Objective: Hormonal contraception has been associated with an increased prevalence of cervical shedding of HIV-1 DNA among infected women. We conducted this study to evaluate the effect of the use of an intrauterine device (IUD) on the detection of HIV-1 DNA in cervical secretions.
Design: A prospective study of HIV-1-seropositive women undergoing IUD insertion at two public family planning clinics in Nairobi, Kenya.
Methods: Cervical swab samples were collected before IUD insertion and approximately 4 months thereafter for the detection of HIV-1-infected cells using polymerase chain reaction (PCR) amplification of HIV-1 gag DNA sequences.
Results: Ninety-eight women were enrolled and followed after IUD insertion. The prevalence of HIV-1 DNA cervical shedding was 50% at baseline and 43% at follow-up [odds ratio (OR) 0.8, 95% confidence interval (CI) 0.5-1.2]. There was no statistically significant difference between the baseline and follow-up shedding rates in a multivariate model that controlled for previous hormonal contraceptive use, condom use, cervical ectopy, friable cervix, cervical infections at an interim visit, and CD4 lymphocyte levels (OR 0.6, 95% CI 0.3-1.1).
Conclusion: The insertion of an IUD did not significantly alter the prevalence of cervical shedding of HIV-1-infected cells. The use of IUDs, in conjunction with condoms, may be an appropriate method of contraception for HIV-1-infected women from the standpoint of potential infectivity to the male partner through exposure to genital HIV-1.
From the aDepartments of Biostatistics, Microbiology, Medicine, and Epidemiology, University of Washington, Seattle, Washington, USA; bFamily Health International, Research Triangle Park, NC, USA; and cDepartment of Obstetrics and Gynaecology, University of Nairobi, Nairobi, Kenya.
Disclaimer: The views expressed in this article do not necessarily reflect those of Family Health International (FHI), the US Agency for International Development (USAID), or the American Foundation for AIDS Research (AmFAR).
Sponsorship: This work was supported by FHI with funds from USAID and from AmFAR (grant no. 02243-16-RG).
Corresponding author and requests for reprints: Barbra Richardson, Box 359909, University of Washington, Seattle, WA 98195, USA.
Received: 26 January 1999; revised: 1 July 1999; accepted: 26 July 1999.
By January 1999, approximately 43% of the 32 million adults infected with HIV worldwide were women, and this proportion is rising. Approximately half of new adult HIV infections are in adults aged 15-24 years, and the majority of adult infections stem from heterosexual transmission. The rising number of women of reproductive age infected with HIV emphasizes the need for safe and effective contraceptive methods for HIV-1-infected women. Male condoms should ideally be used during all episodes of sexual intercourse by HIV-1-seropositive women to protect the male partner. In the event of inconsistent condom use or condom breakage, a second contraceptive method may be advisable for improved pregnancy prevention. The contraceptive options most appropriate for HIV-1-infected women are currently unknown.
Intrauterine devices (IUDs) are a highly effective, long lasting, and inexpensive contraceptive method. The major complication associated with their use is pelvic inflammatory disease (PID) during the first 20 days after insertion. We recently completed a study that showed no difference in the incidence of IUD-related complications between HIV-1-infected and uninfected women during the 4 months after IUD insertion. The use of an IUD results in an inflammatory response in the endometrium and alters the microenvironment of the uterus, oviduct, and cervix. The use of a copper IUD increases menstrual blood flow by 50-75% and increases the copper content of the cervical mucus[5,6]. Both the recruitment of inflammatory cells and increased blood lymphocytes and macrophages provide potential targets for HIV-1 replication. It is thus biologically plausible that the insertion of an IUD could alter the prevalence of cervical shedding of HIV-1 DNA in infected women because of these or other factors. However, no data are available on the effect of IUD use on the infectivity of HIV-1-infected women, as evidenced by the genital shedding of virus or virus-infected cells.
Genital shedding of HIV-1 DNA is one potential marker of the infectivity of an HIV-1-infected person because the increased shedding of infected cells in the genital secretions may increase the dose of virus to which a sexual partner is exposed[7,8]. Recently, Iverson et al.  reported a positive relationship between the genital shedding of cell-associated HIV-1 DNA and cell-free HIV-1 RNA in the cervical secretions. One potential indicator of the effect of a contraceptive method on the infectiousness of an HIV-1-infected woman is thus its association with the cervical or vaginal shedding of HIV-1 DNA. Several studies have investigated the relationship between the cervical or vaginal shedding of HIV-1-infected cells and hormonal contraceptive use. Two studies [10,11] found increased cervical shedding of HIV-1 DNA among hormonal contraceptive users, and two studies [12,13] found no association. However, only one study  has examined the association between IUD use and HIV-1 DNA cervical shedding. Although that study found no association, it had inadequate power to detect an association because only six study participants used IUDs.
A prospective cohort study of the short-term complications of IUD use  was conducted among HIV-1-infected and uninfected women attending two public family planning clinics in Nairobi, Kenya. As part of that larger study, the effects of IUD use on the cervical shedding of HIV-1 DNA were prospectively investigated. The substudy was designed to test the hypothesis that IUD use increases the prevalence of the cervical shedding of HIV-1-infected cells 4 months after IUD insertion.
Materials and methods
Study population and procedures
Details of the study methods have been described elsewhere. Briefly, women attending two family planning clinics in 1994-1995 in Nairobi, Kenya, requesting an IUD were referred to study staff. After written informed consent was obtained and HIV pre-test counselling was conducted, blood was drawn for HIV-1 serological testing. Study nurses conducted a brief interview to gather demographic, recent sexual behavior, and contraceptive use information. Physicians conducted a physical examination including a pelvic examination. Before the collection of any other cervical samples, a Dacron swab was inserted gently approximately 1cm into the cervical os and rotated 360° to collect a sample for the detection of HIV-1 DNA. To avoid blood on the swab, physicians were instructed to collect the cervical samples at a time when the woman was not menstruating. If the woman met the eligibility criteria, the physician then inserted a Copper T 380A IUD (Ortho Pharmaceutical Corporation, Raritan, NJ, USA).
At a follow-up visit 1 month after insertion, nurses informed study participants of their HIV-1 test results and provided post-test counselling. Interviews and physical examinations were repeated. Blood specimens were drawn for CD4 lymphocyte determination and endocervical specimens were collected for the diagnosis of Chlamydia trachomatis and Neisseria gonorrhoeae. At a 4-month follow-up visit, an interview, physical examination, and endocervical swab collection for HIV-1 DNA testing were repeated. Study physicians responsible for the physical examination and specimen collection were masked to participants‚ HIV-1 infection status throughout the study.
Serum samples were tested for HIV-1 antibodies using an enzyme-linked immunosorbent assay (ELISA) kit (Organon Vironostika HIV-1 Microelisa; Organon Teknika, Durham, NC, USA) and positive or indeterminate tests were confirmed using a second ELISA based on a different antigen preparation (HIV-1 Recombigen; Cambridge Bioscience, Worcester, MA, USA). C. trachomatis antigen testing was performed using a Syva MicroTrak II enzyme immunoassay (Syva Company, Brussels, Belgium). Endocervical swab specimens were inoculated onto Thayer-Martin medium for the culture of N. gonorrhoeae.
Endocervical swabs for HIV-1 DNA detection were transported to the laboratory within 4h of collection, stored at -70°C, and subsequently shipped to the University of Washington on dry ice. Samples were batched for testing so that approximately equal numbers of baseline and follow-up samples were tested on each testing date. In addition, cervical specimens from HIV-1-seronegative women were tested as controls and all polymerase chain reaction (PCR) results for the control specimens were negative. Laboratory technicians were blinded to the visit number of the sample, and the serostatus of the women. After thawing, the swabs were vortexed briefly in lysis buffer [10mM TRIS HCl (pH8.3) 50mM KCl, 0.01% (wt/vol) gelatin, 0.45% NP-40, 0.45% Tween 20, and 0.6mg/ml proteinase K], incubated at 56°C for 90min, and then boiled for 15min to inactivate proteinase K. After the removal of the swabs, nested PCR assays were conducted using primers specific for HIV-1 gag DNA, as previously described. The primers used for PCR were designed to detect non-clade B viruses using available gag consensus sequences; these primers have been shown consistently to detect subtypes A, C, and D HIV-1, which are the subtypes found in Kenya.
To ensure that the IUD did not affect the sensitivity of the PCR assay, we performed spiking experiments using DNA from the ACH-2 cell standards (either from one infected cell or 10 infected cells) typically used as controls for the PCR assay[15,16]. We detected both the one copy and the 10 copy standard at the same frequency in the spiked swab samples as were detected using the ACH-2 control DNAs alone. Moreover, there was no significant difference in the prevalence of detection at the baseline visit compared with the follow-up visit at either spiking level, indicating no evidence of an effect of the IUD on the sensitivity of the PCR assay.
The S-Plus statistical package (Version 3.4, StatSci, a division of MathSoft, Inc., Seattle WA, USA) was used for all analyses. To investigate the correlates of the cervical shedding of HIV-1-infected cells at baseline, we used Pearson‚s Chi-square test or Fisher‚s exact test. Generalized estimating equations (GEE) with robust variance estimates, a logit link, and exchangeable correlation structure were used to obtain P values for the comparison of baseline and follow-up binary data, and the Wilcoxon rank sum test was used for continuous data.
To estimate the adjusted odd ratio (OR) for the effect of IUD insertion on the cervical shedding of HIV-1 DNA, we constructed a multivariate GEE model with a logit link and exchangeable correlation structure containing potential confounding variables. In particular, we chose variables to include in the model that were risk factors for shedding at baseline or that were significantly different between the baseline and follow-up visit (cervical friability and condom use). In addition, because recent studies have shown an increased risk of cervical shedding of HIV-1 DNA in women with gonorrhea, cervical ectopy, or hormonal contraceptive use[10,11], and an inverse relationship between cervical shedding and CD4 cell count[11,14], we included reported hormonal contraceptive use in the 3 months before enrolment, cervical ectopy, cervical infection at the 1 month follow-up visit (gonorrhea or chlamydial cervicitis), and log10 CD4 cell count at the 1 month follow-up visit as covariates.
A total of 156 HIV-1-seropositive women were enrolled in the study. Twenty-seven women were excluded from the analysis comparing shedding rates before and after IUD insertion, 14 because they had used an IUD in the 3 months before enrolment, eight because their IUD was removed, expelled, or displaced before their 4 month visit, and five because they did not have baseline cervical HIV-1 DNA results. Of the remaining 129 women, 98 (76%) returned for their 4 month follow-up visit and were included in the analysis. Women with follow-up were less likely to have used injectable contraceptives in the 3 months before enrolment (11 versus 31%, P=0.01) and to have cervical ectopy at enrolment (55 versus 87%, P=0.001) than women without follow-up. All other baseline demographic, sexual history, and physical examination variables were similar between those who returned and those who did not return at 4 months (data not shown).
The median age of the 98 women included in the analysis was 26 years. Eighty-six per cent were currently or previously married. All women were parous, and the median time since the last pregnancy was 1 year. Forty-three per cent had used hormonal contraception during the 3 months before enrolment. Median sexual frequency was one to two times per week, and all women had had two or fewer sexual partners in the 3 months before enrolment. Eight per cent of the women reported having had a sexually transmitted disease in the previous year and most (92%) reported no condom use in the previous 3 months. At the 1 month follow-up visit, 66% of the women had CD4 cell counts of over 500, 27% had counts between 200 and 499, and 7% had counts of less than 200. One woman (1%) had gonorrhea at her 1 month follow-up visit and was treated at her 4 month visit, and six women (6%) had chlamydial cervicitis, one of whom was treated at her 4 month visit. The median time between the baseline visit and the 4 month follow-up visit was 4.3 months (range 3.2-11.9).
No demographic, sexual history, physical examination, or laboratory variables were significantly associated with the cervical shedding of HIV-1-infected cells at baseline (Table 1).
No statistically significant differences were found between the baseline and 4 month follow-up visits with respect to the prevalence of cervical ulcers, PID, cervical ectopy, or cervical edema (Table 2). The prevalence of cervical friability was higher at enrolment than at follow-up (28 versus 14%, P=0.02), and the prevalence of any reported condom use in the past 3 months was higher at enrolment than at follow-up (8 versus 1%, P=0.05) (Table 2).
The prevalence of cervical shedding of HIV-1 DNA was 50% (49/98) at baseline and 43% (42/98) at the 4 month visit [OR 0.8, 95% confidence interval (CI) 0.5-1.2] (Table 3). Assuming a two-sided test with alpha of 0.05, 50% shedding at baseline, and 98 women in the study, we have 80% power to detect a 1.4-fold or greater change in shedding after the insertion of an IUD. Of the 49 women who had cervical HIV-1 DNA detected at baseline, 26 (53%) had persistence of shedding at the 4 month visit. Conversely, of the 49 women who were negative for cervical HIV-1 DNA at baseline, 33 (67%) remained negative. Therefore, 39 (40%) of the 98 women changed their shedding status between the two examinations (Table 4).
To estimate an adjusted OR for shedding after IUD insertion, we constructed a multivariate GEE model containing potential confounding variables. No statistically significant difference between baseline and follow-up HIV-1 cervical shedding prevalences were present after controlling for potential confounding variables (OR 0.6, 95% CI 0.3-1.1) (Table 3). Of the other variables included in the multivariate model, hormonal contraceptive use during the 3 months before enrolment, condom use in the past 3 months, friable cervix, cervical ectopy, or gonorrhea or chlamydial cervicitis during the interim visit were not significantly related to cervical shedding (data not shown). However, we found an inverse relationship between log10 CD4 cell count and the cervical shedding of HIV-1 DNA (OR 0.2, 95% CI 0.0-0.8), indicating that each 10-fold increase in the CD4 cell count was associated with an 80% decrease in the odds of cervical shedding of HIV-1 DNA.
We found no statistically significant difference in the prevalence of HIV-1 DNA cervical shedding among 98 HIV-1-infected women before and 4 months after IUD insertion. This held true after controlling for previous hormonal contraceptive use, condom use, friable cervix, cervical ectopy, cervical infection at an interim visit, and CD4 lymphocyte levels. Only one previous study  has included an analysis of HIV-1 DNA cervical shedding and IUD use. No association was found, but the number of IUD users in that study (six women) was too small to draw any conclusions. Our results are consistent with an epidemiological study  that found no association between IUD use and female-to-male HIV-1 transmission. Our results suggest that IUDs may not increase the HIV-1 exposure and potential transmission risk from an infected woman to her sex partner.
HIV-1-infected women have a critical need for safe and effective contraception to avoid unwanted pregnancy and to prevent vertical transmission. For many women, this means a choice between the available long-acting, reversible contraceptives, primarily IUDs, oral contraceptives, and injectable progesterone. Two previous studies [10,11] have found an increased prevalence of HIV-1 DNA cervical shedding among oral contraceptive users, and two studies [12,13] found no association. The only published study to date among injectable progesterone users  found that its use was associated with increased cervical shedding after adjusting for CD4 lymphocyte levels. Our current prospective data suggest that women are not at an increased risk of cervical shedding of HIV-1 DNA after IUD insertion, and that IUD use is at least as safe as hormonal contraceptive use from the standpoint of infectivity to a sex partner. However, it should be emphasized that the most effective contraceptives from the point of view of pregnancy prevention (e.g. hormonal contraception and IUD) confer no protection with regard to HIV or other sexually transmitted disease transmission, and thus need to be combined with condoms during sexual encounters of risk. In our study of HIV-1-seropositive women, only 8% reported condom use during the previous 3 months at their enrolment interview and this decreased to 1% at the follow-up interview. This underscores the importance of condom promotion for HIV-1-infected women to protect their partners, irrespective of their choice of other contraceptive methods.
The strengths of our study include its prospective nature with each HIV-1-infected woman serving as her own control. Second, the study had adequate power to detect a change in cervical HIV-1 DNA detection that would be predicted to be of clinical relevance (80% power to detect an OR of 1.4 or greater). Third, our measurement of other factors that might affect HIV-1 DNA cervical shedding, such as immune status (CD4 lymphocyte levels), cervical ectopy, and previous hormonal contraceptive use, allowed us to adjust for these variables. Finally, laboratory technicians conducting HIV-1 DNA PCR assays were masked to the timing of specimens (whether a specimen was taken before or after IUD insertion). The limitations of the study include the qualitative nature of our HIV-1 DNA PCR assay and the fact that the follow-up measurements were taken at only one time point (4 months) after IUD insertion. Our results may thus have differed if specimens had been collected at other times. Specifically, any menorrhagia related to IUD use may increase the risk of HIV-1 infection for sexual partners. However, we were unable to detect this as a result of our sampling scheme. In addition, our measurements of immune status and cervical infections were limited to the 1 month visit and thus we could not consider their effects across time.
The side-effects or complications associated with contraception among HIV-1-infected women are important. The International Planned Parenthood Federation and a World Health Organization working group recommend against IUD use by HIV-1-infected women[19,20]. This recommendation is based on theoretical concerns about pelvic infection and increased blood loss. However, in the larger longitudinal study from which these data are drawn, we found that HIV-1-infected women were not at increased risk of complications (PID, IUD expulsions, IUD removal for infection, pain or bleeding, or pregnancies) compared with HIV-1-seronegative women using IUDs. In countries with a high prevalence of HIV-1, the effect of contraceptive use on HIV-1 acquisition by uninfected women is also of critical importance. Cross-sectional studies [21-26] of the correlates of HIV-1 seropositivity have yielded inconsistent results regarding an IUD/HIV-1 relationship, and the only longitudinal study of HIV-1 seroconversion  found no association between IUD use and HIV-1 acquisition (rate ratio 0.8; 95% CI 0.4-1.7). Together, these data suggest that IUDs may be a suitable contraceptive method for women, irrespective of HIV-1 infection status, and that countries with a high HIV-1 prevalence should continue to provide IUDs as an integral part of their family planning service delivery.
The authors thank Ms Susan Chen for data management and the staff of the University of Nairobi research team for data collection and data entry activities. In addition, the authors would like to thank the staff and clients of the Kenyatta National Hospital Family Planning Clinic (Clinic 66) and the Riruta Nairobi City Commission Health Center for their support and participation.
1. Joint United Nations Programme on HIV/AIDS and the World Health Organization. Report on the global HIV/AIDS epidemic, December 1998.
Geneva: Joint United Nations Programme on HIV/AIDS and the World Health Organization, 1999.
2. Odlind V. Modern intra-uterine devices. Baillieres Clin Obstet Gynaecol
3. Farley TM, Rosenberg MJ, Rowe PJ, Chen JH, Meirik O. Intrauterine devices and pelvic inflammatory disease: an international perspective. Lancet
4. Sinei SK, Morrison CS, Sekadde K-C, Allen M, Kokonya D. Complications of use of intrauterine devices among HIV-1-infected women. Lancet
5. Ortiz ME, Croxatto HB, Bardin CW. Mechanisms of action of intrauterine devices. Obstet Gynecol Surv
6. Rybo G. IUDs and endometrial bleeding, qualitative and quantitative aspects.
In: IUD, pathology, and management
. Hafez E, Os W (editors). Lancaster: MTP Press; 1980. pp. 63-74.
7. Mostad SB, Kreiss JK. Shedding of HIV-1 in the genital tract. AIDS
8. Critchlow CW, Kiviat NB. Detection of human immunodeficiency virus type 1 and type 2 in the female genital tract: implications for the understanding of virus transmission. Obstet Gynecol Surv
9. Iversen AK, Larsen AR, Jensen T, et al
. Distinct determinants of human immunodeficiency virus type 1 RNA and DNA loads in vaginal and cervical secretions. J Infect Dis
10. Clemetson DB, Moss GB, Willerford DM, et al
. Detection of HIV DNA in cervical and vaginal secretions. Prevalence and correlates among women in Nairobi, Kenya. JAMA
11. Mostad SB, Overbaugh J, DeVange DM, et al
. Hormonal contraception, vitamin A deficiency, and other risk factors for shedding of HIV-1 infected cells from the cervix and vagina. Lancet
12. Kreiss J, Willerford DM, Hensel M, et al
. Association between cervical inflammation and cervical shedding of human immunodeficiency virus DNA. J Infect Dis
13. 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
14. John GC, Nduati RW, Mbori-Ngacha D, et al
. Genital shedding of human immunodeficiency virus type 1 DNA during pregnancy: association with immunosuppression, abnormal cervical or vaginal discharge, and severe vitamin A deficiency. J Infect Dis
15. Panteleeff DD, John G, Nduati R, et al
. Rapid method for screening dried blood samples on filter paper for human immunodeficiency virus type 1 DNA. J Clin Microbiol
16. Moss GB, Overbaugh J, Welch M, et al
. Human immunodeficiency virus DNA in urethral secretions in men: association with gonococcal urethritis and CD4 cell depletion. J Infect Dis
17. Zeger SL, Liang KY. Longitudinal data analysis for discrete and continuous outcomes. Biometrics
18. European Study Group on Heterosexual Transmission of HIV. Comparison of female to male and male to female transmission of HIV in 563 stable couples. BMJ
19. International Planned Parenthood Federation International Medical Advisory Panel. Statement on contraception for clients who are HIV positive. IPPF Med Bull
20. World Health Organization Scientific Working Group on Improving Access to Quality Care in Family Planning. Medical eligibility criteria for initiating and continuing use of contraceptive methods.
Geneva: World Health Organization, 1996.
21. Musicco M, Nicolosi A, Saracco A, Lazarrin A. IUD use and man to woman sexual transmission of HIV-1.
In: Fourth International Conference on IUDs.
Bardin CW, Mishell DR (editors). Boston: Butterworth-Heinemann; 1994.
22. Lazzarin A, Saracco A, Musicco M, Nicolosi A. Man-to-woman sexual transmission of the human immunodeficiency virus. Risk factors related to sexual behavior, man‚s infectiousness, and woman‚s susceptibility. Italian Study Group on HIV Heterosexual Transmission. Arch Intern Med
23. Mati JK, Hunter DJ, Maggwa BN, Tukei PM. Contraceptive use and the risk of HIV infection in Nairobi, Kenya. Int J Gynaecol Obstet
24. European Study Group. Risk factors for male to female transmission of HIV. BMJ
25. Caraël M, Van de Perre PH, Lepage PH, et al
. Human immunodeficiency virus transmission among heterosexual couples in Central Africa. AIDS
26. Plourde PJ, Plummer FA, Pepin J, et al
. Human immunodeficiency virus type 1 infection in women attending a sexually transmitted diseases clinic in Kenya. J Infect Dis
27. Kapiga SH, Lyamuya EF, Lwihula GK, Hunter DJ. The incidence of HIV infection among women using family planning methods in Dar es Salaam, Tanzania. AIDS
Keywords:© 1999 Lippincott Williams & Wilkins, Inc.
Cervical shedding; contraceptive methods; HIV-1; IUD