Objective: Roles for Chlamydia trachomatis and Neisseria gonorrhoeae infections in pelvic inflammatory disease pathogenesis are well delineated; however, the etiologic contributions of herpes simplex virus type 2 (HSV-2) and Trichomonas vaginalis have been underexplored.
Goal: The goal of this study was to investigate the association between acute and plasma cell endometritis, fallopian tube obstruction, HSV-2 serology, and T. vaginalis infection.
Study Design: The authors conducted a cross-sectional secondary analysis of 736 women at risk for bacterial sexually transmitted diseases that used endometrial biopsy data obtained at enrollment as well as hysterosalpingography results obtained 12 weeks after enrollment.
Results: Women diagnosed with T. vaginalis at enrollment were more likely to have histologic evidence of acute endometritis. Both plasma cell and acute endometritis were significantly more common among women with positive serology HSV-2; furthermore, women coinfected with HSV-2 and C. trachomatis, N. gonorrhoeae, T. vaginalis, or bacterial vaginosis were much more likely to be diagnosed with acute endometritis than were women infected with HSV-2 or one of these pathogens alone. Among women with available HSV-2 serology and hysterosalpingogram results, HSV-2 was the only genital tract pathogen infection associated with fallopian tube obstruction.
Conclusions: Our analyses demonstrate that T. vaginalis infection and positive HSV-2 serology are associated with endometritis. Further work will be needed to determine the specific roles these pathogens may play in pelvic inflammatory disease pathogenesis.
Cross-sectional investigation of women at high risk for the acquisition of sexually transmitted diseases found that acute endometritis was more frequently detected among study participants with Trichomonas vaginalis infection, while both acute and plasma cell endometritis and fallopian tube blockage were significantly more common among those with positive HSV-2 serology.
From the *Department of Obstetrics, Gynecology, and Reproductive Sciences, the †Division of Molecular Diagnostics, and ‡Magee-Women’s Research Institute, University of Pittsburgh, Pittsburgh, Pennsylvania
The authors thank Christine M. Rullo and Lloyd Clarke for assistance in manuscript preparation.
This study was supported by the National Institute of Allergy and Infectious Diseases, National Institutes of Health (NIH [Bethesda, MD]) grants R01 AI41624 and 1 K23 AI064396-01 and the AIDS and Molecular Microbiology/Epidemiology Training Program, NIH (grant T32-AI07333 to TLC).
Correspondence: Thomas L. Cherpes, MD, Magee-Women’s Hospital, Room 3307, 300 Halket Street, Pittsburgh, PA 15213. E-mail: firstname.lastname@example.org
Received for publication December 25, 2005, and accepted February 23, 2006.
PELVIC INFLAMMATORY DISEASE (PID) IS defined as an inflammatory disorder of the female upper genital tract that includes variable combinations of endometritis, salpingitis, pelvic peritonitis, and tuboovarian abscess.1 Long-term sequelae of PID include tubal infertility, ectopic pregnancy, and chronic pelvic pain. Neisseria gonorrhoeae, Chlamydia trachomatis, and endogenous facultative and anaerobic bacteria are frequently isolated from the upper genital tract of women with acute PID, but often times no putative etiologic agent is recovered.2
Herpes simplex virus type 2 (HSV-2) has been detected in the endometrium and fallopian tubes of women with acute PID,3 but its role in PID pathogenesis remains largely unexplored. Prior investigations that suggested a possible association between HSV-2 and PID include a 1990 case–control study that found ectopic pregnancies were significantly more common among women with HSV-2,4 and a 2003 cross-sectional study that found women with a self-reported history of PID were approximately 63 times more likely to be HSV-2-seropositive than women who denied such history.5 In 1985, Lehtinen described a series of women with laparoscopically confirmed acute salpingitis who had HSV-2 isolated from their upper genital tracts.6 Nine years later, an observational study of 72 women with acute PID (verified by laparoscopy and endometrial histopathology) reported that 8 (11%) of the women had HSV-2 isolated from either their fallopian tubes or endometrium.7
These previous findings actuated our current search for possible correlations among HSV-2 infection, endometritis, and fallopian tube pathology among a large cohort of women at high risk for the acquisition of sexually transmitted diseases. This study cohort also provided an opportunity to investigate possible associations among Trichomonas vaginalis infection, endometritis, and PID.
Materials and Methods
From 1998 to 2000, we enrolled 736 women from Pittsburgh-area sexually transmitted disease and gynecology clinics into a cross-sectional investigation of the association between lower genital tract bacterial infections and PID. Although a prior publication involving this cohort used a convenience sample of 556 women in the analysis,8 for the current investigation, all study participants enrolled were included. Nonpregnant women between 15 and 30 years of age with purulent cervical discharge, untreated N. gonorrhoeae or C. trachomatis infection, symptomatic bacterial vaginosis (BV), or sexual contact with a male diagnosed with gonorrheal, chlamydial, or nongonococcal urethritis were eligible for participation. Women with a history of symptomatic PID or those who presented with a diagnostic triad suggestive of acute PID (lower abdominal pain, cervical motion tenderness, and adnexal tenderness) were not enrolled. However, symptomatic women not meeting all three criteria were eligible for participation.
After participants signed informed consent, demographic and behavioral interview data, vaginal fluid for the Gram stain diagnosis of BV and T. vaginalis culture, and endocervical specimens for N. gonorrhoeae culture and C. trachomatis polymerase chain reaction testing were obtained.
Each woman also underwent endometrial biopsy. Using the criteria developed by Kiviat et al, histologic acute endometritis was defined as the presence of ≥5 neutrophils per 400× field of endometrial superficial epithelium together with ≥1 plasma cell per 120× field of endometrial tissue, whereas plasma cell endometritis was defined as the presence of ≥1 endometrial stromal plasma cells in the absence of neutrophils.9
Study participants were also instructed to return approximately 12 weeks after enrollment for hysterosalpingograms (for the radiographic detection of fallopian tube patency or damage).
The methodologies used to detect HSV-2 seroprevalence in this study population have been previously described.10 Sera frozen and stored at −70°C were tested for the presence of type-specific HSV-2 antibodies using commercially available, U.S. Food and Drug Administration-approved enzyme-linked immunosorbent assays (ELISAs) (HerpeSelect 2; Focus Technologies, Cypress, CA). Kits were used according to manufacturer’s instructions, and serum specimens with index values ≥1.10 were considered positive (see package insert). Approval for the use of these stored serum samples for an investigation of the correlates of HSV-2 infection and the relationships among HSV-2 infection, endometritis, and PID was obtained from the Institutional Review Board of Magee-Womens Hospital in Pittsburgh, Pennsylvania.
The data were analyzed with SPSS statistical software, release 12.0.1 (SPSS, Inc., Chicago, IL). Univariate associations were examined between baseline characteristics and the histologic diagnosis of acute and plasma cell endometritis or fallopian tube blockage using, when appropriate, Pearson’s chi-square test, Chi-square test for linear trend, or Fisher exact test. All statistical tests were evaluated at the 0.05 level of significance. Variables with a P value of <0.1 were considered for inclusion in binary or multinomial logistic regression models. Multinomial logistic regression models were used to compare women with a histologic diagnosis of plasma cell or acute endometritis with those women without histologic evidence of endometritis, and binary logistic regression models were used to identify factors associated with fallopian tube blockage. Both the binary and multinomial logistic regression models were developed using forward stepwise regression based on the likelihood ratio test statistic. Variables were retained in the model if the Wald chi-squared test statistic had a P value of ≤0.05.
Among women with HSV-2 serology available, HSV-2 prevalence was 42.6% (308 of 725). The high prevalence of HSV-2 among women at high risk for genital tract infections is a similar frequency as that reported among women in a recent large-scale prevalence study from a number of sexually transmitted disease clinics in the United States.11
In the parent investigation (before inclusion of the HSV-2 serology data), acute histologic endometritis was found associated with C. trachomatis, N. gonorrhea, and BV, whereas plasma cell endometritis was not associated with any lower genital tract bacterial infection.8 We found that 47 of the 86 women (55%) diagnosed with acute endometritis had antibodies against HSV-2, whereas 70 of the 136 women (51%) diagnosed with plasma cell endometritis were HSV-2-seropositive. By univariate analysis, our investigation again detected significant associations between acute endometritis and C. trachomatis, N. gonorrhea, and BV, as well as with T. vaginalis infection, serum HSV-2 antibodies, and black race (Table 1). On the other hand, we found no statistically significant associations between the diagnosis of acute endometritis and age of the study participant, annual household income, number of lifetime sexual partners, introduction of a new sex partner in the 3 months before enrollment, frequency of sexual intercourse, or a self-reported history of N. gonorrhea, C. trachomatis, or BV (Table 1).
The current multinomial logistic regression model analyses included all 595 study participants who had complete test results available for the enrollment diagnoses of C. trachomatis, N. gonorrhea, T. vaginalis, BV, and HSV-2. In these analyses, we found that while controlling for the effects of race, statistically significant, independent associations between acute endometritis and C. trachomatis, N. gonorrhea, T. vaginalis, and HSV-2 were retained in our model (Table 2). Also, while controlling for the effects of race, HSV-2 was the sole reproductive tract infection significantly associated with plasma cell endometritis (odds ratio [OR] = 1.5; 95% confidence interval [CI] = 1.0–2.3; P = 0.05) (Table 2).
In several African countries12,13 and in a cohort of men in the United States with a low prevalence of HSV-2,14 the Focus HSV-2 ELISA, compared with Western blot, had a higher frequency of false-positive results. In these populations, false-positives were detected more frequently from sera that had ELISA index values in the low positive range (1.1–3.5) compared with index values ≥3.5.15 Although use of the higher positive index value increased specificity and positive predictive value of the assay, its sensitivity was decreased.13,14 When our multinomial logistic regression analyses were repeated using a higher cutoff (≥3.5) for the positive index value, the adjusted odds ratios continued to demonstrate positive associations between HSV-2 antibody and both acute and plasma cell endometritis, but statistical significance was lost (Table 2).
Additional analyses were then performed to investigate if coinfection with HSV-2 and another lower genital tract pathogen increased the likelihood for the histologic diagnosis of acute endometritis. These analyses demonstrated that the odds ratio of acute endometritis increased from 5.0 with chlamydial infection and 2.6 for HSV-2 infection alone to an odds ratio of 7.3 among women coinfected with both C. trachomatis and HSV-2 (Fig. 1). Similarly, the odds ratio for a histologic diagnosis of acute endometritis increased from 4.2 with gonorrheal infection and 2.1 infection for HSV-2 alone to an odds ratio of 6.0 when coinfection with N. gonorrhoeae and HSV-2 was present (Fig. 1). Analogous results were also seen for BV and T. vaginalis infections because the odds ratios for acute endometritis were significantly higher among women coinfected with HSV-2 and either of these lower genital tract infections in comparison to the odds ratios for acute endometritis when BV or trichomoniasis alone was present (Fig. 1).
To explore the association between HSV-2 infection and fallopian tube pathology, we next analyzed results from the 471 study participants who had both HSV-2 and hysterosalpingogram data available, and found a statistically significant association between HSV-2 infection and fallopian tube obstruction (OR = 1.7; 95% CI = 1.0–2.8; P = 0.04) (data not shown). Furthermore, when the analysis was reperformed using a higher cutoff (≥3.5) for the HSV-2 ELISA positive index value, we found that although 22% (38 of 176) of women with positive HSV-2 serology had evidence of fallopian tube blockage on hysterosalpingogram, only 11% (33 of 295) of women with negative HSV-2 serology had fallopian tube blockage (OR = 2.2; 95% CI = 1.3–3.6; P = 0.003) (data not shown). No other reproductive tract infections in our analyses were found to be significantly associated with fallopian tube obstruction.
Our investigation among young, nonpregnant women at high risk for other sexually transmitted diseases demonstrates that acute endometritis is more often detected among women diagnosed with C. trachomatis, N. gonorrhoeae, T. vaginalis, and HSV-2. In addition, women coinfected with HSV-2 and one of these lower genital tract diseases are more likely to be diagnosed with acute endometritis than are women infected with either HSV-2 or one of these diseases alone. HSV-2 is also the only genital tract infection in our investigation significantly associated with plasma cell endometritis, and we report a previously unidentified correlation between HSV-2 infection and fallopian tube obstruction.
Both endometritis and PID have polymicrobial etiologies, and an ascending route of infection is the shared etiologic mechanism by which lower genital tract infections are thought to cause disease. The most frequent nonobstetric cause of acute endometritis is C. trachomatis; although weaker, but positive, associations are present with gonococcal infections and BV.16–18 Similarly, C. trachomatis and N. gonorrhoeae are isolated from at least one third of the women with PID, and BV-associated organisms are cultured more often from the upper genital tract of women with PID.19–24
We previously reported that a histologic diagnosis of acute endometritis was more common among women with C. trachomatis and N. gonorrhoeae infections.8 This study extends these findings by detecting additional associations between acute endometritis and T. vaginalis and HSV-2. Although no direct causal connection between T. vaginalis infection and PID is established, trichomonads have been isolated from the fallopian tubes and peritoneal cavity of women with acute salpingitis.25 Also, an association between vaginal trichomoniasis and PID among women colonized with C. trachomatis has been reported.26 Among HIV-infected women attending a sexually transmitted disease clinic in South Africa, patients with trichomoniasis had a significantly higher risk of PID than those without trichomoniasis.27 Finally, the risk of tubal infertility was almost twice as high in women with a self-reported history of trichomoniasis compared with women who denied such infection.28 As the evidence for a role of T. vaginalis in the etiology of PID mounts, it is clear that further research is needed to better delineate its specific involvement.
In our current investigation, positive HSV-2 serology is associated with the histologic diagnosis of acute endometritis. As mentioned previously, other studies have suggested the possibility of an association between HSV-2 and PID,3–7 and the results of our investigation support the plausibility of these earlier observations. It should be emphasized, however, that our cross-sectional investigation is not designed to answer questions related to HSV-2 and PID pathogenesis, and therefore, our results are not conclusive.
Our investigation also detects an association between HSV-2 infection and plasma cell endometritis. The presence of plasma cells in the endometrial stroma, in the absence of other inflammatory cells, is considered a characteristic of chronic PID.29 However, chronic PID remains a poorly defined entity. For example, there is no specific clinical syndrome associated with the detection of endometrial plasma cells.30 Furthermore, plasma cells can be found in the endometria of healthy women with no apparent upper genital tract infection.31 Our observations suggest that HSV-2 infection is associated with the presence of plasma cells in the endometrium and may help clarify why plasma cells are sometimes detected in the endometrium of otherwise healthy women. Future studies will be needed to determine how often the presence of endometrial plasma cells represents a biologic marker for genital herpes infection or if their presence is associated with adverse clinical outcomes.
Our study also suggests mechanisms by which HSV-2 infection may result in fallopian tube damage. Genital herpes infections are characterized by frequent lower genital tract ulcerations, and these ulcers could disrupt the endocervical canal barrier and increase the spread of lower genital tract pathogens to the upper genital tract, facilitating PID. This scenario is consistent with our study that demonstrates that coinfection of HSV-2 with C. trachomatis, N. gonorrhoeae, T. vaginalis, or BV all increase the likelihood of acute endometritis. On the other hand, genital herpes infections may result in the increased presence of inflammatory cells in the upper genital tract, and host inflammatory responses to the virus could increase the likelihood for fallopian tube damage. Our results, although both intriguing and hypothesis-generating, cannot determine a mechanism by which HSV-2 may contribute to upper genital tract pathology.
Our study has several other limitations. We are obviously limited by the cross-sectional design of the parent investigation. It is not possible to determine if the presence of inflammatory cells at the time of endometrial biopsy is a consequence of a genital herpes infection, a lower genital tract infection, or PID itself. Similarly, we cannot determine when HSV-2 was acquired in relation to the acquisition of another genital infection and therefore cannot determine the direct effects of HSV-2 infection on PID pathogenesis. Furthermore, it is possible that the higher prevalence of T. vaginalis among women with acute endometritis and the higher HSV-2 prevalence among women with acute and plasma cell endometritis or fallopian tube blockage may represent a marker for sexual activity or the coacquisition of a PID-associated sexually transmitted disease rather than an etiologic role for the HSV-2 in PID. Hysterosalpingography, the methodology used in this investigation, has a high specificity but low sensitivity for the diagnosis of fallopian tube obstruction.32 Its low sensitivity makes it likely that some cases of tubal impatency were missed with our reliance on this procedure. However, its high specificity makes it less likely we have incorrectly diagnosed women with tubal obstruction.33
Our findings may also be limited by the fact that the associations between HSV-2 and acute and plasma cell endometritis lost statistical significance with use of the higher positive ELISA index value. Although there is justification for use of higher positive index values in some populations,13,14 its relevancy to our cohort of women in the United States with a high prevalence of HSV-2 is uncertain. When the standard positive index value (≥1.1) was applied to another cohort of women in the United States with a HSV-2 seroprevalence (44%) quite similar to ours, the Focus HSV-2 ELISA sensitivity and specificity were 98% and 95%, respectively, compared with a confirmatory immunoblot assay.34 Although we might anticipate a similarly performance of the ELISA in our investigation, we are unable to determine the proportion of women with positive HSV-2 serology who actually have disease.
In conclusion, we found that current T. vaginalis infection is associated with acute endometritis, and that positive HSV-2 serology is associated with acute and plasma cell endometritis as well as fallopian tube obstruction. Clearly, further research will be needed to validate these results. It is also clear that our findings may have considerable clinical significance. For example, if HSV-2 ulceration results in disruption of the endocervical canal barrier, facilitating bacterial influx from the lower to upper genital tract, or if local inflammatory changes associated with genital herpes reactivation and shedding increase the likelihood that tubal scarring occurs, then long-term antiviral suppressive therapy could be expected to decrease the incidence of PID and its adverse sequelae.
1. Centers for Disease Control and Prevention. Pelvic inflammatory disease. Sexually transmitted disease treatment guidelines—2002. MMWR Morb Mortal Wkly Rep 2002; 51:78–84.
2. Svenson L, Westrom L, Ripa KT, Mardh PA. Differences in some clinical and laboratory parameters in acute salpingitis related to culture and serologic findings. Am J Obstet Gynecol 1980; 138:1017–1021.
3. Paavonen J, Teisala K, Heinonen PK, et al. Endometritis and acute salpingitis associated with Chlamydia trachomatis
and herpes simplex virus type two. Obstet Gynecol 1985; 65:288–291.
4. Tuomivaara LM. Ectopic pregnancy and genital infections: A case control study. Ann Med 1990; 22:21–24.
5. Evans BA, Kell PD, Bond RA, MacRae KD, Slomka MJ, Brown DWG. Predictors of seropositivity to herpes simplex virus type 2 in women. Int J STD AIDS 2003; 14:30–36.
6. Lehtinen M, Rantala I, Teisala K, et al. Detection of herpes simplex virus in women with acute pelvic inflammatory disease. J Infect Dis 1985; 152:78–82.
7. Heinonen PK, Miettinen A. Laparoscopic study on the microbiology and severity of acute pelvic inflammatory disease. Eur J Obstet Gynecol Reprod Biol 1994; 57:85–89.
8. Wiesenfeld HC, Hillier SL, Krohn MA, et al. Lower genital tract infection and endometritis: Insight into subclinical pelvic inflammatory disease. Obstet Gynecol 2002; 100:456–463.
9. Kiviat NB, Wølner-Hanssen P, Eschenbach DA, et al. Endometrial histopathology in patients with culture-proved upper genital tract infection and laparoscopically diagnosed acute salpingitis. Am J Surg Pathol 1990; 14:167–175.
10. Cherpes TL, Meyn LA, Hillier SL. Plasma versus serum for the detection of herpes simplex virus type 2-specific immunoglobulin G antibodies with a glycoprotein G2-based enzyme immunoassay. J Clin Microbiol 2003; 41:2758–2759.
11. Gottlieb SL, Douglas JM, Schmidt DS, et al., for the Project RESPECT Study Group. Seroprevalence and correlates of herpes simplex virus type 2 infection in five sexually transmitted-disease clinics. J Infect Dis 2002; 186:1381–1389.
12. Hogrefe W, Su X, Song J, Ashley R, Kong L. Detection of herpes simplex virus-2 immunoglobulin G antibodies in African sera by using recombinant gG2, Western blotting, and gG2 inhibition. J Clin Microbiol 2002; 40:3635–3640.
13. Morrow RA, Nolkamper J, Robinson NJ, Bishop N, Smith J. Performance of Focus ELISA tests for herpes simplex virus type 1 (HSV-1) and HSV-2 antibodies among women in ten diverse geographic locations. Clin Microbiol Infect 2004; 10:530–536.
14. Golden MR, Ashley-Morrow R, Swenson P, et al. Herpes simplex virus type 2 (HSV-2) Western blot confirmatory testing among men testing positive for HSV-2 using the Focus enzyme-linked immunosorbent assay in a sexually transmitted disease clinic. Sex Transm Dis 2005; 32:771–777.
15. Ashley-Morrow R, Friedrich D, Meier A, Corey L. Use of ‘biokit HSV-2 Rapid Assay’ to improve the positive predictive value of Focus HerpeSelect HSV-2 ELISA. BMC Infect Dis. 2005; 5:84.
16. Paukku M, Puolakkainen M, Paavonen T, Paavonen J. Plasma cell endometritis is associated with Chlamydia trachomatis
infection. Am J Clin Pathol 1999; 112:211.
17. Hillier SL, Kiviat NB, Hawes SE, et al. Role of bacterial vaginosis-associated microorganisms in endometritis. Am J Obstet Gynecol 1996; 175:435.
18. Peipert JF, Montagno AB, Cooper AS, Sung CJ. Bacterial vaginosis as a risk factor for upper genital tract infection. Am J Obstet 1997; 177:1184.
19. Eschenbach DA, Buchanan TM, Pollock HM, et al. Polymicrobial etiology of acute pelvic inflammatory disease. N Engl J Med 1975; 293:166–71.
20. Forslin L, Falk V, Danielsson D. Changes in the incidence of acute gonococcal and nongonococcal salpingitis. Br J Vener Dis 1978; 554:247–250.
21. Thompson SE, Hager WD, Wong KH, et al. The microbiology and acute therapy of acute pelvic inflammatory disease in hospitalized patients. Am J Obstet Gynecol 1980; 136:179–86.
22. Bowie WR, Jones H. Acute pelvic inflammatory disease in outpatients: Association with Chlamydia trachomatis
and Neisseria gonorrhoeae
. Ann Intern Med 1981; 95:685–688.
23. Brunham RC, Binns B, Guijon F, et al. Etiology and outcome of acute pelvic inflammatory disease. J Infect Dis 1988; 158:510–517.
24. Haggerty CL, Hillier SL, Bass DC, et al., for the PID Evaluation and Clinical Health (PEACH) Study Investigators. Bacterial vaginosis and anaerobic bacteria are associated with endometritis. Clin Infect Dis 2004; 39:990–995.
25. Märdh P, Weström L. Tubal and cervical cultures in acute salpingitis with special reference to Mycoplasma hominis
and T strain mycoplasmas. Br J Vener Dis 1970; 46:179.
26. Paisarntantiwong R, Brockmann S, Clarke L, Landesman S, Feldman J, Minkoff H. The relationship of vaginal trichomoniasis and pelvic inflammatory disease among women colonized with Chlamydia trachomatis
. Sex Transm Dis 1995; 22:344–347.
27. Moodley P, Wilkinson D, Connolly C, Moodley J, Sturm AW. Trichomonas vaginalis
is associated with pelvic inflammatory dis-ease in women infected with human immunodeficiency virus. Clin Infect Dis 2002; 34:519–522.
28. Grodstein F, Goldman M, Cramer D. Relation of tubal infertility to history of sexually transmitted diseases. Am J Epidemiol 1993; 137:577–584.
29. Greenwood SM, Moran JJ. Chronic endometritis: Morphologic and clinical observations. Obstet Gynecol 1981; 58:176–184.
30. Heatley MK. The association between clinical and pathological features in histologically identified chronic endometritis. J Obstet Gynaecol 2004; 24:801–803.
31. Achilles SL, Amortegui AJ, Wiesenfeld HC. Endometrial plasma cells: Do they indicate subclinical pelvic inflammatory disease? Sex Transm Dis 2005; 32:185–188.
32. Mol BW, Swart P, Bossuyt PM, van Beurden M, van der Veen F. Reproducibility of the interpretation of hysterosalpingography in the diagnosis of tubal pathology. Hum Reprod 1996; 11:1204–1208.
33. Swart P, Mol BW, van der Veen F, van Beurden M, Redekop WK, Bossuyt PM. The accuracy of hysterosalpingography in the diagnosis of tubal pathology: A meta-analysis. Fertil Steril 1995; 64:486–491.
34. Turner KR, Wong E, Kent CK, Klausner JD. Serologic herpes testing in the real world: Validation of new type-specific serologic herpes simplex virus tests in a public health laboratory. Sex Transm Dis 2002; 29:422–425.