Dietary Protein Reduction During Isocaloric Conditions Reduces Body Weight in Men With Overweight or Obesity Authors: Aslak E Lyster, Amalie S Frederiksen, Jacob K Jensen, Annemarie Lundsgaard, Nicoline R Andersen, Kasper S Jørgensen, Daniela Weber, Lukas Nachtigall, Erik A Richter, Andreas M Fritzen, Bente Kiens Published: 2026 ABSTRACT ABSTRACTObjectiveDietary protein reduction increases plasma fibroblast growth factor 21 (FGF21) and energy requirements in lean men under eucaloric conditions. Whether these metabolic effects translate to weight loss during isocaloric conditions in men with overweight or obesity remains unclear.MethodsSeventeen men with overweight or obesity were randomly allocated to two different highly controlled, isocaloric diets for 5 weeks, receiving either a protein‐reduced diet (0.9 g kg−1 BW day−1) substituted with increased carbohydrates or a habitual diet (1.8 g kg−1 BW day−1).ResultsProtein reduction induced a 2.0 ± 0.6 kg weight loss and increased plasma FGF21 concentration compared with higher protein intake, without caloric restriction. Changes in plasma FGF21 concentration were inversely associated with changes in body weight (r = −0.59, p = 0.02). Expression of key components in the FGF21 receptor complex, FGFR1 and β‐klotho, and downstream targets in subcutaneous adipose tissue remained unchanged. Markers of skeletal muscle mitochondrial oxidative phosphorylation were also unaltered. Fat mass decreased significantly, whereas this level of protein reduction only tended to lower lean body mass.ConclusionsThese findings suggest that reducing protein intake (0.9 g kg−1 BW day−1), substituted with increased carbohydrates, compared with a habitual protein intake (1.8 g kg−1 BW day−1), can induce weight loss without restricting total energy intake in men with overweight or obesity.Trial Registration ClinicalTrials.gov identifier: NCT06263725 Keywords: adipose tissue, energy metabolism, FGF21, protein reduction, weight loss Introduction Obesity is increasing worldwide, and the number of adults living with obesity has more than doubled since 1990 [1]. This growing epidemic represents a major health concern and poses substantial economic challenges in society due to its association with multiple chronic diseases [1]. Dietary intervention is a cornerstone in managing obesity, with caloric restriction creating an energy deficit being fundamental for weight loss. Intriguingly, emerging evidence suggests that the macronutrient composition of the diet per se also influences body weight regulation and metabolic health beyond the effects of caloric intake [2]. Among various dietary strategies, high‐protein (HP) diets have gained attention in both mainstream media and scientific research for their potential role in weight loss and particularly protection of lean body mass during weight loss [3, 4]. However, several prospective cohort studies have demonstrated that high total protein intake, in some cases coupled with low carbohydrate intake, is associated with an increased risk of cardiovascular diseases and all‐cause mortality [5, 6, 7, 8]. These findings raise concerns regarding the long‐term health risks associated with high protein intake. In rodent models, the effects of reducing dietary protein intake have been rigorously studied, and beneficial metabolic and health effects of consuming diets with low protein content are well established [9, 10, 11, 12]. In rodents, a protein‐reduced diet promotes thermogenesis by energy‐wasting processes, such as browning of white adipose tissue (WAT) [13] and activation of brown adipose tissue (BAT), which elevate energy expenditure (EE) and promote weight loss [14, 15]. Furthermore, reducing protein intake potently increases insulin sensitivity in rodents [11, 16]. These adaptive responses to protein reduction in rodents are largely mediated by the liver‐derived hormone fibroblast growth factor 21 (FGF21) [9, 11, 17], a potent regulator of metabolism [18], which is released in response to reduced protein intake in both rodents and humans [11, 19, 20]. FGF21 exerts its effects by binding to the cell surface receptor complex of a FGF receptor (FGFR), mainly FGFR1c, and the requisite co‐receptor, β‐klotho, acting and initiating intracellular signaling mainly in the central nervous system and adipose tissue [17, 21, 22]. In humans, the effects of reducing dietary protein intake have been less studied. In our recent study, lean men, consuming a eucaloric diet with reduced protein content meeting minimum protein requirements (0.9 g protein kg−1 BW day−1) for 5 weeks, increased energy requirements by 21% to maintain body weight and elevated plasma FGF21 levels [20]. Importantly, this effect of reduced protein intake occurred independently of whether the reduction of protein was substituted by carbohydrates or fat [20]. Subsequently switching back to a eucaloric HP diet for 5 weeks decreased energy requirements and plasma FGF21 concentrations [20]. Notably, the increase in energy requirements correlated positively with the increase in plasma FGF21 levels [20]. Based on these findings, an important question arises: can the effects of reduced protein intake, still within the minimum requirements of 0.8 g kg−1 BW day−1 set by the World Health Organization (WHO), be translated to men with overweight or obesity and induce weight loss during isocaloric conditions? Furthermore, it remains unclear whether protein reduction may adversely affect lean body mass during weight loss. In healthy individuals, a protein‐reduced diet (5E%) under energy‐surplus conditions attenuated weight gain compared with higher protein intake (25E%); however, this effect was driven by a lack of lean body mass gain [23]. Previous human studies implementing an isocaloric, protein‐reduced diet substituted with carbohydrates have reported divergent findings regarding changes in lean body mass [19, 24]. In the present study, we investigated whether reducing protein intake to 0.9 g kg−1 BW day−1, substituted with increased carbohydrate intake, compared to higher protein intake (1.8 g kg−1 BW day−1), can have an impact on body weight regulation, body composition, metabolism, and glucose homeostasis in men with overweight or obesity, while still consuming the same amount of total energy. Additionally, we examined plasma FGF21 levels and potential adaptations in FGF21 signaling in adipose tissue and markers of mitochondrial oxidative phosphorylation in skeletal muscle to elucidate the effects of dietary protein reduction. Methods Research Design Seventeen healthy males, age 25–45 years, with BMI between 26 and 35 kg/m2, a habitual dietary protein intake of ~20E%, and an inactive lifestyle (≤ 1 training per week) and who were nonsmokers were recruited for this study (characteristics in Table 1). All participants received written and oral information about the study and provided written consent prior to participating in the study, executed at the University of Copenhagen, Department for Nutrition, Exercise and Sports. The study was approved by the Ethical Committee for the Capital Region of Denmark (H‐21015808) and conducted in accordance with the Helsinki Declaration. TABLE 1Participant characteristics.Total (n = 17)HP (n = 7)RP (n = 10)HP vs. RP (p)Age (years)33 ± 1.235 ± 2.032 ± 1.40.228Height (m)1.86 ± 0.021.84 ± 0.021.88 ± 0.020.263Body weight (kg)102.8 ± 2.697.7 ± 2.7106.3 ± 3.80.109BMI29.7 ± 0.529.0 ± 0.630.2 ± 0.70.264Body fat percentage (%)33.5 ± 0.933.4 ± 1.533.6 ± 1.30.932Body fat mass (kg)33.4 ± 1.431.6 ± 1.534.7 ± 2.10.291Android fat (%)42.1 ± 1.442.0 ± 1.142.2 ± 2.30.923Visceral fat (g)1504 ± 1311357 ± 1571608 ± 1940.364LBM (kg)65.9 ± 1.662.9 ± 2.268.0 ± 2.00.119VO2peak (mL min−1 kg−1)32.6 ± 1.133.5 ± 1.832.0 ± 1.40.532 Note: Values are means ± SEM. A two‐tailed unpaired t‐test was applied to test for differences between groups (HP vs. RP).Abbreviation: VO2peak, peak oxygen uptake related to body mass. Participant characteristics. Note: Values are means ± SEM. A two‐tailed unpaired t‐test was applied to test for differences between groups (HP vs. RP). Abbreviation: VO2peak, peak oxygen uptake related to body mass. Study Design The study was a randomized, parallel dietary intervention comparing a HP diet with a reduced‐protein (RP) diet, in which protein was substituted with increased carbohydrate intake. The study was initiated by a 2‐week run‐in period followed by a 5‐week highly controlled, isocaloric dietary intervention, where all food was provided to the participants (Figure 1). During run‐in, participants' fitness level (peak oxygen uptake, VO2peak) and habitual diet were assessed, and participants were familiarized with the indirect calorimetry procedure. Participants' daily physical activity level was monitored, and body weight was measured each morning in an overnight‐fasted state during the run‐in and intervention period. The intervention included weekly test‐days (WK0‐WK5) (Figure 1). On the day before each test‐day, participants collected a feces sample, which was immediately frozen and stored at −20°C until further analysis. Outline of study design. Participants completed a 2‐week run‐in period before being randomly allocated to a 5‐week dietary intervention with either a reduced‐protein diet (RP, n = 10) or a high‐protein diet (HP, n = 7). Tests were conducted on a weekly basis. Participants arrived in the morning in a fasted state (~10 h). Fecal samples were collected the day prior to the test day. Biopsies were collected at WK0 (Pre) and WK5 (Post). Dual‐energy X‐ray (DXA) absorptiometry was collected at WK0, WK3, and WK5. Blood samples and resting metabolic rate (RMR) were collected weekly. Physical activity registration with accelerometer (SENS motion). SAT: subcutaneous adipose tissue. Created in https://BioRender.com. Weekly Tests Participants arrived in the morning after an overnight fast (~10 h) by passive transport. At WK0, WK3, and WK5, body composition was determined using dual‐energy X‐ray absorptiometry (DXA) (Lunar DPX‐IQ, GE HealthCare), and body water content was measured by bioimpedance (InBody270, InBody UK Ltd.). Resting metabolic rate (RMR) was measured by indirect calorimetry using a ventilated hood (Jaeger Vyntus CPX, Intramedic) after at least 30 min of rest in the supine position under thermoneutral conditions. Blood samples were obtained from a catheter inserted into an antecubital vein. At WK0 and WK5 a biopsy was obtained from periumbilical subcutaneous adipose tissue (SAT) and from the vastus lateralis muscle using a modified Bergström needle with suction [25] under local anesthesia (~2 mL of 1% xylocaine; AstraZeneca A/S). The samples were immediately flushed with ice‐cold, sterile saline and snap‐frozen in liquid nitrogen and stored at −80°C until analysis. Habitual and Experimental Diet Macronutrient composition and energy content of the participants' habitual diet were determined based on a 4‐day diet registration, where all food and beverages were weighed to 1 g of accuracy by the participants. The daily energy requirement for each participant was determined by multiplying the calculated RMR [26] with participants' physical activity level estimated during run‐in. The experimental diets were isocaloric and energy intake was fixed at the calculated energy requirements of the individual participant. The macronutrient composition (E%) of the HP diet was designed to reflect the participant's habitual diet to act as a control (see habitual diet in Table S1 and experimental diet in Table 2). For the RP diet, protein intake was reduced from 20E% (habitual intake) to 11E%. Despite this reduction, protein intake was kept within the minimum requirement defined by the WHO (> 0.8 g kg−1 BW day−1) to avoid potential protein deficiency. The reduction in protein content was achieved by lowering mainly meat and dairy consumption, which was replaced by plant‐based alternatives. The lower protein content was substituted with higher carbohydrate intake, increasing carbohydrate intake from 42E% (habitual diet) to 52E% during the intervention, while fat intake was matched to both the HP diet (37E%) and habitual intake to minimize changes in overall dietary composition (Table S1). To ensure similar fiber intake between groups, despite the higher carbohydrate intake in the RP diet, the proportion of simple carbohydrates was increased. All nutrients were provided through the diet, without the use of dietary supplements. TABLE 2Experimental diets.HP diet (n = 7)RP diet (n = 10)HP vs. RP (p)Daily energy intake (MJ day−1)14.4 ± 0.314.9 ± 0.50.487Protein (E%)21 ± 0.011 ± 0.0—Carbohydrate (E%)43 ± 0.052 ± 0.0—Added sugar (E%)4 ± 0.01 ± 0.0—Fat (E%)37 ± 0.037 ± 0.0—Saturated fatty acids (E%)13 ± 0.015 ± 0.0—Monounsaturated fatty acids (E%)10 ± 0.010 ± 0.0—Polyunsaturated fatty acids (E%)4 ± 0.05 ± 0.0—Protein (g day−1)175 ± 492 ± 3***< 0.001Carbohydrate (g day−1)341 ± 7437 ± 14***< 0.001Added sugar (g day−1)9 ± 031 ± 1***< 0.001Dietary fiber (g day−1)44 ± 144 ± 10.875Fat (g day−1)144 ± 3149 ± 50.486Protein (g kg BW−1 day−1)1.80 ± 0.030.87 ± 0.01***< 0.001 Note: Values are means ± SEM. Two‐tailed unpaired t‐test was applied to test for differences between groups (HP vs. RP). Different from HP diet, ***p < 0.001. Statistics were not done on macronutrient E%, due to the nature of the intervention design. Experimental diets. Note: Values are means ± SEM. Two‐tailed unpaired t‐test was applied to test for differences between groups (HP vs. RP). Different from HP diet, ***p < 0.001. Statistics were not done on macronutrient E%, due to the nature of the intervention design. All food items were weighed to 1 g of accuracy and provided to the participants in a 7‐day rotation. The participants did not ingest any other food items or drinks throughout the intervention. Habitual coffee and tea intake was allowed. Participants filled out a daily food checklist to ensure compliance. The habitual and experimental diets were analyzed using the software program VITAKOST (VITAKOST ApS). Physical Activity and Peak Oxygen Uptake Peak oxygen uptake (VO2peak) was measured using a metabolic online system (Jaeger Vyntus CPX, Intramedic) during an incremental test to exhaustion on a bicycle ergometer (Monark Ergomedic 839 E, Monark Exercise AB) [27]. To monitor physical activity, participants wore a triaxial accelerometer (SENS motion activity sensor) attached to the skin on the lateral side of the thigh. SENS motion quantifies accelerometer data as time spent in the following categories: lying and sitting, upright standing, sporadic walking, walking, cycling, and running. Each category was assigned a Metabolic Equivalent (MET) value to determine the energy cost of physical activity, which was used to calculate the physical activity level of the participants [28]. The participants were instructed to maintain their habitual physical activity level throughout the study. Plasma Analysis Plasma glucose concentrations were measured using ABL800 FLEX (Radiometer Medical) from heparin‐coated syringes. Plasma concentrations of urea (cat. no. A11A01641) were measured with an enzymatic colorimetric method on a Pentra C400 analyzer (Horiba ABX). Plasma concentrations of insulin (Insulin kit, cat. no. 10‐1113‐01, Mercodia), total amino acids (L‐amino acid assay kit, cat. no. ab65347, Abcam), and FGF21 (FGF21 kit, cat. no. DF2100, R&D Systems) were measured by enzyme‐linked immunosorbent assays (ELISA). Individual amino acids were analyzed using UPLC‐MS/MS as previously described by Henning and colleagues [29], with minor modifications described in detail in online Supporting Information, “Method Expanded.” MRM transitions are summarized in Table S2. All analyses were performed in accordance with the manufacturers' instructions. Preparation and Western Blotting Analysis SAT and skeletal muscle were homogenized using a Qiagen Tissuelyser II as previously described [30]. Protein concentrations were determined in triplicate using the BCA method (Thermo Fisher Scientific), separated by SDS‐PAGE using either self‐cast 6%–8% gels or 12% precast gels (Bio‐Rad), and transferred to PVDF membranes (Bio‐Rad) using semidry blotting. Membranes were blocked before incubation with primary antibodies (Table S3), followed by HRP‐conjugated secondary antibodies. Membranes were visualized by chemiluminescence (Millipore) using a Bio‐Rad ChemiDoc system (Hercules). SAT protein expression was normalized to Pierce total protein quantification (Thermo Fisher) and also presented relative to GAPDH protein content. Muscle protein expression was normalized to a pooled internal standard loaded on each gel and also presented relative to GAPDH protein content and total protein content assessed by Coomassie staining. RNA Isolation and Gene Expression SAT was homogenized in QiaZol reagent with a Qiagen Tissuelyser II. Total RNA was extracted using the RNeasy Mini Kit (Qiagen) according to the manufacturer's protocol. RNA quality and quantity were assessed using Nanodrop 2000 (Thermo Fisher), and cDNA was synthesized using qScript cDNA SuperMix kit (Quantabio). Gene expression was determined using SYBR (Sybr Green PCR Master Mix, Thermo Fisher). The primers (TAG Copenhagen) are described in Table S4. RT‐qPCR was performed on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems). Each sample was determined in triplicate, and the expression of target genes was normalized to the expression of the, in human adipose tissue, validated housekeeping gene, RPLP0 [31, 32]. Fecal Energy Content Energy content of feces samples was measured with a C6000 isoperibol oxygen bomb calorimeter (IKA‐Werke GmbH & Co. KG). Benzoic acid was used to calibrate the bomb calorimeter before each use. Feces samples were dried for 48 h at 50°C and then weighed prior to combustion in the bomb calorimeter. Calculations The Weir equation was used to calculate RMR (MJ day−1) [33]. The energy content of dried feces samples was calculated as follows: C v : Thermal capacity of calorimeter, ∆T: Measured and corrected increase in temperature, Q ext: All external energy, m: Mass of dry sample. Insulin resistance was estimated using the homeostasis model assessment–insulin resistance index (HOMA‐IR) [34]. Statistical Analysis Data are presented as means ±SEM. Statistical analyses were performed in GraphPad Prism (GraphPad Software v10). Two‐way repeated measures (RM) ANOVA with factors for groups, time, and interactions was conducted. In case of missing values, a mixed‐effects model was applied. Significant interactions were followed by a Tukey–Kramer post hoc test to account for sample size differences. Within‐group changes in body composition were evaluated with one‐way RM ANOVA for HP and RP, respectively, followed by Bonferroni's post hoc test for multiple comparisons. Associations between changes in FGF21 and body weight were examined by Spearman's rank correlation. For single end‐point comparisons (ΔWK0–WK5), unpaired Student's t‐test or Mann–Whitney test was used, depending on data distribution. The significance level was p < 0.05. Results Participant Characteristics The study included 17 men with a mean BMI of 29.7 ± 0.5 (Table 1), VO2peak of 32.6 ± 1.1 mL kg−1 min−1 (Table 1), and a habitual protein intake of 20 ± 0.8E% (Table S1). Experimental Diets In HP and RP, the daily energy intake (14.4 ± 0.3 and 14.9 ± 0.5 MJ day−1, respectively), dietary fat intake (144 ± 3 and 149 ± 5 g day−1, respectively), and dietary fiber intake (44 ± 1 and 44 ± 1 g day−1, respectively) were similar between groups (Table 2). Daily protein intake averaged 1.80 ± 0.03 and 0.87 ± 0.01 g kg−1 BW day−1 in HP and RP, respectively (Table 2, p < 0.001). Daily carbohydrate intake was significantly higher in RP compared to HP (437 ± 14 vs. 341 ± 7 g day−1, p < 0.001). Body Weight Regulation In RP, body weight was reduced after 26 days (p = 0.037) and remained significantly reduced throughout the remainder of the intervention compared with baseline (Figure 2a). Body weight was significantly lower in RP compared with HP after 23 days and forward (Figure 2a, p = 0.033). After 5 weeks, the reduction in body weight was 2.0 ± 0.6 kg greater in RP compared with HP (Figure 2a,b, p = 0.011). In HP, body weight remained unchanged throughout the intervention (Figure 2a). Habitual physical activity level was maintained in both groups (Figure 2c). Energy content of the feces samples averaged 21.8 ± 0.7 and 21.9 ± 0.6 kJ g−1 dry weight at baseline in HP and RP, respectively, and did not change during the intervention (Figure 2d). The effects of protein restriction on body weight (BW). (a) Daily change in BW (kg), (b) change in BW after the intervention, (c) calculated physical activity level in run‐in and during the intervention, and (d) energy content of dried feces samples (kJ/g). HP group (∆) and RP group (●). Statistical analysis: in panels a, c, d, a mixed‐effects model was applied with a post hoc Turkey‐Kramer multiple comparison test; in panel b, a two‐tailed unpaired t‐test was applied. In panel a, different from HP group p values indicated above x‐axis, and different from RP baseline p values indicated below x‐axis. Values are means ± SEM. (a–c) HP: n = 7, RP: (n = 10). (d) HP: n = 6, one participant was excluded due to insufficient sample material, RP: n = 10. Plasma FGF21 Levels Fasting plasma FGF21 concentrations averaged 192 ± 35 and 268 ± 58 pg mL−1 (p = 0.33) at baseline in HP and RP, respectively. After the intervention, plasma FGF21 levels were significantly increased in RP compared to HP (Figure 3a, p = 0.025). The changes in plasma FGF21 concentrations were inversely correlated with the changes in body weight following the interventions (Figure 3b, r = −0.59, p = 0.015). Association between FGF21 and body weight regulation. (a) Change in plasma FGF21 percentage weekly and after intervention, (b) correlation between changes in plasma FGF21 and body weight, (c) resting metabolic rate (RMR), and (d) respiratory exchange rate (RER) during intervention. HP group (∆) and RP group (●). In panel a, post intervention: Following visual inspection, a Shapiro–Wilk test was conducted to test normality, which demonstrated that changes in FGF21 in the HP group did not meet the assumption of normality. Hence, a nonparametric Mann–Whitney test, comparing ranks, was applied for statistical analysis. Significance level, p < 0.05. Values are means ±SEM. HP: n = 7, RP: n = 10. In panel b, associations were analyzed using nonparametric Spearman's rank correlation; lines represent linear trends for visualization. n = 17. In panels c, d, weekly measurements: A mixed‐effects model was applied for statistical analysis. Post intervention changes: An unpaired t‐test was applied for statistical analysis. (c) HP: n = 7, RP: n = 10. (d) HP: n = 7, RP: n = 8, two values at baseline were excluded as RER values reached unphysiological values. RMR averaged 7.7 ± 1.3 and 8.3 ± 1.1 MJ day−1 at baseline in HP and RP, respectively, and was unaltered during the interventions (Figure 3c). Fasting respiratory exchange ratio (RER) averaged 0.84 ± 0.02 in HP and 0.83 ± 0.02 in RP at baseline and was not different between groups during the intervention (Figure 3d). Gene and Protein Content in Saturated Adipose Tissue and Skeletal Muscle SAT mRNA expression of the FGF21 receptor, FGFR1c, and co‐receptor, β‐klotho, did not change in HP or RP (Figure 4a,b). Protein levels of FGFR1 and β‐klotho in SAT also remained unaltered in both groups both when normalized to total protein by Pierce staining (Figure 4d,e) and to GAPDH protein content (Figure S2a,b). Extracellular signal‐regulated kinase 1 and 2 (ERK1/2) phosphorylation levels and total levels in SAT remained unchanged in both HP and RP (Figure 4f,g). Accordingly, ERK1/2 phosphorylation relative to total protein was also unchanged (Figure S2c). Protein restriction's impact on gene expression and protein content in adipose and muscle tissue. Adipose tissue: (a) FGFR1 and (b) β‐klotho mRNA expression relative to RPLP0, (c) representative Western blots demonstrating FGFR1, β‐klotho, phosphorylated ERK1/2, total ERK1/2 protein, total protein staining (Pierce), and GAPDH in adipose tissue before and after in RP group and HP group, (d) FGFR1 protein content, (e) β‐klotho protein content, (f) ERK1/2 protein content, and (g) phosphorylated ERK1/2 content normalized to total protein by Pierce staining. Skeletal muscle tissue: (h) representative Western blots illustrating Oxphos complexes I‐V, GAPDH, and total protein staining (Coomassie) in skeletal muscle tissue before and after intervention and (i) muscle Oxphos complexes I‐V protein content normalized to a pooled internal standard loaded on each gel. Measurements were obtained before and after the intervention in HP and RP groups. HP group (bar with stripes) and RP group (full bar). Two‐way RM ANOVA was applied for statistical analysis. Values are means and individual values for Pre and Post. HP: n = 7, RP: n = 10. The protein content of Oxphos complex I‐V in vastus lateralis muscle remained unaltered in both groups when expressed to a pooled internal standard (Figure 4h,i), as well as when normalized to both GAPDH (Figure S2d) and total protein by Coomassie staining (Figure S2e). Metabolic Plasma Markers Baseline concentrations of plasma glucose averaged 5.3 ± 0.1 mmol L−1 in both groups (Figure 5a). Baseline fasting plasma insulin averaged 7.3 ± 0.9 and 8.5 ± 1.1 μU mL−1 in HP and RP, respectively (Figure 5b), and HOMA‐IR averaged 1.7 ± 0.2 in HP and 2.0 ± 0.3 in RP (Figure 5c). No between‐group differences were observed after the intervention (Figure 5a–c). The total plasma amino acid concentrations averaged 1471 ± 58 and 1405 ± 56 μmol L−1 at baseline in HP and RP, respectively, and remained unchanged in both groups during the intervention (Figure 5d). Moreover, plasma concentrations of individual amino acids and 3‐methylhistidine did not change with RP compared to baseline (Figure 5f–j, Figure S1). The HP diet induced a threefold increase in plasma 3‐methylhistidine levels after 1 week, and 3‐methylhistidine was consistently higher during HP compared to RP (Figure 5j). Plasma urea concentrations averaged 5.0 ± 0.5 in HP and 4.4 ± 0.2 mmol L−1 in RP at baseline and were significantly higher in HP compared with RP during all weeks in the intervention (Figure 5e, p < 0.001). In RP, plasma urea concentrations were reduced at Weeks 1 and 2 compared with baseline values (Figure 5e, p = 0.04). Fasting plasma parameters with varying protein content. (a) Fasting plasma glucose, (b) fasting plasma insulin, (c) HOMA‐IR before and after diet intervention, (d) total fasting plasma amino acid, (e) fasting plasma urea, (f) fasting plasma methionine, (g) fasting plasma leucine, (h) fasting plasma tryptophan, (i) fasting plasma threonine, and (j) fasting plasma 3‐methylhistidine concentrations during 5 weeks on a HP diet or RP diet. HP group (bar with stripes and ∆) and RP group (full bar and ●). Two‐way RM ANOVA or mixed‐effects model was applied for statistical analysis with post hoc Tukey‐Kramer multiple comparison test. In panels a‐c, data are presented as means and individual values. In panels d‐j, data are presented as means ±SEM. HP: n = 7, RP: n = 10. Body Composition Fat mass (FM) was reduced by 1.1 ± 0.3 kg in RP (Table 3, p = 0.01), while FM remained unchanged in HP. This reduction in FM occurred from WK3 to WK5. FM percentage (FM%) averaged 33.4% ± 1.5% and 33.6% ± 1.3% in HP and RP, respectively, and remained unchanged by diets (Table 3). In RP, lean body mass (LBM) averaged 68.0 ± 2 kg at baseline and after the intervention it tended to be reduced by 1.3 ± 0.7 kg (Table 3, p = 0.07). The reduction occurred from WK0to WK3, while no changes were observed from WK3 to WK5. In HP, no changes were observed in LBM (Table 3). TABLE 3Body composition.HP dietRP dietWK0 (n = 7)WK3 (n = 7)WK5 (n = 7) p WK0 (n = 10)WK3 (n = 10)WK5 (n = 10) p Fat masskg31.6 ± 1.631.6 ± 1.731.6 ± 1.80.9534.7 ± 2.134.7 ± 2.033.6 ± 2.3*0.01∆ (kg)0.1 ± 0.40.0 ± 0.50.950.0 ± 0.3−1.1 ± 0.3*0.01Fat %%33.4 ± 1.533.5 ± 1.633.3 ± 1.60.9233.6 ± 1.334.0 ± 1.133.2 ± 1.20.09∆ (%)0.0 ± 0.4−0.1 ± 0.50.920.3 ± 0.3−0.4 ± 0.30.09Visceral fatg1357 ± 1571347 ± 1131235 ± 1230.111608 ± 61941631 ± 1971557 ± 1920.36∆ (g)−9 ± 58−122 ± 660.1123 ± 57−50 ± 500.36LBMkg62.9 ± 2.262.9 ± 2.463.0 ± 2.40.9268.0 ± 2.066.9 ± 2.066.7 ± 2.00.07∆ (kg)−0.1 ± 0.40.1 ± 0.60.93−1.1 ± 0.5−1.3 ± 0.70.07LBM %%66.6 ± 1.466.6 ± 1.666.7 ± 1.60.9266.4 ± 1.466.0 ± 1.266.8 ± 1.30.07∆ (%)0.0 ± 0.40.1 ± 0.50.92−0.4 ± 0.30.4 ± 0.30.06TBW %%54.1 ± 1.754.2 ± 1.754.3 ± 1.70.6153.4 ± 0.853.3 ± 0.853.8 ± 0.90.13∆ (%)0.2 ± 0.30.2 ± 0.40.62−0.1 ± 0.20.4 ± 0.20.13 Note: Values are means ± SEM. Change over time in each group: one‐way RM ANOVA was applied for statistical analysis with a post hoc Bonferroni multiple comparisons test. Different from Pre, *p < 0.05.Abbreviations: FM, fat mass; LBM, lean body mass; TBW, total body water. Body composition. Note: Values are means ± SEM. Change over time in each group: one‐way RM ANOVA was applied for statistical analysis with a post hoc Bonferroni multiple comparisons test. Different from Pre, *p < 0.05. Abbreviations: FM, fat mass; LBM, lean body mass; TBW, total body water. Discussion In the present study, men with overweight or obesity consuming a diet reduced in protein, meeting dietary protein minimum requirements, and increased in carbohydrate intake reduced body weight by 2.0 kg over 5 weeks, whereas those assigned to the isocaloric HP diet maintained baseline body weight. Notably, this divergence occurred despite identical individualized energy prescriptions across groups and random allocation to intervention arms. These findings suggest that the weight loss observed in the RP group reflects a physiological response to reduced protein and increased carbohydrate intake rather than inaccuracies in energy estimation or study design. As both diets appeared to meet baseline energy requirements of the participants, any changes in body weight would, according to the energy balance model, be attributable to alterations in EE or energy loss in feces [35]. Physical activity was meticulously monitored with an accelerometer, and despite natural daily variation in physical activity, no significant changes were observed on a weekly basis in either group. Importantly, the lack of increments in physical activity levels in the RP group rules out unintentional and consistent increases in physical activity as a plausible factor driving weight loss. Fecal energy excretion, assessed by caloric bomb calorimetry [36], remained unchanged in both groups during the intervention, consistent with previous observations in mice consuming a protein‐restricted diet [11]. Weight loss or weight gain resistance mediated by reducing protein intake is well documented in rodent studies [9, 11, 14, 15, 16]. However, evidence in humans remains limited. Nevertheless, a few studies have reported weight loss following reduced dietary protein intake substituted with higher carbohydrate intake in individuals with obesity [19, 24, 37]. However, these studies implemented a greater reduction of protein intake (≤ 0.7 g protein kg−1 day−1), well below the minimum protein intake requirements set by WHO, which would be expected to lead to protein deficiency over time. The reduced protein content in the RP diet was substituted with higher carbohydrate content, which may also have influenced our findings. However, in our previous study in lean men, the metabolic effects of reducing dietary protein occurred independently of whether the reduced protein was replaced by carbohydrates or fat under eucaloric conditions [20]. Nevertheless, in the present cohort we cannot exclude the possibility that carbohydrate intake may have contributed to the observed results. Overall, our findings may indicate that reducing dietary protein, still meeting minimum requirements, substituted with increased carbohydrate intake promotes meaningful weight loss in men with overweight or obesity, even during isocaloric conditions and without restricting total energy intake. A concern with reducing dietary protein intake could be loss of LBM. However, in the present study, the RP group experienced a significant reduction in FM, while LBM only tended to be lower after the intervention. The change in body compartments was consistent with an approximately one‐to‐one fractional loss of FM and LBM, in line with previous studies in individuals with overweight or obesity during an isocaloric, protein‐reduced diet for 6 weeks [19], as well as observations following 5 weeks on calorie‐restricted diets [38]. LBM decreased in the present study primarily during the first 3 weeks (p = 0.07), whereas FM loss occurred from Week 3 to Week 5 (p = 0.01), suggesting a time‐dependent adaptation to reduced protein intake. This interpretation is supported by an initial reduction in plasma urea levels in the RP group, which returned to baseline levels after 3 weeks. In line with this, fasting total and individual plasma amino acid concentrations and 3‐methylhistidine, a potential marker of elevated muscle breakdown [39], remained stable in the RP group during the intervention despite the significantly lower protein intake. These findings may reflect an adaptive physiological response to conserve amino acids and attenuate protein turnover under conditions of reduced protein intake [40], potentially limiting further LBM loss over time as evident by the sustained LBM in the last 2 weeks. On the other hand, the consistently high protein intake in the HP group had no measurable effect on body composition. Overall, these findings suggest that reducing protein, substituted with increased carbohydrate intake for 5 weeks, lowers FM and does not lead to a greater loss of LBM compared with 5 weeks on a traditional calorie‐restricted diet in men with overweight or obesity. Reduced dietary protein intake has consistently been shown to increase circulating levels of FGF21 in rodent studies [9, 16] and humans following an acute low‐protein meal [20] and longer‐term protein‐reduced diets (from 24 h up to 6 weeks) [11, 19, 20, 41]. In the present study, fasting plasma FGF21 levels increased significantly in RP compared with HP. This observation supports the notion that humans increase circulating FGF21 levels in response to longer‐term reduction of protein, substituted with increased carbohydrates. Intriguingly, this cohort exhibited elevated baseline levels of FGF21 (mean: 237 ± 37 pg mL−1), exceeding the range seen in lean individuals (100–200 pg mL−1) [42], a finding consistent with prior reports in individuals with obesity [43]. This has been hypothesized to represent a state of obesity‐related “FGF21‐resistance” characterized by high circulating levels of FGF21 and reduced β‐klotho expression in SAT [44]. Although the physiological relevance of this remains unclear [45], it prompted us to explore local FGF21 signaling in SAT. We examined the gene and protein expression of FGFR1 and β‐klotho and the downstream signaling via ERK1/2 protein expression and its phosphorylation. However, receptor expression and downstream signaling were not affected indicating that FGF21 SAT signaling remains unaltered with RP. In rodents, reduced protein intake promotes weight loss by increasing EE, primarily through direct and indirect FGF21‐mediated thermogenesis in adipose tissue [9, 15, 16, 17]. In humans, FM has been shown to be a significant contributor to RMR, explaining at least 6% of its variance [46]. While this contribution appears modest compared with the weight loss observed in the present study, observations in burn patients have demonstrated increased mitochondrial leak in the respiratory electron transport chain of WAT [47], a phenotype comparable to that observed in the WAT proteome of lean men consuming a protein‐reduced diet for 5 weeks [20], which has been estimated to increase the energetic contribution of WAT up to threefold [48]. As proposed by Nicolaisen and colleagues [20], such a mechanism could, at least in part, explain the effects of reduced dietary protein on EE and, specifically, the weight loss observed in the present study. Nevertheless, given the pronounced weight loss in RP, other tissues, particularly skeletal muscle, might contribute to the effects of RP. Thus, we examined indicators of myocellular mitochondrial capacity by the assessment of protein abundance of the oxidative phosphorylation complexes. However, protein levels remained unchanged in both groups and did not imply any major effects on changed muscle mitochondrial energy utilization with RP. In the present study, increases in plasma FGF21 levels were significantly correlated with decreases in body weight. This relationship is supported by our previous findings in lean men, where increases in plasma FGF21 positively correlated with increases in energy required to maintain weight stability [20]. Vinales and colleagues [41] reported that increased plasma FGF21 levels were positively associated with elevated EE during a 24‐h period on a hypercaloric, low‐protein, high‐carbohydrate diet in individuals with obesity. Intriguingly, in the present study, no changes in RMR were detected. However, a FGF21‐induced increase in RMR might have been too subtle to be detected with the present methodology. Another potential explanation could be that RP‐induced increases in EE mainly occur postprandially. Such an interpretation is supported by our recent findings showing that a meal reduced in protein and increased in carbohydrates elicited enhanced postprandial EE compared with a meal higher in protein content in lean men [20]. Collectively, these findings support the notion that FGF21 may play a role in mediating the metabolic effects of reduction of dietary protein, substituted with an increase in carbohydrate intake in men with overweight or obesity. Glucose homeostasis is a key marker of metabolic health. In the present study, fasting surrogate indices of glucose homeostasis remained unchanged in both groups. In lean men, reducing dietary protein and increasing carbohydrate intake have previously been shown to markedly improve whole‐body insulin sensitivity, as evidenced by a higher glucose infusion rate during a hyperinsulinemic–euglycemic clamp, while HOMA‐IR remained unchanged [20]. Hence, a hyperinsulinemic–euglycemic clamp in individuals with overweight or obesity may reveal effects of reducing dietary protein and substituting it with increased carbohydrates on glucose regulation that were not captured by fasting measures in the present study. The present study only included men, in order to investigate if our previous findings in lean men consuming a eucaloric protein‐restricted diet [20] would translate into weight loss in men with overweight or obesity consuming an isocaloric diet. However, sex differences in metabolic regulation are well established [49]. Therefore, the inclusion of only men in the present study limits the extent to which our findings can be applied to the wider population. Notably, our results are consistent with previous studies investigating reducing dietary protein combined with higher carbohydrate intake in cohorts including both men and women [24, 37]. Still, future studies including both sexes are warranted to determine whether similar responses occur across sexes. Furthermore, in the present study, indirect calorimetry was used to estimate RMR. In future studies, determining total EE over a minimum of a 24‐h period or applying the doubly labeled water technique would potentially allow us to detect changes in diet‐induced thermogenesis, RMR, or both. In addition, due to the highly controlled nature of the intervention, in which an isocaloric diet with fixed energy intake was implemented, the findings should be verified in a free‐living setting. In such conditions, satiety responses and changes in energy intake may influence the outcomes. Longer and more diverse interventions conducted under free‐living conditions are therefore needed to determine the sustainability and broader applicability of the observed effects. Conclusion Reducing dietary protein intake to 0.9 g protein kg−1 BW day−1 for 5 weeks, with the reduced protein substituted with increased carbohydrate intake and without caloric restriction, resulted in significant weight loss in men with overweight or obesity under controlled conditions. This weight loss correlated with the increases in plasma FGF21 levels. Notably, this level of reduced protein intake reduced FM, without excessive LBM loss. Author Contributions B.K., A.L., and A.E.L. designed the study. A.E.L. designed the diets. A.E.L., J.K.J., and A.S.F. distributed and analyzed the diets. A.E.L., A.S.F., J.K.J., E.A.R., N.R.A., K.S.J., D.W., L.N., A.M.F., and B.K. collected and/or analyzed and/or interpreted the data. A.E.L., A.M.F., A.S.F. and B.K. drafted the manuscript. All authors contributed to the reviewing and finalization of the manuscript. Funding B.K. and A.E.L. were funded by the Independent Research Fund Denmark (DFF) (grant no. 9039‐00257B). A.M.F. was funded by the Novo Nordisk Foundation (grant no. NNF22OC0074110). Conflicts of Interest The authors declare no conflicts of interest. Supporting information Figure S1: Fasting plasma amino acid concentrations with varying protein content. (a) Fasting plasma isoleucine, (b) fasting plasma valine, (c) fasting plasma lysine, (d) fasting plasma phenylalanine, (e) fasting plasma histidine, (f) fasting plasma alanine, (g) fasting plasma serine, (h) fasting plasma proline, (i) fasting plasma arginine, (j) fasting plasma glutamic acid, (k) fasting plasma tyrosine, (j) fasting plasma 1‐methylhistidine concentrations during 5 weeks on a HP diet or RP diet. HP group (∆) and RP group (●). Two‐way RM ANOVA was applied for statistical analysis with post hoc Tukey‐Kramer multiple comparison test. Data are presented as means ±SEM and individual values. HP: n = 7, and RP: n = 10. Figure S2: Expression of selected protein from adipose and muscle tissue. (a) Representative Western blots demonstrating FGFR1, β‐klotho, phosphorylated ERK1/2, total ERK1/2 protein, total protein staining (Pierce) and GAPDH before and after in RP group and HP group, (b) adipose β‐klotho normalized to GAPDH protein content, (c) adipose FGFR1 normalized to GAPDH protein content, (d) adipose ERK1/2 phosphorylation related total ERK1/2 protein content, (e) representative Western blots illustrating Oxphos complexes I‐V, GAPDH and total protein staining (Coomassie) before and after in RP and HP group, (f) muscle Oxphos complexes I‐V protein contents normalized to GAPDH protein content, and (g) muscle Oxphos complexes I‐V protein content normalized to total protein assessed by Coomassie staining before and after 5 weeks on a HP diet or RP diet. HP group in stripes and RP group in filled boxes. Two‐way RM ANOVA was applied for statistical analysis. Data are presented as means and individual values. HP: n = 6, and RP: n = 10. Table S1: Self‐registered habitual diet. Table S2: Multiple reaction monitoring (MRM) transitions and parameters for individual amino acid analysis by UPLC‐MS/MS. Table S3: Overview of Western blotting proteins and antibodies. Table S4: Primers for RT‐qPCR analysis. Acknowledgments We are thankful to Irene Bech Nielsen, Betina Bolmgren, Kim Sjøberg (The August Krogh Section for Human and Molecular Physiology), and Bodil Madsen (DTU Food) for skilled technical assistance, Le Lene Stevner for assistance with the ethics application, and the experimental contributions from Nida Ghauri, Christina Nielsen, and Nana Friis (University of Copenhagen). We also thank all the participants for participating in the study. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Deidentified participant data will be provided upon a methodologically sound proposal, in accordance with ethical approvals and data use agreements. Any shared data will be limited to what is deemed necessary to achieve the aims of the proposed analysis. References Associated Data Supplementary Materials Figure S1: Fasting plasma amino acid concentrations with varying protein content. (a) Fasting plasma isoleucine, (b) fasting plasma valine, (c) fasting plasma lysine, (d) fasting plasma phenylalanine, (e) fasting plasma histidine, (f) fasting plasma alanine, (g) fasting plasma serine, (h) fasting plasma proline, (i) fasting plasma arginine, (j) fasting plasma glutamic acid, (k) fasting plasma tyrosine, (j) fasting plasma 1‐methylhistidine concentrations during 5 weeks on a HP diet or RP diet. HP group (∆) and RP group (●). Two‐way RM ANOVA was applied for statistical analysis with post hoc Tukey‐Kramer multiple comparison test. Data are presented as means ±SEM and individual values. HP: n = 7, and RP: n = 10. Figure S2: Expression of selected protein from adipose and muscle tissue. (a) Representative Western blots demonstrating FGFR1, β‐klotho, phosphorylated ERK1/2, total ERK1/2 protein, total protein staining (Pierce) and GAPDH before and after in RP group and HP group, (b) adipose β‐klotho normalized to GAPDH protein content, (c) adipose FGFR1 normalized to GAPDH protein content, (d) adipose ERK1/2 phosphorylation related total ERK1/2 protein content, (e) representative Western blots illustrating Oxphos complexes I‐V, GAPDH and total protein staining (Coomassie) before and after in RP and HP group, (f) muscle Oxphos complexes I‐V protein contents normalized to GAPDH protein content, and (g) muscle Oxphos complexes I‐V protein content normalized to total protein assessed by Coomassie staining before and after 5 weeks on a HP diet or RP diet. HP group in stripes and RP group in filled boxes. Two‐way RM ANOVA was applied for statistical analysis. Data are presented as means and individual values. HP: n = 6, and RP: n = 10. Table S1: Self‐registered habitual diet. Table S2: Multiple reaction monitoring (MRM) transitions and parameters for individual amino acid analysis by UPLC‐MS/MS. Table S3: Overview of Western blotting proteins and antibodies. Table S4: Primers for RT‐qPCR analysis. Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request. Deidentified participant data will be provided upon a methodologically sound proposal, in accordance with ethical approvals and data use agreements. Any shared data will be limited to what is deemed necessary to achieve the aims of the proposed analysis.