Abstract
BACKGROUND
A higher prevalence of hypertension is reported among Afro-descendants compared with other ethnic groups in high-income countries; however, there is a paucity of information in low- and medium-income countries.
METHODS
We evaluated 3,745 adults from 3 ethnic groups (552 White, 2,746 Mestizos, 447 Afro-descendants) enrolled in the prospective population-based cohort study (PURE)—Colombia. We assessed associations between anthropometric, socioeconomic, behavioral factors, and hypertension.
RESULTS
The overall prevalence of hypertension was 39.2% and was higher in Afro-descendants (46.3%) than in Mestizos (37.6%) and Whites (41.5%), differences that were due to the higher prevalence in Afro-descendant women. Hypertension was associated with older age, increased body mass index, waist circumference and waist-to-hip ratio, independent of ethnicity. Low education was associated with hypertension in all ethnic groups, and particularly in Afro-descendants, for whom it was the factor with the strongest association with prevalence. Notably, 70% of Afro-descendants had a low level of education, compared with 52% of Whites—26% of Whites were university graduates while only 7% of Afro-descendants were. We did not find that education level alone had a mediator effect, suggesting that it is not a causal risk factor for hypertension but is an indicator of socioeconomic status, itself an important determinant of hypertension prevalence.
CONCLUSIONS
We found that a higher prevalence of hypertension in Colombian Afro-descendants than other ethnic groups. This was principally associated with their lower mean educational level, an indicator of lower socioeconomic status.
Keywords: Colombia, education level, ethnicity, hand grip strength, hypertension
Graphical Abstract
Graphical Abstract.
The overall worldwide prevalence of hypertension in adults is estimated to be 31.1% and the majority of the approximately 1.3 billion people affected live in low- and middle-income countries (LMICs).1,2 This disproportionate burden in LMICs could be explained by multiple sociodemographic, economic, and behavioral factors. Ethnicity is one of the factors proposed to contribute to these differences, with evidence from large population studies such as the National Health and Nutrition Examination Survey (NHANES) showing a greater prevalence of hypertension in African Americans than non-Hispanic Whites, Hispanic and Asians.3–5 However, less information is available in LMICs such as Colombia, a multiethnic middle-income country.6,7 Therefore, this study aimed to determine the prevalence of hypertension across the 3 ethnic groups that predominate in Colombia: White, Mestizo, and Afro-descendant, and evaluate associations with socioeconomic, metabolic, anthropometric, and behavioral factors which could contribute to potential differences.
METHODS
Population and study design
Data from the Colombian subjects included in the global Prospective Urban Rural Epidemiology (PURE) study were analyzed.8 With the aim of obtaining an adequate geographic and social representation, participants were selected from Colombian urban and rural communities from 11 of the most populated departments (Atlántico, Bolívar, Caldas, Casanare, Cauca, Cesar, Cundinamarca, Nariño, Quindío, Santander, and Tolima) that comprise 51.29% of the Colombian population as previously described.9 Briefly, a multistage convenience sample survey was used. During the first and second stages, the departments and communities were selected. In the third stage, a representative sample of households was recruited, using a community-sampling framework. Households were eligible if at least one member was 35–70 years old and if the members intended to continue living at that address for 4 years or more. The sociodemographic, behavioral characteristics, and health status of all participants were collected in a specific format. This report includes 3,745 adults between 35 and 70 years of age with complete socioeconomic data and who by self-recognition were assigned to an ethnic group based on the parameters used by the Colombia’s National Administrative Department of Statistics (DANE).7 The local ethics committees approved the study protocol, and all participants signed written consent.
Procedures
The sociodemographic characteristics of all consenting participants, including date of birth, cardiovascular disease (CVD) risk factors (such us smoking, hypertension, diabetes, psychosocial factors, and alcohol consumption), monthly income, and the use of home polluting cooking fuels (e.g., wood, charcoal, animal dung, coal) were recorded, and a basic physical examination was performed. Blood pressure, anthropometric, and handgrip measurements were taken. Those with a university diploma were classified as having a high educational level, a secondary or technical diploma as a middle-level education, and those without a schooling history, primary or unknown education status as a low level of education. Current smokers were those who reported consuming a daily tobacco product in the last 12 months or reported quitting smoking in the last year. Self-reported alcohol abstinence was considered as a never drinker, former drinkers as those having ceased alcohol consumption for a year or more, and current drinkers as those who reported alcohol consumption in the past year. We registered daily meal consumption through a food frequency questionnaire (FFQ).10 Obtaining ethnic group data is typically a complex process, due to its subjective nature-self-defined ethnicity tends to evolve according to the context, in particular, of the social and political attitudes of the population. We used the self-recognition of membership of one of the ethnic groups as Mestizo (a mix of White with aboriginal), White, or Afro-descendant (including Black and a mix of Black with White and Black with aboriginal), based on the DANE recommendations.7 This implies that the individual recognizes themselves as belonging to one of the 3 ethnic groups listed or none of them. It refers to the sense of belonging that a person expresses in front of a collective according to their identity and forms of interaction in and with the world. History of hypertension or consumption of antihypertensive medication was also recorded. Trained research assistants used a sphygmomanometer (Omron HEM-757) with a 14 × 48 cm cuff to take blood pressure measurements. Each participant was instructed not to consume any food or drink, no cigarettes or alcohol consumption or physical activity in the previous half-hour and was asked to rest in a seated position 5 minutes before having their blood pressure measurements. Blood pressure was taken twice, in the health facility, with participants sitting upright and the right arm supported at heart level, with a 5-minute interval between each measurement. The mean of the 2 measures was used in the analysis. Anthropometric measurements were acquired following the PURE standardized protocol.8 Bodyweight was obtained using a digital balance, ensuring the patient was wearing light clothing, and height was obtained using a measuring tape approximating each measurement to the closest millimeter. The patient was instructed to be barefoot for both measurements. Waist circumference (WC) and hip circumference (HC) were measured with tape over the patient’s skin. Waist-to-hip ratio (WHR) was calculated by dividing WC by HC and body mass index (BMI) by dividing body weight (in kilograms) by the height (in meters) squared. Handgrip strength (HGS) was measured utilizing a Jamar dynamometer (Sammons Preston, Bolingbrook, IL) following the previously described protocol.11 Briefly, participants were asked to stand, holding the dynamometer on their body’s side with their elbow flexed at a 90° angle and then asked to squeeze the instrument as hard as possible for 3 seconds. Each measurement was repeated 3 times after a 30-second resting interval. Participants were then classified in tertiles according to their BMI, WHR, WC, and HGS. The first urine sample of the morning was taken after a night of fasting for determination of sodium and potassium. Samples were frozen at −20°C to then be sent to the laboratory for analysis using standardized methods. We used the Kawasaki formula to estimate 24-hour sodium and potassium urine excretion. This estimate has been validated as a substitute for determining sodium and potassium (Na/K) daily intake.12
Statistical Analysis
A descriptive analysis estimating the measures of central tendency and dispersion was carried out. Categorical variables are presented as frequencies and percentages, and continuous variables are presented as means and standard deviations. The presence of hypertension, defined as a systolic blood pressure (SBP) of 140 mm Hg or more, and/or a diastolic blood pressure (DBP) of 90 mm Hg or more, or a previous diagnosis of hypertension, or the use of antihypertensive medications, was analyzed as a dichotomized categorical variable. Participants were grouped according to their self-identified ethnicity (Whites, Mestizos, Afro-descendant). We developed 3 paths for the statistical analysis. First, we analyzed separately White, Mestizos, and Afro-descendant using univariate and multivariate Poisson regression models with robust standard errors to evaluate the association (prevalence ratios [PR] and 95% confidence intervals [95% CI] between hypertension and sex, age, education level, BMI by tertiles, WC by tertiles, WHR by tertiles, HGW by tertiles (handgrip strength adjusted by body weight, calculating tertiles separately for men and woman), location, income, and solid fuel cooking. Second, we analyzed separately by sex and self-identified ethnicity using univariate and multivariate logistic regression to evaluate the association of hypertension to the same variables described above.
Third, we employed a mediation effects analysis (Baron and Kenny model13) to evaluate the mediation effect of education on the association of hypertension and ethnicity. The variables used for the mediation analysis were hypertension (binary response with 2 levels: yes, no), self-identified ethnicity (binary direct effect with 2 levels: Afro-descendant and a composite of Mestizo and White), education (mediator effect, with 3 levels: none, primary; secondary school; and university), and using solid fuel cooking, sex, age, and income as covariates. R statistical software version 3.6 was employed for the analysis, and the mediation R package version 4.5. Statistical significance was set at P < 0.05 for a 2-tailed test.
RESULTS
The analysis included 3,745 subjects from the PURE Colombia study with well self-identified ethnicity and complete information about socioeconomic factors, of which 63.5% were women and the mean age was 50.8 ± 9.4 years. Overall, 73.3% of the population were defined as Mestizo, 14.7% as White, and 11.9% as Afro-descendant. Table 1 shows descriptive characteristics by self-identified ethnicity group. In the whole population, only 12.6% had a high educational level, been lowest in the Afro-descendant population (7.8%) compared with Whites (26.6%) and Mestizos (10.5%). There were no ethnic differences in cigarette smoking, alcohol consumption, daily calories intake, fat, protein, and Na/K intake. Most Whites lived in urban areas (61.6%), while the majority of Afro-descendant (75.6%) and Mestizos (58.2%) lived in rural areas. Income was different between the self-identified ethnic groups with 81.2% of Afro-descents, 70.8% of Mestizos and 48.4% of Whites having an income lower that 350 USD. In the whole population the mean SBP was 128.6 ± 21.8 mm Hg, and the mean DBP was 80.6 ± 13.1 mm Hg, and the prevalence of hypertension was slightly higher in women (39.4%) than in men (38.5%). The prevalence of hypertension in each ethnic group stratified by sex is shown in Figure 1; a greater prevalence was observed in Afro-descendant compared with Mestizos and Whites.
Table 1.
Baseline characteristics of study participants according to ethnic groups: Whites, Mestizo, and Afro-descendant.
| Characteristics | Overall | White | Mestizo | Afro-descendant |
|---|---|---|---|---|
| Participants, n (%) | 3,745 | 552 (14.7) | 2,746 (73.3) | 447 (11.9) |
| Age, years, mean (SD) | 50.8 (9.40) | 51.8 (9.29) | 50.6 (9.44) | 50.3 (9.23) |
| Sex | ||||
| Men, n (%) | 1,367 (36.5) | 195 (35.3) | 1,016 (37.0) | 156 (34.9) |
| Women, n (%) | 2,378 (63.5) | 357 (64.7) | 1,730 (63.0) | 291 (65.1) |
| Location | ||||
| Urban, n (%) | 1,598 (42.7) | 340 (61.6) | 1,149 (41.8) | 109 (24.4) |
| Rural, n (%) | 2,147 (57.3) | 212 (38.4) | 1,597 (58.2) | 338 (75.6) |
| Blood pressure measurement | ||||
| SBP, mm Hg, mean (SD) | 128.6 (21.8) | 129.7 (22.0) | 127.7 (21.4) | 133.2 (22.9) |
| DBP, mm Hg, mean (SD) | 80.6 (13.1) | 81.2 (11.1) | 80.0 (13.5) | 83.1 (13.0) |
| Hypertension, n (%) | 1,464 (39.1) | 229 (41.5) | 1,031 (37.5) | 204 (45.6) |
| Anthropometric measurements | ||||
| HGS, kg,mean (SD) | 27.16 (10.53) | 26.78 (9.59) | 27.15 (10.71) | 27.67 (10.73) |
| Height, cm, mean (SD) | 158.4 (9.05) | 158.9 (8.98) | 157.8 (9.11) | 161.7 (8.49) |
| Weight, kg, mean (SD) | 66.1 (13.25) | 67.0 (13.56) | 65.6 (13.29) | 68.0 (12.24) |
| BMI, kg/cm2, mean (SD) | 26.35 (4.95) | 26.55 (4.96) | 26.36 (4.98) | 26.06 (4.76) |
| WC, cm, mean (SD) | 86.15 (11.36) | 85.86 (11.74) | 86.15 (11.25) | 86.47 (11.59) |
| HC, cm, mean (SD) | 97.21 (10.13) | 97.28 (9.77) | 96.72 (10.16) | 100.17 (9.88) |
| WHR, mean (SD) | 0.89 (0.09) | 0.88 (0.09) | 0.89 (0.10) | 0.86 (0.08) |
| Education level | ||||
| Low, n (%) | 2,553 (68.2) | 289 (52.4) | 1,955 (71.2) | 309 (69.1) |
| Middle, n (%) | 721 (19.3) | 116 (21.0) | 502 (18.3) | 103 (23.0) |
| High, n (%) | 471 (12.6) | 147 (26.6) | 289 (10.5) | 35 (7.8) |
| Behavioral characteristics | ||||
| Cigarette smoking, n (%) | ||||
| Never | 2,491 (66.5) | 356 (64.5) | 1,811 (66.0) | 324 (72.5) |
| Current | 500 (13.4) | 56 (10.1) | 384 (14.0) | 60 (13.4) |
| Ex-smoker | 754 (20.1) | 140 (25.4) | 551 (20.1) | 63 (14.1) |
| Consumption of alcohol, n (%) | ||||
| Never | 2,082 (55.6) | 335 (60.7) | 1,484 (54.0) | 263 (58.8) |
| Current | 1,102 (29.4) | 145 (26.3) | 808 (29.4) | 149 (33.3) |
| Ex-drinker | 561 (15.0) | 72 (13.0) | 454 (16.5) | 35 (7.8) |
| Daily calories intake, mean (SD) | 2,251 (1034) | 2,219 (925) | 2,269 (1058) | 2,178 (1011) |
| Na/K intake, mean (SD) | 0.72 (0.23) | 0.69 (0.22) | 0.71 (0.22) | 0.77 (0.27) |
| Fat intake, %, mean (SD) | 18.29 (4.84) | 18.80 (4.83) | 18.28 (4.92) | 17.74 (4.32) |
| Protein intake,%, mean (SD) | 15.80 (3.27) | 16.04 (3.01) | 15.50 (3.16) | 17.32 (3.80) |
| Solid fuel cooking n (%) | 1,018 (27.2) | 103 (18.7) | 859 (31.3) | 56 (12.5) |
| Income | ||||
| ≥350 USD, n (%) | 1,170 (31.2) | 285 (51.6) | 801 (29.2) | 84 (18.8) |
| <350USD, n (%) | 2,575 (68.8) | 267 (48.4) | 1,945 (70.8) | 363 (81.2) |
Data are presented as the mean and standard deviation (SD) or as number (n) and percentage (%). Abbreviations: SBP, systolic blood pressure; DBP, diastolic blood pressure; HGS, hand grip strength; BMI, body mass index; WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; Na/K intake, sodium/potassium intake.
Figure 1.
Prevalence of hypertension among the ethnic groups: Mestizo, White, and Afro-descendent categorized by sex.
Table 2 shows the association between hypertension and sociodemographic factors and anthropometric measurements, analyzed separately by self-identified ethnicity group. Based on the results of the multivariate analysis, older age (>50 years) was associated with higher prevalence of hypertension in all self-identified ethnicities. In Mestizos, the upper tertile of BMI presented association with higher prevalence of hypertension than participants in first tertile of BMI. White subjects in the middle tertile of WC exhibited association with higher prevalence of hypertension than participants in lower tertile of WC, while in Mestizos, people on upper tertile of WC showed association with higher prevalence of hypertension than subjects in lower tertile. In Mestizos and Afro-descendants, those with a low education level and Whites with low and middle education levels displayed association for Mestizos and Afro-Descendant respectively; and for middle education level and for low education level in White subjects) with higher prevalence of hypertension than subjects having high education level.
Table 2.
Association between hypertension, sociodemographic factors, and anthropometric measures according to ethnicity
| Univariate* | Multivariate* | |||||
|---|---|---|---|---|---|---|
| White | Mestizo | Afro-descendant | White | Mestizo | Afro-descendant | |
| PR (95% CI) | PR (95% CI) | PR (95% CI) | PR (95% CI) | PR (95% CI) | PR (95% CI) | |
| Sex | ||||||
| Female | Reference | Reference | Reference | Reference | Reference | Reference |
| Male | 1.02 (0.83 to 1.26) | 0.98 (0.88 to 1.08) | 0.93 (0.75 to 1.16) | 0.99 (0.81 to 1.2) | 0.98 (0.89 to 1.07) | 0.92 (0.75 to 1.13) |
| Age | ||||||
| <50 years | Reference | Reference | Reference | Reference | Reference | Reference |
| ≥50 years | 1.98 (1.57 to 2.48) | 2.36 (2.12 to 2.64) | 1.99 (1.6 to 2.47) | 1.91 (1.52 to 2.4) | 2.17 (1.94 to 2.43) | 1.84 (1.46 to 2.32) |
| BMI (kg/m2) | ||||||
| F: (12.9–24.5); M: (16.4–23.2) | Reference | Reference | Reference | Reference | Reference | Reference |
| F: (24.6–28.3); M: (23.3 -26.4) | 1.27 (0.96 to 1.68) | 1.32 (1.15 to 1.51) | 1.27 (0.98 to 1.65) | 1.03 (0.78 to 1.37) | 1.17 (1.01 to 1.35) | 1.09 (0.8 to 1.48) |
| F: (28.4–76.8); M: (26.5–66.6) | 1.69 (1.31 to 2.18) | 1.67 (1.47 to 1.89) | 1.35 (1.05 to 1.74) | 1.34 (0.98 to 1.82) | 1.28 (1.08 to 1.51) | 1.03 (0.7 to 1.51) |
| WC (cm) | ||||||
| F: (47.1–79.5); M: (45–83.0) | Reference | Reference | Reference | Reference | Reference | Reference |
| F: (79.6–89.0); M: (83.1–92.0) | 1.84 (1.38 to 2.45) | 1.28 (1.12 to 1.47) | 1.43 (1.07 to 1.92) | 1.49 (1.06 to 2.08) | 1.08 (0.92 to 1.26) | 1.25 (0.87 to 1.8) |
| F: (89.1–150.7); M: (92.1–138) | 1.93 (1.45 to 2.55) | 1.82 (1.6 to 2.06) | 1.56 (1.19 to 2.05) | 1.34 (0.9 to 2.0) | 1.3 (1.07 to 1.56) | 1.27 (0.8 to 2.03) |
| WHR | ||||||
| F: (0.46–0.83); M: (0.55–0.91) | Reference | Reference | Reference | Reference | Reference | Reference |
| F: (0.84–0.88); M: (0.92–0.96) | 1.54 (1.18 to 2.01) | 1.16 (1.01 to 1.32) | 1.32 (1.04 to 1.68) | 1.1 (0.82 to 1.48) | 0.99 (0.87 to 1.14) | 1.23 (0.93 to 1.62) |
| F: (0.89–2.22); M: (0.97–2.38) | 1.58 (1.22 to 2.05) | 1.5 (1.33 to 1.7) | 1.33 (1.02 to 1.72) | 1.07 (0.78 to 1.45) | 1.01 (0.87 to 1.16) | 0.9 (0.66 to 1.24) |
| HGW | ||||||
| F: (0.40–1.45); M: (0.57–1.61) | Reference | Reference | Reference | Reference | Reference | Reference |
| F: (0.31–0.39); M: (0.45–0.56) | 1.08 (0.83 to 1.4) | 1.1 (0.97 to 1.26) | 1.37 (1.04 to 1.8) | 0.86 (0.67 to 1.09) | 0.91 (0.8 to 1.03) | 1.12 (0.86 to 1.47) |
| F: (0.05–0.30); M: (0.11–0.44) | 1.25 (0.98 to 1.6) | 1.5 (1.33 to 1.69) | 1.55 (1.19 to 2.0) | 0.83 (0.65 to 1.07) | 1.01 (0.89 to 1.14) | 1.18 (0.89 to 1.57) |
| Education level | ||||||
| High | Reference | Reference | Reference | Reference | Reference | Reference |
| Middle | 1.45 (1.07 to 1.96) | 1.0 (0.81 to 1.25) | 1.16 (0.64 to 2.09) | 1.57 (1.16 to 2.11) | 1.11 (0.9 to 1.36) | 1.44 (0.85 to 2.41) |
| Low | 1.34 (1.02 to 1.74) | 1.31 (1.09 to 1.57) | 1.81 (1.06 to 3.09) | 1.62 (1.21 to 2.17) | 1.46 (1.21 to 1.75) | 2.14 (1.3 to 3.53) |
| Location | ||||||
| Urban | Reference | Reference | Reference | Reference | Reference | Reference |
| Rural | 0.8 (0.64 to 0.99) | 0.89 (0.8 to 0.98) | 0.92 (0.73 to 1.15) | 0.95 (0.74 to 1.23) | 1.01 (0.9 to 1.13) | 0.84 (0.66 to 1.07) |
| Income | ||||||
| ≥350 USD | Reference | Reference | Reference | Reference | Reference | Reference |
| <350 USD | 0.94 (0.77 to 1.14) | 0.87 (0.79 to 0.97) | 0.95 (0.74 to 1.22) | 1.0 (0.79 to 1.27) | 0.84 (0.75 to 0.94) | 0.95 (0.72 to 1.26) |
| Solid fuel cooking | ||||||
| No | Reference | Reference | Reference | Reference | Reference | Reference |
| Yes | 0.87 (0.78 to 0.97) | 0.76 (0.53 to 1.1) | 0.6 (0.41 to 0.86) | 0.93 (0.82 to 1.05) | 0.85 (0.58 to 1.24) | 0.87 (0.78 to 0.97) |
*Prevalence ratios and 95% CI from the univariate and multivariate using Poisson regression models with robust standard errors. In multivariate analysis prevalence ratios and 95% CI were adjusted by age, location, income, home cooking with solid fuels, body mass index, waist circumference, waist-to-hip ratio and handgrip adjusted by bodyweight. Bold values are statistically significant. Abbreviations: BMI, body mass index; WC, waist circumference, WHR, waist-to-hip ratio; HGW, hand grip strength adjusted by weight; F , female; M, male; PR, prevalence ratio; 95% CI, 95% confidence interval.
In Mestizos, those with a monthly income below 350 USD had significantly lower prevalence of hypertension compared with participants with monthly income above 350 USD, while in Afro-descendant participants, cooking using solid fuel sources was associated with lower prevalence of hypertension compared with cooking using other fuel sources.
Table 3 shows the results of the multivariate logistic regression models fitted separately by sex and self-identified ethnicity for the association of education effects with hypertension, adjusted by the covariates described above. We found that in female and male Whites and Mestizos, and female Afro-descendants low education level was associated with high prevalence of hypertension compared with subjects with high education level, while male Afro-descendant did not show that association.
Table 3.
Association between education level and hypertension
| Female White | Female Mestizo | Female Afro-descendant | |
|---|---|---|---|
| PR (95% CI) | PR (95% CI) | PR (95% CI) | |
| Education level | |||
| High | Reference | ||
| Middle | 1.53 (1.03 to 2.26) | 1.13 (0.85 to 1.49) | 1.87 (0.94 to 3.72) |
| Low | 1.49 (1.03 to 2.14) | 1.5 (1.16 to 1.93) | 2.82 (1.44 to 5.53) |
| Education level | Male White | Male Mestizo | Male Afro-descendant |
| PR (95% CI) | PR (95% CI) | PR (95% CI) | |
| High | Reference | ||
| Middle | 1.76 (1.14 to 2.74) | 1.06 (0.78 to 1.45) | 0.82 (0.34 to 1.95) |
| Low | 1.73 (1.03 to 2.91) | 1.35 (1.03 to 1.76) | 1.33 (0.6 to 2.94) |
The table shows the association obtained from Poisson regression models with robust standard errors separately by sex and self-recognized ethnicity. PRs and 95% CI were adjusted by age, location, income, home cooking with solid fuels, body mass index, waist circumference, waist-to-hip ratio and handgrip adjusted by bodyweight. Bold values are statistically significant. Abbreviations: PR, prevalence ratio; 95% CI, 95% confidence interval.
Table 4 shows the proportion of the mediator effect obtained from the mediation analysis results for the association of hypertension with self-identified ethnicity mediated by education, and adjusted by age, sex, cooking with solid fuels, and income. Model 1 did not reveal a mediator effect of education level alone, but when home polluting cooking fuels was included in the mediation model (Model 2), education level exhibited an important proportion of mediation effect (0.176, [95% CI: 0.059; 0.714]). We observed the same effect on the proportion of the mediator effect when sex (Model 3) (0.192, [95% CI: 0.063; 0.724]) and age (Model 4) (0.135, [95% CI: 0.046; 0.359]) were included, but the mediation effect of education level disappeared when income was included as a covariate (Model 5) (−0.022, [95% CI: −0.135; 0.081]). These results could be explained by the strong association existing between the income variable and cooking with solid fuels (chi-square test of income cross-tabulated with polluting cooking fuels, P < 0.001), making the model estimates and results for the mediation effect of education unstable when both variables (income and cooking with solid fuels) are included in the mediation model. Model 6 replaces the variable of cooking using solid fuel by the income variable, but the proportion of the mediator effect of education level was not significant but close to the rejection threshold (P = 0.062), supporting the idea that income is not only associated with cooking with solid fuels, but also exerts less influence on the mediator effect of education level on the association of hypertension and self-identified ethnicity.
Table 4.
Association between self-recognized ethnicity and hypertension mediated by education level.
| Model | Proportion of the mediator effect estimate (95% CI) | p-value |
|---|---|---|
| Model 1 | 0.038 (−0.019 to 0.144) | 0.150 |
| Model 2 | 0.176 (0.059 to 0.714) | 0.014 |
| Model 3 | 0.192 (0.063 to 0.724) | 0.018 |
| Model 4 | 0.135 (0.046 to 0.359) | <0.001 |
| Model 5 | −0.022 (−0.135 to 0.081) | 0.798 |
| Model 6 | −0.054 (−0.232 to 0.003) | 0.062 |
Bold values are statistically significant.
Mediation analysis adjusted for selected covariates.
Model 1: Response variable: hypertension; direct effect: ethnicity; the mediator effect: education.
Model 2: Model 1 plus coal cooking as covariate.
Model 3: Model 2 plus sex as covariate.
Model 4: Model 3 plus age as covariate.
Model 5: Model 4 plus income as covariate.
Model 6: Model 1 plus sex, age and income as covariates.
DISCUSSION
In this study, we describe hypertension prevalence in a sample of Colombian subjects and establish the association with sociodemographic, anthropometric, and metabolic variables. We found that the association between self-identified ethnicity (categorized as White, Mestizo, and Afro-descendant) with hypertension rates is mediated by education level and influenced by socioeconomic disparities such as income and cooking using solid fuel sources. In this adult Colombian sample with a mean age of 50 years, the overall prevalence of hypertension was 39.2% and was slightly higher in women (39.7%) than in men (38.4%). Prevalence was higher in Afro-descendants (45.6%) than in Mestizos (37.5%) and higher than in Whites (41.5%). These differences were due to the higher prevalence observed in the Afro-descendant women, since in men differences between the ethnic groups were not evident. A previous study in a Colombian population aged over 60 years showed that the prevalence of hypertension was 57.7%, an expected difference given the older age of that study in comparison with the present population. They used the color palette tool to identify ethnicity and did not observe differences in prevalence by skin color but did find a higher prevalence in women compared with men, particularly in dark-skinned women.14 The different methods used to classify the ethnic groups, to identify the hypertensive subjects (in Barrera, year 95% was by interview) and difference in the mean age of the sample may explain our contrasting findings.14 Although population studies such as the NHANES3–5 and the Center for Disease Control and Prevention (CDC) database in USA show a higher prevalence of hypertension in Afro-Americans than in other ethnicities,15 ethnicity has not been consistently identified as a risk factor for hypertension in other populations. For example, a study in Cuba16 reported a similar prevalence of hypertension in populations classified as White or Afro-descendants, while in a study in Puerto Rico, Afro-descendants had higher prevalence of hypertension compared with Whites.17 In Brazil, a study using self-report ethnicity observed a higher prevalence among the Afro-descendants, followed by Whites and Mixed race,18 a similar pattern to the present study. A recent report in American adults showed that adult Afro-descendants have a higher prevalence of hypertension (45.3% versus 31.4%, adjusted OR: 2.24 [95% CI: 1.97–2.56]), while Hispanics have a similar prevalence as Whites.19 Taken together, these results suggest that other factors may mediate hypertension risk in different ethnic groups.
The presence of hypertension in the Colombian population was associated with known risk factors such as older age, increased BMI, increased WC, and greater WHR, with WHR showing a stronger association than BMI in all ethnic groups.9 Low education was associated with hypertension in all ethnic groups, and particularly in Afro-descendants in whom it was the modifiable risk factor with the strongest association. Notably, 70% of Afro-descendants had a low level of education compared with 52% of Whites. While 26.6% of Whites attended university, only 7.8% of Afro-descendants did so. Afro-descendants represent 10.6% of the population and have a lower socioeconomic level compared with the remainder of the population. According to data from the Colombian National Survey (DANE 2019), the multidimensional poverty index for the Afro-descendant population was 30.6% compared with 19.6% for the whole population.20 A higher poverty index is associated with higher unemployment, informal work, lower access to the health system a lower educational level, and a higher educational gap. Differences in educational level can be an accurate indicator of social inequality21 and as we previously reported, social inequality is an important risk factor for hypertension.9 Therefore, the higher prevalence of hypertension in Colombia’s Afro-descendant population might be explained by social inequality reflected in their lower mean educational level and use of solid fuels in cooking. The mediation analysis showed that hypertension is associated with ethnicity, and that this association is mediated by education level adjusted by socioeconomic factors such as coal cooking and income, sex, and age. Supporting this view, we found that the previously described22 association between low handgrip strength and hypertension, was evident in Afro-descendant and Mestizos but not in Whites, an ethnicity with higher educational level and socioeconomic status. This suggests that the ethnic differences in the association between HGS and hypertension may also be related to social inequality. Few studies have examined ethnic and socioeconomic differences in handgrip strength and its association with cardiovascular risk factors.23–26 In British adults, Ntuk et al.27 reported ethnic differences in the association between HGS and the risk of diabetes, observing similar mean HGS in Afro-descendants as Whites, but a higher attributable risk for diabetes associated with low HGS in the Afro-descendants males. In a cross-sectional study comparing anthropometric and cardiometabolic risk factors in Chilean schoolchildren from 2 different ethnicities (Native Andean Mapuches and White European descendants), Mapuche children had higher blood pressure, lower adiposity levels and lower levels of HGS,28 results that suggest that the interaction between adiposity, muscle strength and blood pressure may also vary between ethnic groups.11
In the present study, we did not observe ethnic differences in food intake or in the nutritional parameters analyzed. The REGARDS cohort29 that included 30,239 participants with a 9-year follow-up found that Afro-descendants had a greater incidence of hypertension, which was partially explained by a higher salt consumption, a higher Na/K intake, and higher processed food consumption. This discrepancy may be related to differences between the REGARDS study population and our population in the amount and quality of food consumption.
Our study has some limitations. The cross-sectional nature of our study prevents us from establishing causality in the associations we observed. Moreover, 65% of the included population were women, limiting the representativeness of the sample. However, we identified ethnic differences in the association between specific risk factors and the prevalence of hypertension y new information which may guide the development of priorities within programs aimed at reducing the prevalence of hypertension. In addition, the mediation analysis showed that education level plays a direct role in the differences in the prevalence of hypertension between ethnic groups. Educational level has been demonstrated as a reliable marker of socioeconomic status and social inequality, better than self-reported income level.20 Therefore, addressing social inequities and increasing access and quality to education in low- and middle-income countries like Colombia is critical to improve population health status.
ACKNOWLEDGMENTS
To Paul A Camacho; Gregorio Sanchez-Vallejo; Edgar Arcos; Claudia Narvaez; Henry Garcia; Dora I. Molina; Carlos Cure; Aristides Sotomayor; Alvaro Rico; Eric Hernandez-Triana; Myriam Duran; Fresia Cotes by the help with the development of study in the Departments of Colombia.
FUNDING
The main PURE study is funded by the Population Health Research Institute, the Canadian Institutes of Health Research and the Heart and Stroke Foundation of Ontario. In Colombia, the study had a partial financial support of COLCIENCIAS Grants 6566-04-18062 and 6517-777-58228.
DISCLOSURE
The authors declared no conflict of interest.
REFERENCES
- 1. Forouzanfar MH, Liu P, Roth GA, Ng M, Biryukov S, Marczak L, Alexander L, Estep K, Abate KH, Akinyemiju TF, Ali R, Alvis-Guzman N, Azzopardi P, Banerjee A, Bärnighausen T, Basu A, Bekele T, Bennett DA, Biadgilign S, Catalá-López F, Feigin VL, Fernandes JC, Fischer F, Gebru AA, Gona P, Gupta R, Hankey GH, Jonas JB, Judd SE, Khang Y-H, Khosravi A, Kim YJ, Kimokoti RW, Kokubo Y, Kolte D, Lopez A, Lotufo PA, Malekzadeh R, Melaku YA, Mensah GA, Misganaw A, Mokdad AH, Moran AE, Nawaz H, Neal B, Ngalesoni FN, Ohkubo T, Pourmalek F, Rafay A, Rai RK, Rojas-Rueda D, Sampson UK, Santos IS, Sawhney M, Schutte AE, Sepanlou SG, Shifa GT, Shiue I, Tedla BA, Thrift AG, Tonelli M, Truelsen T, Tsilimparis N, Ukwaja KN, Uthman OA, Vasankari T, Venketasubramanian N, Vlassov VV, Vos T, Westerman R, Yan LL, Yano Y, Yonemoto N, El Sayed Zaki M, Murray CJL. Global burden of hypertension and systolic blood pressure of at least 110 to 115 mm Hg, 1990-2015. JAMA 2017; 317:165–182. [DOI] [PubMed] [Google Scholar]
- 2. Chow CK, Teo KK, Rangarajan S, Islam S, Gupta R, Avezum A, Bahonar A, Chifamba J, Dagenais G, Diaz R, Kazmi K, Lanas F, Wei L, Lopez-Jaramillo P, Fanghong L, Ismail NH, Puoane T, Rosengren A, Szuba A, Temizhan A, Wielgosz A, Yusuf R, Yusufali A, McKee M, Liu L, Mony P, Yusuf S. Prevalence, awareness, treatment, and control of hypertension in rural and urban communities in high-, middle-, and low-income countries. JAMA 2013; 310:959–968. [DOI] [PubMed] [Google Scholar]
- 3. Dorans KS, Mills KT, Liu Y, He J. Trends in prevalence and control of hypertension according to the 2017 American College of Cardiology/American Heart Association (ACC/AHA) guideline. J Am Heart Assoc 2018; 7:e008888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Whelton PK, Einhorn PT, Muntner P, Appel LJ, Cushman WC, Diez Roux AV, Ferdinand KC, Rahman M, Taylor HA, Ard J, Arnett DK, Carter BL, Davis BR, Freedman BI, Cooper LA, Cooper R, Desvigne-Nickens P, Gavini N, Go AS, Hyman DJ, Kimmel PL, Margolis KL, Miller ER 3rd, Mills KT, Mensah GA, Navar AM, Ogedegbe G, Rakotz MK, Thomas G, Tobin JN, Wrigth JT, Yoon SSS, Cutler JA. Research needs to improve hypertension treatment and control in African Americans. Hypertension 2016; 68:1066–1072. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Yoon SS, Gu Q, Nwankwo T, Wright JD, Hong Y, Burt V. Trends in blood pressure among adults with hypertension: United States, 2003 to 2012. Hypertension 2015; 65:54–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Munoz AM, Velasquez CM, Bedoya G. Cardio-metabolic parameters are associated with genetic admixture estimates in a pediatric population from Colombia. BMC Genet 2016;17(1):93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. DANE. Resultados del Censo Nacional de Población y Vivienda 2018. 2019. https://www.dane.gov.co/index.php/estadisticas-por-tema/demografia-y-poblacion/censo-nacional-de-poblacion-y-vivenda-2018/cuantos-somos.
- 8. Teo K, Chow CK, Vaz M, Rangarajan S, Yusuf S, PURE Investigators-Writing Group. The Prospective Urban Rural Epidemiology (PURE) study: examining the impact of societal influences on chronic noncommunicable diseases in low-, middle-, and high-income countries. Am Heart J 2009; 158:1–7.e1. [DOI] [PubMed] [Google Scholar]
- 9. Camacho PA, Gomez-Arbelaez D, Molina DI, Sanchez G, Arcos E, Narvaez C, Garcia H, Perez M, Hernandez EA, Duran M, Cure C, Sotomayor A, Rico A, David TM, Cohen DD, Rangarajan S, Yusuf S, Lopez-Jaramillo P. Social disparities explain differences in hypertension prevalence, detection and control in Colombia. J Hypertens 2016; 34:2344–2352. [DOI] [PubMed] [Google Scholar]
- 10. Dehghan M, Lopez-Jaramillo P, Duenas R, Anaya LL, Garcia RG, Zhang X, Islam S, Merchant AT. Development and validation of a quantitative food frequency questionnaire among rural- and urban-dwelling adults in Colombia. J Nutr Educ Behav 2012; 44:609–613. [DOI] [PubMed] [Google Scholar]
- 11. Leong DP, Teo KK, Rangarajan S, Lopez-Jaramillo P, Avezum A, Orlandini A, Seron P, Ahmed SH, Rosengren A, Kelishadi R, Rahman O, Swaminathan S, Iqbal R, Gupta R, Lear SA, Oguz A, Yusoff K, Zatonska K, Chifamba J, Igumbor E, Mohan V, Anjani RM, Gu H, Li W, Yusuf S. Prognostic value of grip strength: findings from the Prospective Urban Rural Epidemiology (PURE) study. Lancet 2015; 386:266–73. [DOI] [PubMed] [Google Scholar]
- 12. Mente A, O’Donnell MJ, Dagenais G, Wielgosz A, Lear SA, McQueen MJ Jiang Y, Xingyu W, Jian B, Calik KBT, Akalin AA, Mony P, Devanath A, Yusufali AH, Lopez-Jaramillo P, Avezum A, Yusoff K, Rosengren A, Kruger L, Orlandini A, Rangarajan S, Teo K, Yusuf S. Validation and comparison of three formulae to estimate sodium and potassium excretion from a single morning fasting urine compared to 24-h measures in 11 countries. J Hypertens 2014; 32:1005–1014. [DOI] [PubMed] [Google Scholar]
- 13. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol 1986; 51:1173–1182. [DOI] [PubMed] [Google Scholar]
- 14. Barrera L, Gomez F, Ortega-Lenis D, Ojeda JC, Mendez F. Prevalence, awareness, treatment and control of high blood pressure in the elderly according to the ethnic group. Colombian survey. Colomb Med (Cali) 2019; 50:115–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Center for Disease Control and Prevention. Hypertension Among Adults Aged 20 and Over, by Selected Characteristics: United States, Selected Years 1988–1994 Through 2013–2016. 2018; ( 3): 1–2 pp. https://www.cdc.gov/nchs/data/hus/2018/022.pdf
- 16. Ordúñez P, Kaufman JS, Benet M, Morejon A, Silva LC, Shoham DA, Cooper RS. Blacks and whites in the Cuba have equal prevalence of hypertension: confirmation from a new population survey. BMC Public Health. 2013;13:169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Gravlee CC, Dressler WW, Bernard HR. Skin color, social classification, and blood pressure in southeastern Puerto Rico. Am J Public Health. 2005;95(12):2191–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Juliao NA, de Souza A, Guimaraes RRM. Trends in the prevalence of systemic arterial hypertension and health care service use in Brazil over a decade (2008–2019). Cien Saude Colet. 2021;26(9):4007–4019. [DOI] [PubMed] [Google Scholar]
- 19. Aggarwal R, Chiu N, Moran AE, Raber I, Shen C, Yeh R, Kazi DS. Racial/ethnic disparities in hypertension prevalence, awareness, treatment, and control in the United States, 2013 to 2018. Hypertension 2021; 78:1719–1726. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. DANE. Población negra, afrocolombiana, raizal y palenquera: resultados del censo nacional de población y vivienda 2018. 2019; ( 69). https://www.dane.gov.co/files/investigaciones/boletines/grupos-etnicos/presentacion-grupos-etnicos-poblacion-NARP-2019.pdf.
- 21. Gamboa L, Londoño E. Assessing Educational Unfair Inequalities at a Regional Level in Colombia. Lecturas Econ 2015;23:97–133. [Google Scholar]
- 22. Mainous AG, 3rd, Tanner RJ, Anton SD, Jo A. Grip strength as a marker of hypertension and diabetes in healthy weight adults. Am J Prev Med 2015;49(6):850–858. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Thorpe RJ, Simonsick E, Zonderman A, Evans MK. Association between race, household income and grip strength in middle- and older-aged adults. Ethn Dis 2016; 26:493–500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Leong DP, Teo KK, Rangarajan S, Kutty VR, Lanas F, Hui C, Quanyong X, Zhenzhen Q, Jinhua T, Noorhassim I, AlHabib KF, Moss SJ, Rosengren A, Akalin AA, Rahman O, Chifamba J, Orlandini A, Kumar R, Yeates K, Gupta R, Yusufali A, Dans A, Avezum A, Lopez-Jaramillo P, Poirier P, Heidari H, Zatonska K, Iqbal R, Khatib R, Yusuf S. Reference ranges of handgrip strength from 125,462 healthy adults in 21 countries: a prospective urban rural epidemiologic (PURE) study. J Cachexia Sarcopenia Muscle 2016; 7:535–546. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Otero J, Cohen DD, Herrera VM, Camacho PA, Bernal O, Lopez-Jaramillo P. Sociodemographic factors related to handgrip strength in children and adolescents in a middle-income country: the SALUS study. Am J Hum Biol 2017; 29: 1–10. [DOI] [PubMed] [Google Scholar]
- 26. Griffin FR, Mode NA, Ejiogu N, Zonderman AB, Evans MK. Frailty in a racially and socioeconomically diverse sample of middle-aged Americans in Baltimore. PLoS One 2018; 13:e0195637. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ntuk UE, Celis-Morales CA, Mackay DF, Sattar N, Pell JP, Gill JMR. Association between grip strength and diabetes prevalence in black, South-Asian, and white European ethnic groups: a cross-sectional analysis of 418 656 participants in the UK Biobank study. Diabet Med 2017; 34:1120–1128. [DOI] [PubMed] [Google Scholar]
- 28. Alvarez C, Flores-Opazo M, Mancilla R, Martinez-Salazar C, Mangiamarchi P, Sade-Calles F, Ramírez-Campillo R. Gender differences in blood pressure and body composition in schoolchildren ascendants from Amerindian and European. Ethn Health 2019; 1:12. [DOI] [PubMed] [Google Scholar]
- 29. Howard G, Cushman M, Moy CS, Oparil S, Muntner P, Lackland DT, Manly JJ, Flaherty ML, Judd SE, Wadley VG, Long DL, Howard VJ. Association of clinical and social factors with excess hypertension risk in black compared with white US Adults. JAMA 2018; 320:1338–1348. [DOI] [PMC free article] [PubMed] [Google Scholar]