While several studies elsewhere have reported on the prevalence and risk factors for preterm delivery and low birth weight in HIV positive pregnant women, no information on the subject currently exists from our sub region [6–11]. The closest information are reports from the central African countries of Cameroun and Rwanda [6–8]. In addition, the sample size for the two studies from Cameroun and Rwanda were rather small making generalization to Nigeria with the 2nd highest global HIV burden difficult. In this study we determined the incidence and risk factors of preterm delivery using information extracted from prospectively collected data of 1626 HIV positive pregnant women over a 6-year period. Information obtained will assist us and others within our sub region or elsewhere to plan strategies to control and prevent preterm delivery and its associated morbidity and mortality in HIV positive pregnant women.
The preterm delivery rate of 11.1% among the HIV positive pregnant women in this study was similar to 13.5% and 13.6% reported by Towsend and colleagues  from United Kingdom and Taguebue and colleagues  from Cameroun but much lower than 23.5% reported by Dreyfuss and colleagues from Tanzania, East Africa . The relatively low rate of the preterm delivery in our cohort may be related to the fact that most of our patients were diagnosed early as part of routine voluntary HIV testing during antenatal care, allowing for early initiation of treatment of opportunistic infection which could trigger preterm contractions. In addition a significant number was already on treatment before present pregnancy and only a small proportion of them were on PI based HAART(3.5%) which has been shown to be associated with preterm delivery. The cohorts from Tanzania had a high rate of nutritional deficiency and were in more advanced stages of HIV disease compared to our cohort  and thus a high rate of over 23%.
The preterm rate was however significantly higher than rates of 5.8% and 8.5% reported among Nigeria women of unknown HIV status by Etuk  and Chike Obi  from Nigeria . This higher rate may be attributable to HIV, confirming the findings of the association between HIV and preterm delivery.
In previous studies several factors were identified as risk factors for the development of preterm delivery in HIV positive pregnant women [6–11], However only multiple pregnancy, stressful work, presence of opportunistic infection at delivery and first trimester exposure to PI based triple ARV therapy were found to be associated with preterm delivery in this cohort. Our study utilized a larger sample size and methodologically collected data and thus may have been able to remove the background noise that may have influenced the result of other studies. In addition those studies were from regions with different HIV epidemiology. While the HIV incidence is in single digits in west Africa, in east Africa it is in double digits and the epidemic there is more matured. The viral subtypes and genotypes are also quite different.
The most important predictor of preterm delivery in this cohort was multiple pregnancies (aOR: 8.6; CI: 6.73 – 12.9). Though this finding is similar to findings of Taguebue and colleagues from central Africa , studies from UK  and East Africa  did not identify multiple pregnancies as contributing to preterm delivery in HIV positive women. It is also important to note that in our earlier study among HIV negative population from the same region of Nigeria where the current study was conducted, we identified multiple pregnancies as a major risk factor for preterm delivery . Our study thus showed that irrespective of HIV status, multiple pregnancies remains a significant cause of preterm delivery in our country. Women with multiple pregnancies tend to develop preterm delivery because of uterine over distention (malpresentation, polyhydramionos), and premature rupture of membrane, antepartum haemorrhage and intrauterine growth retardation [1, 4, 5], which may trigger uterine contractions.
Opportunistic infections at delivery (aOR: 1.9; CI: 1.1 – 5.7), stressful work (aOR 1.2; CI: 1.05- 4.8) and first trimester exposure to PI based HAART (aOR: 2.4; CI: 1.07 – 3.15) were the other factors shown to be associated with preterm delivery in our cohort. While the association of the presence of OIs and stressful work with preterm delivery in HIV positive women seems obvious, that of first trimester PI based HAART exposure was not immediately obvious. However it is in keeping with previous reports which showed that PI based HAART is associated with preterm delivery . It is also important to state that while other studies were not able to associate the period of HAART and preterm delivery we were able to show that only first trimester PI based HAART exposure and not Non PI based HAART or its duration was associated with preterm delivery. There has been ongoing debate on the role of PI based regimen on preterm delivery rate [18–26]. While some studies reported an association between PI use and preterm delivery [20–23], others have reported no association [24, 25] or have been underpowered to detect a modest increase in preterm deliveries . The results of this study supports other studies that linked preterm delivery to use of PI based regimen [20–24]. In our study we not only used a large sample size but also controlled for possible confounders. Our study was also sufficiently powered to detect a two times risk of developing preterm delivery in pregnant HIV positive women who were exposed to first trimester PI based HAART.
The presence of reproductive tract infection at delivery which was found to be associated with preterm birth at univariate analysis did not retain its significant association with preterm delivery after controlling for the effect of co-morbidity of opportunistic infection. In addition the small proportion of women with reproductive tract infection at delivery in this cohort may have contributed to loss of significance. As a result of the great aversion for caesarean section among Nigerian women , we actively screen for and treat vaginal infection before referral at 36 weeks even when the women have been previously treated. The active screening and prompt treatment of reproductive tract infection ensured that only few women still had reproductive tract infection at delivery.
Though our study tried to eliminate the weakness of other previous studies, we had the limitation of its being a retrospective study. However the data used for this analysis, though retrospectively assembled is part of a methodologically collected data base of a PMTCT follow up study in our facility. Data used were extracted from a prospectively collected data using case record form designed for the PMTCT follow up study thus ensuring the quality and completeness of data used in this study.