Accepted on 26 Jul 2022            Submitted on 15 Mar 2022

Introduction

The 21st century marked the onset of an epidemiological transition in Africa, with a surge of noncommunicable diseases (NCDs) taking the lead as the primary cause of death [1]. With the contemporary double burden of communicable and non-communicable diseases, cardiovascular diseases (CVDs) are rising in sub-Saharan Africa (SSA) [2, 3], predominantly driven by increased rates of hypertension, smoking, and obesity [4]. In SSA, CVDs are the most frequent causes of NCDs deaths, responsible for approximately 13% of all deaths and 37% of all NCDs deaths [2]. Between 1990 and 2013, SSA remained the only region of the world where CVD related mortality increased [3]. Of concern, the burden of CVD is expected to double by 2030 [3]. The significant increase in the number of deaths from CVDs, namely hypertensive heart disease, stroke, and heart failure, raises a possibility of an emerging epidemic in SSA that is preventable [5]. Moreover, an upsurge of the CVDs poses an additional burden on an already over-burdened healthcare systems [6] that needs to be addressed.

Earlier studies have established sex differences in the risk and occurrence of CVDs [7, 8]; however, they have mostly ignored the intersectionality of sex, gender, and other demographic characteristics. A review of the evidence for Ghana, Nigeria, South Africa, Sudan, and Tanzania indicated a poor health system response to the increasing risk of CVDs with no discussion on the risk associated with sex and gender [9]. Whilst sex-based studies, especially in NCDs, are gaining more advocacy globally, the sex and gender research gap and data paucity is highly visible in the African context. While sex differences in cardiovascular risk and events have been investigated, their association with sociocultural gender has not received the same attention. Unfortunately, many studies use the terms gender and sex interchangeably. Furthermore, studies with sex-stratified analysis that lack considerations of the influences of gender may not present a complete description of cardiovascular health/events. This directly impacts the design and implementation of interventions designed to meet the targets for Sustainable Development Goal 3 and achieve a 30% reduction in premature mortality due to NCDs [10]. Therefore, this study aimed to investigate and describe the associations between biological sex, gender-related variables, and cardiovascular health risk factors in SSA countries.

Methods Study design

This is a cross-sectional study using deidentified data from the population-wide, sub-national/national World Health Organization’s (WHO)-STEPwise Approach to surveillance of risk factors for non-communicable disease (STEPS) survey data [11]. The design of the STEPS survey has been published elsewhere [11]. Briefly, the survey included Step I: administration of a questionnaire to elicit demographic information and lifestyle behaviors; Step II: physical measurements (height, weight, waist/hip circumference, blood pressure, and heart rate); and Step III: biochemical assessments (fasting blood glucose, total cholesterol, and urine sample for testing of sodium, potassium, and creatinine levels) of respondents. The measurements were carried out at the home of the survey participants immediately after the conclusion of Step I, and assessments were carried out the next day at a common place for all the participants. The surveys were conducted between 2006–2014. The inclusion criteria consisted of participants of working age (18–69 years). For the purposes of this analysis, we selected STEPS data from five SSA countries (Botswana, Gambia, Ghana, Guinea, and Mali) for which data were available in a Microdata repository.

Explanatory variables

Self-reported biological sex and gender-related variables (i.e., education, occupation/employment, marital status, household income, and household size) were identified using the GOING-FWD methodology [12]. Because of country-specific differences in data coding or collection, data harmonization and recategorization was done for the identified gender-related variables. Biological sex in this study was defined as binary, male or female; education level was categorized into three groups: less than secondary, secondary, and post-secondary; marital status was categorized into three groups: single/never married, widow/separated/divorced, and married/common in law; work status was categorized into three groups: paid, unpaid, and unemployed; and household size was categorized into five groups: 1, 2, 3, 4, and ≥5 people in the household. Age of the respondents was recorded as a continuous variable in the dataset and was categorized into six groups: <20, 20–29, 30–39, 40–49, 50–59, and 60–69 years.

Outcome variable (STEPS-HEART index)

Cardiovascular health (CVH) derived from the STEPS-HEART index was the main outcome variable in this study. Variables measured in the survey (i.e., smoking, hypertension, diabetes, obesity/overweight, and daily consumption of fruits and vegetables) were used to construct a composite labeled the STEPS-HEART index; value ranging from 0 to 5 [13]. The STEPS-HEART index is an adaptation of the previously published CANHEART health index tool [14]. It provides a composite quantitative measure (of 0 [worst] to 5 [best]) of CVH factors— smoking, diabetes mellitus, elevated blood pressure, obesity/overweight, and daily consumption of fruits and vegetables [14]. Performance of ideal CVH behavior within each factor was positively coded with a score of 1, while non-ideal performances were scored 0. Each metric of ideal cardiovascular health is defined in Appendix, Table 1. These definitions were collectively adapted from the American Heart Association, World Health Organization, and/or Center of Disease Control and Prevention definitions or recommendations. Later, a binary variable of STEPS-HEART index (value of 4 and 5—better CVH coded as 1, and value less than 4—poorer CVH coded as 0) was formed for the purpose of analysis.

Table 1

Descriptive characteristics of survey respondents, by biological sex (n = 15356).

N1 OVERALL BIOLOGICAL SEX FEMALE (N = 9425) MALE (N = 5931) N (%) OR MEAN [SD] N (%) OR MEAN [SD] N (%) OR MEAN [SD] Demographic variablesCountry/Survey year 15356     Ghana, 2006 2662 (17.3) 1773 (18.8) 889 (15.0)     Gambia, 2010 4103 (26.7) 2335 (24.8) 1768 (29.8)     Mali, 2007 2401 (15.6) 1441 (15.3) 960 (16.2)     Guinea, 2009 2288 (14.9) 1230 (13.1) 1058 (17.8)     Botswana, 2014 3902 (25.4) 2646 (28.1) 1256 (21.2) Age (in years) 15356 37.5 [12.5] 36.9 [12.3] 38.5 [12.8] Age categories 15356     <20 917 (6.0) 568 (6.0) 349 (5.9)     20–29 4789 (31.2) 3138 (33.3) 1651 (27.8)     30–39 4121 (26.8) 2559 (27.2) 1562 (26.3)     40–49 2770 (18.0) 1069 (17.1) 1161 (19.6)     50–59 1962 (12.8) 1101 (11.7) 861 (14.5)     60–69 797 (5.2) 459 (4.8) 347 (5.9) Gender variablesEducation level 15295     Less than secondary 8350 (54.6) 5548 (59.1) 2802 (47.4)     Secondary 5514 (36.1) 3195 (34.0) 2319 (39.3)     Post-secondary 1431 (9.4) 645 (6.9) 786 (13.3) Marital status 15278     Single/Never married 4575 (29.9) 2486 (26.5) 2089 (35.4)     Married/Common in law 9484 (62.1) 5888 (62.8) 3596 (61.0)     Widow/Separated/Divorced 1219 (8.0) 1006 (10.7) 213 (3.6) Work status 14559     Paid 8363 (57.4) 4051 (45.9) 4312 (75.2)     Unpaid 4114 (28.3) 3244 (36.8) 870 (15.2)     Unemployed 2082 (14.3) 1527 (17.3) 555 (9.7) Household size 14955     1 person 2530 (16.9) 1512 (16.5) 1018 (17.6)     2 persons 4224 (28.2) 2667 (29.1) 1557 (26.8)     3 persons 2733 (18.3) 1776 (19.4) 957 (16.5)     4 persons 1623 (10.9) 963 (10.5) 660 (11.4)     ≥5 persons 3845 (25.7) 2238 (24.4) 1607 (27.7) Annual household income (USD) 11819     <20000 9106 (77.0) 5312 (75.5) 3794 (79.4)     ≥20000 2713 (23.0) 56 (24.5) 67 (20.6) Cardiovascular risk-factorsSmoking 15179     Smoker 1859 (12.2) 210 (2.2) 1649 (28.4)     Non-smoker 13320 (87.8) 9169 (97.8) 4151 (71.6) Overweight/obesity 14673     BMI < 25 8767 (59.7) 4642 (51.7) 4125 (72.5)     BMI ≥ 25 5906 (40.3) 4343 (48.3) 1563 (27.5) Leisure physical activity 1576     <30 mins activity 226 (14.3) 138 (20.0) 88 (9.9)     ≥30 mins activity 1350 (85.7) 552 (80.0) 798 (90.1) Fruit and vegetable consumption 10700     <5 servings/day 7205 (67.3) 4490 (68.1) 2715 (66.2)     ≥5 servings/day 3495 (32.7) 2107 (31.9) 1388 (33.8) Hypertension 11907     Yes 2247 (18.9) 1671 (21.6) 576 (13.8)     No 9660 (81.1) 6057 (78.4) 3603 (86.2) Diabetes 6807     Yes 311 (4.6) 215 (4.9) 96 (4.0)     No 6496 (95.4) 4162 (95.1) 2334 (96.0) Outcome variableSTEPS-HEART Index (Africa) 4271     0 4 (0.1) 2 (0.1) 2 (0.1)     1 80 (1.9) 26 (1.8) 54 (1.9)     2 415 (9.7) 109 (7.4) 306 (10.9)     3 1464 (34.3) 441 (29.9) 1023 (36.6)     4 1696 (39.7) 605 (41.0) 1091 (39.0)     5 612 (14.3) 291 (19.7) 321 (11.5)

Note: BMI (Body Mass Index); SD (Standard Deviation).

1 Number of complete observations (excludes missing data).

STEPS-HEART index value of ≥4 indicates better CVH and value <4 indicates poor CVH.

Statistical Analysis

Means, standard deviation, frequencies and percentages were used to describe the study population. Multivariate logistic regression models were used to investigate the association between CVH and gender-related factors in the African context. Sex-stratified logistic regression models were fitted to identify the impact of the association between the various gender variables and the CVH within each sex stratum. Age was included in all the models. Two-way interaction between the sex and gender-related factors were tested by including an interaction term in bivariate models. We also conducted p for trend test to look at a potential trend between age range and household size across CVH categories. Missing data were imputed using the multiple imputation method. All the statistical analyses were executed using SAS software version 9.4 (SAS Institute, Cary, North Carolina).

Data imputation

A complete case analysis when data are not missing completely at random is inefficient and can potentially lead to biased results [15]; hence, we employed multiple imputation to account for the missing data. Missing values were imputed using SAS, which employs a fully conditional specification (FCS) algorithm [16]. The FCS method is an iterative Markov Chain Monte Carlo procedure that sequentially imputes missing values for all covariates included starting from the first variable with missing values by specifying a linear regression or logistic regression model for each continuous or categorical variable, respectively. We used 30 imputations with 20 (by default) iterations and included work status (6% missing), household size (3% missing), and annual household income (23% missing) variables for the imputation.

Results Descriptive characteristics of respondents

Descriptive characteristics of the study population stratified by biological sex are presented in Table 1. Among 15,356 respondents, 61.4% were females (mean age 36.9 years). A quarter of the respondents were from Gambia (26.7%) and Botswana (25.4%) each. Female respondents were more likely to be married (62.8% vs. 61.0%), to have lower formal education (59.1% vs. 47.4%), and report less paid work (45.9% vs. 75.2%) compared to their male counterparts.

Over a third of the study population had three or more cardiovascular risk factors with a higher prevalence in females than males (Figure 1). Although close to 80% of the study population was less than 50 years of age, the prevalence of hypertension was high, especially in women (with a 21.6% reported prevalence). Diabetes prevalence was similar between sexes but higher than reported population prevalence for these countries, likely related to the very high prevalence of overweight/obesity, especially among women (48.3%). Smoking prevalence was very high in males, almost a third (28.4%), while it remained low in females (2.2%). Of the persons reporting engagement in leisure physical activities, a majority (85.7%) reported meeting the ideal daily physical activity, however, in the wider population, access to fruits and vegetables was low in both sexes (31.9% and 33.8%); only a third had access to at least 5 servings per day (Table 1).

Figure 1 

Cardiovascular risk-factors of the respondents by biological sex.

Association of gender-related factors with CVH by sex

The respondents had a mean CVH value of 3.5 and a median value of 4, indicating intermediate to good cardiovascular health across the study population. Overall, females showed poorer CVH compared to males (ORfemale = 0.95, 95% CI: 0.91–0.99). CVH decreased with increasing age for both sexes. Females in the oldest age group (ORfemale = 0.37, 95% CI: 0.26–0.53) had poorer CVH compared to their male counterparts (ORmale = 0.67, 95% CI: 0.47–0.97, pinteraction < 0.01). While unpaid workers showed better CVH (ORfemale = 1.08, 95% CI: 1.01–1.17 and ORmale = 1.28, 95% CI: 1.12–1.47), those who were unemployed (ORfemale = 0.87, 95% CI: 0.79–0.95 and ORmale = 0.78, 95% CI: 0.68–0.89) showed poorer CVH for both males and females. Females with household size of ≥5 showed better CVH compared to males (ORfemale = 1.39, 95% CI: 1.24–1.56 vs. ORmale = 1.18, 95% CI: 1.05–1.34, pinteraction < 0.05) (Table 2).

Table 2

Multiple imputation logistic regression model for assessing association of gender-related variables with cardiovascular health, overall and by biological sex.

CARDIOVASCULAR HEALTH (STEPS-HEART INDEX) OVERALL FEMALE MALE P INTERACTION OR (95% CI) P-VALUE OR (95% CI) P-VALUE OR (95% CI) P-VALUE Sex (Female) 0.95 (0.91–0.99) 0.01 Age     <20 2.06 (0.50–8.48) 0.32 2.35 (0.56–9.81) 0.25 1.54 (0.36–6.55) 0.56 0.15     20–29 (ref) – – – – – – –     30–39 1.14 (0.85–1.51) 0.39 1.25 (0.93–1.68) 0.16 1.07 (0.79–1.44) 0.68 <0.05     40–49 0.76 (0.56–1.02) 0.07 0.77 (0.57–1.04) 0.09 0.79 (0.57–1.08) 0.14 0.60     50–59 0.65 (0.49–0.87) 0.01 0.58 (0.43–0.78) <0.001 0.80 (0.58–1.11) 0.18 <0.01     60–69 0.48 (0.35–0.66) <0.001 0.37 (0.26–0.53) <0.001 0.67 (0.47–0.97) <0.05 <0.001 Education level     Less than secondary (ref) – – –     Secondary 0.99 (0.94–1.05) 0.81 0.93 (0.86–1.02) 0.11 1.10 (1.01–1.20) <0.05 <0.01     Post-secondary 0.91 (0.84–0.99) <0.05 0.87 (0.77–0.98) <0.05 0.92 (0.83–1.03) 0.16 0.54 Marital status     Single/Never married (ref) – – –     Married/Common in law 1.11 (1.04–1.18) <0.01 1.05 (0.98–1.14) 0.16 1.09 (0.96–1.24) 0.19 <0.01     Widow/Separated/Divorced 0.99 (0.89–1.09) 0.79 1.03 (0.93–1.15) 0.57 0.93 (0.75–1.16) 0.54 0.59 Work status     Paid (ref) – – –     Unpaid 1.14 (1.08–1.21) <0.001 1.08 (1.01–1.17) <0.05 1.28 (1.12–1.47) <0.001 0.23     Unemployed 0.84 (0.79–0.90) <0.001 0.87 (0.79–0.95) <0.01 0.78 (0.68–0.89) <0.001 0.60 Household size     1 (ref) – – –     2 0.93 (0.86–0.99) <0.05 0.88 (0.81–0.96) <0.01 0.99 (0.89–1.11) 0.97 0.79     3 0.96 (0.88–1.05) 0.36 0.96 (0.87–1.07) 0.47 0.95 (0.84–1.09) 0.49 0.29     4 0.99 (0.90–1.09) 0.92 0.99 (0.88–1.13) 0.98 0.98 (0.84–1.15) 0.83 0.93     ≥5 1.30 (1.19–1.43) <0.001 1.39 (1.24–1.56) <0.001 1.18 (1.05–1.34) <0.01 <0.05

Note: OR (Odds Ratio); CI (Confidence Interval); STEPS-HEART index (4 or 5: Better CVH vs. <4: Poorer CVH).

Association of gender-related factors with CVH by country

Table 3 demonstrates the country specific CVH of the respondents. Females showed poorer CVH in all countries except in Gambia, where the association was in reverse (OR = 1.07, 95% CI: 0.97–1.19); however, the finding is not statistically significant. CVH decreased with increasing age in all five countries (p for trend < 0.001). Respondents with post-secondary education as compared to less than secondary education in Gambia showed significantly poorer CVH (OR = 0.67, 95% CI: 0.53–0.83). Respondents who were married or common in law had better CVH compared to single or never married individuals, mainly in Gambia (OR = 1.12, 95% CI: 0.97–1.28) and Mali (OR = 1.19, 95% CI: 0.99–1.43). Unpaid respondents compared to their paid counterparts had better CVH except in Guinea, where unpaid respondents had poorer CVH (OR = 0.91, 95% CI: 0.72–1.15). CVH seemed to be better with increasing number of household members in all countries (p for trend < 0.001). There was no evidence of heterogeneity in the associations between sexes (pinteraction > 0.05), except for age and education level in Gambia.

Table 3

Multiple imputation logistic regression model for assessing association of gender-related variables with cardiovascular health, by country.

CARDIOVASCULAR HEALTH (STEPS-HEART INDEX) GHANA GAMBIA MALI GUINEA BOTSWANA OR (95% CI) P-VALUE OR (95% CI) P-VALUE OR (95% CI) P-VALUE OR (95% CI) P-VALUE OR (95% CI) P-VALUE Sex (Female) 0.73 (0.66–0.80) <0.001 1.07 (0.97–1.19) 0.15 0.98 (0.88–1.09) 0.67 0.99 (0.89–1.11) 0.94 1.00 (0.93–1.09) 0.93 Age     <20 NA NA 2.76 (1.87–4.06) <0.001 1.57 (0.17–14.18) 0.69 1.34 (0.15–11.52) 0.79     20–29 (ref) – – – – – – – – – –     30–39 1.28 (1.09–1.49) <0.01 1.38 (1.19–1.59) <0.001 1.04 (0.82–1.32) 0.76 1.18 (0.72–1.95) 0.51 1.14 (0.72–1.81) 0.57     40–49 0.90 (0.76–1.06) 0.21 0.88 (0.74–1.03) 0.11 0.61 (0.48–0.77) <0.001 0.83 (0.51–1.36) 0.47 0.97 (0.61–1.55) 0.91     50–59 0.87 (0.72–1.04) 0.12 0.77 (0.63–0.93) <0.01 0.72 (0.55–0.94) <0.05 0.68 (0.41–1.13) 0.14 0.59 (0.37–0.95) <0.05     60–69 0.56 (0.41–0.75) <0.001 0.59 (0.42–0.82) <0.01 0.50 (0.34–0.73) <0.001 0.51 (0.28–0.91) <0.05 0.57 (0.34–0.98) <0.05 Education level     Less than secondary (ref) – – – – –     Secondary 0.89 (0.79–1.01) 0.07 1.16 (0.99–1.37) 0.06 0.96 (0.69–1.35) 0.82 0.89 (0.73–1.08) 0.24 0.99 (0.89–1.11) 0.90     Post-secondary 1.07 (0.89–1.29) 0.46 0.67 (0.53–0.83) <0.001 0.84 (0.45–1.57) 0.59 1.08 (0.84–1.39) 0.54 1.03 (0.90–1.17) 0.66 Marital status     Single/Never married (ref) – – – – –     Married/Common in law 0.99 (0.89–1.13) 0.93 1.12 (0.97–1.28) 0.11 1.19 (0.99–1.43) 0.06 0.99 (0.84–1.19) 0.98 0.95 (0.82–1.11) 0.53     Widow/Separate/Divorce 0.95 (0.79–1.13) 0.56 0.89 (0.70–1.12) 0.32 0.88 (0.63–1.22) 0.42 0.75 (0.56–1.01) 0.06 0.91 (0.72–1.15) 0.44 Work status     Paid (ref) – – – – –     Unpaid 1.03 (0.79–1.34) 0.85 1.19 (1.02–1.39) <0.05 1.19 (0.97–1.48) 0.09 0.91 (0.72–1.15) 0.44 1.14 (1.01–1.29) <0.05     Unemployed 0.98 (0.77–1.23) 0.84 0.80 (0.63–1.03) 0.07 0.87 (0.71–1.07) 0.19 1.11 (0.76–1.61) 0.58 0.99 (0.89–1.11) 0.92 Household size     1 (ref) – – – – –     2 0.93 (0.81–1.08) 0.35 1.03 (0.90–1.17) 0.66 0.97 (0.72–1.31) 0.86 0.99 (0.81–1.21) 0.94 0.95 (0.82–1.09) 0.44     3 0.93 (0.78–1.09) 0.36 1.02 (0.89–1.16) 0.80 0.92 (0.66–1.29) 0.64 1.02 (0.82–1.27) 0.85 1.02 (0.86–1.21) 0.81     4 0.99 (0.81–1.23) 0.97 1.03 (0.86–1.22) 0.77 0.99 (0.69–1.43) 0.99 0.98 (0.76–1.28) 0.93 0.95 (0.76–1.18) 0.62     ≥5 1.02 (0.84–1.24) 0.84 1.25 (1.09–1.42) <0.01 1.16 (0.95–1.42) 0.15 1.17 (0.96–1.44) 0.12 1.17 (0.94–1.47) 0.14

Note: NA (Not available); OR (Odds Ratio); CI (Confidence Interval); STEPS-HEART index (4 or 5: Better CVH vs. <4: Poorer CVH).

Discussion

In this SSA population cohort, the cardiovascular risk factor prevalence was high especially in females who exhibited high levels of hypertension (21.6%) and almost half of whom were overweight and/or obese (48.3%). The cumulative cardiovascular risk factor level was higher in females than males, and this association was compounded by deleterious gender-related factors, mostly socioeconomic, which put females at a greater disadvantage compared to males. In the SSA countries surveyed, females reported worse CVH than males in all countries.

High prevalence of hypertension among females in our study cohort is in contrast with the studies that show men having higher blood pressure regardless of their race and ethnicity across all age groups [17]. The high prevalence of hypertension among the females in SSA population could be related to the high prevalence of overweight/obesity. Studies conducted in SSA countries have shown existing sex disparities in overweight/obesity where the increase in prevalence was accounted for almost entirely by females [18, 19]. Furthermore, the relationship between obesity and hypertension is well established [20, 21]. The rising prevalence of overweight/obesity in this young population of females is worrisome as it is associated with increasing risks of CVDs, particularly hypertension and diabetes that are known risk factors of myocardial infarction, heart failure, stroke, and cognitive decline [21]. In order to impact the rising epidemic of hypertension in women in SSA, gender factors should be considered in the interventions aimed at the prevention and or control of hypertension. Gendered lifestyle factors, such as low consumption of fruits and vegetables and limited physical activity, that are predominant risk factors for women [22], should be taken into account to address the epidemic of CVD risk.

Marital status is an important gendered factor known to predict a range of health outcomes, including CVD. This study identified that marital status was associated with better CVH in both sexes, marriage being more beneficial fo