Assessing the Risk of Bias in Randomized Clinical Trials With Large Language Models

Importance

Large language models (LLMs) may facilitate the labor-intensive process of systematic reviews. However, the exact methods and reliability remain uncertain.

Objective

To explore the feasibility and reliability of using LLMs to assess risk of bias (ROB) in randomized clinical trials (RCTs).

Design, Setting, and Participants

A survey study was conducted between August 10, 2023, and October 30, 2023. Thirty RCTs were selected from published systematic reviews.

Main Outcomes and Measures

A structured prompt was developed to guide ChatGPT (LLM 1) and Claude (LLM 2) in assessing the ROB in these RCTs using a modified version of the Cochrane ROB tool developed by the CLARITY group at McMaster University. Each RCT was assessed twice by both models, and the results were documented. The results were compared with an assessment by 3 experts, which was considered a criterion standard. Correct assessment rates, sensitivity, specificity, and F1 scores were calculated to reflect accuracy, both overall and for each domain of the Cochrane ROB tool; consistent assessment rates and Cohen κ were calculated to gauge consistency; and assessment time was calculated to measure efficiency. Performance between the 2 models was compared using risk differences.

Results

Both models demonstrated high correct assessment rates. LLM 1 reached a mean correct assessment rate of 84.5% (95% CI, 81.5%-87.3%), and LLM 2 reached a significantly higher rate of 89.5% (95% CI, 87.0%-91.8%). The risk difference between the 2 models was 0.05 (95% CI, 0.01-0.09). In most domains, domain-specific correct rates were around 80% to 90%; however, sensitivity below 0.80 was observed in domains 1 (random sequence generation), 2 (allocation concealment), and 6 (other concerns). Domains 4 (missing outcome data), 5 (selective outcome reporting), and 6 had F1 scores below 0.50. The consistent rates between the 2 assessments were 84.0% for LLM 1 and 87.3% for LLM 2. LLM 1's κ exceeded 0.80 in 7 and LLM 2's in 8 domains. The mean (SD) time needed for assessment was 77 (16) seconds for LLM 1 and 53 (12) seconds for LLM 2.

Conclusions

In this survey study of applying LLMs for ROB assessment, LLM 1 and LLM 2 demonstrated substantial accuracy and consistency in evaluating RCTs, suggesting their potential as supportive tools in systematic review processes.

Impact of Bariatric Surgery on the Stability of the Genetic Material, Oxidation, and Repair of DNA and Telomere Lengths

Obesity causes genetic instability, which plays a key-role in the etiology of cancer and aging. We investigated the impact of bariatric surgery (BS) on DNA repair, oxidative DNA damage, telomere lengths, alterations of antioxidant enzymes and, selected proteins which reflect inflammation. The study was realized with BS patients (n = 35). DNA damage, base oxidation, BER, and NER were measured before and 1 month and 6 months after surgery with the single-cell gel electrophoresis technique. SOD and GPx were quantified spectrophotometrically, malondealdehyde (MDA) was quantified by HPLC. Telomere lengths were determined with qPCR, and plasma proteome profiling was performed with high-resolution mass spectrophotometry. Six months after the operations, reduction of body weight by 27.5% was observed. DNA damage decreased after this period, this effect was paralleled by reduced formation of oxidized DNA bases, a decline in the MDA levels and of BER and NER, and an increase in the telomere lengths. The activities of antioxidant enzymes were not altered. Clear downregulation of certain proteins (CRP, SAA1) which reflect inflammation and cancer risks was observed. Our findings show that BS causes reduced oxidative damage of DNA bases, possibly as a consequence of reduction of inflammation and lipid peroxidation, and indicate that the surgery has beneficial long-term health effects.

Identification of the hub genes related to adipose tissue metabolism of bovine

Due to the demand for high-quality animal protein, there has been consistent interest in how to obtain more high-quality beef. As well-known, the adipose content of beef has a close connection with the taste and quality of beef, and cattle with different energy or protein diet have corresponding effects on the lipid metabolism of beef. Thus, we performed weighted gene co-expression network analysis (WGCNA) with subcutaneous adipose genes from Norwegian red heifers fed different diets to identify hub genes regulating bovine lipid metabolism. For this purpose, the RNA sequencing data of subcutaneous adipose tissue of 12-month-old Norwegian red heifers (n = 48) with different energy or protein levels were selected from the GEO database, and 7,630 genes with the largest variation were selected for WGCNA analysis. Then, three modules were selected as hub genes candidate modules according to the correlation between modules and phenotypes, including pink, magenta and grey60 modules. GO and KEGG enrichment analysis showed that genes were related to metabolism, and participated in Rap, MAPK, AMPK, VEGF signaling pathways, and so forth. Combined gene interaction network analysis using Cytoscape software, eight hub genes of lipid metabolism were identified, including TIA1, LOC516108, SNAPC4, CPSF2, ZNF574, CLASRP, MED15 and U2AF2. Further, the expression levels of hub genes in the cattle tissue were also measured to verify the results, and we found hub genes in higher expression in muscle and adipose tissue in adult cattle. In summary, we predicted the key genes of lipid metabolism in the subcutaneous adipose tissue that were affected by the intake of various energy diets to find the hub genes that coordinate lipid metabolism, which provide a theoretical basis for regulating beef quality.

Low-carbohydrate versus balanced-carbohydrate diets for reducing weight and cardiovascular risk

BACKGROUND

Debates on effective and safe diets for managing obesity in adults are ongoing. Low-carbohydrate weight-reducing diets (also known as 'low-carb diets') continue to be widely promoted, marketed and commercialised as being more effective for weight loss, and healthier, than 'balanced'-carbohydrate weight-reducing diets.

OBJECTIVES

To compare the effects of low-carbohydrate weight-reducing diets to weight-reducing diets with balanced ranges of carbohydrates, in relation to changes in weight and cardiovascular risk, in overweight and obese adults without and with type 2 diabetes mellitus (T2DM).

SEARCH METHODS

We searched MEDLINE (PubMed), Embase (Ovid), the Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science Core Collection (Clarivate Analytics), ClinicalTrials.gov and WHO International Clinical Trials Registry Platform (ICTRP) up to 25 June 2021, and screened reference lists of included trials and relevant systematic reviews. Language or publication restrictions were not applied.

SELECTION CRITERIA

We included randomised controlled trials (RCTs) in adults (18 years+) who were overweight or living with obesity, without or with T2DM, and without or with cardiovascular conditions or risk factors. Trials had to compare low-carbohydrate weight-reducing diets to balanced-carbohydrate (45% to 65% of total energy (TE)) weight-reducing diets, have a weight-reducing phase of 2 weeks or longer and be explicitly implemented for the primary purpose of reducing weight, with or without advice to restrict energy intake.  DATA COLLECTION AND ANALYSIS: Two review authors independently screened titles and abstracts and full-text articles to determine eligibility; and independently extracted data, assessed risk of bias using RoB 2 and assessed the certainty of the evidence using GRADE. We stratified analyses by participants without and with T2DM, and by diets with weight-reducing phases only and those with weight-reducing phases followed by weight-maintenance phases. Primary outcomes were change in body weight (kg) and the number of participants per group with weight loss of at least 5%, assessed at short- (three months to < 12 months) and long-term (≥ 12 months) follow-up.

MAIN RESULTS

We included 61 parallel-arm RCTs that randomised 6925 participants to either low-carbohydrate or balanced-carbohydrate weight-reducing diets. All trials were conducted in high-income countries except for one in China. Most participants (n = 5118 randomised) did not have T2DM. Mean baseline weight across trials was 95 kg (range 66 to 132 kg). Participants with T2DM were older (mean 57 years, range 50 to 65) than those without T2DM (mean 45 years, range 22 to 62). Most trials included men and women (42/61; 3/19 men only; 16/19 women only), and people without baseline cardiovascular conditions, risk factors or events (36/61). Mean baseline diastolic blood pressure (DBP) and low-density lipoprotein (LDL) cholesterol across trials were within normal ranges. The longest weight-reducing phase of diets was two years in participants without and with T2DM. Evidence from studies with weight-reducing phases followed by weight-maintenance phases was limited. Most trials investigated low-carbohydrate diets (> 50 g to 150 g per day or < 45% of TE; n = 42), followed by very low (≤ 50 g per day or < 10% of TE; n = 14), and then incremental increases from very low to low (n = 5). The most common diets compared were low-carbohydrate, balanced-fat (20 to 35% of TE) and high-protein (> 20% of TE) treatment diets versus control diets balanced for the three macronutrients (24/61). In most trials (45/61) the energy prescription or approach used to restrict energy intake was similar in both groups. We assessed the overall risk of bias of outcomes across trials as predominantly high, mostly from bias due to missing outcome data. Using GRADE, we assessed the certainty of evidence as moderate to very low across outcomes.  Participants without and with T2DM lost weight when following weight-reducing phases of both diets at the short (range: 12.2 to 0.33 kg) and long term (range: 13.1 to 1.7 kg).  In overweight and obese participants without T2DM: low-carbohydrate weight-reducing diets compared to balanced-carbohydrate weight-reducing diets (weight-reducing phases only) probably result in little to no difference in change in body weight over three to 8.5 months (mean difference (MD) -1.07 kg, (95% confidence interval (CI) -1.55 to -0.59, I2 = 51%, 3286 participants, 37 RCTs, moderate-certainty evidence) and over one to two years (MD -0.93 kg, 95% CI -1.81 to -0.04, I2 = 40%, 1805 participants, 14 RCTs, moderate-certainty evidence); as well as change in DBP and LDL cholesterol over one to two years. The evidence is very uncertain about whether there is a difference in the number of participants per group with weight loss of at least 5% at one year (risk ratio (RR) 1.11, 95% CI 0.94 to 1.31, I2 = 17%, 137 participants, 2 RCTs, very low-certainty evidence).  In overweight and obese participants with T2DM: low-carbohydrate weight-reducing diets compared to balanced-carbohydrate weight-reducing diets (weight-reducing phases only) probably result in little to no difference in change in body weight over three to six months (MD -1.26 kg, 95% CI -2.44 to -0.09, I2 = 47%, 1114 participants, 14 RCTs, moderate-certainty evidence) and over one to two years (MD -0.33 kg, 95% CI -2.13 to 1.46, I2 = 10%, 813 participants, 7 RCTs, moderate-certainty evidence); as well in change in DBP, HbA1c and LDL cholesterol over 1 to 2 years. The evidence is very uncertain about whether there is a difference in the number of participants per group with weight loss of at least 5% at one to two years (RR 0.90, 95% CI 0.68 to 1.20, I2 = 0%, 106 participants, 2 RCTs, very low-certainty evidence).  Evidence on participant-reported adverse effects was limited, and we could not draw any conclusions about these.  AUTHORS' CONCLUSIONS: There is probably little to no difference in weight reduction and changes in cardiovascular risk factors up to two years' follow-up, when overweight and obese participants without and with T2DM are randomised to either low-carbohydrate or balanced-carbohydrate weight-reducing diets.

Genotoxic effect of meat consumption: A mini review

In 2015, the International Agency for Research on Cancer classified the consumption of processed meat as carcinogenic to humans (Group 1) and red meat as probably carcinogenic to humans (Group 2A) based on sufficient data from animal models and epidemiological studies. However, research characterising the mechanisms underlying this carcinogenic process in humans are limited, particularly with respect to measures of direct DNA damage. The current review sought to evaluate and summarize the recent literature, published since 2000, regarding the associations of meat consumption and three biomarkers of genotoxicity in humans: DNA strand breaks (measured using the comet assay), DNA adducts, and micronucleus formation. After screening 230 potential articles, 35 were included, and then were classified as experimental or observational in design, the latter of which were further categorized according to their dietary assessment approach. Among the 30 observational studies, 4 of which used two different assays, 3 of 5 comet assay studies, 13 of 20 DNA adduct studies, and 7 of 9 micronucleus studies reported a positive association between meat consumption and DNA damage. Among the 5 experimental studies, 1 of 1 using the comet assay, 3 of 3 measuring DNA adducts and 0 of 1 measuring micronuclei reported significant positive associations with meat consumption. Nevertheless, common limitations among the selected publications included small sample size, and poor methodological reporting of both exposure and outcome measures. Moreover, the vast majority of studies only measured DNA damage in one biological sample using a single assay and we cannot exclude the possibility of publication bias. Ultimately, our review of the literature, published since 2000, revealed a preponderance of studies that support mechanisms of genotoxicity in playing an important role in the meat-cancer association.

Comparative efficacy of a 12 week yoga-based lifestyle intervention and dietary intervention on adipokines, inflammation, and oxidative stress in adults with metabolic syndrome: a randomized controlled trial

The present randomized controlled trial (RCT) evaluated the comparative efficacy of 12 week yoga-based lifestyle intervention (YBLI) and dietary intervention (DI) alone on adipokines, inflammation, and oxidative stress in Indian adults with metabolic syndrome (Met S). A parallel, two arm, RCT was conducted in Integral Health Clinic (IHC), All India Institute of Medical Sciences, India from 2012 to 2014. IHC is an outpatient facility conducting YBLI programs for prevention and management of chronic diseases. Two hundred sixty men and women (20-45 years) visiting the outpatient department of a tertiary care hospital were diagnosed with Met S and randomized 1:1 to receive 12 week YBLI (n = 130) or DI (n = 130). Primary outcomes were change in plasma levels of adipokines (leptin, adiponectin, and leptin:adiponectin ratio), markers of inflammation (tumor necrosis factor [TNF]-α, interleukin [IL]-6), markers of oxidative stress (thiobarbituric acid reactive substances [TBARS], 8-hydroxy-2'-deoxyguanosine [8-OHdG], and superoxide dismutase [SOD]) measured at baseline, 2 weeks, and 12 weeks. YBLI group showed a significant decrease in leptin, leptin:adiponectin ratio, IL-6, 8-OHdG, and TBARS levels, whereas there was a significant increase in adiponectin and SOD levels. No significant changes were noticed in DI alone group. YBLI showed significantly greater reduction in TBARS levels than in DI group, suggestive of reduced oxidative stress in adults with Met S. A 12 week YBLI had a positive impact on oxidative stress versus DI alone in adults with Met S.

Influence of bariatric surgery induced weight loss on oxidative DNA damage

Obesity is associated with elevated cancer risk, which may be represented by elevated genomic damage. Oxidative stress plays a key role in obesity related detrimental health consequences including DNA oxidation damage. The elevated cancer risk in obesity may be a consequence. Weight loss has been shown to reduce genomic damage, but the role of oxidative stress in that has not been clarified. The aim of this study is therefore to investigate the influence of bariatric surgery induced weight loss on DNA oxidation damage in morbidly obese subjects. For this aim, we used cryopreserved peripheral blood mononuclear cells in the FPG comet assay. Advanced protein oxidation products and 3-nitrotyrosine were measured as oxidative and nitrative protein stress markers. Furthermore, expression of oxidative stress related proteins HSP70 and Nrf2 as well as mitochondrial enzyme citrate synthase and NADPH oxidase subunit p22 phox were analysed. Our findings revealed significantly reduced DNA strand breaks, but DNA base oxidation was not reduced. We observed significant reduction in plasma AOPPs and 3-nitrotyrosine, which indicated an improvement in oxidative/nitrative stress. However, expression of HSP70 and Nrf2 were not altered after weight loss. In addition, expression of citrate synthase and p22 phox were also unaltered. Overall, bariatric surgery induced significant reduction in excess body weight and improved the patients' health status, including reduced DNA strand breaks and slightly improved antioxidant status in some of the investigated endpoints, while cellular ROS formation and DNA oxidation damage stayed unaltered. This complex situation may be due to combined beneficial effects of weight loss and burdening of the body with fat breakdown products. In the future, collecting samples two years after surgery, when patients have been in a weight plateau for some time, might be a promising approach.

Effect of Lower Versus Higher Red Meat Intake on Cardiometabolic and Cancer Outcomes: A Systematic Review of Randomized Trials

This article has been corrected. The original version (PDF) is appended to this article as a Supplement.

BACKGROUND

Few randomized trials have evaluated the effect of reducing red meat intake on clinically important outcomes.

PURPOSE

To summarize the effect of lower versus higher red meat intake on the incidence of cardiometabolic and cancer outcomes in adults.

DATA SOURCES

EMBASE, CENTRAL, CINAHL, Web of Science, and ProQuest from inception to July 2018 and MEDLINE from inception to April 2019, without language restrictions.

STUDY SELECTION

Randomized trials (published in any language) comparing diets lower in red meat with diets higher in red meat that differed by a gradient of at least 1 serving per week for 6 months or more.

DATA EXTRACTION

Teams of 2 reviewers independently extracted data and assessed the risk of bias and the certainty of the evidence.

DATA SYNTHESIS

Of 12 eligible trials, a single trial enrolling 48 835 women provided the most credible, though still low-certainty, evidence that diets lower in red meat may have little or no effect on all-cause mortality (hazard ratio [HR], 0.99 [95% CI, 0.95 to 1.03]), cardiovascular mortality (HR, 0.98 [CI, 0.91 to 1.06]), and cardiovascular disease (HR, 0.99 [CI, 0.94 to 1.05]). That trial also provided low- to very-low-certainty evidence that diets lower in red meat may have little or no effect on total cancer mortality (HR, 0.95 [CI, 0.89 to 1.01]) and the incidence of cancer, including colorectal cancer (HR, 1.04 [CI, 0.90 to 1.20]) and breast cancer (HR, 0.97 [0.90 to 1.04]).

LIMITATIONS

There were few trials, most addressing only surrogate outcomes, with heterogeneous comparators and small gradients in red meat consumption between lower versus higher intake groups.

CONCLUSION

Low- to very-low-certainty evidence suggests that diets restricted in red meat may have little or no effect on major cardiometabolic outcomes and cancer mortality and incidence.

PRIMARY FUNDING SOURCE

None (PROSPERO: CRD42017074074).

No effect of caloric restriction or exercise on radiation repair capacity

INTRODUCTION

Maintenance of normal weight and higher levels of physical activity are associated with a reduced risk of several types of cancer. Because genomic instability is regarded as a hallmark of cancer development, one proposed mechanism is improvement of DNA repair function. We investigated links between dietary weight loss, exercise, and strand break rejoining in an ancillary study to a randomized-controlled trial.

METHODS

Overweight/obese postmenopausal women (n = 439) were randomized to the following: a) reduced calorie weight loss diet ("diet," n = 118), b) moderate- to vigorous-intensity aerobic exercise ("exercise," n = 117), c) a combination ("diet + exercise," n = 117), or d) control (n = 87). The reduced calorie diet had a 10% weight loss goal. The exercise intervention consisted of 45 min of moderate to vigorous aerobic activity 5 d·wk for 12 months. DNA repair capacity was measured in a subset of 226 women at baseline and 12 months from cryopreserved peripheral mononuclear cells using the comet assay. Anthropometric and body composition measures were performed at baseline and 12 months.

RESULTS

DNA repair capacity did not change significantly with any of the 12-month interventions compared with control; there were also no significant changes when stratified by changes in body composition or aerobic fitness (V˙O2max). At baseline, DNA repair capacity was positively associated with weight, body mass index, and fat mass (r = 0.20, P = 0.003; r = 0.19, P = 0.004; r = 0.13, P = 0.04, respectively) and inversely with lean body mass (r = -0.14, P = 0.04).

CONCLUSION

In conclusion, DNA repair capacity in cryopreserved PBMCs (Comet Assay) did not change with dietary weight loss or exercise interventions in postmenopausal women within a period of 12 months. Other assays that capture different facets of DNA repair function may be needed.

Clinical Effectiveness of Weight Loss and Weight Maintenance Interventions for Men: A Systematic Review of Men-Only Randomized Controlled Trials (The ROMEO Project)

Men are underrepresented in obesity services, suggesting current weight loss service provision is suboptimal. This systematic review evaluated evidence-based strategies for treating obesity in men. Eight bibliographic databases and four clinical trials' registers were searched to identify randomized controlled trials (RCTs) of weight loss interventions in men only, with mean/median body mass index of ≥30 kg/m² (or ≥28 kg/m² with cardiac risk factors), with a minimum mean/median duration of ≥52 weeks. Interventions included diet, physical activity, behavior change techniques, orlistat, or combinations of these; compared against each other, placebo, or a no intervention control group; in any setting. Twenty-one reports from 14 RCTs were identified. Reducing diets produced more favorable weight loss than physical activity alone (mean weight change after 1 year from a reducing diet compared with an exercise program -3.2 kg, 95% confidence interval -4.8 to -1.6 kg, reported p < .01). The most effective interventions combined reducing diets, exercise, and behavior change techniques (mean difference in weight at 1 year compared with no intervention was -4.9 kg, 95% confidence interval -5.9 to -4.0, reported p < .0001). Group interventions produced favorable weight loss results. The average reported participant retention rate was 78.2%, ranging from 44% to 100% retention, indicating that, once engaged, men remained committed to a weight loss intervention. Weight loss for men is best achieved and maintained with the combination of a reducing diet, increased physical activity, and behavior change techniques. Strategies to increase engagement of men with weight loss services to improve the reach of interventions are needed.

Dairy foods and dairy protein consumption is inversely related to markers of adiposity in obese men and women

A number of intervention studies have reported that the prevalence of obesity may be in part inversely related to dairy food consumption while others report no association. We sought to examine relationships between energy, protein and calcium consumption from dairy foods (milk, yoghurt, cheese, dairy spreads, ice-cream) and adiposity including body mass index (BMI), waist (WC) and hip circumference (HC), and direct measures of body composition using dual energy X-ray absorptiometry (% body fat and abdominal fat) in an opportunistic sample of 720 overweight/obese Australian men and women. Mean (SD) age, weight and BMI of the population were 51 ± 10 year, 94 ± 18 kg and 32.4 ± 5.7 kg/m², respectively. Reduced fat milk was the most commonly consumed dairy product (235 ± 200 g/day), followed by whole milk (63 ± 128 g/day) and yoghurt (53 ± 66 g/day). Overall dairy food consumption (g/day) was inversely associated with BMI, % body fat and WC (all p < 0.05). Dairy protein and dairy calcium (g/day) were both inversely associated with all adiposity measures (all p < 0.05). Yoghurt consumption (g/day) was inversely associated with % body fat, abdominal fat, WC and HC (all p < 0.05), while reduced fat milk consumption was inversely associated with BMI, WC, HC and % body fat (all p < 0.05). Within a sample of obese adults, consumption of dairy products, dairy protein, and calcium was associated with more favourable body composition.

The effect of age, gender, diet and lifestyle on DNA damage measured using micronucleus frequency in human peripheral blood lymphocytes

Micronucleus (MN) frequency in cytokinesis-blocked peripheral blood lymphocytes (PBL) has become one of the best-established biomarkers for studying DNA damage occurring in vivo in humans. The application of this method in population biomonitoring studies requires a deep understanding of how lifestyle and common host variables may influence MN frequency in PBL. In this mini-review, an update is provided on results from studies reporting on the impact of age, gender, diet and lifestyle factors (e.g. exercise, alcohol, smoking and recreational drugs) on this biomarker. Evidence from these studies shows that each of these factors, either in isolation or in combination, can significantly influence MN frequency. Proper control for these factors is required to enable better measurement of the impact of other conditions, such as environmental exposure to genotoxins or a susceptible genetic background, on MN frequency in PBL.

Moderate energy restriction-induced weight loss affects circulating IGF levels independent of dietary composition

BACKGROUND

Obesity is associated with major changes in the circulating IGF system. However, it is not clear to what extent the IGF system is normalized following diet, and the possible role of different types of diet is also unknown.

OBJECTIVE

To compare changes in the circulating IGF system following 12 weeks of moderate energy restriction (7000 kJ/day) in overweight or obese males on a high protein high red meat diet (HP) or a high carbohydrate diet (HC).

DESIGN

Seventy-six men (mean age, 51+/-1.0 years; body mass index, 32.8+/-0.5 kg/m(2)) were allocated to matched groups treated with isocaloric diets of HP (n=34) or HC (n=42). Outcome measures were weight, body composition, IGF-related peptides, homoeostasis model assessment of insulin resistance (HOMA1-IR) and adipokines.

RESULTS

Weight loss did not differ between diets (HP 8.5+/-0.6 kg; HC 8.2+/-0.6 kg, P>0.05). IGF-related peptides increased total IGF1 (HP 23%; HC 18%, P<0.0001), bioactive IGF1 (HP 18%; HC 15%, P<0.002), IGF1:IGF-binding protein-3 (IGFBP-3; HP 29%; HC 22%, P<0.0001) and IGFBP-1 (HP 24%; HC 25%, P<0.01). By contrast, decreases were observed in IGFBP-3 (HP -4%; HC -3%, P<0.01), pro-IGF2 (HP -3%; HC -6%, P=0.001), total IGF2 (HP -7%; HC -3%, P=0.001) and sIGF2R (HP -10%; HC -6%, P<0.005). Only IGFBP-2 increased differentially by diet (HP 34%; HC 50%, P<0.0001, diet P<0.05). Adiponectin increased in both diets, but leptin and HOMA-IR decreased (P<0.001).

CONCLUSIONS

Weight loss induced by moderate energy restriction modulated the IGF system independent of dietary protein or red meat content. The effect of diet on IGFBP-2 appeared to have limited biological effect as total IGF2 and pro-IGF2 did not change.