Physical Activity and Obesity Risk in Adults in Colombia: The Prospective Urban Rural Epidemiology (PURE) Study

PURPOSE

To investigate longitudinal associations between physical activity levels and obesity in adults in Colombia, where participation in large amounts of light-intensity physical activity is a necessity for many people.

METHODS

Participation in moderate- and vigorous-intensity physical activity was assessed from 2005 to 2009 and obesity was assessed from 2011 to 2019 in men and women from the Prospective Urban Rural Epidemiology (PURE) study. Total physical activity level was categorized as low (<600 MET-min/week), medium (600-3,000 MET-min/week), or high (>3,000 MET-min/week) (600 MET-min/week is equivalent to 150 minutes of moderate activity or 75 minutes of vigorous activity per week). Obesity was defined as body mass index ≥30 kg/m2. Analyses were adjusted for age, sex, smoking, socioeconomic status, diet, alcohol, sedentary time, and sleep.

RESULTS

The main analysis included 3,086 men and women aged 51 ± 9 years at baseline (mean ± SD). Compared to the low physical activity group, the odds ratio (95% confidence interval) for obesity was 0.67 (0.53, 0.85) in the medium physical activity group, and 0.78 (0.62, 0.98) in the high physical activity group after adjustment for potential confounders. Smoking is probably a major confounder and it is noteworthy that similar associations were observed in participants who reported never smoking.

CONCLUSIONS

The PURE study is the only prospective cohort study in Colombia. The present analysis is important because it suggests that even the busy people of Colombia could substantially reduce their risk of obesity by participating in moderate- and vigorous-intensity physical activity.

Sex Differences in Cardiovascular Disease Risk Factor Prevalence, Morbidity, and Mortality in Colombia: Findings from the Prospective Urban Rural Epidemiology (PURE) Study

Background

Controversies exist on whether the presence of cardiovascular risk factors and their association with major cardiovascular events (MACE) is different between men and women. Most of the evidence comes from high-income countries, hindering extrapolation of sociocultural and demographic factors of other regions.

Objective

To evaluate sex differences in the prevalence of cardiovascular risk factors and the incidence of MACE and diabetes in Colombian adults.

Methods

We performed a survival analysis from women and men aged 35-70 belonging to the Prospective Urban Rural Epidemiology-Colombia prospective study. Incidence rates for MACE composite (myocardial infarction, stroke, heart failure, death) and each outcome and diabetes were calculated. Kaplan-Meier curves and log-rank tests were performed. The association between demographic, behavioral, and metabolic variables with MACE and diabetes were evaluated with Cox proportional hazards models.

Results

7,552 participants (50±9.7 years) were included; 64% were women. Women had higher hypertension prevalence, body mass index, levels of total cholesterol, LDL-c, and HDL-c but lower triglycerides levels. Women were more sedentary but fewer smokers or active alcohol consumers and had higher educational levels. After 12-year mean follow-up (SD 2.3), the incidence rate of MACE composite was higher in men [4.2 (3.6-4.9) vs. 3.2 (2.8-3.7) cases per 1000 person-years]. Diabetes had the greatest association with MACE (HR = 2.63 95%CI:1.85;3.76), followed by hypertension (HR = 1.75 95%CI:1.30;2.35), low relative grip strength (HR = 1.53 95%CI:1.15;2.02), smoking (HR = 1.47 95%CI: 1.11;1.93), low physical activity (HR = 1.42 95%CI: 1.03;1.96). When evaluating risk factors by sex, only an increased waist-to-hip ratio was more strongly associated with MACE in men (p-interaction <0.05).

Conclusions

The composite MACE outcome was higher in men despite having a lower overall burden of risk factors. The risk factors contribution was similar, leading us to reconsider the need to carrying out differentiated cardiovascular risk prevention and management campaigns, at least in our region.

Waist circumference cut-off points to identify major cardiovascular events and incident diabetes in Latin America: findings from the prospective Urban rural epidemiology study Colombia

Background

Abdominal obesity (AO) indirectly represents visceral adiposity and can be assessed by waist circumference (WC) measurement. In Latin America, cut-off points for the diagnosis of AO are based on Asian population data. We aim to establish the WC cut-off points to predict major cardiovascular events (MACE) and incident diabetes.

Methods

We analyzed data from the cohort PURE study in Colombia. WC cut-off points were defined according to the maximum Youden index. Multivariate logistic regression was used to obtain associations between WC and MACE, diabetes, and cumulative incidence of outcomes visualized using Kaplan-Meier curves.

Results

After a mean follow-up of 12 years, 6,580 individuals with a mean age of 50.7 ± 9.7 years were included; 64.2% were women, and 53.5% were from rural areas. The mean WC was 85.2 ± 11.6 cm and 88.3 ± 11.1 cm in women and men, respectively. There were 635 cases of the MACE composite plus incident diabetes (5.25 events per 1,000 person-years). Using a cut-off value of 88.85 cm in men (sensitivity = 0.565) and 85.65 cm in women (sensitivity = 0.558) resulted in the highest value for the prediction of the main outcome. These values were associated with a 1.76 and 1.41-fold increased risk of presenting the composite outcome in men and women, respectively.

Conclusions

We defined WC cut-off points of 89 cm in men and 86 cm in women to identify the elevated risk of MACE and incident diabetes. Therefore, we suggest using these values in cardiovascular risk assessment in Latin America.

Personal and household PM2.5 and black carbon exposure measures and respiratory symptoms in 8 low- and middle-income countries

BACKGROUND

Household air pollution (HAP) from cooking with solid fuels has been associated with adverse respiratory effects, but most studies use surveys of fuel use to define HAP exposure, rather than on actual air pollution exposure measurements.

OBJECTIVE

To examine associations between household and personal fine particulate matter (PM_2.5) and black carbon (BC) measures and respiratory symptoms.

METHODS

As part of the Prospective Urban and Rural Epidemiology Air Pollution study, we analyzed 48-h household and personal PM_2.5 and BC measurements for 870 individuals using different cooking fuels from 62 communities in 8 countries (Bangladesh, Chile, China, Colombia, India, Pakistan, Tanzania, and Zimbabwe). Self-reported respiratory symptoms were collected after monitoring. Associations between PM_2.5 and BC exposures and respiratory symptoms were examined using logistic regression models, controlling for individual, household, and community covariates.

RESULTS

The median (interquartile range) of household and personal PM_2.5 was 73.5 (119.1) and 65.3 (91.5) μg/m³, and for household and personal BC was 3.4 (8.3) and 2.5 (4.9) x10^-5 m^-1, respectively. We observed associations between household PM_2.5 and wheeze (OR: 1.25; 95%CI: 1.07, 1.46), cough (OR: 1.22; 95%CI: 1.06, 1.39), and sputum (OR: 1.26; 95%CI: 1.10, 1.44), as well as exposure to household BC and wheeze (OR: 1.20; 95%CI: 1.03, 1.39) and sputum (OR: 1.20; 95%CI: 1.05, 1.36), per IQR increase. We observed associations between personal PM_2.5 and wheeze (OR: 1.23; 95%CI: 1.00, 1.50) and sputum (OR: 1.19; 95%CI: 1.00, 1.41). For household PM_2.5 and BC, associations were generally stronger for females compared to males. Models using an indicator variable of solid versus clean fuels resulted in larger OR estimates with less precision.

CONCLUSIONS

We used measurements of household and personal air pollution for individuals using different cooking fuels and documented strong associations with respiratory symptoms.

Measuring and predicting personal and household Black Carbon levels from 88 communities in eight countries

Black Carbon (BC) is an important component of household air pollution (HAP) in low- and middle- income countries (LMICs), but levels and drivers of exposure are poorly understood. As part of the Prospective Urban and Rural Epidemiological (PURE) study, we analyzed 48-hour BC measurements for 1187 individual and 2242 household samples from 88 communities in 8 LMICs (Bangladesh, Chile, China, Colombia, India, Pakistan, Tanzania, and Zimbabwe). Light absorbance (10^-5 m^-1) of collected PM_2.5 filters, a proxy for BC concentrations, was calculated via an image-based reflectance method. Surveys of household/personal characteristics and behaviors were collected after monitoring. The geometric mean (GM) of personal and household BC measures was 2.4 (3.3) and 3.5 (3.9)·10^-5 m^-1, respectively. The correlation between BC and PM_2.5 was r = 0.76 for personal and r = 0.82 for household measures. A gradient of increasing BC concentrations was observed for cooking fuels: BC increased 53% (95%CI: 30, 79) for coal, 142% (95%CI: 117, 169) for wood, and 190% (95%CI: 149, 238) for other biomass, compared to gas. Each hour of cooking was associated with an increase in household (5%, 95%CI: 3, 7) and personal (5%, 95%CI: 2, 8) BC; having a window in the kitchen was associated with a decrease in household (-38%, 95%CI: -45, -30) and personal (-31%, 95%CI: -44, -15) BC; and cooking on a mud stove, compared to a clean stove, was associated with an increase in household (125%, 95%CI: 96, 160) and personal (117%, 95%CI: 71, 117) BC. Male participants only had slightly lower personal BC (-0.6%, 95%CI: -1, 0.0) compared to females. In multivariate models, we were able to explain 46-60% of household BC variation and 33-54% of personal BC variation. These data and models provide new information on exposure to BC in LMICs, which can be incorporated into future exposure assessments, health research, and policy surrounding HAP and BC.

Ethnic Differences in the Prevalence of Hypertension in Colombia: Association With Education Level

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.

Variations between women and men in risk factors, treatments, cardiovascular disease incidence, and death in 27 high-income, middle-income, and low-income countries (PURE): a prospective cohort study

BACKGROUND

Some studies, mainly from high-income countries (HICs), report that women receive less care (investigations and treatments) for cardiovascular disease than do men and might have a higher risk of death. However, very few studies systematically report risk factors, use of primary or secondary prevention medications, incidence of cardiovascular disease, or death in populations drawn from the community. Given that most cardiovascular disease occurs in low-income and middle-income countries (LMICs), there is a need for comprehensive information comparing treatments and outcomes between women and men in HICs, middle-income countries, and low-income countries from community-based population studies.

METHODS

In the Prospective Urban Rural Epidemiological study (PURE), individuals aged 35-70 years from urban and rural communities in 27 countries were considered for inclusion. We recorded information on participants' sociodemographic characteristics, risk factors, medication use, cardiac investigations, and interventions. 168 490 participants who enrolled in the first two of the three phases of PURE were followed up prospectively for incident cardiovascular disease and death.

FINDINGS

From Jan 6, 2005 to May 6, 2019, 202 072 individuals were recruited to the study. The mean age of women included in the study was 50·8 (SD 9·9) years compared with 51·7 (10) years for men. Participants were followed up for a median of 9·5 (IQR 8·5-10·9) years. Women had a lower cardiovascular disease risk factor burden using two different risk scores (INTERHEART and Framingham). Primary prevention strategies, such as adoption of several healthy lifestyle behaviours and use of proven medicines, were more frequent in women than men. Incidence of cardiovascular disease (4·1 [95% CI 4·0-4·2] for women vs 6·4 [6·2-6·6] for men per 1000 person-years; adjusted hazard ratio [aHR] 0·75 [95% CI 0·72-0·79]) and all-cause death (4·5 [95% CI 4·4-4·7] for women vs 7·4 [7·2-7·7] for men per 1000 person-years; aHR 0·62 [95% CI 0·60-0·65]) were also lower in women. By contrast, secondary prevention treatments, cardiac investigations, and coronary revascularisation were less frequent in women than men with coronary artery disease in all groups of countries. Despite this, women had lower risk of recurrent cardiovascular disease events (20·0 [95% CI 18·2-21·7] versus 27·7 [95% CI 25·6-29·8] per 1000 person-years in men, adjusted hazard ratio 0·73 [95% CI 0·64-0·83]) and women had lower 30-day mortality after a new cardiovascular disease event compared with men (22% in women versus 28% in men; p<0·0001). Differences between women and men in treatments and outcomes were more marked in LMICs with little differences in HICs in those with or without previous cardiovascular disease.

INTERPRETATION

Treatments for cardiovascular disease are more common in women than men in primary prevention, but the reverse is seen in secondary prevention. However, consistently better outcomes are observed in women than in men, both in those with and without previous cardiovascular disease. Improving cardiovascular disease prevention and treatment, especially in LMICs, should be vigorously pursued in both women and men.

FUNDING

Full funding sources are listed at the end of the paper (see Acknowledgments).

The Phospholipid Linoleoylglycerophosphocholine as a Biomarker of Directly Measured Insulin Resistance

BACKGROUND

Plasma concentrations of some lysophospholipids correlate with metabolic alterations in humans, but their potential as biomarkers of insulin resistance (IR) is insufficiently known. We aimed to explore the association between plasma linoleoylglycerophosphocholine (LGPC) and objective measures of IR in adults with different metabolic profiles.

METHODS

We studied 62 men and women, ages 30 to 69 years, (29% normal weight, 59% overweight, 12% obese). Participants underwent a 5-point oral glucose tolerance test (5p-OGTT) from which we calculated multiple indices of IR and insulin secretion. Fifteen participants additionally underwent a hyperinsulinemic-euglycemic clamp for estimation of insulin-stimulated glucose disposal. Plasma LGPC was determined using high performance liquid chromatography/time-of-flight mass spectrometry. Plasma LGPC was compared across quartiles defined by the IR indices.

RESULTS

Mean LGPC was 15.4±7.6 ng/mL in women and 14.1±7.3 ng/mL in men. LGPC did not correlate with body mass in-dex, percent body fat, waist circumference, blood pressure, glycosylated hemoglobin, log-triglycerides, or high density lipoprotein cholesterol. Plasma LGPC concentrations was not systematically associated with any of the studied 5p-OGTT-derived IR indices. However, LGPC exhibited a significant negative correlation with glucose disposal in the clamp (Spearman r=-0.56, P=0.029). Despite not being diabetic, participants with higher plasma LGPC exhibited significantly higher post-challenge plasma glucose excursions in the 5p-OGTT (P trend=0.021 for the increase in glucose area under the curve across quartiles of plasma LGPC).

CONCLUSION

In our sample of Latino adults without known diabetes, LGPC showed potential as a biomarker of IR and impaired glucose metabolism.