Discussion
This study demonstrates that metabolically unhealthy lean and obese groups had similarity in overall fat distribution with increased android to gynoid fat ratio compared with respective healthy groups. However, the main abnormality in lean unhealthy groups was a higher android percent fat mass, while the main abnormality in the unhealthy obese groups was a lower gynoid percent fat mass. This suggests that despite several similarities, there are subtle differences in fat distribution between individuals who are lean versus those who are obese in the presence of metabolic abnormalities. In particular, android fat may play a larger role in lean individuals, and lower gynoid fat may play a larger role in those with obesity, suggesting a difference in the underlying mechanisms responsible for metabolic abnormalities in lean individuals as compared with those with obesity.
It is well established that increased android fat mass is associated with an increase in insulin resistance and MetS.3 22 Decreased gynoid or gluteofemoral fat mass is also associated with MetS abnormalities. In one study, increased thigh fat was associated with lower glucose and lipid levels after accounting for abdominal fat that was associated with high glucose and lipid levels.23 In the Women’s Health Initiative study, in postmenopausal women with normal BMI, lower gluteofemoral fat mass, estimated by DXA, was associated with a higher incidence of CVD independent of increased trunk fat mass.15 Evidence from precise phenotyping studies and from genetic studies shows that increased gluteofemoral and leg fat mass is protective of cardiometabolic diseases.22 24 Increased gluteofemoral fat mass is also independently associated with a better lipid and glucose profile, as well as a decrease in cardiovascular and metabolic risks.18 One recent study showed higher vaspin levels were associated with decreased gluteofemoral fat and increased risk of T2D.25 Thus, decreased gluteofemoral fat mass may explain some of the metabolic abnormalities in MUO individuals.
This study confirms the findings from our previous study comparing MUL and MHL individuals,16 but it expands the findings by including a comparison between the lean population and that with obesity. The study draws attention to the lean population with metabolic abnormalities while much of the published literature has focused on the population with obesity. In general, individuals with obesity, with or without MetS, are at a higher risk of T2D and CVD than lean individuals.26 Among individuals with obesity without metabolic abnormalities, BMI is independently associated with increased risk of T2D and CVD,8 but the risk is increased by almost twofold in the presence of metabolic abnormalities associated with increased visceral fat and decreased gynoid fat.27 Thus, there is a variable contribution of increased subcutaneous fat, increased visceral fat and decreased gynoid fat to the causation of metabolic abnormalities in people with obesity.14 However, in lean or normal-weight individuals, an increased proportion of visceral fat seems to be the main cause of metabolic abnormalities. Although we saw increased gynoid fat in MUL individuals that may be related to increased total fat, we think the overall balance was more toward dominant android fat in this group. Thus, the negative effect of much higher android fat might have neutralized the positive effect of increased gynoid fat on metabolic parameters in lean individuals.
Our observations are important because a significant percentage of lean or normal-weight adults have metabolic abnormalities, highlighting a need for different treatment strategies among MUL and MUO individuals. Despite the concept of metabolically obese lean or metabolically unhealthy normal-weight individuals being recognized for a long time,28 29 treatment strategies have not been studied in this population. Current treatment of metabolic abnormalities is derived mostly from studies in the obese populations. We suggest further studies be conducted focused specifically on the unhealthy lean population.
The strength of this study lies in its large, nationally representative sample from the NHANES dataset, with body fat distribution measured using DXA, the gold standard technique. Although a priori power calculations were not feasible due to the fixed survey design, subgroup sizes were substantial, including 120 to over 1000 participants even within the MUL group, making the study unlikely to be underpowered. In fact, some comparisons may be overpowered, and results should be interpreted with attention to both statistical and clinical significance. The main limitation is missing variables in a substantial number because DXA was done in only a subset of participants. The observed differences in fat distribution between unhealthy and healthy groups were modest and may partly reflect measurement variability, though our findings align with prior studies. Additionally, the cross-sectional nature of NHANES limits causal inference, and despite adjustment for key covariates and stratified analyses, residual confounding cannot be entirely ruled out. While the overall sample was large, some subgroup comparisons, particularly those involving smaller strata such as younger females, may have limited power for detecting interaction effects. Therefore, such findings should be considered hypothesis generating and require validation in prospective or mechanistic studies to clarify their clinical implications. In addition, there was a significant age difference between healthy and unhealthy groups across all BMI categories, with unhealthy people being older than healthy people. The age difference was wider in lower BMI categories than higher BMI categories and may have contributed to unhealth in the lean group. Despite these limitations, the patterns observed were consistent across analyses and suggest biologically plausible differences that warrant further investigation in prospective and mechanistic studies. Our goal was to investigate the differences between unhealthy and healthy phenotypes irrespective of the events leading to metabolic unhealth.
We conclude that there are subtle differences in fat distribution between MUL and MUO individuals, suggesting different mechanisms leading to metabolic abnormalities. The differences are small and may not be directly clinically relevant but will lead to further studies to investigate the biological or genetic basis of our findings. One previous study linked metabolic abnormalities in lean individuals to lipodystrophy genes.30 There may also be hormonal differences (eg, 11-beta-hydroxysteroid dehydrogenase activity) between the MUL and MUO individuals. However, further research is needed to understand these mechanisms and to develop treatment strategies for lean individuals with MetS.
Leave a Reply