To examine the socio-economic and lifestyle determinants of cooking oil choice in Costa Rica during the last decade (1994–2004).

Cross-sectional study. Subjects (total n = 2274) belonged to the control population of a large case–control study; they were recruited yearly. Data about type of oil used for cooking, dietary intake, socio-economic and demographic characteristics were collected.

A dietitian visited all subjects and conducted the interviews at their homes; all subjects lived in the Costa Rican central valley region.

Adult, free-living, rural and urban Costa Ricans with no history of myocardial infarction and physical or mental disability.

The odds of choosing soybean over palm oil increased significantly each year (P < 0.05) and was determined by high socio-economic status (SES) and variables that suggest health awareness (self-reported history of hypertension, high cholesterol, multivitamin use and intake of green leafy vegetables). The odds of choosing other unsaturated oils, namely corn and sunflower, over soybean oil also increased yearly (P < 0.05) and was associated with the same two factors (high SES and health awareness). Palm oil users remained in the lowest SES tertile and were more likely to live in rural areas. Across all SES tertiles, high health awareness determined the odds of choosing other unsaturated oils over palm oil, and soybean oil (P < 0.05).

These data show that, in addition to SES, health awareness is associated with the selection of unsaturated oils over palm oil in a developing country undergoing transition. These data should be considered when targeting nutrition messages and policies that promote better dietary choices.

Stature is a powerful indicator of poor nutrition early in life in nations where undernutrition is a public health problem. Hypertension in adults has been associated with factors present early in life such as low birth weight. We tested the hypothesis that short stature is associated with hypertension among adults.

A household survey of representative adults in Rio de Janeiro city, Brazil was carried out in 1996.

Blood pressure and anthropometric measures were collected from 2802 adults in their own households. Prevalence estimates and modelling incorporated the sample design and weights.

Age-adjusted prevalence of hypertension for both sexes was lower in the third quartile of stature distribution. In women, but not in men, the odds ratio comparing the first quartile of stature with the fourth quartile was statistically significant with an odds ratio of 1.68 (95% CI 1.02–2.76). Adjusting for known risk factors for hypertension such as age, income, smoking, sodium and alcohol intake and race, the association among women, comparing the first with the fourth quartile for stature, was 1.84 (95% CI 1.03–3.30). With further adjustment for residual of weight on height the ratio reduced to 1.76 (95% CI 0.97–3.19, P value of trend = 0.03). Systolic blood pressure showed a U-shaped association with quartiles of stature, mainly among women, with a β-coefficient significantly lower at the third quartile.

This association of stature with hypertension supports the theory of an important ontogenetic dependence of adult blood pressure, at least among women.