess. National data from chlamydia case reports and estimates of chlamydia positivity measured in family planning clinics participating in the Infertility Prevention Project (IPP) are used to monitor chlamydia burden; however, both surveillance methodologies suffer from considerable limitations and biases including changing test technologies, changing populations under study, and nonstandardized screening and testing protocols, making interpretation of temporal trends difficult.5
We describe trends in chlamydia prevalence estimates from representative samples of the civilian, noninstitutionalized US population aged 14 to 39 years from the National Health and Nutrition Examination Survey (NHANES), from 1999 to 2008, using data from the five 2-year survey cycles. These estimates are derived from survey data using consistent and representative sampling methods over time, allowing evaluation of trends.
The NHANES is a series of cross-sectional surveys designed to provide national statistics on the health and nutritional status of the general household population through household interviews and standardized physical examinations with collection of biologic samples in special mobile examination centers. In 1999, the NHANES became a continuous survey, with data released in 2-year cycles. The sampling plan of the survey is a stratified, multistage, probability cluster design that selects a sample representative of the US civilian, noninstitutionalized population. Participants were asked to self-identify race and ethnicity, sex, and age. Non-Hispanic blacks, Mexican Americans, adolescents (oversampled from 1999–2006 only), and low-income persons were sampled at higher frequencies than other persons to improve stability of estimates for these subpopulations.
All participants aged ≥18 years provided written informed consent. For minors (age <18 years), parents gave written consent, accompanied by the minor's assent. Protocols for test result notifications have been described in a previous report.6 The study protocol was reviewed and approved by an institutional review board at CDC.
Urine specimens were collected from participants aged 14 to 39 years. Specimens were processed in the examination centers and shipped to the CDC for C. trachomatis testing using nucleic acid amplification tests. The Abbott LCx assay (Abbott Laboratories, Abbott Park, IL) was used to test samples from the two 2-year cycles from 1999 to 2002 and a different nucleic acid amplification test, the Becton Dickinson ProbeTec (BD, Franklin Lakes, NJ) was used to test samples from 2003 to 2008 because manufacture of the Abbott LCx was discontinued.6 Both assays were run according to the manufacturers' instructions. No samples were tested with both assays.
Prevalence estimates were calculated using weights to incorporate the complex sample design. The weights were constructed to account for the unequal probability of selection and nonresponse to the household interview and physical examination. The weights were further ratio adjusted by age, sex, and race/ethnicity to the US population control estimates from the Current Population Survey (US Census Bureau) adjusted for undercounts. Ninety-five percent confidence intervals (CIs) for the prevalence estimates for each subgroup were calculated using a logit transformation based on a t-statistic using the corrected number of degrees of freedom calculated by subtracting the number of primary sampling units from the number of strata represented in the subgroup of interest. Estimates with relative standard errors (RSEs) of greater than or equal to 30%, greater than or equal to 40%, or based on fewer than 10 positive persons were noted and are considered unstable and should be interpreted with caution.7 Estimates with an RSE greater than 50% are considered the least stable and are reported with exact RSE and number of persons testing positive. Estimates and standard errors were calculated using SUDAAN (Research Triangle Institute, Research Triangle Park, NC), a statistical package that accounts for the complex sample design. All other calculations were completed in SAS v 9.2 (Cary, NC).
Prevalence estimates are reported stratified by age, gender, and race/ethnicity. Individual logistic regression models including only survey cycle (time) as a continuous variable were run on the total population and each subgroup. The average percent change in chlamydia prevalence and corresponding 95% CI over the five 2-year cycles for each subgroup was determined using the odds ratio (and its 95% CI) from each individual model with the assumption that the odds ratio will approximate the prevalence ratio given lower disease prevalence over one survey cycle. The odds ratio was multiplied by itself 4 times to estimate the reduction over the five cycles, subtracted from 1, and converted to a percent to estimate the percent change in prevalence. Changes in prevalence whose 95% CI did not include zero were considered statistically significant. The estimated percent change assumed a linear relationship over the 5 survey cycles and the true relationship may be nonlinear.
To provide more in-depth results for the most current prevalence estimates, differences in prevalence between subgroups in the 2007–2008 cycle were evaluated using a univariate t statistic obtained from a general linear contrast in SUDAAN.8 No corrections for multiple comparisons were made.
Among the 20,836 people aged 14 to 39 years originally selected to participate in NHANES 1999–2008, 17,190 (83%) were interviewed in the home, and 16,481 (96% of those interviewed) were examined. Of these, 15,885 (96% of those examined) had results of C. trachomatis testing.9
In the 2007–2008 cycle, the overall chlamydia prevalence was 1.6% (95% CI: 1.1%–2.4%) (Table 1). Prevalence was higher among females (2.2%, 95% CI: 1.4%–3.4%) than males (1.1%, 95% CI: 0.7%–1.7%) (P < 0.05). Prevalence was higher among adolescents (2.5%, 95% CI: 1.6%–3.8%) and persons aged 20 to 25 years (2.8%, 95% CI: 1.7%–4.7%) than persons aged 26 years or older (0.8%, 95% CI: 0.3%–1.8%) (P < 0.05). Prevalence among non-Hispanic black persons (6.7%, 95% CI: 4.6%–9.9%) was higher than the prevalence among non-Hispanic white persons (0.3%, 95% CI: 0.1%–1.2%) (P < 0.05) and Mexican American persons (2.4%, 95% CI: 1.5%–4.0%) (P < 0.05).
The prevalence of chlamydia decreased over the five 2-year cycles (Table 1). On the basis of regression analysis, there was a 40% reduction (95% CI: 8%–61%) in chlamydia prevalence overall (P < 0.05). Although an estimated reduction in prevalence was found for each gender, age, and race/ethnicity subgroup except non-Hispanic blacks, the relative size, precision, and statistical significance of the estimated reduction varied (Table 1). Decreases in prevalence were most notable in men (53% reduction, 95% CI: 19%–72%), in adolescents aged 14 to 19 years (48% reduction, 95% CI: 11%–70%), and in non-Hispanic white persons (77% reduction, 95% CI: 48%–91%). There was no change in prevalence among non-Hispanic black persons. The prevalence estimates for non-Hispanic white persons have high RSEs (>40% in 2001–2002 and 60% in 2007–2008) and are based on a small number of positive sample persons (n = 3 in 2007–2008) so results for this subgroup should be interpreted with caution (Table 1).
As chlamydia prevalence is highest among adolescents and adolescents are oversampled, trends in prevalence by race/ethnicity and by gender were analyzed in this age subgroup. Among 14- to 19-year-old males, prevalence decreased significantly (70% reduction, 95% CI: 22%–89%) over the 5 survey cycles, primarily because of a decrease from 1999–2000 to 2001–2002. (Fig. 1) Prevalence among adolescent females appeared to be decreasing through 2005–2006, but increased in the last survey cycle (31% overall reduction, 95% CI: −36%–65%). Among 14- to 19-year-old non-Hispanic blacks, prevalence decreased 45% (95% CI: 4%–70%); however, prevalence in the 2007–2008 cycle continued to be higher than prevalence among non-Hispanic whites (P < 0.05) (Fig. 2). The estimate for 14- to 19-year-old non-Hispanic whites in 2007–2008 is based on only 2 positive sample persons with RSE >60% so results should be interpreted with caution.
The prevalence among 14- to 25-year-old females, the targeted population for screening, was 4.1% (95% CI: 2.4%–6.8%) in 1999–2000; 2.8% (95% CI: 1.8%–4.5%) in 2001–2002; 4.3% (95% CI: 2.7%–6.7%) in 2003–2004; 1.8% (95% CI: 1.1%–2.9%) in 2005–2006; 3.8% (95% CI: 2.4%–6.0%) in 2007–2008. The estimated percent change over the 5 survey cycles was a 19% reduction (95% CI: −57–57%, P = 0.54).
As a sensitivity analysis to our temporal trend analysis using data in 2-year cycles, we also investigated changes in prevalence by aggregating survey cycles to increase sample size and compared estimates from 1999–2002 with 2003–2008 (4 years vs. 6 years), as well as from 1999–2002 with 2004–2008 (4 years vs. 4 years), for each subgroup (results not shown). All reported findings from the trend analysis of the five 2-year cycles were similar to those resulting from analyses of the aggregated data.
This analysis represents the first population-based assessment of trends in chlamydia prevalence in the U.S. Data were gathered using a consistent and well-established methodology, over consecutive time intervals. The overall chlamydia prevalence decreased approximately 40% over the 10-year time period with current prevalence among 14- to 39-year-olds estimated to be less than 2%. Prevalence decreased among 14- to 19-year-olds, particularly among non-Hispanic blacks; however, racial disparities persist. Among 14- to 25-year-old females, the group targeted by screening recommendations, prevalence remained at 3.8% in the 2007–2008 survey cycle.
Findings from this analysis of prevalence trends contrast with two other continuously monitored indices of chlamydia burden in the United States, national case reporting and chlamydia positivity data from the IPP.10 Both case report rates and chlamydia positivity among women aged 15 to 24 years attending IPP family planning clinics have increased over the past decade. However, case report and IPP data have limitations that compromise their reliability for assessing trends in chlamydia burden. The consistent increases in chlamydia case rates over the past 20 years are likely attributed to ongoing increases in chlamydia screening and improvements in test technology.10 Chlamydia positivity data from IPP are collected among women seeking healthcare in family planning clinics10 and the characteristics of the population served may change over time.5 IPP data lack standardization regarding clinic participation, test technology used, and protocols for screening and testing. Such limitations, and the resulting biases, have been described by Miller, who concluded that the most reliable information on trends in chlamydia should be assessed using population-based surveys to monitor prevalence.5
NHANES estimates are based on a population-based probability sample. Sampling methods, test assays, and testing protocols are standardized and provide valid prevalence estimates which can be used for monitoring trends in disease burden. However, the NHANES sample was not designed to reliably estimate conditions of low prevalence for each 2-year cycle by all demographic subgroups. Estimates of low prevalence may not be precise or stable (e.g., when RSE >30% or when the number of positives are <10 sample persons); thus, we had limited statistical power to report prevalence and investigate trends in all subpopulations.7
One additional methodologic consideration during evaluation of these data across the 10-year time period is the change in assay used. LCx was used from 1999 to 2002 and Probe Tec from 2003 to 2008. Although no published studies present data directly comparing the test performance characteristics of these assays, one study compared each with a third standard, a combination of culture and nucleic acid amplification tests.11 Van der Pol and colleagues report that with urine specimens, Probe-Tec had similar sensitivity and specificity as LCx among females and similar sensitivity but slightly lower specificity than LCx among males. Since prevalence appeared to decrease overall, it is unlikely that differential test performance alone would account for the observed decrease in prevalence in NHANES. Furthermore, had the assay change been the primary reason for the observed decrease in prevalence, such changes would be expected to be equally distributed across different age and racial/ethnic groups. However, we observed disproportionate changes in prevalence over time, which were pronounced across racial/ethnic groups.
We document a decreasing trend in chlamydia prevalence over the 10-year time period, particularly among adolescents. During this time, the US chlamydia control program focused on increasing screening and treatment of asymptomatic females as well as the treatment of sex partners of females and males with chlamydia. The percentage of sexually active females enrolled in commercial and Medicaid healthcare plans aged 16 to 25 years who were screened for chlamydia increased from 25.4% in 2000 to 44.7% in 2008.4 Decreases in prevalence in NHANES coincided with these increases in opportunistic screening. While only ecologically linked, increases in prevention efforts, such as chlamydia screening and subsequent treatment, may have contributed to the observed decline in prevalence. In addition, overall trends in NHANES are consistent with declines in prevalence observed among socioeconomically disadvantaged young men and women who entered the National Job Training Program. From 1998 to 2007, there was a decrease in chlamydia prevalence among both males and females screened in the National Job Training Program, after controlling for testing protocols that changed to use more sensitive assays.12,13 These decreases were consistent across age, race/ethnicity, and geographic strata.
Although we document overall decreases in prevalence, we did not observe decreases in females aged 14 to 25 years, the population targeted for annual screening. It is possible that statistical power was too limited as a result of low prevalence to detect a true change in prevalence over time in the trend analysis. It is also possible that prevalence in this population has remained stable because of inadequate or ineffective control strategies. Although screening coverage in the United States has increased over the last decade, it is still suboptimal. In the National Survey of Adolescent Health, only 27% of sexually active females aged 18 to 26 years reported having a chlamydia test in the past 12 months.14 In the National Hospital Ambulatory Medical Care Survey, only 16% of preventive visits by asymptomatic women aged 15 to 25 years included chlamydia screening.15 The latest HEDIS data indicate that only 45% of sexually active women aged 16 to 26 years were screened in 2008.4 In addition, it has been suggested that screening programs may not be effective at reducing chlamydia prevalence16 and that the impact of screening on prevalence may be muted because of arrested development of immunity with treatment.17 Additional control strategies, including efforts to increase adherence to rescreening guidelines and partner treatment to prevent reinfection, may be needed to reduce population prevalence.
Of additional concern are the increasing racial disparities in chlamydia prevalence. As prevalence decreased among non-Hispanic whites aged 14 to 39 years and did not change among non-Hispanic blacks aged 14 to 39 years, both absolute and relative racial disparities in prevalence increased from 1999–2000 to 2007–2008. Reducing racial disparities in STDs is an important objective.18 The disparity associated with chlamydia case reports is the second highest (after gonorrhea) of any nationally notifiable disease (excluding malaria with most cases being imported from outside the United States) in the United States.19 Our findings show that chlamydia prevalence is also highest in non-Hispanic blacks. Although we document a reduction in prevalence among non-Hispanic black adolescents, it is clear that racial disparities in chlamydia infection still exist.
In summary, it appears the overall chlamydia burden in the United States decreased from 1999 to 2008. This conclusion is based on population estimates of chlamydia prevalence which are not subject to the biases associated with surveillance relying on facility-based data sources. Given the limitations of surveillance data for tracking chlamydia trends, continued monitoring of population-level trends in prevalence is warranted; however, unstable estimates resulting from low disease prevalence and small sample sizes necessitates caution when interpreting trends. Our findings suggest that existing activities addressing chlamydia prevention may be on the right track, but there remains a need to reduce prevalence in populations at greatest risk as well as to reduce racial disparities.
1. Holmes KK, Sparling PF, Stamm WE, et al., eds. Sexually Transmitted Diseases. 4th ed. McGraw Hill, 2007.
2. Recommendations for the prevention and management of Chlamydia trachomatis infections, 1993. Morb Mort Wkly Rep 1993; 42(RR-12).
3. Workowski K, Berman S. Sexually transmitted diseases treatment guidelines, 2006. MMWR Recomm Rep 2006; 55.
4. Centers for Disease Control and Prevention (CDC). Percentage of sexually active female enrollees aged 16–25 years* who were screened for Chlamydia trachomatis
infection by state and year – Healthcare Effectiveness Data and Information Set (HEDIS), United States, 2000–2008. Available online at: http://www.cdc.gov/std/chlamydia/female-enrollees-00–08.htm
. Accessed February 17, 2011.
5. Miller W. Epidemiology of chlamydial infection: are we losing ground? Sex Transm Infect 2008; 84: 82.
6. Datta S, Sternberg M, Johnson R, et al.. Gonorrhea and chlamydia in the United States among persons 14 to 39 years of age, 1999 to 2002. Ann Intern Med 2007; 147: 89.
10. Centers for Disease Control and Prevention. Sexually Transmitted Disease Surveillance 2008. Atlanta: U.S. Department of Health and Human Services; 2009.
11. Van Der Pol B, Ferrero D, Buck-Barrington L, et al.. Multicenter evaluation of the BDProbeTec ET System for detection of Chlamydia trachomatis and Neisseria gonorrhoeae in urine specimens, female endocervical swabs, and male urethral swabs. J Clin Microbiol 2001; 39: 1008.
12. Joesoef M, Mosure D. Prevalence trends in chlamydial infections among young women entering the National Job Training Program, 1998–2004. Sex Transm Dis 2006; 33: 571.
13. Satterwhite C, Tian L, Braxton J, et al.. Chlamydia Prevalence Among Women and Men Entering the National Job Training Program: United States, 2003–2007. Sex Transm Dis 2009; 36: 1.
14. Nguyen TQ, Ford CA, Kaufman JS, et al.. Infrequent chlamydial testing among young adults: Financial and regional differences. Sex Transm Dis 2008; 35: 725–730.
15. Hoover K, Tao G, Kent C. Low rates of both asymptomatic chlamydia screening and diagnostic testing of women in US outpatient clinics. Obstet Gynecol 2008; 112: 891–898.
16. Low N, Bender N, Nartey L, et al.. Effectiveness of chlamydia screening: Systematic review. Int J Epidemiol 2009; 38: 435–448.
17. Brunham RC, Pourbohloul B, Mak S, et al.. The unexpected impact of a Chlamydia trachomatis infection control program on susceptibility to reinfection. J Infect Dis 2005; 192: 1836–1844.
18. Newman L, Berman S. Epidemiology of STD disparities in African Am communities. Sex Transm Dis 2008; 35: S4.
19. Racial disparities in nationally notifiable diseases—United States, 2002. MMWR Morb Mortal Wkly Rep 2005; 54: 9–11.