The majority of Chlamydia trachomatis genital infections in humans are asymptomatic and without clinical evidence of complications at the time of diagnosis. The natural history of chlamydial infection in humans, including the duration of infection and factors influencing resolution of infection, is not yet completely understood. This is in part attributable to the inherent challenges and ethical considerations in studying untreated chlamydia in humans. An improved understanding of the natural history of chlamydia in humans has implications for chlamydia screening and treatment recommendations. In April 2008, the Centers for Disease Control and Prevention convened an advisory group for the Chlamydia Immunology and Control Expert Advisory Meeting, in which studies related to chlamydia natural history, pathogenesis, and immunobiology were reviewed and gaps in our knowledge that would have implications for prevention and control of C. trachomatis infection were identified. This article summarizes the key questions posed and the evidence reviewed on the duration of untreated, uncomplicated genital chlamydial infection in humans and the factors associated with chlamydia resolution.
Genital Chlamydia trachomatis infection is the most frequently reported bacterial sexually transmitted disease in the United States, and the number of cases reported continues to increase each year. In 2007, >1.1 million chlamydial infections were reported to the Centers for Disease Control and Prevention (CDC) . Increasing numbers of reported cases may simply reflect increased screening efforts rather than a true increase in disease burden, but this remains unclear. Further advances in C. trachomatis prevention and control ef-forts have been hampered in part by our limited knowledge of the natural history of untreated genital chlamydial infections in humans. Specifically, we need more information on the duration of untreated infection and the factors that influence outcomes, including spontaneous resolution (ie, chlamydia clearance without antichlamydial therapy), persistence, and the development of complications. An improved understanding of the natural history of genital chlamydial infections in humans could have implications for determining screening and rescreening intervals, timing of notification of test results and treatment, and optimizing partner notification and treatment recommendations. Differences in assumptions concerning the natural history of C. trachomatis infection will significantly affect estimates of screening strategy cost-effectiveness .
In April 2008, the CDC convened an advisory group for the Chlamydia Immunology and Control Expert Advisory Meeting, in which in vitro, animal, and humans studies related to C. trachomatis natural history, pathogenesis, and immunobiology were reviewed and knowledge gaps that would have implications for prevention and control of chlamydia were identified. This article summarizes the key questions posed and evidence reviewed on the duration of untreated, uncomplicated genital C. trachomatis infection in humans and factors associated with clearance. For each key question, we also address limitations of current studies and considerations in designing future studies (including ethical challenges). Animal and in vitro studies of the duration of untreated chlamydial infection that were reviewed at the advisory meeting are discussed elsewhere in this supplement [3, 4], as are human studies related to chlamydia pathogenesis and protective immunity [5–7].
Participants and Methods
Two searches of the literature were conducted on 28 January 2008 with the Medline computerized database of the US National Library of Medicine. Both searches were limited to humans, to persons ⩾13 years of age, to the English language, and to the tag terms titles or abstracts. The first search was undertaken to investigate the epidemiology of the natural history of untreated, uncomplicated chlamydial infection in humans and the factors influencing chlamydia clearance. Initially, the subject heading “chlamydia” was entered and yielded 5316 citations. These results were sequentially combined and exploded with results of searches, using the following subject headings: natural history, untreated, spontaneous resolution, spontaneous remission, duration, clearance, eradication, resolved, recurrence, relapse, remission, repeat, persistence, follow-up, and subsequent. The first search yielded a total set of 909 citations. The second search sought to investigate host and organism factors influencing susceptibility to or outcomes of uncomplicated chlamydial infection with respect to duration of untreated chlamydial infection in humans. As with the first search, the subject heading “chlamydia” was initially entered, and these results were sequentially combined and exploded with results of searches, using the following subject headings: immune, host response, genetic, microbial, biologic, and molecular. The second search yielded a total set of 418 citations.
From both sets of citations, articles were then removed if they exclusively discussed (1) Chlamydia species other than C. trachomatis, (2) complicated genital chlamydial infection (eg, pelvic inflammatory disease [PID], epididymitis, infertility), (3) chlamydia-associated reactive arthritis, or (4) in vitro research on cell lines rather than on specimens from human participants with current or prior chlamydial infection. Articles on trachoma were not excluded if they also included research on genital chlamydial infections. Citations were then reviewed individually and selected for the final set for each search on the basis of their relevance to key questions that were formulated to address the natural history of untreated, uncomplicated chlamydia in humans. Nine articles from the first search and 10 from the second were reviewed and included in this article. Two additional relevant articles during the search period that were missed by the searches were also reviewed and included, bringing the total articles included to 21 citations. In this article, the key questions and evidence reviewed are summarized in the text. Evidence tables are also included that highlight the study designs and populations, methodology, and key findings.
Key question 1: what is the duration of untreated, uncomplicated genital chlamydial infection in humans? Studies evaluating the natural history of untreated, uncomplicated genital chlamydial infections in humans are summarized in Table 1. The majority of chlamydia natural history studies have prospectively or retrospectively evaluated for infection resolution in the relatively short interval between initial screening and the follow-up visit for treatment or for reassessment as part of a protocol focused on another infection [8–14]. Participants in these studies did not receive specific chlamydia treatment at the time of screening because they had no indications for treatment (eg, examination findings of urethritis or cervicitis or being a contact to a sexual partner with chlamydia). Some participants received antibiotics with poor or no antichlamydial activity for other sexually transmitted diseases (STDs; eg, bacterial vaginosis, trichomoniasis, or syphilis). Follow-up intervals in most of these studies were a few weeks, although in a few instances, they were as long as several months. Chlamydia resolution rates in these studies were 11%–44%, as measured by C. trachomatis-specific commercial nucleic acid amplification tests or research polymerase chain reaction (PCR) assays [8–14]. Although many of these studies evaluated nonpregnant women in conventional clinical settings (eg, STD clinics, emergency rooms, and general practice clinics) [8–12], exceptions included (1) a study by Sheffield et al  that evaluated pregnant adolescents and adults for whom chlamydia testing was performed on archived specimens collected at baseline and follow-up visits in a multicenter bacterial vaginosis treatment trial and (2) a study by van Valkengoed et al  in which initial chlamydia screening was performed on a mailed urine specimen and repeat testing was performed when patients visited a general practice clinic for treatment.
Three studies evaluated the natural history of untreated, uncomplicated genital C. trachomatis infection in humans during intervals ⩾1 year [15–17]. In a study performed >30 years ago, McCormack et al  evaluated C. trachomatis clearance by using culture among women who were initially examined and tested for chlamydia at a university health service and then underwent reexamination and testing 15–25 months later. Of 7 chlamydia-infected women who had not received antichla mydial therapy in the interim, 4 (57%) were infected 16–17 months after the initial testing. Molano et al  retrospectively evaluated C. trachomatis clearance with use of PCR performed on stored cervical specimens from Colombian adolescent and adult female participants who had been enrolled in a human papillomavirus natural history study and had cervical specimens collected every 6–9 months for a mean follow-up time of >5 years. Their analyses were limited to 82 female participants with normal cervical cytological findings who tested positive for C. trachomatis at baseline and had ⩾1 follow-up visit. Chlamydia resolution occurred in 54% of participants at 1 year of follow-up, 83% at 2 years, 91% at 3 years, and 95% at 4 years. Morré et al  prospectively followed 30 asymptomatic chlamydia-infected female participants for 1 year without treatment, and C. trachomatis nucleic acid amplification tests were performed on urine samples mailed with clinical and behavioral data questionnaires at 1, 6, and 12 months after the baseline clinic visit. The chlamydia resolution rate at 1 year was 45%, and none of the women reported clinical symptoms of PID.
A major limitation of all chlamydia natural history studies reviewed was the lack of knowledge about when participants were initially infected; this limits the accuracy of estimates for the duration of chlamydia infection. Another major limitation in most studies was not performing typing of the C. trachomatis outer membrane protein A (OmpA) on both the initial and subsequent chlamydia-positive specimens to ensure a more accurate classification of continued infection as either persistent (with strains of the same OmpA type) or new (with strains of different OmpA types). Geisler et al  previously reported that 5% of participants initially classified as having persistent chlamydial infection were subsequently found to have discordant OmpA types at the screening and treatment visits, suggesting that they had acquired a new chlamydial infection. Another limitation of most of the studies reviewed concerned the generalizability of the observations, because most studies were conducted in populations with high chlamydia prevalence rates. In summary, human studies on the duration of untreated, uncomplicated genital C. trachomatis infections have shown that chlamydia clearance increases over time, with approximately half of infections spontaneously resolving ∼1 year after initial chlamydia testing. However, lack of precise data on the timing of infection acquisition, lack of OmpA typing data (or other C. trachomatis strain typing methods), and lack of generalizable findings are major limitations of most chlamydia natural history studies to date. There have been no studies on the natural history of untreated, uncomplicated chlamydial infection in men with follow-up periods of more than a few months.
Ethical considerations are a major challenge in studying the natural history of untreated chlamydia. Untreated genital chlamydia can infrequently progress to symptomatic upper genital tract complications in men (eg, epididymitis) and women (eg, PID). There is also concern that untreated chlamydia that remains asymptomatic can still progress to cause undetected upper genital tract inflammation and damage, which has been associated with additional long-term complications (eg, ectopic pregnancy and infertility). Therefore, participants who are identified as chlamydia-infected should be treated promptly. Taking these ethical challenges into consideration, designing longitudinal studies on the natural history of chlamydia will be difficult.
Key question 2: which clinical factors influence resolution of untreated, uncomplicated genital chlamydial infection in humans? In the absence of antibiotic therapy, C. trachomatis clearance in humans is presumed to be mediated by host immune responses. As discussed above, one factor that has been associated with chlamydia resolution in most studies is a longer interval between the time of initial chlamydia testing and subsequent treatment, suggesting that the immune system clears the infection over time. However, it is important to understand which clinical factors are associated with a more rapid resolution of chlamydial infection, because these may be surrogates for protective immune responses. It has been postulated that prior chlamydial infections may provide protective immunity against subsequent infections. Older age, which may be a surrogate for prior chlamydial infections, was associated with chla mydia resolution in 2 studies [8, 13], but this association was not consistently found in all studies reviewed. Two studies [8, 9] evaluated whether a history of genital chlamydia influenced the duration of chlamydial infection. The results were contradictory; one study demonstrated a trend toward a higher frequency of chlamydia resolution among those with a history of chlamydia , but the other did not find a difference in resolution rates . Both studies were limited, because determination of a prior chlamydia infection was based only on review of clinical and laboratory records. Because serum C. trachomatis antibody testing was not performed, prior chlamydia infection status may have been misclassified in patients without a reported or documented history of chlamydia. Both studies were also limited by small sample sizes, short follow-up periods, and the inability to determine precisely the duration of chlamydia infection. Moreover, 1 of the 2 studies did not incorporate C. trachomatis ompA typing in the classification of outcomes . Two studies demonstrating an association of sex with chlamydia resolution had contradictory results; the first found a trend toward more frequent chlamydia resolution in men (36% vs 16%) , and the other found more frequent resolution in women (23% vs 6%) . A single study reported an association between race or ethnicity and infection resolution, with infections resolved more often in white than in nonwhite participants . Sheffield et al  evaluated whether bacterial vaginosis influences chlamydia resolution and demonstrated no association with either clinical findings of bacterial vaginosis or treatment of the condition. Finally, Molano et al  reported an association between oral contraceptive use and chlamydia resolution, with women who had ever used oral contraceptives clearing the infection more rapidly.
In summary, we have limited knowledge about the clinical factors that influence the duration of untreated, uncomplicated genital chlamydial infections in humans. Prior chlamydial infections may afford some degree of protective immunity and, thereby, shorten the duration of subsequent infections. Future studies addressing the influence of prior chlamydial infection on the duration of subsequent infections will need to incorporate C. trachomatis serological testing to define prior exposure more accurately. If possible, such studies should be designed to precisely define the timing of chlamydia acquisition and use larger samples of participants to provide sufficient power to evaluate associations of clinical factors with chlamydial infection resolution. To define more precisely the timing of chlamydia acquisition, studies would probably have to incorporate very frequent collection of genital tract samples for chlamydia testing and a diary of sexual activities in a population of participants who are negative for C. trachomatis at baseline but are at high risk for STD acquisition.
Key question 3: which host immune responses occur in uncomplicated genital chlamydial infections in humans, and which genetic determinants of the host modulate these immune responses? Critical to our understanding of the duration of genital chlamydial infection in humans is knowledge of the systemic (peripheral blood) and mucosal host immune responses that facilitate C. trachomatis clearance and the genetic determinants that modulate these immune responses. In contrast to vast amounts of animal and in vitro data on the host immune responses facilitating C. trachomatis clearance, there has been sparse research into the human immune responses to genital chlamydial infections, and most of the available information does not address the relationship of host immune responses to resolution of uncomplicated chlamydial infection.
Key findings from the human immune response studies reviewed (Table 2) are that, during active genital chlamydial infection, (1) serum and genital mucosal antibodies (mainly immunoglobulin [Ig] A and IgG) to specific C. trachomatis proteins and to C. trachomatis elementary bodies (EBs) are usually detected [18–21]; (2) there is an increase in genital mucosal concentrations of select cytokines [18, 21, 22]; (3) urethral polymorphonuclear (PMN) cell counts in men are highly variable, with almost 20% of men having <5 PMN cells per oil immersion field ; (4) PMN cell counts in cervicovaginal lavage samples are not consistently elevated ; (5) an increase in cervical T cell but not B cell phenotypes has been demonstrated, and there is no change in immunophenotypes systemically ; and (6) both systemic and mucosal lymphoproliferative responses of peripheral blood mononuclear cells to C. trachomatis antigens have been demonstrated [18, 20]. These studies assessed host immune responses during active chlamydial infection.
In a longitudinal study of host immune responses in commercial sex workers in Nairobi who were at high risk of chlamydia, Cohen et al  found that select peripheral blood mononuclear cell lymphoproliferative responses to C. trachomatis EBs and heat-shock protein 60 (cHSP60), measured in participants at baseline, correlated with protection against incident chlamydia, but neither cervical nor serum antibodies against C. trachomatis EBs and cHSP60 were associated with a decrease in incident chlamydial infection. As noted in Table 2, the study by Agrawal et al  also evaluated differences in host immune responses between female participants with primary genital chlamydial infection and those with recurrent genital chlamydial infections (distinguished by absence or presence, respectively, of serum C. trachomatis IgG). These authors found that cervical lymphoproliferative responses to cHSP10 were higher in recurrent infection, whereas lymphoproliferative responses to OmpA were higher in primary infection; another key finding was that interferon-γ levels were higher in cervical wash samples from women with recurrent infection.
A major limitation of many of the studies evaluating host immune responses in uncomplicated genital chlamydial infections in humans was the lack of control for other potential pathogens (eg, herpes simplex virus, Mycoplasma, and Ureaplasma) that could induce immune responses (eg, cytokines and cellular phenotypes) that are not specific for C. trachomatis. Another limitation in some studies was lack of generalizability of the observations beyond the patient populations studied, such as studies involving human immunodeficiency virus (HIV)-infected participants or participants with multiple prior STDs.
Understanding host immune responses in uncomplicated genital chlamydial infections in humans is just as important as understanding the host genetic determinants that modulate these immune responses. It is often difficult to link specific genetic determinants to specific host immune responses, and the 2 studies that evaluated the relationship of genetic determinants and uncomplicated genital chlamydial infections have primarily focused on how the genetic determinant affected risk of chlamydia acquisition  or recurrence , rather than on host immune responses to the infection (Table 2).
These 2 studies, by Geisler et al  and Wang et al , evaluated data from 485 adolescents who were enrolled in the Reaching for Excellence in Adolescent Care and Health (REACH) study, a multicenter, longitudinal study of adolescent health. Genetic variants determined at baseline included HLA class I and II variants and 16 single-nucleotide polymorphisms from 7 cytokine genes. C. trachomatis ligase chain reaction testing was performed every ∼6 months for a median followup period of 3 years. The majority of participants were female, black, and HIV infected. Geisler et al  found that DQB1*06 and/or B*44-Cw*04 predicted incident chlamydial infection (defined as a new chlamydial infection during follow-up in participants negative for chlamydia at baseline) independent of prior chlamydial infection. Wang et al  found that an interleukin-10 gene promoter variant (promoter positions -1082, -819, and -592) was protective against recurrent chlamydia (defined as ⩾2 chlamydial infections during the study separated by a visit during which the patient tested negative for chlamydia), and female participants without this genetic variant had higher endocervical interleukin-10 levels at the time of chlamydia acquisition. The major limitations of these studies were the inclusion of a significant number of HIV-infected participants, which could have confounded immune responses, and the fact that the populations were primarily black, which limited the generalizability to persons with other races, especially with regard to differences in genetic backgrounds. Another limitation of the REACH cohort studies was analysis of a small number of genetic variants.
In summary, there are few studies of host immune responses in humans with uncomplicated genital chlamydial infections. These studies evaluated immune responses during active chlamydial infection but did not evaluate which specific host responses affect infection resolution versus persistence. There are also few studies on the role of immunogenetic determinants of the outcome of uncomplicated genital chlamydial infection (in contrast, trachoma or tubal factor infertility have been studied more extensively). Future studies of the relationship of host immune responses and genetic determinants with uncomplicated chlamydial infection resolution will need to address the limitations of studies to date. Such future studies will need (1) a more exhaustive evaluation for other potential pathogens that could influence immune responses, (2) a healthy, diverse patient population, (3) more detailed evaluations of host immune responses against a wider range of C. trachomatis antigens, and (4) use of high-throughput DNA sequencing technology to screen a much larger number of genetic determinants (not only HLA and cytokine gene variants but also, perhaps, repetitive DNA elements).
Key question 4: which biological properties of C. trachomatis influence resolution of uncomplicated genital chlamydial infections in humans? Although there have been considerable advancements in understanding of the biological and functional characteristics of C. trachomatis, there has been sparse translational research evaluating how the biological characteristics of C. trachomatis impact uncomplicated genital chlamydial infection outcomes in humans. The primary biological characteristic of C. trachomatis of which the relationship to resolution of uncomplicated genital chlamydial infection in humans has been studied is the OmpA protein (previously referred to as major outer membrane protein, or omp1). OmpA is an abundant C. trachomatis protein that is exposed on the surface of the organism and is highly immunogenic. Three studies used ompA genotyping (by different methods) to evaluate the relationship of C. trachomatis OmpA genotype with uncomplicated genital chlamydial infection resolution or persistence in women [16, 17, 29]. Molano et al  demonstrated that female participants infected with C. trachomatis serogroup B and C OmpA types had a longer duration of genital chlamydial infection (ie, persistence). Morré et al  reported that female participants infected with C. trachomatis OmpA type E were more likely to have chlamydia persistence after 1 year of follow-up than were those infected with the other genotypes. Both of these studies were limited, because OmpA genotyping was not performed at all visits and men were not evaluated.
Dean et al  described several female participants with ⩾3 consecutive, uncomplicated genital chlamydial infections with the same OmpA genotype (identical ompA gene sequences or only 1 or 2 amino acid changes) during 12 years of followup at a Seattle STD clinic and suggested that this was evidence for persistence, although reinfection from an untreated partner could not be excluded. The minimum inhibitory concentrations of doxycycline and azithromycin for the C. trachomatis strains isolated from these participants were not elevated, ar guing against antimicrobial resistance contributing to persistence.
A study by Gomes et al  evaluated the relationship of OmpA genotype with another biological characteristic of C. trachomatis infection, the infectious load (a quantitative measure of organism burden and presumably a surrogate for chlamydia replication), and how genotype correlated with genital chlamydial infection outcomes. Chlamydia load was based on the copy number of organisms per copy number of eukaryotic cells, both measured by real-time quantitative PCR. Rather than study the relationship of OmpA genotype and chlamydia load to chlamydial infection resolution over time, these authors studied differences in the chlamydia load in those with prior chlamydia (controlling for OmpA genotype) and changes in the chlamydia load with subsequent chlamydia infections. Their study demonstrated that chlamydial infectious burden was initially not associated with OmpA genotype but was lower in subsequent chlamydia episodes, regardless of genotype, among in patients with prior infections, although this difference did not reach statistical significance. Neither clinical signs nor symptoms of infection were associated with organism load, and the proportion of patients in this study who had clinical findings of PID is unclear.
In summary, we have very limited knowledge of the impact of biological characteristics of C. trachomatis on outcomes of uncomplicated genital chlamydial infection in humans, including most importantly, the duration of infection. From the studies reviewed, it is reasonable to speculate that humans do not develop complete protective immunity to a specific OmpA genotype but that they develop partial protective immunity, as evidenced by the lower organism loads with subsequent genital chlamydial infection. Since the sequencing of the C. trachomatis genome, there has been an acceleration in the in vitro work describing biological characteristics and possible life cycle variants of C. trachomatis. Future studies must translate these findings into human data to improve understanding of the clinical implications in the natural host of this important sexually transmitted pathogen (eg, which C. trachomatis proteins stimulate immune responses in humans, resulting in more rapid chlamydia resolution and/or protection against recurrent infection).