ABSTRACT
The present study investigated the predictive values of urine paraquat (PQ) concentration, dose of poison, arterial blood lactate and Acute Physiology and Chronic Health Evaluation (APACHE) II score in the prognosis of patients with acute PQ poisoning. A total of 194 patients with acute PQ poisoning, hospitalized between April 2012 and January 2014 at the First Affiliated Hospital of P.R. China Medical University (Shenyang, China), were selected and divided into survival and mortality groups. Logistic regression analysis, receiver operator characteristic (ROC) curve analysis and Kaplan-Meier curve were applied to evaluate the values of urine paraquat (PQ) concentration, dose of poison, arterial blood lactate and (APACHE) II score for predicting the prognosis of patients with acute PQ poisoning. Initial urine PQ concentration (C0), dose of poison, arterial blood lactate and APACHE II score of patients in the mortality group were significantly higher compared with the survival group (all P<0.05). Logistic regression analysis revealed that C0, dose of poison and arterial blood lactate correlated with mortality risk of acute PQ poisoning (all P<0.05). ROC curve analysis suggested that the areas under the curve (AUC) values of C0, dose of poison, arterial blood lactate and APACHE II score in predicting the mortality of patients within 28 days were 0.921, 0.887, 0.808 and 0.648, respectively. The AUC of C0 for predicting early and delayed mortality were 0.890 and 0.764, respectively. The AUC values of urine paraquat concentration the day after poisoning (Csec) and the rebound rate of urine paraquat concentration in predicting the mortality of patients within 28 days were 0.919 and 0.805, respectively. The 28-day survival rate of patients with C0 ≤32.2 µg/ml (42/71; 59.2%) was significantly higher when compared with patients with C0 >32.2 µg/ml (38/123; 30.9%). These results suggest that the initial urine PQ concentration may be the optimal index for predicting the prognosis of patients with acute PQ poisoning. Additionally, dose of poison, arterial blood lactate, Csec and rebound rate also have referential significance.
ABSTRACT
Objective To evaluate the prognostic value of urine paraquat (PQ) concentrations combined with poisoning time and creatinine clearance rate (CCr) on prognosis of patients with acute paraquat poisoning (APP). Methods A retrospective case control study was conducted. Clinical data of 96 patients with APP admitted to Department of Emergency of Shengjing Hospital of China Medical University from March 2014 to May 2016 were analyzed. The gender, age, body weight, urine PQ concentrations (determined by semi-quantitative colorimetric method), poisoning time (time from oral poison to urine detection) and CCr of patients were collected, and poisoning index (poisoning index = urine PQ concentrations × poisoning time/CCr) and simplified poisoning index (simplified poisoning index = urine PQ concentrations × poisoning time) were calculated. The patients were divided into death group and survival group according to 2-month outcome after poisoned with clinical data and telephone follow-up. The urine PQ concentrations, poisoning index, and simplified poisoning index between the two groups were compared. Binary classification logistic regression was used to analyze the risk factors affecting prognosis. Receiver-operating characteristic curve (ROC) and diagnostic test were used to analyze the prognostic value of the parameters. Results Compared with survival group, the urine PQ concentrations [mg/L: 30.00 (10.00, 100.00) vs. 10.00 (3.00, 10.00)], poisoning index [mg·h-1·μmol-1: 12.72 (1.86, 33.75) vs. 0.56 (0.18, 1.12)], and simplified poisoning index [mg·h-1·L-1: 600.00 (150.00, 1 000.00) vs. 60.00 (18.00, 120.00)] in death group were significantly increased (all P < 0.01). It was shown by logistic regression analysis that both urine PQ concentrations [odds ratio (OR) = 1.046, 95% confidence interval (95%CI) = 1.006-1.087, P = 0.022] and poisoning index (OR = 1.353, 95%CI = 0.029-1.815, P = 0.031) were independent risk factors affecting the prognosis of patients with APP. It was shown by ROC curve and diagnostic test that the poisoning index had greater area under ROC curve (AUC was 0.902) for evaluating the prognosis of patients with APP. When the best cut-off value was greater than 1.23 mg·h-1·μmol-1, the sensitivity was 90.91%, and the specificity was 73.08%. The AUC of urine PQ concentrations for evaluating the prognosis was 0.759. When the best cut-off value was greater than 20.00 mg/L, the sensitivity was 63.64%, and the specificity was 76.92%. The AUC of simplified poisoning index for evaluating the prognosis was 0.846. When the best cut-off value was greater than 135.00 mg·h-1·L-1, the sensitivity was 81.82%, and the specificity was 76.92%. Conclusion The poisoning index calculated with urine PQ concentrations combined with poisoning time and CCr has prognostic value for prognosis of APP patients, and the prognostic value of poisoning index is greater than that of the urine PQ concentrations alone.
