Obesity-related patterns of multimorbidity and their association with health-related quality of life in Iranian adults: a cross-sectional study

Abstract

Objective

To identify patterns of non-communicable disease (NCD) multimorbidity among Iranian adults and assess their associations with sociodemographic factors and health-related quality of life (HRQOL).

Design

Cross-sectional analysis using data from the 2021 Iranian STEPS Survey, which follows the WHO STEPwise approach to collecting information on NCD risk factors.

Setting

A community-based, nationally representative household survey conducted across Iran.

Participants

A total of 17 517 Iranian adults with complete survey data were included. Of these, 56.7% were women and 67.4% lived in urban areas.

Primary outcome measures

Latent class analysis was used to identify multimorbidity clusters based on chronic disease profiles, including myocardial infarction, stroke, asthma/chronic obstructive pulmonary disease, cancer, obesity (abdominal and defined by the body mass index), hypertension, diabetes, chronic kidney disease and dyslipidaemia. Associations between cluster membership, demographic characteristics and HRQOL (measured using the EuroQol Visual Analogue Scale, EQ-VAS) were examined using multinomial logistic regression.

Results

Four distinct multimorbidity clusters were identified: obesity with severe metabolic syndrome (OSMS, 13.0%), obesity with early metabolic syndrome (OEMS, 39.3%), low comorbidity (LC, 42.4%) and non-obese cardiometabolic multimorbidity (NOCM, 5.2%). Cluster membership varied significantly by sex, age, education, occupation and insurance status. Older age, female sex, lower education, unpaid work and urban residence were associated with OSMS and OEMS, while male sex, older age, retirement and complementary insurance were associated with NOCM. EQ-VAS scores were significantly lower in the OSMS (β = −8.5; 95% CI −10.1 to −6.9), OEMS (β = −2.1; 95% CI −3.2 to −1.1) and NOCM (β = −6.2; 95% CI −8.2 to −4.2) clusters compared with the LC cluster.

Conclusions

Multimorbidity clusters in Iran are heterogeneous and linked to demographic and socioeconomic factors. Obesity and social disadvantage strongly shape disease patterns and HRQOL. Findings highlight the need for integrated, equity-oriented policies addressing both clinical complexity and social determinants through prevention and coordinated care.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• This study is based on a nationally representative 2021 WHO STEPS Survey with rigorous sampling, standardised protocols, and objective clinical and laboratory measurements, ensuring robust and generalisable findings for Iranian adults.

• The study used latent class analysis to identify data-driven multimorbidity clusters, combining statistical fit indices and clinical interpretability for model selection.

• The study applied complex survey weighting and adjusted multinomial regression models to account for population structure and confounding.

• The cross-sectional design of the study limits causal inference and temporal interpretation of associations between risk factors, multimorbidity and health-related quality of life (HRQOL).

• Reliance on some self-reported diagnoses, exclusion of certain chronic and mental conditions, use of the EuroQol Visual Analogue Scale alone for HRQOL, and complete-case analysis may introduce misclassification and selection bias.

Introduction

Over recent decades, the global burden of disease has shifted from communicable to non-communicable diseases (NCDs), with this transition occurring most rapidly in low-income and middle-income countries (LMICs).¹ Currently, NCDs are responsible for about 74% of all deaths worldwide, and four major groups—cardiovascular diseases (CVDs), cancers, chronic respiratory diseases and diabetes mellitus (DM)—account for more than 80% of these fatalities.² Consistent with global patterns, Iran recorded approximately 310 000 NCD-related deaths in 2021, alongside an estimated 80 million prevalent cases, equivalent to one chronic condition per resident.³ This immense burden places increasing pressure on the nation’s healthcare system.

An especially complex aspect of this burden is multimorbidity, defined as the coexistence of two or more chronic conditions in a single individual.⁴ In high-income countries, between 65% and 98% of adults aged 65 years and older live with multiple chronic diseases.⁵ While national data in Iran remain limited, available evidence suggests a comparable concern: multimorbidity affects between 19.4% and 21.1% of Iranian adults overall, and 40% of those aged 65 years and above, with individuals over 60 years having more than twice the odds compared with those under 50 years.^{6 7} Multimorbidity not only adds to the health burden but also amplifies it by significantly reducing individuals’ quality of life and increasing healthcare utilisation, prolonging hospital stays and increasing the frequency of outpatient visits.^{5 8} Despite this growing challenge, Iran’s healthcare infrastructure remains largely disease-specific. Clinical guidelines, financing structures and health information systems typically target individual conditions such as hypertension, DM or cancer-in isolation,⁹ mirroring the global pattern.^10–12 This fragmented, vertical approach lacks the integrated care pathways necessary for effective multimorbidity management, often leaving physicians without practical tools to treat patients with complex, overlapping conditions. Moreover, the few existing Iranian studies on multimorbidity are generally small, hospital-based or condition-specific and therefore fail to characterise population-level patterns of chronic disease co-occurrence or assess their impact on quality of life.^{7 9}

An essential yet underexplored aspect of multimorbidity is how chronic conditions cluster within individuals. These clusters may reflect shared biological, behavioural or social determinants, and their identification can inform screening programmes, guide the development of integrated care pathways for patients with similar risk profiles, and enable more efficient resource allocation. Studying these clusters in relation to health-related quality of life (HRQOL) provides additional insight by identifying the groups that experience the greatest burden, thereby helping to prioritise high-risk populations for targeted interventions and supportive care. However, no nationally representative study has yet investigated multimorbidity clustering patterns in Iran or assessed their relationship with HRQOL. To address this gap, we applied latent class analysis (LCA) to data from the nationally representative 2021 Iranian STEPwise Approach to NCD Risk Factor Surveillance (STEPS) survey.¹³ LCA is a robust statistical method for identifying unobserved population subgroups based on shared characteristics. Specifically, we aimed to identify common patterns of co-occurring NCDs (disease clusters), examine their variation across age, sex and sociodemographic factors, and assess the HRQOL profile associated with each cluster. By translating complex multimorbidity patterns into clear, actionable clinical groupings, and spotlighting those clusters most strongly associated with poor quality of life, this study aims to equip policy makers with evidence to guide priority setting, resource allocation and the design of integrated NCD strategies tailored to the Iranian context.

Methods

Study design and participants

This study analysed data collected from the Iran STEPS 2021 Survey, which provides a nationally and regionally representative sample.¹³ The survey followed the WHO’s STEPwise approach for gathering data on risk factors related to NCDs. The study was divided into three components: (1) Questionnaire-based information gathering, (2) Physical measurements of participants, and (3) Laboratory tests. A systematic cluster random sampling method was employed to ensure that the sample accurately reflected the demographic distribution at both the national and provincial levels, with appropriate adjustments made for relative population sizes. A total of 28 821 individuals aged ≥18 years were selected for the survey, of whom 27 874 completed the questionnaire and 27 745 underwent physical measurements. Laboratory assessments were performed only among participants aged ≥25 years, with 18 119 individuals completing this step. Individuals unable to complete the questionnaire because of psychological conditions, those with physical limitations preventing participation in measurements or laboratory tests, and pregnant women were excluded.¹³ For the analyses in this study, we considered a complete-case approach and included only participants with no missing values for any of the variables included, resulting in a final analytical sample of 17 517 individuals (figure 1).

Figure 1. Flow diagram of selection of study participants.

Data collection

The questionnaire covered a range of topics, including demographic information, lifestyle habits, medical history, diet, HRQOL, as well as information about household assets. In the second phase, participants underwent a series of physical measurements. These included the recording of their height, weight, waist and hip measurements, blood pressure and heart rate. Height was measured using a standard measuring tape, ensuring that the participant stood upright against a wall, aligning their body properly. Weight was measured with a calibrated digital scale (Inofit), with participants wearing light clothing and no shoes. Waist circumference was measured using a non-stretchable tape at the midpoint between the lowest rib and the top of the iliac crest, with the participant standing upright and relaxed. Blood pressure was recorded after participants rested in a seated position for 15 min, with three separate measurements taken at 3 min intervals. The final blood pressure value was determined by averaging the second and third readings. The third phase involved laboratory testing, which included the analysis of total serum cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), fasting plasma glucose (FPG), haemoglobin A1c and the urine albumin-to-creatinine ratio (UACR). Samples were transported under appropriate cold chain conditions to a central laboratory for analysis.

Definition of variables

Several demographic variables were considered, including age, categorised into three epidemiologically relevant groups: 25–40 years, 41–60 years and above 60 years, broadly representing younger, middle-aged and older adults, respectively.^{14 15} Residency was categorised based on whether participants lived in urban or rural areas. Education level was classified by the number of completed school years, grouped as 0, 1–6, 7–11 and 12 or more years. The wealth index was derived using principal component analysis based on household asset ownership¹³ and divided into five quintiles, ranging from the poorest (quintile 1) to the wealthiest (quintile 5). Marital status was categorised as single, married, widowed and divorced. Occupation was categorised as employed, engaged in unpaid work, retired or unemployed. Insurance status was categorised as no insurance, basic insurance or complementary insurance.

Abdominal obesity was defined by a waist circumference of >102 cm in men and >88 cm in women. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in metres. It was classified as <25 kg/m² (underweight/normal), 25 kg/m² to <30 kg/m² (overweight) and ≥30 kg/m² (obese).¹⁶ Hypertension was defined as having a systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or the use of any antihypertensive medication.¹⁷ DM was defined as having an FPG ≥126 mg/dL or the use of any antidiabetic drugs.¹⁸ Dyslipidaemia was defined based on several criteria: total TC ≥200 mg/dL, TG ≥150 mg/dL, LDL-C ≥130 mg/dL, non-HDL-C ≥160 mg/dL or HDL-C <40 mg/dL in men and <50 mg/dL in women, or the use of any lipid-lowering drugs.¹⁹ Chronic kidney disease (CKD) was defined as having an estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m² or a UACR ≥30 mg/g.²⁰ The eGFR was calculated using the CKD-EPI Creatinine Equation (2009).²¹ Myocardial infarction (MI) was defined by a positive response to the question: ‘Have you ever been told by a doctor or healthcare worker that you are having a heart attack?’. Stroke was defined as responding ‘yes’ to the question: ‘Have you ever been told by a doctor or healthcare worker that you have had a stroke?’. CVD was defined as a positive response to either MI or stroke questions. Cancer was assessed with the question: ‘During the past 12 months, did a doctor or other health worker tell you that you had cancer?’. Asthma/chronic obstructive pulmonary disease (COPD) were identified by the question: ‘During the past 12 months, did a doctor or other health worker tell you that you had asthma or COPD?’.

Quality of life was measured using the EuroQol Visual Analogue Scale (EQ-VAS), a component of the EuroQol Five-Dimension Scale. Participants rated their current overall health on a vertical scale from 0 (worst imaginable) to 100 (best imaginable).

Data analysis

Data cleaning and weighting procedures were conducted in accordance with the STEPS 2021 Study protocol to ensure representativeness of both national and provincial populations. A four-step weighting process was applied: (1) Adjustment for overall non-response, (2) Adjustment for non-response at each survey stage, (3) Weighting by age, sex and geographical area within each province, and (4) Final calibration incorporating all previous weights. All prevalence estimates reported are weighted and accompanied by 95% CIs. Baseline participant characteristics are presented as means with SE for continuous variables, and as percentages with corresponding 95% CIs for categorical variables, stratified by sex.

LCA was performed using the R package ‘poLCA’. Ten NCD multimorbidities served as observed indicators: MI, stroke, COPD/asthma, cancer, abdominal obesity, BMI-defined obesity, hypertension, DM, CKD and dyslipidaemia. The optimal number of latent classes was determined by comparing models based on the adjusted Bayesian Information Criterion (BIC) and the Akaike Information Criterion (AIC), both widely recognised as robust metrics for class enumeration in categorical data. The model exhibiting the lowest adjusted BIC and AIC values was selected as the best fit. In addition to statistical criteria, clinical interpretability and expert judgement informed the final model selection. On selection of the optimal model, each participant was assigned to the latent class corresponding to their highest posterior probability of membership. Average posterior probabilities exceeding 70% were considered indicative of good model fit. Latent classes were labelled according to the NCDs whose prevalence within each class exceeded that observed in the overall sample. The identified latent classes were subsequently compared in terms of comorbidities and baseline demographic and sociodemographic characteristics using weighted prevalence estimates and 95% CIs (online supplemental table 1).

To explore associations between sociodemographic as well as lifestyle factors and multimorbidity classes, multinomial logistic regression models were employed. ORs with 95% CIs were reported, using Cluster 3, the healthiest class, as the reference group. Furthermore, linear regression models were fitted to examine the association between multimorbidity classes and HRQOL, using EQ-VAS as a continuous outcome. Results are presented as β coefficients with 95% CIs, again referencing Cluster 3. Three sequential models were constructed for regression analyses: Model 1 was unadjusted; Model 2 adjusted for age and sex; Model 3 further adjusted for residency, wealth index, marital status, education and occupation.

All statistical analyses were conducted using Stata V.17 (StataCorp, College Station, Texas, USA) and R V.4.3.1. A two-sided value of p<0.05 was considered statistically significant.

Patient and public involvement

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

Results

Baseline characteristics of the study population

Table 1 displays baseline characteristics for 17 517 participants, mostly middle-aged, with 34.5% (n=6045) aged 25–40 years and 43.9% (7685) aged 41–60 years. About 56.72% (9937) are women and 67.4% (11 802) reside in urban areas. Wealth classes were distributed evenly. Regarding employment, 36.3% were employed and 49.1% reported unpaid work. The majority are married (83.3%, n=14 598) and have some form of insurance, with 64.3% (11 268) holding basic coverage.

Table 1. Baseline characteristics of the study population.

Variables	Women n=9937	Men n=7580	Total n=17 517
Age, years
25–40	3561 (35.8)	2484 (32.8)	6045 (34.5)
41–60	4445 (44.7)	3240 (42.7)	7685 (43.9)
above 60	1931 (19.4)	1856 (24.5)	3787 (21.6)
Residency
Rural	3255 (32.8)	2460 (32.5)	5715 (32.6)
Urban	6682 (67.2)	5120 (67.6)	11 802 (67.4)
Education
Zero	2173 (22.0)	899 (11.9)	3072 (17.7)
1–6 years	2981 (30.2)	1902 (25.3)	4883 (28.1)
7–11 years	1598 (16.2)	1664 (22.1)	3262 (18.7)
12 years and over	3124 (31.6)	3065 (40.7)	6189 (35.6)
Wealth Index
Class 1 (poorest)	2129 (23.3)	1429 (19.2)	3558 (21.4)
Class 2	1897 (20.7)	1342 (18.0)	3239 (19.5)
Class 3	1871 (20.5)	1686 (22.6)	3557 (21.4)
Class 4	1751 (19.1)	1613 (21.6)	3364 (20.3)
Class 5 (wealthiest)	1498 (16.4)	1384 (18.6)	2882 (17.4)
Marital status
Single	715 (7.2)	640 (8.4)	1355 (7.7)
Married	7844 (78.9)	6754 (89.1)	14 598 (83.3)
Divorced/separated with partner	297 (3.0)	93 (1.2)	390 (2.2)
Widow	1081 (10.9)	93 (1.2)	1174 (6.7)
Occupation
Employed	1048 (10.6)	5264 (69.9)	6312 (36.3)
Unpaid work	8406 (85.1)	136 (1.8)	8542 (49.1)
Retired	243 (2.5)	1443 (19.2)	1686 (9.7)
Unemployed	179 (1.8)	687 (9.1)	866 (5.0)
Insurance
No	782 (7.9)	748 (9.9)	1530 (8.7)
Basic	6495 (65.4)	4773 (63.0)	11 268 (64.3)
Complementary	2660 (26.8)	2059 (27.2)	4719 (26.9)

Clusters description

Cluster 1: Obesity with severe metabolic syndrome

About 13% of participants (n=2278) belong to the obesity with severe metabolic syndrome (OSMS) cluster. They show extreme obesity, abdominal obesity (96.3%; 95% CI 95.09% to 97.20%) and BMI-defined obesity (55.5%; 95% CI 52.46% to 58.44%), together with a severe metabolic profile: hypertension (92.4%; 95% CI 90.69% to 93.83%), DM (61.0%; 95% CI 57.96% to 63.98%), CKD (49.8%; 95% CI 46.75% to 52.79%) and dyslipidaemia (95.2%; 95% CI 93.42% to 96.47%), plus frequent cardiovascular events (table 2, figure 2).

Table 2. Prevalence of non-communicable diseases in each cluster.

	OSMS	OEMS	LC	NOCM
Cancer	1.80 (1.30 to 2.50)	1.44 (1.12 to 1.86)	1.04 (0.73 to 1.48)	3.19 (2.18 to 4.66)
Myocardial infarction	32.87 (30.13 to 35.73)	1.74 (1.28 to 2.35)	1.84 (1.50 to 2.26)	40.32 (35.91 to 44.91)
Stroke	5.68 (4.54 to 7.07)	0.00	0.28 (0.17 to 0.44)	11.88 (9.2 to 15.18)
Cardiovascular disease	35.21 (32.43 to 38.09)	1.74 (1.28 to 2.35)	2.12 (1.76 to 2.56)	45.01 (40.49 to 49.61)
Asthma and COPD	5.28 (4.00 to 6.95)	3.10 (2.62 to 3.67)	1.91 (1.53 to 2.38)	3.90 (2.73 to 5.55)
Abdominal obesity	96.28 (95.09 to 97.20)	95.13 (94.23 to 95.90)	3.36 (2.79 to 4.05)	5.14 (3.21 to 8.12)
BMI obesity	55.47 (52.46 to 58.44)	50.76 (49.10 to 52.41)	0.31 (0.20 to 0.50)	0.04 (0.01 to 0.25)
Hypertension	92.41 (90.69 to 93.83)	32.17 (30.61 to 33.77)	16.09 (14.98 to 17.27)	85.54 (82.31 to 88.26)
Diabetes	61.01 (57.96 to 63.98)	5.93 (5.19 to 6.77)	3.29 (2.72 to 3.99)	46.50 (41.96 to 51.09)
Chronic kidney disease	49.77 (46.75 to 52.79)	3.16 (2.63 to 3.79)	3.63 (3.04 to 4.34)	51.80 (47.22 to 56.34)
Dyslipidaemia	95.17 (93.42 to 96.47)	94.97 (94.13 to 95.70)	77.22 (75.86 to 78.52)	86.89 (83.72 to 89.52)
Cluster size, %	13 %	39.3 %	42.4 %	5.2 %

Values are reported as weighted percentages (95% CIs).

BMI, body mass index; COPD, chronic obstructive pulmonary disease; LC, low comorbidity; NOCM, non-obese cardiometabolic multimorbidity; OEMS, obesity with early metabolic syndrome; OSMS, obesity with severe metabolic syndrome.

Figure 2. Non-communicable disease prevalence in each cluster: (A) Heatmap and (B) grouped bar chart showing the weighted prevalence (%) of non-communicable disease indicators across clusters. BMI, body mass index; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CVD, cardiovascular disease; HTN, hypertension; LC, low comorbidity; MI, myocardial infarction; NOCM, non-obese cardiometabolic multimorbidity; OEMS, obesity with early metabolic syndrome; OSMS, obesity with severe metabolic syndrome.

Cluster 2: Obesity with early metabolic syndrome

Representing 39.3% of the cohort (n=6879), the obesity with early metabolic syndrome (OEMS) cluster also displays high abdominal obesity (95.1%; 95% CI 94.23% to 95.90%) and BMI-defined obesity (50.8%; 95% CI 49.10% to 52.41%), but milder metabolic disturbance: hypertension (32.2%; 95% CI 30.61% to 33.77%), DM (5.9%; 95% CI 5.19% to 6.77%) and CKD (3.2%; 95% CI 2.63% to 3.79%), consistent with an early metabolic syndrome state.

Cluster 3: Low comorbidity

This largest cluster, low comorbidity (LC) with 42.4% of participants (n=7 424), shows minimal comorbidity: abdominal obesity (3.4%; 95% CI 2.79% to 4.05%) and BMI-defined obesity (0.3%; 95% CI 0.20% to 0.50%); hypertension (16.1%; 95% CI 14.98% to 17.27%), DM (3.3%; 95% CI 2.72% to 3.99%) and CKD (3.6%; 95% CI 3.04% to 4.34%) are likewise low.

Cluster 4: Non-obese cardiometabolic multimorbidity

Comprising 5.2% of the sample (n=911), the non-obese cardiometabolic multimorbidity (NOCM) cluster has negligible abdominal obesity (5.1%; 95% CI 3.21% to 8.12%) and BMI-defined obesity (0.04%; 95% CI 0.01% to 0.25%), yet high cardiometabolic burden: hypertension (85.5%; 95% CI 82.31% to 88.26%), DM (46.5%; 95% CI 41.96% to 51.09) and CKD (51.8%; 95% CI 47.22% to 56.34%), along with increased cardiovascular events and cancer prevalence.

Values are reported as weighted percentages (95% CIs).

Demographic distribution across clusters

Table 3 shows the differences among all demographic variables in cluster membership. In terms of sex, women were more common in OSMS (68.9%; 95% CI 66.0% to 71.5%) and OEMS (75.7%; 95% CI 74.1% to 77.2%), whereas men predominated in LC (62.7%; 95% CI 61.1% to 64.3) and NOCM (78.5%; 95% CI 74.8% to 81.7%). Regarding age distribution, LC had the youngest participants, with 49.4% (47.7%–51.0%) aged 25–40 years, while OSMS (55.3% (52.3%–58.3%)) and NOCM (58.3% (53.6%– 62.9%)) had older populations (>60 years). With regard to education, illiteracy was most common in OSMS (31.1% (28.7%–33.7%)) and NOCM (27.4% (23.9%–31.2%)), while higher education level (≥12 years) was most frequent in LC (50.9% (49.3%–52.5%)). In terms of occupation, unpaid work predominated in OSMS (61.9% (58.9%–64.8%)) and OEMS (65.6% (64.0%–67.2%)), employment was most common in LC (56.9% (55.3%–58.5%)), and retirement was most frequent in NOCM (38.0% (33.6%–42.7%)).

Table 3. The baseline information of participants in each cluster.

Variables	OSMS	OEMS	LC	NOCM
Sex
Female	68.86 (66.04–to 71.54)	75.67 (74.11– to 77.16)	37.27 (35.70– to 38.88)	21.52 (18.27– to 25.17)
Male	31.14 (28.46 to 33.96)	24.33 (22.84 to 25.89)	62.73 (61.12 to 64.30)	78.48 (74.83 to 81.73)
Age, years
25-40	4.48 (3.41 to 5.86)	33.27 (31.70 to 34.88)	49.35 (47.72 to 50.97)	5.87 (4.08 to 8.36)
41-60	40.18 (37.29 to 43.14)	51.39 (49.72 to 53.05)	37.40 (35.86 to 38.96)	35.80 (31.25 to 40.61)
above 60	55.34 (52.34 to 58.31)	15.34 (14.15 to 16.62)	13.26 (12.14 to 14.46)	58.34 (53.59 to 62.94)
Residency
Rural	21.03 (19.15 to 23.05)	24.30 (23.11 to 25.52)	26.77 (25.59 to 27.98)	23.11 (20.08 to 26.45)
Urban	78.97 (76.95 to 80.85)	75.70 (74.48 to 76.89)	73.23 (72.02 to 74.41)	76.89 (73.55 to 79.92)
Education
Zero	31.13 (28.66 to 33.71)	14.04 (13.04 to 15.10)	8.72 (8.01 to 9.49)	27.37 (23.89 to 31.15)
1–6 years	35.04 (32.12 to 38.09)	28.95 (27.56 to 30.37)	19.69 (18.55 to 20.87)	26.58 (23.16 to 30.30)
7–11 years	12.88 (10.92 to 15.13)	20.60 (19.31 to 21.95)	20.72 (19.50 to 21.99)	14.69 (11.46 to 18.63)
12 years and over	20.95 (18.41 to 23.74)	36.42 (34.74 to 38.13)	50.88 (49.25 to 52.50)	31.36 (26.78 to 36.34)
WealthIndex
Class 1 (poorest)	22.47 (19.79 to 25.40)	17.56 (16.35 to 18.83)	19.47 (18.25 to 20.76)	23.43 (19.82 to 27.48)
Class 2	21.99 (19.51 to 24.69)	20.20 (18.80 to 21.67)	19.20 (17.94 to 20.52)	23.71 (20.20 to 27.62)
Class 3	18.52 (16.51 to 20.71)	20.03 (18.84 to 21.28)	18.24 (17.09 to 19.46)	16.66 (14.03 to 19.67)
Class 4	20.78 (18.43 to 23.35)	20.79 (19.48 to 22.17)	20.05 (18.84 to 21.32)	17.28 (13.87 to 21.31)
Class 5 (wealthiest)	16.24 (14.01 to 18.74)	21.41 (19.85 to 23.07)	23.03 (21.47 to 24.68)	18.92 (14.99 to 23.59)
Marital status
Single	1.71 (0.91 to 3.20)	5.34 (4.65 to 6.12)	13.52 (12.39 to 14.73)	2.91 (1.59 to 5.29)
Married	77.69 (75.13 to 80.06)	84.96 (83.70 to 86.14)	81.54 (80.22 to 82.78)	84.87 (80.66 to 88.30)
Divorced/separate with partner	2.20 (1.33 to 3.61)	2.6 (2.12 to 3.25)	2.38 (1.97 to 2.86)	2.81 (1.05 to 7.31)
Widow	18.40 (16.36 to 20.63)	7.07(6.23 to 8.03)	2.57 (2.14 to 3.08)	9.40 (7.16 to 12.26)
Occupation
Employed	15.83 (13.79 to 18.11)	25.73(24.26 to 27.26)	56.93 (55.32 to 58.53)	30.44 (26.21 to 35.03)
Unpaid work	61.89 (58.92 to 64.77)	65.62 (63.97 to 67.24)	28.59 (27.17 to 30.05)	19.7 5(16.72 to 23.19)
Retired	18.09 (15.87 to 20.55)	6.40 (5.48 to 7.47)	8.61 (7.71 to 9.61)	38.04 (33.58 to 42.71)
Unemployed	4.19 (3.17 to 5.51)	2.24 (1.85 to 2.71)	5.87 (5.21 to 6.60)	11.77 (9.52 to 14.47)
Insurance
No	6.83 (5.33 to 8.72)	9.60 (8.61 to 10.70)	11.67 (10.61 to 12.83)	4.73 (3.42 to 6.51)
Basic	48.77 (45.75 to 51.79)	62.33 (60.70 to 63.93)	65.31 (63.70 to 66.89)	53.40 (48.79 to 57.95)
Complementary	44.40 (41.44 to 47.41)	28.07 (26.61 to 29.58)	23.02 (21.62 to 24.48)	41.87 (37.35 to 46.52)

Values are reported as weighted percentages (95% CIs)

LC, low comorbidity; NOCM, non-obese cardiometabolic multimorbidity; OEMS, obesity with early metabolic syndrome; OSMS, obesity with severe metabolic syndrome.

Predictors of cluster membership: multinomial logistic regression results

Table 4 presents the results of multinomial logistic regression with the LC cluster as the reference. In terms of sex, men had lower odds of belonging to the OSMS (OR=0.35 (0.25 to 0.48)) and OEMS (OR=0.28 (0.23 to 0.34)) clusters compared with LC, whereas they had higher odds of belonging to NOCM (OR=2.68 (1.56 to 4.61)), relative to women. For age, compared with younger adults (25–40 years), those aged 41–60 years and over 60 years had higher odds of belonging to both OSMS (OR=8.64 (6.06 to 12.33); OR=23.03 (15.40 to 34.44)) and NOCM (OR=5.46 (3.46 to 8.61); OR=13.81 (8.48 to 22.47)). For education, compared with illiterate individuals, those with ≥12 years of schooling had lower odds of belonging to OSMS (OR=0.50 (0.36 to 0.68)) and NOCM (OR=0.49 (0.34 to 0.71)). Regarding wealth, compared with the poorest group (Class 1), individuals in higher wealth classes had higher odds of belonging to OEMS (OR=1.20 (1.01 to 1.42) for Class 2; OR=1.39 (1.19 to 1.64) for Class 3; OR=1.34 (1.12 to 1.59) for Class 4; OR=1.30 (1.06 to 1.61) for Class 5). A similar pattern was observed for OSMS, with higher odds in Class 3 (OR=1.31 (1.03 to 1.68)) and Class 4 (OR=1.40 (1.07 to 1.84)). In contrast, no significant associations were observed between wealth and NOCM. Full results for other sociodemographic variables, including occupation, marital status and insurance coverage, are shown in table 4.

Table 4. Association between sociodemographic and comorbidity clusters using a multivariate multinomial logistic model.

Variables	OSMS	OEMS	LC	NOCM
Sex (female)
Male	0.35 (0.25 to 0.48)	0.28 (0.23 to 0.34)	1.0 (reference)	2.68 (1.56 to 4.61)
Age (25–40) years
41–60	8.64 (6.06 to 12.33)	1.83 (1.60 to 2.11)	1.0 (reference)	5.46 (3.46 to 8.61)
above 60	23.03 (15.40 to 34.44)	1.53 (1.24 to 1.90)	1.0 (reference)	13.81 (8.48 to 22.47)
Residency (rural)
Urban	1.47 (1.24 to 1.74)	1.27 (1.14 to 1.42)	1.0 (reference)	1.31 (1.05 to 1.63)
Education (zero)
1–6 years	0.96 (0.76 to 1.20)	1.16 (0.98 to 1.38)	1.0 (reference)	0.60 (0.46 to 0.79)
7–11 years	0.68 (0.51 to 0.92)	0.99 (0.80 to 1.21)	1.0 (reference)	0.55 (0.37 to 0.80)
12 years and over	0.50 (0.36 to 0.68)	0.70 (0.57 to 0.87)	1.0 (reference)	0.49 (0.34 to 0.71)
Wealth Index (Class 1 (poorest))
Class 2	1.14 (0.88 to 1.47)	1.20 (1.01 to 1.42)	1.0 (reference)	1.11 (0.81 to 1.52)
Class 3	1.31 (1.03 to 1.68)	1.39 (1.19 to 1.64)	1.0 (reference)	0.96 (0.70 to 1.31)
Class 4	1.40 (1.07 to 1.84)	1.34 (1.12 to 1.59)	1.0 (reference)	0.94 (0.65 to 1.37)
Class 5 (wealthiest)	1.09 (0.80 to 1.49)	1.30 (1.06 to 1.61)	1.0 (reference)	0.86 (0.56 to 1.32)
Marital status (single)
Married	1.73 (0.86 to 3.49)	1.70 (1.38 to 2.10)	1.0 (reference)	1.14 (0.56 to 2.34)
Divorced/separate with partner	2.27 (1.09 to 4.75)	1.84 (1.38 to 2.45)	1.0 (reference)	1.95 (0.83 to 4.58)
Occupation (employed)
Unpaid work	1.78 (1.29 to 2.46)	1.69 (1.39 to 2.05)	1.0 (reference)	1.90 (1.11 to 3.24)
Retired	1.81 (1.34 to 2.45)	1.18 (0.93 to 1.50)	1.0 (reference)	2.50 (1.78 to 3.51)
Unemployed	1.27 (0.87 to 1.84)	0.90 (0.69 to 1.17)	1.0 (reference)	2.16 (1.51 to 3.08)
Insurance (no)
Basic	0.90 (0.64 to 1.28)	0.90 (0.73 to 1.10)	1.0 (reference)	1.70 (1.11 to 2.59)
Complementary	1.44 (0.99 to 2.09)	1.06 (0.84 to 1.34)	1.0 (reference)	2.06 (1.32 to 3.22)

Values are ORs (95% CIs).

LC, low comorbidity; NOCM, non-obese cardiometabolic multimorbidity; OEMS, obesity with early metabolic syndrome; OSMS, obesity with severe metabolic syndrome.

Association between cluster membership and low quality of life

Table 5 shows the association between multimorbidity classes and EQ-VAS scores. In the fully adjusted model, EQ-VAS scores were significantly lower in the OSMS (β = −8.5; 95% CI −10.1 to −6.9), OEMS (β = −2.1; 95% CI −3.2 to −1.1) and NOCM (β = −6.2; 95% CI −8.2 to −4.2) clusters compared with the LC cluster.

Table 5. Association between cluster membership and health-related quality of life.

Outcome	Model 1	Model 2	Model 3
EQ-VAS (continuous)	β (95 % CI)	β (95 % CI)	β (95 % CI)
OSMS	−13.4 (−14.8 to −12.0)	−8.9 (−10.4 to −7.3)	−8.5 (−10.1 to −6.9)
OEMS	−4.5 (−5.3 to −3.6)	−2.4 (−3.4 to −1.4)	−2.1 (−3.2 to −1.1)
LC	Reference	Reference	Reference
NOCM	−10.1 (−11.9 to −8.2)	−6.9 (−8.9 to −4.9)	−6.2 (−8.2 to −4.2)

Values are β coefficients with 95% CIs from Model 1 (crude), Model 2 (adjusted for sex and age) and Model 3 (further adjusted for residency, wealth index, marital status, education, occupation and insurance)

EQ-VAS, EuroQol Visual Analogue Scale; LC, low comorbidity; NOCM, non-obese cardiometabolic multimorbidity; OEMS, obesity with early metabolic syndrome; OSMS, obesity with severe metabolic syndrome.

Discussion

Four distinct multimorbidity clusters were identified among Iranian adults: OSMS, OEMS, LC and NOCM. The OSMS and NOCM clusters were associated with the poorest HRQOL. Membership in these high-burden clusters was strongly associated with advancing age, lower educational attainment and urban residence. Although multimorbidity was most prevalent in those over 60 years, a considerable number of middle-aged adults fell into the OEMS cluster, highlighting the need for early preventive efforts.^{4 22–24}

The prevalence and composition of multimorbidity patterns vary considerably across studies, reflecting differences in disease definitions, the number and types of conditions included, population age structures, data sources and sampling frameworks.^{6 25–29} By using LCA on a nationally representative sample, this study mitigated some of these methodological challenges; however, comparisons with other populations must consider such heterogeneity.

Obesity, particularly abdominal obesity and BMI-defined obesity, was central to the OSMS and OEMS clusters, reinforcing its role as a principal driver of cardiometabolic multimorbidity.^{6 26 30} Conversely, the clustering of cardiometabolic disorders among non-obese individuals may reflect non-adiposity pathways, including renal-vascular dysfunction, hypertension, dyslipidaemia, smoking, ageing, genetic susceptibility, environmental exposures, physical inactivity, and unequal access to early diagnosis and treatment, rather than obesity-mediated metabolic risk alone.^31–34 Zoghi et al estimated the prevalence of metabolically unhealthy normal-weight phenotypes at around 6%–7% nationally,³⁵ and these individuals often have outcomes as poor as, or worse than, their obese metabolically unhealthy counterparts.^{36 37} This finding highlights the need for Iran’s National Committee for Prevention and Control of NCDs to move beyond BMI-centred strategies by incorporating factors such as family history, waist circumference, early life nutrition and physical activity into risk prediction models.^{38 39} Clinically, integrating targeted screening tools like lipid and glucose monitoring or body fat assessment in normal-weight individuals could enable earlier detection. This approach is especially important in provinces with food insecurity and childhood undernutrition and would support national goals for reducing NCD mortality.^{40 41} A pronounced female predominance was observed within both obesity-driven clusters, with women accounting for over two-thirds of these groups. This aligns with previous data showing consistently higher rates of multimorbidity among women across all age groups.^{6 7 42 43} Several interrelated factors contribute to this disparity. Women tend to engage more frequently in healthcare-seeking behaviours and thus receive diagnoses more often. However, they also face disproportionate exposure to adverse social determinants, such as lower income, limited educational opportunities and caregiving burdens, that collectively increase disease risk and negatively impact quality of life.^{43 44} Additionally, women exhibit higher prevalence of obesity, physical inactivity, poorer self-rated health and greater psychosocial stressors including loneliness, which further compound their vulnerability to multimorbidity.^{23 45 46} This vulnerability is reflected in the persistent gap in quality of life between women and men, supporting earlier findings that women report poorer self-rated health, a consequence of both biological susceptibilities and amplified socioeconomic hardship.^{26 43} These findings also highlight the need for culturally sensitive and feasible interventions targeting diet, physical activity and cardiometabolic risk among women. In Iran, where low physical activity and metabolic risk factors remain major NCD priorities, strategies such as primary-care-based counselling, women-friendly physical activity opportunities, home-based exercise programmes, peer or family support and affordable nutrition education may be more practical.^{47 48} However, caregiving responsibilities, cost, mobility constraints, lack of access to safe and acceptable facilities, and broader sociocultural or regulatory barriers may limit intervention reach and should be considered in future intervention-design research.^{49 50}

Advancing age remained a strong predictor of cluster membership, with individuals over 60 years exhibiting the greatest odds of belonging to the OSMS and NOCM groups.^{6 22} Lower educational attainment was independently associated with these clusters, reflecting the critical role of health literacy and preventive behaviours in chronic disease development.^{10 22 25 42} The association between wealth status and obesity-related multimorbidity should also be interpreted within the socioeconomic context of Iran and other developing settings, where the relationship between socioeconomic position and obesity may differ from patterns commonly observed in high-income countries.⁵¹ Previous studies in Iran and other developing countries have suggested that obesity and its metabolic consequences may be more common among socioeconomically advantaged groups, potentially reflecting greater access to energy-dense foods, more sedentary lifestyles and lower occupational physical activity.^51–53 Additionally, urban residence was linked to higher multimorbidity, likely attributable to lifestyle-related risk factors and differential healthcare access within LMIC contexts.⁵⁴

Consistent with prior research, multimorbidity was significantly associated with reduced HRQOL, particularly among individuals in the OSMS and NOCM clusters.^{10 25 43 55 56} The pathways through which multimorbidity diminishes subjective well-being include high treatment burden, psychological stress and physical limitations.^{30 43} Despite higher healthcare utilisation, many individuals experience unmet needs due to suboptimal management and out-of-pocket costs, with disproportionate effects on those of lower socioeconomic status.²⁶ Integrated care models, characterised by multidisciplinary teams, coordinated services and patient-centred management, can mitigate these challenges by reducing treatment burden, aligning care with patient priorities and addressing psychosocial complexity.^{57 58} Evidence indicates that such approaches improve physical and mental health outcomes, enhance self-management, increase patient satisfaction and provide financial protection.^{57 59 60}

Several strengths support the study’s validity. A large, nationally representative sample was used, with data collected following standardised WHO protocols by trained interviewers. Physical measurements and laboratory tests complemented self-reports, improving disease identification. LCA identified real world multimorbidity patterns, and linking these to quality-of-life outcomes offers important insights for clinicians and policy makers. To our knowledge, this is one of the first nationally representative studies in Iran examining multimorbidity clusters alongside HRQOL.

However, some limitations should be noted. The cross-sectional design prevents causal inference, and residual confounding may remain because not all relevant individual, clinical, behavioural and environmental factors were available in the data set. Important chronic conditions, including mental health and rheumatological diseases, were excluded, though a recent meta-analysis suggests their inclusion may not alter multimorbidity patterns substantially.⁶¹ In addition, several relevant risk factors were not assessed, including sleep quality, short or long sleep duration, and obstructive sleep apnoea, which may act both as a complication of obesity and as a risk factor for cardiometabolic disease.^{62 63} Environmental exposures such as air pollution and light-at-night exposure, which have been associated with cardiometabolic outcomes and may therefore contribute to multimorbidity patterns, were also not included.^{64 65} Reliance on self-reported diagnoses may have introduced recall bias and inflated associations between social determinants and multimorbidity.^{61 66} The EQ-VAS, while widely used, does not capture all quality-of-life domains such as mental and social functioning.

Despite these limitations, the study provides valuable evidence on the complex association between multimorbidity, sociodemographic factors and quality of life in Iran. Future research should use large prospective cohorts with repeated assessments to clarify causal pathways, examine transitions between multimorbidity clusters and assess whether cluster membership predicts incident CVD, CKD, diabetes, disability, hospitalisation, mortality and quality-of-life decline. Future studies should incorporate more objective measures, such as measured biomarkers, medication use, physician-diagnosed conditions, registry or electronic health record linkage, accelerometer-based physical activity and detailed dietary or sleep assessments. Including a broader range of diseases, particularly mental, rheumatological and acute-on-chronic conditions, together with behavioural, social and environmental risk factors, would strengthen latent cluster analyses and improve prospective risk prediction.

Our findings highlight the need for targeted interventions focused on high-risk groups such as older adults, women with low socioeconomic status and urban residents. Strategies must move beyond single-disease approaches to integrated care models that address both multimorbidity’s medical complexity and its social determinants. The association between urban residence and higher multimorbidity in Iran may reflect the impact of urbanisation, including more sedentary lifestyles, greater consumption of processed foods⁶⁷ and increased exposure to air pollution,⁶⁸ alongside greater healthcare access.⁶⁹ The high prevalence of both abdominal obesity and BMI-defined obesity, particularly among women, further emphasises the importance of prevention strategies that promote healthier diets and physical activity. In this context, health system planning should prioritise integrated NCD prevention and management, with attention to heterogeneous multimorbidity profiles and socioeconomic disparities in access, utilisation and outcomes. Such approaches may help reduce the burden of multimorbidity and improve quality of life, particularly among disadvantaged and high-risk populations.

Conclusion

This study identified four distinct patterns of multimorbidity among Iranian adults, each with unique demographic, socioeconomic and clinical profiles. The presence of severe and early stage obesity-driven clusters, alongside a lean but high-risk cardiometabolic group, underscores the complex and heterogeneous nature of multimorbidity. These findings highlight the urgent need for early, targeted and integrated interventions that consider both medical complexity and the broader social determinants of health. Population-level prevention programmes should prioritise obesity reduction and improved screening for cardiometabolic risks among normal-weight individuals. Strengthening integrated care and prevention strategies will not only reduce the burden of multimorbidity but also directly enhance HRQOL and overall well-being.

Data availability statement

Data may be obtained from a third party and are not publicly available.