Dietary protein intake in relation to metabolic syndrome reversion: findings from the Tehran lipid and glucose study

Abstract

Dietary protein quantity and quality may influence metabolic regulation and cardio-metabolic outcomes, yet little is known about their role in metabolic syndrome (MetS) reversion, a beneficial transition linked to reduced disease risk. This study examined associations between total and source-specific dietary protein intake and MetS reversion in Iranian adults. Adults with MetS at baseline (Phase 3, 2006–2008) of the Tehran Lipid and Glucose Study were followed until Phase 6. Dietary intake was assessed using a validated 168-item food frequency questionnaire. Total, animal and plant-based protein, as well as protein food-specific sources (red meat, processed meat, poultry, dairy, legumes, and nuts) was categorized into tertiles. MetS reversion was defined as the transition from MetS to non-MetS during follow-up, categorized by timing: final phase, any phase, early, and sustained reversion. Cox proportional hazards regression models estimated hazard ratios (HRs) and 95% confidence intervals (CIs), adjusted for covariates. In the fully adjusted models, participants in the second tertile of animal protein intake had a significant lower HR for “any phase MetS reversion” (HR: 0.68; 95% CI: 0.48–0.96), compared to the participants in the first tertile. Also, participants in the second tertile of poultry intake had higher HRs for “any phase MetS reversion” (HR: 1.47; 95% CI: 1.03–2.09) and “early phase MetS reversion” (HR: 1.58; 95% CI: 1.01–2.46), compared to the participants in the first tertile. Furthermore, higher intakes of dairy products and nuts were associated with 48% (HR: 0.52; 95% CI: 0.31–0.86) and 43% (HR: 0.57; 95% CI: 0.35–0.92) lower HRs for “early phase MetS reversion”, respectively. No significant associations were observed between dietary protein sources and sustained MetS reversion; however, the small number of sustained reversion events (n = 31, 5.4%) limited statistical power for this outcome, precluding definitive conclusions about long-term metabolic maintenance.

Introduction

Metabolic syndrome (MetS) represents a constellation of interconnected cardio-metabolic risk factors, including abdominal obesity, elevated blood pressure, hyperglycemia, hypertriglyceridemia, and reduced high-density lipoprotein cholesterol (HDL-C)^1,2. The global prevalence of MetS has reached epidemic proportions, affecting approximately 20–25% of the adult population worldwide, with considerable regional variations³. While much research has focused on MetS incidence and prevention, emerging evidence suggests that MetS is not necessarily irreversible. Studies have demonstrated that individuals with MetS can achieve reversion to a metabolically healthy state through lifestyle modifications, particularly dietary interventions^4–6. Understanding the dietary factors that facilitate MetS reversion is therefore of paramount public health importance.

Several longitudinal studies have examined factors associated with MetS reversion, identifying lifestyle modifications as key determinants. A systematic review by Yamaoka and Tango⁷ demonstrated that lifestyle interventions, including dietary modifications and physical activity, can achieve MetS reversion rates ranging from 30% to 50% over 1–3 years. Dietary factors investigated in relation to MetS reversion include overall dietary patterns (such as Mediterranean diet and DASH diet), macronutrient composition, fiber intake, and specific food groups^8,9. For instance, adherence to Mediterranean dietary patterns has been associated with higher rates of MetS reversion in European cohorts¹⁰, while increased whole grain and fiber consumption have shown beneficial effects in Asian populations^11,12.

Dietary protein has garnered increasing attention as a key macronutrient influencing cardio-metabolic health. Protein intake affects multiple metabolic pathways relevant to MetS components, including satiety regulation, glucose homeostasis, lipid metabolism, blood pressure control, and body weight management^13,14. However, the relationship between protein intake and metabolic health is complex and may depend not only on the quantity but also on the quality and source of protein consumed^15,16. Animal-derived proteins (such as red meat, poultry, and dairy) differ substantially from plant-derived proteins (such as legumes and nuts) in their amino acid profiles, associated nutrients, and potential health effects^17,18.

Recent meta-analyses and systematic reviews have yielded inconsistent findings regarding the associations between different protein sources and cardio-metabolic outcomes. While some studies suggest that higher intake of plant-based proteins is associated with reduced MetS risk and improved metabolic profiles^19,20, others indicate that certain animal protein sources, particularly poultry and dairy products, may also confer metabolic benefits^21,22. Conversely, red and processed meat consumption has been consistently linked to increased cardio-metabolic risk^23,24. These discrepancies may reflect differences in study populations, dietary assessment methods, and the complex interplay between protein sources and overall dietary patterns.

Despite the growing body of evidence on dietary protein and MetS incidence, research on MetS reversion (MetS recovery) is extremely scarce, particularly in Middle Eastern populations; where dietary patterns and protein sources differ substantially. To our knowledge, no previous study has investigated this relationship in an Iranian population. Therefore, the present study aimed to prospectively examine the associations between total protein intake and protein from various sources (red meat, processed meat, poultry, dairy, legumes, and nuts) with the likelihood of MetS reversion among participants of the Tehran Lipid and Glucose Study. Fish and eggs were not analyzed separately due to very low consumption levels in this population.

Materials and methods

Study population

This investigation was carried out using data from the Tehran Lipid and Glucose Study (TLGS), a long-term population-based cohort study that monitors risk factors for chronic non-communicable diseases among residents of district 13 in Tehran, Iran²⁵. The TLGS began in 1999 and has continued through several examination cycles performed approximately every three years. For the present analysis, 10,225 adult men and women aged ≥ 19 years who participated in the third examination were considered eligible (Fig. 1). Individuals were excluded if they lacked complete dietary data (n = 7,139), had missing anthropometric, biochemical, or demographic measurements (n = 849), reported implausible energy intakes (< 800 or > 4,200 kcal/d; n = 126), had a history of cardiovascular disease at baseline (n = 54), did not have MetS at baseline (n = 1,475), or were lost to follow-up (n = 6). The final population included 576 adults with MetS. Participants were followed until sixth examination, with a median follow-up duration of 7.6 years.

Fig. 1.

Flow chart of the study.

Assessment of metabolic syndrome and its reversion

MetS was defined using the Joint Interim Statement (JIS) criteria¹, with a modified waist circumference cutoff (≥ 95 cm) appropriate for Iranian adults²⁶. MetS reversion was defined as transitioning from meeting ≥ 3 diagnostic criteria at baseline to meeting fewer than 3 criteria during follow-up. Four distinct reversion patterns were operationalized based on previous longitudinal studies examining MetS dynamics²⁷ and the clinical importance of distinguishing transient from sustained metabolic improvements²⁸. These patterns capture different trajectories of metabolic health and allow assessment of both early response and long-term maintenance of metabolic improvements, which may have distinct determinants and prognostic implications²⁹.

Four reversion patterns were operationalized:

• Final-phase reversion: MetS at baseline and non-MetS at phase 6, irrespective of phases 4–5.

• Any-phase reversion: Becoming non-MetS in at least one subsequent phase (4, 5, or 6).

• Early reversion: Reverting to non-MetS by phase 4.

• Sustained reversion: Maintaining non-MetS status across phases 4, 5, and 6 after baseline MetS.

Dietary assessment

Dietary intakes were assessed at baseline using a validated 168-item semi-quantitative food frequency questionnaire (FFQ)³⁰. Participants reported their usual frequency and portion size for each food over the previous year. Reported intakes were converted into grams per day and nutrient estimates were derived mainly from the USDA food composition database, supplemented with Iranian food composition tables when needed. Daily intakes of protein, animal protein (protein from animal sources), plant-based protein (protein from plant sources), as well as daily intake of each of protein sources including red meat, processed meat, poultry, dairy products, legumes and nuts, were calculated in grams per day, and categorized into tertiles for analysis.

Assessment of covariates

Demographic information was collected by trained interviewers using structured questionnaires²⁵. Anthropometric measurements (weight, height, waist circumference) were taken following standard TLGS measurement protocols using calibrated digital scales and non-stretch tape measures. BMI was calculated as weight (kg)/height (m²). Blood pressure was measured on the right arm using a mercury sphygmomanometer that had been calibrated by the Iranian Institute of Standards and Industrial Research³¹.

Physical activity was assessed using the validated Modifiable Activity Questionnaire (MAQ), expressed as MET-min/week. Activity levels ≤ 600 MET-min/week were categorized as low, and values > 600 MET-min/week as moderate-to-high. The psychometric properties of the Persian MAQ have been published previously³².

Fasting blood samples were collected in the morning after an overnight fast. Plasma glucose and lipid markers were analyzed by enzymatic colorimetric methods at the TLGS laboratory using standardized kits (Pars Azmoon, Iran) and a Selectra 2 auto-analyzer. Quality control procedures ensured that intra- and inter-assay variability remained below 5%.

Statistical analysis

Baseline characteristics of participants were described according to MetS reversion status, using means (± SD) for continuous variables and frequencies (percentages) for categorical variables. Dietary intake of total protein, animal protein, plant-based protein, and protein from specific sources (red meat, processed meat, poultry, dairy products, legumes, and nuts) were expressed in grams per day and categorized into tertiles, with the first tertile serving as the reference category.

Cox proportional hazards regression models were used to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between protein intake (in tertiles) and MetS reversion. Person-years of follow-up served as the underlying time metric. For each of the four reversion outcomes, time-to-event was operationalized as follows:

• Final-phase reversion: Time from baseline (phase 3) to phase 6 for those who achieved non-MetS status at phase 6; participants not meeting this criterion were censored at their last examination.

• Any-phase reversion: Time from baseline to the first occurrence of non-MetS status at any follow-up examination (phase 4, 5, or 6); participants who never achieved reversion were censored at their last examination.

• Early reversion: Time from baseline to phase 4 for those who achieved non-MetS status by phase 4; participants not meeting this criterion were censored at phase 4 or their last examination if lost to follow-up before phase 4.

• Sustained reversion: Time from baseline to phase 4 (the first examination where sustained reversion could be determined) for those who maintained non-MetS status continuously across phases 4, 5, and 6; participants not meeting this criterion were censored at their last examination.

For all models, participants were censored if they were lost to follow-up or did not experience the specific reversion outcome by the end of observation. The assumption of proportional hazards was evaluated using Schoenfeld residuals and was not violated for any exposure-outcome combination.

Potential covariates were selected a priori based on established evidence from previous literature documenting their associations with both dietary protein intake and metabolic health outcomes^20,33,34. These covariates included: age and sex (demographic factors), body mass index (BMI; adiposity measure), smoking status (lifestyle factor), physical activity level (lifestyle factor), and total energy, total fat, total carbohydrate, and total fiber intake (dietary factors). These variables were identified through a systematic review of prior studies examining dietary exposures and MetS-related outcomes and represent established or potential covariates that could bias the exposure-outcome relationship if not adjusted for. Covariates meeting a significance threshold of P < 0.20 in univariate screening models were retained in multivariable-adjusted analyses. Three sequential models with increasing levels of adjustment were constructed:

• Crude model: Unadjusted associations between protein intake tertiles and MetS reversion outcomes.

• Model 1: Adjusted for demographic and lifestyle factors including age (continuous, years), sex (male/female), smoking status (current smoker/non-smoker), and physical activity level (low: ≤600 MET-min/week vs. moderate-to-high: >600 MET-min/week).

• Model 2: Additionally adjusted for body mass index (BMI, continuous, kg/m²) and dietary factors including total energy intake (continuous, kcal/day), total fat intake (continuous, g/day), total carbohydrate intake (continuous, g/day), and total fiber intake (continuous, g/day). This model represents the fully adjusted analysis accounting for all identified potential covariates.

All statistical analyses were conducted using SPSS version 20 (IBM Corp., Armonk, NY, USA). Two-sided P-values < 0.05 were considered statistically significant.

Results

A total of 576 participants with MetS at baseline were included in the analysis. Over a median follow-up of 7.6 years, the total number of participants with final-phase, any-phase, early and sustained reversion of MetS were 89 (15.5%), 209 (36.3%), 129 (22.4%) and 31 (5.4%), respectively.

Baseline characteristics of participants according to the four definitions of MetS reversion are summarized in Table 1. Individuals who continued to meet MetS criteria through phase 6 showed higher baseline BMI compared with those who achieved reversion only in the final phase. Participants categorized as having reversion in any phase were generally younger and exhibited lower WC at baseline, relative to those who maintained MetS across all phases. Likewise, individuals who reverted by phase 4 were younger and had lower BMI, WC, and total protein intake at baseline compared with participants who still had MetS at phase 4. No meaningful differences were observed between individuals with and without sustained MetS reversion.

Table 1.

Baseline characteristics of the participants across metabolic syndrome reversion types.

Variable	Final-phase reversion		Any-phase reversion		Early reversion		Sustained reversion
	Yes(n=89)	No (n=487)	Yes(n=209)	No (n=367)	Yes(n=129)	No(n=447)	Yes(n=31)	No (n=545)
Age (year)	45.47±11.98	47.35±12.33	44.19±12.38	48.70±11.94*	0.84 (0.56–1.27)	47.96±12.02*	43.58±12.64	47.26±12.25
Male (%)	61.8	52.8	57.4	52.3	0.83 (0.55–1.25)	53.0	45.2	54.7
Smoking (%)	9.0	10.1	11.0	9.3	0.86 (0.52–1.43)	10.3	3.2	10.3
Low PA † (%)	25.8	23.4	30.1	20.2	0.75 (0.50–1.13)	21.9	25.8	23.7
BMI (m2/kg)	29.36±3.83	30.30±4.23*	29.82±4.23	30.35±4.15	0.73 (0.48–1.11)	30.36±4.13*	29.42±4.70	30.20±4.16
WC (cm)	98.15±8.04	99.90±8.64	98.65±8.18	100.2±8.74*	0.78 (0.50–1.20)	100.0±8.64*	96.71±7.67	99.79±8.59
Total energy intake (kcal/d)	2347±795.8	2244±741.7	2224±731.0	2280±761.7	1.05 (0.69–1.61)	2275±755.4	2298±820.7	2257±747.1
Total carbohydrate intake (g/d)	338.4±122.4	329.4±121.6	323.9±117.3	334.7±124.1	1.08 (0.70–1.64)	332.7±122.0	323.4±122.5	331.2±121.7
Total protein intake (g/d)	78.08±28.87	77.67±29.21	76.17±26.88	79.20±30.28	73.41±25.97	78.99±29.90*	73.11±25.63	78.01±29.32
Total fat intake (g/d)	82.81±36.82	75.19±30.14	76.50±31.66	76.50±31.66	74.99±31.45	76.76±31.35	85.69±43.78	75.84±30.46

Data are mean ± SD.

*P<0.05.

PA, physical activity; BMI, body mass index; WC, waist circumference.

† Modifiable activity questionnaire (MAQ) scores ≤ 600 METs-min/week was considered as low physical activity.

The HRs and 95% CIs of MetS reversion across tertile categories of protein intake are shown in Tables 2 and 3. In the crude models, there were inverse associations between animal protein intake and “any-phase reversion”, and between dairy intake and nuts intake and “early reversion”, and also a positive association between poultry intake and “any-phase reversion”. After adjustment for covariates, participants in the second tertile of animal protein (range: 17.29–26.39 g/day) intake had a significant lower HR for “any phase MetS reversion” (HR: 0.68; 95% CI: 0.48–0.96; P = 0.029), compared to the participants in the first tertile (≤ 17.28 g/day). Also, participants in the second tertile of poultry (range: 32.22–56.76 g/day) intake had higher HRs for “any phase MetS reversion” (HR: 1.47; 95% CI: 1.03–2.09; P = 0.033) and “early phase MetS reversion” (HR: 1.58; 95% CI: 1.01–2.46; P = 0.045), compared to the participants in the first tertile (≤ 32.21 g/day). Furthermore, higher intakes of dairy products (≥ 529.9 g/day) and nuts (≥ 6.86 g/day) were associated with 48% (HR: 0.52; 95% CI: 0.31–0.86; P = 0.011) and 43% (HR: 0.57; 95% CI: 0.35–0.92; P = 0.022) lower HRs for “early phase MetS reversion”, respectively. There were no significant associations between total protein, plant-based protein, red meat, processed meat, and legumes intake and HRs of each of MetS reversion types, in the crude and adjusted models.

Table 2.

HR (95% CI) of metabolic syndrome reversion across tertiles of dietary protein intake in participants of Tehran lipid and glucose Study.

MetS reversion	Final-phase reversion		Any-phase reversion		Early reversion		Sustained reversion
	T2	T3	T2	T3	T2	T3	T2	T3
Total protein
Crude	0.76 (0.45–1.29)	1.17 (0.71–1.94)	0.78 (0.56–1.09)	0.87 (0.62–1.21)	0.84 (0.56–1.27)	0.74 (0.48–1.14)	0.69 (0.29–1.64)	0.80 (0.35–1.86)
Model 1	0.74 (0.44–1.27)	1.09 (0.65–1.80)	0.75 (0.54–1.05)	0.81 (0.58–1.13)	0.83 (0.55–1.25)	0.69 (0.45–1.06)	0.70 (0.29–1.67)	0.78 (0.33–1.82)
Model 2	0.81 (0.43–1.52)	1.38 (0.54–3.57)	0.82 (0.55–1.22)	0.98 (0.52–1.85)	0.86 (0.52–1.43)	0.79 (0.35–1.81)	0.78 (0.27–2.20)	1.10 (0.21–5.64)
Animal protein
Crude	0.79 (0.48–1.31)	0.78 (0.47–1.31)	0.71 (0.51–0.99)	0.78 (0.57–1.09)	0.75 (0.50–1.13)	0.72 (0.48–1.10)	1.15 (0.50–2.66)	0.88 (0.36–2.16)
Model 1	0.78 (0.47–1.29)	0.73 (0.43–1.23)	0.68 (0.49–0.95)	0.71 (0.51–0.99)	0.73 (0.48–1.11)	0.66 (0.43–1.02)	1.15 (0.50–2.69)	0.85 (0.34–2.12)
Model 2	0.73 (0.43–1.24)	0.59 (0.31–1.14)	0.68 (0.48–0.96)	0.71 (0.47–1.08)	0.78 (0.50–1.20)	0.76 (0.45–1.29)	1.25 (0.52–3.01)	0.99 (0.31–3.12)
Plant protein
Crude	1.10 (0.65–1.85)	1.27 (0.75–2.14)	0.95 (0.69–1.33)	0.98 (0.70–1.37)	1.05 (0.69–1.61)	1.07 (0.70–1.64)	1.28 (0.54–3.03)	1.13 (0.46–2.79)
Model 1	1.09 (0.64–1.84)	1.13 (0.67–1.92)	0.98 (0.70–1.36)	0.90 (0.64–1.27)	1.08 (0.70–1.64)	0.97 (0.63–1.49)	1.34 (0.57–3.20)	1.05 (0.42–2.61)
Model 2	1.00 (0.55–1.82)	1.14 (0.50–2.60)	1.10 (0.75–1.61)	1.20 (0.69–2.10)	1.33 (0.81–2.18)	1.54 (0.76–3.13)	1.47 (0.53–4.10)	1.59 (0.37–2.75)

Data are hazard ratio (95% CI); proportional hazard Cox regression was used. CI, confidence interval; MetS, metabolic syndrome.

Model 1 was adjusted for sex, age, BMI, smoking status, physical activity; model 2 was additionally adjusted for total energy (kcal/d), total fat (g/d), total carbohydrate (g/d) and total fiber (g/d) intake.

Definitions of reversion:

Final-phase reversion: Reversion from phase 6 to 3, non-MetS at phase 6 and MetS at phase 3, regardless of status in phases 4 or 5.

Any-phase reversion: Reversion at any phase, transition from MetS at baseline (phase 3) to non-MetS in any subsequent phase (4, 5, or 6).

Early reversion: Reversion from phase 4 to 3 only: non-MetS at phase 4 and MetS at phase 3.

Sustained reversion: non-MetS status maintained across phases 4, 5, and 6 after having MetS at phase 3.

Table 3.

Risk of metabolic syndrome reversion across tertiles of different sources of dietary protein intake in participants of Tehran lipid and glucose Study.

MetS reversion	Final-phase reversion		Any-phase reversion		Early reversion		Sustained reversion
	T2	T3	T2	T3	T2	T3	T2	T3
Red meat
Crude	0.90 (0.52–1.53)	1.22 (0.75-2.00)	1.21 (0.86–1.71)	1.25 (0.89–1.74)	1.26 (0.82–1.94)	1.20 (0.78–1.84)	1.59 (0.65–3.90)	1.36 (0.55–3.38)
Model 1	0.88 (0.52–1.51)	1.10 (0.66–1.83)	1.17 (0.83–1.65)	1.08 (0.77–1.53)	1.19 (0.77–1.83)	1.06 (0.68–1.65)	1.51 (0.62–3.73)	1.39 (0.54–3.54)
Model 2	0.87 (0.50–1.50)	1.07 (0.60–1.89)	1.21 (0.85–1.72)	1.20 (0.82–1.75)	1.27 (0.81–1.98)	1.29 (0.79–2.09)	1.61 (0.65-4.00)	1.66 (0.59–4.63)
Processed meat
Crude	0.92 (0.56–1.53)	0.89 (0.53–1.48)	1.08 (0.77–1.50)	0.96 (0.68–1.34)	1.02 (0.67–1.54)	0.89 (0.58–1.38)	1.17 (0.50–2.70)	0.91 (0.37–2.23)
Model 1	0.76 (0.44–1.28)	0.72 (0.42–1.23)	0.82 (0.58–1.15)	0.71 (0.49–1.01)	0.78 (0.51–1.20)	0.65 (0.41–1.04)	0.86 (0.36–2.08)	0.70 (0.27–1.84)
Model 2	0.69 (0.40–1.18)	0.62 (0.36–1.09)	0.82 (0.58 − 0.15)	0.70 (0.48–1.02)	0.80 (0.52–1.23)	0.68 (0.43–1.10)	0.79 (0.32–1.94)	0.63 (0.24–1.68)
Poultry
Crude	1.18 (0.72–1.96)	0.96 (0.57–1.62)	1.43 (1.02-2.00)	1.18 (0.83–1.66)	1.44 (0.94–2.20)	1.10 (0.71–1.72)	1.47 (0.65–3.31)	0.70 (0.27–1.84)
Model 1	1.26 (0.76–2.09)	0.97 (0.57–1.65)	1.42 (1.01–1.99)	1.17 (0.83–1.67)	1.47 (0.96–2.25)	1.15 (0.73–1.82)	1.54 (0.68–3.49)	0.81 (0.30–2.15)
Model 2	1.16 (0.68–1.96)	0.94 (0.51–1.74)	1.47 (1.03–2.09)	1.34 (0.89-2.00)	1.58 (1.01–2.46)	1.50 (0.89–2.54)	1.48 (0.64–3.45)	0.88 (0.29–2.66)
Dairy
Crude	1.19 (0.72–1.98)	0.89 (0.52–1.50)	1.04 (0.75–1.43)	0.72 (0.51–1.02)	0.92 (0.62–1.36)	0.52 (0.33–0.82)	1.10 (0.47–2.59)	0.93 (0.39–2.24)
Model 1	1.11 (0.67–1.85)	0.87 (0.51–1.48)	1.00 (0.72–1.39)	0.74 (0.53–1.05)	0.86 (0.58–1.28)	0.52 (0.33–0.82)	1.01 (0.42–2.40)	0.89 (0.37–2.15)
Model 2	1.08 (0.64–1.81)	0.75 (0.42–1.35)	1.00 (0.72–1.38)	0.72 (0.49–1.06)	0.87 (0.58–1.30)	0.52 (0.31–0.86)	1.06 (0.44–2.56)	0.92 (0.33–2.51)
Legumes
Crude	0.90 (0.54–1.50)	0.82 (0.49–1.37)	0.90 (0.65–1.25)	0.76 (0.55–1.07)	0.80 (0.53–1.20)	0.70 (0.46–1.07)	0.82 (0.33–2.02)	1.03 (0.44–2.39)
Model 1	0.85 (0.51–1.43)	0.77 (0.46–1.29)	0.87 (0.62–1.21)	0.73 (0.52–1.02)	0.75 (0.50–1.14)	0.67 (0.44–1.02)	0.73 (0.29–1.80)	0.91 (0.39–2.11)
Model 2	0.84 (0.50–1.42)	0.76 (0.44–1.33)	0.87 (0.62–1.21)	0.78 (0.54–1.12)	0.77 (0.51–1.17)	0.78 (0.49–1.23)	0.77 (0.31–1.94)	1.09 (0.44–2.68)
Nuts
Crude	1.13 (0.68–1.88)	0.97 (0.57–1.62)	1.00 (0.73–1.39)	0.75 (0.53–1.05)	1.07 (0.72–1.58)	0.60 (0.38–0.95)	1.79 (0.75–4.26)	1.05 (0.40–2.72)
Model 1	1.06 (0.64–1.77)	0.94 (0.55–1.58)	0.96 (0.70–1.33)	0.75 (0.53–1.05)	1.02 (0.68–1.52)	0.60 (0.38–0.94)	1.71 (0.71–4.11)	1.09 (0.42–2.85)
Model 2	1.01 (0.60–1.68)	0.75 (0.43–1.30)	0.98 (0.71–1.36)	0.73 (0.51–1.05)	1.04 (0.69–1.57)	0.57 (0.35–0.92)	1.59 (0.66–3.85)	0.83 (0.30–2.30)

Data are hazard ratio (95% CI); proportional hazard Cox regression was used. CI, confidence interval; MetS, metabolic syndrome.

Model 1 was adjusted for sex, age, BMI, smoking status, physical activity; model 2 was additionally adjusted for total energy (kcal/d), total fat (g/d), total carbohydrate (g/d) and total fiber (g/d) intake.

Definitions of reversion:

Final-phase reversion: Reversion from phase 6 to 3, non-MetS at phase 6 and MetS at phase 3, regardless of status in phases 4 or 5.

Any-phase reversion: Reversion at any phase, transition from MetS at baseline (phase 3) to non-MetS in any subsequent phase (4, 5, or 6).

Early reversion: Reversion from phase 4 to 3 only: non-MetS at phase 4 and MetS at phase 3.

Sustained reversion: non-MetS status maintained across phases 4, 5, and 6 after having MetS at phase 3.

There were no significant associations between any protein sources and sustained MetS reversion. However, given the small number of sustained reversion events (n = 31, 5.4% of the cohort), statistical power was limited for detecting associations with this outcome.

Discussion

The present study examined the association between dietary protein intake from different sources and MetS reversion over a median follow-up of 7.6 years in Iranian adults participating in the Tehran Lipid and Glucose Study. Our findings revealed complex and source-specific associations: moderate animal protein intake was associated with lower likelihood of any-phase MetS reversion (HR: 0.68), while moderate poultry consumption was linked to higher rates of both any-phase (HR: 1.47) and early reversion (HR: 1.58). Conversely, higher intakes of dairy products and nuts were associated with reduced early-phase reversion (HR: 0.52 and 0.57, respectively). These findings highlight the importance of considering protein quality and source rather than total intake when designing dietary interventions for metabolic health.

Interpretation of main findings

The inverse association between moderate animal protein intake and any-phase MetS reversion warrants careful interpretation. Animal proteins, particularly from red meat sources, contain saturated fatty acids and heme iron, both implicated in insulin resistance and systemic inflammation³⁵. Additionally, animal proteins have higher proportions of branched-chain amino acids (BCAAs) and aromatic amino acids, which have been linked to impaired insulin signaling and increased diabetes risk³⁶. However, our findings contrast with a recent Iranian cohort study by Hajihashemi et al. (2021) that reported inverse associations between higher animal protein intake and incident MetS³⁵. This discrepancy may reflect fundamental differences between MetS incidence and reversion as distinct outcomes. While protein intake may influence the development of new MetS through effects on satiety and body weight, the reversal of established metabolic dysfunction may be more sensitive to the pro-inflammatory potential of certain animal protein sources²¹. The non-linear relationship observed—with only the second tertile showing significant associations—suggests a potential threshold effect that warrants further investigation.

The positive association between moderate poultry intake and MetS reversion is biologically plausible and aligns with evidence supporting metabolic benefits of lean white meat. Poultry, particularly when consumed without skin and prepared using healthier cooking methods, provides high-quality protein with substantially lower saturated fat content compared to red meat²⁴. This favorable macronutrient profile may contribute to improved insulin sensitivity, better glycemic control, and favorable lipid profiles—all key components of MetS reversal³⁷. A recent meta-analysis by Hidayat et al. (2021) found that poultry consumption was associated with neutral to modestly inverse associations with MetS risk, particularly when compared to red and processed meat³⁸. The substitution of red meat with poultry in intervention studies has been shown to improve markers of insulin sensitivity and reduce inflammatory biomarkers³⁹. However, the dose-response pattern observed in our study—with benefits apparent in the second but not the third tertile—suggests that moderate rather than high poultry consumption may be optimal for metabolic recovery⁴⁰.

The unexpected inverse associations between higher intakes of dairy products and nuts with early-phase MetS reversion appear to contradict substantial evidence supporting their metabolic benefits. The PURE study, which included 147,812 individuals from 21 countries, found that higher dairy consumption was associated with lower MetS prevalence and reduced risk of hypertension and diabetes^41,42. Similarly, nuts have been consistently linked to improved metabolic health through their content of unsaturated fatty acids, fiber, magnesium, and polyphenols⁴³. Several factors may explain these discordant findings. First, the heterogeneity of dairy products is substantial—fermented dairy like yogurt has been more consistently associated with metabolic benefits compared to cheese⁴⁴. If dairy intake in our population was predominantly from high-fat cheese or consumed in less healthy dietary patterns, this could explain the unexpected associations. Second, the distinction between incident MetS and MetS reversion is crucial. While these foods may help prevent MetS development through effects on weight management in healthy individuals, their role in reversing established metabolic dysfunction may differ⁴⁵. Third, residual confounding and reverse causation are important considerations. Individuals with more severe metabolic dysfunction may have increased their dairy or nut intake as part of health improvement efforts, creating an apparent inverse association that reflects confounding by indication rather than a true adverse effect⁴⁶. Additionally, nuts are energy-dense foods, and excessive consumption without appropriate dietary compensation could lead to positive energy balance, potentially impeding metabolic recovery⁴⁷.

While we adjusted for BMI—a key indicator of overall metabolic burden—in our multivariable models, and baseline characteristics (Table 1) showed no meaningful differences across reversion groups, residual confounding by unmeasured aspects of baseline metabolic severity cannot be entirely excluded. Individuals with more severe metabolic dysfunction may have selectively increased their intake of foods perceived as healthy (dairy, nuts) as part of self-directed health improvement efforts, creating confounding by indication that could partially explain the unexpected inverse associations observed. Additionally, if higher dairy or nut consumers had subclinical differences in metabolic parameters not fully captured by our covariate adjustments, this could contribute to residual confounding. These limitations underscore the observational nature of our study and highlight the need for randomized controlled trials to establish causal relationships between specific protein sources and MetS reversion.

Biological mechanisms

Several interconnected pathways likely mediate the associations observed in our study. The amino acid composition of different protein sources varies substantially and may have distinct metabolic effects. Animal proteins contain higher proportions of BCAAs and aromatic amino acids linked to insulin resistance, while plant proteins are rich in amino acids such as arginine and glycine that have been associated with improved insulin sensitivity^48,49. The fat content and fatty acid composition of protein-rich foods are critical determinants of their metabolic effects. Red meat and high-fat dairy contain substantial saturated fatty acids implicated in insulin resistance and dyslipidemia, while poultry contains lower amounts of saturated fat⁵⁰. Interestingly, recent evidence suggests that the metabolic effects of saturated fat may vary depending on the food source, with dairy-derived saturated fat showing different associations compared to meat-derived saturated fat—a “food matrix effect”⁵¹.

Chronic low-grade inflammation is a central feature of MetS and a key target for dietary interventions. Red and processed meats have been associated with elevated inflammatory markers, potentially due to heme iron, advanced glycation end products formed during cooking, and pro-inflammatory fatty acids⁵². Conversely, plant protein sources and nuts contain anti-inflammatory compounds that may reduce systemic inflammation⁵³. A recent Iranian study by Pourmontaseri et al. (2024) found that the dietary inflammatory index was associated with increased MetS risk, suggesting that inflammatory potential of the diet is a key pathway linking dietary patterns to metabolic outcomes⁵⁴. Emerging evidence also highlights the role of gut microbiota, with animal-based diets potentially promoting production of trimethylamine N-oxide (TMAO), a metabolite linked to insulin resistance, while plant proteins and fermented dairy may promote beneficial gut bacteria producing anti-inflammatory short-chain fatty acids^55,56.

Strengths and implications

Our study has several important strengths, including the use of data from the well-established TLGS cohort with rigorous standardized protocols, prospective design with long follow-up allowing examination of different reversion patterns, and assessment of multiple protein sources separately⁵⁷. The findings have several implications for clinical practice and public health nutrition. Our results support the importance of considering protein source and quality, not just quantity, when providing dietary guidance to individuals with MetS⁵⁸. The positive association between moderate poultry consumption and MetS reversion suggests that replacing red meat with lean poultry may be a practical dietary strategy, consistent with major dietary guidelines⁵⁹. However, given the unexpected inverse associations for dairy and nuts and the substantial evidence from other cohorts supporting their metabolic benefits, it would be premature to recommend limiting these foods based solely on our findings⁶⁰. Dietary recommendations should continue to emphasize overall healthy dietary patterns—such as Mediterranean-style diets—that have been shown to improve metabolic health and include moderate amounts of nuts and fermented dairy products⁶¹. From a public health perspective, our findings contribute to evidence that dietary interventions for metabolic health should focus on food-based recommendations considering the entire food matrix rather than isolated nutrients⁶².

Limitations and future research directions

Several limitations must be acknowledged. First, dietary intake was assessed using a food frequency questionnaire at baseline only, which has several important implications. FFQs are subject to measurement error and recall bias⁶³. More importantly, the use of a single baseline dietary assessment does not capture changes in dietary intake that may have occurred during follow-up. This is a particularly salient concern in our study, as participants were diagnosed with MetS at baseline and may have subsequently modified their diets in response to their diagnosis or as part of health improvement efforts. If participants who successfully achieved MetS reversion did so partly through dietary changes initiated after baseline (e.g., reducing red meat, increasing vegetables), our baseline dietary assessment would not capture these changes, leading to non-differential misclassification of exposure. Such misclassification would generally bias associations toward the null, potentially obscuring true associations between dietary changes and MetS reversion⁶⁴. Conversely, if individuals with more severe metabolic dysfunction selectively increased intake of perceived “healthy” foods (dairy, nuts) after baseline as part of self-directed health efforts, this could create spurious inverse associations through confounding by indication. Future studies should incorporate repeated dietary assessments throughout follow-up to better capture the dynamic relationship between dietary changes and metabolic recovery. Second, despite adjustment for multiple covariates, residual confounding and reverse causation remain concerns⁶⁴. Third, our findings are based on an Iranian population with specific dietary patterns that may limit generalizability³⁰. Finally, the relatively small number of participants experiencing sustained MetS reversion (n = 31, 5.4%) substantially limited statistical power for this outcome. This small event rate precluded meaningful analysis of long-term metabolic maintenance and represents an important limitation in interpreting null findings for this outcome. The study should therefore not be interpreted as providing evidence against associations between protein intake and sustained reversion, but rather as underpowered to detect such associations if they exist.

Several important research priorities emerge from our findings. First, randomized controlled trials examining the effects of specific protein source substitutions on MetS components and reversion are urgently needed to provide stronger causal evidence⁶⁵. Second, future studies should incorporate objective biomarkers of protein intake and protein source to reduce measurement error inherent in self-reported dietary assessment⁶⁶. Third, dose-response relationships and potential threshold effects for different protein sources should be more carefully characterized, as our finding of non-linear associations suggests complex relationships⁶⁷. Fourth, mechanistic studies examining the pathways through which protein sources influence metabolic health—including effects on insulin sensitivity, inflammation, gut microbiota, and metabolite profiles—are needed to move beyond associations to biological understanding⁶⁸. Finally, the specific question of what dietary factors promote sustained MetS reversion deserves focused investigation, as this outcome likely represents the most clinically meaningful metabolic improvements⁶⁹.

Conclusion

In this prospective cohort study of Iranian adults with metabolic syndrome, we found that moderate consumption of poultry was associated with higher likelihood of MetS reversion, while moderate animal protein intake and higher intakes of dairy and nuts showed inverse associations with reversion. These findings highlight the complexity of relationships between dietary protein sources and metabolic recovery and underscore the importance of considering protein quality, food matrix, and source-specific effects. The beneficial association observed for poultry supports its potential role as a preferred animal protein source for individuals with MetS, particularly when consumed as part of an overall healthy dietary pattern. However, the unexpected inverse associations for dairy and nuts require replication and mechanistic investigation before they can inform dietary recommendations. Overall, our results emphasize that dietary strategies for metabolic health should be food-based, considering the entire nutrient and bioactive compound profile of protein-rich foods, and should be tailored to population-specific dietary patterns and preferences. Importantly, the limited number of sustained reversion events precluded definitive conclusions about dietary protein’s role in long-term metabolic maintenance, highlighting the need for larger studies with extended follow-up.

Abbreviations

BMI

Body mass index

CI

Confidence interval

DBP

Diastolic blood pressure

FFQ

Food frequency questionnaire

HR

Hazard ratio

MET

Metabolic equivalent

MetS

Metabolic syndrome

SBP

Systolic blood pressure

TLGS

Tehran lipid and glucose study

WC

Waist circumference

Author contributions

Z.G and S.M designed the study. P.M analyzed the data. Z.G and S.M wrote the manuscript. F.A supervised the study and revised the manuscript. All authors read and approved the final manuscript.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

The study protocol was approved by the ethics committee of the Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran. All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki Declaration and its later amendments. Written informed consent was obtained from all individual participants included in the study.