Prognostic factors for suicidal carbon monoxide poisoning in a regional hospital in Hong Kong

Abstract

Background Acute carbon monoxide (CO) poisoning is a common and unique way of attempting suicide in Hong Kong. Majority of cases are accomplished by charcoal burning at their homes or in rented hotel rooms. All patients suffering from CO poisoning are managed initially by emergency physicians and subsequently by intensivists or physicians when they are admitted to the hospitals. Days or weeks after the acute intoxication, some of these patients develop delayed neuropsychiatric sequelae (DNS), which include a wide spectrum of temporary to permanent disabilities, involving impairment in cognitive, neurological and psychiatric aspects. However, it is often unclear from the initial presentation, which kind of patients will develop DNS. Nevertheless, many prognostic factors have been identified to be associated with development of DNS. But, there is scarce information about these prognostic factors for such patients in Hong Kong.

Concerning the treatment for CO poisoning, there are a lot of debates in the past decades. Hyperbaric oxygen (HBO) therapy has been suggested to prevent the development of DNS by many clinicians. HBO therapy is the administration of 100% oxygen under increased atmospheric pressure, usually performed in a specialized designed chamber. HBO accelerates the elimination of CO in the body, and is thus postulated to be the effective treatment for CO poisoning. However, it has its own risks, like ear damage, and pneumothorax. Moreover, the results of different clinical trials concerning HBO therapy are contradicting. According to the best evidence available, the use of HBO therapy for patients with CO poisoning still remains controversial. Furthermore, there is no effective specific treatment for DNS, once it occurs after CO poisoning. Some authors have suggested that HBO may help to prevent the development of DNS. Yet, there is no concrete evidence to support this. Similarly, there are not many clinical trials concerning the impact of HBO therapy on DNS in patients with CO poisoning in Hong Kong.

Objectives The primary objective of this study is to identify possible prognostic factors associated with the development of DNS in patients with suicidal CO poisoning admitted to a regional hospital in Hong Kong. The secondary objective is to evaluate the impact of HBO therapy on the prevention of DNS in these patients, as well as its complications.

Methods This is a retrospective cohort study demonstrating the clinical features of patients with acute CO poisoning admitted and treated in a regional hospital in Hong Kong within a 10-year period. The clinical records of patients from all age and sex admitted to the regional hospital through the emergency department within that period with a coded diagnosis of CO poisoning were retrieved for analysis. This was performed with the help of a hospital-based clinical electronic database called Clinical Data Analysis and Reporting System, using the International Classification of Diseases 9 coding system. By means of the electronic patient record (ePR), all the demographic data, initial presenting features including level of consciousness, electrocardiogram (ECG), and initial sets of blood test results and all other essential clinical information were retrieved. The diagnosis of DNS in each patient was confirmed with the clinical notes traced from the ePR together with relevant radiological imaging reports.

Results From January 2004 to December 2013, 87 cases of suicidal CO poisoning were recruited for the analysis; 8 patients (9.2%) developed DNS while 23 patients (26.4%) were treated with HBO therapy. Patients treated with HBO therapy did not develop DNS. Univariate analysis showed that Glasgow coma scale (GCS) of 3 points, presence of loss of consciousness, elevated troponin, and elevated creatinine kinase were prognostic factors for DNS development. A descriptive summary of 7 cases with abnormal ECG changes on presentation showed that abnormal ECG changes might be another possible risk factor for DNS. From the multivariate analysis, the association of HBO treatment and DNS did not reach statistical significance, owing to a limited sample size.

Conclusion In patients with acute suicidal CO poisoning, the presence of LOC, low GCS, abnormal ECG changes, elevated troponin and creatinine kinase on presentation were possible predicting factors associated with the development of DNS. Although, our study did not demonstrate statistical significance of HBO in preventing DNS, it did show that HBO therapy has certain degree of protection against the development of DNS for the patients in our study. This demonstrated the clinical significance of HBO therapy in patients with CO poisoning. Further studies with larger sample size might be needed to demonstrate statistical significance.