Does the relationship between body mass index and health-related quality of life vary depending on early, middle and later adulthood?

Abstract

Objective

This study investigated the association between body mass index (BMI) and health-related quality of life (HRQoL) in a Korean adult population using a nationwide population-based survey.

Design

Cross-sectional study.

Setting

This study analysed data from the 2021 Community Health Survey in South Korea.

Participants

A total of 223 336 respondents were aged ≥19 years.

Results

Underweight status was consistently associated with lower EuroQol HRQoL (EQ-5D) index scores across all age groups (early, middle and later adulthood), with the detrimental impact being most pronounced in men and the elderly (≥65 years). Specifically, underweight men aged 40–64 years exhibited the highest risk for problems in anxiety/depression (OR 2.48; 95% CI 1.86 to 3.30).

The association between obesity and HRQoL showed different patterns depending on sex. In women, obesity was negatively associated with HRQoL across all life stages; notably, young women (19–39) with preobesity or obesity reported difficulties in all five EQ-5D dimensions. Conversely, in men aged ≥40 years, obesity was associated with better HRQoL outcomes and lower odds of anxiety/depression among those aged 40–64 years (OR 0.85; 95% CI 0.77 to 0.94) and those aged ≥65 years (OR 0.88; 95% CI 0.79 to 0.98).

Conclusions

The association between BMI and HRQoL differs according to sex and age. Policy attention regarding HRQoL may benefit from considering the maintenance of proper weight and the promotion of nutritional health, tailored to sex and age.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Body mass index (BMI) is associated with health-related quality of life (HRQoL). While previous studies focused on obesity in younger or mixed-age groups, this study included underweight and adults aged 19 years and older.

WHAT THIS STUDY ADDS

• Results showed that underweight was associated with lower EuroQol HRQoL (EQ-5D) index scores in both men and women, especially in later adults. Women with obesity had lower HRQoL, but middle and later adult men with obesity had better HRQoL than those with normal weight. Women aged 19–39 years with obesity were more likely to report problems in all five EQ-5D dimensions compared with those with normal weight. Conversely, men aged ≥40 years with underweight were more likely to experience issues in all five dimensions.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• The association between BMI and HRQoL differed by sex and age. The findings suggest that policy attention must be directed toward maintaining proper weight and promoting nutritional health to improve HRQoL based on sex and age.

Introduction

The global public health burden of obesity has increased with socioeconomic development. Owing to numerous health risks and tremendous increases in prevalence, overweight and obesity have gained recognition as major public health concerns.¹ Meanwhile, underweight is often ignored in high-income countries because of its low prevalence.² However, the prevalence of underweight remains high, especially in vulnerable populations, such as children, adolescents, pregnant women and older adults, even in high-income countries.^{3 4}

In contrast to solid evidence that the association between BMI and mortality is U-shaped or J-shaped, with the lowest mortality in the healthy weight category,⁵ little is known about the relationship between BMI and health-related quality of life (HRQoL).⁵ HRQoL is a multidimensional concept that captures the dimensions of physical functioning and psychological and social well-being.⁵ Deviations from the normal weight range are associated with lower HRQoL,⁶ particularly worse physical functioning and physical well-being.⁷ Meanwhile, the majority of previously published studies have demonstrated no or only a weak association between BMI categories and anxiety/depression.⁸

Most studies on overweight and obesity and quality of life have relied on data from young, middle-aged or mixed-age group populations. Research on persons aged 65 years and older (≥ 65 years) is limited and has examined only physical functioning.⁴

Obesity is linked to chronic conditions that further diminish HRQoL, particularly in older populations.⁹ Obesity in older Chinese adults has shown a suboptimal quality of life.¹⁰ Being underweight is linked to poor HRQoL among older Japanese adults, suggesting the importance of maintaining proper weight and avoiding nutritional risks at an advanced age.¹¹

The association between BMI and disease may differ by geographical region, even among Asian.¹² Most previous studies have been conducted in Western countries, which have a high prevalence of overweight and obesity.¹³ Therefore, this study aimed to investigate the association between BMI and HRQoL among Koreans aged ≥19 years based on national population-based data, considering the sex and age differences in the BMI-HRQoL association.

Methods

Data source and study population

Data were analysed using the 2021 Community Health Survey (CHS) of the Regional Public Health Act in South Korea.¹⁴ The CHS is a nationwide population-based survey conducted annually since 2008 by the Korea Disease Control and Prevention Agency (KDCA) targeting adults aged ≥19 years. In the CHS, stratified cluster sampling and systematic sampling were used to select sample areas and households, respectively.¹⁵ Trained surveyors visited each household and conducted face-to-face computer-assisted personal interviews. The surveyors attempted to prevent the spread of COVID-19 during the interviews, which were conducted from August 16 to 31 October 2021.¹⁶

Of the 229 242 participants, individuals with missing data (5906 (2.58%)) were excluded. The final study population comprised 223 336 participants aged ≥19 years. Participants aged between 19 and 39 years were defined as early adults. Middle adults were considered those aged 40–64 years. Adults aged ≥65 years were defined as later adults.

Dependent variable

HRQoL was assessed using the EuroQoL HRQoL scale (EQ-5D-3L), the most widely used generic index measures of HRQoL.¹⁷ The EQ-5D questionnaire assessed five health dimensions: mobility, self-care, usual activities, pain/discomfort and anxiety/depression. Each dimension had three levels: no problems, moderate problems and severe problems. These levels can describe 243 different health conditions.¹⁷ Problems were categorised into two groups: have problems (have moderate or severe problems) and those with no problems.

The EQ-5D index score, converted from EQ-5D-3L, was calculated based on the South Korea values set by the KDCA.¹⁸ The maximum score for this value is 1, and a score close to 1 indicates better health. The Korean version of EQ-5D has proven to be valid and reliable.¹⁹

Independent variable

BMI was used as an independent variable of interest and was calculated as self-reported current weight (kg)/(height (m))² based on the Obesity Guideline by the WHO for the Asia Pacific Region and the Korean Society for the Study of Obesity. Participants were divided into four groups according to their BMI values: underweight (<18.5 kg/m²), normal (18.5–22.9 kg/m²), preobesity (23–24.9 kg/m²), obesity (≥25.0 kg/m²).^{20 21}

Covariates

Annual household income was categorised into three groups based on US$: <$4700, $4700–$15 664 and >$15 664. An annual household income of $4700 was the standard for livelihood benefits for one-person households in 2021 under the National Basic Living Security Act 8 in South Korea. Under this Act, an annual household income of $15 664 is the standard median income for one-person households.²² Education was categorised into three groups: uneducated (including illiterate), elementary school graduate, junior high school graduate or high school graduate, and college graduate or university graduate or higher.

Marital status was categorised into two groups: currently married and never married/divorced/widowed. Administrative district units in South Korea are divided into eup, myeon and dong.¹⁴ Residential areas were classified as ‘eup/myeon’ (rural area) and ‘dong’ (urban area).²³

Smoking status was categorised into three groups: current (cigarette or e-cigarette), past smoker and never-smoker. The participants were categorised into high-drinking, moderate-drinking and non-drinking groups. High-risk drinking was defined as drinking more than seven glasses of alcoholic drinks for men and more than five glasses for women two times per week or more.²⁴ Moderate drinking was defined as drinking less than six glasses of alcoholic drinks for men and less than four glasses for women two times per week or more, or regardless of the amount of alcohol consumed, drinking less than four times a month or less. The non-drinking group was defined as those who had not consumed any alcohol in the previous year. The frequency of breakfast consumption was categorised into two groups: 5–7/week and 0–4/week.²⁵ Regular walking was calculated as the activity during the last week, divided into ‘yes’ (regular walking for at least 30 min 5 times/week in the past week) and no.²⁶ Participants were classified as ‘yes’ if they had a clinical diagnosis of either diabetes or hypertension, and ‘no’ otherwise.

Statistical analyses

Sampling weights based on the sample design of the South Korean CHS were applied to statistical analyses to present unbiased estimated representative data for the entire South Korean population. Multiple logistic regression was used to analyse the ORs for moderate or severe problems across the five EQ-5D dimensions.

In addition, multiple linear regression analysis was performed to identify the association between EQ-5D scores and BMI after adjusting for age, marital status, income, education, health behaviours and diabetes or hypertension.

We also conducted a stratified analysis according to sex and age to examine patterns within each subgroup. Age was divided into three categories: early adulthood (19–39 years), middle adulthood (40–64 years) and later adulthood (≥ 65 years). All analyses were performed using SAS V.9.4 (SAS Institute, Cary, North Carolina), and the statistical significance level was set at p<0.05.

Ethical considerations

The study design and survey contents were approved by Statistics South Korea (No. 117075). This study was not subjected to deliberation by the research ethics committee as it was conducted directly or commissioned by the state or local government to review and evaluate public welfare or service programmes (Enforcement Rule of Bioethics and Safety Act, Article 2).

Patient and public partnership

Patients and members of the public were not involved in the design, conduct or reporting of this study.

Results

Table 1 presents the participants’ sociodemographic characteristics, health behaviours, BMI and diabetes or hypertension status. Of the 223 336 participants, 102 945 (46.1%) were men and 70 658 (31.6%) were aged ≥65 years. Among the participants, 31.0% and 50.5% of the men and women, respectively, were of normal weight. The proportion of underweight was 3.3% for men and 5.1% for women aged ≥65 years.

Table 1. General characteristics of the participants (N=223 336).

Variables	n (%)
	Men (n=102 945)				Women (n=120 391)
	19–39 years (n=24 234)	40–64 years (n=47 974)	≥ 65 years (n=30 737)	Total	19–39 years (n=25 616)	40–64 years (n=54 854)	≥ 65 years (n=39 921)	Total
Marital status
Currently married	30.2	77.4	83.5	62.6	40.0	77.1	49.8	59.7
Never married, separated,  divorced, widowed	69.8	22.6	16.5	37.4	60.0	22.9	50.2	40.3
Annual household income
≥15 664	92.5	90.9	56.3	85.0	93.4	88.3	47.0	80.9
4700–15 664	6.1	7.3	35.9	12.2	5.4	9.9	39.2	14.8
<4700	1.5	1.8	7.8	2.8	1.2	1.9	13.7	4.3
Education
≥College	56.3	51.9	19.6	47.4	66.0	39.8	6.5	41.2
High, junior high	43.6	44.8	51.4	45.6	33.7	54.2	34.6	43.0
≤Elementary	0.1	3.3	29.0	7.0	0.3	6.0	58.9	15.8
Residence area
Rural	14.6	19.5	26.6	19.2	12.8	17.4	26.2	17.9
Urban	85.4	80.5	73.4	80.8	87.2	82.6	73.8	82.1
Smoking status
Current smoker	31.4	38.7	18.3	32.4	4.4	2.8	1.6	3.0
Past smoker	18.9	37.5	55.7	34.6	5.8	2.9	1.6	3.5
Never smoker	49.8	23.8	26.0	33.0	89.8	94.3	96.7	93.4
Drinking status
High-risk drinking	13.4	20.7	7.0	15.7	7.2	3.7	0.4	4.1
Moderate drinking	69.6	58.4	44.2	59.5	66.8	51.9	19.7	49.6
Non-drinking	17.0	20.9	48.9	24.8	26.0	44.5	79.9	46.3
Breakfast
0~4 /week	70.0	39.9	7.6	44.1	70.4	38.5	9.4	42.3
5–7/week	30.0	60.1	92.4	55.9	29.6	61.5	90.6	57.7
Regular walking
≥5 day/week	48.5	45.7	50.9	47.6	41.8	43.7	43.5	43.0
<5 day/week	51.5	54.3	49.1	52.4	58.2	56.3	56.5	57.0
Body mass index
Underweight	2.2	1.0	3.3	1.8	11.9	4.4	5.1	6.9
Normal	30.1	28.5	39.1	31.0	58.4	49.8	40.7	50.5
Preobesity	23.4	28.7	31.0	27.3	13.7	23.6	25.6	20.8
Obesity	44.3	41.9	26.6	39.9	16.0	22.2	28.7	21.7
Diabetes or hypertension
Yes (≥1)	5.3	30.9	59.3	27.5	1.9	20.9	62.1	23.8
No	94.7	69.1	40.7	72.5	98.1	79.1	37.9	76.2

Table 2 describes the distribution of experiences with ‘moderate or severe’ problems across all five EQ-5D dimensions and the mean EQ-5D index scores by BMI. For men, the mean EQ-5D index score was the lowest in the underweight category across all age groups. Meanwhile, women aged ≥65 years who were underweight had the lowest EQ-5D scores across all age groups of women.

Table 2. Distribution (%) of moderate or severe problems of health-related quality (EQ-5D) of life according to body mass index.

Variables	Mobility	Self-care	Usual activities	Pain/discomfort	Anxiety/depression	EQ-5D index
	%	%	%	%	%	Score (mean, SE)
Men
19–39 years	1.54	0.39	1.18	11.80	7.25	0.98±0.000
Normal	1.14	0.34	0.90	9.52	7.13	0.98±0.000
Underweight	2.45	0.94	3.01	9.79	11.68	0.97±0.001
Preobesity	1.23	0.36	0.95	10.56	6.33	0.98±0.001
Obesity	1.92	0.41	1.38	13.99	7.58	0.98±0.001
40–64 years	4.85	1.57	3.72	21.70	8.45	0.97±0.000
Normal	5.13	1.85	4.42	21.87	9.38	0.96±0.001
Underweight	15.37	6.48	14.44	30.93	22.04	0.91±0.004
Preobesity	4.08	1.25	3.12	19.98	7.73	0.97±0.001
Obesity	4.89	1.47	3.35	22.52	7.91	0.97±0.001
≥65 years	24.54	8.89	18.33	38.55	12.63	0.91±0.001
Normal	26.43	10.34	20.49	40.08	13.66	0.90±0.002
Underweight	44.57	20.71	38.53	53.27	25.27	0.83±0.003
Preobesity	20.65	7.02	15.12	35.68	11.28	0.93±0.001
Obesity	23.18	7.11	15.76	37.34	10.74	0.91±0.002
Women
19–39 years	1.62	0.32	1.34	14.89	14.72	0.97±0.000
Normal	1.27	0.20	1.00	13.33	13.27	0.97±0.001
Underweight	1.38	0.41	1.38	12.96	16.37	0.97±0.001
Preobesity	1.91	0.39	1.69	15.65	14.79	0.97±0.001
Obesity	2.69	0.61	2.13	20.72	18.38	0.96±0.001
40–64 years	6.52	1.51	4.24	31.01	14.77	0.95±0.000
Normal	4.35	1.14	3.19	26.96	14.43	0.96±0.001
Underweight	6.21	1.86	5.71	28.07	18.46	0.95±0.001
Preobesity	6.19	1.39	3.69	31.96	13.88	0.95±0.001
Obesity	11.14	2.30	6.60	38.43	15.70	0.94±0.001
≥65 years	44.07	15.14	32.51	62.83	22.79	0.85±0.001
Normal	42.25	15.58	32.53	61.67	23.00	0.86±0.001
Underweight	57.13	26.61	47.20	70.84	31.24	0.79±0.003
Preobesity	39.70	12.28	28.45	60.48	21.07	0.86±0.002
Obesity	48.03	14.53	32.93	65.01	22.17	0.84±0.002

EQ-5D, EuroQol health-related quality of life scale.

Table 3 and figure 1 summarise the findings of the logistic regression analyses of the reported problems for each of the five EQ-5D dimensions. Regarding the association between BMI and HRQoL, underweight men in both age groups (40–64 and ≥65 years) were significantly more likely to report problems across all five EQ-5D dimensions. Specifically, among men aged 40–64 years, the ORs were highest for anxiety/depression (OR 2.48; 95% CI 1.86 to 3.30) and usual activities (OR 2.30; 95% CI 1.55 to 3.42). Similarly, men aged ≥65 years exhibited significantly elevated odds for all dimensions, with higher odds observed for usual activities (OR 1.71; 95% CI 1.42 to 2.06) and anxiety/depression (OR 1.65; 95% CI 1.35 to 2.03). In contrast, obesity was associated with lower odds of reporting problems in both the 40–64 group (OR 0.85; 95% CI 0.77 to 0.94) and the ≥65 group (OR 0.88; 95% CI 0.79 to 0.98).

Table 3. Adjusted odds of experiencing ‘moderate or severe’ problems across EQ-5D dimensions (logistic regression).

Variables	Mobility	Self-care	Usual activities	Pain/discomfort	Anxiety/depression
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Men
19–39 years
Normal	1.00	1.00	1.00	1.00	1.00
Underweight	2.06 (1.02 to 4.15)	2.54 (0.79 to 8.21)	2.95 (1.55 to 5.60)	1.18 (0.85 to 1.63)	1.97 (1.45 to 2.69)
Preobesity	1.33 (0.93 to 1.89)	1.40 (0.70 to 2.81)	1.24 (0.80 to 1.91)	1.15 (1.01 to 1.31)	0.88 (0.75 to 1.02)
Obesity	2.01 (1.51 to 2.69)	1.60 (0.90 to 2.85)	1.75 (1.27 to 2.42)	1.42 (1.28 to 1.59)	1.05 (0.92 to 1.19)
40–64 years
Normal	1.00	1.00	1.00	1.00	1.00
Underweight	2.21 (1.54 to 3.15)	1.75 (1.11 to 2.78)	2.30 (1.55 to 3.42)	1.36 (1.07 to 1.72)	2.48 (1.86 to 3.30)
Preobesity	0.97 (0.83 to 1.14)	0.79 (0.60 to 1.05)	0.87 (0.73 to 1.04)	0.93 (0.86 to 1.00)	0.89 (0.80 to 1.00)
Obesity	1.14 (0.99 to 1.31)	0.95 (0.76 to 1.19)	0.89 (0.76 to 1.04)	1.09 (1.02 to 1.17)	0.85 (0.77 to 0.94)
≥ 65 years
Normal	1.00	1.00	1.00	1.00	1.00
Underweight	1.64 (1.35 to 1.98)	1.60 (1.27 to 2.01)	1.71 (1.41 to 2.06)	1.38 (1.17 to 1.64)	1.65 (1.35 to 2.03)
Preobesity	0.86 (0.78 to 0.94)	0.74 (0.65 to 0.84)	0.79 (0.71 to 0.87)	0.83 (0.78 to 0.90)	0.84 (0.78 to 0.93)
Obesity	1.20 (1.09 to 1.32)	0.94 (0.82 to 1.08)	0.97 (0.87 to 1.08)	1.04 (0.96 to 1.12)	0.88 (0.79 to 0.98)
Women
19–39 years
Normal	1.00	1.00	1.00	1.00	1.00
Underweight	0.95 (0.64 to 1.41)	1.69 (0.79 to 3.61)	1.13 (0.78 to 1.66)	1.03 (0.90 to 1.18)	1.32 (1.17 to 1.49)
Preobesity	1.49 (1.10 to 2.02)	2.01 (1.00 to 4.02)	1.49 (1.07 to 2.07)	1.16 (1.03 to 1.30)	1.13 (1.01 to 1.27)
Obesity	1.77 (1.34 to 2.33)	2.28 (1.24 to 4.20)	1.58 (1.15 to 2.17)	1.48 (1.33 to 1.64)	1.30 (1.17 to 1.45)
40–64 years
Normal	1.00	1.00	1.00	1.00	1.00
Underweight	1.66 (1.31 to 2.09)	1.31 (0.84 to 2.03)	1.92 (1.49 to 2.48)	1.08 (0.97 to 1.22)	1.28 (1.12 to 1.48)
Preobesity	1.30 (1.15 to 1.48)	1.09 (0.84 to 1.40)	1.05 (0.90 to 1.24)	1.17 (1.10 to 1.25)	1.01 (0.93 to 1.10)
Obesity	2.14 (1.91 to 2.39)	1.38 (1.11 to 1.71)	1.61 (1.41 to 1.84)	1.45 (1.37 to 1.54)	1.06 (0.98 to 1.14)
≥65 years
Normal	1.00	1.00	1.00	1.00	1.00
Underweight	1.23 (1.07 to 1.40)	1.49 (1.28 to 1.74)	1.39 (1.22 to 1.59)	1.21 (1.05 to 1.39)	1.50 (1.32 to 1.71)
Preobesity	1.18 (1.09 to 1.28)	1.02 (0.92 to 1.14)	1.09 (1.01 to 1.19)	1.08 (1.00 to 1.16)	0.97 (0.90 to 1.06)
Obesity	1.74 (1.62 to 1.87)	1.26 (1.15 to 1.39)	1.34 (1.24 to 1.45)	1.26 (1.18 to 1.35)	0.95 (0.88 to 1.03)

EQ-5D, EuroQol health-related quality of life scale.

Figure 1. Adjusted odds of experiencing ‘moderate or severe’ problems across EQ-5D dimensions (mobility, self-care, usual activities, pain/discomfort, anxiety/depression) by BMI groups. Adjusted for age (continuous), marital status, annual household income, education, residence area, smoking status, drinking status, breakfast, regular walking and diabetes or hypertension. (A) men aged 19–39, (B) men aged 40–64, (C) men aged 65 or over, (D) women aged 19–39, (E) women aged 40–64, (F) women aged 65 or over. BMI, body mass index; EQ-5D, EuroQol health-related quality of life scale.

Among women, however, women aged 19–39 years with pre-obesity or obesity were more likely to report problems in all five dimensions. Furthermore, being underweight was a significant risk factor for women aged ≥65 years, who showed elevated odds for problems in mobility (OR 1.23; 95% CI 1.07 to 1.40), self-care (OR 1.49; 95% CI 1.28 to 1.74), usual activities (OR 1.39; 95% CI 1.22 to 1.59), pain/discomfort (OR 1.21; 95% CI 1.05 to 1.39) and anxiety/depression (OR 1.50; 95% CI 1.32 to 1.71).

Table 4 summarises the total EQ-5D index scores according to BMI, considering the sociodemographic characteristics, health behaviours and diabetes and hypertension statuses of the participants. Being underweight was negatively associated with EQ-5D index scores in both men and women across all age groups: aged 19–39 years (men: β=−0.0094, women: β=−0.0042), aged 40–64 years (men: β=−0.0440, women: β=−0.0099), aged ≥65 years (men: β=−0.0436, women: β=−0.0282). Obesity was negatively associated with EQ-5D index scores in women across all age groups.

Table 4. Multiple linear regression exploring the relationship between BMI and EQ-5D index scores.

Variables	19–39 years		40–64 years		≥ 65 years
	Estimate	(95% CI)	Estimate	(95% CI)	Estimate	(95% CI)
Men
BMI
18.5–22.9 (normal)	Reference		Reference		Reference
<18.5 (underweight)	−0.0094	(−0.014 to −0.005)	−0.0440	(−0.051 to −0.037)	−0.0436	(−0.052 to −0.035)
23.0–24.9 (preobesity)	−0.0001	(−0.002 to 0.002)	0.0032	(0.001 to 0.005)	0.0127	(0.009 to 0.016)
≥25.0 (obesity)	−0.0036	(−0.005 to −0.002)	0.0011	(−0.001 to 0.003)	0.0051	(0.001 to 0.009)
Women
BMI
18.5–22.9 (normal)	Reference		Reference		Reference
<18.5 (underweight)	−0.0042	(−0.007 to −0.002)	−0.0099	(−0.013 to −0.006)	−0.0282	(−0.035 to −0.022)
23.0–24.9 (preobesity)	−0.0027	(−0.005 to −0.001)	−0.0019	(−0.004 to −0.001)	−0.0001	(−0.004 to 0.004)
≥25.0 (obesity)	−0.0103	(−0.012 to −0.008)	−0.0127	(−0.014 to −0.011)	−0.0192	(−0.023 to −0.015)

BMI, body mass index; EQ-5D, EuroQol health-related quality of life scale.

Discussion

This study investigated the relationship between BMI and HRQoL using a nationally representative sample of South Korean adults. Among the men participants, 1.8% were underweight and 39.9% were obese. Among the women participants, 6.9% were underweight and 21.7% were obese. Among Asians aged 24–95 years, the obesity prevalence was 35.8% and 34.0% in men and women, respectively, and the underweight prevalence was 5.2% and 9.1% in men and women, respectively.²⁷ We found a higher prevalence of obesity in men and a lower prevalence of obesity in women in our study than in other Asian studies as well as a lower prevalence of underweight in men and women.

When we considered the quality of life score, both men and women who were underweight reported a decreased quality of life. The quality of life of middle-aged and older men with obesity was likely to be better than that of men with normal weight. However, these findings should be interpreted with caution, as overweight and obesity in this demographic are established risk factors for adverse health outcomes, including a higher cumulative lifetime risk of incident cardiovascular disease, cardiovascular mortality²⁸ and oesophageal cancer.²⁹

Our findings demonstrate that the negative association between underweight and HRQoL is notably observed among individuals aged ≥65 years. These results align with previous studies, indicating a stronger link between underweight status and reduced quality of life, particularly in older men.³⁰ Furthermore, we observed that while being underweight was associated with worse HRQoL in men aged 40 and older, being overweight was linked to better outcomes. This finding is consistent with a Canadian study of adults over 40, which reported that underweight status imposed a persistent burden on quality of life, whereas being overweight at age 57 was associated with higher quality of life compared with normal weight.³¹

In contrast, women with obesity reported lower HRQoL than their normal-weight counterparts across all age groups (≥ 19 years). This association may be attributed to women’s greater vulnerability to stress related to body image and cultural weight perceptions.³² Such stress likely stems from the higher levels of discrimination women face in both professional and social environments due to obesity.³³ Similarly, the tendency of older Korean women with obesity to underestimate their body size can also be interpreted within this sociocultural framework.³⁴

The associations between EQ-5D scores and BMI were similar to the odds of reporting a problem on EQ-5D dimensions. Adults aged ≥65 years who were underweight were more likely to have problems in all five dimensions in both men and women.^{7 35} This is consistent with a previous study that demonstrated that being underweight is associated with poor HRQoL.³⁶ Both men and women with underweight across all age groups were more likely to have moderate or severe problems in the anxiety/depression dimension.³⁶

Sex differences exist in the associations between mental health and obesity.³⁷ This study showed that men aged 40–64 years with obesity were less likely to experience problems in the anxiety/depression dimension than those with normal weight; however, women with obesity were more likely to have problems than those with normal weight. It is plausible that the healthy BMI range for women may differ from that for men. More in-depth studies on potential sex differences and weight changes over time are needed.⁵

Many developed countries are battling preobesity or obesity, and the high prevalence of obesity in South Korea is also a major public health problem. According to the 5th National Health Plan in South Korea, obesity is a subject of prevention and management.³⁸ While obesity in young adults is primarily driven by unhealthy lifestyle practices,³⁹ obesity in middle-aged and older adults stems from the concurrence of multiple factors, such as inactivity, poor nutritional habits and basal metabolism and nutritional need reduction.⁴⁰

However, our study found that middle-adulthood and later-adulthood men with preobesity had positive associations, whereas adults with underweight showed a negative association. The negative association between underweight status and EQ-5D index scores was observed in both men and women, with notable patterns identified within the man subgroups. The associations between BMI categories and HRQoL were clearly evident in both men and women among the older age groups. A previous study suggested that older adults who are underweight are expected to live the shortest lives and spend the fewest years in an active state.¹¹ The underweight of later adults should serve as a more important health problem than that of early or middle adults.

Older adults are particularly vulnerable to health risks associated with being underweight, such as sarcopenia and an increased risk of hip fracture.⁴¹ Moreover, nutritional problems or inadequate community food environment for older adults are important causes of food insecurity, especially among South Korean single rural older adults.⁴² Addressing the underweight status among older adults is crucial for promoting their overall well-being and reducing negative health outcomes. Therefore, underweight status should be considered important, especially for older adults in Korea, whose population is rapidly ageing.

Governments need sustainable policy interventions to establish health habits. Therefore, a better understanding of the association found in this study, which identified the associations between BMI and HRQoL considering sex and age, could help health practitioners understand the association between BMI and HRQoL, potentially informing more tailored public health messages for different sex and age groups.

This study has some limitations. First, weight and height were self-reported, possibly leading to the underestimation of weight, overestimation of height and underestimation of overall BMI.⁴³ Nevertheless, self-reported BMI is widely accepted as a reliable proxy for assessing health-related associations in large-scale epidemiological studies.⁴⁴ Furthermore, given that previous research has consistently identified an adverse impact of self-reported underweight status on quality of life,⁴⁵ we consider our findings to be valid despite these measurement limitations. Importantly, any potential misclassification of BMI categories suggests that our results may actually underestimate the true strength of the association.⁴³ For instance, if healthy individuals in the normal weight range were misclassified as underweight, the observed negative impact on HRQoL would be diluted. Thus, the actual decrement in quality of life associated with being underweight might be even more pronounced than reported here. Second, we did not consider all possible confounding factors, such as pain, osteoarthritis and long-standing health conditions.⁴⁶ Previous studies suggest that the number of pain sites serves as a critical mediator of the association between obesity and diminished physical function.⁴⁷ Third, body composition and nutritional status are key components for maintaining good general health and longevity.⁴⁸ Although BMI is widely used by clinicians and nutritionists, it does not provide information on the relative contribution of fat mass and fat-free mass. Fat mass and fat-free mass could not be considered because information on waist circumference, waist-to-hip ratio and nutritional status was not available. Given that men with overweight or obesity tend to exhibit greater muscle and lean body mass compared with women.⁴⁹ Fourth, the definition of obesity varies globally, which limits generalisability. The obesity cut-off is ≥25.0 kg/m² in South Korea and Japan, whereas it is ≥28.0 kg/m² in China and ≥30.0 kg/m² in the United States.⁵⁰ Therefore, our results should be interpreted with caution when applied to populations subject to different diagnostic standards and sociocultural contexts. Finally, causality could not be inferred due to the cross-sectional study design. Future longitudinal studies are therefore warranted to address bidirectional relationships and control for unmeasured confounders. Nevertheless, a major strength of this study is the use of nationally representative data.

Data availability statement

Data are available upon reasonable request.