It has recently been suggested that infection with human immunodeficiency virus (HIV) in pregnant women might reduce the risk of developing preeclampsia.1 One of the pathogenetic explanations for preeclampsia is maladaptation of the immune system to paternal antigens.2 Local immune hyperreactivity may cause incomplete trophoblastic invasion of the uterine spiral arteries, thus exposing the placental site to vasoconstriction and the eventual development of preeclampsia. Immune deficiency, induced by HIV or any other cause, could therefore inhibit a tendency to immune hyperreactivity and thus prevent the development of preeclampsia. The study by Wimalasundera et al1 described the rarity of preeclampsia in untreated HIV-positive women compared with HIV-negative women and HIV-positive patients on antiretroviral drugs and raised intriguing questions about the roles of HIV infection, impaired immunity, and antiretroviral drugs in the pathogenesis of preeclampsia. In view of the high rates of both HIV seroprevalence (29.3%)3 and preeclampsia-eclampsia (7.8%)4 in our region, we felt well placed to investigate the relationship between the 2 conditions. The objective of this study was to estimate whether untreated HIV-positive pregnant women had a lower rate of preeclampsia-eclampsia than HIV-negative women.
MATERIALS AND METHODS
The study was set in Soweto, a large urban settlement to the southwest of Johannesburg, South Africa, with a population of approximately 2 million. Chris Hani Baragwanath Hospital, a 3,200-bed institution, provides hospital care to the residents of this area and also functions as a tertiary referral center for hospitals in neighboring regions. Several primary care clinics in Soweto refer patients to this hospital when necessary. Five of these clinics, Zola, Dobsonville, Mofolo, Lillian Ngoyi, and Chiawelo, operate midwife obstetric units where deliveries are performed. All complicated and high-risk deliveries are transferred to Chris Hani Baragwanath Hospital for further management. The referral criteria are set out in manuals distributed to each clinic.5 Women with hypertension in pregnancy are referred to this hospital for further management. Human immunodeficiency virus seropositivity is not considered a high-risk obstetric condition, and HIV-positive women are managed by midwives if they have no obstetric reason for referral to the hospital. In Soweto, all antenatal HIV testing is conducted by the Perinatal HIV Research Unit at Chris Hani Baragwanath Hospital. Testing is done on blood using Determine HIV-1/2 (Abbott Japan Co, Ltd, Fukui, Japan) rapid immunochromatographic test, with positive results submitted for confirmation using Capillus HIV-1/HIV-2 (Trinity Biotech PLC, Bray, Ireland) rapid latex aggregation test. Women who are HIV-positive are given single-dose intrapartum nevirapine to prevent mother-to-child transmission, and their newborns receive a single dose of nevirapine shortly after delivery, in line with the HIVNET 012 protocol.6 Counting of CD4 cells is not routinely performed because maternal antiretroviral therapy is not as yet given to HIV-positive patients in South African government hospitals.
The study population consisted of pregnant Soweto residents, who delivered in public health facilities (Chris Hani Baragwanath Hospital and the Soweto clinics) from March to December 2002. Only women who attended antenatal clinics and underwent HIV testing, and who gave birth at a gestational age of 20 weeks or more, were included in the study. Non-Sowetan patients were excluded because many of these are high-risk obstetric referrals from other hospitals. Referral may be related to both HIV infection and preeclampsia-eclampsia, and inclusion of these patients might have introduced confounding. The sample size was calculated using Epi-Info 6 statistical software (Centers for Disease Control and Prevention, Atlanta, GA). To attain statistical significance (P = .05 at a power of 80%), we required a sample size of 2,588 to show a decrease in the rate of preeclampsia-eclampsia from 8% in HIV-negative women to 5% in those who were HIV-positive, assuming an HIV seroprevalence rate of 30%. This number was rounded up to 2,600 and divided into representative proportions of hospital and clinic deliveries, based on the previous years’ total confinements. This calculation yielded a sample size that required us to use 1,797 hospital births and 803 clinic deliveries. For the hospital cohort, randomly selected clusters were used, based on the last 3 digits of the 8-digit hospital number. Number allocation of files in the hospital bears no relationship to demographic or clinical characteristics, and the use of such clusters was not expected to result in any bias or design effect in analysis. We selected each cluster from an electronically generated list of random 3-digit numbers. All files ending with these numbers (clusters) were drawn from the hospital record room, and individual files in each cluster were individually assessed for eligibility. For the clinic cohort, the sample was stratified to include a representative number of women from each of the 5 clinics, based on the annual number of deliveries. At each clinic, we used a simple random sample based on numbered deliveries in the birth register.
Data collection was done by record review. For the hospital confinements, patient files were used to obtain HIV results, which are usually recorded once a woman has agreed to be tested. For the clinic confinements, demographic and obstetric data were obtained from birth registers, and the HIV results were obtained from the results books in the Perinatal HIV Research Unit. Data collected included maternal age, parity, gestation at delivery, weight, and hemoglobin concentration. Weight and hemoglobin concentrations were not obtainable for the clinic cohort because this information is not routinely recorded in the birth registers. A search was made in each case for hypertensive disease, the presence or absence of proteinuria, and whether the condition had its onset during the pregnancy. All data collection was undertaken by the 3 authors, using a data sheet designed for the study.
The definitions of hypertensive disorders are based on those proposed by Davey and MacGillivray in 1988.7 Proteinuria was defined as 1+ on reagent strip on at least 2 occasions, or 300 mg in a 24-hour specimen of urine. Hypertension was defined as a diastolic blood pressure of 110 mm Hg or greater on any one occasion, or a diastolic blood pressure of 90 mm Hg or greater on 2 or more occasions, 4 or more hours apart. Preeclampsia was defined as hypertension associated with proteinuria, which developed after 20 weeks of pregnancy. Where the mother initiated antenatal care after 20 weeks of pregnancy, preeclampsia was defined if it was detected subsequent to a normal blood pressure reading at a previous antenatal visit. The term preeclampsia-eclampsia includes all patients with preeclampsia, severe preeclampsia, eclampsia, and superimposed preeclampsia.
We analyzed data on Epi-Info 6 statistical software, using the χ2 test and Fisher exact test for comparison of frequencies, and Student t test and the Mann-Whitney test for comparison of means and medians, respectively. A P value < .05 was accepted as indicating statistical significance. To exclude confounding explanatory variables, we performed stratified analysis using the Mantel-Haenszel test. Permission to undertake the study was obtained from the University of the Witwatersrand's Committee for Research on Human Subjects and from the hospital authorities.
To achieve the necessary sample size of 1,797 hospital subjects, 3,333 files were reviewed. Of these, 878 were excluded because they belonged to women who were not resident in Soweto, and 658 had no results, either because they were not tested (404), were tested but results could not be found (160), or declined HIV testing (94). For the clinic cohort of 803 subjects, the number of exclusions was not recorded. The total sample was 2,600. In the hospital cohort, 516 women were HIV-positive (28.7%), and in the clinic cohort, 188 women (23.4%) were HIV-positive. The HIV seroprevalence rate for the whole sample was 27.1% (95% confidence interval 25.4–28.8%). CD4 counts were available for 13 women (0.5%). Ten women had acquired immunodeficiency syndrome (AIDS), 6 on the basis of a CD4 count of less than 200/mm3, and 4 because they had AIDS-defining conditions. Two of these women died in the puerperium. Baseline obstetric information is shown in Table 1, which compares basic clinical data between HIV-positive and HIV-negative women. Primiparity was more frequent in HIV-negative women (42.5% versus 33.9%; P < .001). Women who were HIV-positive had lower mean hemoglobin values (10.9 versus 11.5 g/dL; P < .001) and lower mean weights (69.1 versus 71.5 kg; P = .008) than HIV-negative women, although these data were available only for hospital confinements.
Four hundred fifty women (17.3%) were found to be hypertensive, with 150 (5.8%) having proteinuric hypertension. One hundred eight HIV-negative women had proteinuric hypertension (5.7%), compared with 42 HIV-positive women (6.0%; P = .75). The corresponding frequencies for preeclampsia-eclampsia (preeclampsia, eclampsia, and superimposed preeclampsia) were 98 (5.2%), compared with 40 (5.7%; P = .61), respectively. The frequencies of the different grades of hypertension are shown in Table 2.
Mantel-Haenszel stratified analysis was performed to identify confounding explanatory variables, and examined the influence of age less than 30 years, primiparity, weight less than 60 kg, hemoglobin level less than 10 mg/dL, clinic delivery, and delivery at less than 37 weeks, on the development of proteinuric hypertension. Only gestational age at delivery proved to be confounding. Women who gave birth before 37 weeks had a 15.4% rate of proteinuric hypertension if they were HIV-negative, compared with 8.4% if they were HIV-positive (P = .025). Of those women who delivered at term (≥ 37 weeks), HIV-negative and HIV-positive women did not differ significantly in rates of proteinuric hypertension (3.3% versus 5.1%, respectively). This is shown in Table 3.
Our study failed to show any association between HIV seropositivity and the risk of developing preeclampsia-eclampsia. This lack of association was observed for all grades of hypertension, whether mild or severe, nonproteinuric or proteinuric. The study sample was representative of an entire community of public health service users and was not restricted to high-risk obstetric cases delivered in hospitals. Although HIV-positive women were slightly older, were less likely to be primiparous, and had lower weights and lower hemoglobin levels than HIV-negative women, these variables did not confound the results. The confounding influence of preterm delivery is interesting but has a simple explanation. Women who are HIV-positive are known to have a higher risk of preterm birth,8,9 perhaps as a result of systemic infection or chorioamnionitis, and those who are destined to become preeclamptic may not develop the condition by the time they give birth. Human immunodeficiency virus seropositivity, or any condition causing preterm birth, can therefore “protect” against preeclampsia. It should be noted that our study could not specifically investigate the relationship between immune deficiency and preeclampsia-eclampsia, because of the very low rate of CD4 count testing. We were thus unable to distinguish immune-deficient women in our population. Patients without AIDS are essentially immunocompetent, and a protective effect of HIV against preeclampsia may only be noted in patients with AIDS or with subnormal CD4 counts.
The findings of Wimalasundera et al,1 which prompted our research, can be questioned on the basis of small sample size, high-risk referral hospital population, and possibly the influence of preterm delivery as described above. In their article, the authors cited an American study that described an unusually low rate of preeclampsia (0.7%) in a population of HIV-positive pregnant women.10 However, on closer reading of that article, we found that the authors specifically excluded preeclamptic women who required elective delivery, thus removing the bulk of these patients from their study population and invalidating any inferences about HIV and preeclampsia.
Some limitations are inherent in our research. We necessarily excluded all women who did not attend antenatal clinics. This number is, however, less than 5% in our area11 and should not materially affect the results. We also had to exclude women who refused HIV testing and whose results could not be retrieved. In terms of its value as a local audit, our study exposed a deficiency in retrieving and recording HIV results. It is difficult to speculate about whether this shortfall is health care provider- or patient-related, but the consequences of this are potentially serious for HIV-positive women for whom the necessary precautions to prevent HIV transmission are not taken, especially considering the local HIV seroprevalence rate of 27.1%.
Human immunodeficiency virus infection may be protective by resulting in preterm delivery before the clinical onset of preeclampsia-eclampsia. We could, however, find no direct association between HIV infection and a reduction in the risk of developing preeclampsia. Any further study should investigate the role of HIV-related immune deficiency by including indices of immunity, such as CD4 counts, and a large enough sample size.
1.Wimalasundera RC, Larbalestier N, Smith JH, de Ruiter A, McG Thom SA, Hughes AD, et al. Pre-eclampsia, antiretroviral therapy, and immune reconstitution. Lancet 2002;360:1152–4.
2.Pridjian G, Puschett JB. Preeclampsia. Part 2: experimental and genetic considerations. Obstet Gynecol Surv 2002;57:619–40.
3.Department of Health. National HIV and syphilis sero-prevalence survey of women attending antenatal clinics in South Africa, 2000. Pretoria, South Africa: Department of Health; 2001.
4.Buchmann EJ. Perinatal health in the Chiawelo district of Soweto: maternal characteristics and clinical outcomes. CHASA J Compr Health 1996;7:118–23.
5.Buchmann E, Thembekile N, Moabelo E. Primary care midwifery: practical guidelines for nurses at clinics and midwife obstetric units. 2nd ed. Johannesburg, South Africa: Chris Hani Baragwanath Hospital and Central Wits Region Community Health Services; 2001.
6.Guay LA, Musoke P, Fleming T, Bagenda D, Allen M, Nakabiito C, et al. Intrapartum and neonatal single-dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet 1999;354:795–802.
7.Davey DA, MacGillivray I. The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 1988;158:892–8.
8.Ellis J, Williams H, Graves W, Lindsay MK. Human immunodeficiency virus infection is a risk factor for adverse perinatal outcome. Am J Obstet Gynecol 2002;186:903–6.
9.Coley JL, Msamanga GI, Fawzi MC, Kaaya S, Hertzmark E, Kapiga S, et al. The association between maternal HIV-1 infection and pregnancy outcomes in Dar es Salaam, Tanzania. BJOG 2001;108:1125–33.
10.Stratton P, Tuomala RE, Abboud R, Rodriguez E, Rich K, Pitt J, et al. Obstetric and newborn outcomes in a cohort of HIV-infected pregnant women: a report of the Women and Infants Transmission Study. J Acquir Immune Defic Syndr Hum Retrovirol 1999;20:179–86.
11.Buchmann EJ. Perinatal health in the Chiawelo district of Soweto: health services, patient referral and clinical management. CHASA J Compr Health 1997;8:31–5.