Noninvasive End Tidal CO2 Is Unhelpful in the Prediction of Complications in Deliberate Drug Poisoning

We study the performance of capnometry in the detection of early complications after deliberate drug poisoning.

This was a prospective cohort study of self-poisoned adult patients who presented at an emergency department (ED) between April 20, 2012, and May 6, 2014. Patients who ingested at least 1 neurologic or respiratory depressant drug were included. The primary outcome was the predictive value of an end tidal CO2 (etco2) measurement greater than or equal to 50 mm Hg for the detection of early complications defined a priori by hypoxia requiring oxygen greater than or equal to 3 L/min, bradypnea less than or equal to 10 breaths/min, or ICU admission after intubation or antidote administration because of unresponsiveness to pain or respiratory arrest. Consciousness scales and clinical data were recorded at admission and every 30 minutes. Noninvasive etco2 was continuously measured for 2 hours after inclusion unless the patient was admitted to the ICU. Patients and physicians were blinded to etco2 values.

Two hundred one patients were included, 35 of whom exhibited at least 1 complication. An etco2 measurement greater than or equal to 50 mm Hg predicted the onset of a complication, with a sensitivity of 46% (95% confidence interval [CI] 29% to 63%) and a specificity of 80% (95% CI 73% to 86%), leading to a positive predictive value of 33% (95% CI 20% to 48%) and a negative predictive value of 88% (95% CI 81% to 92%). etco2 was less able to predict complications than the Glasgow Coma Scale score at inclusion.

Capnometry in isolation does not provide adequate prediction of early complications in self-poisoned patients referred to the ED. A dynamic minute-by-minute assessment of etco2 could be more predictive.