Dietary supplements : trends, demographics and mortality among users

Abstract

Dietary supplements have received growing attention in society and sales numbers indicate increased use. Still, time trends in prevalence of users have been unknown. In order to accurately explore associations between diet and diseases, valid assessment methods for micronutrient intake from foods and dietary supplements are needed.

The prevalence of dietary supplement users in Sweden, estimated in three nationally representative surveys, conducted during the last decades, increased by 70% and was 22% and 33% for men and women, respectively, in 1997. Use of natural remedies increased threefold. The increase was observed in all age groups, all socio-economic groups, and all parts of Sweden. The use of these preparations was associated with several sociodemographic and health behaviour factors. The best predictors of use were age, sex and subjective health. Women and older individuals were more likely to be users as well as people reporting a poor health status. Subjects who were underweight and subjects reporting moderate to high levels of exercise were also more often users.

To evaluate if a self-reported food frequency questionnaire (FFQ) can measure use of dietary supplements and the total micronutrient intake, sensitivity, specificity and validity was compared to fourteen 24-hour recalls (collected monthly by telephone) in 248 men (40-75 years old) in a random sample from central Sweden. The sensitivity of the FFQ for measuring use of dietary supplements was 78% and the specificity 93%. Validity for micronutrient intake (Spearman correlation coefficients) increased by 13% (from 0.49 to 0.62), when nutrients from supplements were included. The classification of micronutrient intake from the FFQ into corresponding quintile in the 24-hour recalls was also improved from 37% to 5 1 % when supplements were included. The average reproducibility (intraclass correlation coefficients) for total nutrient intake was 0.61 between two FFQs completed one year apart. The observed increase in the validity of micronutrient estimates due to inclusion of supplements has important implications for epidemiological studies.

In a population-based prospective cohort of 48,645 Swedish men, we investigated the association between use of dietary supplement and all-cause, cardiovascular and cancer mortality. By linking the cohort to the Death Register and the Register of Causes of Death, 2499 deaths (all cause) was ascertained during 246,522 person-years of follow-up, while 540 deaths occurred due to cardiovascular diseases and 398 deaths due to cancer during 153,796 person-years of follow-up for the specific cause mortality. In the multivariate adjusted model, there was no association observed between use of dietary supplements and all-cause mortality, cardiovascular or cancer mortality. However, both confounding and reversed causality seemed to be important issues to consider in this type of study. Independent of adjusting for several lifestyle factors, there was still an adverse association among former smokers, between regular use of dietary supplements and all-cause mortality and a suggested association among current smokers.

In conclusion, the use of dietary supplements is associated with several lifestyle factors and the prevalence of users has increased during the last decades. Thus, use of dietary supplements has to be considered in nutritional surveys and when assessing associations between diet and diseases. The adverse association found between use of dietary supplements and mortality among former/current smokers needs further evaluation.