Determinants of preeclampsia and eclampsia among women delivering in county hospitals in Nairobi, Kenya [version 1; peer review: 1 not approved]

Background: Preeclampsia is defined as the onset of a new episode of high blood pressure in a woman usually after 20 weeks gestation plus proteinuria, whereas eclampsia is defined as generalized seizures in a pregnant woman who generally has preeclampsia criteria. Preeclampsia and eclampsia are hypertensive disorders of pregnancy and thus, among the top causes of maternal death worldwide. The objective of this study was to investigate risk factors for preeclampsia/eclampsia. Methods: This was a hospital-based unmatched case-control study carried out among women of reproductive age (15-49 years) who have given birth at Nairobi County Hospitals and admitted to the postnatal ward July-September 2019 with a sample size of 352 participants (88 cases and 264 controls). All cases were selected, while controls were simple random sampled, as per eligibility criteria. Information or data were gathered using a structured interviewer-administered questionnaire and data abstraction tool. Descriptive analysis was carried out, where, categorical variables were presented in percentages or proportions, whereas; continuous variables were presented in means, standard deviations, and range. This was followed by a bivariable mixed-effect logistic regression analysis and a multivariable mixed-effect logistic regression analysis using the significant variables from bivariable analysis. Results: Of all the 88 cases enrolled in the study 5 (5.68%) had eclampsia and 83 (94.32%) had preeclampsia. There was a significant association between personal history of hypertension (AOR=7.1; 95% CI: 2.6-19.3, p=0.001), Occupation as a housewife (AOR=3.1; 95% CI: 1.1-8.8, p=0.034), nulliparity (AOR=7.5; 95% CI: 1.5-37.5, p=0.015) , Open Peer Review


Introduction
Preeclampsia affects 2-10% of pregnant women globally and eclampsia 0.03-0.05% 1 . However, the overall prevalence of preeclampsia according to studies varies from 4.5% to 23% 2 . Preeclampsia affects about 4% of pregnancies in the United States of America 3 . In Kenya, the incidence of preeclampsia is about 0.3% 4 while the prevalence is 6.1% 5 . Similarly in Ethiopia, the prevalence of preeclampsia is about 5% 6 . In Nigeria, the prevalence of preeclampsia ranges between 2 and 16.7% 7 .
The maternal mortality ratio in Sub-Saharan Africa is estimated to be 510 maternal deaths per 100,00 live births 8 . Maternal and fetal complications in addition to maternal and fetal mortality are much greater in mothers with pre-eclampsia and eclampsia than those without 2 . Preeclampsia is the second top cause of maternal death globally, which may lead to grave maternal complications (stroke, eclampsia, and organ failure), and poor perinatal outcome for the fetus and infant especially intrauterine development restriction, low birth weight, and stillbirth 9 . Preeclampsia develops in 20% of first pregnancies and entails more than 40% of premature birth resulting from treatment 2 . About 98% of maternal, fetal, and neonatal death related to preeclampsia and eclampsia occur in low-income countries, with much of this mortality in South Asia and Sub-Saharan Africa 10 . Preeclampsia also accounts for increasing maternal & perinatal/infant mortality, and it is a major cause of maternal mortality 15-20% in developed countries 11 .
Preeclampsia and eclampsia are among the hypertensive disorders of pregnancy. About 13% of maternal mortality globally is due to hypertensive disorders of pregnancy and the number is even higher in developing countries with an estimate between 20-80% in Africa and Latin America 1 .
Many risk factors have been associated with preeclampsia/ eclampsia 2 . Nevertheless, a full account of the risk factors of preeclampsia/eclampsia has not been well established in the Kenyan population 4 . Therefore, this study was conducted to investigate some risk factors of preeclampsia/eclampsia namely; socio-demographic, reproductive and obstetric, clinical, behavioral and family history-related factors, with a view of informing policy, creating awareness, and formulating strategies to improve antenatal care and delivery services among women of reproductive age in Kenya.

Study design and setting
This was a hospital-based unmatched case-control study carried out at Nairobi County Hospitals' postnatal ward. This study design was chosen because it is appropriate for identifying risk factors associated with pre-eclampsia/eclampsia, which are latent conditions among gravid women.
This study was carried out at Mbagathi District Hospital (Kibra), Mama Lucy Kibaki Hospital (Embakasi Central), and Pumwani Maternity Hospital (Kamukunji) in Nairobi County, Kenya; with a catchment population of 4.6 million people, 58% of which are slum dwellers. These are public county hospitals offering primary, secondary, and tertiary health care services to Nairobi residents and neighboring counties such as Kajiado, Kiambu and Machakos counties.

Study population and eligibility of participants
The study population was women of reproductive age (15-49 years) who have given birth at Nairobi County Hospitals and admitted to the postnatal ward from July-September, 2019. The controls and cases were selected from the study population-based outlined criteria for eligibility. A postnatal ward mother that met/satisfied the definition of a case or control and consented to participation was included. Those aged less than 18 years old were included in the study, but both the parent/guardian and patient were required to sign consent and assent forms, respectively. A postnatal ward mother that did not meet the definition of a case or control was excluded from the study.
Case definition and recruitment A case was defined as a woman of reproductive age (15-49 years) admitted in postnatal wards having been diagnosed of preeclampsia (high blood pressure during pregnancy, ≥140/90 mmHg, plus proteinuria ≥300 mg/24 hours or ≥ 1+ dipstick) or eclampsia (high blood pressure during pregnancy, proteinuria plus generalized seizures) at Nairobi County Hospitals July-September, 2019. All cases that met the case definition were selected during the study period, July-September 2019, to obtain the required sample of 88 cases and the response rate was 100%. Thus, each case was selected after the physician (medical officer or consultant obstetrician/ gynecologist) had made a diagnosis (preeclampsia/eclampsia) and client admitted to the postnatal ward.

Control definition and recruitment
Control was defined as a woman of reproductive age (15-49 years) admitted to the postnatal ward without preeclampsia/ eclampsia at Nairobi County Hospitals from July-September 2019. Controls were simple randomly sampled from a list of women admitted to each of the three hospitals' post-natal ward without preeclampsia/eclampsia which was recruited at the time of selection of the cases. A total of three controls for each case were selected by the time of selection to cases. The sampling frame was updated as per deliveries taking place in study hospitals from July-September 2019.

Sample size determination
The sample size of 352 participants (264 controls & 88 cases) was determined as specified by Kelsey, Jennifer, and others 12 for case-control studies as follow: , Since the outcome is rare in the Kenyan population, the ratio of 3 controls to 1 case was used in calculating the sample size but still maintains the statistical power of the study; N 1 is the number of cases and N 2 is the number of controls. Furthermore, P 1 is the proportion of cases exposed thus, was the proportion of women with pre-eclampsia/eclampsia that attend less than four antenatal clinic (ANC) visits; P 2 is the proportion of controls exposed and this was the proportion of women without pre-eclampsia/eclampsia that attends less than four ANC visit, that was set at 42% 13 ; Z α (1.96) and Z β (-0.84) were the required values specifying the two-tailed confidence interval (95%) and statistical power (80%) desired respectively. The odds for attending less than four ANC visits-preeclampsia/ eclampsia association was been set at 2 (universally acceptable) and r=3, is the ratio of controls to cases. Given these figures, the desired sample size of 352 participants (88 cases and 264 controls) was computed.

Study variables
The dependent variable was pre-eclampsia/eclampsia status measured as a binary categorical variable. The independent variables of interest were maternal age, education, residence, marital status, occupation, ethnicity, religion, maternal age, age at first marriage, age at first pregnancy, number of antenatal care visit, time/trimester of first ANC visit, gravidity, parity, child sex, anaemia in pregnancy, urinary tract infection (UTI) in pregnancy, alcohol use, tobacco use, family history of hypertension, family history of diabetes, personal history of diabetes, personal history of hypertension, and use of traditional treatment or medicine. All of these explanatory variables were more or less potential confounders and a determination of which one is a confounder was made during analysis. A pretested structured interview questionnaire was used to collect primary data from postnatal mothers and a data abstraction tool was also used to collect some secondary data, both available as Extended data 14 .

Ethical considerations
The

Minimization of biases
Before starting data collection, six research assistants (two for each of the three health facilities) were trained on the sampling technique for cases and controls and standard interview skills to reduce systemic error and interview bias. Research assistants were also trained to get information such as the number of antenatal care (ANC) visits, time/trimester of first ANC visit from the patient medical record using data abstraction tool in an attempt to reduce recall bias. Misclassification bias which could affect both cases and controls was minimized by including cases and controls determined by the obstetrician.

Statistical analysis
The filled questionnaires were first checked for completeness then followed by the entry of data collected from the field into Microsoft Excel spreadsheets, cleaned, formatted, coded and audited for quality and consistency using Epi-info software before exporting the dataset to Stata -version 14.0 computer programming software. Descriptive analysis was carried out, where, categorical variables were presented in percentages or proportions, whereas, continuous variables were presented in means and standard deviations. This was followed by bivariable mixed-effect logistic regression analysis and a multivariable mixed-effect logistic regression analysis using the significant variables (p ≤0.20) from the bivariable analysis. A modified Hosmer-Lemeshow goodness-of-fit test was carried out; after which, variables with p≤0.05 were considered as factors associated with preeclampsia/eclampsia.

Results and discussion
Results

Socio-demographic characteristics of the study participants.
The study participants were 352, of whom 88 were cases and 264 were controls. Among the cases, preeclampsia accounted for 83 (94.3%) and eclampsia 5 (5.7%  Table 2. Postnatal mothers from Nairobi County were at increased risk of preeclampsia/eclampsia compared to those from other counties (OR=3.0, 95% CI=0.7-13.1, p=0.152). The associations between socio-demographic factors and preeclampsia/eclampsia are summarized in Table 3.

The reproductive and obstetric factors associated with preeclampsia/eclampsia
We hypothesized that reproductive and obstetrics factors likely associated with preeclampsia/eclampsia were maternal age, age at first marriage, age at first pregnancy, number of ANC visits, time/trimester of first ANC visit, gravidity, parity, and child-Sex. Compared to mothers whose first antenatal care visit was in the first trimester of pregnancy, mothers whose first antenatal care visit was in the second and third trimesters were at reduced risk of preeclampsia/eclampsia (OR=0.8, 95% CI=0.5-1.4, p=0.473 and OR=0.5, 95% CI=0.2-1.4, p=0.191).
Mothers whose age at first marriage was <20 years old were 10% less likely to develop preeclampsia/eclampsia than those that were ≥20 years old (OR=0.9, 95% CI=0.5-1.6, p=0.659) whereas mothers whose age at first pregnancy was <20 years old were 0.7 times more likely to develop preeclampsia/eclampsia than those that were ≥20 years old (OR=0.7, 95% CI=0.4-1.3, p=0.199). The associations between reproductive and obstetric factors and preeclampsia/eclampsia are summarized in Table 4.

The clinical factors associated with preeclampsia/eclampsia
Anemia and urinary tract infections in pregnancy were hypothesized to relate significantly with the occurrence of preeclampsia and eclampsia. Compared to mothers with a normal number of pus cells (0-5 hpf) on admission for delivery, mothers with mild-moderate and severe UTI (6-10 hpf and >10 hpf) in pregnancy on admission were 1.7 times more likely to develop preeclampsia/eclampsia (OR=1.7, 95% CI=0.4-7.0, p=0.454). Mothers with mild-moderate and severe levels (7-10.9 g/dl and <7 g/dl) of anemia on admission for delivery were 0.9 times as likely to suffer from preeclampsia/eclampsia than those with normal hemoglobin level (≥11 g/dl); (OR=0.9, 95% CI=0.5-1.6, p=0.625). Associations  Table 5.

The behavioural and family history-related factors associated with preeclampsia/eclampsia
Further, the relationships between alcohol use, tobacco use, traditional treatment use, diabetes, and hypertension with preeclampsia/eclampsia were assessed. History of diabetes and hypertension were personal and family history.
Postnatal mothers who used tobacco were 3.9 times more likely to develop preeclampsia/eclampsia than those who did not use tobacco (OR=3.9, 95% CI=1.0-14.9, p=0.046). Postnatal mothers who used tobacco were at increased risk of suffering preeclampsia/eclampsia; however, the association was not statistically significant in that the 95% confidence included 1 despite a p-value of 0.046 (OR=3.9, 95% CI=1.0-14.9, p=0.046). On the other hand, mothers with a personal history of hypertension were 6.3 times more likely to suffer from preeclampsia/eclampsia than those without a personal history of hypertension (OR=6.3, 95% CI=2.7-14.8, p=0.001). In addition, there was a significant association between personal history of hypertension and development of preeclampsia/eclampsia (p<0.001). The associations between behavioural and family history-related factors and preeclampsia/eclampsia are presented in Table 6.

Multivariable mixed-effect logistic regression on risk factors for preeclampsia/eclampsia.
A multivariable mixed-effect logistic regression model was used to determine the effect of various socio-demographic, reproductive and obstetric, clinical, behavioral and family history-related factors. Only variables Table 5. Association between clinical factors and preeclampsia/eclampsia status.
Mothers with preeclampsia/eclampsia were 7.1 times more likely to have personal history of hypertension comparing to their counterparts (AOR=7.1, 95% CI=2.6-19.3, p=0.001). The associations between risk factors and preeclampsia/eclampsia are presented in Table 7.

Discussion
Women with preeclampsia/eclampsia were more likely to be 35-49 years of age comparing to their counterparts (AOR=5.9, 95% CI=1.1-33.3, p=0.042). This was concurrent with other studies that showed an increased risk in a similar age group 11,15,16 . A study conducted in Iran has linked the obstetric danger of advanced maternal age to aging-mediated vascular damage 2 . This is because as a woman gets advanced age, she is more likely to develop heart/blood vessel related problems chiefly due to the steady failure of compliance of said vessels that are mainly linked with the ageing of uterine blood vessels and arterial firmness. Thus, the reason why preeclampsia/eclampsia is more likely in older women is biologically plausible.
The odds of developing preeclampsia/eclampsia were 3.1 times higher in postnatal mothers with occupation as housewife comparing to mothers with other occupations (AOR=3.1, 95% CI=1.1-8.8, p=0.034). The association between occupation as housewife and preeclampsia/eclampsia was statistically significant. A similar finding was obtained from a study done in Nigeria 17 . This could be due to stress-related factors arising from low socio-economic status and low wealth index.
According to the results of our study, women who were nulliparous or primiparous were at increased risk of developing preeclampsia/eclampsia than those who were multiparous. This was concurrent with several studies that found the same findings 1,11,18 . Several hypotheses have linked nulliparity to a maternal immune maladaptation 2 . This is for the reason that nulliparity is due to early trophoblastic invasion and how the mother reacts to it. The breakdown or malfunction of the normal invasion of trophoblastic cells leads to maladaptation of the coiled arterioles, which are linked to the causation of preeclampsia 1 .
Mothers that attended less than four ANC visits were at increased risk of developing preeclampsia/eclampsia when compared with those that attended four or more ANC visits. This is comparable to a study conducted in Nigeria 17 . Longdistance to access health facility, limited knowledge about antenatal care (ANC) services coupled with low socio-economic status could possibly be responsible for postnatal mothers not meeting the WHO recommended four or more ANC visits; a situation which could predispose mothers to obstetric complications such as preeclampsia and eclampsia. In this study however, the association between a number of ANC visits and the outcome (preeclampsia/eclampsia) was not statistically significant in that the 95% confidence included 1 despite a p-value of 0.041 (AOR = 1.8; 95% CI: 1.0-3.3, p=0.041).
The study results found out that anemia and UTI in pregnancy were not significantly associated with preeclampsia/eclampsia; however, this finding contradicts studies conducted in Egypt, Sudan, and by WHO 19-21 .
Women with personal history of hypertension were more likely to suffer from preeclampsia/eclampsia compared with their counterparts (AOR=7.

Conclusions
This study indicated that a personal history of hypertension, older/advanced maternal age (35-49 years), occupation, and parity (nulliparous/primiparous) were factors significantly associated with preeclampsia/eclampsia. In this study, only 66 out of 352 participants (18.8%) had their urinary pus cells recorded on admission for delivery and whether UTI in pregnancy is actually not associated with preeclampsia/eclampsia

Results:
All your tables can be summarized in two tables only: Table 1 all comparing all variables between preeclampsia and controls Table 2 logistic regression with non -adjusted and adjusted values ○ Is the work clearly and accurately presented and does it cite the current literature? Yes

Is the study design appropriate and is the work technically sound? Yes
Are sufficient details of methods and analysis provided to allow replication by others? Partly

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Partly