Effect of intrauterine device use on cervical shedding of HIV-1 DNA.

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.

Use of sexually transmitted disease risk assessment algorithms for selection of intrauterine device candidates.

Sexually transmitted diseases (STD) are an important contraindication for intrauterine device (IUD) insertion. Nevertheless, laboratory testing for STD is not possible in many settings. The objective of this study is to evaluate the use of risk assessment algorithms to predict STD and subsequent IUD-related complications among IUD candidates. Among 615 IUD users in Kenya, the following algorithms were evaluated: 1) an STD algorithm based on US Agency for International Development (USAID) Technical Working Group guidelines: 2) a Centers for Disease Control and Prevention (CDC) algorithm for management of chlamydia; and 3) a data-derived algorithm modeled from study data. Algorithms were evaluated for prediction of chlamydial and gonococcal infection at 1 month and complications (pelvic inflammatory disease [PID], IUD removals, and IUD expulsions) over 4 months. Women with STD were more likely to develop complications than women without STD (19% vs 6%; risk ratio = 2.9; 95% CI 1.3-6.5). For STD prediction, the USAID algorithm was 75% sensitive and 48% specific, with a positive likelihood ratio (LR+) of 1.4. The CDC algorithm was 44% sensitive and 72% specific, LR+ = 1.6. The data-derived algorithm was 91% sensitive and 56% specific, with LR+ = 2.0 and LR- = 0.2. Category-specific LR for this algorithm identified women with very low (< 1%) and very high (29%) infection probabilities. The data-derived algorithm was also the best predictor of IUD-related complications. These results suggest that use of STD algorithms may improve selection of IUD users. Women at high risk for STD could be counseled to avoid IUD, whereas women at moderate risk should be monitored closely and counseled to use condoms.

PIP: This study aimed to evaluate the effectiveness of using risk assessment algorithms in predicting sexually transmitted disease (STD) and subsequent IUD-related complications among IUD candidates. The study population was selected among women who desired an IUD insertion in Nairobi, Kenya. The following algorithms drawn from the study of IUD use and HIV infection among these 615 IUD users were evaluated: 1) an STD algorithm based on US Agency for International Development (USAID) Technical Working Group guidelines; 2) a Centers for Disease Control and Prevention (CDC) algorithm for management of chlamydia; 3) a data-derived algorithm modeled from data. Algorithms were also evaluated for prediction of chlamydial and gonococcal infection at 1 month and complications (pelvic inflammatory disease, IUD removals, and IUD expulsions) at 4 months. Results showed that women with STDs were more likely to develop complications than women without STDs (19% vs. 6% risk ratio = 2.9; 95% CI, 1.3-6.5). In STD prediction, the USAID algorithm was 91% sensitive and 56% specific, with LR+ = 2.0 and LR- = 0.2. Category-specific LR for this algorithm identified women with very low (1%) and very high (29%) infection probabilities. Thus, sexually transmitted disease was associated with increased risk for complications after IUD insertion. Moreover, it may be concluded that simple risk assessment criteria can assist in the identification of women at high and low risk for STD among women presenting for IUD insertion; it may also be concluded that the use of simple risk assessment tools may facilitate the identification of women who require close observation, thus reducing the incidence of IUD-related complications.

The role of maternal syphilis, gonorrhoea and HIV-1 infections in spontaneous abortion.

The role of the human immunodeficiency virus (HIV) and other sexually transmitted diseases (STDs) as risk factor for spontaneous abortion was investigated in a case-control study in Nairobi, Kenya. Cases (n = 195) were women admitted with clinical signs and symptoms of spontaneous abortion, before 20 weeks of gestation. Patients with induced or clinically septic abortion were excluded. Controls were unselected pregnant women in their second or third trimester (n = 195). Spontaneous abortion was independently associated with maternal HIV-1 antibody (14.4% versus 6.2%; adjusted odds ratio, 2.3; 95% confidence interval, 1.1-4.8), with maternal syphilis seroreactivity (6.8% vs 2.1%, adjusted odds ratio, 4.3; 95% confidence interval 1.2-15.6), and with vaginal colonization with group B streptococci (15.6% vs 5.2%; adjusted odds ratio, 3.2; 95% confidence interval, 1.5-6.7). No significant association was found between maternal infections with gonococci and genital mycoplasmas, and spontaneous abortion.

Contraceptive use and HIV infection in Kenyan family planning clinic attenders.

This pilot study aimed to determine the feasibility of a larger study of contraception and risk of HIV infection in women. We also measured risk factors for and occurrence of HIV infection in the participants. A cohort of 1537 seronegative women attending a family planning clinic in Nairobi, Kenya was enrolled and followed for up to 12 months per woman. HIV testing was done quarterly. A nested case-control analysis was done with seroconverting women (cases) and 3 matched controls per case, who had detailed interviews and received physical examinations and STD tests. The prevalence of HIV at enrollment was 6.1%; seropositive women were excluded from further analysis. The 12-month life-table cumulative incidence of HIV was 2.1 per 100 women (95% confidence interval [CI] 1.1-3.2). In the nested case-control analysis (17 cases and 51 controls), the crude odds ratio of HIV infection comparing oral contraceptive (OC) users with other women was 3.5 (95%) CI 0.8-21.5), which persisted after control for single confounders at a time. The putative association between OC use sand HIV infection is critical to public health policy, yet no study has been conducted specifically to measure it, yielding weak and conflicting evidence. We intend to conduct a larger study with a similar design as the current pilot study, which confirmed the feasibility of a more definitive project.

Cervical incompetence: assessment of a scoring system for patient selection for cervical cerclage.

Forty-seven antenatal patients with clinical diagnosis of cervical incompetence were subjected to a structured scoring system before insertion of McDonald Stitch. All other possible causes of midtrimester abortion were ruled out and the patients followed up to the end of their pregnancies. Significantly there were fewer preterm deliveries in the high score group (P less than 0.01) and no abortions. The mean cerclage-delivery interval was shown to rise with increasing scores. Significantly better fetal salvage ratio was also seen in the high score group (P less than 0.05). The scoring system shows promise as a selective and prognostic tool and is recommended for trial on a wider scale.

The role of chlamydia trachomatis in the aetiology of acute pelvic inflammatory disease (PID) in Nairobi area, Kenya: a preliminary report.

A preliminary report of 50 cases of acute pelvic inflammatory disease (PID) studied at the Kenyatta National Hospital to find out the role of chlamydia trachomatis in its aetiology is presented. Endocervical swab cultures were done. There were 2/50 (4.0%) cases positive for chlamydia trachomatis. The role of this pathogen in the aetiology of acute PID is discussed.

Experience with IUCD insertion outside of menses in Kenya.

OBJECTIVE: To determine if women receiving intrauterine devices (IUCDs) outside of menses have an acceptable rate of insertion problems and subsequent IUCD-related complications. DESIGN: Cross-sectional and prospective cohort study of insertions at times other than during menses. SETTING: The study was carried out in two government family planning (FP) clinics in Nairobi, Kenya. SUBJECTS: After appropriate pre-test and post-test HIV counselling, 1686 women requesting IUCDs at two FP clinics between 1994 and 1995 in Nairobi were enrolled at baseline into a study examining the effect of human immuno-deficiency virus (HIV) infection on IUCD-related complications. Six hundred and forty nine women (156 HIV-infected and 493 HIV-uninfected) were selected for the four month follow up study. They were classified according to their menstrual cycle status at time of IUCD insertion. MAIN OUTCOME MEASURES: Problems at the time of insertion (pain, bleeding, immediate expulsion) and IUCD-related complications through four months. RESULTS: Rates of immediate insertion problems were low in the women who had insertions during menses (7.0%), outside of menses (4.0%) or had oligomenorrhea/amenorrhea (2.6%). The adjusted odds ratios for IUCD insertion problems outside of menses and in oligomenorrhea/amenorrhea (versus women with insertion during menses) were 0.54 (95 % CI 0.18-1.59) and 0.39 (95% CI 0.12-1.29) respectively. IUCD-related complications were higher in the oligomenorrhea/amenorrhea (11.5%) or insertion outside of menses (6.9%), than the within menses (4.3%) groups. However, the differences were not statistically significant. Adjusted odds ratios for IUCD outside of menses and oligomenorrhoea/amenorrhea groups were 1.65 (95% CI 0.21-12.91) and 2.72 (95% CI 0.34-21.71) respectively. CONCLUSION: The results confirm that the IUCD can be safely inserted outside of menses with minimal insertion difficulties and subsequent complications. Availability of IUCDs outside of menses may enhance IUCD acceptance in Kenya and create better opportunity for visual screening of the cervix for sexually transmitted infections.

Complications of use of intrauterine devices among HIV-1-infected women.

BACKGROUND: A WHO expert group and the International Planned Parenthood Federation recommend against use of intrauterine devices (IUDs) in HIV-1-infected women based on theoretical concerns about pelvic infection and increased blood loss. We investigated whether the risk of complications after IUD insertion is higher in HIV-1-infected women than in non-infected women. METHODS: 649 (156 HIV-1 infected 493 non-infected) women in Nairobi, Kenya, who requested and met local eligibility criteria for insertion of an IUD were enrolled. We gathered information on IUD-related complications, including pelvic inflammatory disease, removals due to infection, pain, or bleeding, expulsions, and pregnancies at 1 and 4 months after insertion. Patients' HIV-1 status was masked from physicians. FINDINGS: Complications were identified in 48 of 615 women (11 [7.6%] HIV-1-infected women, 37 [7.9%] non-infected). Incident pelvic inflammatory disease (two [1.4%] HIV-1 infected, one [0.2%] non-infected) and infection-related complications (any tenderness, removal of IUD for infection or pain; ten [6.9%] HIV-1 infected, 27 [5.7%] non-infected) were also rare and similar in the two groups. Complication rates were similar by CD4 (immune) status. Multivariate analyses suggested no association between HIV-1 infection and increased risks for overall complications (odds ratio 0.8 [95% CI 0.4-1.7]) or infection-related complications (1.0 [0.5-2.3]), adjusted for marital status, study site, previous IUD use, ethnic origin, and frequency of sexual intercourse, but a slight increase cannot be ruled out. INTERPRETATION: Our data suggest that IUDs may be a safe contraceptive method for appropriately selected HIV-1-infected women with continuing access to medical services.

PIP: Both the International Planned Parenthood Federation and the World Health Organization have warned against use of IUDs in HIV-infected women due to theoretical concerns about pelvic infection and increased blood loss. No published studies have investigated this concern, however. The validity of this recommendation was investigated in a comparative study of 156 HIV-1-infected and 493 non-infected women from two public family planning clinics (Kenyatta National Hospital and Riruta City Clinic) in Nairobi, Kenya, who requested and met local eligibility criteria for IUD insertion. At 1 and 4 months after insertion, information was collected from physicians--blinded as to the patient's HIV status--on IUD-related complications such as pelvic inflammatory disease (PID), removals, expulsions, and pregnancies. Complications were identified in 11 (7.6%) HIV-positive and 37 (7.9%) HIV-negative women. There were only 3 incident cases of PID, 2 of which occurred in HIV-infected women. IUD removal due to pain or infection occurred in 10 (6.9%) HIV-infected and 27 (5.7%) noninfected women. There were no differences in overall IUD complications in HIV-1-infected women by CD4 status (severely, moderately, or mildly immunocompromised). After adjustment for marital status, study site, previous IUD use, ethnic origin, and frequency of sexual intercourse, multivariate analysis suggested no association between HIV-1 infection and increased risks for overall IUD-related complications (odds ratio (OR), 0.8; 95% confidence interval (CI), 0.4-1.7) or infection-related complications (OR, 1.0; 95% CI, 0.5-2.3). These findings suggest that the IUD may be a safe contraceptive method for appropriately selected HIV-infected women with continuing access to medical services.

Is the intrauterine device appropriate contraception for HIV-1-infected women?

OBJECTIVE: To assess whether the risk of complications is higher in HIV-1-infected women compared with non-infected women in the two years following insertion of the intrauterine contraceptive device. DESIGN: Prospective cohort study. POPULATION: Six hundred and forty-nine women (156 HIV-1-infected, 493 non-infected) in Nairobi, Kenya who requested an intrauterine contraceptive device and met local eligibility criteria. METHODS: We gathered information on complications related to the use of the intrauterine contraceptive device, including pelvic inflammatory disease, removals due to infection, pain or bleeding, expulsions, and pregnancies at one, four, and 24 months after insertion by study physicians masked to participants' HIV-1 status. Cox regression was used to estimate hazard ratios. RESULTS: Complications were identified in 94 of 636 women returning for follow up (14.7% of HIV-1-infected, 14.8% of non-infected). The incidence of pelvic inflammatory disease was rare in both infected (2.0%) and non-infected (0.4%) groups. Multivariate analyses suggested no association between HIV-1 infection and increased risk of overall complications (hazard ratio = 1.0; 95% CI 0.6-1.6). Infection-related complications (e.g. any pelvic tenderness, removal for infection or pain) were also similar between groups (10.7% of HIV-1-infected, 8.8% of non-infected; P = 0.50), although there was a non-significant increase in infection-related complications among HIV-1-infected women with use of the intrauterine contraceptive device longer than five months (hazard ratio = 1.8; 95% CI 0.8-4.4). Neither overall nor infection-related complications differed by CD4 (immune) status. CONCLUSIONS: HIV-1-infected women often have a critical need for safe and effective contraception. The intrauterine contraceptive device may be an appropriate contraceptive method for HIV-1-infected women with ongoing access to medical services.

The prevalence of Neisseria gonorrhoea and Chlamydia trachomatis in intra-uterine contraceptive acceptors in Kenya.

PIP: The 1725 women presenting at Kenyatta National Hospital in 1984-86 for IUD insertion were screened for cervical Chlamydia trachomatis and Neisseria gonorrhoea before the IUD insertion. 207 (12%) cases of chlamydia trachomatis and 61 (3.5%) cases of Neisseria gonorrhoea were detected. There was no association between the ages of the women and the prevalence of these 2 sexually transmitted pathogens; however, there was a significant relationship between the prevalence of N gonorrhoea and marital status. N gonorrhoea was detected in 6.2% of never-married and 5.2% of formerly married women compared with 2.3% of currently married subjects (p0.001). Although there was no significant relationship between parity and the rate of isolation of the 2 pathogens, infection tended to be lower in women with 5 or more children. Educational attainment was significantly associated with N gonorrhoea infection: 5.1% in women who had 0-7 years of schooling compared with 3.0% in those with 8 or more years of education (p0.05). 12 women with C trachomatis infection were also positive for N gonorrhoea. There was no significant relationship between C trachomatis infection and any of the demographic variables examined. Given the finding that the greatest risk of pelvic inflammatory disease occurs in the 1st month of IUD use, it can be speculated that pathogens are inserted into the uterine cavity at the time of IUD insertion. It is therefore recommended that clients--especially the unmarried, the formerly unmarried, and those with low levels of education--be screened and treated for N gonorrhoea and C trachomatis before an IUD is inserted.^ieng

Towards developing a diagnostic algorithm for Chlamydia trachomatis and Neisseria gonorrhoeae cervicitis in pregnancy.

C trachomatis and N gonorrhoeae are major causes of maternal and neonatal morbidity and mortality in developing countries. To identify characteristics predictive of cervical infection, we examined pregnant women attending an antenatal clinic in Nairobi, Kenya. C trachomatis was isolated from 14/178 (8%), and N gonorrhoeae from 17 (10%); cervical infection with either pathogen was present in 28 (16%). Two characteristics were independently predictive of cervical infection by logistic regression analysis; the presence of either endocervical mucopus or induced endocervical bleeding, relative risk 4.2 (95% confidence interval (CI) 2.2 to 8.0) and having more than one sexual partner during pregnancy, relative risk 3.3 (95% CI 1.4 to 7.6). A screening programme for cervical infection which tested women with one or both risk markers would have a sensitivity of 68% (95% CI 51 to 85%) and a positive predictive value of 0.35 (95% CI 0.22 to 0.47). In countries where resources are limited, diagnostic algorithms incorporating clinical signs and behavioural characteristics may be useful in identifying pregnant women at high risk of cervical infection.

PIP: Chlamydia trachomatis and Neisseria gonorrhoea are major causes of maternal and neonatal morbidity in developing countries. 178 pregnant women attending an antenatal clinic in Nairobi, Kenya, were therefore examined in the attempt to identify characteristics predictive of cervical infection. Chlamydia trachomatis was isolated from 14 (8%) and Neisseria gonorrhoea from 17 (10%); cervical infection with either pathogen was present in 28 (16%). Logistic regression analysis found that the presence of either endocervical mucopus or induced endocervical bleeding and having more than 1 sex partner during pregnancy were independently predictive of cervical infection. A screening program for cervical infection which tested women with 1 or both risk markers wold have a sensitivity of 68% and a positive predictive value of 0.35. The authors conclude that countries with limited resources may use diagnostic algorithms and incorporate clinical signs and behavioral characteristics to help identify pregnant women at high risk of cervical infection.

Preventing IUCD-related pelvic infection: the efficacy of prophylactic doxycycline at insertion.

Most of the small increased risk in pelvic inflammatory disease (PID) associated with the intrauterine contraceptive device (IUCD) appears to be caused by bacterial contamination of the endometrial cavity at the time of insertion. This randomized clinical trial of 1813 women in Nairobi, Kenya, assessed the effectiveness of 200 mg of doxycycline given orally at the time of insertion in reducing the occurrence of PID. The rate of this infection in the doxycycline-treated group was 31% lower than that in the placebo-treated group (1.3 and 1.9%, respectively; RR 0.69; 95% CI 0.32 to 1.5). The rate of an unplanned IUCD-related visit to the clinic was also 31% lower in the doxycycline-treated group (RR 0.69; 95% CI 0.52 to 0.91). Although the significance level (P = 0.17) for the reduction is PID does not meet the conventional standard of 0.05, the results may be suggestive of an effect. Moreover, the reduction in IUCD-related visits (P = 0.004) not only represents an important decrease in morbidity but also substantiates the reduction found for PID. Further studies are needed to corroborate these results. Consideration should be given to the prophylactic use of doxycycline at the time of IUCD insertion as an approach to preventing PID and other IUCD-related morbidity.

PIP: This double-blind, randomized clinical trial was conducted to investigate whether the use of prophylactic doxycycline at intrauterine contraceptive device (IUCD) insertion can reduce the incidence of pelvic inflammatory disease (PID) in women. 1813 women in Nairobi, Kenya, were given 200 mg of doxycycline, taken orally at the time of IUCD insertion. Analysis of the data collected show that the rate of PID infection in the doxycycline-treated group was 31% lower than that in the placebo-treated group. The rate of an unplanned IUCD-related visit to the clinic was also 31% lower in the doxycycline-treated group. Although the significance level (P = 0.17) for the reduction in PID does not meet the conventional standard of 0.05, the results may be suggestive of an effect. In addition, the reduction in IUCD-related visits (P = 0.004) not only represents an important decrease in morbidity, but also substantiates the reduction found for PID. To conclude, the prophylactic use of doxycycline at the time of IUCD insertion appears effective, well tolerated, and cost-effective. Further studies are needed to corroborate these results and consideration should be given to the prophylactic use of doxycycline at the time of IUCD insertion as an approach in preventing PID and other IUCD-related morbidity.