Causes of coma and their evolution in the medical intensive care unit

Epidemiological data regarding the frequency of the different causes of coma and their evolution over time in the ICU are scarce [1–4]. In the 1970s, Plum and Posner [2] described the causes of 500 successive cases of coma. However, since then no specific study focusing on the causes of coma and their evolution over time has been published although the ICU management of brain injuries has progressed tremendously. Some data are available from prospective and retrospective studies assessing coma outcome [3–8], but the different causes of coma were excluded in these studies. These studies, therefore, do not provide a global description of the causes of coma and their evolution. The aim of this study was to describe the causes of coma and their evolution over an 8-year time period in a medical ICU. From January 2001 to December 2008, we retrospectively identified patients who presented with coma in the first 24 h of ICU admission, defined as a Glasgow Coma Scale below 8 in the absence of sedative drugs, and we retrieved their demographic data (supplementary Table). For statistical purposes, four time periods of 2 years each were defined. A total of 2,189 out of 4,482 patients were evaluated. Baseline characteristics of the patients can be found in the supplementary Table. Detailed causes of coma, initial Glasgow Coma Scale and ICU survival are shown Table 1. The frequency of anoxo-ischemic encephalopathy and shock of any origin increased as the causes of coma over the 8-year time period is shown in Table 2. At the same time, the frequency of the other causes of coma remained stable. Among the 2,189 comatose patients, 1,139 (52%) survived to ICU discharge (Table 1). Survival rates according to each cause of coma are given in the Table 1. To our knowledge, this is the largest existing cohort that focused on the causes of coma and their evolution. The study population was comparable in age, sex ratio, and the Glasgow Coma Scale score to previously reported populations of comatose patients as in the Study to Understand Prognoses and Preferences for Outcomes and Risks of Treatments (SUPPORT) cohort. SUPPORT, the largest previously available cohort of 596 non-traumatic comatose patients, studied factors such as high risk of death and severe disability [1–8]. However, the SUPPORT cohort excluded drug intoxication and metabolic causes of coma [4]. Plum and Posner [2] found that the most common causes of coma were drug poisons (30%), cerebral hemorrhage including intracerebral, epidural, and subdural hemorrhages (15%), and brainstem infarction (8%). Cardiac arrest represented only 2% in the Plum and Posner cohort. This discrepancy could be somewhat explained by the fact that the Plum and Posner’s study was performed in a neurological ICU, whereas our study was performed in a general medical ICU. Moreover, our study was subject to regular updates of the guidelines for cardiac arrest resuscitation which has improved immediate survival over the last 30 years. This hypothesis is corroborated by the findings of the other previous reports [4–8]. The reason for the Electronic supplementary material The online version of this article (doi:10.1007/s00415-011-6388-z) contains supplementary material, which is available to authorized users.