This prospective cohort study was carried out from April to October 2021 at three public health facilities in three regions of Unguja, Zanzibar. One of these facilities is the Mnazi Mmoja hospital in the Urban region, which has an annual live birth delivery count of 13,291. Kivunge hospital is located in the North “A” region, serving 4000 live births annually, and Makunduchi hospital in the South region, with 1000 live births per year [32] . During the study period, these were the only public health facilities that offered Comprehensive Emergency Obstetric and New-born Care (CEmONC) in Unguja. Moreover, the Urban region recorded the highest proportion of women of childbearing age at 60.9%, with Unguja’s south region coming next at 54.0% [23] . The participants were enrolled at antenatal clinics (ANC) following their antenatal care visits in the selected health facilities in Unguja, Zanzibar.

Eligible participants were pregnant women at ≥20 weeks of gestation attending ANC visits at selected facilities. Inclusion criteria were willingness or intention to deliver within Unguja and their availability for follow-up up to 28 days postpartum/ during the study period. Systematic random sampling was used to select participants from the selected health facilities. Enrolment was undertaken after the women had given consent to participate in the study.

A sample size of 728 pregnant women was estimated to be required, using Kish’s formula (Kish, 1965), based on a 19.5% estimated prevalence of PE/E [33] , 95% confidence level, 4% margin of error, design effect of 1.5, 20% anticipated loss to follow-up and a 5% non-response rate. Allocation across sites was based on the average number of women booking at ANC per month. 488 from Mnazi Mmoja, 124 from Kivunge, and 116 from Makunduchi. Pregnant women were followed across three distinct phases. Phase one involved three antenatal follow-up visits at 20, 28, and 36 weeks of gestation. Phase two occurred within 24 hours after childbirth. Phase three included postnatal follow-up conducted at 72 hours, 14 days, and 28 days after childbirth.

In this study, a total of seven follow-up visits were carried out: three ANC visits at 20, 28, and 36 weeks of gestation; one visit within 24 hours of delivery; and three visits at 72 hours, 14 days, and 28 days postpartum. All data collected were entered into a prepared database.

To enhance follow-up feasibility, participants were given a follow-up card containing contact information, visit schedules, and the research team’s telephone numbers. Participants were reminded of the next ANC visits and were advised to report to the hospital in case of any complications during pregnancy. A reminder call was made to participants one week before and after their expected delivery date (EDD) based on the ultrasound scan, which worked well for most women. The research team, along with obstetricians, increased the timing of ANC and follow-up visits for women suspected of having preeclampsia based on the severity of their condition. The research team made sure that the women were contacted and followed up weekly, and they were advised to go to the hospital if they noticed any warning signs.

Maternal outcome measures were defined as adverse health effects experienced by pregnant women diagnosed with or without preeclampsia in Unguja. These outcomes included chronic hypertension, eclampsia, HELLP syndrome, renal failure, and maternal death. Data were collected from the ANC (during the recruitment) phase through a 28-day postpartum follow-up. Neonatal outcomes were assessed for new-borns and fetuses born to the same cohort of women (mothers with or without preeclampsia). These included preterm birth, low birth weight, neonatal intensive care unit (NICU) admissions, stillbirth, and neonatal death. These outcomes were monitored from the time of birth up to 28 days postnatal. All outcomes were prospectively recorded as part of the present study data collection process.

Six bilingual (Swahili and English) female and male nurse officers were trained as research assistants (RAs). The training lasted for 3 days, and they covered the procedure for data collection, ethical conduct, and research tools. The training was facilitated by the principal investigator (PI). A pilot study was conducted at Alrahma Private Hospital before the commencement of data collection. This health facility (Alrahma Private Hospital), was not selected for the study data collection. After making the needed changes and corrections to the questionnaire, Cronbach’s alpha was used to check how reliable the instrument was, and it scored 0.70, which is generally seen as acceptable. PI reviewed the data daily to ensure completeness, consistency, and accuracy. Regular team meetings addressed emerging issues, including blood pressure (B/P) readings, urine protein tests, and laboratory examinations such as creatinine and liver function tests. Standardized protocols were followed for clinical measurements. This included using the correct cuff size for B/P readings and ensuring proper patient positioning. Additionally, the research team discussed any concerns related to the mother and baby.

Later on, the group of pregnant women with preeclampsia was obtained due to the expert categorization where by preeclampsia was defined as a combination of having a high blood pressure (systolic reading of 140 mmHg or higher) and diastolic reading of 90 mmHg or higher, along with albumin levels exceeding 30 mg/L in the urine.

Data analysis was performed using SPSS software (version 26.0). Descriptive statistics (means, standard deviations, frequencies) were used to summarize the baseline characteristics. Chi-square or Fisher’s exact tests were used to determine the statistical association between a set of independent variables and PE/E in a bivariate analysis. Incidence and risk ratios were calculated to compare outcomes between the exposed (PE/E) and non-exposed groups, using a multivariable binomial regression model. A p-value of <0.05 was considered statistically significant, and 95% confidence intervals were reported where applicable.

Most women were aged 20 - 34 (77.6%), married (97.4%), and had secondary education (76.2%). More than half (59.3%) were housewives with a median monthly income of 16,000 Tsh (approximately $6), (IQR 7,000 - 56,500). Two-thirds (66.9%) resided in urban areas ( Table 1 ).

Of the new-borns, 51.5% were male. The mean birth weight was 3.1 kg (SD ±0.5), and 82.8% weighed between 2.5 - 3.9 kg. Their mean gestational age was 36.9 weeks (SD ±2.4), with a mean Apgar score at 1 minute of 8.3 (SD ±2.2) ( Table 2 ).

* = Missing due to 9 Abortion, 5 Loss of follow up.

Of the enrolled women, 728 were followed up through 28 days post-delivery ( Figure 1 ). Among these, 5 pregnant mothers were lost during the first phase. About 723 (99%) of pregnant women were followed until delivery, of whom 55 developed preeclampsia, and nine women had an abortion and could not be followed up, resulting in a cohort of 714 women followed up until 28 days post-delivery. None of the studied reproductive factors showed a statistically significant association with preeclampsia in either crude or adjusted models (p > 0.05 for all comparisons) ( Table 3 ).

Women with a history of chronic hypertension had 3.6 times higher odds of developing preeclampsia (aOR= 3.69, 95%CI: 1.01-13.54, p =0.049). Similarly, a previous history of preeclampsia (aOR 3.19, 95% CI: 1.04 - 9.72, p = 0.041) and a family history of hypertension (aOR = 1.92, 95% CI: 1.05 - 3.51, p = 0.033) were independently associated with a higher risk of preeclampsia ( Table 4 ).

Women with a previous history of preeclampsia or eclampsia were significantly more likely to develop preeclampsia in their subsequent pregnancy than their counterparts (aOR = 3.11, 95% CI: 1.11 - 8.71, p = 0.031) ( Table 5 ).

In adjusted analysis, high blood pressure (SBP_ ≥ 140 mmHg) at first ANC visit was strongly associated with developing PE/E, (aOR = 12.23, 95% CI: 1.02 - 146.77, p = 0.048) compared to low blood pressure of SBP < 140 mmHg. Proteinuria at any level was also a significant predictor: for preeclampsia (aOR = 81.65 (95% CI: 18.89 - 353.02) (1+); 50.62 (95% CI: 3.16 - 809.57) (2+), and 273.79 (95%CI: 20.40 - 3674.21) (3+), all p < 0.001). On the other hand, women who gave birth to an infant weighing 2.5 - 3.9 kg had significantly lower odds of developing preeclampsia (aOR = 0.35, 95% CI: 0.15 - 0.82, p = 0.016) as compared to their counterparts who gave birth to low-birth-weight infants ( Table 6 ).

Of the enrolled women, 714 were followed up until 28 days post-delivery. Among them, 345 (48.3%) experienced premature labour; of these, 45 (6.3%) had premature labour between 28 and 34 weeks of gestation, while 302 (42.3%) had premature labour during 35 to 36 weeks of pregnancy. Among the 714 babies born, 45 (6.3%) died within 28 days post-delivery. None of these maternal and neonatal adverse birth outcomes studied showed a statistically significant association with preeclampsia in either crude or adjusted models (p > 0.05 for all comparisons) ( Table 6 ).

In summary, individual and family history of hypertensive disorders, high BP at first ANC visit, and proteinuria were the important predictors for preeclampsia. None of the sociodemographic and reproductive variables showed a statistically significant association with preeclampsia.

Additionally, in the current study, changes in paternity were not statistically associated with the development of preeclampsia. This finding contrasts with a previous study conducted in Ethiopia, which reported a significant association between paternity change and the risk of preeclampsia [39] . The contrasts could have been caused by the inherited family pattern, which may play a role in the development of pre-eclampsia, as the placenta plays a significant role in the disease’s pathogenesis [40] . Moreover, the current findings contrast with those of another study carried out in Zanzibar, which suggested that women who changed partners had a 7.7 times greater risk of developing preeclampsia [19] .

This variation could be attributed to the fact that the present study was limited to three public hospitals in Unguja, whereas the other study encompassed both Unguja and Pemba islands. Moreover, the variance may arise from the different study designs and populations: our research involved the enrollment and follow-up of pregnant women from antenatal care (ANC) through to the postnatal period, whereas the other study concentrated on postnatal mothers only. On the other hand, our research indicated that the risk of developing preeclampsia was 1.2 times greater for women with previous pregnancies than for those without any prior pregnancies. This could be attributed to the disease’s aetiology, as multigravida women have an increased risk of developing preeclampsia, considering they may have experienced this condition before.

Similar findings were reported in other sub-Saharan African studies, where the multigravida women had 1.89 times higher risks for preeclampsia compared to the nullipara women [5] . However, this contrasts with a study from Ghana, where primigravida status was not associated with preeclampsia [25] . This variation may be attributed to the sociodemographic characteristics of study participants.

The present research demonstrated that the likelihood of developing preeclampsia was 3.7 times higher for women with a history of chronic hypertension and 3.9 times higher for those with a family history of hypertension. Similar to the studies conducted in China, Malaysia, Colombia, Pakistan, Ethiopia, Brazil, Sierra Leone, and Zanzibar, which reported that a history of chronic hypertension and a family history of hypertension were predictors of preeclampsia [6] [10] [19] [26] [41] - [45] . On the other hand, the current findings were inconsistent with the studies done in Bangladesh, Uganda, and Ghana, which revealed that a previous history of hypertension was not associated with preeclampsia [4] [25] [46] . This could be due to the geographical area of the study participants.

Furthermore, women who had a prior history of preeclampsia or eclampsia were 3.2 times more likely to develop preeclampsia in their following pregnancy compared to their counterparts. The findings were in line with the findings reported in Colombia and Bangladesh women with a previous history of preeclampsia had a 4 to 5 times higher risk of developing preeclampsia [47] [48] . This could be due to the genetic factors that might contribute to the physiological predisposition of preeclampsia.

In a recent study, the present findings demonstrate that women with preeclampsia had higher odds of having low-birth-weight infants compared to their counterparts. The finding were in line with other studies done in Malaysia, Cameroon, Nigeria, Uganda, Zimbabwe, Southern and Northeast Ethiopia, China and Thailand, reported that babies born to preeclamptic women were at a higher risk of perinatal death, low birth weight, birth asphyxia, neonatal intensive care unit (NICU) admission and preterm birth than babies born to normotensive women [1] [30] [42] [49] - [53] . On the other hand, premature labour and neonatal deaths were not significantly associated with preeclampsia in the present study, as reported in the previous studies. The divergences may result from methodological factors of the study utilized, such as the study setting, sample size, and study design. The current finding suggests a statistical correlation for women with high blood pressure (SBP_≥ 140 mmHg) during their first ANC visit, and had 12 times higher risk of preeclampsia than those with SBP < 140 mmHg. However, the results contradict those of a prospective study conducted in Uganda, which identified diastolic hypertension ≥ 90 mmHg as significant to preeclampsia [46] .