Association between four indirect indicators of insulin resistance and lung cancer: evidence from UK biobank and Mendelian randomization analysis
Menée à l'aide d'une méthode de randomisation mendélienne et de données de la "UK Biobank" portant sur 388 013 personnes (durée médiane de suivi : 13,2 ans), cette étude analyse l’association entre la résistance à l'insuline évaluée par 4 indicateurs (index triglycérides-glucose (TyG), TyG combiné à l’indice de masse corporelle (TyG-IMC), rapport triglycérides/cholestérol HDL et score métabolique d’insulino-résistance) et le risque de cancer du poumon (3 527 cas)
Background: Insulin resistance (IR) has been increasingly linked to cancer development and progression, but its association with lung cancer (LC) remains unclear. This prospective study investigated the association between four indirect IR indicators—triglyceride-glucose (TyG) index, TyG combined with body mass index (TyG-BMI), triglyceride to high-density lipoprotein cholesterol ratio (TG/HDL-C), and metabolic score for IR (METS-IR), and LC risk.
Methods: A total of 388,013 participants without previous cancer at baseline from the UK Biobank were included. Cox hazards regression models were used to examine the associations between four indirect indicators of IR and LC. Restricted cubic splines (RCS) evaluated nonlinear associations. The discriminatory ability of each index was assessed using the area under the receiver operating characteristic (ROC) curve. Sensitivity analyses were performed to test robustness. In addition, Mendelian randomization (MR) analysis was performed to explore potential causal relationships.
Results: Among 388,013 cancer-free participants from UK Biobank, 3,527 incident LC cases were documented during a median follow-up of 13.25 years. Kaplan-Meier analysis revealed associations between higher levels of four indicators and increased cumulative LC incidence. In fully adjusted Cox regression models, METS-IR exhibited the strongest association with overall LC, with the highest quartile associated with an increased risk (hazard ratio: 1.24, 95% confidence interval: 1.05–1.47). No statistically significant associations were observed for lung adenocarcinoma or lung squamous cell carcinoma. In contrast, several indicators remained associated with small cell lung cancer (SCLC), with elevated risks observed for TG/HDL-C (1.52, 1.03–2.25), METS-IR (2.30, 1.29–4.11), and TyG-BMI (2.22, 1.20–4.08) in the highest quartiles. RCS analyses suggested predominantly linear associations. ROC curve analyses indicated modest discriminatory performance, with relatively higher discrimination for SCLC, particularly for the TG/HDL-C ratio and TyG index. However, MR analyses did not provide evidence supporting a causal association.
Conclusions: This prospective cohort study suggested that indirect indicators of insulin resistance are associated with the risk of overall lung cancer, particularly small cell lung cancer, whereas no significant associations were observed for lung adenocarcinoma or lung squamous cell carcinoma. However, MR analyses did not support a causal relationship. These findings suggest that these indicators may reflect underlying metabolic vulnerability rather than direct causal mechanisms.
BMC Cancer , article en libre accès, 2026