Maternal physical and social characteristics that influence the occurrence of preeclampsia/eclampsia and hemorrhage in Eastern Region of Ghana. A prospective cohort study

Abstract

Background:

Preeclampsia/eclampsia places grave barriers to the successful reduction of maternal mortality and morbidity. These barriers have significant adverse outcomes for most women especially in sub-Saharan Africa, with Ghana contributing 88.7% of preeclampsia cases in Africa.

Aim:

The aim of this study is to evaluate the maternal, physical and social characteristics that influence the occurrence of preeclampsia/eclampsia in Eastern Region of Ghana.

Methods:

A prospective cohort study was conducted among pregnant women with gestational age > 28 weeks seeking antenatal care in seven hospitals in the Eastern Region of Ghana from October 2022 to March 2023. Using a simple random sampling technique, 445 patients were recruited at the antenatal care clinics using the antenatal care register as the reference point, and their delivery outcomes were evaluated after 13 weeks of follow-up.

Results:

The study shows that pregnant women with poor personal hygiene are 10 times the odds (aOR: 10.169, p < 0.001, 95% CI: 4.896–21.121) of developing preeclampsia/eclampsia compared to those with good personal hygiene and about five times the risk of haemorrhage (RRR: 5.12, p-value = 0.014, 95% CI: 1.393–18.815) compared to those who are normal. Those who opted for Jadelle contraceptives had 3.4 times the odds (aOR: 3.419, p = 0.028, 95% CI: 1.143–10.227) of developing preeclampsia/eclampsia compared to those who used Depo-Provera contraceptives. Women who depended on borehole as their source of drinking water were 6.95 times the odds (aOR: 6.951, p = 0.013, 95% CI: 1.511–31.981) of developing preeclampsia/eclampsia compared to those who used sachet water. Furthermore, there was statistically significant association between comorbidity conditions and haemorrhage.

Conclusion:

The study showed that pregnant women with poor personal hygiene were 10 times more likely to develop preeclampsia/eclampsia compared to those with good personal hygiene. The use of some family planning methods may influence the development of preeclampsia/eclampsia. Women need to seek medical advice on contraceptives before and during pregnancy before choosing contraceptives methods. Also, effective community-based health planning and services concept should be implemented to improve access to family planning services in the rural communities.

Introduction

Preliminary data available show that preeclampsia/eclampsia (PE-E) and haemorrhage are the main predisposing factors contributing to maternal morbidity and mortality. ¹ About 1%–5% of all pregnant women develop PE-E, ² leading to perinatal morbidity and mortality. ³ Preeclampsia develops over a period of time during pregnancy, ⁴ but women with a history of preeclampsia are more at risk in subsequent pregnancies. ⁵ This is why early screening and diagnosis during a woman’s first pregnancy is considered essential to preventing preeclampsia. ⁶ The traditional system of diagnosing preeclampsia has always been the reliance on blood pressure greater than or equal to 140/90 with proteinuria. ⁷ However, recent suggestions by professional bodies indicate that in a multisystem involvement, diagnosing PE-E can be without proteinuria but with end-organ impairment such as thrombocytopenia, impaired liver function, new renal insufficiency, pulmonary oedema or new-onset cerebral or visual disturbances. ⁸ Additionally, the contribution of angiogenic factors have been cited as significant biomarkers in preeclampsia. ⁹ The instability of the angiogenic factors is vital in determining the pathogenesis of the maternal health condition leading to preeclampsia. ¹⁰ They exhibit different signs and symptoms such as headache, generalized oedema and epigastric pain with protein in urine. ⁶ More so, eclampsia presents vomiting, blurred vision and convulsion. ¹¹ There are grave consequences of severe PE-E when it comes to maternal health. ¹² These diseases contribute about 50,000–100,000 maternal deaths globally every year, ¹³ and is much more prevalent in developing countries.^14,15 Meanwhile, the United Nations sustainable development goal (SDG) 3.1 is geared towards fighting the maternal mortality battle, especially in Africa. ¹⁶ Notwithstanding, efforts to achieve zero maternal mortality in Africa have been challenging, ¹⁷ and earlier research have documented the contribution of maternal characteristics as barriers to fostering SDG three. ¹⁸ For instance, approximately 5%–10% of maternal mortalities as a result of PE-E are due to foetal growth restriction, ¹⁷ placental abruption, ¹⁹ history of abortion/miscarriage, early pregnancy ²⁰ and gestational age. ²¹ Other characteristics such as personal hygiene, ²² history of hypertensive disorders, ³ history and choice of contraceptives^23,24 and homocysteine metabolism disorder, ²⁵ are also well-documented linkages to the development of PE-E. This affects about 2%–8% of all pregnancies worldwide ²⁶ and increases the rate of dying among pregnant women in Africa. ²⁷ , ²⁸ For instance, the chances of a pregnant woman dying from PE-E in the developing world increases by 40 folds. ¹⁷ That is, the majority (66%) of the deaths attributable to PE-E occur in sub-Saharan Africa, ¹⁵ with Ghana reporting a PE-E rate of 88% and 11%, respectively, ²⁴ and 76.9% and 12.4%, respectively, in Ethiopia. ²⁸ The trend of maternal deaths in the Eastern region is as follows: 85 deaths in 2022, 70 in 2023 and 68 as of November 2024. Out of these deaths, post-partum haemorrhage and antepartum haemorrhage contributed 30.5% in 2022, 33.9% in 2023 and 32.3% as of November 2024. Hypertensive disease (PE-E) contributed 17.5% in 2022, 23.2% in 2023 and 30.9% as of November 2024. Although the actual cause of PE-E is unknown, and measurement standards for detecting early risk of eclampsia are unstandardised, ²⁹ and there is enough evidence to show that maternal, physical and social factors play a major role in the occurrence of PE-E. ³⁰ However, data on the exact maternal factors facilitating the occurrence of PE-E are limited, ¹¹ with few published literature being skewed towards investigating the clinical causes of PE-E. ²⁶ Based on this backdrop, this study aims to evaluate the maternal, physical and social characteristics that influence the occurrence of PE-E in Eastern Region of Ghana. We hypothesize that 35% of PE-E are influenced by maternal, physical and social characteristics in Eastern Region of Ghana.

Methods

Study location

The Eastern Region is one of the 16 administrative regions of Ghana and shares common boundaries with the Greater Accra, Central, Ashanti, Brong Ahafo, Oti and Volta Regions of the country. The population of the region is 2.9 million. ¹⁹ With a growth rate of 2.1%, the population is estimated to increase to 4.5 million by 2040. Presently, 49.2% of the population is males, while females represent 50.8%. The age structure of the region indicates that the proportion of the population aged 0–14 (under 15 years) is 38.4%, while those aged 15–64 and 65+ are 55.9% and 5.7%, respectively. The region occupies a total land mass of approximately 19,323 km², which makes it the sixth largest region of the country in terms of land size. It has a population density of 136.3 people per km² and is 43.4% urban with an annual urban growth rate of 3.7%. ³¹ Regarding the economy, the labour force participation rate for the population aged 15–64 is almost 74.2%. The study was conducted in New Juaben South Municipality at the Eastern Regional Hospital, Koforidua, Nsawam Adoagyiri Municipality (Nsawam Government Hospital), Birim Central Municipality (Oda Government Hospital), Akwapim North District (Tetteh Quarshie Memorial Hospital at Mampong), Suhum Municipality (Suhum Government Hospital), Denkyembour District (St. Dominic Hospital) and Asamankese Government Hospital at West Akim Municipality. ³² The region has 17 hospitals out of which 5 belong to the Christian Health Association of Ghana. All these facilities, in addition to all health centres and Community-based Health Planning and Services (CHPSs) centres provide prenatal care.

Study design

The study was a prospective cohort study involving pregnant women at gestational age ⩾28 weeks seeking antenatal care (ANC) in seven major hospitals of the Eastern Region of Ghana. Preeclampsia is rarely diagnosed before 20 weeks, and most cases develop between 28 and 37 weeks of gestation. Starting the study at 28 weeks focuses data collection on the period of highest clinical relevance, when risk becomes more apparent and symptoms may begin to emerge. The ANC registers were reviewed to identify qualified attendants (gestational age ⩾ 28 and should not have developed pre-eclampsia at the time of taking part in the study) for ANC services. The study recruited qualified pregnant women during ANC visiting days to the ANC clinic. Participants were first orally invited to participate in the study and were later asked to sign a consent form before the commencement of the interview. This was done by research assistants who were mostly midwives, dieticians and public health officers, until the sample size of 445 was reached. The ANC register was used as the reference population, and study participants were selected from the ANC register. Researchers evaluated the delivery outcomes of participants in terms of PE-E and haemorrhage after 13 weeks of follow-up. The researchers made a minimum of three contacts to participants during the follow-up period. Participants who were recruited but failed to attend subsequent ANC clinic visits were contacted through a telephone call and self-reported information on personal hygiene was obtained. The delivery outcomes were extracted from the medical records of the clients and some through telephone calls made by the midwives. We examined their personal hygiene (oral hygiene, fingernails, body odour and stench from clothes) status, history of abortion, history of miscarriage, early pregnancy, previous infection with sexually transmitted infections (STIs), birth spacing, use of family planning methods, type of family planning method used, use of condom and presence of comorbidity. The sample size was calculated using Cochran method. ³³ This was done based on the prevalence of preeclampsia in the region. N = Z² * P (1 − P)/d2, where N = sample size, Z = score for 95% confidence interval which is 1.96, P = regional prevalence of preeclampsia was 32.9% q = 1 − p (proportion of people without preeclampsia), d = margin of error set at 5%. Using a loss to follow-up rate of 30%, ³⁴ and a non-response rate of 10%, ³⁵ the total sample size was estimated.

$$N={\left(1.96\right)}^2\times 0.329\times \left(1-0.329\right)/{\left(0.05\right)}^2$$

$$N=\left(3.8416\times 0.329\times 0.671\right)/0.0025=339.2$$

Plus 30% loss to follow-up and 10% non-response rate increases the sample size by 135.6. Therefore, 339.2 + 135.6 = 474.8 which is approximately = 475.

These numbers were proportionately assigned to all the seven participating hospitals within the Eastern Region of Ghana. The loss of follow-up rate was 6.3% leaving only 445 respondents to work with (Table 1).

Table 1.

Participating districts and their estimated sample sizes.

Municipality/district	Expected pregnancy	Estimation of sample size	Sample size
Birim Central	3632	52.034	52
New Juaben South	6275	89.899	90
West Akim	5454	78.137	78
Akwapim North	4891	70.072	70
Nsawam Adoagyiri	4389	62.879	63
Suhum	4523	64.799	65
Denkyembour	3991	57.177	57
Total	33,155	475	475

The sample size was proportionately allocated to the seven hospitals within the Eastern Region of Ghana as tabulated above.

Conceptual framework of the study

The conceptual framework of the study included the most basic factors that contribute to the occurrence of PE-E and haemorrhage which are maternal factors such as parity, gravidity, anaemia in pregnancy, history of abortion/miscarriage, early pregnancy, antenatal status, gestational age and birth spacing are spacing. Physical factors also have the propensity to influence the occurrence of PE-E and haemorrhage. Most of these factors may manifest in the form of aerosolized particles that can easily enter the nasopharyngeal regions by way of air, through water, in the form of perfumes and deodorants. Socio-economic factors can tremendously trigger the development of PE-E and haemorrhage. If not addressed, these factors are likely to destabilise the pregnant woman’s mind; thus, inducing emotional and psychological stress which may trigger the development of complications.

Study variables

The demographic variables in this study were age, ethnicity, education level, marital status and religion. Other maternal variables were body mass index (BMI), gestational age, parity, mid-upper arm circumference, history of abortion, history of miscarriage, previous pregnancy, gravidity, birth spacing, comorbidity, work-related stressful, source of water, means of transport, previously diagnosis of STIs, seasonal variation, type of family planning method, use of condoms and personal hygiene. Poor personal hygiene was assessed as the presence of one or more of the following: body odour, wearing dirty and smelly clothes, smelly mouth and vaginal odour.

Ethical clearance and consent

The study protocol was reviewed and study consent granted by the Ghana Health Service Ethics Review Committee with approval number (GHS-ERC): GHS-ERC: (007/05/21). The written informed consent form for this study was read and interpreted in local dialects of participants without formal education before the data were collected. With regards participants who were under the age of 18, their legally authorized representatives identified by minor participants also granted written informed consent before the start of the study. This was done after the written informed consent form and study objectives were read and explained to them in both English and their local dialect. All study participants granted written informed consent before the start of the study.

Recruitment

The principal investigator led the research assistants to the various hospitals’ ANC clinics. He introduced the team to the in-charges of the ANC clinics and explained the purpose of the study to them. The ANC registers were reviewed to identify the qualified attendants for ANC services. A respondent was excluded from the study if the outcome of interest had already occurred and if the respondent had a gestational age < 28 weeks. All these attendants formed the sampling frame. Thereafter, using the ANC register as the reference point, research assistants who were mostly midwives, dieticians and public health officers, used a simple random process to recruit 475 qualified pregnant women (⩾28 weeks gestation) to participate in the study. The study ran for 6 months in total. However, the pregnant women were monitored for 14 weeks, starting at 28 weeks of gestation and ending at birth.

Pilot test and study implementation

We piloted (pretested) the questionnaire among pregnant women at Jubilee Hospital in the Birim Central Municipality of Ghana. This hospital was not part of the study area. Feedback on the piloted questionnaire was discussed for updates and reviews to be made before its use. This was done to validate and fine-tune the data collection tool to ensure reliability of the tool before its use. Results for the pilot test are not presented here because they were for the purpose for instrument modification and not for the preliminary results. The data collection tools were modified to remove sensitive questions and rearrange questions to start with less sensitive questions to more sensitive questions. The questionnaire was tested among ANC mothers at the community hospital. About 60% of the ANC mothers present participated in the pilot study. The study was piloted from 19 May to 29 June 2022, and changes to data collection instruments and field notes were resubmitted to the GHS-ERC for further review before study implementation. The final study protocol was implemented in 2023.

Follow-ups

Research assistants, including midwives, dieticians and public health officers, performed follow-ups at each ANC visit for 13 weeks. Among other things, the follow-ups evaluated the pregnant women’s routine ANC services, such as fetal presentation, HB levels, oedema and urine protein checks. The rate of loss to follow-up was 6.3%. During the follow-up period, participants received the same normal drugs as other pregnant women and were evaluated on the ANC services they received. Participants received no additional information other than heightened awareness of the independent variables and post-delivery outcomes.

Sampling

A lottery method probability sampling process was used. It involved a simple random selection of 7 government hospitals out of 17 in the Eastern Region. The names of all hospitals in the Eastern Region were written on pieces of paper and then mixed evenly and 7 selected randomly out of the 17. In the participating hospitals, a numbered list of ANC attendants with gestational ages ⩾ 28 weeks from each hospital was obtained from the ANC registers. The numbered list was captured on pieces of papers and dropped into a container and properly mixed manually. Then, chits were randomly picked out of the container until the sample size was reached. The study also followed the Strenthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines to improve the transparency and quality of reporting.

Data analysis

We entered the data into EPI Data software (EpiData Association) and transported it into IBM (International Business Machines Corporation) SPSS version 20 for editing and re-categorization. Data were then exported into STATA for the final analysis. Descriptive statistics of the independent variables were generated in the form of frequencies and percentages in tabular form. Pearson’s Chi-square was conducted to determine the association between the outcome variables and the independent variables. To determine the strength of association between the outcome variables and the independent variables, logistic regression was computed for each independent variable against the outcomes. Afterwards, we adjusted for all other variables that might cause the associations generated between each individual independent variable and the outcomes. A multinomial logistic regression was finally conducted to determine the associations between the independent variables and outcomes. To determine multicollinearity and interaction between variables, we performed correlation matrix to reduce dimensionality and computed the variance inflation factor (VIF) for each variable with a mean VIF of 1.43. The results were presented in the form of frequencies, percentages, odds ratios and relative risk ratios at 95% CI with a p-value < 0.05.

Results

Demographic characteristics of respondents

The regional prevalence of preeclampsia was 32.9%. Majority of the participants were between ages 20 and 29 years, 211 (47.5%) followed by those who were between 30 and 39 years, 199 (44.8%), and 40+ age group were the least with 15 (3.4%). For ethnicity, Akuapem ethnic group dominated the respondents with 101 (22.7%). Respondents who attained up to secondary level education formed the majority 195 (43.8%) with those without any formal education being 21 (4.7%). A little over half of the respondents were married 304 (68.3%) with 2 (0.5%) being separated/divorce. Eight in 10 of the respondents were Christians 392 (88.1%), Muslims were 52 (11.7%) and Traditionalist accounted for 1 (0.2%; Table 2). Less than half of the respondents ever had an abortion 114 (25.7%), with 94 (21.1%) ever experiencing a miscarriage and 20 (4.5%) ever having an early pregnancy. Over half of the respondents 313 (70.3%) had greater than or equal to 2 years birth spacing, 40 (8.9%) had comorbidity, 50 (11.2%) had poor personal hygiene, 76 (17.2%) had stressful work and 385 (86.5%) used sachet water. Additionally, 107 (24.0%) of respondents walk on foot, and 11 (2.4%) ever had gonorrhoea. Pregnant women who used the different types of family planning methods were 84 (18.9%) for Depo-Provera, 66 (14.8%) for Jadelle and 23 (5.2%) for Implanon. Those who use other methods, including hormonal contraception (HC), were the majority with 257 (57.8%; Table 2).

Table 2.

Demographic characteristics of respondents.

Variable	Frequency	Percent	p-Value
Age
10–19	19	4.3	0.56
20–29	211	47.5
30–39	199	44.8
40+	15	3.4
Ethnicity
Ewe	63	14.2	0.09
Asante	49	11.0
Akyem	78	17.5
Akuapem	101	22.7
Krobo	28	6.3
Others	126	28.3
Educational level
Primary	133	29.9	0.06
Secondary/O’level	195	43.8
Tertiary	96	21.6
No formal education	21	4.7
Marital status
Married	304	68.3	0.38
Unmarried	139	31.2
Separated/divorced	2	0.5
Religion
Christian	392	88.1	0.05
Muslim	52	11.7
Traditionalist	1	0.2
Abortion
No	330	74.3	0.95
Yes	114	25.7
Miscarriage
No	351	78.9	0.55
Yes	94	21.1
Early pregnancy
No	424	95.3	0.92
Yes	20	4.5
Birth spacing
<2 years	62	13.9	0.36
⩾2 years	313	70.3
N/A	70	15.7
Comorbidity
Yes	40	8.9	0.65
No	405	91.0
Personal hygiene
Good	395	88.8	<0.01
Poor	50	11.2
Worksite
Stressful	76	17.2	0.09
Very stressful	89	20.1
Less stressful	278	62.8
Which season
Dry season	21	4.7	0.02
Rainy season	35	7.9
N/A	389	87.4
Source of drinking water
Sachet	385	86.5	0.05
Tap	24	5.4
Bore hole	10	2.3
Well	7	1.6
Others	19	4.3
Transport
Foot	107	24.0	0.29
Vehicle	337	75.7
Ever had STIs
Gonorrhoea	11	2.5	<0.01
Chlamydia	7	1.6
Genital warts	9	2.0
Genital herpes	11	2.5
Syphilis	3	0.7
HIV	2	0.5
None	402	90.3
Type of FP method
Depo-Provera	84	18.9	0.03
Jadelle	66	14.8
Implanon	23	5.2
Sayana press	1	0.2
Microgynon	12	2.7
Microlut	2	0.5
Other (pills, IUD and injectables, HC)	257	57.8
Use of condom
No	412	92.6	0.24
Yes	33	7.4

Logistic regression depicting the association between selected maternal, social and physical characteristics and the occurrence of PE-E

Table 3 describes the regression analysis output of study participants and their association with the development of PE-E. The results show that poor personal hygiene, family planning uptake of Jadelle, Implanon, use of condom and water from borehole are statistically significant. The logistic regression output shows that pregnant women with poor personal hygiene are ten times (AOR 10.17, CI: 4.89-21.12) more likely to develop PE-E compared to those with good personal hygiene. Pregnant women who indicated that their main source of drinking water was from the borehole had 4.366 times the odds of PE-E compared to those who used sachet water. For respondents on family planning uptake, it was observed that those who opted for Jadelle were three times more likely to develop PE-E compared to those who chose Depo-Provera. Additionally, those who opted for Implanon were about seven times (6.933) the odds of developing PE-E compared to those who used Depo-Provera and those who used condom (2.289). After adjusting for age, gestational age, BMI, educational level and marital status, the odds of those with poor personal hygiene increased to 10.169 times, those who used Jadelle increased to 3.419 times, and those on Implanon increased to 7.761 times. Those who usually use condoms increased to 3.31 times, and those who depended on borehole as their source of water increased to 6.951 times the odds of developing PE-E compared to those who used sachet water, respectively (Table 3).

Table 3.

Logistic regression depicting the association between selected maternal, social and physical characteristics and the occurrence of PE-E.

Preeclampsia/eclampsia	COR	p-Value	95% CI	aOR	p-Value	95% CI
Personal hygiene
Good	Reference
Poor	7.56	0.01	3.99–14.31	10.17	0.01	4.89–21.12
Type of family planning methods
Depo	Reference
Jadelle	2.89	0.05	1.02–8.17	3.42	0.03	1.14–10.23
Implanon	6.93	0.01	2.10–22.88	7.76	0.01	2.17–27.74
Source of water
Sachet	Reference
Tap	1.72	0.29	0.62–4.82	2.81	0.07	0.92–8.55
Borehole	4.37	0.03	1.19–16.00	6.95	0.01	1.51–31.98
Well	2.62	0.26	0.49–13.86	1.99	0.44	0.35–11.48
Others	2.34	0.12	0.81–6.77	2.39	0.13	0.78–7.32
Use of condom
No	Reference
Yes	2.29	0.05	1.01–5.17	3.31	0.01	1.37–7.99

Variables adjusted for age, gestational age, BMI, educational level and marital status.

Multinomial logistic regression showing association between demographic, maternal and physical factors and the occurrence of PE-E and haemorrhage

The multinomial logistic regression on Table 4 shows statistical significance for independent variables including educational level, BMI, ever had STIs, haemoglobin levels (HBs) levels, source of water, ever use family planning, type of family planning method and personal hygiene in respect of PE-E. For haemorrhage, the significant independent variables are HB levels, comorbidity, ever use family planning and personal hygiene. The results show that educational level, source of drinking water, BMI, type of family planning method and personal hygiene increase the risk of PE-E by about 1.5, 1.4, 1.5, 1.3 and 9, respectively, among pregnant women with PE-E compared to those who were normal. However, the study further shows HB levels, those who ever had STIs and those who ever used modern family planning methods are 32.7%, 18.8% and 73.3% less likely to develop PE-E compared to those who were normal.

Table 4.

Multinomial logistic regression showing association between demographic, maternal and physical factors and the occurrence of PE-E and haemorrhage.

Outcome of delivery	RRR	p-Value	95% CI	RRR	p-Value	95% CI
Preeclampsia/eclampsia	Reference = normal			Haemorrhage	Reference = normal
Age	1.56	0.15	0.86–2.84	1.85	0.18	0.75–4.57
Gestational age	1.06	0.79	0.68–1.65	1.12	0.76	0.55–2.31
HB	0.67	0.02	0.49–0.93	0.57	0.04	0.33–0.99
BMI	1.49	0.04	1.02–2.20	0.79	0.44	0.44–1.43
Source of water	1.42	0.01	1.09–1.85	0.98	0.96	0.54–1.80
Educational level	1.46	0.05	1.01–2.11	0.96	0.90	0.52–1.79
Marital status	1.09	0.83	0.52–2.26	0.69	0.55	0.21–2.29
Abortion	0.75	0.44	0.36–1.55	1.36	0.63	0.39–4.76
Miscarriage	0.87	0.73	0.40–1.89	3.73	0.11	0.74–18.91
ANC visits	1.02	0.94	0.59–1.73	0.75	0.44	0.37–1.55
Gravida	0.72	0.06	0.51–1.01	0.89	0.69	0.55–1.54
Ever had STIs	0.81	0.02	0.68–0.97	1.04	0.85	0.69–1.57
Birth spacing	0.77	0.59	0.29–2.02	1.81	0.43	0.42–7.72
Comorbidity	0.68	0.48	0.23–2.00	0.24	0.03	0.06–0.88
Personal hygiene	9.09	<0.01	4.26–19.42	5.12	0.01	1.39–18.82
Ever used FP	0.27	0.02	0.09–0.82	0.15	0.02	0.03–0.74
Type of family planning	1.29	0.03	1.03–1.63	1.22	0.19	0.91–1.65

In terms of haemorrhage, the study shows that HB levels, comorbidity and those who ever used modern family planning methods are less likely to develop haemorrhage by 43.1%, 76.3% and 84.6%, respectively. On the contrary, personal hygiene increases the risk of haemorrhage by 5.12 times compared to those who are normal (Table 4).

Discussion

The study discusses the maternal, physical and social characteristics on the occurrence of PE-E and haemorrhage. The findings show that there is a relationship between maternal, physical and social characteristics and the occurrence of PE-E and haemorrhage. Family planning is very important in controlling family size and unwanted pregnancies. ³⁶ In this study, the analysis showed that 18.9% of the pregnant women used Depo-Provera prior to getting pregnant, and 14.8% Jadelle. This is in contrast with the findings of a study conducted in Nigeria where the findings showed that 21.4% of the participants used Depo-Provera. ¹⁷ Additionally, pregnant women who used Implanon were 5.2% as against 2.3% reported by Rana et al. ¹² The variations observed here may be attributable to the study locations and the coverage of the studies. In Ghana, however, study findings showed that 43.3% of pregnant women used Depo-Provera before getting pregnant as reported by Boggess et al. ²⁴ The reasons for the difference observed here could be due to the type of study, study locations and coverage. Asare et al.’s ²⁶ study was a randomized case–control study conducted at Comboni Mission Hospital, Sogakope, in the South Tongu District of the Volta Region of Ghana as against a prospective cohort study conducted in seven hospitals in the Eastern Region of Ghana. Surely, this current study is more representative of the Eastern Region compared to a facility-based study in one hospital in the Volta Region. Also, findings from this study show that Depo-Provera is the most accepted contraceptive method used by the pregnant women in the Eastern Region. Twenty-five percent (25.7%) of the pregnant women who participated in this study reported ever having an abortion, compared to 13% reported in Nigeria. ¹⁷ Those who had miscarriage were 21.1% as against 8.7% reported by Bujang. ³⁷

From the regression analysis, the study findings show that poor personal hygiene, family planning uptake of Jadelle, Implanon, use of condom and water from borehole were significantly associated with PE-E. The logistic regression output shows that pregnant women with poor personal hygiene were more than seven (7.6) times the odds of developing PE-E compared to those with good personal hygiene. This agrees with the findings in previous studies where pregnant women with history of poor oral hygiene were about four (3.7) times the odds of developing PE-E compared to those with good oral hygiene. ²² After adjusting for age, gestational age, BMI, educational level and marital status, the odds of those with poor personal hygiene increased to 10 times the risk of developing PE-E compared to those with good personal hygiene. Meanwhile, maternal personal hygiene is expected to be maintained and is known to reduce the risk of infections during pregnancy. ³⁸ Poor personal and oral hygiene is mostly associated with limited resources, personal neglect and lack of understanding on the importance of personal hygiene. ⁹ Policymakers should formulate interventions leading to policy decisions towards improved personal hygiene during ANC visits. Additionally, clinicians and midwives should council ANC clients on proper hygienic practices during visits to the ANC clinics. The SDGs 3.1, 5 and 6 are all linked, and there is a relationship between safe drinking water, sanitation and maternal health. ¹⁶ Provision of available and safe drinking water, coupled with good sanitary conditions and the usage of acetylsalicylic acid about 13 weeks could prevent the recurrence of pre-eclampsia in pregnancies for women and lead to healthier pregnancy, safe delivery and improved maternal health for all. ¹

Results from the multinomial logistic regression reveal that educational level, source of drinking water, BMI, type of family planning method and personal hygiene increase the risk of PE-E by about 1.5, 1.4, 1.5, 1.3 and 9, respectively, among pregnant women with PE-E compared to those who were normal. On the other hand, HB levels, those who ever had STIs, and those who ever used modern family planning methods are 32.7%, 18.8% and 73.3% less likely to develop PE-E compared to those who were normal. This means, poor personal hygiene is likely to increase the risk of infection which may contribute to an increased risk of developing PE-E as reported in other studies. ³⁹ The biological mechanism connecting poor personal hygiene and PE-E involves increased systemic inflammation, nutritional deficiency, altered gut microbiome, oxidative stress, maternal immune activation and disrupted fetal development. ⁴⁰ These are complex mechanisms that explain how poor personal hygiene can increase the odds of developing PE-E. Poor personal hygiene leads to infections such as parasitic and bacterial infections. The variation in odds of occurrence may be due to the difference in study area or part of the body where this study assessed.

Additionally, HB levels, comorbidity and those who ever used modern family planning methods are less likely to develop haemorrhage by 95.6%, 96.9%, and 98.1%, respectively. Also, personal hygiene increases the risk of haemorrhage by 5.12 times compared to those who are normal. This is because severe infections can trigger disseminated intravascular coagulation, a life-threatening condition where the mother’s body’s clotting system is disrupted. Again, family planning reduces menstrual blood flow and the general duration of menstruation. This can reduce the risk of anaemia and its related complications. Haemorrhage can develop in different ways; thus, either antepartum, intrapartum or postpartum. This is why the continuous medical monitoring of pregnant women and newly delivered mothers is an essential part of methods to reduce maternal mortality and post-partum haemorrhage. ¹²

For family planning uptake, there were observed variations among pregnant women who opted for Jadelle and those who used Depo-Provera. For instance, those who used Jadelle were three times the odds of developing PE-E compared to those who chose Depo-Provera. After adjusting for age, gestational age, BMI, educational level and marital status, those who used Jadelle remained about three (3.4) times the odds of PE-E compared to those who used Depo-Provera. Jadelle which contains levonorgestrel as the main active ingredient can interact with other drugs such as phenytoin, oxcarbazepine, rifampicin and many others. ²⁵ However, evidence suggests that residues of the hormone, levonorgestrel may interact with certain medications of a previous client which may lead to higher risk of PE-E. ²³ Additionally, after adjusting for age, gestational age, BMI, educational level and marital status, those who used Implanon increased to about eight (7.8) times the odds of developing PE-E compared to those who used Depo-Provera. Meanwhile, the use of Implanon may cause inconvenience in many users and a few may get infections. ⁴¹ Research conducted in Indonesia disclosed that the use of HC was associated with the development of preeclampsia. ⁴² This is consistent with the findings in this study. The few adverse events following the use of Implanon may be responsible for the associations with PE-E observed in this study.

In a multivariate logistic regression of factors associated with preeclampsia, study findings demonstrated that the general use of contraception was significantly associated with the development of preeclampsia and that pregnant women who used contraceptives were 26 times more likely to develop preeclampsia. ²⁴ This is clearly in consonance with the findings in this study where the use of Jadelle, condom and Implanon were all associated with the development of PE-E. However, this study is at variance with the findings by Roberts et al. ⁴³ and Firoz et al. ⁴¹ that the use of modern contraceptive methods decreased the risk of preeclampsia. While Spagnoletti et al. ⁴⁴ was a case–control study conducted in the United Kingdom, this current study is a prospective cohort study conducted in seven hospitals in the Eastern Region of Ghana. Again, Spagnoletti et al.’s study was also a case–control study conducted in the United Kingdom among pregnant women. The genetic make-up and the type of food eaten in these countries differ significantly and may be contributing to how women in these countries will respond to the different family planning commodities used. ⁴⁵ Study findings have shown that the use of some family planning contraceptives have adverse events which include abnormal vaginal bleeding which may be associated with the increase in haemorrhage reported in this study. ¹⁹

Pregnant women who depended on borehole as their source of drinking water were about seven (6.9) times the odds of developing PE-E compared to those who used sachet water. This is concurrent with the findings in other case–control studies conducted in Bangladesh where in both crude and adjusted odds ratios, there were increased risks (almost 2 times the odds) of developing PE-E among pregnant women who drunk borehole water. ³⁶ The difference in odds of association may be due to the method of study conducted and the location of the studies. This is further explained by evidence provided by other studies, indicating the link between drinking water in Bangladesh and other hypertensive disorders. ⁴⁴ The results also measured the risk of haemorrhage, and the significant independent variables were comorbidity, and personal hygiene. The results show that for every 1 unit increase in the educational level among cases of PE-E, there is a corresponding 0.4 times increased risk of developing PE-E compared to those who were normal. Conversely, data available suggest that educational attainment is not a significant predictor of preeclampsia and eclampsia. ²⁸ The difference observed is that the study by Mareg et al. ³⁰ and the current study may be due to differences in methodology, considering that both studies were conducted among women attending ANC services. However, findings from our study are in conformity with studies conducted by Parker et al. ⁴⁵ where poorly educated women were more likely to develop PE-E. The difference observed here is that Parker et al. was a national population-based study as compared to the current study. This means that educated women in this current study may be well informed to seek medical care more often and to identify danger signs in pregnancy earlier through antenatal visits as compared to home-based interviewed women who may include large groups of uneducated women. Mareg et al. ³⁰ study also agrees with the findings that illiteracy increases the risk of developing PE-E. ²⁸ This study was a case–control study conducted in Gedeo Zone, Southern Ethiopia, as compared to our current study.

Regarding gravidity, findings show that for every 1 unit increase in gravidity among PE-E women, there is a corresponding 0.34 times decrease in the risk of developing PE-E compared to those who were normal. This is contrary to the findings conducted elsewhere where the odds of developing PE-E among women with primigravida was higher compared to women with multigravida. ³⁰ The reason for the variation in association could be due to the difference in study design and location of study. The study by Boutin et al. ³² was conducted in a single university hospital in Quebec as compared to a prospective cohort study conducted in the Eastern Region of Ghana. Our results also show that a unit increase in comorbidity leads to a corresponding 1.5 times decrease in the risk of developing haemorrhage compared to those who were normal. This is in contrast with the multivariate analysis conducted in Sweden which showed that there was higher risk of haemorrhage among pregnant women with comorbidity compared to those without comorbidity. ⁴⁶ The study in Sweden extracted the data from the already established Swedish Register for Gallstone Surgery and ERCP (GallRiks), whereas this current study was conducted in seven hospitals in the Eastern Region of Ghana with different socio-economic characteristics and poor health systems.

Limitations

The main limitation in this study was about the absolute dependence on the respondents’ ability and trustworthiness to give the right answers. Apart from this, blood pressure measurements and women already on treatment were not added to the analysis. The conceptual framework used for this study is also subject to limitations. The framework is literature based but specifically adopted for this study, considering the study purpose and objectives. Variables included in this framework are not exhaustive of all literature-based variables. This framework can be modified for future research but should provide comprehensive information beyond what is used in this study.

Conclusion

The findings in this study showed that pregnant women with poor personal hygiene were 10 times more likely to develop PE-E compared to those with good personal hygiene. The use of some family planning methods may influence the development of PE-E and therefore, those who are at risk of PE-E may need to seek medical advice before choosing a family planning method. Additionally, personal hygiene practice should be considered a critical factor in routine ANC counselling to help minimise the risk of PE-E. Also, effective CHPSs concept should be implemented to improve access to family planning services in the rural communities. Based on the findings of this study, we recommend that further research, especially clinical trials on the effects of personal hygiene on PE-E, should be conducted.