Early consciousness disorder in acute ischemic stroke: incidence, risk factors and outcome

Stroke is the second most frequent cause of death and the most frequent cause of disability worldwide. In China, stroke has become the leading cause of death and disability in both urban and rural populations, with incidence increasing every year [1]. Approximately 80 % of all strokes are ischemic [1].

Consciousness, which refers to awareness of self and environment, depends on levels of arousal and wakefulness, as well as stimulus content [2]. Many patients in early stages of acute ischemic stroke show acute disorder of consciousness, a condition known as early consciousness disorder (ECD). In fact, stroke is one of the three most frequent causes of conscious disturbance in emergency rooms, together with trauma and hypoglycemia [3]. Data from various stroke registries suggest that 4–38 % of stroke patients experience decreased level of consciousness or coma, and 13–48 % experience confusion or delirium [4‐11].

Little is known about the incidence or risk factors of ECD in patients who experience acute ischemic stroke. This is an important question, because early consciousness disorder may disturb physician’ history-taking in clinical practice and ECD in the acute stroke period may predict stroke progression and increase risk of stroke-associated pneumonia [12‐14]. Studies have found several possible risk factors of acute disorder of consciousness after stroke: age, sex, previous stroke, atrial fibrillation, diabetes mellitus, alcohol consumption, stroke severity, infarct sites and, according to some reports, massive cerebral infarct and multiple brain infarcts [15‐19]. However, currently available information are rough data extracted from registries for different population. Direct comparative data between stroke patients with ECD and without ECD especially on prospective study are scarce. Further work is needed to identify a reliable set of ECD predictors.

Also unclear is whether ECD influences stroke-related complications and clinical outcomes in stroke patients. This is made more important by the fact that the presence of ECD in a stroke patient can affect decisions about treatment and management.

To address these questions, the present prospective study analyzed potential ECD risk factors and associated outcomes in a mid-sized cohort of Chinese patients with acute ischemic stroke.

Between 1 January 2009 and 31 December 2010, 1,506 patients with acute ischemic stroke were consecutively and prospectively registered in the Chengdu Stroke Registry. Of those, 569 (37.8 %) were admitted to our hospital within 24 h of stroke onset and were therefore included in the present study. Of this group, comprising 303 men and 266 women aged 64.75 ± 14.02 years, 199 (35 %) had ECD on admission, of whom 38 (6.7 %) were Coma. Among the 143 patients in our cohort with hyper-acute ischemic stroke who were admitted within 6 h of symptom onset, ECD occurred in 52 (36.4 %). Among the 81 patients admitted within 4.5 h of symptom onset, ECD occurred in 29 (35.8 %).

Although a sizeable proportion of stroke patients are admitted to the neurology ward with conscious disturbance, the incidence and risk factors of ECD are poorly understood, as are potential relationships between ECD and clinical outcomes in acute ischemic stroke. Here we provide evidence that ECD is common in Chinese patients with acute ischemic stroke, occurring in more than one of every three patients in our cohort that was admitted within 24 h of symptom onset. In this subset of patients who had ECD, nearly a third was in a state of decreased consciousness or coma on admission. This incidence is within the range of 4.3–37.7 % reported by other stroke registries [4‐9]. In contrast, the 5.3 % incidence of delirium in our study is lower than the range of 13–48 % reported for other ethnic groups [10, 11]. Differences in the incidence of consciousness disorder between our study and others may easily arise from heterogeneity in the time between stroke onset and hospital admission, which most studies in the literature did not strictly control. Such heterogeneity increases the risk that pre-hospital care, disease course and stroke-related complications contributed to observed conscious disturbance. Another simple explanation for differences between our study and others is that ours involved a single hospital, highlighting the fact that our findings should be validated in multi-site populations of Asian and non-Asian ethnicities.

Incidence of ECD was highest among the subgroup of patients with cardioembolism, and lowest among the subgroup with small-artery occlusion. These results are consistent with a study showing that altered consciousness was 3.2-fold more likely to occur in patients with cardioembolic infarction than in patients with atherothrombotic infarction [28]. In addition, the European Community Stroke Project reported that the incidence of coma and confusion within a week of onset of acute ischemic stroke was higher in the presence of atrial fibrillation than in its absence [17]. This may be because cardiac emboli are often large and therefore more likely to affect the main arteries of the cerebral vascular system, which then causes severe stroke [29].

In our cohort, age, NIHSS score on admission, massive cerebral infarction, high serum glucose on admission and history of alcohol consumption were independent risk factors of ECD. These results are consistent with a cross-sectional study of 9,044 Caucasian prehospital patients that showed that patients with abnormal GCS scores were more likely than patients with normal scores to be male and slightly older and to have a history of alcohol use, diabetes, substance abuse, stroke/transient ischemic attack and seizure [19]. Contrary to studies in French cohorts [18], our results do not support the idea that involvement of anterior circulation and multiple infarctions independently influence early conscious state in patients with acute ischemic stroke. In our cohort, CT/MRI imaging confirmed massive cerebral infarction in nearly two-thirds of patients with ECD, and 20 % of these patients suffered from hemorrhagic transformation during hospitalization. And for this reason, ECD should be considered an important sign in patients with acute ischemic stroke, and it should be taken into account when designing treatment and management strategies.

Our results suggest that poor glycemic control is a predictor of ECD in diabetic patients with acute ischemic stroke. History of diabetes mellitus was identified as an independent predictor of ECD in patients with acute ischemic stroke only when the multivariate logistic regression model did not include serum glucose concentration on admission. This result may reflect the fact that hyperglycemia increases coagulation and inhibits fibrinolysis, decreasing reperfusion of ischemic tissue and increasing infarct volumes. Hyperglycemia inhibits vasodilation, increasing the production of reactive oxygen species and nicotinamide adenine dinucleotide phosphate. Hyperglycemia may also inhibit mitochondrial function in the ischemic penumbra, causing substantial intracellular acidosis [30]. Previous work has associated hyperglycemia with a higher rate of hemorrhagic complications in patients with acute ischemic stroke [30]. Our results, together with previous work, suggest that close monitoring and control of serum glucose in diabetic patients with ischemic stroke may reduce ECD incidence and improve prognosis. However, as patients in our cohort were not given hemoglobin A1c (HbA1c) detection routinely, the interpretation of our results should be cautious because we couldn’t rule out the interference from stress hyperglycemia as a component of the acute phase response caused by severe stroke.

Interestingly, we found evidence that high systolic pressure on admission protects against ECD in patients with acute ischemic stroke. This may reflect auto-regulation of cerebral blood flow during the acute stage of infarction. Indeed, an earlier prospective study from our group showed that a decrease in blood pressure during the first 24 h after admission independently predicts one-month death in Chinese patients with acute stroke [31]. A prospective study in Spanish patients with acute ischemic stroke found that both higher and lower blood pressure during the first 24 h after stroke onset are associated with early neurological deterioration and poor prognosis [32]. Together, these studies highlight the importance of closely monitoring and controlling blood pressure during the first 24 h after stroke. In particular, they highlight the need to prevent a reduction in blood pressure in early stages of infarction. A study of 22 older American patients with delirium using 99mTc HMPAO SPECT scanning associated this conscious state with reduced blood flow in the left inferior frontal, right temporal, right occipital, and pontine regions [33]. It may be that a decrease in systemic blood pressure reduces cerebral perfusion pressure and blood flow to the ischemic penumbra, particularly in the absence of normal auto-regulation [34]. Preventing a reduction in blood pressure soon after infarction is consistent with official guidelines for early management of acute ischemic stroke [35].

Post-stroke complications, both neurological and medical, account for 23–50 % of deaths related to acute ischemic stroke. In particular, brain edema is a major cause of death during the first week after stroke [15]. Medical and neurological complications can hinder rehabilitation, prolong hospitalization, increase healthcare costs and contribute to poor functional outcomes or death [36]. The present study associates ECD with significantly elevated risk of stroke-related complications. The most frequent neurological complication in our patients with ECD was malignant edema, followed by hemorrhagic transformation and post-stroke epilepsy. Regarding post-stroke seizures or epilepsy, the overall frequency is similar to previous reports [15]. As seizures or epilepsy is not the primary issue of the present study, patients in our cohort were not given electroencephalography routinely and we didn’t record detailed information about acute symptomatic seizures which could result in early consciousness disorder. Pulmonary infection was the most common medical complication. Our results are consistent with work in US stroke patients associating greater neurological deficit with higher incidence of medical complications; indeed, that study found neurological impairment to be the strongest predictor of medical complications [37]. Since NIHSS score on admission was significantly higher among our patients with ECD than among patients without it, we speculate that more severe stroke at least partly explains the higher frequency of stroke-related complications in stroke patients with ECD.

Many complications are preventable or treatable if recognized in time. Therefore clinicians should pay close attention to patients with acute ischemic stroke and ECD in order to detect stroke-related complications as soon as possible, and thereby improve stroke outcome. Our results, and those of previous studies, highlight the need for large trials addressing the prevention and treatment of complications and their effect on stroke outcome in this stroke patient subpopulation. Indeed, the present study is one of the few to directly compare characteristics and prognosis of acute ischemic stroke patients with and without ECD. Our results suggest that the two populations differ in several clinically important respects, justifying large, well-designed studies of early management of stroke patients with ECD.

In our cohort, ECD independently predicted death during hospitalization and 3-month follow-up according to a multivariate regression model that did not take into account NIHSS score on admission or stroke-related complications. When these two parameters were included as inputs in the model, ECD was no longer an independent predictor of these outcomes, although it remained an independent predictor of 3-month death/disability. These different results echo discrepancies between previous studies showing associations of ECD with post-stroke death and poor functional outcome [38] or no such associations [39]. While these discrepancies may reflect differences in hospital populations and enrolment criteria, they are consistent with the idea that neurological deficit and stroke-related complications confound the association of ECD with death. The present study, to our knowledge, provides the first evidence that ECD elevates (3.3-fold) the risk of 3-month death/disability in patients with acute ischemic stroke, even after adjusting for confounding factors. This result is consistent with a study in French patients with acute stroke that associated acute confusional state with worse 6-month functional outcome [40].

The results of the present study should be interpreted with caution given its limitations. First, the cohort came from a single hospital, so the results may not be representative of other populations. Second, we conducted follow-up at a single time point of only 3 months, and follow-up involved a telephone interview or mailed questionnaire instead of a clinical visit, which may increase the risk of reporting bias.

The present study provides clear evidence that at least in Chinese populations, ECD may occur in as many as one-third of patients with acute ischemic stroke, particularly those with cardio-embolism stroke. Our analysis identifies age, stroke severity, and massive cerebral infarction as potential risk factors of ECD in ischemic stroke patients. Although ECD appears to affect anterior circulation much more often than posterior circulation, involvement of anterior circulation does not appear to independently predict ECD in patients with acute ischemic stroke. ECD in such patients is associated with malignant edema, hemorrhagic transformation and epilepsy, as well as with death and disability within 3 months after stroke onset.