Pneumomediastinum predicts early mortality in acute paraquat poisoning

Abstract

Context. In paraquat (PQ) poisoning, death often occurs after the appearance of pneumomediastinum (PM). However, the clinical features and eventual outcome of PM in PQ intoxication remains unclear. Objective. We aimed to characterize PM following PQ poisoning and its prognostic value for predicting mortality. Materials and methods. Enrolled PQ-poisoned patients (n = 75) were divided into two groups according to whether PM could be detected by chest computed tomography or not. The study outcomes included 5- and 90-day death after intoxication. Survival curves were derived using the Kaplan–Meier method, and mortality risk factors were analyzed by forward stepwise Cox regression analysis. Results. PM was documented in 21.3% of the patients (16/75); in 13 of them PM set in within 3 days of PQ ingestion. 15 patients died within 3 days of appearance of PM. Compared with patients without PM, those with PM were younger (P = 0.011), and had higher scores of Acute Physiology and Chronic Health Evaluation (P < 0.001) and Sequential Organ Failure Assessment (P = 0.003). In addition, patients with PM had a higher incidence of acute renal failure (P = 0.001), toxic hepatitis (P = 0.008), and respiratory insufficiency (P = 0.003). PM predicted an increased risk of 90-day death (93.8% of patients with PM vs. 40.7% among those without PM; hazard ratio [HR], 2.4; 95% confidence interval [CI], 1.0–5.6; P = 0.045), and increased risk of 5-day death (81.3% vs. 27.1%; HR, 3.2; 95% CI, 1.2–8.1; P = 0.017). Discussion and conclusion. Early PM, occurring within 8 days, is a specific predictor of mortality in PQ poisoning.

Keywords:

• Pneumomediastinum
• Paraquat
• Mortality
• Computed tomography

Funding

Supported by grants from the National Natural Science Foundation of China (81171792, 81270136, 81101414, and 81301626) and the National Global Experts Recruitment Program of China (WQ20133100165).

Acknowledgments

We thank Prof. Zi-Sheng Ai (Department of Medical Statistics, College of Medicine, Tongji University) for the assistance with statistical analysis.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.