Menu
Menu
Menu
Menu
Background: Maternal mortality in South Africa is rising, and heart conditions currently account for 41 per cent of indirect causes of deaths. Little is known about the burden of heart disease in pregnant South Africans.Methods: We systematically reviewed the contemporary epidemiology and peripartum outcomes of heart disease in South African women attending antenatal care. Searches were performed in PubMed, ISI Web of Science, the EBSCO Africa-Wide database, the South African Union Catalogue, and the Current and Completed Research database (South Africa). References of included articles were also hand-searched. Studies reporting epidemiologic data on antenatal heart disease in South Africa were included. Data on morbidity and mortality were also collected.Results: Seven studies were included in the systematic review. The prevalence of heart disease ranged from 123 to 943 per 100,000 deliveries, with a median prevalence of 616 per 100,000. Rheumatic valvular lesions were the commonest abnormalities, although cardiomyopathies were disproportionately high in comparison with other developing countries. Peripartum case-fatality rates were as high as 9.5 per cent in areas with limited access to care. The most frequent complications were pulmonary oedema, thromboembolism, and major bleeding with warfarin use. Perinatal mortality ranged from 8.9 to 23.8 per cent, whilst mitral lesions were associated with low birth weight. Meta-analysis could not be performed due to clinical and statistical heterogeneity of the included studies.Conclusion: Approximately 0.6 per cent of pregnant South Africans have pre-existing cardiac abnormalities, with rheumatic lesions being the commonest. Maternal and perinatal morbidity and mortality continue to be very high. We conclude this review by summarising limitations of the current literature and recommending standard reporting criteria for future reports. © 2012 Watkins et al; licensee BioMed Central Ltd.
Any study investigating antenatal heart diseases in pregnant South African women was considered for review. In order to obtain the best epidemiological data, only studies that were designed to measure hospital- or community-based burden of heart disease were included in the final analysis. Studies were excluded if they did not provide any information regarding a reference population of pregnant women without heart disease. The review was limited to studies conducted in participants over 18 years of age. Although we secondarily collected data on maternal and foetal outcomes, studies were not included or excluded on this basis. Language of publication was restricted to English articles only. Editorials and review articles were excluded. Two independent reviewers (DAW and MS) reviewed lists of articles obtained from several databases relevant to the South African population. The pre-specified search strategy for each database was as follows: MEDLINE was searched with the term “Heart Diseases”[MeSH] AND pregnan*[TIAB] AND “South Africa”[All Fields]; ISI Web of Science was searched with the term TS = HEART DISEASE (and) TS = PREGNAN* (and) CU = SOUTH AFRICA; and the EBSCO Africa-Wide database was searched with the term SU = heart or cardiac or cardiovascular (and) TI = disease* (and) SU = pregnan* (and) TX = South Africa. To search for South African conference proceedings, theses, and abstracts, two internal databases at the University of Cape Town Health Sciences Library were searched. Current and Completed Research (South Africa) [8] was searched using “heart AND pregnancy” and the South African Union Catalogue (SACat) section on South African Theses [8] was searched using the term (“heart” OR “cardiac”) AND “pregnancy.” All databases were accessed during the month of March 2011. Articles were selected on the basis of relevant title with relevant abstract and full text articles were obtained from potentially eligible reports. In the final stage of the search, reference lists of full-text articles were hand-searched. Discrepancies were resolved by consensus discussion between the two reviewers with adjudication by the senior author (BMM) as necessary. Two reviewers (DAW and MS) used a standardised data extraction form to obtain information on study design, patient demographics, and total numbers of each cardiac lesion. Secondarily, basic information was obtained on the rates of specific outcomes: maternal death, pulmonary oedema, thrombosis, haemorrhage, and foetal demise. Again, discrepancies were adjudicated by consensus discussion between the two reviewers with the assistance of the senior author (BMM) as necessary. When study data were incomplete or contradictory with regard to the primary objective, the original author of the manuscript was contacted to clarify his or her findings. Prevalences, case-fatality rates, and when applicable, perinatal mortality rates, are expressed per standard 100,000 deliveries. In hospital-based series and cohorts, we have reported the “prevalence” of antenatal heart disease as the number of patients in the study (all with heart disease) divided by the total number of deliveries at the institution, multiplied by 100,000. In some reports, cases of heart disease were reported in terms of “incidence”; however as most of the cardiac lesions likely predated pregnancy, we felt this term to be imprecise. We have also reported death rates as case-fatality in spite of the fact that papers tended to use the term “maternal mortality” because in our assessment, deaths were almost exclusively due to complications of the cardiac lesion (e.g., pulmonary oedema or coagulopathies related to warfarin misuse). We report “perinatal mortality” as any stillbirth, peripartum or neonatal death (i.e., less than 1 month). Unfortunately some studies did not have full records for all patients [9,10] and thus, we analysed perinatal deaths in a subset of women with cardiac disease, likely inflating mortality rates. We intended to perform a meta-analysis in an attempt to provide an aggregate of the prevalence data from the included studies. All reported numerical data were stored and analysed using Microsoft Excel 2010.
N/A
