The number of women diagnosed with AIDS has increased over the last 10 years such that, in 1997, 20% of all AIDS patients in Europe were women . The predominant mode of HIV transmission in women is sexual intercourse. In addition to the risk of HIV infection, women are also at risk for other sexually transmitted infections (STI). Several studies showed that sexual transmission of HIV is enhanced by co-infection with STI [2–4] and that HIV-infected persons may experience prolonged duration and higher recurrence of STI. HIV-infected women have been frequently diagnosed with vaginal candidiasis [5,6] and bacterial vaginosis [7,8]. Gonorrhoea, trichomoniasis, genital herpes and genital warts are more frequent in HIV-infected women, in comparison with women who are not infected with HIV [9,10]. HIV-infected women play an important role in the further spread of HIV. As the transmission of HIV could be enhanced by co-existence of an STI, the prevention and early treatment of STI in these women can reduce their infectious role. Investigation of risk factors for the occurrence of STI within the group of HIV-infected women might give more insight to the specific groups at risk. Although risk factors for STI have been studied in HIV-infected women, to our knowledge, none of the studies have taken into account the duration of infection.
The primary objective of the study was to estimate the prevalence and incidence of STI in a group of European HIV-infected women with a known interval of seroconversion. In addition, we investigated behavioural factors associated with the occurrence of STI and the relation between STI and HIV disease progression.
The European study on the natural history of HIV infection in women is a prospective multicentre cohort of HIV-infected women with a known interval of seroconversion. In total, 487 women have been included since 1993 from 31 centres in twelve European countries. Women were asked about (lifetime) obstetric and gynaecological history, sexual behaviour, history of STI, long-term treatment for HIV and other infections, and socio-demographic characteristics. Follow-up visits were scheduled every 6 months. At each visit a clinical examination was performed including examination for STI, assessment of the progression of HIV infection and a Pap smear. Blood was drawn at each visit to determine immunologic and virologic markers.
All women were HIV positive when they entered the study and their interval of seroconversion was determined retrospectively based on the existence of a negative HIV test less than 2 years before a positive HIV test or on one limited unbroken period (maximum 2 consecutive years) of reported HIV risk behaviour since 1980. The midpoint of the seroconversion interval (estimated as the interval between the last HIV negative and the first HIV positive test or between the start-date and stop-date of a period with risk behaviour for HIV) was used as the date of seroconversion. Because the effect of time since seroconversion on the occurrence of STI was the subject of the present study, we decided to limit the current analyses to women with a seroconversion interval based on serological tests. In addition, centres with more than 50% lost to follow-up were excluded. In total, 229 of 487 women were included in the present study.
Diagnostic criteria and procedures
A positive wet mount was used to diagnose the presence of vaginal candidiasis and trichomoniasis. Both genital ulcerations and genital warts were diagnosed clinically during gynaecological examination. Fifty per cent of the diagnoses of gonorrhoea were based on Gram stain detection and the other 50% on a positive culture. Primary genital ulcerations and primary genital warts were defined as the first known episode of that infection. Chlamydia trachomatis (chlamydial) infection was diagnosed in case of a positive enzyme-linked immunosorbent assay, polymerase chain reaction/LCR or IF performed on an endocervical sample depending on the test available in the centre. Of the 28 centres, 11 centres screened systematically for all three acute STI (chlamydial infection, trichomoniasis and gonorrhoea), six centres for one or two acute STI and 11 centres screened only in case of symptoms. Sixteen centres screened systematically for vaginal candidiasis, whereas 12 centres did not. However, no differences in prevalence and incidence were observed between centres screening systematically and centres screening in case of symptoms. All centres screened systematically on genital warts and genital ulcerations.
Self-reported data on STI were gathered over the year prior to study entry and over the period since the last study visit. These data were combined with clinically diagnosed STI. Therefore, women could have more than one episode of a specific STI in the period between two consecutive visits assuming that they received treatment and were cured from the first episode. An exception was made for genital warts by counting only one episode per visit since warts may persist longer.
Time-dependent covariates related to HIV disease progression were CD4 cell count and time since seroconversion. To ensure the right direction of cause (i.e. immunosuppression) and effect, CD4 cell counts were taken from the visit preceding assessment of an STI. The occurrence of AIDS, the use of antiretroviral treatment, Pneumocystis carinii pneumonia (PCP) prophylaxis and prophylactic treatment for oral and vaginal candidiasis were treated as time-dependent covariates. The following methodological covariates were examined: lag time between seroconversion and study entry, setting of the centre, and methods of screening. Socio-demographic covariates included age at seroconversion, continent of birth, marital status, migration, age at end of education and risk group. Geographic region was divided into three categories: South (Italy, Portugal, Spain), Central (Belgium, France, The Netherlands, Switzerland) and North (Denmark, Finland, Norway, Sweden). Sexual behaviour was measured through the number of past sexual partners, history of STI, history of prostitution and age at first vaginal intercourse. Number of partners, condom use with regular and casual partners, use of contraceptive pill and engagement in unprotected sex with an HIV-positive partner since the last study visit were treated as time-dependent covariates.
The prevalence of specific STI was calculated at entry for 229 women. The incidence and risk factor analyses were based on follow-up visits. Since not all women had a follow-up visit, 192 women with 639 visits (until July 1998) were available for analyses. The prevalence of STI at entry did not differ significantly between women with and without follow-up visits.
Because of the small numbers of specific STI, acute STI were grouped together and other STI were grouped together to study risk factors for the occurrence of these infections. The group of acute STI consisted of chlamydial infection, trichomoniasis and gonorrhoea. Other STI included genital ulcerations and warts. Vaginal candidiasis was considered as a separate group.
Incidence, which was expressed as number of episodes per 100 person-years, and 95% confidence intervals (95% CI) were calculated using Poisson regression. Subjects were considered to be at risk from entry until the end of the follow-up period. The generalized estimating equation method (GEE) was used to evaluate covariates associated with the occurrence of an STI. This method corrects for dependency between multiple visits of one woman using a covariance structure that assumed an equal correlation between all visits of one woman. Since a specific STI could be diagnosed twice between two visits and to take into account person-years of follow-up, we used a Poisson distribution in the GEE model . Covariates which were significantly associated with an STI in univariate analyses were selected for multivariate analyses together with covariates that appeared to be significant after controlling for CD4 cell count or time since seroconversion. Both time since seroconversion and CD4 cell count were included in multivariate analyses, independently of the results of the univariate analyses. The `final' multivariate model appeared after a stepwise backward procedure. Interactions between covariates that were included in the multivariate models were also examined (P < 0.05). In addition, all non-significant covariates were added once more to the final model to investigate possible confounding.
Of the 229 HIV-infected women with a known interval of seroconversion, 69% were infected through heterosexual contact (Table 1). Almost half of the women lived in southern Europe. The median CD4 cell count at entry was 417 × 106 cells/l [interquartile range (IQR), 280–642 × 106 cells/l]. At entry, 17% of the women received prophylactic treatment for PCP and 4% received oral treatment for candidiasis. The median number of women included per centre was 5 (IQR, 2–14), the median interval of seroconversion was 7 months (IQR, 3–12 months) and the median number of visits per woman was three (IQR, 1–5). The median interval between visits was six months (IQR, 5–8 months). The number of partners since the last study visit decreased with ongoing time since seroconversion, whereas the percentage of women reporting unprotected sexual intercourse remained stable over time since seroconversion.
Prevalence and incidence
Table 2 shows the prevalence and incidence of acute and other STI, and vaginal candidiasis based on self-report and clinical examination. At entry, 15% of the women were diagnosed with at least one acute STI, 10% with at least one other STI and 13% with vaginal candidiasis. Chlamydial infection was the most frequently diagnosed and reported acute STI during follow-up with 6.1 (95% CI, 3.9–8.8) diagnosed and 3.5 (95% CI, 2.0–5.7) reported episodes per 100 person–years. The incidence of vaginal candidiasis was high: 24 diagnoses and 41 self-reports of episodes per 100 person–years. Most genital ulcerations were reported, whereas most genital warts were diagnosed.
As described in the methods, STI were grouped into acute STI (chlamydial infection, trichomoniasis and gonorrhoea), other STI (genital warts and genital ulcerations) and vaginal candidiasis since the number of each specific infection alone was insufficient. In addition to time since seroconversion and CD4 cell count, only covariates significantly associated in univariate analyses are shown in Tables 3, 4 and 5. No covariates including sexual behaviour seriously confounded the effects presented in the three multivariate models and no statistically significant interactions were found.
In univariate analyses, the risk of an acute STI decreased with ongoing time since seroconversion and a decreasing CD4 cell count, but only significantly for a count below 200 × 106 cells/l (Table 3). Furthermore, use of antiretroviral therapy, use of PCP prophylaxis, number of lifetime partners, history of prostitution, number of partners since the last study visit and irregular condom use with either casual or regular partners were all significantly associated with an acute STI in univariate analyses. After correction for time since seroconversion, older age at end of education significantly decreased the risk of an acute STI. In multivariate analyses, only history of prostitution and irregular condom use with casual partners remained independently associated with an acute STI. Women with a history of prostitution had a two-fold higher risk (95% CI, 1.20–3.33) compared to women without one. Women who used condoms in less than 50% of their sexual contacts with casual partners had an almost eight-fold higher risk (95% CI, 3.52–17.0), and those who used condoms in 50–99% of the contacts had a two-fold higher risk (95% CI, 1.01–4.00) than women who did not have a casual partner. The risk of an acute STI still decreased with ongoing time since seroconversion and a decreasing CD4 cell count, although relative risks increased towards unity after correction for sexual behaviour.
Ongoing time since seroconversion and a CD4 cell count below 200 × 106 cells/l increased the risk of other STI in univariate analyses, although not significantly (Table 4). Women who had AIDS or who used PCP prophylaxis and women with no more than one partner since the last study visit were at increased risk for other STI in univariate analyses. After correction for CD4 cell count and time since seroconversion, unprotected sex with an HIV-positive partner appeared to be significantly related with other STI. Multivariate analyses showed that only having unprotected sex with an HIV-positive partner since the last study visit remained significantly associated with other STI with a relative risk of 2.23 (95% CI, 1.02–4.85). In comparison with univariate analyses, the relative risk increased from unity for time since seroconversion and towards unity for CD4 cell counts below 200 × 106 cells/l after correction for unprotected sex with an HIV-positive partner. However, these associations were still not significant.
In univariate analyses, both increasing time since seroconversion and a long lag time between seroconversion and entry were significantly but inversely associated with the risk of vaginal candidiasis (Table 5). Prescribed treatment for vaginal candidiasis was positively associated with vaginal candidiasis indicating the recurrent character of this condition. Neither the use of oral treatment for candidiasis nor CD4 cell count were significantly associated with vaginal candidiasis in both univariate and multivariate analyses. In multivariate analyses, the relationship with time since seroconversion remained significant. Also, prescribed treatment for vaginal candidiasis at the preceding study visit remained positively associated with the occurrence of a recurrent infection (relative risk, 1.50; 95% CI, 1.02–2.21).
Although the women knew that they were HIV positive, transmission of acute STI occurred frequently at entry and during follow-up. The occurrence of other STI and vaginal candidiasis was also high. At entry, 15% of the women were diagnosed with at least one acute STI, 10% with at least one other STI and 13% with vaginal candidiasis. The prevalence at study entry might be relatively high for several reasons. First, the reason for study entry could have been related to the presence of an STI, especially when this infection was symptomatic. Unfortunately, information on the reason for study entry is lacking. Second, the first study visit could have been the first medical check-up for a long time and therefore the presence of infections might be high since asymptomatic infections could exist for an extended period of time. Nevertheless, the prevalence of STI in our study is comparable to those observed in previous studies among HIV-infected women [8–10,12–15].
Women with a history of prostitution or irregular condom use with casual partners were at increased risk of an acute STI, which is an indication for the validity of self-reported sexual behaviour. These results are comparable with other studies among HIV-infected persons showing a relationship between sexual behaviour and acute STI [10,13,16]. In contrast with some studies [14,15], we did not find differences between women who became infected through heterosexual intercourse and through intravenous drug use. The results of the present study suggest that the risk of an acute STI decreased with ongoing time since seroconversion and with a decrease in CD4 cell count after controlling for sexual behaviour, although it was not statistically significant. Both covariates could have been a surrogate marker for sexual risk behaviour assuming that women become less sexually active with ongoing time since infection because after correction for sexual behaviour the relative risks changed towards one. However, the decreased risk with ongoing time since seroconversion and decreasing CD4 cell count in multivariate analyses might indicate a residual confounding which suggests that not all relevant sexual behaviour was (adequately) measured.
In contrast, the occurrence of other STI was increased with ongoing time since seroconversion and a CD4 cell count below 200 × 106 cells/l, but not significantly. Once a woman is infected (through sexual intercourse) with human papilloma virus, which might cause genital warts, or with herpes simplex virus, which might cause genital ulcerations, reactivation of these infections could take place especially in immunosuppressed women [10,17,18]. In addition, the occurrence of other STI was higher in women who had unprotected sexual contacts with an HIV-positive partner. One explanation might be that consistent re-infection and/or chronic irritation, due to unprotected sexual contacts, induces ganglion triggering that results in the recurrence of ulcerations or the persistence of genital warts .
Vaginal candidiasis was not associated with sexual risk behaviour, which is in accordance to the literature. Women who received treatment for vaginal candidiasis were at increased risk for vaginal candidiasis indicating the recurrent character of this condition. The occurrence of vaginal candidiasis significantly decreased with time since HIV infection even after correction for treatment for vaginal and systemic candidiasis. It is known that oral treatment for candidiasis prevents vaginal candidiasis in immunosuppressed HIV-infected women . So, it might be that prescription of oral treatment for candidiasis, which increased with ongoing time since HIV infection, was not consequently reported. Indeed, additional analyses showed that oral treatment was only reported for 17% of the women with oesophageal candidiasis. Our finding is in contrast with studies demonstrating that the occurrence of vaginal candidiasis increased with a decreasing CD4 cell count [6,14,21]. However in line with our findings, a study of Imam et al.  demonstrated that Candida often infects vaginal mucosa when there is no significant reduction in CD4 cell counts.
Several limitations of the present study should be noted. First, since not all centres screened systematically on STI and did not use a sensitive test for the detection of chlamydial infections and gonorrhoea, the prevalence and incidence could have been underestimated. However, no significant differences in incidence of STI were found between centres screening systematically and centres screening only in case of symptoms. Second, we combined clinically diagnosed STI with self-reported events. It has been shown that self-reported infections are reliable  although this appeared to be untrue for HIV-positive drug-using prostitutes . Since only 6% of the women in our study were drug-using prostitutes, we think it is very unlikely that our results have been affected seriously. Self-reported and diagnosed infections were treated as separate infections, assuming that a woman was treated and cured within the period between two visits. Nevertheless, it might be possible that a woman reported an STI that was previously diagnosed. However, the number of visits at which a specific STI was diagnosed followed by a visit at which the same STI was reported was small (less than 10%). This is the reason why overestimation of prevalence and incidence is unlikely. For genital warts, we counted only one episode per visit since these infections could persist for a long time. Nevertheless, it cannot be ruled out that the same episode was measured twice.
In conclusion, among HIV-infected women the occurrence of other STI and vaginal candidiasis was high and, although women knew their serostatus, acute STI were frequently acquired. Results suggested that the risk of acute STI and vaginal candidiasis decreased and the risk of other STI increased with ongoing time since seroconversion. The present study was the first study that describes these associations and therefore these findings should be confirmed by other seroconverter studies. Transmission of HIV is enhanced by co-infection with STI. Therefore, we strongly recommend regular gynaecological examinations of and early treatment for HIV-infected women next to prevention measures, particularly in those with a history of prostitution and irregular condom use, to reduce morbidity and further HIV transmission.
We thank all women registered in the European multicenter study on the natural history of HIV-infection in women for their ongoing participation, and Roel Coutinho, Han Fennema, Udi Davidovich, Miranda Langendam and Françoise Hamers for critical reading of the manuscript.
2. Wasserheit JN. Epidemiological synergy.
:Interrelationships between human immunodeficiency virus infection and other sexually transmitted diseases.
Sex Transm Dis 1992, 19: 61 –77.
3. Wald A, Corey L, Handsfield HH, Holmes KK. Influence of HIV infection on manifestations and natural history of other sexually transmitted diseases.
Annu Rev Public Health 1993, 14: 19 –42.
4. Ghys PD, Fransen K, Diallo MO. et al
. The associations between cervicovaginal HIV shedding, sexually transmitted diseases and immunosuppression in female sex workers in Abidjan, Cote d'Ivoire.
AIDS 1997, 11: F85 –F93.
5. Imam N, Carpenter CC, Mayer KH, Fisher A, Stein M, Danforth SB. Hierarchical pattern of mucosal candida infections in HIV-seropositive women.
Am J Med 1990, 89: 142 –146.
6. Rhoads JL, Wright DC, Redfield RR, Burke DS. Chronic vaginal candidiasis in women with human immunodeficiency virus infection.
JAMA 1987, 257: 3105 –3107.
7. Sewankambo N, Gray RH, Wawer MJ. et al
. HIV-1 infection associated with abnormal vaginal flora morphology and bacterial vaginosis.
Lancet 1997, 350: 546 –550.
8. Frankel RE, Selwyn PA, Mezger J, Andrews S. High prevalence of gynecologic disease among hospitalized women with human immunodeficiency virus infection.
Clin Infect Dis 1997, 25: 706 –712.
9. Chirgwin KD, Feldman J, Augenbraun M, Landesman S, Minkoff H. Incidence of venereal warts in human immunodeficiency virus-infected and uninfected women.
J Infect Dis 1995, 172: 235 –238.
10. Fennema JSA, van Ameijden EJC, Coutinho RA, van den Hoek JAR. HIV, sexually transmitted diseases and gynaecologic disorders in women: increased risk for genital herpes and warts among HIV-infected prostitutes in Amsterdam.
AIDS 1995, 9: 1071 –1078.
11. SAS Institute. SAS/STAT Software: Changes enhancements through release 6.12.
Cary, NC: SAS Institute; 1996.
12. Bersoff-Matcha SJ, Horgan MM, Fraser VJ, Mundy LM, Stoner BP. Sexually transmitted disease acquisition among women infected with human immunodeficiency virus type 1.
J Infect Dis 1998, 178: 1174 –1177.
13. Capps L, Peng G, Doyle M, el-Sadr W, Neaton JD. Sexually transmitted infections in women infected with the human immunodeficiency virus.
:Terry Beirn Community Programs for Clinical Research on AIDS (CPCRA).
Sex Transm Dis 1998, 25: 443 –447.
14. Clark RA, Brandon W, Dumestre J. et al
. Clinical manifestations of infection with the Human Immunodeficiency Virus in women in Louisiana.
Clin Infect Dis 1993, 17: 165 –172.
15. Kissinger P, Clark R, Dumestre J, Bessinger R. Incidence of three sexually transmitted diseases during a safer sex promotion program for HIV-infected women.
J Gen Intern Med 1996, 11: 750 –752.
16. Sorvillo F, Kovacs A, Kerndt P, Stek A, Muderspach L, Sanchez-Keeland L. Risk factors for trichomoniasis among women with human immunodeficiency virus (HIV) infection at a public clinic in Los Angeles county, California: implications for HIV prevention.
Am J Trop Med Hyg 1998, 58: 495 –500.
17. Ghys PD, Diallo MO, Ettiegne-Traore V. et al
. Genital ulcers associated with human immunodeficiency virus-related immunosuppression in female sex workers in Abidjan, Ivory Coast.
J Infect Dis 1995, 172: 1371 –1374.
18. Cameron DW, Ngugi EN, Ronald AR. et al
. Condom use prevent genital ulcers in women working as prostitutes.
Sex Transm Dis 1991, 18: 188 –191.
19. Corey L. Genital herpes.
In Sexually Transmitted Diseases
. Edited by Holmes KK, Mårdh PA, Sparking PF, et al
. New York: McGraw-Hill Inc.; 1990: 452.
20. Schuman P, Capps L, Peng G. et al
. Weekly fluconazole for the prevention of mucosal candidiasis in women with HIV infection.
:A randomized, double-blind, placebo-controlled trial. Terry Beirn Community Programs for Clinical Research on AIDS.
Ann Intern Med 1997, 126: 689 –696.
21. Duerr A, Sierra MF, Feldman J, Clarke LM, Ehrlich I, DeHovitz J. Immune compromise and prevalence of Candida vulvovaginitis in human immunodeficiency virus-infected women.
Obstet Gynecol 1997, 90: 252 –256.
22. van Duynhoven YT, Nagelkerke NJ, van de Laar MJ. Reliability of self-reported sexual histories: test-retest and interpartner comparison in a sexually transmitted diseases clinic.
Sex Transm Dis 1999, 26: 33 –42.
23. Fennema JSA, van Ameijden EJC, Coutinho RA, van den Hoek JAR. Validity of self-reported sexually transmitted diseases in a cohort of drug-using prostitutes in Amsterdam: trends from 1986 to 1992.
Int J Epidemiol 1995, 24: 1034 –1041.
The following investigators participate in the European Study on the Natural History of HIV infection in Women: Yolanda Pelgrom, Institute of Tropical Medicine, Antwerp, Belgium; Birgit Kvinesdal, Hvidover Hospital, Hvidover, Denmark; Jorma Paavonen, Hannele Savonius, Department of Obstetrics and Gynecology, University of Helsinki, Finland; Catherine Marimoutou, Groupe d'Epidémiologie Clinique du SIDA en Aquitaine, Bordeaux, France; Marie-Emmanuelle Mars, La Conception Hospital, Marseilles, France; Jean-Albert Gastaut, Dominique Sperandeo (steering committee), Saint-Marguerite Hospital, Marseilles, France; Christine Bergeron, CERBA, Cergy-Pontoise, France; Catherine Crenn-Hébert, Louis Mourier Hospital, Colombes, France; Françoise Meier, Louis Mourier Hospital, Colombes, France (steering committee); Paul Cesbron, Laënnec Hospital, Creil, France; Marie-Laure Babut, Henri Mondor Hospital, Créteil, France; Anne Odier, private practitioner, Paris, France; Jean-Dominique Poveda, Pasteur Institute, Paris, France; Christophe Griffault, Françoise Hamers, European Centre for the Epidemiological Monitoring of AIDS, Saint-Maurice, France; Hélène Vrillon, Saint-Maurice Hospital, Saint-Maurice, France; Alain Berrébi, La Grave Hospital, Toulouse, France; Anastasia Roumeliotou, Athens School of Public Health, Athens, Greece; J. Fiore, Achiropita Lepera, Bari University, Bari, Italy; Alberto Matteelli, Alessandra Tebaldi, Brescia University, Brescia, Italy; Alberto Agarossi, Mario Conti, Daniele Federici, Sacco Hospital, Milan, Italy; Annarosa Del Mistro, Oncology Institute, Padova, Italy; Barbara Suligoi, Instituto Superiore di Sanità, Rome, Italy (steering committee); Ingrid Spijkerman, Joke Bax, Municipal Health Service, Amsterdam, The Netherlands; Elisabeth von der Lippe, Ullevål Hospital, Oslo, Norway; Lucia de Pinho, Coimbra Hospital, Coimbra, Portugal; Jorge Cardoso, Centro de Salude da Lapa, Lisbon, Portugal; Manuela Doroana, Santa Maria Hospital, Lisbon, Portugal; Soledad García Pérez, Sandoval Centre, Madrid, Spain; José María Peña, La Paz Hospital, Madrid, Spain; Jesús Grande, 12 de Octubre Hospital, Madrid, Spain; Elisa Pérez Cecilia, San Carlos Hospital, Madrid, Spain; Alberto Reche Rosado, Carlos Haya Hospital, Malaga, Spain; José Manuel Agud, Txagorritxu Hospital, Vitoria, Spain; Bo Anzén, Danderyd Hospital, Danderyd, Sweden (steering committee); Kristina Elfgren, Pehr Olof Pehrson, Huddinge University Hospital, Huddinge, Sweden; Pietro Vernazza, Kantonsspital, San Gallen, Switzerland (steering committee).