Impact of Preoperative Anemia on Short-Term Outcomes After Colorectal Cancer Surgery: How the Choice of Diagnostic Criteria Leads to Conflicting Findings

Le résumé fourni par la source

Background: Whether preoperative anemia is independently associated with short-term postoperative outcomes in colorectal cancer (CRC) remains controversial. This study investigated the association between preoperative anemia and 30-day postoperative complications (PO30C) and evaluated how different anemia diagnostic criteria influence this association. Methods: In this retrospective cohort study, patients who underwent radical CRC surgery from January 2018 to January 2021 were consecutively enrolled. Firth's regression was used for univariate and multivariable analysis to evaluate the impact of preoperative anemia on short-term outcomes after CRC surgery. Dose-response analysis assessed the relationship between anemia severity and PO30C. Sensitivity analysis compared the association under WHO and China's anemia criteria, and misclassification analysis characterized criteria-related differences. Population-based examination of the misclassification mechanism was performed using China Health and Retirement Longitudinal Study (CHARLS) data. Internal validation used 10-fold cross-validation and bootstrap resampling. Results: Among 231 patients, 108 (46.8%) had preoperative anemia and 29 (12.6%) developed PO30C. Multivariable analysis showed that preoperative anemia (OR = 2.39, 95% CI: 1.04-5.81, P = 0.039) and male sex (OR = 3.11, 95% CI: 1.23-9.24, P = 0.015) were independently associated with PO30C. A significant dose-response relationship was observed across anemia severity groups (P-trend = 0.005), with moderate-to-severe anemia independently associated with PO30C (OR = 2.88, 95% CI: 1.04-8.03, P = 0.042). The association was significant under China criteria (P = 0.039) but not under WHO criteria (P = 0.075). Misclassification analysis revealed that WHO criteria reclassified 33 borderline patients as anemic, whose PO30C rate (9.1%) resembled the normal group (6.7%) rather than the anemia group (18.5%), thereby attenuating the association. CHARLS data corroborated this misclassification mechanism. The model showed moderate discrimination (optimism-corrected C-index = 0.704). Conclusion: Preoperative anemia was independently associated with PO30C under the China criteria, whereas WHO criteria attenuated this association through misclassification of low-risk borderline patients. Population-appropriate diagnostic criteria are important for perioperative risk estimation and may explain conflicting findings in the existing literature.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.