PDF
ePub
Citation
Print
INTRODUCTION
Despite nutritional recommendations and food guidelines, the pandemic of chronic disease continues to be a leading cause of mortality [1]. Therefore, it is important to evaluate major modifiable lifestyle risk factors such as diets of individuals or groups in order to prevent chronic diseases [2]. Diet quality indices are a popular topic in nutritional epidemiology. People consume food as a complex mixture of various food products or as a certain type of diet rather than as a single nutrient product. In addition, a strong intercorrelation among some nutrients and foods makes it difficult to detect their isolated effects as a single nutrient or food item [34]. The diet quality index (DQI) was developed as a composite assessment of diet, including measures for eight food groups and nutrient-based recommendations, by Patterson and colleagues in 1994 in order to examine clusters of healthy and unhealthy lifestyle characteristics [5]. The index presents summary scores of an individual's overall dietary intake pattern rather than just one aspect of that diet; therefore, it is useful in estimating variability within a diet and overall dietary quality for individual populations.
A great quantity of research has been done regarding the relationship between individual nutrients and mortality. However, the entire diet must be examined since nutrients are not consumed independently but together within a variety of foods in the diet. Thus, a simple and inexpensive method for monitoring diet quality in various populations must be developed. Dietary quality indices or indicators aim to evaluate overall diet and categorize individuals according to the “healthiness” of their eating behavior [6]. They were developed to assess overall dietary quality and are interpreted in a manner in which high diet quality can predict chronic disease risk [78].
In recent years, research has examined the association of overall diet quality with risk of mortality [91011]. Previous publications have detected significant risk reductions of 10–15% in total mortality based on the healthy diet indicator (HDI) [1213]. In the National Institutes of health-aarp cohorts, higher diet quality based on the healthy eating index (HEI), the alternate healthy eating index (AHEI), and dietary approaches to stop hypertension (DASH) scores has been reported to be associated with a 17–24% reduction in all-cause mortality [14]. In the EPIC- Spain cohort, high diet quality was shown to reduce all-cause mortality by 21% [15]. In a recent meta-analysis consisting of 13 cohort studies, the pooled relative risk estimate for overall 6.5 to 20-year mortality was 0.78, with a variation from 0.76 to 0.80 [16].
The diet pattern of Koreans is becoming more Westernized due to increased consumption of animal products and fat [17]. However, Koreans still have an insufficient energy ratio acquired from fat. A revised diet quality index for Koreans is necessary in order to evaluate dietary intake and nutrition promotion activities for the Korean population.
In this study, we examined the association of all-cause mortality with the diet quality index for Koreans (DQI-K). We hypothesize that low diet quality is associated with increased risk of mortality. This assumption is based on evidence from association studies between individual foods as well as nutrients and mortality.
SUBJECTS AND METHODS
Study design and population
This study was based on a large-scale genomic cohort study known as the Health Examinees (HEXA) study. For the baseline study, a total of 169,718 subjects aged between 40–69 years old were recruited from 38 general hospitals and health examination centers throughout Korea between 2004 and 2013. HEXA data are publicly available for use. The Institutional Review Board of the Seoul National University Hospital, Seoul, Korea approved its use for statistical analysis (IRB No. E-1503-103-657). The HEXA study design has been described in detail in previous studies [1819].
Updated from previously published HEXA studies, the current study used the Health Examinees-Gem (HEXA-G) participant sample, which includes additional eligibility criteria on participating sites (i.e., health examination centers and general hospitals). The HEXA-G criteria selected 21 sites from the 38 original sites based on the following exclusion criteria: pilot study participants in 2004-2006 (eight sites, n = 9,345); unsuitable HEXA biospecimen quality control criteria (eight sites, n = 12,232); participated in the HEXA study for less than 2 years (five sites, n = 8,799). Except for those who did not comply with the HEXA-G criteria, 139,342 participants remained [20]. From the HEXA-G population, we excluded subjects who were missing information in the food frequency questionnaire (FFQ) (n = 2,511) and who reported implausible energy intakes (outside the range of 800–4,000 kcal for men or 500–3,500 kcal for women, n = 2,290). A total of 134,541 subjects remained eligible (Fig. 1).
Mortality data were obtained from Statistics Korea by matching with HEXA-G subjects until December 2015. In total, 99.5% of the HEXA-G participants were successfully linked with mortality data.
Components of diet quality index for Koreans (DQI-K)
For the current study, DQI-K was developed based on the original DQI and a literature review. Among diet quality scores, the HEI, DQI, and HDI are widely used to examine the association between various outcomes and diet quality. The purpose of the HEI is to promote health guidelines, whereas the DQI and HEI were devised to assess health-related outcomes. The DQI-K has adapted the components of these indicators, which are based on American dietary recommendations, to the Korean model. The index is based on Korean dietary recommendations and dietary reference intakes for Koreans. As shown in Table 1, the DQI-K consists of eight components from the original DQI and replaces carbohydrates with whole grains.
The DQI-K is calculated based on eight components: 1) daily protein intake, 2) percent of energy from fat, 3) percent of energy from saturated fat, 4) daily cholesterol intake, 5) daily whole-grain intake, 6) daily fruit intake, 7) daily vegetable intake, and 8) daily sodium intake (Table 1). DQI-K scores of 0, 1, and 2 or 0 and 1 were assigned to each component, and scores for individual components were summed to determine each participant's overall DQI-K score, which ranged from 0 to 9; a lower DQI-K score reflects a higher quality diet.
Our subjects were divided into low DQI-K (total DQI-K score of 0–4) and high DQI-K (total DQI-K score of 5–9) groups. We further divided the DQI-K scores into four groups for detailed analysis (0–2, 3–4, 5–6, and 7–9).
Statistical analysis
Chi-square tests (for categorical variables) and linear regressions (for continuous variables) were used to analyze the selected characteristics of the study population based on DQI-K scores.
Person-months of follow-up were calculated from the enrollment date at recruitment until the date of death for cases or the last date of follow-up (December 31, 2015). To assess the association between the DQI-K and all-cause mortality, hazard ratios (HRs) and 95% confidence intervals (95% CI) were estimated via Cox proportional hazard regression models [21]. We tested the Cox proportional hazards assumption using visual evaluation of log[log(survival)] plots [22].
The first multivariable model was adjusted for age (continuous), gender, income (< 2,000,000 and ≥ 2,000,0000 won), smoking (never and ever), alcohol drinking (never, past, and current), BMI (continuous, kg/m2), and total calories (continuous, kcal). We additionally analyzed the incidence of diabetes, which may have altered dietary habits.
All statistical analyses were performed using SAS statistical software version 9.4 (SAS Institute, Cary, NC). Statistical significance was determined as P < 0.05.
RESULTS
General characteristics
In our study, 2,165 subjects died during the 6.6-year follow-up period (1,309 men and 856 women). The mean age at enrolment was 53.6 years in men and 52.3 years in women.
The baseline characteristics of the study subjects are summarized in Table 2. In this study, subjects with high scores were more likely to have ever been smokers and drinkers, were less likely to be physically active, and more likely to have higher energy intake compared to those with low scores.
Association between the DQI-K and mortality in study participants
To compare diet quality, the participants were divided into two groups based on their DQI-K (scores 0–4 and 5–9). A high DQI-K score indicates poor diet quality. The associations between the DQI-K and mortality in the HEXA-G are shown in Table 3. Poor diet quality (score 5–9) was associated with an increased risk of all-cause mortality (HR = 1.22, 95% CI = 1.06–1.40; multivariable-adjusted HR = 1.31, 95% CI = 1.08–1.91). We also examined the DQI-K as a continuous variable and observed a 6% increase in multivariable-adjusted all-cause mortality (multivariable-adjusted HR = 1.06, 95% CI = 1.02–1.11). The highest DQI-K scores (7–9) were associated with an increased risk of mortality compared to the lowest scores (0–2) (multivariable-adjusted HR = 1.69, 95% CI = 1.15–2.48). Among the components of the DQI-K, the percentage of energy from fat was the most influential factor on the results.
We also conducted analyses to include the prevalence of diabetes into the first multivariable model, and this association was still significant (HR = 1.23, 95% CI = 1.06–1.43).
DISCUSSION
In the present study, the DQI-K was modified based on the original DQI as well as Korean dietary recommendations and dietary reference intakes for Koreans. In previous studies, older participants had higher diet quality than younger participants [232425262728] and non-drinkers [262930]. Regarding food intake, diet quality has been shown to be positively associated with consumption of fruits and vegetables [53132]. Our findings, which parallel those of previous research, support that poor diet quality significantly increases the risk of all-cause mortality [910111213153334353637].
The DQI-K captures overall diet quality by reflecting the dietary patterns of the Korean population as well as variations within individual components. On the other hand, single diet component measures such as intake of daily cholesterol, high sodium consumption, and low fruit intake are limited in explaining overall diet behaviors.
This study has several limitations. One recognized limitation is that the dietary data were collected by a single FFQ at recruitment for assessment of overall diet quality. However, adult dietary patterns appear to remain relatively stable over time [383940414243]. Although the FFQ cannot be used to evaluate the details of sodium intake in individuals, it is still used to measure sodium intake to assess its relationship to chronic diseases, especially in large epidemiologic studies. Further, the FFQ is a useful tool for measuring typical intakes of ingredients to assess their relationships with chronic diseases, especially in large epidemiologic studies [44]. Second, there is the probability of dietary measurement errors. However, we used an FFQ that has shown considerable validity and reliability [45]. Third, the subjects were asked about their dietary intake for the year preceding the diagnosis using the FFQ. Thus, measurement errors likely occurred due to poor recall despite the validity and reproducibility evidence of the questionnaire.
This study also has several strengths. The diet pattern of Koreans is distinct from those of Western and other Asian people. The DQI-K, which is based on the diet guidelines for Koreans, can serve as a simple scoring system that enables rapid evaluation of an individual's diet quality. Additionally, scoring one's own diet quality can help increase awareness of the importance of a healthy diet. The DQI-K can also serve as a basis for education and counseling programs for those who need guidance. Other advantages of the present study include its prospective design with a sufficient number of subjects, which takes into account plausible confounders. This large prospective study supports the important effect that diet quality based on diet guidelines has on risk of mortality. Our results support the health benefits of compliance with diet guidelines to achieve improved health outcomes. Furthermore, the DQI-K and Dietary Reference Intakes for Koreans are closely interrelated and can be used to complement one another in establishing Korean diet guidelines.
In our study, poor diet quality (higher DQI-K scores) according to the DQI-K was associated with an increased risk of all-cause mortality. The DQI-K in the present study can be used to assess the diet quality of Koreans. It is possible to examine the effect of compliance with diet guidelines for Koreans on health outcomes using the DQI-K.
XML Download