Systematic review of observational studies of the impact of cardiovascular risk factors on preeclampsia in sub-saharan Africa

Background Maternal mortality is a public health issue, particularly in low- and middle-income countries (LMIC). Sub-Saharan Africa (SSA) is the region most affected worldwide by maternal mortality, and preeclampsia is one of the main causes. We performed a systematic review of observational studies to identify the impact of cardiovascular risk factors on preeclampsia in SSA with a more representative sample. Methods Databases: PubMed and Google Scholar were searched to identify published studies. Studies were included if they reported results on the link between at least one cardiovascular risk factor and preeclampsia. Relevant studies quality was assessed with the Newcastle-Ottawa Scale (NOS). Odds ratios and relative risk (RR) were reported with their confidence intervals. Results Twelve articles (8 case-controls, 3 cohorts, 1 cross-sectional) were included in this review, with a total of 24,369 pregnant women. Cardiovascular risk factors such as chronic hypertension, overweight, obesity, diabetes and alcohol were significantly associated with a high risk of preeclampsia. Very few data were available concerning some risk factors. None of the articles reported tobacco consumption as a preeclampsia risk factor. There is a lack of data from French-speaking SSA countries. Conclusion Cardiovascular risk factors increase the risk of preeclampsia. Our results suggest the need for prospective cohort studies to ascertain this association in order to reduce maternal mortality due to preeclampsia.

through both fetal and maternal complications. In the literature, authors reported the involvement of many risk factors; the most frequently cited are advanced age, multiple pregnancies, nulliparity, personal history of preeclampsia and pre-pregnancy medical conditions such as chronic high blood pressure, type 2 diabetes and renal failure [6,7]. In the last years, with the spread of cardiovascular diseases, many studies have reported that cardiovascular risk factors are involved in the onset of preeclampsia [8][9][10][11]. However, most of these studies were conducted in developed countries. In SSA, a large majority were conducted in English-speaking countries. To estimate the real contribution of cardiovascular risk factors, it is necessary to provide representative data. Moreover, especially about SSA countries, that have one of the highest maternal mortality rates worldwide. Since most of these risk factors are preventable, early identification of women at risk appears to be mandatory for optimal management.
This study aimed to systematically review all observational evidence regarding the impact of cardiovascular risk factors on preeclampsia in SSA.

Search strategy
This review was conducted according to PRISMA guidelines [12]. The study protocol was approved by the review team before searching was performed. Studies were retrieved through internet research in PubMed and Google Scholar databases from March 6th to April 30th, 2019. Searching was performed by using terms in line with the "PICO" method. The search was done with the following terms: hypertension, diabetes mellitus, overweight, obesity, tobacco use, alcohol consumption, dyslipidemia, preeclampsia, hypertensive disorders, Africa South of Sahara, and the names of all SSA countries. Keywords were used as follow: (("hypertension" OR "diabetes mellitus" OR "overweight" OR "obesity" OR "alcohol consumption" OR "tobacco use" OR "dyslipidemia") AND (preeclampsia OR "hypertensive disorders")) and all SSA countries names. References of relevant studies were manually screened for additional articles. Titles and abstracts were screened first. Then a full quality assessment review of articles was done before selection for final review. Two authors independently assessed the articles for inclusion in the review. Any discrepancy which arose between the authors in the review process was resolved through discussion for consensus.

Eligibility criteria
Studies were included in the review if: the design was cross-sectional, case-control, prospective or retrospective cohorts, it evaluated the link between at least one cardiovascular risk factor (chronic hypertension, diabetes, dyslipidemia, obesity, overweight, tobacco, alcohol) and preeclampsia. it was published in English or French.
We did not use a restriction on the years of publication. Only studies meeting our inclusion criteria were selected. The quality of retrieved studies was assessed using the Newcastle-Ottawa scale (NOS) [13]. Selection, comparability, and outcome/exposure assessment were rated for case-control and cohort studies separately. The rating system scores from 0 (highest degree of bias) to 9 (lowest degree of bias) allocated among three criteria: selection (0 to 4), comparability (0 to 2) and outcome/exposure assessment (0 to 3). Studies with a total score of 0 for one of the three criteria were excluded.

Data extraction
Selected studies were analyzed for data extraction. We extracted data concerning: authors, years and areas of studies, study populations, study design, sample size, sampling methods, inclusion and exclusion criteria, preeclampsia definition, risk factors, length of follow up for cohort studies and exposure variables for all studies. We also reported all studied variables per each article. Targeted risk factors were chronic hypertension, diabetes mellitus, overweight, obesity, alcohol, tobacco, and dyslipidemia. Preeclampsia was a main outcome reported in the included studies. Bias, confounding and their management were assessed in included studies by reviewing each study methods and discussion sections. Studies with significant bias were not included in this review. When available, results were reported as odds ratios or relative risks with 95% confidence intervals.

Outcomes
The main outcome was preeclampsia. It was defined as new onset of systolic blood pressure of 140 mmHg or more and diastolic blood pressure of 90 mmHg or more, at least twice on two occasions at least 4 h apart, with either 24 h proteinuria ≥300 mg or urinary protein strip showing ≥1+ without urinary tract infection, according to the American College of Obstetricians and Gynecologists, in almost all the included studies. However, one study used a higher blood pressure cut-off of 160/110 associated with random proteinuria ≥2+ [14].

Results
A total of 770 articles were retrieved through electronic searching in databases and 4 studies from references. Of this, 44 duplicates were removed and 554 studies were removed because they were conducted in a region other than SSA, or preeclampsia (hypertensive disorders) was not a main outcome. One hundred and seventy-six titles and abstracts were screened. Finally, twelve studies (8 case-control, 3 cohorts and one cross-sectional) were included in this systematic review. The study selection process is summarized in Fig. 1.
Most studies had high quality case definition and subject selection. Cases and controls were selected from the same population. There was no information on the proportion of non-response and the representativeness of the sample in some studies. All included studies scored 7 and above at quality assessment. Table 1 and Table 2 present NOS-based evaluation of case-control and cohort studies, respectively.

Studies description
The characteristics of included studies are shown in Table 3. They were conducted between 1998 and 2018. Participants were recruited during pregnancy or after delivery. Preeclampsia standards were not identical among all included studies. They were based either on the American College of Obstetrician and Gynecologists or the International Society for the Study of Hypertension in Pregnancy (ISSHP) standards. Only one study was conducted in a French-speaking SSA country. In cohort studies, pregnant women were included before 20 weeks of pregnancy. Case-control study populations mostly included women who were more than 20 weeks pregnant or those who had already given birth. Some cardiovascular risk factors such as dyslipidemia and tobacco use were not studied. Few studies assessed tobacco (4/12). Sample size was calculated in almost all the included studies. The odds ratios are summarized according to the risk factors in Table 4.

Diabetes
In our review, only one case-control study reported the role of pregestational diabetes in the occurrence of preeclampsia. Pregnant women with pre-gestational diabetes were four times at risk of developing preeclampsia [19]. Pregnant women with known diabetes before pregnancy were not included in some studies [22].

Overweight and obesity
Three case-control studies reported that being overweight before pregnancy increased the risk of developing preeclampsia by up to seven times [10,15,22]. All cohort studies [20][21][22] and two case-control studies [10,17] in this review reported that obesity was an important risk factor for preeclampsia.

Alcohol
One case-control study suggested an association between alcohol and preeclampsia. In this study, women who reported alcohol consumption during pregnancy were more likely to develop preeclampsia compared to those who did not drink alcohol in multivariable analysis (aOR = 3.97, 95% CI = 1.80-8.75) [18]. However, Kiondo et al. found no association between alcohol and preeclampsia [16].

Discussion
The prevalence of cardiovascular risk factors in SSA women is not negligeable. The median prevalences were:

Selection
Representativeness of the exposed cohort 1 1 1 Selection of the non-exposed cohort 0 1 1   hypertension (29%), diabetes (7%), overweight (35%), obesity (11%), alcohol consumption (13%), tobacco (2%) [24][25][26]. Apart from tobacco, there is a significant prevalence of cardiovascular risk factors in SSA women. This review presents evidence from published observational studies on the association between cardiovascular risk factors and preeclampsia. Chronic hypertension can increase the risk of developing preeclampsia during pregnancy by 3 to 10-fold. This association was also found among Latin American and Caribbean women and is generally present in LMIC [8,27]. A possible explanation for this is the lack of follow-up of chronic hypertension among women. These results emphasize the need for detecting and managing hypertension among women in SSA.
Other high-risk factors for preeclampsia are overweight and obesity. These findings corroborate Proorolajal's et al. meta-analysis, which concluded that overweight and obesity are predictors of preeclampsia. Paré et al., in a prospective cohort study, reported that  the risk of preeclampsia increases with the body mass index [7,28]. These results require clarification, as the BMI in the pregestational period could not be evaluated in most cases. In LMIC, women are not always able to accurately report their body mass index or weight, so some authors have used mid-upper arm circumference to assess nutritional status, because this parameter may be stable during pregnancy [29]. This methodological particularity could lead to an overestimation of obesity cases, because pregnancy is responsible for weight gain in most women. Other authors have evaluated BMI based on the weight reported by pregnant women, which is an important methodological limitation. Obesity is associated with oxidative stress, increased levels of circulating inflammation markers, dyslipidemia, insulin resistance and impaired endothelial function. These metabolic and biochemical disturbances likely predispose to an intrauterine environment favorable to placental perfusion disorders and endothelial dysfunction in preeclampsia [30]. Although only one study reported a significant association between pre-gestational diabetes and preeclampsia, our findings are similar to those reported by Duckitt et al. in a systematic review. Being diabetic before pregnancy increases the risk of preeclampsia by 4 times [31]. However, one study alone does not permit generalizing these findings. Like diabetes, alcohol consumption and dyslipidemia have been poorly studied. Both factors may increase the risk of preeclampsia [18]. However, Kiondo et al. found no association between alcohol consumption and preeclampsia in Uganda [16]. These results may be explained by the fact that alcohol consumption was selfreported, and in SSA and worldwide, alcohol consumption by pregnant women is not considered acceptable, or is even prohibited. No studies evaluated the impact of tobacco on the occurrence of preeclampsia. This may be explained by the fact that in SSA, tobacco use is not common among women. However, tobacco in its smoking form reduces the risk of preeclampsia by 35%, although the mechanism remains controversial [32]. This protective effect of smoking on preeclampsia risk is limited to low-risk pregnancies only. Smoking is an independent risk factor for superimposed preeclampsia in chronic hypertension, which is significant in the context of other cardiovascular risk factors [33]. Despite this lack of information, the harmful effects of tobacco on health, especially in arteriosclerosis, should not be neglected.
In LMIC, there is an epidemiological transition from communicable disease to non-communicable disease, led by cardiovascular disease, which accounted for 33.2% of deaths worldwide in 2008 [26]. Since cardiovascular risk factors are involved in the occurrence of preeclampsia, the screening of women at risk could help reduce mortality related to hypertensive complications of pregnancy, in particular preeclampsia. Reducing preeclampsia and cardiovascular risk factors would lessen long-term cardiovascular complications and therefore female mortality linked to non-communicable disease. Indeed, the occurrence of preeclampsia can increase by four the risk of heart failure and by two the risk of stroke, coronary heart disease or death from a cardiovascular event in the medium and long term.
However, this review has some limitations. The studies were all conducted in hospital settings. Therefore, there might be a selection bias because of home deliveries. The lack of studies in French speaking Sub-Saharan African countries does not permit generalizing our results. The self-reported data collected in the studies does not guarantee the reliability of the results. Some authors did not consider possible confounding factors that could lead to misestimation of their results. We included studies from various geographical areas, conducted with specific methodologies. One limitation of our review is that we report results from studies with potential limitations and which did not focus exclusively on preeclampsia.

Conclusion
Cardiovascular risk factors are involved in the occurrence of preeclampsia in SSA. This review highlights the role of chronic hypertension, diabetes, overweight, obesity, alcohol. Some of the included studies have methodological limitations. The lack of data suggests a need for prospective cohort studies to assess the impact of cardiovascular risk factors on the occurrence of preeclampsia in SSA. Pre-pregnancy cardiovascular risk factor screening is needed to prevent preeclampsia among women.
Abbreviations SSA: Sub-Saharan Africa; LMIC: Low-and middle-income countries; NOS: Newcastle-Ottawa scale; WHO: World health Organization; PRIS MA: Preferred reporting items for systematic reviews and meta-analyses; ISSHP: International society for the study of hypertension in pregnancy; BMI: Body mass index