OBJECTIVE

To investigate the incidence and independent risk factors of lower extremity deep vein thrombosis (DVT) in patients with acute carbon monoxide poisoning and to evaluate the predictive performance of a risk factor model to support early identification and individualized intervention for high-risk patients.

METHODS

This retrospective cohort study included 180 patients diagnosed with acute carbon monoxide poisoning from January 2021 to December 2023. Lower extremity venous ultrasound was performed within 48 hours of admission to detect DVT, and patients were categorized into DVT and non-DVT groups. Clinical data - including demographic characteristics, poisoning-related variables, and biochemical markers - were collected and analyzed. Variables with statistical significance were subjected to logistic regression analysis to identify independent risk factors for DVT. Spearman correlation analysis and receiver operating characteristic (ROC) curve analysis were further conducted to assess variable relationships and the predictive performance of the risk model.

RESULTS

Among the 180 patients, 23 (12.78%) developed DVT. Spearman correlation analysis showed that coma duration, carboxyhemoglobin concentration, C-reactive protein, procalcitonin, D-dimer, creatinine, blood urea nitrogen, lactate dehydrogenase, myoglobin, and creatine kinase were positively correlated with DVT (all P < 0.001), whereas earlier initiation of hyperbaric oxygen therapy and higher albumin levels were negatively correlated (r = -0.397, P < 0.001). Logistic regression identified coma duration, D-dimer level, and carboxyhemoglobin concentration as independent risk factors for DVT. The ROC curve demonstrated good predictive performance, with an area under the curve of 0.888 (95% CI: 0.827-0.948, P < 0.05).

CONCLUSION

Lower extremity DVT is relatively common in ptients with acute carbon monoxide poisoning. Coma duration, D-dimer levels, and delayed initiation of hyperbaric oxygen therapy are significantly associated with increased risk. The proposed risk factor model demonstrates strong predictive value and may assist in early clinical detection and targeted prevention strategies.