THE SPECTRUM OF UROGENITAL DISEASE caused by Chlamydia trachomatis ranges from asymptomatic clinically inapparent infections to symptomatic clinically complex syndromes. The major outer-membrane protein (MOMP) of C trachomatis contains specific antigens that differentiate chlamydial strains by serovar (based on antigenic cross-reactivity on microimmunofluorescence) or genotype (based on nucleotide sequencing of the OMP 1 gene). Whether different serovars or genotypes of C trachomatis demonstrate different virulence potentials is unclear. With the development of C trachomatis serotyping and genotyping techniques, studies have been performed to assess whether specific nonlymphogranuloma venerum (non-LGV) serovars or genotypes are associated with specific clinical manifestations. 1–7 These studies have not always demonstrated consistent relationships and have in fact produced contradictory results. Possible reasons for differing outcomes include insufficient study sample sizes to reliably detect associations, if they exist, and failure to address potential confounding factors such as age, race, and presence of the recently described nonfusing C trachomatis phenotypes. 3,4,8–10
Better defining the relationship between chlamydial serovars and specific clinical manifestations may improve our understanding of the pathogenesis and epidemiology of C trachomatis urogenital infections. Utilizing a large database that allowed for adjustment of potential confounding factors, we conducted a cross-sectional study to assess the relationship of chlamydial serovars to clinical manifestations of urogenital disease, and we also reviewed prior studies assessing this relationship.
Study Population and Data Collection
The study population consisted of 671 women and 933 heterosexual men who had culture-documented C trachomatis urogenital infection diagnosed at the Public Health–Seattle and King County Sexually Transmitted Diseases (PHSKC-STD) Clinic during the period of May 1995 through December 1999. The PHSKC-STD Clinic computerized database system, which comprises gender-specific standardized medical record forms, was used in the collection of data, including information on demographics, sexual practices, hormonal contraceptive use, history of prior sexually transmitted diseases (STDs), presence of urogenital symptoms, physical examination findings, and clinical diagnoses of STDs, pelvic inflammatory disease (PID), and epididymitis. Laboratory data collected included C trachomatis and Neisseria gonorrhoeae urethral (male) and cervical cultures, C trachomatis serovar, and number of polymorphonuclear cells (PMNs) per high-power field (hpf) on gram-stained smear. Individuals with coexisting gonococcal infection, infections due to nonfusing C trachomatis phenotypes or mixed chlamydial serovars, and repeated infections with the same serovar within 1 month were excluded. Use of collected data for research purposes was previously approved through the University of Washington Human Subjects Committee. Before analyses, all patient identifiers were removed from the clinical records.
Source and Characterization of C trachomatis Isolates
The methods used for specimen collection, isolation of C trachomatis, and serotyping have been previously described in detail. 9,11 In brief, specimens were collected from the anterior urethra in men and endocervix in women, stored in chlamydia transport medium at 4 °C, and transported within 24 hours to the laboratory. Specimens were inoculated onto McCoy cells and overlaid with growth medium. For serotyping, a low-passage microtiter plate culture method was used, with a panel of 20 subspecies- and serovar-specific monoclonal antibodies on mature fixed inclusions. Monoclonal antibodies were run individually or in noncross-reactive pools. Wells were then reacted with a fluorescein isothiocyanate conjugate and read with an epifluorescence microscope. Final reaction patterns were compared with prototype strain reaction patterns to determine serotype. Serovars were grouped in accordance with results of studies by Yuan et al, 12 Wang et al, 13 and Lampe et al 14: B complex (serovars B, Ba, D, Da, D-, E, L1, L2, L2a); intermediate group F/G (serovars F and G); and C complex (serovars A, C, H, I, Ia, I-, J, K, L3). The nonfusing chlamydial phenotype was detected visually by observing inclusion morphology during serotyping. 9
Study Design and Statistical Analysis
Statistical analyses were conducted with Stata software (Release 6.0; Stata Corp., College Station, TX). Separate analyses were performed for women and men. Association of chlamydial serovars to demographic and behavioral characteristics was assessed with the chi-square or Fisher exact test. Multivariate logistic regression was used to assess the relationship of specific signs and symptoms to chlamydial serovars, while controlling for the covariates age and race. Serovars representing <5% of the total isolates were analyzed only in their corresponding serovar class because of the effect of such small sample sizes on the ability to detect associations, if they exist.
Characteristics of the Study Population and Distribution of C trachomatis Serovars
Inclusion criteria were met by 480 women and 700 heterosexual men. The mean ages (± SD) of women and men were 22.2 (± 6.0) years and 26.5 (± 7.0) years, respectively. Among women, 34% were white, 32% black, and 34% other races. Among men, 24% were white, 52% black, and 24% other races. Of the women, 12% used oral contraceptives, 6% were pregnant, 8% had trichomoniasis, 36% had bacterial vaginosis, 14% reported a prior chlamydial infection, and 14% reported a prior gonococcal infection. In men, prior chlamydial or gonococcal infections were reported by 17% and 23%, respectively. Women (89%) and men (86%) were infected predominately with chlamydial serovars D, E, F, Ia, and J. Serovars recovered in women and men, respectively, included: B (both 1%), D (15% versus 17%), D− (both 2%), E (30% versus 33%), F (20% versus 17%), G (3% versus 4%), H (both 2%), I (0% versus 0.4%), Ia (12% versus 9%), J (13% versus 10%), and K (3% versus 5%). Serovar classes B complex (48% versus 53%), intermediate group F/G (23% versus 21%), and C complex (29% versus 26%) were recovered at similar frequencies in women and men, respectively.
Relationship of C trachomatis Serovars to Demographic, Behavioral, and Clinical Characteristics
Compared with women infected with other predominant serovars or serovar classes, women infected with serovar F and intermediate group F/G serovars were more often pregnant (P = 0.01 and P = 0.03, respectively) (Table 1). Women infected with serovars D and Ia were more likely to be older than 21 years than those infected with serovars E, F, and J (P = 0.05).
Compared with men infected with other predominant serovars or serovar classes, men infected with serovars E and F were more often younger than 26 years (P = 0.04) (Table 2). Distribution of races differed among serovars and serovar classes (P = 0.003 and P = 0.001, respectively), with men infected with C complex serovars more often black or white than other races, compared with those infected with B complex or intermediate group F/G serovars. A prior gonococcal infection was reported most often by men infected with C complex serovars, followed by those infected with B complex and then those infected with intermediate group F/G serovars (P = 0.01).
Relationship of C trachomatis Serovars to Specific Clinical Manifestations
After controlling for age and race on multivariate analysis, we noted that abdominal pain and/or dyspareunia was more often associated with infection due to serovar F in women, compared with infections due to serovars D, E, Ia, and J (P = 0.048) (Table 3). Other symptoms and signs were not associated with chlamydial serovar in women. Diagnosis of neither MPC nor PID was associated with chlamydial serovar.
For men, univariate analysis showed that those infected with serovar Ia more often reported urethral discharge than did those infected with serovars D, E, F, and J (P = 0.025) (Table 4). This association and other symptoms or signs investigated in men were not statistically significant after multivariate analysis was done after controlling for age and race.
The clinical manifestations of urogenital chlamydia are likely due to both the host immune response (cell-mediated and humoral-mediated response) and to virulence characteristics of the organism itself. Understanding the relationship of C trachomatis serovars to clinical manifestations of urogenital disease may improve our understanding of the epidemiology and pathogenesis of chlamydial infections. Prior studies assessing the relationship of non-LGV chlamydial serovars or genotypes to clinical manifestations have produced contradictory results (Table 5). Some reports have not described associations between chlamydial serovars and clinical or laboratory findings in women. 1–3 However, Morré et al 4 reported associations of serovar K infections with more frequent complaint of abnormal vaginal discharge and serovar Ia infections with asymptomatic disease. Workowski et al 5 reported lower frequencies of cervical friability, mucopus on examination, and mucopurulent cervicitis by colposcopy in women infected with intermediate group F/G serovars. van Duynhoven et al 6 reported that women infected with intermediate group F/G genotypes more often complained of lower abdominal pain. Similarly, in this study, women infected with serovar F reported abdominal pain and/or dyspareunia more often than those infected with other predominant serovars. Abdominal pain could be a clinical manifestation of endometritis and/or salpingitis. One hypothesis might be that if serovar F strains less often produce clinically evident MPC (as in the study of Workowski et al 5), they may be more likely to go unrecognized and thus could progress to infect the upper genital tract. In contrast to abdominal pain and/or dyspareunia, which was reported by 96 women, examination findings suggestive of PID, including adnexal tenderness (23 women), fundal tenderness (18 women), and cervical motion tenderness (15 women), were infrequent, and we were not able to detect relationships between these examination findings and chlamydial serovars.
Genotypic variation has been demonstrated within most known serovars. 7 It is possible that individual genotypes within known serovars could demonstrate different virulence potentials. Dean et al 7 reported that variant F genotypes were associated with symptomatic, severe endometrial disease on endometrial biopsy and E genotypes were associated with asymptomatic, milder infections in women.
In men, Morré et al 4 reported an association of serovar Ga with symptoms, especially dysuria, while van de Laar et al 3 reported men infected with intermediate group F/G serovars less often complained of urethral discharge and less often had ≥10 PMNs per hpf on gram-stained smear than men infected with other serovar classes. Batteiger et al 2 reported men with ≤3 PMNs per hpf on gram-stained smear had fewer isolates of intermediate group F/G serovars than men with ≥10 PMNs per hpf, while van Duynhoven et al 6 found men with genotype G/Ga infections were more likely to have <10 PMNs per Hpf and those with infections due to genotypes H and J more often complained of dysuria and urethral discharge. Our study was unable to identify associations of clinical or laboratory findings with chlamydial serovars in men.
The contradictory nature of results from prior studies may be due to several factors. Sample sizes may have been insufficient to reliably detect associations of clinical and laboratory findings with chlamydial serovars. In addition, some previous studies did not control for potential confounding factors. Adjusting for age and race in our analyses appeared important, as men infected with serovar Ia more often complained of urethral discharge and men with C complex serovars were somewhat more likely to have urethral discharge on examination on univariate analysis, but these associations were not significant on multivariate analysis. Age and race also had some influence on the relationship of clinical manifestations with serovars in women, though overall statistical significance was not affected.
Identifying infections with nonfusing C trachomatis phenotypes or concurrent gonorrhea and addressing them as potential confounders through exclusion or controlling for them in a multivariate analysis may also be important. Individuals with these infections were excluded from our study. Nonfuser C trachomatis strains occupy nonfusogenic inclusions and lack IncA protein in inclusion membranes. 9 A recent study of the clinical manifestations of infection with nonfusing C trachomatis strains demonstrated that women and men infected with these mutant strains were less likely to have signs of infection than were those infected with wild-type fusing strains, and women were also less likely to have symptoms of infection. 10 The prevalence of nonfusing chlamydial strains is about 1.5% in the PHSKC-STD Clinic population we studied, 9 but it is possible that infections caused by these strains may be more prevalent in other populations and, as a result, may have more influence on the results of studies relating clinical manifestations to chlamydial serovars in these populations.
When comparing patient characteristics by serovar, we found that women infected with serovar F were more often pregnant than were those infected with other serovars. Given our finding of more frequent abdominal pain and/or dyspareunia in women infected with serovar F, we explored the relationship of pregnancy to this finding. Pregnancy was not associated with abdominal pain and/or dyspareunia on univariate or multivariate analysis and also did not significantly contribute, on the basis of a likelihood ratio test, to the multivariate model assessing the relationship between serovars and abdominal pain and/or dyspareunia. It is unclear why women infected with serovar F were more often pregnant. Persson et al 1 reported that of women attending an obstetrics and gynecology clinic, those infected with B complex chlamydial serovars were more often pregnant (40%) than those infected with intermediate group F/G (27%) or C complex (33%) serovars. If intermediate group F/G serovars more often cause asymptomatic and chronic infections, they may be more likely to be discovered at pregnancy-associated screening for chlamydia. Additionally, perhaps pregnancy-associated immunosuppression allows proliferation and better detection of these strains.
Our study has limitations that should be noted. Because the population we studied consisted of a high-risk group of patients attending an STD clinic, our findings may not be generalizable to populations with different demographic or behavioral characteristics. Because our study was retrospective and involved data collection from standardized forms, we were unable to differentiate the individual symptoms of abdominal pain versus dyspareunia, as these two symptoms were recorded as one summary variable. Nondifferential misclassification may have been another reason for the contradictory nature of results of our study and other studies. This form of misclassification bias results from the degree of inaccuracy that characterizes how data are collected from a study population. The usual result of the misclassification is attenuation of significant findings, and one may be less likely to detect an association even if it really exists. 15 In our study and possibly others, the use of data derived from history and physical examinations from multiple clinicians may have led to misclassification bias. Having only one clinician collecting data in a standardized manner (as was done in the study of Workowski et al 5) would be ideal but is not often feasible.
In conclusion, we found that women who reported abdominal pain and/or dyspareunia were more often infected with serovar F. We did not identify an association of specific clinical manifestations with chlamydial serovars in men. Evaluating a large sample size and addressing potential confounding factors in the study design and statistical analyses strengthen the validity of our results. Misclassification bias may have influenced the findings of our study and other studies and should be minimized when possible. We conclude that clinical manifestations of C trachomatis urogenital infection are not strongly influenced by the infecting serovar.
1. Persson K, Osser S. Lack of evidence of a relationship between genital symptoms, cervicitis and salpingitis and different serovars of Chlamydia trachomatis
. Eur J Clin Microbiol Infect Dis 1993; 12: 195–199.
2. Batteiger BE, Lennington W, Newhall WJ, Katz BP, Morrison HT, Jones RB. Correlation of infecting serovar and local inflammation in genital chlamydial infections. J Infect Dis 1989; 160: 332–336.
3. van de Laar MJW, Lan J, van Duynhoven YTHP, et al. Differences in clinical manifestations of genital chlamydial infections related to serovars. Genitourin Med 1996; 72: 261–265.
4. Morré SA, Rozendaal L, van Valkengoed IGM, et al. Urogenital Chlamydia trachomatis
serovars in men and women with a symptomatic or asymptomatic infection: an association with clinical manifestations? J Clin Microbiol 2000; 38: 2292–2296.
5. Workowski KA, Stevens CE, Suchland RJ, et al. Clinical manifestations of genital infection due to Chlamydia trachomatis
in women: differences related to serovar. Clin Infect Dis 1994; 19: 756–760.
6. van Duynhoven YTHP, Ossewaarde JM, Derksen-Nawrocki RP, van der Meijden, van de Laar MJW. Chlamydia trachomatis
genotypes: correlation with clinical manifestations of infection and patients’ characteristics. Clin Infect Dis 1998; 26: 314–322.
7. Dean D, Oudens E, Bolan G, Padian N, Schachter J. Major outer membrane protein variants of Chlamydia trachomatis
are associated with severe upper genital tract infections and histopathology in San Francisco. J Infect Dis 1995; 172: 1013–1022.
8. Workowski KA, Suchland RJ, Pettinger MB, Stamm WE. Association of genital infection with specific Chlamydia trachomatis
serovars and race. J Infect Dis 1992; 166: 1445–1449.
9. Suchland RJ, Rockey DD, Bannantine JP, Stamm WE. Isolates of Chlamydia trachomatis
that occupy nonfusogenic inclusions lack IncA, a protein localized to the inclusion membrane. Infect Immun 2000; 68: 360–367.
10. Geisler WM, Suchland RJ, Rockey DD, Stamm WE. Epidemiology and clinical manifestations of unique Chlamydia trachomatis
isolates that occupy nonfusogenic inclusions. J Infect Dis 2001; 184: 879–884.
11. Eckert LO, Suchland RJ, Hawes SE, Stamm WE. Quantitative Chlamydia trachomatis
cultures: correlation of chlamydial inclusion forming units with serovar, age, gender, and race. J Infect Dis 2000; 182: 540–544.
12. Yuan Y, Zhang YX, Watkins NG, Caldwell H. Nucleotide and deduced amino acid sequences for the 4 variable domains of the major outer membrane proteins of the 15 Chlamydia trachomatis
serovars. Infect Immun 1989; 57: 1040–1049.
13. Wang SP, Grayston JT. Three new serovars of Chlamydia trachomatis
: Da, Ia, and L2a. J Infect Dis 1991; 163: 403–405.
14. Lampe MF, Suchland RJ, Stamm WE. Nucleotide sequence of the variable domains within the major outer membrane protein gene from serovariants of Chlamydia trachomatis
. Infect Immun 1993; 61: 213–219.
15. Gordis L. Epidemiology. Philadelphia: WB Saunders, 1996: 183–184.