Secondary Logo

Journal Logo

Original Studies

Biological and Behavioral Factors Associated With Positive Chlamydia Retests

Smith, Kirsty S. MPH*; Guy, Rebecca PhD*; Danielewski, Jennifer PhD†‡; Tabrizi, Sepehr N. PhD†‡§; Fairley, Christopher K. PhD¶∥; McNulty, Anna M. MMed**††; Rawlinson, William PhD‡‡§§; Saville, Marion MBChB§¶¶; Garland, Suzanne M. MD†‡§; Donovan, Basil MD*; Kaldor, John M. PhD*; Hocking, Jane S. PhD∥∥; on behalf of the REACT investigators

Author Information
Sexually Transmitted Diseases: July 2017 - Volume 44 - Issue 7 - p 417-422
doi: 10.1097/OLQ.0000000000000616
  • Free

Anogenital Chlamydia trachomatis infection (chlamydia) is the most commonly notified sexually transmissible infection (STI) in high income countries.1–3 Among sexually active 15 to 24 year olds, population-based prevalence is estimated to be 2% to 6%.4–7 High prevalence has also been reported in men who have sex with men (MSM), at 3% to 6% for urethral infections and 5 % to 9% for rectal infections.8

Repeat detection is common after treatment of chlamydia. A systematic review found the overall median proportion of females retesting positive was 13.9%.9 Repeat infections also appear common in MSM.10 Repeat infection is associated with increased risk of chlamydia-related sequelae, such as pelvic inflammatory disease and infertility in women,11 and in MSM, observational data have shown that 2 or more past rectal chlamydia or gonorrhoea infections were associated with HIV seroconversion.12

Single-dose azithromycin has been the treatment of choice for chlamydia for over 20 years; however, treatment failure rates of 5% to 14% in genital infection and 6% to 21% in asymptomatic rectal infection have recently been reported.13 The reasons for azithromycin treatment failure are not known. A cohort study of young women found some evidence that higher organism loads predicted treatment failure14 and a recent study in MSM detected an association between high rectal chlamydia organism load and repeat positivity.15

In 2011 to 2013, we conducted REACT, a randomized trial to determine if a postal home-collection kit plus short-message-service reminder at 3 months would increase the retesting rate in patients treated for chlamydia.16 In the context of this trial, we examined repeat chlamydia and investigated behavioral and organism factors (organism load and genovar) associated with repeat positive tests. We also used genotyping and behavioral data to differentiate between treatment failure and reinfection.

MATERIALS AND METHODS

Ethics Statement

The study was approved by the Human Research Ethics Committees of Alfred Health, South Eastern Sydney and Illawarra Area Health Service, and the University of New South Wales. All participants provided consent to participate in this study.

Study Populations and Procedures

The eligibility and exclusion criteria and consent process for the REACT trial have been described previously.16 It was conducted through 2 metropolitan public sexual health clinics in Australia and included a sample size of 600: 200 each of women, heterosexual men, and MSM.

Baseline and retest specimens collected at the study clinics were tested with either the Cobas 4800 duplex real-time polymerase chain reaction (PCR) for C. trachomatis and Neisseria gonorrhoeae, (Roche Diagnostics, Mannheim, Germany); or the BDProbeTec ET C. trachomatis and N. gonorrhoeae Amplified DNA Assay (BD Biosciences, Sparks, MD). Specimens collected at home were tested using the Artus C. trachomatis Plus RG PCR kit (Qiagen, Hilden, Germany). Chlamydia positive samples were stored by laboratories at −80°C for research purposes.

All REACT participants had confirmed chlamydia infection at baseline and were prescribed a single 1-g dose of azithromycin as per routine practice at that time.17 Participants were then sent a short-message-service reminder at 3 months for repeat testing. Half of the participants (stratified into three categories: heterosexual men, women, and MSM) were randomly selected to be mailed a home collection kit for self-sampling, whereas the other half were reminded to retest at the clinic. Those in the home testing arm could also retest at the clinic if they preferred. Samples in both arms included urine for heterosexual men, urine and rectal swabs for MSM, and vaginal/cervical swabs or urine for women.

Data Collection

Data on testing, treatment, and baseline characteristics (age, country of birth, previous chlamydia diagnoses, condom use, number of partners in the previous 3 months, and STI symptoms) were extracted from the patient management system for each participant. Participants were also asked to undertake an online survey 4 months after enrolment, capturing treatment and sexual behavior data (if sexually active and if so, condom use) since the baseline chlamydia diagnosis.

Laboratory Testing and Interpretation

Chlamydia Testing

Stored samples were sent to the Department of Microbiology and Infectious Diseases at the Royal Women’s Hospital, Melbourne, Australia, for nucleic acid extraction, quantification of chlamydial load and genovar determination. The automated MagNA Pure 96 system (Roche Applied Science, Mannheim, Germany) was used to extract nucleic acid from one set of samples. For urines or Uriswabs 1 mL was spun at 13,000g for 15 minutes, the supernatant removed and the pellet resuspended in 200 μL of PBS. Swabs were agitated into 500 μL of PBS. 200 μL was then taken to extract nucleic acid on the MagNA Pure 96 system according to the manufacturer's instructions. The total nucleic acid was eluted in MagNA Pure 96 elution buffer in a final volume of 100 μL. Nucleic acid for urine, vaginal or rectal swabs placed into Cobas media at the time of collection, was isolated using the Roche Cobas 4800 CT/NG assay (Roche Molecular Diagnostics, Branchburg, NJ), according to the manufacturer's protocol. The Cobas platform removes 400 μL of sample (swabs) or 850 μL of sample (urine) for extraction, which was subsequently eluted into a final volume of 100 μL of elution buffer.

Quantification of Organism Load and Genovar Determination

Quantitative real-time PCR targeting the ompA gene was performed18 to estimate C. trachomatis load and classify positive samples into one of three broad phylogenetic groups. C. trachomatis load was quantified by comparing the crossing-threshold of each sample to the crossing-threshold of a standard curve constructed by amplifying a range of known copy numbers of the ompA gene. In addition, quantitative PCR of the human β-globin gene was used to quantify eukaryotic cell DNA copy number, allowing for variation in load estimates due to specimen type as described previously.14 The chlamydial genomic copy number was divided by the eukaryotic cellular DNA copy number and expressed as the number of organisms present per 100 eukaryotic cells.14,19 Results were calculated in copies/mL for urine or copies/swab for dry swabs. A second set of real-time PCRs was then used to determine specific genovars18 for each sample and determine whether any mixed infections were present.

Multilocus Sequence Typing Analysis

For individuals who had 2 consecutive samples with the same chlamydia genovar, multilocus sequence typing (MLST) was used in conjunction with behavioral data to help differentiate treatment failure from reinfection. These specimens were amplified for 5 of the most variable regions of the C. trachomatis genome; hctB, CT058, CT144, CT172, pbpB according to the current protocol.20 The PCR amplicons for each region were assessed and then sent to the Australian Genome Research Facility for Sanger Sequencing. Sequence analysis was performed using the CLC Main Workbench (version 7; CLC Bio, Aarhus, Denmark). The Uppsala University C trachomatis MLST database (http://mlstdb.bmc.uu.se) was used to perform allele and sequence type queries to assess whether the 2 positive specimens represent the same or different C. trachomatis strains.

Classification of Repeat Positive Cases

To differentiate between chlamydia reinfection and treatment failure, a modified version of an algorithm developed by Batteiger et al21 and adapted by Walker et al.14 was used (Fig. 1).

Figure 1
Figure 1:
Algorithm to differentiate between reinfection and treatment failure. *Of the 12 individuals with incomplete genovar data, 5 were classified as reinfections: 3 on the basis of self-reporting condomless sex, and 2 on the basis of a positive retest at a different anatomical site to the baseline infection (applies to MSM only); and 7 cases could not be classified as genovar data were not available. #Of these, 2 cases were differentiated by MLST.

A reinfection was defined by: a positive retest with a different genovar; or if there was a self-reported history of condomless sex since the baseline infection regardless of genovar; or in MSM, if the second infection was detected at a different anatomical site. Treatment failure was defined by a positive retest with the same genovar, and a history of no sex or consistent condom use since baseline. A persisting infection was defined as a positive retest with the same genovar without documented treatment between the 2 tests.

Statistical Analyses

Descriptive statistics were used to describe the characteristics of participants. The proportion and 95% confidence intervals (CI) of participants with a positive retest were calculated using exact binomial methods. Organism load estimates were log transformed for analysis. Univariate and multivariate logistic regressions were used to investigate factors associated with retest positivity including organism load, patient demographics, sexual behavior, past chlamydia diagnoses and presence of symptoms. Univariate logistic regression was used to investigate the association of organism load with treatment failure or reinfection.

Variables included in our multivariate models were selected on the basis of clinical relevance and the likelihood ratio test. Box plots were generated to compare distributions in organism load between different groups. χ2 tests were used to examine the associations between categorical variables and Mann-Whitney U or Kruskall-Wallis tests were used to investigate associations between categorical and continuous variables. Analyses were performed using STATA (version 13.0; StataCorp, College Station, Tex).

RESULTS

A total of 290 participants (100 women, 89 heterosexual men and 101 MSM) were retested between 1 and 4 months after their initial diagnosis and had baseline specimens stored for genovar and organism load testing. There was no difference in age, risk group or number of sex partners last 3 months between those who retested and those who did not. Of those who retested, baseline genovar was determined for 262 (87 [87%] women, 86 [97%] heterosexual men and 89 [88%] MSM), and baseline organism load was quantified for 266 individuals (89 [89%] women, 87 [98%] heterosexual men and 90 [89%] MSM). Urine specimens were provided by 125 men and 41 women, urethral swabs by 2 men, vaginal/cervical swabs by 59 women, and rectal swabs by 63 men.

Sample Characteristics of Participants Who Retested Between 1 and 4 Months

The median age of participants was 27 years (interquartile range [IQR], 23–31 years): women, 24 years (IQR, 21–27 years); heterosexual men, 27 (IQR, 24–31 years); and MSM, 30 years (IQR, 27–37 years) (P < 0.01). There was no difference in time to retest between patient groups (Table 1).

TABLE 1
TABLE 1:
Characteristics of Participants Who Retested Between 1 and 4 Months at Recruitment

Baseline Genovar and MLST Results

The most common genovar among heterosexuals was E (49%) followed by F (20%), whereas for MSM, it was G (35%), D (29%), or J (20%). There was a significant difference in the genovar distribution between MSM and heterosexual men and women (P < 0.01) but no difference between heterosexual men and women (P = 0.47) (Table 1). Among MSM, in rectal swabs, genovars G and D were most commonly detected (32.8% each), whereas there was a greater diversity of genovars in urine specimens, with genovar G most commonly detected (29.0%).

In a total of 43 participants, 14.8% (95% CI, 10.9%–19.4%) retested positive for chlamydia 1 to 4 months after their initial diagnosis. Repeat positivity was 10.0% (95% CI, 4.9%–17.6%) for women, 13.5% (95% CI, 7.2%–22.3%) for heterosexual men, and 20.8% (95% CI, 13.3%–30.0%) for MSM. There was no difference in repeat positivity between heterosexual men and women (P = 0.46), but it was significantly higher for MSM compared with heterosexual men and women (20.8% vs 11.6%, P = 0.04).

Among the 43 individuals with a repeat positive test, both baseline and retest samples for 31 (8 women, 7 heterosexual men and 16 MSM) were available for genotyping. The majority of missing specimens were not stored and for 3 men, genovar could not be identified on retesting. Of these 31 pairs, 27 (87.1%) presented with an identical genovar (7 women, 6 heterosexual men, 14 MSM), and 4 (12.9%) with a different genovar (1 woman, 1 heterosexual man, 2 MSM). Multilocus sequence typing was performed on all 27 repeat positive cases with identical genovars and of these, 2 cases (1 woman, 1 MSM) were identified as different (new source) infections. Among the 21 MSM with positive retests, 16 retested positive at the same anatomical site as initial infection (14 rectal, 2 urethral infections) and 5 at a different site. There was no difference in repeat positivity among MSM by whether the initial infection was rectal or urethral chlamydia (P = 0.33).

Organism Load

Median organism load at baseline was significantly higher in cervical/vaginal swabs compared with urine specimens (P < 0.01) and rectal swabs (P < 0.01) (Fig. 2A). Once we accounted for sampling variability and standardised by the number of eukaryotic cells present, there was no difference in median organism load between specimen types (P = 0.90) (Fig. 2B). There was also no difference in the standardised load between the three population groups (P = 0.99) (Fig. 2C).

Figure 2
Figure 2:
Chlamydia trachomatis load by specimen type and risk group. A, Load copies per mL/load copies per swab by specimen type. B, Load copies per 100 cells by specimen type. C, Load copies per 100 cells by risk group.

Median organism load (per 100 cells [log10]) was significantly higher for B complex than C complex genovars (P < 0.01) and higher for Intermediate than C complex (P < 0.01), but there was no difference between intermediate and B complex genovars (P = 0.40).

Factors Associated With a Positive Retest

Participants were classified as either MSM or heterosexual for this analysis because of the small number of repeat positive cases for heterosexual men and women separately. Regression analyses were stratified by risk group (heterosexual versus MSM) because of a significant interaction between organism load (per 100 cells [log10] and risk group (P < 0.01).

Univariate analysis showed that baseline organism load was the only variable associated with repeat positivity among MSM. Multivariate analysis, adjusting for time since first diagnosis and number of sex partners in last 3 months, found that the odds of repeat positivity increased by 90% with each additional log organism load (adjusted odds ratio [aOR], 1.9; 95% CI, 1.1–3.2). No other variables were associated with repeat positivity among MSM. Among heterosexual men and women, univariate analysis found that country of birth and time since diagnosis were associated with repeat positivity; organism load was not associated. Multivariate analysis adjusting for gender, time since first diagnosis, country of birth and organism load, found that being born in Australia (aOR, 2.6; 95% CI, 1.1–6.7) and time since diagnosis (aOR, 0.9; 95% CI, 0.8–0.9), were associated with repeat positivity; organism load was not associated (Table 2).

TABLE 2
TABLE 2:
Factors Associated With a Positive Retest

Treatment Failure and Reinfections

Of the 43 repeat positive cases, 9 were classified as treatment failure with an overall failure percentage of 3.1% (95% CI, 1.4%–5.8%), 26 as reinfections (9.0%; 95% CI, 6.0%–12.9%), 1 as a persisting infection (0.3%; 95% CI, 0.0%–1.9%) and 7 could not be classified because genovar data were unavailable (Fig. 1). Treatment failure was more common among MSM than heterosexuals (6.9% vs 1.1%, P = 0.01), but there was no significant difference in reinfection (11.9% vs 7.4%; P = 0.21) (Table 3). Of the 7 MSM classified as having treatment failure, all had rectal infections at baseline, and one of these had rectal and urethral coinfection at baseline, and urethral infection at follow-up. Of these, 5 reported using condoms always and 2 reported no sex since their baseline infection.

TABLE 3
TABLE 3:
Treatment Outcomes at 1–4 Months

Among MSM, treatment failure was associated with organism load (OR, 3.3; 95% CI, 1.4–7.9 for each additional log load per 100 cells [log10]), but reinfection (OR, 1.6; 95% CI, 0.9–2.8) was not. Among heterosexual men and women, neither treatment failure (OR, 1.2; 95% CI, 0.5–3.2) nor reinfection (OR, 0.8; 95% CI, 0.6–1.2) were associated with organism load. Those classified as having a reinfection were more likely to retest earlier than those who were negative on retest (10.9 weeks vs 13.0 weeks on average, P = 0.01).

DISCUSSION

In this cohort of heterosexual men, women, and MSM, we found high rates of positive chlamydia retests within 4 months of initial diagnosis, particularly among MSM. Using MLST and sexual behavior data, we further differentiated between treatment failure and reinfection and found that treatment failure was higher among MSM (6.9% vs 1.1%), and these were predominantly rectal infections.

Treatment failure rates in our study were lower than some others in which rates of up to 21% have been reported in MSM22 and 8% in women.21 There are several possible explanations. Variability in the definition, assessment and timing of positive retests in the available studies make it difficult to compare the findings. Not all studies used genotyping in addition to behavioral data to help differentiate reinfections from treatment failure, and self-reported behavior may have been inaccurate, therefore some cases may have been incorrectly classified. Not all studies excluded repeat tests collected within a month of treatment as we did, and there is the possibility that some cases were false positive diagnoses due to residual DNA.23 Our estimates of treatment failure are conservative; all positive retests in individuals reporting condomless sex are classified as reinfections. However, some cases of treatment failure in MSM may have been misclassified in those who reported always using condoms, as rectal chlamydia has been shown to be associated with sexual activities other than penile-anal intercourse.24 Also, 5 of the 7 repeat positive cases classified as unknown, were from individuals reporting no sex or always using condoms, and some may have been treatment failures. Further, it is possible that the NAAT used in this study did not detect low copy number chlamydial DNA, therefore an infection at another site may have been missed at baseline but picked up at repeat testing.

We observed that higher organism load was associated with repeat chlamydia positivity in MSM, and on further classification of repeat infection, that higher load was associated with treatment failure, but not with reinfection. This is consistent with another recent study.15 A possible explanation, particularly with regard to treatment failure, is heterotypic resistance where, if organism load is sufficiently high, the less susceptible organisms can survive treatment thereby allowing the infection to persist after treatment.25 This raises the question of whether other antibiotics or extended doses of azithromycin are needed.26 We found no statistical association between organism load and repeat infection in heterosexual men and women, whereas 2 previous studies have. One of these19 was based on just 4 participants; the other was a cohort of young women tested at regular intervals.14 We chose to standardise load by the number of eukaryotic cells to allow for variation in load between different sample types. However, chlamydia load may have been underestimated in individuals with high load and a stronger inflammatory response, in which a higher proportion of inflammatory cells will be included among sampled cells.27 Further, our small sample size and degradation of frozen samples over time28 may have affected estimates of organism load contributing to differences with other studies.

We found different distributions of circulating genovars between heterosexuals and MSM, consistent with Bom et al.20 We did not find any association between genovar and repeat positivity, but we did find C-complex cases (H, I, J, K) had a lower organism load. Evidence for an association between load and genovar is conflicting with other studies finding load was highest for serovar D- and B-complex serovars.29

Among heterosexual men and women, the odds of repeat positivity decreased by 10% with each additional week delay in being retested for chlamydia, but no association was found among MSM. Over a quarter of home arm participants in the REACT trial retested early at the clinic, and among heterosexual men there was a higher positivity compared to those who used the postal home collection system,30 suggesting they perceived they were at high risk or had symptoms, and opted to retest earlier rather than waiting for the kit to be posted.

In conclusion, high positive retest rates, particularly among MSM, highlight the importance of retesting around 3 months after treatment. Treatment failure appears to be more common in MSM with rectal chlamydia, particularly those with higher organism loads, further reinforcing concerns about whether single-dose azithromycin is the optimal treatment.

REFERENCES

1. Centers for Disease Control and Prevention (CDC). Sexually Transmitted Disease Surveillance 2011. Atlanta: United States Department of Health and Human Services 2012.
2. European Centre for Disease Prevention and Control. Sexually transmitted infections in Europe 2013. Stockholm: ECDC; 2015.
3. Kirby Institute. HIV, viral hepatitis and sexually transmitted infections in Australia Annual Surveillance Report 2016. Sydney: Kirby Institute, UNSW Australia; 2016.
4. Hocking JS, Willis J, Tabrizi S, et al. A chlamydia prevalence survey of young women living in Melbourne, Victoria. Sex Health 2006; 3:235–240.
5. Goulet V, de Barbeyrac B, Raherison S, et al. Prevalence of Chlamydia trachomatis: Results from the first national population-based survey in France. Sex Transm Infect 2010; 86:263–270.
6. Eggleston E, Rogers SM, Turner CF, et al. Chlamydia trachomatis infection among 15- to 35-year-olds in Baltimore, MD. Sex Transm Dis 2011; 38:743–749.
7. Sonnenberg P, Clifton S, Beddows S, et al. Prevalence, risk factors, and uptake of interventions for sexually transmitted infections in Britain: Findings from the National Surveys of Sexual Attitudes and Lifestyles (Natsal). Lancet 2013; 382:1795–1806.
8. Hocking JS, Guy R, Walker J, et al. Advances in sampling and screening for chlamydia. Future Microbiol 2013; 8:367–386.
9. Hosenfeld CB, Workowski KA, Berman S, et al. Repeat infection with Chlamydia and gonorrhea among females: A systematic review of the literature. Sex Transm Dis 2009; 36:478–489.
10. Harte D, Mercey D, Jarman J, et al. Is the recall of men who have sex with men (MSM) diagnosed as having bacterial sexually transmitted infections (STIs) for re-screening a feasible and effective strategy? Sex Transm Infect 2011; 87:577–582.
11. Haggerty CL, Gottlieb SL, Taylor BD, et al. Risk of sequelae after Chlamydia trachomatis genital infection in women. J Infect Dis 2010; 201(Suppl 2):S134–S155.
12. Bernstein KT, Marcus JL, Nieri G, et al. Rectal gonorrhea and chlamydia reinfection is associated with increased risk of HIV seroconversion. J Acquir Immune Defic Syndr 2010; 53:537–543.
13. Dukers-Muijrers NH, Speksnijder AG, Morré SA, et al. Detection of anorectal and cervicovaginal Chlamydia trachomatis infections following azithromycin treatment: Prospective cohort study with multiple time-sequential measures of rRNA, DNA, quantitative load and symptoms. PLoS One 2013; 8:e81236.
14. Walker J, Tabrizi SN, Fairley CK, et al. Chlamydia trachomatis incidence and re-infection among young women—behavioural and microbiological characteristics. PLoS One 2012; 7:e37778.
15. Kong FY, Tabrizi SN, Fairley CK, et al. Higher organism load associated with failure of azithromycin to treat rectal chlamydia. Epidemiol Infect 2016; 144:2587–2596.
16. Smith KS, Hocking JS, Chen M, et al. Rationale and design of REACT: A randomised controlled trial assessing the effectiveness of home-collection to increase chlamydia retesting and detect repeat positive tests. BMC Infect Dis 2014; 14:223.
17. Australian Sexual Health Alliance 2016. Accessed at http://www.sti.guidelines.org.au/. Accessed October 10th, 2016.
18. Stevens MP, Twin J, Fairley CK, et al. Development and evaluation of an ompA quantitative real-time PCR assay for Chlamydia trachomatis serovar determination. J Clin Microbiol 2010; 48:2060–2065.
19. Gomes JP, Borrego MJ, Atik B, et al. Correlating Chlamydia trachomatis infectious load with urogenital ecological success and disease pathogenesis. Microbes Infect 2006; 8:16–26.
20. Bom RJ, Christerson L, Schim van der Loeff MF, et al. Evaluation of high-resolution typing methods for Chlamydia trachomatis in samples from heterosexual couples. J Clin Microbiol 2011; 49:2844–2853.
21. Batteiger BE, Tu W, Ofner S, et al. Repeated Chlamydia trachomatis genital infections in adolescent women. J Infect Dis 2010; 201:42–51.
22. Hathorn E, Opie C, Goold P. What is the appropriate treatment for the management of rectal Chlamydia trachomatis in men and women? Sex Transm Infect 2012; 88:352–354.
23. Dukers-Muijrers NH, Morré SA, Speksnijder A, et al. Chlamydia trachomatis test-of-cure cannot be based on a single highly sensitive laboratory test taken at least 3 weeks after treatment. PLoS One 2012; 7:e34108.
24. Jin F, Prestage GP, Mao L, et al. Incidence and risk factors for urethral and anal gonorrhoea and chlamydia in a cohort of HIV-negative homosexual men: The Health in Men Study. Sex Transm Infect 2007; 83:113–119.
25. Wang SA, Papp JR, Stamm WE, et al. Evaluation of antimicrobial resistance and treatment failures for Chlamydia trachomatis: a meeting report. J Infect Dis 2005; 191:917–923.
26. Horner PJ. Azithromycin antimicrobial resistance and genital Chlamydia trachomatis infection: Duration of therapy may be the key to improving efficacy. Sex Transm Infect 2012; 88:154–156.
27. Wiggins R, Horner PJ, Whittington K, et al. Quantitative analysis of epithelial cells in urine from men with and without urethritis: Implications for studying epithelial: pathogen interactions in vivo. BMC Res Notes 2009; 2:139.
28. van Dommelen L, Wolffs PF, van Tiel FH, et al. Influence of temperature, medium, and storage duration on Chlamydia trachomatis DNA detection by PCR. J Clin Microbiol 2013; 51:990–992.
29. Vodstrcil LA, McIver R, Huston WM, et al. The epidemiology of Chlamydia trachomatis organism load during genital infection: a systematic review. J Infect Dis 2015; 211:1628–1645.
30. Smith KS, Hocking JS, Chen MY, et al. Dual intervention to increase chlamydia retesting: a randomized controlled trial in three populations. Am J Prev Med 2015; 49:1–11.
Copyright © 2017 American Sexually Transmitted Diseases Association All rights reserved.