It's worth a moment to step back and consider how this objective is not typical of clinical trials, and instead is an attempt to better assess the presence of healthy function rather than the presence of disease or disease-based biomarkers. A shift toward viewing health not just as the absence of disease but as the presence of flexibility, adaptability, elasticity, etc. (in response to environmental changes/challenges) has led to a search for “biomarkers of health” rather than “biomarkers of disease.” When thinking about metabolic health, this has fueled the hunt for good tools for assessing adaptability; for example, an oral glucose tolerance test (OGTT) can evaluate how effectively and rapidly a person responds to a glucose load, and this functional assessment is quite different than a fasting plasma glucose (FPG). Although FPGs have the advantage of being quick, easy, and inexpensive, they are also quite poor tests when considering both metabolic and phenotypic flexibility, and even for accurately screening for diabetes, a much later stage on the spectrum of metabolic dysfunction.
For example, in a systematic review and meta-analysis published in BMJ in 2017, an FPG was found to have a specificity of 94% for diagnosing diabetes, indicating that if an FPG is elevated, diabetes is very likely. However, this analysis of over 138 studies found that an FPG had a sensitivity (a measure of how well it screens for diabetes) of only 25%, making it an extremely poor method for catching undiagnosed cases of diabetes, and certainly not helpful for assessing impaired glucose tolerance. In contrast, the degree of postprandial hyperglycemia (the rise in blood glucose following a meal), is a much better predictor of pathology and disease risk (as well as poor metabolic flexibility) than an FPG. Postprandial glucose is linked to cardiovascular mortality in both diabetics and non-diabetics, and has been associated with “vascular stiffness, brachial-ankle pulse-wave velocity, carotid intima thickness, and left ventricular hypertrophy” to a much greater degree than FPG or even HbA1c.
Also intrinsic to this more function-based assessment is the focus on the interaction and communication between body systems, and the complexity involved in regulating glucose homeostasis. For example, a previous report indicates that the metabolites that change the most in response to a glucose challenge are primary and secondary bile acids, which have been reported to increase up to 8-fold in concentration compared to fasting levels. In a metabolomic study previously published in Diabetes, Framingham participants were found to have significant changes in 91 of 110 metabolites tested during an OGTT, including amino acids, Kreb’s cycle intermediates, bile acids, and more. Serotonin levels were suppressed 2-fold during the challenge, with significant drops in metabolites related to vitamins B1, B2, B3, B5, and B6. “Metabolic derailing” (alterations in adaptive processes) occurs in as little as 4 weeks in response to a high-fat and high-calorie diet even among healthy individuals, marked by changes in insulin, C-peptide, glucagon, leptin, cholesterol, etc., which drive low-grade inflammation and insulin resistance, and even plaque formation, changes more easily observed in response to challenge versus fasting levels. Thus, functional responses observed only in response to challenges (e.g., an OGTT vs. an FPG) are considerably complex and involve much more than changes in insulin secretion.
Given this background, the study published in AJCN contributes to our understanding of weight loss, and the interindividual variability in response to it. There is much to be gleaned from this study, but a quick overview and key highlights follow. Seventy-two normoglycemic volunteers with overweight/obesity were randomized to either a weight loss group (WLG) by energy restriction, or a control group designed to maintain weight (WMG). An average loss of 6.4% (5.6kg) occurred in the WLG, but a wide range of loss was reported (0.1 to 17.5kg), attributed primarily to dietary compliance. While fasting variables did not appear to change, subtle improvements in markers of phenotypic flexibility were observed in response to OGTT and MMTT (mixed-meal tolerance test), including plasma levels of glucose, insulin, and amino acids markers for insulin resistance (such as isoleucine, leucine, valine, etc.).
A metabolomic signature that was associated with greater metabolic improvements in response to weight loss was also observed; chief among these was 2-methyl-butyryl-carnitine (a branched-chain amino acid derivative), found at baseline concentrations of at least 30% higher levels among those who had the largest improvements in insulin sensitivity following weight loss. Similarly, deoxycholate (a secondary bile acid) was 100% higher during the challenge tests, and plasma urea levels were 15-20% among those with the greatest improvements following weight loss. Given that microbiota are known to be primarily responsible for generating secondary bile acids, this strongly suggests that the microbiome plays an important role in predicting who will improve the most in response to weight loss. Those participants who increased their fiber and reduced their saturated fat intake had the largest improvements in insulin sensitivity. Ideally, this study and others like it will help personalize weight loss interventions, identifying the key modifiable factors that most strongly predict benefit.