Exploring the causal relationship between hemoglobin and pancreatic cancer and its potential mechanisms through bioinformatics and Mendelian randomization.

BACKGROUND

Abnormal hemoglobin (HGB) levels and the onset of malignant tumors have attracted substantial clinical interest. PAAD, a highly fatal malignancy of the digestive system, warrants further investigation regarding its potential link with HGB levels. To explore the genetic relationship between the two, we employed Mendelian randomization in conjunction with transcriptomic analysis to probe their underlying connection.

METHODS

A combined approach utilizing Mendelian randomization (MR) and transcriptomics was adopted to examine the genetic association between HGB levels and PAAD, along with possible mechanistic pathways. Based on GWAS datasets derived from European populations, MR analysis was conducted to evaluate the causal relationship between HGB levels and the risk of PAAD. To test the reliability of the results, heterogeneity and directional pleiotropy were evaluated using the MR-Egger intercept test, Cochran's Q test, and leave-one-out analysis. Transcriptomic datasets from TCGA and GEO were then integrated to identify differentially expressed genes, followed by functional enrichment analysis. LASSO regression was subsequently applied to select characteristic genes and construct a prognostic model, which was then subjected to validation.

RESULTS

MR analysis revealed a negative association between HGB levels and the development of PAAD. Genetically, elevated HGB levels were linked to a reduced risk of PAAD (β_IVW = - 0.40, OR_IVW = 0.66, 95% CI = 0.48-0.92, p = 0.013). Using the PAAD dataset, seven key genes (DNMT3A, TFCP2L1, PPARGC1A, GSTA5, BICC1, NRG4, BCL2L13) were identified through LASSO regression, and HGB scores were computed based on their expression. Kaplan-Meier survival curve analysis indicated that patients with high scores exhibited significantly poorer overall survival (OS) than those in the low-score group (p < 0.0001). The scoring model demonstrated high predictive accuracy for 1-, 3-, and 5-year OS, with AUC values of 0.77, 0.79, and 0.91, respectively. Multivariate Cox regression and prognostic modeling of the seven genes showed that, apart from NRG4, the remaining six were independent risk factors associated with unfavorable prognosis in PAAD (all p < 0.05). The model yielded a C-index of 0.72, reflecting strong predictive power. Column-line plots further confirmed the model's effective performance for predicting 1-, 3-, and 5-year OS. Validation with the GSE85916 and TCGA-PAAD dataset demonstrated consistent robustness of the model in forecasting OS in PAAD patients, reinforcing its reliability and potential applicability.

CONCLUSIONS

This study identified a genetic causal relationship between HGB levels and the risk of PAAD. Through transcriptomic analysis, we constructed a prognostic model based on HGB-associated key genes. The model displayed reliable predictive capacity and offers new perspectives for clinical strategies aimed at preventing PAAD.