With the rise of non-communicable disease—mainly obesity, type 2 diabetes mellitus (T2DM), and cardiovascular disease (CVD) (1–3)—in both developed and developing countries, significant changes were made in the public health priorities for their prevention. Of note, modern dietary guidelines not only aim to prevent nutrient deficiencies but also aim to prevent chronic disease. Conventional “nutrient-based” guidelines showed considerable limitations when extending to chronic disease prevention; thus, the translation of “nutrient-based” into “food-based” guidelines has been accentuated recently (4–6). Indeed, the growing burden of non-communicable diseases has led to demand for a holistic approach to food consumption that can manage the double burden of malnutrition, under-, and over-nutrition. “Food-based” guidelines attempt to reflect the synergistic contributions of dietary patterns, including food composition, preparation methods, and underlying social interactions. To ensure the compliance and effectiveness of established guidelines, quantitative measurements such as diet quality scores have been developed using both empirical and a priori methods such as Mediterranean diet score, Healthy Eating Index (HEI), and Alternative Healthy Eating Index (AHEI) (7, 8).

Several epidemiological studies have examined the association between diet quality and morbidity and mortality from non-communicable disease, where heterogeneous outcomes were found, including significant positive (9–11), inverse (12, 13), or no associations (14). The Nurses' Health Study comprising 80,029 participants reported a statistically significant association between higher AHEI and a lower risk of T2DM (9). However, no significant association was found in a meta-analysis examining the association between dietary diversity score (DDS) and obesity (15). Dietary factors associated with inflammation status have been highlighted, along with the recent development of dietary index predicting inflammation levels (16–20). Chronic inflammation status is recognized as an underlying pathophysiological mechanism linking multiple chronic diseases, including atherosclerosis (21), T2DM (22), CVD (23), and several cancers (24). The heterogeneous results of the previous literature suggest the need for a population-specific diet quality assessment tool regarding the association between non-communicable diseases.

In our previous study, we developed Data Derived Inflammatory Index (DDII) based on high-sensitive C-reactive protein (hsCRP) levels (20). It is a Korean-specific index that reflects both diet and lifestyle factors with major attributes to hsCRP levels. Food and Agriculture Organization (FAO) of the United Nations developed the Minimum Dietary Diversity for Women (MDD-W), a food group-based dichotomous scoring system, designed to assess the micronutrient adequacy of the women of reproductive age in developing countries (25). Several epidemiologic studies found the association between MDD-W and multiple chronic diseases (26–28).

Of note, the highest prevalence of obesity was reported among Filipino women from the immigrant study in Korea (29). Another immigrant study in the United States found that the prevalence of obesity in Filipinos tripled between 1992 and 2011, which was the highest among Asian immigrants (30). Our group found an inverse association between obesity prevalence and higher DDS based on 11 food groups (31). Abdominal obesity is a distinct phenotype of the obesity pandemic in developing countries, strongly associated with metabolic comorbidities (32–34). In addition, Asians may have a higher body fat mass at a given value of BMI compared with Caucasians (35). Based on this rationale, international institutions have proposed population-specific cutoff points for the diagnosis of obesity and metabolic syndrome (19, 36, 37). In tandem, population-specific public health intervention strategies, including dietary guidelines to prevent non-communicable disease, should be enforced concerning the metabolic differences.

Given that Filipino immigrant women are susceptible to lower dietary diversity and metabolic diseases, we aimed to examine whether MDD-W, a developing country-specific score, and DDII, a Korean-specific score, were associated with obesity and other metabolic comorbidities among Filipino immigrant women in Korea. We examined the associations between MDD-W and DDII and indicators of obesity, dyslipidemia, and insulin resistance in the Filipino Women's Diet and Health Study (FiLWHEL).

A total of 413 participants from the baseline population of the FiLWHEL study were included in the analysis, with variation regarding the major outcomes; 413 participants were included for general obesity (no. of cases = 124), 411 participants were included for abdominal obesity (no. of cases = 177), and 405 participants were included in the dyslipidemia and insulin resistance analyses (no. of cases = 39 for TG, 28 for HDL, 76 for LDL, 109 for TC, 43 for TG/HDL, 141 for HbA1c, 43 for FBG, 57 for insulin levels, and 65 for HOMA-IR). The baseline characteristics according to each diet quality score are outlined in Tables 1, 2. MDD-W scores did not vary across DDII scores, nor did DDII scores according to MDD-W scores. The individuals with greater MDD-W scores tended to have higher % energy from carbohydrate intake but lower % energy from fat intake compared to those with lower MDD-W scores. The participants with higher DDII scores tended to be older, had higher BMI, higher % energy from carbohydrate intake but lower % energy from fat intake, and higher fasting blood glucose levels compared to those with lower DDII scores. Engagement in vigorous activity was higher among individuals with high MDD-W scores than among those with low scores. The proportions of current alcohol consumers were the lowest in the top categories of MDD-W and DDII scores, while the proportions of participants who lived more than 8 years in Korea were the highest in the top categories of both scores. The mean (±SD) score was 4.79 (±1.53) for MDD-W and 2.33 (±0.36) for DDII.

We observed a statistically significant linear trend of increased hsCRP levels according to the quintiles of DDII in the age- and energy-adjusted model; LS-means (95% CIs) were 0.79 (0.61–1.02) for quintile 1, 0.96 (0.76–1.22) for quintile 2, 1.02 (0.82–1.28) for quintile 3, 1.14 (0.90–1.45) for quintile 4, and 1.65 (1.30–2.09) for quintile 5 (p for trend < .001) (Supplementary Table 1).

The associations between the tertiles of MDD-W and obesity, dyslipidemia, and insulin resistance markers are presented in Table 3. We observed a statistically significant association between wider dietary variety and lower prevalence of central obesity, while higher prevalence of dyslipidemia. The multivariable adjusted ORs (95% CIs) of the 3rd vs. the 1st tertiles of MDD-W were, 0.58 (0.36–0.94) for central obesity, 2.43 (1.11–5.36) for hypertriglyceridemia, 2.02 (1.17–3.47) for hypercholesterolemia, and 3.37 (1.53–7.43) for abnormal TG/HDL ratio (p for trend: 0.029, 0.029, 0.013, and 0.003, respectively). We also found a suggestive inverse trend between MDD-W and lower prevalence of general obesity but an association or trend was not statistically significant. No statistically significant association was found between MDD-W scores and biomarkers of insulin resistance (HbA1c, FBG, insulin, and HOMA-IR).

We observed that the increased DDII scores were associated with an elevated prevalence of dyslipidemia and insulin resistance (Table 4). The multivariable adjusted ORs (95% CIs) of the 5th vs. 1–3rd quintiles of DDII were 6.44 (2.56–16.20) for triglycerides, 3.90 (1.92–7.90) for LDL cholesterol, 3.36 (1.81–6.24) for total cholesterol, 6.25 (2.53–15.41) for abnormal TG/HDL ratio, 3.59 (1.96–6.59) for HbA1c, 2.61 (1.11–6.17) for fasting blood glucose levels, 9.67 (4.16–22.48) for insulin levels, and 9.73 (4.46–21.25) for HOMA-IR (p for trend: <0.001 for all, except 0.033 for fasting blood glucose).

In the stratified analysis, the association between MDD-W scores and obesity was not modified by age nor length of residence in Korea (Table 5). Although, the association was not statistically significant, we found a more pronounced association for DDII and LDL-cholesterol among the participants under age 35 compared to the participants above age 35 (Table 6). We also found a suggestive pronounced association with LDL cholesterol among the participants with <8 years of residence in Korea compared to the immigrants who moved earlier. The associations of DDII with triglycerides and HOMA-IR did not differ by age, BMI, or length of stay in Korea (Supplementary Table 9).

In the sensitivity analysis for the quartiles of MDD-W using the logistic regression, the trend remained statistically significant for triglycerides, total cholesterol, and TG/HDL ratio (Supplementary Table 2). When MDD-W was examined using the GLM procedure, we found a suggestive inverse trend between lower means of waist circumferences with higher scores of MDD-W (p for trend = 0.052 for 3 categories and 0.071 for 4 categories) (Supplementary Tables 3, 4). The association was retained in the sensitivity analysis of dyslipidemia and insulin resistance by quintiles of DDII using both logistic regression and GLM procedure (Supplementary Tables 5, 6). When we limited the analysis to WRA, we found increased odds of triglycerides, total cholesterol, and TG/HDL ratio (Supplementary Table 7). A similar trend was found for abdominal obesity, total cholesterol, TG/HDL ratio, insulin levels, and HOMA-IR with hsCRP levels (Supplementary Table 8). We found a similar association for MDD-W and DDII, compared to the main results, when we adjusted for BMI as a continuous variable (Supplementary Tables 10, 11).

In this analysis of 413 Filipino immigrant women in Korea, we found a statistically significant association between greater dietary diversity and decreased prevalence of abdominal obesity and increased prevalence of dyslipidemia. We also found a statistically significant association between increased inflammatory index and increased prevalence of dyslipidemia and insulin resistance.

Higher points of MDD-W score were inversely associated with central obesity, which corresponded with previous epidemiological studies on dietary diversity and chronic disease (48–50). A cross-sectional study examining 289 Iranian female participants found a statistically significant inverse association between quartiles of DDS and obesity and abdominal adiposity; ORs (95% CIs): 1.00, 0.43 (0.14, 1.39), 0.34 (0.11, 1.29), and 0.25 (0.09, 0.98) for obesity (p for trend = 0.04), and ORs (95% CIs): 1.00, 0.64 (0.32, 1.65), 0.45 (0.21, 1.09), and 0.28 (0.11, 1.12) for abdominal adiposity (p for trend = 0.04) (51). A prior case-cohort study from European Prospective Investigation into Cancer and Nutrition-Norfolk study found lower risk of T2DM among participants consuming greater variety of fruit (RR: 0.70, 95% CI: 0.53, 0.91), vegetable (RR: 0.77, 95% CI: 0.61, 0.98), and fruits and vegetables combined (RR: 0.61, 95% CI: 0.48, 0.78) (52). Indeed, consuming a variety of food types may be beneficial for sustaining healthy dietary patterns and preventing non-communicable diseases (4, 53). Various international and national guidelines recommend diverse nutrient-dense food consumption across all food groups, with major emphasis on vegetables and protein sources (3, 54, 55).

We found that greater dietary diversity was associated with higher prevalence of hypercholesterolemia. Despite the accumulating literature supporting the association between dietary diversity and obesity, limited evidence exists regarding the lipid profiles (12, 56–58). In the cross-sectional analysis of the 2016 Chinese Urban Adults Diet and Health Study (CUADHS), adherence to Chinese healthy food diversity (HFD) recommendations was not significantly associated with metabolic syndrome and its components after adjusted for potential confounders (59). A U.S. study examining the risk of coronary heart disease using the data from three cohort studies did not observe a significant association with MDD-W (28). Although, the reason we observed high total cholesterol levels with MDD-W scores remains unclear, other dietary indices, including Healthy Eating Index (HEI), Alternate Healthy Eating Index (AHEI), MedDietScore (MDS), and Dietary Approaches to Stop Hypertension (DASH), may be more relevant for CVD-related lipid profiles (60). A report from the Multi-Ethnic Study of Atherosclerosis (MESA) reported a weak correlation between DDS and diet quality scores such as DASH and the AHEI (58). We also observed that MDD-W scores were not correlated with DDII scores in our study population.

Recently, diet quality scores reflecting the diet-related inflammation status have been developed internationally, which showed a decent agreement with previously established measurement schemes (20, 61–63). We found a statistically significant association between proinflammatory DDII scores and increased prevalence of dyslipidemia and insulin resistance in our study. Our findings aligned with recent epidemiological literature investigating the association between dietary inflammation estimates and chronic disease (64–66). A study examining the literature-derived dietary inflammatory index (DII®), consisting of 46 pro- and anti-inflammatory parameters, found a statistically significant association between higher DII® and metabolic syndrome (65).

The inflammatory status induced by surplus energy intake may play a role in the pathophysiology linking diet, obesity, and dyslipidemia (67). The chronic inflammation state of obesity is widely acknowledged as a pivotal contributor to the progression of dyslipidemia, T2DM, CVD, and certain cancers (68, 69). Excess energy mainly stored in adipocytes triggers obesity-associated inflammation through multiple pathways, including inflammatory cytokine upregulation, increased oxidative stress, and low-grade inflammatory response (70, 71). Subsequently, the inflammation status yields the development of insulin resistance and increased IGF-1 (72). Moreover, hypercaloric intake mainly from carbohydrate sources is linked to greater stores of intrahepatic triglycerides, resulting in atherogenic dyslipidemia (73).

Immigrants alter their dietary behaviors as part of their acculturation process, known as “dietary acculturation” (74, 75). It involves a complex multidimensional interaction between cultural, social, economic, and environmental factors, with variations among individuals. In our previous cross-sectional analysis, we found that Filipino immigrants consume a smaller variety of foods compared with Koreans (76). A study examining the dietary diversity of Filipino children from the 2013 Filipino National Nutrition Survey found that mean (± SD) DDS score was 4.1 (± 1.3) out of nine food groups (77). Moreover, another Filipino pediatric study reported that the meals are mainly composed of unhealthy sources (refined rice or cookies), while vegetables, fruits, and meats contribute less to daily nutrient intake (78). With suggestions that maintaining the diet pattern low in food diversity may not be optimal for Filipino immigrants, favorable acculturation with food diversity may be an option. Further, epidemiological investigations are warranted regarding dietary acculturation and health outcomes among Filipino immigrants in Korea.

The investigation of Filipino immigrant women in Korea is unique to the literature; to the best of our knowledge, this is the first study to examine the association between diet quality scores and prevalence of obesity, dyslipidemia, and insulin resistance in the immigrant population of East Asia. The strengths of the current study are that it focuses on immigrant Filipino women in the multicultural society of South Korea and that the data collection methodology used across sites was standardized and uniform, along with centralized sample processing. However, several limitations need to be acknowledged. First, the cross-sectional nature of the study limits our ability to examine temporal or causal relationships. In addition, dietary data were collected through a single-day 24-h recall, which may not reflect the usual daily intake. The study was limited in power due to its relatively small sample size. Moreover, although we controlled for potential confounding factors, the results are still subject to residual confounding.

In conclusion, greater dietary diversity was positively associated with a lower prevalence of obesity, but higher prevalence of dyslipidemia, while proinflammatory patterns were associated with an elevated prevalence of dyslipidemia and insulin resistance. Further epidemiological and clinical studies are warranted to replicate our findings.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Institutional Review Board of Sookmyung Women's University. The patients/participants provided their written informed consent to participate in this study.

HK outlined the initial manuscript and performed data analysis. HL and MK assisted in the manuscript drafting. CL and JL conceptualized, supervised the study process, and received funding for the project. HK, HL, SP, GC, SH, SY, SL, and JL were involved in the data collection and critically revised the intellectual content of the manuscript. All authors made a significant contribution to the interpretation of the data.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.