Application of a Machine Learning Technology in the Definition of Metabolically Healthy and Unhealthy Status: A Retrospective Study of 2567 Subjects Suffering from Obesity with or without Metabolic Syndrome

Phenotyping obesity: A focus on metabolically healthy obesity and metabolically unhealthy normal weight

Over the past 4 decades, research has shown that having a normal body weight does not automatically imply preserved metabolic health and a considerable number of lean individuals harbour metabolic abnormalities typically associated with obesity. Conversely, excess adiposity does not always equate with an abnormal metabolic profile. In fact, evidence exists for the presence of a metabolically unhealthy normal weight (MUHNW) and a metabolically healthy obese (MHO) phenotype. It has become increasingly recognised that different fat depots exert different effects on the metabolic profile of each individual by virtue of their location, structure and function, giving rise to these different body composition phenotypes. Furthermore, other factors have been implicated in the aetiopathogenesis of the body composition phenotypes, including genetics, ethnicity, age and lifestyle/behavioural factors. Even though to date both MHO and MUHNW have been widely investigated and documented in the literature, studies report different outcomes on long-term cardiometabolic morbidity and mortality. Future large-scale, observational and population-based studies are required for better profiling of these phenotypes as well as to further elucidate the pathophysiological role of the adipocyte in the onset of metabolic disorders to allow for better risk stratification and a personalised treatment paradigm.

Differences in the levels of inflammatory markers between metabolically healthy obese and other obesity phenotypes in adults: A systematic review and meta-analysis

AIMS

The aim of this study was to systematically review and analyze differences in the levels of C-reactive protein (CRP), interleukin-6 (IL-6), and tumor necrosis factor-alpha (TNF-α) comparing metabolically healthy but obese (MHO) with metabolically healthy non-obese (MHNO), metabolically unhealthy non-obese (MUNO), and metabolically unhealthy obese (MUO) subjects.

DATA SYNTHESIS

We searched PubMed, Embase, Web of Science, and Scopus for studies that matched the relevant search terms. Differences in inflammatory marker levels between MHO and the other three phenotypes were pooled as standardized mean differences (SMD) or differences of medians (DM) using a random-effects model. We included 91 studies reporting data on 435,007 individuals. The CRP levels were higher in MHO than in MHNO subjects (SMD = 0.63, 95% CI: 0.49, 0.76; DM = 0.83 mg/L, 95% CI: 0.56, 1.11). The CRP levels were higher in MHO than in MUNO subjects (SMD = 0.16, 95% CI: 0.05, 0.28; DM = 0.39 mg/L, 95% CI: 0.09, 0.69). The CRP levels were lower in MHO than in MUO individuals (SMD = -0.43, 95% CI: -0.54, -0.31; DM = -0.82 mg/L, 95% CI: -1.16, -0.48). The IL-6 levels in MHO were higher than in MHNO while lower than in MUO subjects. The TNF-α levels in MHO were higher than in MHNO individuals.

CONCLUSIONS

This review provides evidence that CRP levels in MHO are higher than in MHNO and MUNO subjects but lower than in MUO individuals. Additionally, IL-6 levels in MHO are higher than in MHNO but lower than in MUO subjects, and TNF-α levels in MHO are higher than in MHNO individuals.

SYSTEMATIC REVIEW REGISTRATION

PROSPERO number: CRD42021234948.

Uncovering Predictors of Lipid Goal Attainment in Type 2 Diabetes Outpatients Using Logic Learning Machine: Insights from the AMD Annals and AMD Artificial Intelligence Study Group

Identifying and treating lipid abnormalities is crucial for preventing cardiovascular disease in diabetic patients, yet only two-thirds of patients reach recommended cholesterol levels. Elucidating the factors associated with lipid goal attainment represents an unmet clinical need. To address this knowledge gap, we conducted a real-world analysis of the lipid profiles of 11.252 patients from the Annals of the Italian Association of Medical Diabetologists (AMD) database from 2005 to 2019. We used a Logic Learning Machine (LLM) to extract and classify the most relevant variables predicting the achievement of a low-density lipoprotein cholesterol (LDL-C) value lower than 100 mg/dL (2.60 mmol/L) within two years of the start of lipid-lowering therapy. Our analysis showed that 61.4% of the patients achieved the treatment goal. The LLM model demonstrated good predictive performance, with a precision of 0.78, accuracy of 0.69, recall of 0.70, F1 Score of 0.74, and ROC-AUC of 0.79. The most significant predictors of achieving the treatment goal were LDL-C values at the start of lipid-lowering therapy and their reduction after six months. Other predictors of a greater likelihood of reaching the target included high-density lipoprotein cholesterol, albuminuria, and body mass index at baseline, as well as younger age, male sex, more follow-up visits, no therapy discontinuation, higher Q-score, lower blood glucose and HbA1c levels, and the use of anti-hypertensive medication. At baseline, for each LDL-C range analysed, the LLM model also provided the minimum reduction that needs to be achieved by the next six-month visit to increase the likelihood of reaching the therapeutic goal within two years. These findings could serve as a useful tool to inform therapeutic decisions and to encourage further in-depth analysis and testing.

Transparent machine learning suggests a key driver in the decision to start insulin therapy in individuals with type 2 diabetes

AIMS

The objective of this study is to establish a predictive model using transparent machine learning (ML) to identify any drivers that characterize therapeutic inertia.

METHODS

Data in the form of both descriptive and dynamic variables collected from electronic records of 1.5 million patients seen at clinics within the Italian Association of Medical Diabetologists between 2005-2019 were analyzed using logic learning machine (LLM), a "clear box" ML technique. Data were subjected to a first stage of modeling to allow ML to automatically select the most relevant factors related to inertia, and then four further modeling steps individuated key variables that discriminated the presence or absence of inertia.

RESULTS

The LLM model revealed a key role for average glycated hemoglobin (HbA1c) threshold values correlated with the presence or absence of insulin therapeutic inertia with an accuracy of 0.79. The model indicated that a patient's dynamic rather than static glycemic profile has a greater effect on therapeutic inertia. Specifically, the difference in HbA1c between two consecutive visits, what we call the HbA1c gap, plays a crucial role. Namely, insulin therapeutic inertia is correlated with an HbA1c gap of <6.6 mmol/mol (0.6%), but not with an HbA1c gap of >11 mmol/mol (1.0%).

CONCLUSIONS

The results reveal, for the first time, the interrelationship between a patient's glycemic trend defined by sequential HbA1c measurements and timely or delayed initiation of insulin therapy. The results further demonstrate that LLM can provide insight in support of evidence-based medicine using real world data.

Obesity-Associated Hepatic Steatosis, Somatotropic Axis Impairment, and Ferritin Levels Are Strong Predictors of COVID-19 Severity

The full spectrum of SARS-CoV-2-infected patients has not yet been defined. This study aimed to evaluate which parameters derived from CT, inflammatory, and hormonal markers could explain the clinical variability of COVID-19. We performed a retrospective study including SARS-CoV-2-infected patients hospitalized from March 2020 to May 2021 at the Umberto I Polyclinic of Rome. Patients were divided into four groups according to the degree of respiratory failure. Routine laboratory examinations, BMI, liver steatosis indices, liver CT attenuation, ferritin, and IGF-1 serum levels were assessed and correlated with severity. Analysis of variance between groups showed that patients with worse prognoses had higher BMI and ferritin levels, but lower liver density, albumin, GH, and IGF-1. ROC analysis confirmed the prognostic accuracy of IGF-1 in discriminating between patients who experienced death/severe respiratory failure and those who did not (AUC 0.688, CI: 0.587 to 0.789, p < 0.001). A multivariate analysis considering the degrees of severity of the disease as the dependent variable and ferritin, liver density, and the standard deviation score of IGF-1 as regressors showed that all three parameters were significant predictors. Ferritin, IGF-1, and liver steatosis account for the increased risk of poor prognosis in COVID-19 patients with obesity.

Predictors of weight loss in patients with obesity treated with a Very Low-Calorie Ketogenic Diet

Introduction

The Very Low-Calorie Ketogenic Diet (VLCKD) has emerged as a safe and effective intervention for the management of metabolic disease. Studies examining weight loss predictors are scarce and none has investigated such factors upon VLCKD treatment. Among the molecules involved in energy homeostasis and, more specifically, in metabolic changes induced by ketogenic diets, Fibroblast Growth Factor 21 (FGF21) is a hepatokine with physiology that is still unclear.

Methods

We evaluated the impact of a VLCKD on weight loss and metabolic parameters and assessed weight loss predictors, including FGF21. VLCKD is a severely restricted diet (<800 Kcal/die), characterized by a very low carbohydrate intake (<50 g/day), 1.2-1.5 g protein/kg of ideal body weight and 15-30 g of fat/day. We treated 34 patients with obesity with a VLCKD for 45 days. Anthropometric parameters, body composition, and blood and urine chemistry were measured before and after treatment.

Results

We found a significant improvement in body weight and composition and most metabolic parameters. Circulating FGF21 decreased significantly after the VLCKD [194.0 (137.6-284.6) to 167.8 (90.9-281.5) p < 0.001] and greater weight loss was predicted by lower baseline FGF21 (Beta = -0.410; p = 0.012), male sex (Beta = 0.472; p = 0.011), and central obesity (Beta = 0.481; p = 0.005).

Discussion

VLCKD is a safe and effective treatment for obesity and obesity related metabolic derangements. Men with central obesity and lower circulating FGF21 may benefit more than others in terms of weight loss obtained following this diet. Further studies investigating whether this is specific to this diet or to any caloric restriction are warranted.

Machine learning insights concerning inflammatory and liver-related risk comorbidities in non-communicable and viral diseases

The liver is a key organ involved in a wide range of functions, whose damage can lead to chronic liver disease (CLD). CLD accounts for more than two million deaths worldwide, becoming a social and economic burden for most countries. Among the different factors that can cause CLD, alcohol abuse, viruses, drug treatments, and unhealthy dietary patterns top the list. These conditions prompt and perpetuate an inflammatory environment and oxidative stress imbalance that favor the development of hepatic fibrogenesis. High stages of fibrosis can eventually lead to cirrhosis or hepatocellular carcinoma (HCC). Despite the advances achieved in this field, new approaches are needed for the prevention, diagnosis, treatment, and prognosis of CLD. In this context, the scientific com-munity is using machine learning (ML) algorithms to integrate and process vast amounts of data with unprecedented performance. ML techniques allow the integration of anthropometric, genetic, clinical, biochemical, dietary, lifestyle and omics data, giving new insights to tackle CLD and bringing personalized medicine a step closer. This review summarizes the investigations where ML techniques have been applied to study new approaches that could be used in inflammatory-related, hepatitis viruses-induced, and coronavirus disease 2019-induced liver damage and enlighten the factors involved in CLD development.