Incidence and Clinical Course of Left Ventricular Systolic Dysfunction in Patients with Carbon Monoxide Poisoning

Abstract

Background and objectives: Carbon monoxide (CO) poisoning can cause tissue hypoxia and left ventricular systolic dysfunction (LVSD) requiring intensive medical management. Our objectives were to find incidence and clinical course of LVSD CO intoxicated patients and make a clinical scoring to predict LVSD.

Subjects and methods: We included all consecutive patients with CO exposure in the emergency room. LVSD was defined by LVEF <50% assessed by echocardiography. We compared their clinical, chemical, radiological and electrocardiographic patterns according to the presence of LVSD.

Results: From May 2009 to June 2015, we included a total of 81 patients (48 men, 47±19 years old) with CO exposure in this cohort. LVSD was found in about 25 patients (31%). Nine had regional wall motion abnormality. Follow up echocardiographic examinations were available in 21 patients. Of them, 18 patients showed complete recovery in about 3 days (mean 2.8±1.7 days). Of 3 patients without recovery, 2 had significant coronary artery stenosis. LVSD was significantly associated with initial heart rate (>100/min), pulmonary edema on chest X-ray, serum NT pro-BNP (>100 pg/mL), troponin-I (>0.1 ng/mL) and lactic acid (>4.0 mg/dL) after a univariate analysis. Combining these into a clinical score, according to their beta score after a multivariate analysis (rage=0-16), allowed prediction of LVSD with a sensitivity of 84% and specificity of 91% (reference ≥8, area under the curve=0.952, p<0.001).

Conclusion: About 31% showed LVSD in patients with CO poisoning, and most of them (86%, 18 of 21 patients) recovered within 3 days. Patients with a higher clinical score (≥8) might have LVSD.

Keywords: Carbon monoxide; Left ventricular dysfunction; Poisoning.

Fig. 1. A diagram showing the outcomes of the study population. LVSD: left ventricular systolic dysfunction, RWMA: regional wall motion abnormality.

Fig. 2. Receiver operating curve analysis in the detection of left ventricular systolic dysfunction by a clinical score. Clinical score ≥8 showed optimal sensitivity (84%) and specificity (91%). CI: confidence interval.