Comparative impact of social isolation on mortality in adults aged 40 years and above with versus without metabolic syndrome: evidence from two large cohorts in the U.S. and U.K.

Abstract

Introduction

Social isolation is increasingly recognized as a significant public health concern associated with mortality risk. However, whether the impact of social isolation on mortality differs between individuals with and without metabolic syndrome (MetS) remains unclear. This study aimed to investigate the associations of social isolation with all-cause mortality, cardiovascular mortality (CVDM), cancer mortality (CAM), other cause mortality (OTM), and premature mortality in MetS and non-MetS populations using data from large cohorts in the UK and the US.

Methods

This study analyzed data from 75,190 participants with metabolic syndrome (MetS) and 229,388 participants without MetS in the UK Biobank, as well as 5758 MetS participants and 7448 non-MetS participants from the U.S. National Health and Nutrition Examination Survey (NHANES). All participants included in the study were aged 40 years or above. The identification of MetS was based on a comprehensive assessment of multiple biochemical indicators, including waist circumference, blood glucose, blood pressure, and blood lipid levels. Social isolation was evaluated using information on marital status, household size, frequency of contact with family and friends, and engagement in social activities. The primary outcomes included all-cause mortality, cardiovascular mortality, cancer mortality, other-cause mortality, and premature mortality, defined as death before the age of 70. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the associations between social isolation and various mortality outcomes. In addition, interaction and subgroup analyses were conducted to explore the potential modifying effects of MetS status, as well as lifestyle and other risk factors, on the relationship between social isolation and mortality.

Results

In the UK Biobank, the rates of all-cause mortality, CVDM, CAM, OTM, and premature mortality among participants with MetS were 9.07%, 1.48%, 4.22%, 3.36%, and 1.98%, respectively; the corresponding rates among participants without MetS were 4.81%, 0.51%, 2.61%, 1.68%, and 2.47%. In NHANES, the respective mortality rates among individuals with MetS were 26.20%, 9.24%, 6.15%, 10.85%, and 13.90%, and among those without MetS were 25.80%, 8.13%, 6.31%, 11.30%, and 14.10%. Cox regression analyses showed that, in the fully adjusted models, social isolation was significantly associated with increased risks of all-cause mortality, CVDM, CAM, OTM, and premature mortality in both individuals with and without MetS. In the UK Biobank, the HRs for participants with MetS were 1.30, 1.21, 1.12, 1.38, and 1.39, respectively; for those without MetS, the HRs were 1.51, 1.75, 1.30, 1.76, and 1.54, respectively. In the U.S. NHANES, the HRs for the MetS group were 1.14, 1.54, 1.48, 1.71, and 1.09, respectively; while for the non-MetS group, the HRs were 1.60, 1.75, 1.47, 1.56, and 1.39, respectively. The results of the interaction and sensitivity analyses were consistent.

Conclusions

Compared to individuals without MetS, those with MetS have higher mortality rates. Moreover, social isolation is associated with increased mortality regardless of MetS status. It is a risk factor for all-cause mortality, CVDM, CAM, OTM, and premature mortality in both MetS and non-MetS populations. Notably, the impact of social isolation on all-cause, cardiovascular, and premature mortality is more pronounced in individuals without MetS. Public health strategies should focus on population-wide interventions to reduce social isolation, enhance social engagement, and improve overall health and longevity, rather than targeting only high-risk groups.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12963-025-00414-9.

Background

In 2021, noncommunicable diseases (NCDs) caused at least 43 million deaths, accounting for 75% of global nonpandemic-related deaths [1]. According to the latest data from the World Health Organization (WHO), premature deaths account for as much as 78.82% of all NCD-related deaths [2]. These premature deaths not only result in productivity losses but also have a significant impact on the economy [3].

Metabolic syndrome (MetS) is a syndrome characterized by abdominal obesity, elevated blood glucose, high blood pressure, and dyslipidaemia [4]. The occurrence of MetS significantly increases the risk of developing NCDs, such as cardiovascular diseases [5, 6] and diabetes [7] and is closely associated with an increased mortality rate [7, 8]. Approximately one-third of Americans are affected by this condition, with the prevalence reaching as high as 45.5% in some Middle Eastern countries [9, 10]. MetS affects approximately one-fourth of the population worldwide, impacting more than one billion individuals [9].

“Social isolation” refers to the objective state of lacking interactions with others or having a limited social network. It is defined as an objective lack of relationships and a low frequency of interactions with family, friends, and the broader community [11, 12]. As a globally recognized public health concern, social isolation has been shown to be closely associated with mortality [13] and the onset of various diseases, including type 2 diabetes [12], cardiovascular diseases [14], dementia [15] and others.

Although both social isolation and MetS are independently linked to adverse health outcomes, limited research has examined whether the mortality risk associated with social isolation differs according to MetS status. It remains unclear whether individuals with and without MetS are equally vulnerable to the negative health impacts of social isolation.

To address this gap, the present study utilized data from two large-scale cohorts, the UK Biobank and the U.S. cohort. National Health and Nutrition Examination Survey (NHANES)—to investigate the associations between social isolation and multiple mortality outcomes, including all-cause mortality, cardiovascular mortality (CVDM), cancer mortality (CAM), other-cause mortality (OTM), and premature mortality. Furthermore, we compared these associations between individuals with and without MetS to explore potential effect modification by metabolic health status.

Methods

Study population

This study focused on individuals aged above 40 years, with the study population stratified into groups with and without MetS. The UK Biobank is a large-scale biomedical database and research resource aimed at improving human health, containing detailed genetic, health, and lifestyle information from 500,000 volunteers aged 40 to 69 years at baseline. Among the 502,419 participants, we initially included 152,276 participants who met the diagnostic criteria for MetS at baseline, along with 350,143 participants without MetS. After a series of inclusion and exclusion steps, a total of 75,190 participants with MetS and 229,388 participants without MetS were included in the final analysis (Supplementary Fig. 1).

The U.S. National Health and Nutrition Examination Survey (NHANES) is a complex, weighted, stratified, multistage sampling study. We utilized data from 1988 to 2020, with an initial population of 135,310 participants. After applying a series of inclusion and exclusion criteria, a total of 5758 participants with MetS and 7448 participants without MetS were included in the final analysis (see Supplementary Fig. 2). The detailed inclusion and exclusion procedures for UKB and NHANES are provided in the Supplementary Materials (Supplementary pp. 20–21).

Assessment of social isolation

The Berkman-Syme Social Network Index [16] measures social isolation based on indicators such as marital status, close relationships, attendance at religious services, and social participation. Building on this framework and previous studies [17], a multidimensional social isolation index was also constructed in the NHANES population. In this context,, the social isolation score was derived from four indicators. The scoring for participants in NHANES III prior to 1998 was based on marital status [16, 18], social interactions [16, 19], participation in religious activities, and involvement in clubs or organizations [20, 21]. Specifically, being married or cohabiting with a partner was considered meeting one criterion. Engaging in social interactions or contact with family, friends, or neighbours more than 155 times per year fulfilled another criterion. Attending religious activities more than three times per year and participating in any club or organization, such as religious groups, fraternities, school groups, or sports teams, were also included. Each unmet criterion was scored as one, while meeting the criterion was scored as zero, resulting in a variable ranging from zero to four.

For data collected after 1998 [22], the scoring criteria were slightly modified. Marital status was assessed, with one point assigned for being widowed, divorced, separated, or never married and zero points assigned for being married or living with a partner. Living alone was scored as one point. The participants were also asked about their difficulty “going out to things such as shopping, movies, or sporting events” and “participating in social activities, such as visiting friends, attending clubs or meetings, or going to parties.” For these two questions, responses of “some difficulty,” “much difficulty,” or “unable to do” were scored as one point each. These scores were combined to create a variable ranging from zero to four. In the NHANES, social isolation is defined as a score of three or higher.

In the UKB cohort, based on previous research [12] on social isolation, the assessment was conducted using a three-item score ranging from 0 to 3, derived from household size, frequency of social activities, and contact with others. The specific assessment was as follows: “How often do you visit friends or family, or have them visit you?” (scored as one point for responses of “once a month,” “once every few months,” “never or almost never,” or “no friends or family outside the household;” scored as zero points for responses of “once a week,” “two to four times a week,” or “almost daily”); “Which of the following leisure or social activities do you engage in once a week or more often?” (scored as one point if no activities were selected); and “Including yourself, how many people live in your household? These include those who usually live in the house, such as students living away from home during term time, and partners in professions, such as the armed forces or pilots” (scored as one point if the response was “0,” and zero points otherwise). In the UK Biobank, social isolation is defined as a score of two or higher.

Assessment of MetS

The diagnostic criteria for the MetS population were determined based on previous research [23, 24], the NCEP-ATP III guidelines, and the standards proposed by the International Diabetes Federation (IDF) [25] and the American Heart Association/National Heart, Lung, and Blood Institute (AHA/NHLBI) [26]. A diagnosis of MetS is made if at least three of the following five criteria are met: abdominal obesity (waist circumference ≥ 102 cm in males, or ≥ 88 cm in females); high triglycerides (serum triglycerides ≥ 150 mg/dL or 1.7 mmol/L), or use of lipid-lowering medications; low high-density lipoprotein cholesterol (HDL) (serum HDL cholesterol < 40 mg/dL in males, or < 50 mg/dL in females), or use of lipid-modifying medications (lipid-modifying medications are categorized under the low HDL group to avoid double counting); hypertension (systolic blood pressure (SBP) ≥ 130 mmHg, diastolic blood pressure (DBP) ≥ 85 mmHg, or use of antihypertensive treatment, or diagnosis by a physician), with both SBP and DBP calculated as the arithmetic mean of repeated measurements (up to four times); and high blood glucose (fasting glucose ≥ 100 mg/dL or 5.6 mmol/L), or use of antihyperglycaemic treatments, including insulin, or diagnosis with diabetes.

Assessment of outcome

Mortality data in NHANES were obtained from the publicly available Linked Mortality Files (LMF) released by the National Center for Health Statistics (NCHS), Centers for Disease Control and Prevention (CDC), accessible at https://www.cdc.gov/nchs/data-linkage/index.htm. These files linked NHANES participants to the National Death Index (NDI) using a probabilistic matching algorithm. Participants were followed from the date of survey participation until the date of death or the end of follow-up, whichever occurred first.

Mortality outcomes included all-cause mortality, cardiovascular disease mortality (CVDM), cancer-related mortality (CAM), other-cause mortality (OTM), and premature mortality. All-cause mortality was defined as death from any cause. CVDM was defined as death attributed to any cardiovascular disease, CAM as death due to any type of cancer, and OTM as death from all other causes. According to the WHO [27] and previous studies [28], premature mortality is commonly defined as death occurring before the age of 70. In addition, considering that 65 and 75 years are also widely used thresholds in epidemiological research [29], we included them in sensitivity analyses.

Covariates

Covariates included sex, age, BMI, race/ethnicity (NHANES: non-Hispanic White, other; UKB: White, other), household income (UKB: <£18,000, £18,000–£51,999, and ≥£52,000; NHANES: family income-to-poverty ratio levels: 0–1.0, 1.1–3.0, > 3.0), educational attainment (UKB: high-quality education (college or university degree), medium-quality education (including A levels/AS levels or equivalent, O levels/GCSEs or equivalent, CSEs or equivalent, NVQ or HND or HNC or equivalent, other professional qualifications such as nursing or teaching), and low-quality education (none of the above); NHANES: less than high school, high school or equivalent, college or above), smoking status (nonsemokers and smokers), alcohol consumption (nondrinkers and drinkers, defined as ≥ 1 drink/day for women and ≥ 2 drinks/day for men), physical activity (UKB: engaging in at least 150 min of moderate-intensity physical activity per week or at least 75 min of vigorous-intensity physical activity; NHANES: levels of moderate to vigorous leisure-time physical activity (0, 1–2, or ≥ 3 times/week)), healthy diet score (UKB: meeting five out of seven recommended intakes for fruits, vegetables, whole grains, fish, dairy, plant oils, refined grains, processed meats, unprocessed meats, and sugar-sweetened beverages; NHANES: meeting the top half of the Healthy Eating Index (HEI) distribution), family history (UKB: family history of diabetes, heart disease, and cancer; NHANES: family history of diabetes or myocardial infarction), and comorbidities (including cardiovascular disease, cancer, and chronic obstructive pulmonary disease).

Statistical methods

Baseline characteristics for continuous variables are reported as the means (SD: standard deviation), whereas categorical variables are presented as counts (percentages). Wilcoxon nonparametric tests and χ² tests were used to assess the relationships between continuous and categorical variables and social isolation. All analyses based on NHANES data were conducted using complex survey weighted methods, with the Mobile Examination Center weights applied.

We investigated the associations between social isolation and various mortality outcomes—including all-cause mortality, cause-specific mortality, and premature mortality—separately in populations with and without MetS. Cox proportional hazards models were used to estimate hazard ratios (HRs). Model 1 was unadjusted; Model 2 was adjusted for sex, age, BMI, and race; and Model 3 was further adjusted for income level, education level, smoking status, alcohol consumption, physical activity, dietary quality, family history of diabetes or heart disease, and comorbidities (including cardiovascular disease, cancer, and chronic obstructive pulmonary disease (COPD)).

Interaction terms were used to evaluate multiplicative interactions between social isolation and sex, age group, smoking status, drinking status, physical activity, dietary quality, health status, family history, and comorbidities, with statistically significant multiplicative interaction terms indicating interaction effects. As a sensitivity analysis, we further included the original linear variable of social isolation in the models using data from both the UK Biobank and NHANES.All analyses were conducted via R software (version 4.3.3), with p < 0.05 considered statistically significant. The results are reported with precision up to two decimal places, while p values < 0.001 are reported directly.

Results

Population characteristics

Table 1 shows that, based on the UK Biobank, among individuals with MetS, the all-cause mortality rate was 8.62% in the non-isolated group and 11.49% in the socially isolated group. CVDM rates were 1.38% and 2.03%, CAM rates were 4.10% and 4.86%, OTM rates were 3.13% and 4.61%, and premature mortality (< 70 years) rates were 1.79% and3.07%, respectively. Among those without MetS, the all-cause mortality rate was 4.56% in the non-isolated group and 6.42% in the socially isolated group; corresponding CVDM rates were 0.47% and 0.79%, CAM rates were 2.54% and 3.07%, OTM rates were 1.55% and 2.55%, and premature mortality rates were2.30% and 3.55%.

Table 1.

Baseline characteristics by social isolation status stratified by metabolic syndrome status in the UK biobank and NHANES populations

UK biobank
Population with Mets					Population without Mets
Characteristic	Overall N = 75,190	Not isolate N = 63,512	Social isolated N = 11,678	p	Characteristic	Overall N = 229,388	Not isolate N = 198,877	Social isolated N = 3 0,511	p
Age	58.26 (7.56)	58.41 (7.55)	57.42 (7.57)	< 0.001	Age	55.00 (8.11)	55.06 (8.14)	54.60 (7.87)	< 0.001
BMI (kg/m 2 )	30.69 (4.76)	30.62 (4.69)	31.08 (5.10)	< 0.001	BMI (kg/m 2 )	25.85 (3.74)	25.86 (3.69)	25.77 (4.01)	< 0.001
Sex				< 0.001	Sex				< 0.001
Female	32,176(42.79)	27,376(43.10)	4800 (41.10)		Female	123,159 (53.69)	108,421 (54.52)	14,738 (48.30)
Male	43,014(57.21)	36,136(56.90)	6878 (58.90)		Male	106,229 (46.31)	90,456 (45.48)	15,773 (51.70)
Race				< 0.001	Race				0.06
White	71,357(94.90)	60,430(95.15)	10,927(93.57)		White	220,321 (96.05)	191,523 (96.30)	28,798 (94.39)
Non-White	3833 (5.10)	3082 (4.85)	751 (6.43)		Non-white	9067 (3.95)	7354 (3.70)	1713 (5.61)
Education levels				< 0.001	Education levels				< 0.001
High-quality	13,398(17.82)	11,072(17.43)	2326(19.92)		High-quality	23,578 (10.28)	19,938 (10.03)	3640 (11.93)
Medium-quality	39,652(52.74)	33,794(53.21)	5858(50.16)		Medium-quality	112,046 (48.85)	98,213 (49.38)	13,833 (45.34)
Low-quality	22,140(29.45)	18,646(29.36)	3494(29.92)		Low-quality	93,764 (40.88)	80,726 (40.59)	13,038 (42.73)
Income levels				< 0.001	Income levels				< 0.001
<£18,000	20,255(26.94)	15,823(24.91)	4432(37.95)		<£18,000	39,295 (17.13)	31,112 (15.64)	8183 (26.82)
£18,000–£51,999	39,369(52.36)	34,101(53.69)	5268(45.11)		£18,000–£51,999	118,124 (51.50)	103,476 (52.03)	14,648 (48.01)
≥£52,000	15,566(20.70)	13,588(21.39)	1978(16.94)		≥£52,000	71,969 (31.37)	64,289 (32.33)	7680 (25.17)
No Smoking	44,112(58.67)	37,856(59.60)	6256(53.57)	< 0.001	No Smoking	160,618 (70.02)	140,736 (70.77)	19,882 (65.16)	< 0.001
Healthy drinking	40,221(53.49)	32,758(51.58)	7463(63.91)	< 0.001	Healthy drinking	102,862 (44.84)	86,074 (43.28)	16,788 (55.02)	< 0.001
High quality diet	17,520(23.30)	14,814(23.32)	2706(23.17)	0.719	High quality diet	66,845 (29.14)	58,017 (29.17)	8828 (28.93)	0.39
Physical active	56,312 (74.89)	48,876 (76.96)	7436 (63.68)	< 0.001	Physical active	190,152 (82.90)	168,018 (84.48)	22,134 (72.54)	< 0.001
Comorbidity	38,936 (51.78)	32,783 (51.62)	6153 (52.69)	0.033	Comorbidity	47,281 (20.61)	40,926 (20.58)	6355 (20.83)	0.315
Family history	38,975 (51.84)	33,047 (52.03)	5928 (50.76)	0.012	Family history	77,696 (33.87)	67,401 (33.89)	10,295 (33.74)	0.609
All-cause mortality	6817 (9.07)	5475 (8.62)	1342(11.49)	< 0.001	All-cause mortality	11,026 (4.81)	9068 (4.56)	1958 (6.42)	< 0.001
CVDM	1114 (1.48)	877 (1.38)	237 (2.03)	< 0.001	CVDM	1180 (0.51)	939 (0.47)	241 (0.79)	< 0.001
CAM	3174 (4.22)	2607 (4.10)	567 (4.86)	< 0.001	CAM	5987 (2.61)	5049 (2.54)	938 (3.07)	< 0.001
OTM	2529 (3.36)	1991 (3.13)	538 (4.61)	< 0.001	OTM	3859 (1.68)	3080 (1.55)	779 (2.55)	< 0.001
Premature death < 70	1382(1.98)	1055(1.79)	327 (3.07)	< 0.001	Premature death < 70	5529 (2.47)	4477 (2.30)	1052 (3.55)	< 0.001
NHANES
Characteristic	Overall N = 5758	Not isolate N = 4553	Social isolated N = 1205	p	Characteristic	Overall N = 7448	Not isolate N = 5868	Social isolated N = 1580	p
Age	64.08(10.74)	64.25(10.51)	63.07(12.00)	< 0.001	Age	61.97(11.91)	62.52(11.54)	58.50(13.51)	< 0.001
BMI (kg/m 2 )	32.59(6.23)	32.42(6.06)	33.59(7.06)	< 0.001	BMI (kg/m 2 )	25.96(5.00)	25.99(4.86)	25.77(5.65)	0.024
Sex				< 0.001	Sex				0.013
Female	2356(42.20)	1991(44.50)	365(28.60)		Female	4107(50.40)	3286(51.20)	821(45.50)
Male	3402(57.80)	2562(55.50)	840(71.40)		Male	3341(49.60)	2582(48.80)	759(54.50)
Race				0.06	Race				0.033
Non-Hispanic White	2903(76.90)	2235(22.70)	620(25.90)		Non-hispanic white	3842(78.40)	2817(21.10)	789(24.70)
Others	2855(23.10)	2318(77.30)	585(74.10)		Others	3606(21.60)	3051(78.90)	791(75.30)
Education levels				< 0.001	Education levels				< 0.001
Below high school	1859(20.60)	1383(19.40)	476(27.20)		Below high school	2070(16.90)	1543(15.80)	527(23.90)
High school and equal	1950(31.00)	1482(30.10)	468(36.50)		High school and equal	2506(29.30)	1848(27.80)	658(38.70)
Above high school	1949(48.40)	1688(50.40)	261(36.3)		Above high school	2872(53.80)	2477(56.40)	395(37.40)
Poverty income ratio				< 0.001	Poverty income ratio				< 0.001
0–1.0	746(12.30)	558(10.57)	188(22.50)		0–1.0	624(9.20)	482(8.20)	142(15.60)
1.1-3.0	1551(34.50)	1340(35.20)	211(30.90)		1.1-3.0	1366(25.20)	1237(26.30)	129(18.20)
> 3.0	3461(53.20)	806(54.30)	806(46.60)		> 3.0	5458(65.60)	1309(65.50)	1309(66.20)
No Smoking	2725(44.80)	2182(45.40)	543(41.10)	0.04	No Smoking	3269(44.60)	2711(46.20)	558(34.00)	< 0.001
Healthy drinking	5038(84.60)	3967(84.50)	1071(85.20)	0.66	Healthy drinking	6566(84.80)	5213(85.70)	1353(78.60)	< 0.001
High quality diet	2986(51.20)	2399(52.40)	587(44.10)	0.001	High quality diet	4061(55.00)	3318(56.30)	743(6.80)	< 0.001
Physical active	935(13.40)	742(13.70)	193(12.16)	0.87	Physical active	1310(15.40)	998(15.00)	312(17.50)	0.18
Comorbidity	2166(41.60)	1666(40.70)	500(44.10)	0.60	Comorbidity	2182(32.90)	1726(33.00)	456(31.70)	0.52
Family history	1108(23.70)	865(23.40)	243(25.70)	0.57	Family history	1106(19.90)	880(19.60)	880(19.60)	0.28
All-cause mortality	2453(26.20)	1758(23.40)	695(42.90)	< 0.001	All-cause mortality	3201(25.80)	2276(22.50)	925(46.20)	< 0.001
CVDM	941(9.24)	679(8.31)	262(14.70)	0.04	CVDM	1152(8.13)	803(7.07)	349(14.80)	< 0.001
CAM	526(6.15)	399(5.70)	127(8.81)	0.66	CAM	738(6.31)	526(5.54)	212(11.15)	< 0.001
OTM	986(10.85)	680(9.39)	306(19.40)	0.001	OTM	1311(11.30)	947(9.94)	364(20.20)	< 0.001
Premature death < 70	1329(13.90)	934(12.20)	395(24.10)	< 0.001	Premature death < 70	1658(14.10)	1141(11.70)	517(28.70)	< 0.001

BMI: Body Mass Index; CAM: Cancer Mortality; CVDM: Cardiovascular Disease Mortality; Mets: Metabolic Syndrome; NHANES: National Health and Nutrition Examination Survey; OTM: Other-cause Mortality

*For continuous variables, we represent them as mean (standard deviation), while categorical variables are expressed as counts (percentages)

P-value assessment: Continuous variables are compared using the Kruskal-Wallis rank sum test, and categorical variables are assessed using the Chi-squared test. Additionally, bold text is used to highlight variable names, subgroup headers, and population strata, and carries no statistical implication

Based on the U.S. NHANES data, among participants with MetS, the all-cause mortality rate was 23.40% in the non-isolated group and 42.90% in the socially isolated group. The CVDM rates were 8.31% and 14.70%, CAM rates were 5.70% and 8.81%, OTM rates were 9.39% and 19.40%, and premature mortality rates were 12.20% and 24.10%, respectively. Among those without MetS, the all-cause mortality rate was 22.50% in the non-isolated group and 46.20% in the socially isolated group; corresponding CVDM rates were 7.07% and 14.80%, CAM rates were 5.54% and 11.15%, OTM rates were 9.94% and 20.20%, and premature mortality rates were 11.70% and 28.70%, respectively.

Cox analysis

As shown in Table 2, fully adjusted Cox regression models (Model 3) demonstrated that social isolation was significantly associated with increased risk of all-cause mortality, cause-specific mortality, and premature mortality before age 70, regardless of MetS status, in both the UK Biobank and NHANES cohorts. For all-cause mortality, the HRs for socially isolated versus non-isolated individuals were: UK Biobank – MetS: HR = 1.30 (95% CI: 1.25–1.43), non-MetS: HR = 1.51 (95% CI: 1.43–1.58); NHANES – MetS: HR = 1.14 (95% CI: 1.07–1.21), non-MetS: HR = 1.60 (95% CI: 1.39–1.83). For CVDM, the HRs were: UK Biobank – MetS: HR = 1.21 (95% CI: 1.12–1.31), non-MetS: HR = 1.75 (95% CI: 1.52–2.02); NHANES – MetS: HR = 1.54 (95% CI: 1.20–1.97), non-MetS: HR = 1.75 (95% CI: 1.39–2.18). For CAM, the HRs were: UK Biobank – MetS: HR = 1.12 (95% CI: 1.07–1.17), non-MetS: HR = 1.30 (95% CI: 1.22–1.40); NHANES – MetS: HR = 1.48 (95% CI: 1.11–1.98), non-MetS: HR = 1.47 (95% CI: 1.14–1.90). For OTM, the HRs were: UK Biobank – MetS: HR = 1.38 (95% CI: 1.26–1.53), non-MetS: HR = 1.76 (95% CI: 1.63–1.91); NHANES – MetS: HR = 1.71 (95% CI: 1.39–2.09), non-MetS: HR = 1.56 (95% CI: 1.27–1.91). For premature mortality before age 70, the HRs were: UK Biobank – MetS: HR = 1.39 (95% CI: 1.27–1.51), non-MetS: HR = 1.54 (95% CI: 1.44–1.65); NHANES – MetS: HR = 1.09 (95% CI: 1.00–1.18), non-MetS: HR = 1.39 (95% CI: 1.15–1.69). Supplementary Tables 2 and 3 further present the detailed associations between social isolation and mortality outcomes among individuals with and without MetS in the UK Biobank and NHANES cohorts, both before and after covariate adjustment. Supplementary Table 3 presents detailed information on basic metabolic-related indicators for individuals with and without MetS from the two major databases.

Table 2.

Multivariable Cox regression analyses of social isolation and all-cause, cause-specific, and premature mortality stratified by metabolic syndrome status in the UK biobank and NHANES cohorts

UK biobank					NHANES
	Mets population		Without Mets			Mets population		Without Mets
	HR (95%CI)	p	HR (95%CI)	p		HR (95%CI)	p	HR (95%CI)	p
All-cause mortality					All-cause mortality
Not isolated	Ref	Ref	Ref	Ref	Not isolated	Ref	Ref	Ref	Ref
Isolated	1.30 (1.23 to 1.39)	<0.001	1.51 (1.43 to 1.58)	<0.001	Isolated	1.14(1.07 to 1.21)	<0.001	1.60(1.39 to 1.83)	<0.001
Cardiovascular disease mortality					Cardiovascular disease mortality
Not isolated	Ref	Ref	Ref	Ref	Not isolated	Ref	Ref	Ref	Ref
Isolated	1.21 (1.12 to 1.31)	<0.001	1.75 (1.52 to 2.02)	<0.001	Isolated	1.54(1.20 to 1.97)	<0.001	1.75(1.39 to 2.18)	<0.001
Cancer mortality					Cancer mortality
Not isolated	Ref	Ref	Ref	Ref	Not isolated	Ref	Ref	Ref	Ref
Isolated	1.12 (1.07 to 1.17)	<0.001	1.30 (1.22 to 1.40)	<0.001	Isolated	1.48(1.11 to 1.98)	0.007	1.47(1.14 to 1.90)	0.003
Other-cause mortality					Other-cause mortality
Not isolated	Ref	Ref	Ref	Ref	Not isolated	Ref	Ref	Ref	Ref
Isolated	1.38 (1.26 to 1.53)	<0.001	1.76 (1.63 to 1.91)	<0.001	Isolated	1.71(1.39 to 2.09)	<0.001	1.56(1.27 to 1.91)	<0.001
Premature death < 70					Premature death < 70
Not isolated	Ref	Ref	Ref	Ref	Not isolated	Ref	Ref	Ref	Ref
Isolated	1.39 (1.27 to 1.51)	<0.001	1.54 (1.44 to 1.65)	<0.001	Isolated	1.09(1.00 to 1.18)	0.04	1.39(1.15 to 1.69)	<0.001

CI: Confidence Interval; HR: Hazard Ratio; Ref: reference; Mets: Metabolic Syndrome

This model adjusted for sex, age, body mass index (BMI), race, income level, education level, smoking status, alcohol consumption, physical activity, dietary quality, family history of diabetes or heart disease, and comorbidities (including cardiovascular disease, cancer, and chronic obstructive pulmonary disease). Additionally, in the UK Biobank cohort, adjustment was made for assessment center. Complete tables of unadjusted and adjusted models are provided in Supplementary Tables 1 and 2

Interaction, subgroup and sensitivity analyses

Figures 1 and 2 illustrate the interaction effects between social isolation and common lifestyle factors on various mortality outcomes, as well as the subgroup-specific associations of social isolation with mortality, stratified by MetS status in the UK Biobank and NHANES populations. Subgroup analyses and interaction tests based on NHANES (Fig. 2) showed that, among individuals with MetS, social isolation had significant interactions with non-smoking (p for interaction = 0.01) and non-drinking (p for interaction = 0.01) in relation to CVDM. Moreover, results from both the UK Biobank and NHANES consistently indicated that the adverse impact of social isolation on mortality was more pronounced in individuals without metabolic syndrome (non-MetS). Supplementary Tables 4 and 5 provide detailed results of interaction and subgroup analyses for all mortality outcomes.

Fig. 1.

Interaction and subgroup analyses stratified by metabolic syndrome status in the UK biobank population. Note: (a) Individuals with metabolic syndrome; (b) Individuals without metabolic syndrome. HR = Hazard Ratio

Fig. 2.

Interaction and subgroup analyses stratified by metabolic syndrome status in the NHANES population. Note: (a) Individuals with metabolic syndrome; (b) Individuals without metabolic syndrome. HR = Hazard Ratio; NHANES = National Health and Nutrition Examination Survey

The sensitivity analysis results in Supplementary Table 6 are consistent with previous findings, confirming that the risk associated with social isolation for various mortality outcomes remains consistent in both MetS and non-MetS populations. Furthermore, the risk is more pronounced in the non-MetS group for all-cause mortality, CVDM, and premature death outcomes (all p < 0.001).

Discussion

This study comprehensively examined the associations between social isolation and various mortality outcomes such as all-cause mortality, CVDM, CAM, OTM, and premature mortality in populations with and without MetS using two large-scale cohorts: the UK Biobank and NHANES. Our findings demonstrate that social isolation is consistently associated with increased mortality risks across multiple causes of death regardless of MetS status.

In both cohorts, individuals with metabolic syndrome consistently exhibited higher mortality rates than those without, further supporting the notion that metabolic syndrome is a significant risk factor for mortality. In both the UK Biobank and NHANES, socially isolated individuals exhibited significantly elevated mortality rates across all measured endpoints compared to their non-isolated counterparts, emphasizing the detrimental role of social isolation as an independent risk factor.

Interestingly, the associations between social isolation and mortality outcomes—particularly all-cause mortality, cardiovascular mortality, and premature mortality—were more pronounced in the non-MetS populations. This suggests that while MetS substantially elevates baseline mortality risk, social isolation may exert an even stronger relative influence on mortality among individuals without metabolic abnormalities.

The robustness of these associations across multiple subgroup analyses strengthens the validity of our findings. Our results extend previous research linking social isolation to mortality in elderly and chronically ill populations by demonstrating its broad relevance across metabolic health statuses.

Previous studies have confirmed the associations between social isolation and all-cause mortality in elderly individuals [30] and obese individuals [31]. Although the underlying mechanisms remain unclear, they can still be explained from the following perspectives. First, studies have shown that individuals experiencing social isolation tend to have unhealthy lifestyles [32], including smoking, physical inactivity, and poor dietary habits, and may also neglect health-protective behaviours, such as adhering to medical recommendations. Healthy social interactions and support can help reduce stress, encourage healthy behaviours, and provide emotional support [33]. In contrast, individuals who live alone or lack social connections may face a greater risk of death if they develop acute symptoms, as they may not have a strong network of close companions to encourage them to seek medical care [31, 34]. Moreover, several studies have shown that social isolation can activate the hypothalamic‒pituitary‒adrenal axis [35], enhance sympathetic activity, impair parasympathetic function, and stimulate a proinflammatory immune response and oxidative stress [36]. Animal research on adult baboons further indicates that relative social isolation is associated with elevated basal cortisol levels [37]. Furthermore, experimental studies in prairie voles have shown that both acute (e.g., one hour) and repeated acute (e.g., one hour daily for four weeks) social isolation significantly increase corticosterone levels [38].

Moreover, this study found that among individuals without metabolic syndrome, the adverse effects of social isolation on certain mortality risks—such as all-cause mortality, cardiovascular mortality, and premature death—were more pronounced. While previous research and clinical attention have primarily focused on high-risk populations—such as older adults [39] or individuals with pre-existing chronic conditions [40, 41]. However, individuals without MetS may receive less medical attention due to the absence of overt clinical symptoms. The impact of social isolation on mortality may not be limited to the regulation of endocrine and metabolic mechanisms. Studies [35] have shown that social isolation can also induce functional changes in brain regions, such as enhancing amygdala-mediated social threat responses and impairing regulatory functions of the prefrontal cortex. These neural mechanisms may further exacerbate individuals’ psychosocial stress responses, thereby affecting health outcomes.

The strengths of this study include several aspects. First, this study conducted a comprehensive assessment comparing the impact of social isolation on mortality risk between individuals with and without MetS, revealing the differential effects of social isolation on mortality across varying metabolic health statuses. Second, the inclusion of samples from two major population-based cohorts in the U.K. and the U.S. enhances the persuasiveness and evidential strength of the findings.

Nevertheless, the study has certain limitations. First, there are differences in the assessment of social isolation before and after NHANES III. Although the methods we used have been validated by previous research [21], this discrepancy requires cautious interpretation. Second, while cohort studies offer relatively strong causal inference, the lack of randomized intervention trials (RCTs) limits the strength of the evidence. Future RCTs and clinical studies are still needed. Third, despite adjusting for several potential covariates in our analysis, residual confounding may still exist. Finally, the UK Biobank, while one of the largest biomedical databases in the UK, does not fully represent the general UK population. It should be noted that 94% of participants are of European white ancestry, which may limit the generalizability of findings to other ethnic groups and populations outside the UK.

Conclusion

Social isolation is strongly associated with increased risks of all-cause, cardiovascular, cancer, and premature mortality, regardless of MetS status. Beyond managing physiological factors, interventions should focus on strengthening social connections and support. Notably, the negative impact of social isolation on outcomes such as all-cause, cardiovascular, and premature mortality is more pronounced in individuals without MetS. Public health strategies should adopt population-wide approaches, rather than targeting only high-risk groups, to reduce social isolation, enhance social engagement, and improve overall health and longevity.

Abbreviations

AHA

The American heart association

CDC

Centers for disease control and prevention

COPD

Chronic obstructive pulmonary disease

DBP

Diastolic blood pressure

HDL

High-density lipoprotein cholesterol

HRs

Hazard ratios

ICD-10

10th revision codes

IDF

International diabetes federation

MetS

Metabolic syndrome

NCDs

Noncommunicable diseases

NHANES

National health and nutrition examination survey

NHLBI

National heart, lung, and blood institute

SBP

Systolic blood pressure (SBP)

SD

Standard deviation

Author contributions

YZL, TTL, and QQG are joint corresponding authors. Conceptualization, Writing – original draft: YZL and SYL. Supervision: QQG, TTL and YZL. Formal analysis, Software, and Visualization: SYL and CHL. Data curation, Methodology, Validation, and Writing – review and editing: All authors. Funding acquisition: YZL. All authors approved the final version before submission. YZL is the guarantor and attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.

Funding

This work is supported by the National Natural Science Foundation of China (Grant No. 82470826) and the National Science and Technology Major Project (Grant No. 2024ZD0533403). The funder of the study had no role in the study design, data collection, data analysis, data interpretation, or the writing of the report. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Data availability

The data for this study were sourced from the UK Biobank and NHANES. UK Biobank data are available upon reasonable request, and the official website can be accessed at: [https://www.ukbiobank.ac.uk/](https://www.ukbiobank.ac.uk/). The project IDs involved in this study are 98937 and 206612. NHANES data can be downloaded directly from the official website: [https://www.cdc.gov/nchs/](https://www.cdc.gov/nchs/).

Declarations

Ethics approval and consent to participate

This study utilized publicly accessible secondary data that does not include any identifiable information.

Competing interests

The authors declare no competing interests.