Our study which was conducted between 2002 and 2008, showed an incidence of 5.7/100 wyar (3.3-8.1) at nine months post-partum and an overall HIV incidence of 2.3/100 wyar (95% CI 1.1-4.1) six years post-partum. However a similar study conducted in Harare between November 1997 and January 2001 showed cumulative incidences of 3.4% (3.0-3.8) and 6.5% (5.7-7.4) amongst women followed up for 12 and 24 months post-partum respectively . By comparing our incidence of 5.7/100 wyar (3.3-8.1) at nine months to that of the previous study of 3.4% (3.0-3.8) at 12 months we can see clearly from the confidence intervals that there are no statistically significant differences between the two estimates, indicating that in this study we are unable to provide evidence of a declining incidence. However the overall incidence of 2.3/100 over the entire six years follow-up period appears to indicate a general decline although this can be attributed to following up the same group over a long period. As HIV-1 prevalence continues to decline in Zimbabwe there is need to develop and implement effective local strategies that will result in a reduction of the incidence of HIV.
During the early days of the HIV epidemic acquisition of HIV infection was associated with an increased level of education and high socioeconomic position . Twenty-five years into the epidemic this study confirms that low education among women is now significantly associated with acquisition of HIV-1. In Tanzania this reversal of risk of acquisition of HIV from educated to non-educated was noticed between 1991 and 2005 . If all children, particularly girls, are educated beyond primary school the end-result may lead to a reduction of the incidence of HIV . We also note an association between acquisition of HIV and sexual abuse among women but suggest that this needs a qualitative approach to reveal pertinent aspects of the abuse.
We also showed that women that have children fathered by different men are about three times more likely to acquire HIV compared to women whose children have the same father (Table 3). Having children with different men reflects a high level of risky behavior as it shows a tendency of having unprotected sex with different men. Since about 70% of the participants were below 25 years of age at enrolment such risky behavior could be a result of inability to meet reproductive health needs after sexual debut , as corroborated by the high rate of co-infections in this population. Over 80% of the women had at least two sexually transmitted infections/conditions. STIs are important in that most HIV infections occur in the presence of other STIs [15–18].
The earliest HIV-1 seroconversions were noticed at the six weeks' visit after childbirth (Table 1). Because HIV-1 antibodies can take between three and six weeks to appear in detectable amounts in peripheral blood  the participants that had detectable HIV antibodies at six weeks could have been actually infected before enrollment but had undetectable amounts of antibodies at the time of screening which was on average four weeks before childbirth. The mothers may possibly have been in the acute phase of infection around the time of delivery and yet they did not receive nevirapine for PMTCT and were therefore at risk of transmitting the virus to their babies antepartum, intrapartum or during breastfeeding [7, 20]. Furthermore, because of its traumatic nature, vaginal delivery often causes microscopic ulcerations  which may increase HIV portals of entry/exit. It is important therefore to emphasize condom use in preventing transmission of STIs including HIV-1 during pregnancy and after delivery. Also there is need for use of simple and affordable point-of-care tests that can detect primary infections late in pregnancy in order to reduce the risk of mother to child transmission.
Pregnancy, whether intended or unintended, is an outcome of unprotected sex. An incidence rate of 5.7/100 wyar within nine months of delivery demonstrates that pregnancy is one of the major events following which HIV seroconversion can occur. HIV acquisition may be high during breastfeeding since it is considered a safe period where no other contraceptive methods (including condoms) need to be used. However we hope that the opt-out testing scheme recently introduced in antenatal clinics will increase the number of pregnant women with knowledge of their HIV status  and thereby help reduce transmission of HIV.
About 50% (18/39) of the seroconverters were aged 20 years and below at enrollment but of note are the four that were still teenagers when they acquired HIV. These four teenagers initiated sex, became pregnant, acquired HIV (and possibly other STIs) and gave birth, all in a short period before becoming mature adults. Better health for young women and men is most likely to be achieved if youths delay their sexual debut until they are physically and mentally mature to begin their married life . Instead, the young women have relationships with older partners and the age difference between them and their partners is a risk factor for acquisition of HIV . Furthermore, women that knew that their partners had other sexual partners were also at a greater risk of acquiring HIV demonstrating how powerless women are in negotiating safe sex (Table 4). It is therefore imperative for preventive strategies to engage men, especially when almost 100% of married women report having sex with their spouses only . A high incidence in young adults needs to be addressed if a future AIDS-free generation is to be achieved and this has to be addressed from the perspectives of both the young woman and the older partner. Men can be involved by setting up mobile men-friendly reproductive health clinics in industrial areas where they are free to attend to get treatment and/or counseling. The prevalence of HIV at baseline in this cohort was 25.6%. The prevalence has continued to decline to reach 15.1% in 2008  and this decline is attributed primarily to behavior change  yet our study shows that about 90% of both the seroconverters and non-seroconverters in this study do not believe that abstinence protects against HIV (Table 4) even though abstinence, be faithful and condom use (ABC) has been for a long time a cornerstone of HIV prevention strategies.
Following up the same group of participants for a long time has its limitations which may demand careful interpretation of the data. Firstly, there may be a systematic selection of those participants that are less likely to seroconvert thereby leading to a possible underestimation of the true incidence. Secondly, as young people get older a high risk population is removed from the cohort with no replacement and this may again lead to the underestimation of HIV incidence. In this study the youngest woman at enrolment (15 years) was 21 years at the end of the study. Thirdly, the same participants may become targets of several other HIV prevention studies thereby resulting in an apparent reduction of incidence. Finally, the introduction of VCT facilities and the "opt out" option may result in more participants knowing their HIV status thus causing people to less likely acquire HIV .