Metabolic Scarring: The Persistent Impact of Past Obesity on Long‐Term Metabolic Health Despite Weight Loss

TL;DR

Past obesity leaves lasting metabolic effects even after weight loss, and a new risk score can help identify those at higher risk.

Contribution

A novel clinical risk score integrating weight history to better predict metabolic risk beyond current BMI.

Findings

Formerly obese individuals have significantly higher HbA1c than never-obese peers.

The metabolic scarring risk score achieved an AUC of 0.79 for predicting elevated glycaemic risk.

Weight history contributes up to 4 points in the risk score, highlighting its importance over current BMI alone.

Abstract

Conventional cardiometabolic risk assessment relies primarily on a patient's current body mass index, yet individuals who have lost weight after a period of obesity may continue to harbour elevated metabolic risk. We sought to quantify the persistent impact of past obesity on glycaemic control and to develop a clinical risk score that integrates weight history with current risk factors.

We performed a cross‐sectional analysis of 15,422 adults (≥ 18 years) from the 2011–2020 NHANES cycles. Participants with complete self‐reported weight history (highest adult weight, weight 1 year ago, number of ≥ 5% weight‐loss episodes) and measured BMI were included. Metabolic scarring was defined by elevated haemoglobin A1c (HbA1c ≥ 5.7%) or HOMA‐IR ≥ 2.5. We applied inverse‐probability‐weighted logistic regression to estimate the association between prior obesity and current HbA1c, adjusting for confounders. We then refit a survey‐weighted logistic model using age per decade, current BMI, weight‐history category, sex and race/ethnicity, converting regression coefficients into an integer point‐based score. Discrimination was evaluated by survey‐weighted area under the receiver‐operating characteristic curve (AUC).

Formerly obese individuals exhibited significantly higher HbA1c than never‐obese peers (adjusted β = 0.58%, p < 0.002), indicative of metabolic scarring. The derived risk score ranged from −31 to +90 points (median = 6; IQR = −3 to 16) and achieved an AUC of 0.79 (95% CI 0.77–0.81). Age per decade, BMI, and weight history contributed 4, 1 and up to 4 points, respectively; female sex and Non‐Hispanic White race subtracted points. Calibration across predicted‐risk deciles was excellent (slope = 0.98).

A history of obesity imparts a lasting glycemic risk that is not captured by current BMI alone. Our metabolic scarring risk score offers a pragmatic tool for identifying individuals at elevated metabolic risk despite weight normalisation.

Past obesity leaves a lasting metabolic footprint. Using NHANES data, we show that formerly obese individuals exhibit elevated HbA1c despite weight normalisation. We introduce a metabolic‐scarring risk score that integrates weight history to improve glycaemic risk prediction beyond current BMI.

Figures