Geisler, William M. MD, MPH*†; Wang, Chengbin PhD†; Morrison, Sandra G. BS*; Black, Carolyn M. PhD‡; Bandea, Claudiu I. PhD‡; Hook, Edward W. III MD*†§
CHLAMYDIAL INFECTION is the most commonly reported sexually transmitted bacterial infection. The majority of persons with genital Chlamydia trachomatis infections are asymptomatic and, as a result, detection of infection often relies on screening. Some untreated chlamydial infections resolve without therapy,1–4 and others persist for long periods of time3,4 and a proportion of these will progress to cause complications. The rates of spontaneous resolution, persistence, and progression of untreated chlamydial infections in turn may profoundly influence recommendations for elements of control efforts such as chlamydia screening frequency or time parameters for partner notification.5
Studies of untreated chlamydial infection during the interval between testing for chlamydia (by screening) and treatment provide a glimpse at the natural history of infection in humans. Of particular interest to our group, because of their implications for understanding human host immune responses to infection, are reports that 20% to 54% of infections spontaneously resolve without treatment over periods as long as 1 year.1–4 These studies on the natural history of chlamydia had one or more limitations, including small numbers of participants,1–4 retrospective study design,1,3 inclusion of patients who received antibiotics that may have antichlamydial activity,1 or difficulties in ascertaining whether repeat positive chlamydial tests in untreated persons reflect persistence of infection or reinfection after spontaneous resolution of initial infection.1–4
To better understand the natural history of chlamydial infections and address limitations of prior studies in part, we are prospectively evaluating the natural history of chlamydial infection in the interval between screening and treatment for infections detected by screening in a study named “Human Immune Responses to C. trachomatis Infection” (HIR-CT). In this article, we present findings from our initial epidemiologic investigation of the natural history of chlamydia in the HIR-CT study and discuss some of the implications of our findings.
Materials and Methods
Study Design and Study Population
This ongoing prospective study of chlamydia outcomes before therapy (HIR-CT) began in September 2002. The study population is composed of male and female subjects 16 years or older presenting to the Jefferson County Department of Health Sexually Transmitted Diseases (STD) clinic in Birmingham, AL, for chlamydia treatment within 60 days of a positive chlamydia screening test. Before October 2005, most Jefferson County Department of Health STD clinic patients were screened for chlamydial infection using cell culture. Subsequent to October 2005, a switch was made to the Gen-Probe Aptima Combo 2 (Gen-Probe, San Diego, CA) nucleic acid amplification test (NAAT), which has improved sensitivity over culture.6,7 Although many patients seen in our clinic are empirically treated at the time of initial evaluation for chlamydia-associated syndromes (e.g., urethritis or cervicitis) or other treatment indications (e.g., chlamydia contact), only patients who have not received antibiotics effective for C. trachomatis are eligible for inclusion in the study. For inclusion in the study, a 60-day maximum time interval between screening and treatment visits was chosen to minimize confounding factors such as unreported interval treatment or exposure to a new chlamydial infection. We also exclude those who have received antibiotics with antichlamydial activity since being screened or have had a prior hysterectomy. The study nurse is aware that the patients' reason for visit was for treatment of a positive chlamydia test, but she is not aware of any clinical manifestations or other clinical or behavioral data before approaching the patient for enrollment.
Clinical and Laboratory Data Collection
Participants' clinical records are reviewed to confirm their eligibility to participate in the study, and the following data from both the screening and treatment visits are recorded onto a standardized data collection form: sociodemographic information; sexual history including STD history, risk factors, and contraception; antibiotic use history; genital symptoms; genital examination findings; and routine laboratory data (e.g., wet mount findings, chlamydia and gonorrhea test results, etc.). At the enrollment (treatment) visit, subjects' genital swabs, urethral in men and cervical in women, are collected for gonorrhea and chlamydial cultures, which are performed by using the previously described methods.1,8 For patients included in the data presented in this paper, all genital specimens negative for chlamydia culture were tested by using polymerase chain reaction (PCR) (COBAS AMPLICOR; Roche Diagnostic Systems, Branchburg, NJ). Briefly, chlamydia transport media containing the genital swab (stored at −70°C) is thawed and vortexed, and then 100 μL was aspirated and tested by using PCR following the manufacturer's protocol.
To help ascertain interval infections occurring between screening and treatment visits, C. trachomatis major outer membrane protein (OmpA) genotyping, consisting of amplification and sequencing of the OmpA gene, is performed on C. trachomatis strains from both visits in subjects who tested positive for chlamydia at the treatment visit. OmpA genotyping was performed at the Centers for Disease Control and Prevention through March 2004 by reported methods,9 and since then has been performed at the University of Alabama at Birmingham with slight protocol9,10 modifications and reported primers.9,10 Residual chlamydia transport medium containing genital swabs is also used for the OmpA genotyping.
Analyses presented here were conducted on Stata (Stata Corp. Release 8.0, College Station, TX). The proportion with spontaneous resolution (defined as chlamydia culture and PCR negative) versus persisting infection (defined as chlamydia culture positive or PCR positive for culture-negative specimens) when presenting for treatment and the associated predictors were the primary outcomes of interest. These outcomes were assessed univariately using Fisher exact test or the Mann-Whitney U test as appropriate, then by multivariate modeling. P values are 2-sided. Subjects found to have discordant C. trachomatis OmpA genotypes from the screening and treatment visits were excluded from further analysis because of inability to assign them with certainty as resolution versus persistence of infection.
Characteristics of Study Participants
To date, 134 subjects have been enrolled. Five subjects with persistent chlamydia (testing positive at both screening and treatment visits) had discordant C. trachomatis OmpA genotypes and were excluded from further analyses (Fig. 1). Although both strains could not be typed in 4 subjects or were not available for 9 subjects with persisting chlamydia, these patients were included in our analyses, as they denied new sex partners between visits and subanalyses revealed excluding them did not alter the significance of our findings.
In the 129 subjects analyzed, 115 (89%) were female and 114 (88%) were black. The median age was 21 years (range, 17–54), with men somewhat older than women (24 vs. 21 years; P = 0.099). The median interval between screening and treatment visits was 13 days (range, 4–59), and blacks had a longer median interval between visits than whites (13 vs. 11 days; P = 0.031). One or more prior chlamydial infections were reported by 54 (42%) subjects, and a history of chlamydia was associated with black race (46% vs. 13%; P = 0.024) and an older median age (22 vs. 21 years; P = 0.039) but not gender. Four (3%) subjects had gonococcal coinfection at the screening or treatment visit; all were black women who were without clinical manifestations of cervicitis at the screening visit. No subjects were known to be HIV-positive nor did they test positive for HIV-1 antibodies.
At the time of return for treatment, of 115 female subjects, 23 (20%) had bacterial vaginosis, 17 (15%) candidiasis, and 15 (13%) trichomoniasis. Twenty-six (23%) female subjects were found to have purulent endocervical discharge not noted at the time of initial evaluation, and all but one had persisting infection; those with endocervical discharge tended to have a higher median age (23 vs. 21 years; P = 0.076) but no difference in the median interval between screening and treatment. Two (2%) women, neither with a history of prior chlamydia, were newly diagnosed with pelvic inflammatory disease (PID) at the time of returning for treatment, both being persistently C. trachomatis culture positive and one having become Neisseria gonorrhoeae positive in the interval since her screening visit. The interval between screening and treatment was 25 days for the subject who acquired gonorrhea before returning and 7 days in the other subject.
Similar to observations made in women, 10 (71%) of 14 C. trachomatis infected men presented with new urethral discharge at the treatment visit, 9 of whom were demonstrated to have persisting infection. One 24-year-old black heterosexual man with a prior history of chlamydia was diagnosed with epididymitis at the treatment visit at which time he was without clinical manifestations of urethritis and tested negative for chlamydia by culture and by PCR; the interval between screening and treatment in this subject was 55 days.
The Natural History of Chlamydial Infection and the Relationship to Chlamydial Diagnostic Test and Patient Characteristics
Overall, 38 (29%) subjects had negative chlamydia culture results at the treatment visit. However, 15 of these culture-negative subjects tested chlamydia positive by PCR, yielding a spontaneous resolution rate (i.e., culture-negative and PCR-negative) of 18% (Table 1). Irrespective of the initial screening test, a greater number of subjects tested chlamydia negative by culture versus PCR at the treatment visit (P <0.0001). When stratifying the chlamydia outcomes by the initial screening test, the spontaneous resolution rates were much higher in subjects initially screened by NAAT versus culture/enzyme immunoassay (41% vs. 8%; P <0.0001).
Spontaneous resolution was not significantly associated with age or race. Though not reaching statistical significance, spontaneous resolution was somewhat more common in men (36% vs. 16%; P = 0.13), in subjects who reported prior chlamydia (24% vs. 13%; P = 0.16), or with a longer median interval between screening and treatment visits (14 vs. 12 days; P = 0.079) (Table 2). The significance of these findings was unchanged on multivariate analyses (all P >0.05). Stratifying by gender, men with persisting infection compared with those with spontaneous resolution more often had urethral discharge (P = 0.005), and women with persisting infection were somewhat more likely to have purulent endocervical discharge (P = 0.070). Women with persisting infection were also somewhat more likely to have vaginal candidiasis (18% vs. 0%; P = 0.070), but not other vaginal infections.
The natural history of chlamydial infections remains poorly understood, and its understanding is critical for recommendations relating to the timing of treatment, as it relates to development of complications and to recommendations for rescreening. Natural history studies of untreated chlamydia in humans are difficult to design because ethically, clinical suspicion or the diagnosis of genital chlamydia mandates treatment. An experiment of nature, more or less, provides the opportunity to study the natural history of chlamydia. Patients without chlamydial signs or symptoms or a known contact to an infected partner at the time of initial evaluation (chlamydia screening) are typically tested but not routinely treated; they return for therapy upon notification of a positive chlamydia test.
Taking advantage of these standard practices, we initiated a prospective study that allows us to study the natural history of chlamydia and the epidemiologic, clinical, and biological factors that influence clearance versus persistence of chlamydial infection before treatment. In the first analyses of data from our cohort, we have demonstrated that both progressive development of clinical findings and spontaneous resolution of chlamydial infection before receiving therapy are common. The rate of spontaneous clearance we report (18% overall) parallels those in studies with a similar follow-up time.1,2 In a retrospective study of 74 chlamydia culture-positive STD clinic patients (mostly women) returning for treatment, Parks et al. reported spontaneous resolution in 28% by culture, direct fluorescence antibody, or PCR.1 Similarly, in a study of 94 C. trachomatis PCR-positive STD clinic patients, Joyner et al. reported 19% of men and 22% of women spontaneously cleared their infection.2 In both our analyses and the findings by Parks et al.,1 resolution of chlamydia was associated with a longer interval between the time of testing and treatment. Other studies of the natural history of chlamydia with longer follow-up have provided further evidence that spontaneous resolution of chlamydia seems to occur in a time-dependent fashion. In a retrospective analysis of asymptomatic chlamydia-infected Columbian women who did not receive therapy, Molano et al. reported resolution of infection in 54% at 1 year, 83% at 2 years, 91% at 3 years, and 95% at 4 years.3 Similarly, in a nested case-control study by Morre et al., 30 asymptomatic chlamydia-infected women in Amsterdam had a person/year spontaneous clearance rate of 44.7% per year.4 Of the chlamydia natural history studies discussed above, only the Parks et al. study1 also found an association of spontaneous resolution with older age, an indirect surrogate for the likelihood of prior chlamydia and perhaps protective immunity. We did not find an association of age and resolution; however, there were no significant differences in reported prior history of chlamydia in those with resolution versus persistence.
Unlike earlier studies using this design, we were also able to characterize the progression of infection, manifesting on examination as either interval development of signs of lower genital tract inflammation (e.g., cervical or urethral discharge) or upper genital tract inflammation (e.g., PID or epididymitis). The development of lower genital signs of infection occurred almost entirely in individuals with persisting chlamydial infection and may represent the progression of acute or early chlamydial infection. Among persons with persisting infection, 32% overall (100% of men and 26% of women; Table 2) developed urogenital (urethral or cervical) discharge in the interval between initial evaluation and treatment. Serious complications of chlamydia occurred in 3 subjects (PID 2, 2% and epididymitis 1, 7%) in the interval between screening and treatment, and in 2 of these subjects, over 3 weeks had elapsed before they returned for treatment. Our observations confirm the limited data previously published addressing complications of chlamydia in this clinical setting. Bachmann et al. previously reported that 4% of women developed PID in the interval between testing at an emergency room and return for treatment at a University Hospital in Birmingham, AL,11 and Hook et al. reported that 3% of women developed interval PID between screening and treatment in a study performed at 2 STD clinics in Baltimore.12
Another interesting finding in our study was that all chlamydia culture-negative men were also PCR-negative; however, 12 of 23 chlamydia culture-negative women were PCR-positive. There are several possible explanations: (a) deoxyribonucleic acid from nonviable C. trachomatis may be more rapidly cleared from the urethra than the cervix (the testing interval did not differ by gender); (b) mucosal immune responses in the urethra may more effectively and/or rapidly eradicate C. trachomatis than those in the cervix; and (c) culture-negative persistent chlamydial forms may more likely develop over time in the cervix than the urethra. However, considering the small number of culture-negative men in our analyses to date, this finding should be considered preliminary and will need to be verified with a more suitable sample size of men.
Prior studies on the natural history of chlamydia had one or more limitations that we worked to address in our study, and these included: smaller numbers of participants,1–4 retrospective study design,1,3 inclusion of patients who received antibiotics which may have antichlamydial activity,1 or difficulties in ascertaining whether repeat positive chlamydial tests in untreated persons reflect persistence of infection or new infection following spontaneous resolution.1–4 Our study design addressed the first 3 limitations, and we utilized C. trachomatis OmpA genotyping in patients categorized as persisting infection to address the final limitation, that being whether repeat positive tests were persisting or new infection. We found a small proportion had discordant C. trachomatis OmpA genotypes between their screening and treatment visits. It is possible these subjects spontaneously resolved their initial infection and were then infected with a new C. trachomatis OmpA genotype; 4 of the 6 patients with discordant genotypes provided partner data, and 2 reported new partners since initial testing. It is also possible that these subjects had a mixed C. trachomatis OmpA genotype infection not detected at either visit.
Our study is not without limitations. As is often the case when studying the natural history of sexually transmitted infections in humans, we are unable to precisely determine the duration of infection before spontaneous resolution (without knowing the date of acquisition). Such information would not only be valuable in determining chlamydia screening intervals but would also aid in understanding whether individuals who develop signs of infection or complications have early (acute) chlamydial infection or have progression of more long-standing infection. Because of ethical issues, we are only able to observe the natural history between the time of screening and treatment (the “experiment of nature”) and are otherwise obligated to provide treatment to patients in a timely fashion. Our protocol did not modify standard clinic procedures for promptly notifying patients of a positive chlamydia test and urging them to return for therapy; the majority return between one and 4 weeks following the initial test (median 13 days, which is comparable to that reported elsewhere13). Although OmpA genotyping chlamydial strains helps to confirm persisting infection versus new infection, it does not rule out the possibility that subjects may have resolved their initial infection but were reinfected with the same strain again due to re-exposure from an untreated partner. Another limitation is that although we collect self-reported data and review patients' medical and laboratory records to assess for a history of prior chlamydia, we still cannot exclude the possibility that subjects still might have had a prior chlamydial infection; this information may be important if having prior chlamydia influences one's time to spontaneous resolution (efficiency of chlamydia eradication) in subsequent chlamydial infections.
A final limitation is that we conservatively defined spontaneous resolution as having a negative C. trachomatis PCR at the time of returning for treatment. Chlamydia PCR is more sensitive than culture, and this is supported by our finding of a lower spontaneous clearance rate when measured by PCR compared with chlamydia culture; however, a false-negative result is still possible with PCR. A positive PCR result with a negative culture could represent either identification of residual deoxyribonucleic acid from dead chlamydial organisms or detection of low levels of viable organisms missed by quantitative culture. If PCR is detecting viable chlamydia, but in low levels, the finding of an initial culture-positive but later culture-negative state may still be indicative of an antichlamydial host immune response that is functioning to decrease the quantitative burden of chlamydial organisms over time and will lead to resolution of infection. It is likely that in some instances, C. trachomatis NAATs are measuring dead organisms, and by the time some individuals have their infection detected by a screening NAAT, they have already resolved their infection. This may be the reason we found a higher resolution rate in those initially screened by NAATs (compared with culture) in which many of these individuals either had already resolved their infection or were actively resolving their infection. It is also possible that more of the initially positive NAATs may have resolved because infections with fewer organisms may respond better to host defenses and NAATs presumably detect more infections with fewer organisms. The possible misclassification of chlamydia outcome could influence the significance of the epidemiologic and clinical characteristics evaluated.
In summary, we present evidence that spontaneous resolution of chlamydia in humans in common, and confirm findings by others1,3 that resolution occurs in a time-dependent fashion. A great dilemma in understanding the natural history of genital chlamydial infections is understanding what factors contribute to resolution of infection. Mechanisms of spontanteous resolution are of great importance for efforts to develop chlamydial vaccines, providing opportunities to understand what elements of the human host immune response or chlamydial biology contribute to resolution of chlamydial infection. We are collecting serum and genital specimens to further address the mechanisms of chlamydial resolution in our future studies. A very important take-home point from our current findings is the importance of subjects receiving chlamydia treatment promptly, not only to avert chlamydia complications but also to limit chlamydia transmission to other susceptible individuals. Further studies addressing ways to achieve timely treatment, whether through reduction of turn-around-time or use of strategies such as point of care tests are warranted.
1. Parks KS, Dixon PB, Richey CM, et al. Spontaneous clearance of Chlamydia trachomatis infection in untreated patients. Sex Transm Dis 1997; 24:229–235.
2. Joyner JL, Douglas AM Jr, Foster M, et al. Persistence of Chlamydia trachomatis infection detected by polymerase chain reaction in untreated patients. Sex Transm Dis 2001; 29:199–200.
3. Molano M, Meijer CJ, Weiderpass E, et al. The natural course of Chlamydia trachomatis infection in asymptomatic Colombian women: A 5-year follow-up study. J Infect Dis 2005; 191:907–916.
4. Morré SA, van den Brule AJC, Rozendaal L, et al. The natural course of asymptomatic Chlamydia trachomatis infections: 45% clearance and no development of clinical PID after one-year follow-up. Int J STD AIDS 2002; 13:S12–S18.
5. Hu D, Hook EW III, Goldie SJ. Screening for Chlamydia trachomatis in women 15 to 29 years of age: A cost-effectiveness analysis. Ann Intern Med 2004; 141:501–513.
6. Johnson RE, Newhall WJ, Papp JR, et al. Screening tests to detect Chlamydia trachomatis and Neisseria gonorrhoeae infections. MMWR Recomm Rep 2002:51:1–38.
7. Watson EJ, Templeton A, Russell I, et al. The accuracy and efficacy of screening tests for Chlamydia trachomatis: A systematic review. J Med Microbiol 2002; 51:1021–1031.
8. Smith KR, Ching S, Lee H, et al. Evaluation of ligase chain reaction for use with urine for identification of Neisseria gonorrhoeae in females attending a sexually transmitted disease clinic. J Clin Microbiol 1995; 33:455–457.
9. Bandea CI, Kubota K, Brown TM, et al. Typing of Chlamydia trachomatis strains from urine samples by amplification and sequencing the major outer membrane protein gene (omp1). Sex Transm Infect 2001; 77:419–422.
10. Carlson JH, Hughes S, Hogan D, et al. Polymorphisms in the Chlamydia trachomatis cytotoxin locus associated with ocular and genital Isolates. Infect Immun 2004; 72:7063–7072.
11. Bachmann LH, Richey CM, Waites K, et al. Patterns of Chlamydia trachomatis testing and follow-up at a university hospital medical center. Sex Transm Dis 1999; 26:496–499.
12. Hook EW III, Spitters C, Reichart CA, et al. Use of cell culture and a rapid diagnostic assay for Chlamydia trachomatis screening. JAMA 1994; 272:867–870.
13. Wong D, Berman SM, Furness BW, et al. Time to treatment for women with chlamydial or gonococcal infections: A comparative evaluation of sexually transmitted disease clinics in 3 US cities. Sex Transm Dis 2005; 32:194–198.