Hypertension in women of reproductive age: a cross-sectional analysis of prevalence and risk factors across 21 low-income and middle-income countries using Demographic and Health Surveys (2013–2023)

Abstract

Objective

Women of reproductive age (WRA) in low-income and middle-income countries (LMICs) bear a disproportionate burden of hypertension, with limited pooled analyses exploring its prevalence and associated risk factors. This study investigates hypertension prevalence and key determinants among WRA in 21 LMICs.

Design

Retrospective, cross-sectional study.

Participants

Nationally representative data were obtained from the Demographic and Health Survey conducted in 21 LMICs between 2013 and 2023. This research focused on female participants aged 15–49 who were selected for blood pressure monitoring, resulting in a weighted sample of 818 325 WRA (36 970 pregnant and 781 355 non-pregnant).

Primary outcome measures

The primary outcomes were the prevalence of hypertension (defined as systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg) and the identification of individual, household and community-level risk factors associated with the condition. Descriptive statistics of proportions between pregnant and non-pregnant women were assessed. Multilevel logistic regression identified individual, household and community factors affecting hypertension.

Results

The study found the prevalence of hypertension was 8.20% (95% CI 7.95% to 8.45%) among pregnant women and 10.52% (95% CI 10.42% to 10.62%) among non-pregnant women, with substantial regional disparities. Côte d’Ivoire and Haiti exhibited the highest prevalence (48.00% in pregnant women; 57.30% in non-pregnant women, respectively), while the Philippines reported the lowest (0.00% in pregnant women and 0.50% in non-pregnant women). Among pregnant versus non-pregnant women, risk factors included advanced age (35–49 years) (adjusted OR (aOR) 3.31, 95% CI 2.89 to 3.80 vs 3.69, 95% CI 3.60 to 3.77), low education levels (aOR 1.15, 95% CI 1.02 to 1.30 vs 1.33, 95% CI 1.30 to 1.35), not currently employed (aOR 1.08, 95% CI 1.01 to 1.15 vs 1.05, 95% CI 1.04 to 1.09), higher body mass index (BMI) (aOR 1.79, 95% CI 1.76 to 1.81; non-pregnant women), rural residence (aOR 1.14, 95% CI 1.04 to 1.24 vs 1.14, 95% CI 1.12 to 1.16) and limited healthcare access were linked to higher hypertension rates (aOR 1.03, 95% CI 0.94 to 1.13 vs 1.01, 95% CI 1.00 to 1.03).

Conclusions

The burden of hypertension among WRA is driven by advanced age, lower education, high BMI and rural residence. Policymakers should prioritise targeted interventions addressing key sociodemographic and geographic risk factors. Strengthening education, equitable healthcare access and community-based strategies is essential to reducing hypertension-related risks and associated maternal health complications among WRA in LMICs.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• Uses nationally representative data from 21 low-income and middle-income countries covering 818 325 women of reproductive age, ensuring robust and generalisable findings across diverse populations.

• Analyses individual, household and community-level factors, offering a comprehensive understanding of hypertension risks in both pregnant and non-pregnant women.

• The cross-sectional design prevents causal inference, and reliance on self-reported measures (eg, media exposure, healthcare access) may introduce recall or social desirability bias.

• Excludes important confounders such as diet, physical activity and antihypertensive use and may under-represent regional disparities in healthcare and socioeconomic status.

Introduction

Hypertension, defined as systolic blood pressure (SBP)≥140 mm Hg or diastolic blood pressure (DBP) ≥90 mm Hg, is a leading cause of preventable mortality, accounting for 8.5 million deaths annually from complications such as stroke, renal disease and cardiovascular disorders.¹ Despite global and national initiatives to enhance treatment coverage, hypertension prevalence remains high, particularly in low-income and middle-income countries (LMICs).

The overall prevalence of hypertension in 45 countries is 32.3%, according to a systematic study and meta-analysis.^{2 3} The prevalence is higher in upper-middle-income nations (37.8%) than in lower-middle-income (31.1%) and low-income (23.1%) countries. Urban populations also have higher rates (32.7%) than rural areas (25.2%).^{2 3} However, according to the WHO, the number of persons with hypertension has dramatically increased in LMICs, primarily due to a rise in risk factors for the disease in particular populations. There are regional differences, with the highest rate of hypertension seen in the African Region (27%), according to the WHO.⁴ Optimising healthcare for women with hypertensive disorders is a necessary step toward achieving the Sustainable Development Goals.⁵

Studies also have shown that women in LMICs are more likely than women in countries with higher incomes to suffer hypertension.⁶ In LMICs, women’s awareness of their hypertensive status, defined as having a prior diagnosis by a healthcare professional, was higher than that of men (45% among women vs 31% among men). Additionally, compared with males, women in LMICs (36% women vs 22% men) reported using antihypertensive medications at a higher rate. However, hypertension control remains a significant challenge; in LMICs, only 28% of women and 23% of men successfully controlled their condition, whereas in high-income nations, these proportions were notably higher at 52% and 49%, respectively. Awareness and treatment rates are higher among women than men in LMICs, yet the overall control of hypertension remains low. Contributing factors include the high cost and limited availability of antihypertensive drugs, as well as restricted healthcare access.⁶ These patterns highlight differences in hypertension rates among women of reproductive age (WRA) across demographic and regional lines. Regardless of gender-based comparative trends, hypertension in WRA presents unique risks for maternal-fetal complications, limited healthcare access and socio-economic disparities, necessitating a dedicated analysis of its prevalence and determinants.

The dual burden of addressing chronic hypertension and managing maternal health outcomes underscores the urgent need for targeted interventions in these settings. Previous studies show heterogeneous hypertension patterns across income groups, with a lower prevalence in LMICs than in upper-middle-income countries, but a higher reported risk in LMICs than in high-income countries (HICs). This likely reflects methodological differences and varying stages of the epidemiological transition.^{2 3 6}

While the absolute prevalence of hypertension is generally lower in WRA than in men and older populations, it remains a critical public health concern in LMICs due to several factors that uniquely influence risk within this demographic. In many LMICs, pregnancy-related hypertension is also significant and contributes to maternal mortality.⁷ Furthermore, among WRA, factors such as diabetes, obesity and age (35–44 years) have been linked to an increased risk of hypertension.⁸ Preeclampsia may occur in roughly 10%–25% of women with persistent hypertension, which increases risks for serious complications, such as stroke, renal failure and preterm birth. Moreover, there is an increased risk of placental abruption of up to 2% in women with chronic hypertension.⁹ Women with chronic hypertension also have a much higher risk of life-threatening maternal events, such as stroke, renal failure, pulmonary oedema and mortality.¹⁰

In Lesotho, a study found hypertension in one out of five individuals, with a third categorised as prehypertensive. Women aged 45–49 had nearly tenfold higher odds of hypertension compared with younger groups.¹¹ Ghanaian data showed 7.5% of women reported hypertension, with variations across socioeconomic groups. Higher age, education, marital status, employment and wealth were associated with increased hypertension risk. After adjustment for age, marital status remained a significant independent predictor of hypertension, suggesting that the association may reflect marriage-related social stressors or lifestyle changes rather than life stage alone. The relationship between employment and hypertension appears complex and potentially bidirectional, consistent with a healthy worker effect whereby healthier women are more likely to be employed, while those with pre-existing hypertension may be less able to enter the workforce, despite the potential health benefits associated with employment.¹² Similarly, Ghanaian studies linked hypertension to obesity, age and high salt intake.¹³ In the Gambia, hypertension prevalence was 9.9%, with higher rates in older women, those with greater maternal knowledge and those from lower socioeconomic backgrounds.¹⁴ In India, according to the National Family Health Survey, hypertension among WRA rose from 11.8% in 2015–2016 to 14.8% in 2019–2021. Key risk factors included age, high body mass index (BMI), tobacco and alcohol use, while higher education and urban residence were linked to lower odds of hypertension.¹⁵

Despite extensive evidence on hypertension risk factors, important uncertainties remain regarding how these factors vary across LMIC settings. A significant yet sometimes disregarded problem in LMICs, hypertension in WRA is driven by urbanisation, epidemiological shifts and socioeconomic development. Better healthcare resources and public health initiatives are needed to address this. To direct health system programmes, comparable data on the prevalence of hypertension and related factors are required. Although hypertension among WRA is prevalent in LMICs, comprehensive data covering all 21 LMICs between 2013 and 2023 is limited. Ongoing monitoring of hypertension prevalence and risk factors among WRA is essential to guide healthcare policies and intervention programmes effectively. This study aims to investigate the hypertension prevalence in WRA and its contributing factors in LMICs. This study fills a critical knowledge gap by providing pooled prevalence data from 21 LMICs, leveraging the Demographic and Health Survey (DHS) dataset’s strengths to inform policy interventions.

Methods and materials

Participants and study design

The study used pooled data from the most recent DHS conducted between 2013 and 2023 across 21 LMICs. The included countries are: Bangladesh, Benin, Burundi, Cameroon, Côte d’Ivoire, Gambia, Haiti, India, Kenya, Lesotho, Madagascar, Mali, Mauritania, Mozambique, Nepal, Philippines, Tajikistan, Tanzania, Timor-Leste, Yemen and Zambia. The DHS is a cross-sectional survey that is nationally representative and offers accurate data on men, women and children. To ensure that the results are comparable across countries, it uses standardised questionnaires, sample strategies, data gathering methodologies and coding. To investigate the factors linked to hypertension in WRA (15–49 years), all data were harmonised and integrated into a single pooled dataset for multilevel analysis. The inclusion criteria were WRA (15–49 years) with valid recorded blood pressure measurements. Exclusion criteria were records with missing, incomplete or implausible blood pressure values, as well as missing data on key study variables.

The sampling procedures used in DHS were complex, involving multistage selection, clustering and stratification. The survey generally used a two-stage stratified cluster sampling design based on a sampling frame derived from the most recent national census. In the first stage, primary sampling units (PSUs), defined as census enumeration blocks in urban areas and villages in rural areas, were randomly selected, with probability proportional to size from subnational geographic. In the second stage, the number of households (around 20–30) selected by equal probability systematic sampling in the selected Enumeration Areas (women aged 15–45 were interviewed).¹⁶ From an initial pool of 1 049 503 interviewed WRA across the 21 countries, we excluded individuals with missing or incomplete blood pressure measurements. This resulted in a final analytic sample of 818 325, of whom 781 355 were non-pregnant and 36 970 were pregnant. The specific survey years and DHS phases for each country are detailed in table 1.

Table 1. Overview country-specific DHS data across 21 LMICs.

No	Countries	Survey	Years	DHS phase
1	Bangladesh	Standard DHS	2017–2018	DHS-7
2	Benin	Standard DHS	2017–2018	DHS-7
3	Burundi	Standard DHS	2016–2017	DHS-7
4	Cameroon	Standard DHS	2018	DHS-7
5	Côte d’Ivoire	Standard DHS	2021	DHS-8
6	Gambia	Standard DHS	2019–2020	DHS-8
7	Haiti	Standard DHS	2016–2017	DHS-7
8	India	Standard DHS	2019–2021	DHS-7
9	Kenya	Standard DHS	2022	DHS-8
10	Lesotho	Standard DHS	2014	DHS-7
11	Madagascar	Standard DHS	2021	DHS-8
12	Mali	Standard DHS	2018	DHS-7
13	Mauritania	Standard DHS	2019–2021	DHS-7
14	Mozambique	Standard DHS	2022–2023	DHS-8
15	Nepal	Standard DHS	2022	DHS-8
16	Philippines	Standard DHS	2022	DHS-8
17	Tajikistan	Standard DHS	2017	DHS-7
18	Tanzania	Standard DHS	2022	DHS-8
19	Timor-Leste	Standard DHS	2016	DHS-7
20	Yemen	Standard DHS	2013	DHS-VI
21	Zambia	Standard DHS	2018	DHS-7

All survey data were obtained from the DHS Program and accessed through the DHS Program Survey Search database (https://dhsprogram.com/).

DHS, Demographic and Health Survey; LMICs, low-income and middle-income countries.

Measurements

For pregnant women, blood pressure was classified by specific cut-offs: SBP 120–139 mm Hg or DBP 80–90 mm Hg is considered prehypertension; SBP 140–159 mm Hg or DBP 90–99 mm Hg is considered stage 1 hypertension; and SBP ≥160 mm Hg or DBP ≥100 mm Hg is considered stage 2 hypertension. Due to variations in data collection, women with hypertension stages 1 and 2 were grouped as hypertensive (coded as ‘1’), while those with normal blood pressure and prehypertension were coded as ‘0’. Blood pressure was measured using a standardised protocol.¹⁷ Trained health technicians obtained three blood pressure measurements with participants in a seated position at approximately 10 min intervals. The average of the second and third measurements was used to define SBP and DBP.

The DHS collected data on a wide range of variables from selected households and eligible respondents through face-to-face interviews conducted by trained field personnel using standardised questionnaires. Independent factors were grouped at individual, household and community levels. At level 1 (individual level) included age (15–24, 25–34 and 35–49 years), education (no education, primary, secondary and higher), marital status (never in union, married and widowed/divorced/separated), employment (not currently employed vs employed), history of pregnancy termination (miscarriage, abortion or stillbirth; yes/no), smoking status (yes/no), parity (nulliparous, primiparous, multiparous and grand multiparous), and media exposure (yes/no), as well as use of contraception (yes/no) and BMI (normal, underweight and overweight/obesity) only among non-pregnant women. Perceived access to healthcare was assessed based on respondents’ reported difficulties in seeking medical care and was classified as limited if at least one factor (such as distance to a health facility or cost of care) was identified as a ‘big problem’, and not limited otherwise.

At level 2 (household level), variables included household size (<5, 5–10 and >10 members), sex of the head of the household (male/female) and source of drinking water (improved vs unimproved). Toilet facilities were categorised as improved (eg, flush or pour-flush systems, ventilated improved pit latrines) or unimproved (eg, pit latrines without slabs or open defecation). At level 3 (community level), place of residence (urban/rural) and survey year (2013–2015, 2016–2018, 2019–2021 and 2022–2023) were included as community-level contextual variables. Data were pooled using the DHS platform, which applies standardised sampling, questionnaires and blood pressure measurement protocols across participating countries and survey years. This methodological standardisation ensures that outcome measurements are comparable across the 21 countries studied.

Variable selection

Variable selection was guided by prior literature and DHS analytical frameworks. Conceptual frameworks refer to established theoretical models that describe the hierarchical and causal determinants of hypertension, evidence from previous DHS and population-based studies demonstrating associations between selected independent variables and hypertension. To minimise potential confounding and isolate the independent effects of sociodemographic factors, simultaneous adjustment for key covariates was implemented in the multivariable models.

Statistical analyses

To ensure comparability across DHS surveys, variables were harmonised through recording and standardisation of definitions prior to analysis. This process included verification of variable definitions, standardisation of coding schemes and exclusion of implausible values in accordance with DHS data quality protocols. Only variables with comparable definitions across surveys were included in the pooled analyses. STATA V.16 was used for the data analysis. Sampling weights, adjusted for non-response rates and the unequal probability of selection resulting from the multistage stratified cluster sampling approach, were applied to all analyses. Obtaining the sample’s frequencies and percentage distributions for each variable was a necessary step in the descriptive analysis process. Forest plots were used independently for each group to graphically depict the combined prevalence of hypertension in pregnant and non-pregnant women.

Multilevel binary logistic regression was used to identify variables associated with hypertension while accounting for the hierarchical structure of the DHS data, in which individuals were nested within communities/PSUs and communities were nested within countries. Separate models were fitted for pregnant and non-pregnant women. Random intercepts were specified at the community and country levels to capture unobserved heterogeneity and to explain between-community and between-country variation in hypertension. The log-odds of hypertension were modelled using a three-level mixed-effects framework comprising individuals, communities/PSUs and countries. Four sequential multilevel logistic regression models were fitted. First, a null model without covariates was estimated to assess the extent of between-community and between-country variation. Model I included individual-level variables only. Model II included individual-level and household-level variables. Model III further adjusted for individual-level, household-level and community-level variables, with survey year included as a fixed effect to account for variation in survey timing across countries. Results from the multivariable models are presented as adjusted ORs (aORs) with 95% CIs. Estimates were interpreted as average associations across heterogeneous settings (ie, reflecting underlying variation).¹⁸

Patient and public involvement

No participants were involved in developing the research questions, outcome measures, study design, recruitment or reporting and dissemination plans of this study. The findings will be shared through peer-reviewed academic publications.

Results

This study includes a total weighted sample of 818 325 women in LMICs from 2013 to 2023 (36 970 or 4.52% pregnant and 781 355 or 95.48% non-pregnant) (table 2). Among pregnant respondents, most (47.16%) were aged 15–24, while the largest age group among non-pregnant women was 35–49 (37.31%). One-third of the women, 12 492 (33.79%) pregnant and 281 967 (36.09%) non-pregnant, had no formal education. Almost all the 35 698 (96.56%) pregnant and more than half of non-pregnant (535 798 or 68.57%) were married. More than half of non-pregnant and pregnant women were not currently employed. Almost all the 35 234 (95.30%) pregnant and non-pregnant (733 875 or 93.92%) were non-smokers. Two-thirds of the women, 24 161 (65.35%) pregnant women and 579 902 (74.22%) non-pregnant women, had access to improved toilet facilities. The majority of households were headed by males (83.91% among pregnant women and 81.50% among non-pregnant women). Most women (78.53% pregnant and 83.97% non-pregnant) used an improved drinking water source. Rural residents (76.74% of pregnant and 72.49% of non-pregnant women) made up most participants.

Table 2. Frequency distribution of pregnant and non-pregnant respondents in LMICs, 2013–2023.

Variables	Pregnant women		Non-pregnant women
	n=36 970		n=781 355
	Frequency	Per cent	Frequency	Per cent
Age (years)
15–24	17 436	47.16	258 049	33.03
25–34	16 407	44.38	231 799	29.67
35–49	3127	8.46	291 507	37.31
Women's educational status
Not educated	12 492	33.79	281 967	36.09
Primary	17 752	48.02	355 301	45.47
Secondary	1741	4.71	40 380	5.17
Higher	4985	13.48	103 707	13.27
Marital status
Never in union	837	2.26	204 222	26.14
Married	35 698	96.56	535 798	68.57
Widowed/divorced/separated	435	1.18	41 335	5.29
Access to healthcare
Big problem	6892	18.64	125 052	16
Not a big problem	30 078	81.36	656 303	84
Employment status
Not currently employed	21 182	57.3	446 431	57.14
Employed	15 788	42.7	334 924	42.86
Pregnancy termination
No	31 551	85.34	688 318	88.09
Yes	5419	14.66	93 035	11.91
Smoking status
No	35 234	95.3	733 875	93.92
Yes	1735	4.69	47 368	6.06
Exposure to the media
No	10 539	28.51	189 471	24.25
Yes	26 431	71.49	591 884	75.75
Parity
Nullipara (0)	12 897	34.89	236 720	30.3
Primiparous (1)	10 707	28.96	101 848	13.03
Multiparous (2–4)	10 783	29.17	370 403	47.41
Grand multiparous (≥5)	2583	6.99	72 384	9.26
Use of contraception
No	36 970	100.00	412 825	52.83
Yes	0	0	368 530	47.17
Body mass index (BMI)*
Normal			328 941	42.1
Underweight			125 291	16.04
Overweight/obesity			258 591	33.1
Type of toilet facility
Improved	24 161	65.35	579 902	74.22
Unimproved	12 808	34.64	201 446	25.78
Number of individuals in the family
<5 people	14 627	39.56	298 292	38.18
5–10 people	19 924	53.89	444 366	56.87
>10 people	2419	6.54	38 697	4.95
Sex of household head
Male	31 023	83.91	636 816	81.5
Female	5947	16.09	144 533	18.5
Source of drinking water
Improved	29 031	78.53	656 137	83.97
Unimproved	7172	19.4	96 417	12.34
Residence
Urban	8599	23.26	214 926	27.51
Rural	28 371	76.74	566 429	72.49

^*BMI is not reported because the mother's weight during pregnancy is not valid for analysis, as it is influenced by the baby's weight.

LMICs, low-income and middle-income countries.

The prevalence of hypertension among non-pregnant women was 10.52% (95% CI 10.42% to 10.62%), varying widely across regions. Haiti reported the highest prevalence (57.30%), while the Philippines had the lowest (0.50%). Among pregnant women, hypertension prevalence was 8.20% (95% CI 7.95% to 8.45%), with Côte d'Ivoire exhibiting the highest rates (48.00%). Across the analyses, Côte d’Ivoire and Bangladesh showed consistently higher prevalence of hypertension among pregnant and non-pregnant women. Figures1 2 provide detailed prevalence estimates by country. Substantial heterogeneity was observed in the pooled prevalence of hypertension across the 21 LMICs, with very high between-country variability among both pregnant women (I²=99.3%) and non-pregnant women (I²=100%).

Figure 1. Prevalence of hypertension among pregnant women in LMICs (2013–2023). Côte d’Ivoire recorded the highest prevalence (48.00%), while no cases were reported in the Philippines. Low prevalence (1%–20%) is shown in green, moderate prevalence (21%–40%) in yellow and high prevalence (41%–60%) in red. LMICs, low-income and middle-income countries. ES, Effect Size.

Figure 2. Prevalence of hypertension among non-pregnant women in LMICs (2013–2023). Haiti exhibited the highest prevalence (57.30%), with the Philippines showing the lowest (0.50%). Low prevalence (1%–20%) is shown in green, moderate prevalence (21%–40%) in yellow and high prevalence (41%–60%) in red. LMICs, low-income and middle-income countries. ES, Effect Size.

Using multilevel logistic regression (table 3), four models examined the effects of fixed and random factors, showing that hypertension variability across clusters decreased with model complexity. Hypertension variability across clusters was found to be 21.22% for pregnant women, decreasing to 17.43% in model III, showing that the variability of hypertension was explained more effectively in the more complex models. The Proportional Change in Variance (PCV) explained 12.19%, 20.09% and 21.56% of the variance among pregnant mothers in model I, model II and model III, respectively. For non-pregnant mothers, the cluster-level variance decreased by 1.44%, with a PCV of 18.24% after adjusting for individual and community-level variables. The decreasing log-likelihood values and significant p values (<0.001) across all models affirm the robustness of the fitted multilevel logistic regression. Consequently, model III explained a large number of variances in hypertension, which implies that variables at the community and individual levels vary.

Table 3. Random effect parameters and modal comparison.

	Cluster-level variance	ICC (%)	PCV (%)	Log likelihood	Assumption (χ2 (df), p value)
Models for pregnant women
Null model	0.886	21.22	Reference	−14 031.38	Null model is nested in model I (LR χ2(9)=462.14, <0.001)
Model I	0.778	19.12	12.19	−13 800.31	Model I is nested in model II (LR χ2(4)=98.91 <0.001)
Model II	0.708	17.77	20.09	−13 750.86
Model III	0.695	17.43	21.56	−13 746.70	Model II is nested in model III (LR χ2(1)=8.32, 0.004)
Models for non-pregnant women
Null model	0.318	8.82	Reference	−402 326.15	Null model is nested in model I (LR χ2(11)=713 100.52, <0.001)
Model I	0.384	10.44	20.75	−366 670.89	Model I is nested in model II (LR χ2(2)=1061.53, <0.001)
Model II	0.378	10.30	18.87	−366 140.12
Model III	0.376	10.26	18.24	−365 897.21	Model II is nested in model III (LR χ2(1)=485.83, <0.001)

Table 4 indicates that age significantly influenced hypertension, with higher odds for women aged 35–49 (aOR 3.31, 95% CI 2.89 to 3.80 for pregnant, 3.69, 95% CI 3.60 to 3.77 for non-pregnant). Low education levels (pregnant: aOR 1.15, 95% CI 1.02 to 1.30; non-pregnant 1.33, 95% CI 1.30 to 1.35), not currently employed (pregnant: aOR 1.08, 95% CI 1.01 to 1.15; non-pregnant: aOR 1.05, 95% CI 1.04 to 1.09), limited healthcare access were linked to higher hypertension rates (pregnant: aOR 1.03, 95% CI 0.94 to 1.13; non-pregnant: 1.01, 95% CI 1.00 to 1.03), and BMI also significantly influenced hypertension. Women with primary, secondary or higher education had a lower risk of hypertension than those without formal education. Employment also influenced hypertension risk: pregnant working women had a 7% decrease in odds (aOR 0.93, 95% CI 0.87 to 0.95), and non-pregnant working women had a 5% decrease in odds (aOR 0.95, 95% CI 0.92 to 0.96) compared to their non-working counterparts. Among non-pregnant women, smokers had a 1.06 times higher risk of hypertension (95% CI 1.04 to 1.09) than non-smokers. Women with higher BMIs (overweight/obesity) had nearly double the odds of hypertension with aOR of 1.79 (95% CI 1.76 to 1.81) in non-pregnant women. Community factors like rural residence increased the risk in both groups, with an aOR of 1.14 (95% CI 1.04 to 1.24) for pregnant women and 1.14 (95% CI 1.12 to 1.16) for non-pregnant women.

Table 4. Multilevel binary logistic regression of factors associated with hypertension among pregnant and non-pregnant women in LMICs, DHS 2013–2023.

Variables	Pregnant women	Non-pregnant women
	aOR (95%CI)*	aOR (95% CI)†
Individual level
Age (years)
15–24	1	1
25–34	2.00 (1.84 to 2.17)***	1.73 (1.69 to 1.77)***
35–49	3.31 (2.89 to 3.80)***	3.69 (3.60 to 3.77)***
Women's educational status
Not educated	1	1
Primary	0.98 (0.90 to 1.07)	0.86 (0.85 to 0.88)***
Secondary	0.74 (0.61 to 0.89)**	0.79 (0.76 to 0.82)***
Higher	0.87 (0.77 to 0.98)*	0.75 (0.74 to 0.77)***
Marital status
Never in union	1	1
Married	1.40 (1.05 to 1.86)*	1.24 (1.22 to 1.30)***
Widowed/divorced/separated	1.49 (0.98 to 2.28)	1.35 (1.29 to 1.40)***
Access to healthcare
Big problem	1	1
Not a big problem	0.97 (0.89 to 1.06)	0.99 (0.97 to 1.00)
Employment status
Not currently employed	1	1
Employed	0.93 (0.87 to 0.99)*	0.95 (0.92 to 0.96)***
Pregnancy termination
No	1	1
Yes	0.89 (0.81 to 0.98)*	0.94 (0.93 to 0.96)***
Smoking status
No	1	1
Yes	1.13 (0.98 to 1.32)	1.06 (1.04 to 1.09)***
Exposure to the media
No	1	1
Yes	1.16 (1.06 to 1.26)**	0.98 (0.97 to 0.99)*
Parity for both groups of WRA
Nullipara (0)	1	1
Primiparous (1)	0.79 (0.72 to 0.86)***	0.88 (0.86 to 0.91)***
Multiparous (2–4)	0.74 (0.67 to 0.81)***	0.86 (0.84 to 0.89)***
Grand multiparous (≥5)	0.47 (0.39 to 0.56)**	0.79 (0.76 to 0.81)***
Use of contraception‡
No		1
Yes		1.06 (1.05 to 1.08)***
Body mass index (BMI)§
Normal		1
Underweight		0.67 (0.66 to 0.68)***
Overweight/obesity		1.79 (1.76 to 1.81)***
Household level
Type of toilet facility
Improved	1	1
Unimproved	0.78 (0.72 to 0.85)***	0.89 (0.98 to 0.91)***
Number of individuals in the family
<5 people	1	1
5–10 people	0.83 (0.77 to 0.89)***	0.86 (0.85 to 0.87)***
>10 people	0.85 (0.73 to 0.99)*	0.80 (0.77 to 0.82)***
Sex of household head
Male	1	1
Female	0.83 (0.75 to 0.91)***	0.91 (0.89 to 0.93)***
Source of drinking water
Improved	1	1
Unimproved	0.89 (0.81 to 0.98)*	0.89 (0.87 to 0.91)***
Community level
Residence
Urban	1	1
Rural	1.14 (1.04 to 1.24)**	1.14 (1.12 to 1.16)***
Survey years
2013–2015	1	1
2016–2018	0.57 (0.09 to 3.61)	0.46 (0.42 to 0.49)***
2019–2021	1.33 (0.19 to 9.48)	0.93 (0.86 to 0.99)**
2022–2023	0.25 (0.03 to 1.88)	0.42 (0.39 to 0.45)***

*p<0.05, **p<0.01, ***p<0.001.

^*Adjusted for age, women’s educational status, marital status, access to healthcare, employment status, pregnancy termination, smoking status, exposure to the media, parity for both groups of WRA, type of toilet facility, number of individuals in the family, sex of household head, source of drinking water, residence, and survey years.

^†Adjusted for women’s educational status, marital status, access to healthcare, employment status, pregnancy termination, smoking status, exposure to the media, parity for both groups of WRA, use of contraception, BMI, type of toilet facility, number of individuals in the family, sex of household head, source of drinking water, and residence, and survey years.

^‡There are no pregnant women using contraception, so a valid comparison cannot be made.

^§BMI is not reported because the mother's weight during pregnancy is not valid for analysis, as it is influenced by the baby's weight.

aOR, adjusted OR; DHS, Demographic and Health Survey; LMICs, low-income and middle-income countries; WRA, women of reproductive age.

Discussion

To our knowledge, this study provides one of the first comprehensive analyses of hypertension prevalence and risk factors among pregnant and non-pregnant WRA in LMICs, offering critical insights into public health challenges in these settings. This analysis highlights a significant prevalence of hypertension with profound regional disparities, particularly within the populations of Côte d’Ivoire and Bangladesh. While the data identifies women aged 35–49 years and high BMI as consistent drivers of the condition, it also underscores how education levels and rural residence uniquely modulate hypertension risk across the different maternal statuses of pregnancy and non-pregnancy.

Existing studies show a higher burden of hypertension in LMICs compared with HICs, with men and women both experiencing rising rates.^{6 19} The study found hypertension prevalence to be 8.20% among pregnant and 10.52% among non-pregnant women in LMICs, among women. Overall, hypertension prevalence among women was lowest in HICs (25.3%) and highest in sub-Saharan Africa (36.3%), highlighting marked regional differences.²⁰ This contrast underscores how our pooled analysis, by incorporating lower-prevalence regions such as the Philippines (0.50%), provides a more nuanced global average for WRA than single-region studies, while simultaneously reflecting the profound influence of regional healthcare access, socioeconomic conditions and public health awareness programmes on hypertension outcomes.^{6 19 20} Among the 21 LMICs, Bangladesh showed consistently higher prevalence of hypertension among pregnant (45.80%) and non-pregnant (42.54%) women. Studies indicate young adults living outside Dhaka face higher hypertension risks due to limited socioeconomic opportunities, poverty and isolation. High hypertension prevalence among active age populations in Bangladesh poses significant economic and health burdens.^{21 22}

Significant risk factors between hypertension status in pregnant women and non-pregnant women included older age groups (25–34 and 35–49), less education, limited healthcare access, not working status, women with higher BMI and rural living. Earlier studies have found that hypertension begins to rise during reproductive years (between ages 15 and 49). Our findings confirm that risk increases progressively within this cohort, reaching its peak among women aged 35–49. The risk of hypertension increases in women who are older because of obesity, diabetes and atherosclerosis, which affects the small arteries, causing hypertension. This progression highlights that 'reproductive age’ is not a period of uniform risk; rather, it is a window of increasing physiological vulnerability where the cumulative effects of BMI and socioeconomic stressors manifest as elevated blood pressure.^{23 24} Low educational attainment is linked to unhealthy lifestyle choices, such as an unbalanced diet or a lack of exercise, contributing to higher hypertension risk.²⁵ Using concentration index decomposition, the top three sources of inequality in hypertension were job (15%), educational level (18%) and economic status (21%)—account for over 70% of inequality in hypertension prevalence.²⁶ Higher BMI is also a significant factor, as individuals with obesity are 2–3 times more likely to develop cardiovascular diseases, for which hypertension is an early indicator. Similarly, while Indian and Ghanaian data linked hypertension to advanced age (35–49 years) and high BMI, our multivariable analysis demonstrates that these factors remain the most potent predictors across the entire 21-country sample, with non-pregnant women in the 35–49 age group exhibiting a 3.69-fold increase in risk.

Hypertension awareness is notably low in LMICs compared with HICs.⁶ Our study found that educational status and media exposure among both pregnant and non-pregnant women correlate with hypertension prevalence. Women with education beyond elementary school are less likely to be hypertensive, while less educated women in India, sub-Saharan Africa and Bangladesh exhibit higher rates.27,29 Additionally, those with only primary education or no formal education show lower awareness, impacting healthcare access and increasing hypertension risk associated with higher prevalence of uncontrolled blood pressure.³⁰ Media exposure is linked to lower hypertension rates among non-pregnant women, supporting its role as an educational tool. A study used Facebook to promote a 14 min video on hypertension awareness. Viewers who clicked on the advertisement typically spent around 5 min watching the video, which could encourage them to adopt preventive health behaviours.³¹ Through public service announcements, advertisements and short dramas, mass media provides educational information that raises awareness about maternal care, particularly regarding the risks of hypertension among women.³² Access to healthcare and education influences hypertension control, with regular service users more likely to receive effective treatments.^{33 34} Addressing hypertension thus requires both individual monitoring and broader health education strategies, including health education and policy interventions such as salt taxes in non-essential high energy density foods (eg, processed food, junk food), are essential for hypertension management.³

Hypertension during pregnancy is linked to significant maternal and fetal health risks, including increased morbidity and mortality.³⁵ In the USA, hypertension accounts for up to 15% of maternal deaths, ranking it as the second most common cause of maternal mortality.^{36 37} While in India, hypertension may be responsible for up to one-third of maternal deaths.³⁸ A study in Ethiopia revealed that the perinatal mortality rate was 21.2% for women with hypertensive disorders of pregnancy (HDP) compared with 6.2% for women with normal blood pressure, with all adverse perinatal outcomes being more prevalent among those with HDP.³⁹ Furthermore, women who experience elevated blood pressure during pregnancy are at a greater risk of developing cardiovascular diseases later in life.^{40 41}

Pregnancy-related hypertension disorders, like pre-eclampsia and gestational hypertension, require careful management, including monitoring and treatment to reduce complications. Key risk factors include advanced maternal age (35–49 years), obesity and pre-existing conditions.⁴² Effective management, involving early detection, consistent monitoring, lifestyle interventions, pharmacological treatment and, in severe cases, planned preterm delivery, is crucial for minimising adverse outcomes for both mother and fetus.⁴³ Pre-eclampsia requires close monitoring to keep blood pressure under 140/90 mm Hg, using antihypertensives and magnesium sulfate to prevent seizures, with early delivery recommended in severe cases.⁵ Similarly, gestational hypertension mandates careful monitoring of blood pressure and fetal well-being, initiating pharmacological intervention if levels exceed 140/90 mm Hg to prevent complications.^{44 45}

Education plays a critical role, as women with higher education levels are less likely to develop hypertension due to better health literacy that enables a better understanding of disease risks, improved access to medical information and enhanced ability to make more rational and informed decisions regarding healthy lifestyles, such as diet, physical activity and stress management. Education can also improve an individual’s cognitive abilities, leading to better preventive behaviours, including adherence to health precautions.⁴⁶ Access to healthcare is another key factor, with barriers such as cost and infrastructure limiting timely care. Additionally, the availability of affordable antihypertensive medications remains a challenge, especially in rural areas where healthcare access is poorer.⁴⁷ Community health workers can play a pivotal role in educating rural populations about hypertension risk factors.

Obesity increases the risk of hypertension due to insulin resistance and inflammation leading to higher sympathetic nervous system activity and sodium retention in kidneys which causes high blood pressure.⁴⁸ Poor dietary habits, such as consuming high-sodium and processed foods, contribute to hypertension, while low fruit and vegetable intake limits essential nutrients.^{49 50} Physical inactivity, often driven by urbanisation, further heightens these risks, emphasising the need for targeted interventions.⁵¹ The observed associations are likely modulated by unmeasured behavioural confounders, including increasing sodium consumption and decreasing levels of physical activity associated with shifting labour patterns. The lack of individual-level data on these factors in the DHS prevents a direct correlation; however, the persistent significance of BMI in our models (aOR 1.79) serves as a proxy for the metabolic consequences of these lifestyle shifts. Smoking, though historically lower among women in LMICs, is rising in urban areas, exacerbating hypertension risk. Smoking increases blood pressure primarily through nicotine-induced vasoconstriction, which narrows blood vessels and raises heart rate, both of which increase blood pressure. Additionally, smoking promotes oxidative stress and inflammation, leading to arterial stiffness and atherosclerosis, which contribute to sustained high blood pressure over time.⁵²

A multifaceted approach combining individualised, household and community care is required to increase awareness, treatment and control among WRA. There are many barriers in the healthcare system, including lack of access to care and low health literacy. Identifying innovative strategies to overcome these barriers and deliver effective interventions in LMICs to prevent and control hypertension is urgently needed. Community involvement in hypertension management has been shown to improve outcomes.⁵³ Policymakers concerned about improving the health of disadvantaged WRA focus on improving hypertension prevention and control.³⁴ Furthermore, collaborative efforts between global and local stakeholders will be necessary to improve the control of hypertension by promoting blood pressure screenings and access to affordable healthcare.

This study uses nationally representative datasets from 21 LMICs, providing a robust overview of hypertension prevalence among 818 325 WRA, pregnant and non-pregnant, ensuring reliable and generalisable findings across demographic settings. It accounts for individual, household and community-level factors, providing nuanced insights into hypertension risks. The distinction between pregnant and non-pregnant women highlights unique risks for each group. Despite its strengths, the study has limitations. The cross-sectional design limits causal inference, and reliance on self-reported data (eg, healthcare access, media exposure) may introduce recall bias or social desirability bias. The dataset may not fully capture regional disparities in healthcare access or socioeconomic conditions. It is possible that pre-existing hypertension or related illness contributes to less opportunities for employment rather than employment status solely influencing health outcomes, hence reverse causality between health status and employment cannot be completely ruled out. Some confounders, such as diet, physical activity and antihypertensive treatments, were excluded due to DHS data limitations. The survey does not routinely collect standardised, cross-country comparable data on dietary intake or objectively measured physical activity, and information on antihypertensive use is either unavailable or lacks treatment details. Future research should explore longitudinal studies to confirm temporal relationships, along with regional data, community-based interventions and public health campaigns to identify targeted strategies for LMICs.

Conclusions

This study underscores the significant burden of hypertension among WRA in LMICs and highlights the interplay of sociodemographic, household and community factors. Policymakers and healthcare stakeholders must prioritise targeted interventions to address these disparities, ensuring equitable access to healthcare and strengthening community-based programmes. These efforts are crucial for achieving global targets to reduce hypertension-related morbidity and mortality. Adopting national hypertension screening programmes in LMICs can identify at-risk populations earlier, mitigating long-term health burdens.

Promoting health education and encouraging healthier lifestyle changes are essential for preventing hypertension. Enhanced monitoring and treatment of hypertensive WRA through national surveillance systems can help manage this growing issue. Community-based interventions and telemedicine offer promising avenues for increasing care access, especially in underserved areas. Adaptable strategies, including salt reduction, should be prioritised to meet global targets for reducing hypertension. Regular evaluation will ensure that interventions remain responsive to the evolving needs of WRA in LMICs.

The funder was not involved in the manuscript’s study, drafting, evaluation or approval.

Data availability statement

Data are available in a public, open access repository.