Prognostic factors in unintentional mild carbon monoxide poisoning

Objectives: To identify early predictors of recovery from mild carbon monoxide poisoning, and to search for qualitative interactions between subsets of patients and treatment effects. Design and setting: Inception cohort study from a 4-year, randomised, controlled trial. which compared normobaric oxygen therapy to the combination of normobaric plus hyperbaric oxygen therapy in 307 patients. Study was conducted at a hyperbaric oxygen therapy referral centre. Patients and interventions: Victims of unintentional, non-fire-related, domestic, carbon monoxide intoxication were evaluated if the time elapsed from end of exposure to hospital admission was less than 12 h, the carboxyhaemoglobin level was greater than or equal to 10 % or 5 % for smoker or non-smoker, respectively, and if they did not suffer a loss of consciousness. One course of hyperbaric oxygen therapy or 6-h normobaric oxygen therapy was given. Measurements and results: The main outcome measure was status at the I-month recovery. Of the 307 evaluated patients, 206 (67 %, 95 % CI: 62 % -72 %) had recovered at 1 month. None of the patients died or suffered severe sequelae. Apathy and headaches were the main moderate sequelae. The multivariate analysis selected dizziness before admission (OR = 1.92, [1.17-3.15], p = 0.010) and headaches upon hospital admission (OR = 2.14, [1.09-4.171, p = 0.026) as jointly associated with persistent neurological symptoms. No significant crossover interaction between each combination of these two predictors and treatment effects was observed. Conclusions: At 1 month after mild carbon monoxide intoxication, victims who presented with dizziness before hospital admission or headaches upon admission have an increased risk of minor persistent neurological symptoms, but almost all patients could resume their former occupation.