Original Contribution

Predictors of early clinical deterioration after acute ischemic stroke^?

Leng C. Lin MD ^a,e, J.T. Yang MD, PHD ^b, H.H. Weng MD, PhD ^c,f,

C.T. Hsiao MD ^a,e,?, Shiao L. Lai MD ^d, W.C. Fann MD ^a

^aDepartment of Emergency Medicine, Chang Gung Memorial Hospital, Chiayi 61363,

Chang Gung University College of Medicine, Taiwan

^bDivision of Neurosurgery, Department of Surgery, Chang Gung Memorial Hospital, Chiayi 61363,

Chang Gung University College of Medicine, Taiwan

^cDepartment of Diagnostic Radiology, Chang Gung Memorial Hospital, Chiayi 61363,

Chang Gung University College of Medicine, Taiwan

^dDepartment of Neurology, Chang Gung Memorial Hospital, Chiayi 61363, Chang Gung University College of Medicine, Taiwan ^eDepartment of Nursing, Chang Gung Institute of Technology, Chiayi 61363, Taiwan

^fDepartment of Respiratory Care, Chang Gung Institute of Technology, Chiayi 61363, Taiwan

Received 30 September 2009; revised 14 December 2009; accepted 15 December 2009

Abstract The measurements for predicting early deterioration of stroke patients is controversial. We studied Laboratory measurements and previously identified risk factors to identify factors or predictors of early deterioration after stroke. A prospective observational study of 196 patients with first-time acute ischemic stroke was performed. Demographic data, patient histories, laboratory measurements, and initial Stroke severity assessments were recorded. Patients with early deterioration in National Institutes of Health Stroke Scale scores (increase >=3 points within 3 days) were defined as having stroke-in- evolution (SIE). Thirty patients were diagnosed with SIE. An initial National Institutes of Health Stroke Scale score of 12 or higher, a Glasgow Coma Scale score of 12 or lower, D-dimers more than 1000, or blood urea nitrogen/creatinine (BUN/Cr) ratio higher than 15 were more frequent in SIE patients. After multivariate analysis, only a BUN/Cr higher than 15 was independent predictor of SIE. These patients were 3.41-fold more likely to have SIE (P = .008). These findings suggest that BUN/Cr may be a novel predictor of SIE, potentially useful in emergency departments.

(C) 2011

Introduction

^? This work was supported by Grants CMRPG 660272 and CMRPG 660282 from the Chang Gung Medical Research Council and Grant NMRPG 3191 from the National Science Council of Taiwan.

* Corresponding author. Department of Nursing, Chang Gung Memo-

rial Hospital, Chiayi County 613, Taiwan, ROC. Tel.: +886 05 3623511;

fax: +886 05 3623002.

E-mail address: [email protected] (C.T. Hsiao).

Ischemic stroke is one of the major causes of death worldwide. Among stroke patients, 20% to 40% will experience early deterioration in condition after hospital admission [1]. Timely intervention can dramatically improve outcome and reduce disability [2]. Treatment in a specialized stroke unit has been shown to result in better outcomes than a conventional stroke treatment, particularly with regard to the

0735-6757/$ - see front matter (C) 2011 doi:10.1016/j.ajem.2009.12.019

diagnostic studies“>risk of early Neurologic deterioration [2]. Further to this, early admission to a stroke unit has been shown to positively influence clinical outcome [3].

In previous reports, many predictors of early deterioration after ischemic stroke have been proposed and studied, yet the results remain controversial. For example, several studies have been performed to ascertain whether cerebral or systemic causes are the major determinants of stroke deterioration. The findings, however, have been inconsistent. Measures of overall status and stroke severity (low initial Canadian Stroke Severity score [4], low initial National Institutes of Health Stroke Scale [NIHSS] score [3], early computed tomography [CT] findings of stroke severity [4-7], and changes in cerebral blood flow affecting the ischemic penumbra as measured by positron emission tomography and single-photon emission CT imaging [8,9]) have been shown to predict early deterioration. Laboratory tests for Coagulation markers (fibrinogen [4], D- dimers [10,11]), inflammatory markers (increased interleukin [IL]-6 [4], decreased IL-10 [12]), Serum glucose at admission [3-5], hematocrit [7], and physiologic parameters, such as blood pressure (both elevated and decreased) [3,11,13,14], and body temperature [4], have been identified as predictors of early deterioration. In addition, patient medical history (diabetes mellitus [6], atherosclerosis [7], chronic heart disease [6]) has also been associated with early stroke deterioration. These differing findings may be due to variations in study design, patient populations, selection and exclusion criteria, and the criteria used for defining early deterioration.

A further complication in trying to determine predictors of early deterioration after stroke is that different stroke subtypes may manifest differently in terms of clinical changes before deterioration. Tei et al [7] showed that the probability of early deterioration after stroke was much more frequent for certain stroke subtypes than for others. In most of the aforementioned studies, no attempt was made to differentiate among stroke subtypes, although others have shown that some proposed markers are statistically significant only for certain stroke subtypes [7,12]. A predictor of early deterioration that is equally applicable to all stroke subtypes is needed.

Given this, we performed a prospective observational study in an effort to identify risk factors for early deterioration after stroke (regardless of stroke subtype) in a population of patients. In doing this, our aim was determine predictors that could be readily assessed in an emergency department (ED) setting and potentially help physicians identify high-risk stroke patients who may benefit from more intensive care and observation.

Methods

Patients

Consecutive patients admitted to the Chia-Yi Chang Gung Memorial Hospital (a stroke referral hospital) from

July 2007 to June 2008 with acute ischemic stroke were prospectively recruited to participate in this study. Patients transferred from other hospitals were eligible for enrollment in the study. Patients were excluded if they had a previous acute ischemic stroke, the time between the onset of neurologic symptoms and ED presentation was more than 12 hours, evidence of hemorrhagic stroke assessed by CT, or required fibrino- lytic therapy. The study protocol was approved by the institutional review board of Chang Gung Memorial Hospital. Written informed consent was obtained from all patients approached.

Diagnostic studies

The following data were collected from the participants in the ED on a standardized data collection form: age; sex; symptoms and signs; Glasgow Coma Scale (GCS); history of hypertension, diabetes mellitus, hyperlipidemia, coronary heart disease, or peripheral vascular disease; smoking habits; stroke localization; electrocardiogram; arterial blood pressure on ED admission and every

8 hours for the first 3 days; admission blood glucose; complete blood count; C-reactive protein; D-dimer; Blood urea nitrogen ; creatinine (Cr); BUN/Cr; Liver enzymes aspartate aminotransferase and alanine amino- transferase; triglyceride; total cholesterol; low-density lipoprotein; albumin; and glycosylated hemoglobin. All patients received a Brain CT scan within 6 hours of ED admission. The CT findings were interpreted by a radiologist from the stroke team.

Each patient’s neurologic condition was assessed by a neurologist from the stroke team who was blinded to all other patient data. Stroke severity was assessed using the NIHSS score. Evaluations were performed by physicians trained in NIHSS assessment immediately after admission, within the first 3 days during hospitalization, and before discharge. Patients for whom the NIHSS score returned to 0 within the initial 24 were classified as having a transient ischemic attack and were excluded from the study.

Patients with early deterioration in condition were defined as having stroke-in-evolution (SIE). Stroke-in- evolution was diagnosed for those patients who experi- enced worsening neurologic condition as indicated by an increase of 3 or more points on the NIHSS score within 3 days. Theoretically, any increase in an NIHSS score can be indicative of SIE. However, in some cases, the symptoms are borderline, with the score increasing by only 1 or 2 points. Hence, we chose to use an increase of 3 or more in the present study to diagnose SIE because such an increase is more definitive in terms of indicating worsening patient condition.

The study end point was patient outcome on day 3 as measured by the NIHSS. Data were collected by research assistants and the authors. The patients were divided into 2 groups for data analysis. The first group included the patients

with “stroke-in-evolution” and the second group included those without “stroke-in-evolution.” Demographic charac- teristics, clinical history, clinical manifestations, GCS, and NIHSS were compared between patients diagnosed as having SIE and those who were not. Analysis was performed to identify potential risk factors for SIE. We chose to examine BUN/Cr more than 15 as a potential risk factor because such a ratio is commonly considered to indicate prerenal azotemia and dehydration.

Statistical analysis

Continuous data are expressed as mean +- SD. Categor- ical data are expressed as frequencies and percentage. Demographic characteristics were compared using Student t test or ?2 test. Age, sex, and variables found to be associated with SIE by univariate logistic regression analysis were entered into multivariate logistic regression models. The variables entered into univariate analysis were selected on the basis of previous findings suggesting that these variables may correlate with SIE. Multivariate logistic regression analysis with stepwise selection was used to control for the potential confounding variables of age and sex to determine the factors related to SIE. The results of the univariate and multivariate regression analysis are expressed as odds ratios (ORs) with 95% confidence intervals (CIs). Data were analyzed using SAS 9.0 statistical software (SAS Institute Inc, Cary, NC). A P value less than .05 was considered statistically significant.

Results

Thirty (15.3%) of 196 patients enrolled were diagnosed with SIE. The mean (SD) age of the 196 participants (70 women and 126 men) enrolled was 70.6 (10.4) years, with a range of 44 to 99 years. Table 1 summarizes the demographic and clinical characteristics of the 30 SIE and

166 non-SIE patients. Only the BUN/Cr ratio was statistically significantly related to SIE (P = .006). Patients having D-dimer levels higher 1000 ug/L were much more common in the SIE group, however not statistically significantly so (P = .060).

Initial stroke severity upon admission was measured by assessment of NIHSS and GCS scores. Statistical analysis showed that greater initial stroke severity was significantly more likely in patients with SIE (Table 2). There was a significant difference between the groups regarding NIHSS score (P = .030). The percentage of patients with a prespecified NIHSS score of 12 or more was higher in the SIE group (41.4%) compared with the non-SIE group (22.4%). There was also a significant difference between the groups in GCS scores (P = .012). A higher percentage of SIE patients (33.3%) had a GCS score of 12 or lower than did non-SIE patients (16.3%; P = .028).

Variables (missing no)	Non-SIE (n = 166)	SIE (n = 30)	P
Age a (y) (1)	70.3 +- 10.3	72.2 +- 11.0	.373
Male b	106 (63.9)	20 (66.7)	.767
D-Dimer N1000 b	57 (35.2)	16 (53.3)	.060
ECG-af b	40 (24.1)	8 (26.7)	.763
Systolic blood	167.3 +- 36.9	171.2 +- 31.5	.584
pressure a (3)
Diastolic blood	90.7 +- 17.7	97.3 +- 25.6	.188
pressure a (3)
Blood sugar a (3)	165.5 +- 87.5	174.8 +- 85.9	.594
AST a (66)	28.9 +- 25.4	30.9 +- 33.1	.749
ALT a (5)	27.1 +- 17.7	29.6 +- 20.1	.495
BUN a (50)	16.1 +- 10.5	13.4 +- 9.3	.644
Cr a	1.1 +- 0.8	1.1 +- 0.6	.969
BUN/Cr N15 b (50)	37 (31.1)	16 (59.3)	.006 ?
TG a (41)	119.6 +- 61.5	123.7 +- 91.0	.825
Total Chol a (31)	179.2 +- 40.8	185.9 +- 34.9	.440
Uric acid a (56)	5.5 +- 1.8	5.8 +- 2.0	.386
Albumin a (1)	4.4 +- 3.2	4.0 +- 0.5	.118
WBC a	7.8 +- 2.8	7.8 +- 2.6	.903
Hgb a	13.6 +- 1.7	12.9 +- 2.0	.990
PLT a (1)	206.7 +- 65.2	208.8 +- 79.8	.878
SEG a (2)	0.6 +- 0.1	0.7 +- 0.1	.795
WBC ? (SEG + Band) a (4)	5.3 +- 2.5	5.3 +- 2.4	.876
LDL a (38)	112.5 +- 31.6	121.4 +- 30.6	.202
HbAlc a (95)	0.07 +- 0.02	0.08 +- 0.03	.559
OCSP a (40)	1.9 +- 1.3	2.2 +- 1.4	.286
ECG-af indicates electrocardiogram finding of atrial fibrillation (); AST, aspartate aminotransferase; ALT, alanine aminotransferase; TG, thyro- globulin; Chol. cholesterol; WBC, white blood count; Hgb, hemoglobin; PLT, platelets; SEG, segmented neutrophils; WBC ? (SEG + Band), the white blood Cell differential from the complete blood count; LDL, low- density lipoprotein; HbAlc, glycosylated hemoglobin; OCSP, Oxford- shire Community Stroke Project clinical stroke syndrome classification. a Continuous data expressed as mean +- SD. b Categorical data expressed as number (%). * Statistically significant at P b .05.

Univariate logistic regression analysis was used to test for the factors potentially associated with SIE including age, sex, D-dimer levels, BUN/Cr ratio, NIHSS, and GCS scores (Table 3). Patients with a BUN/Cr ratio higher than

15 were 3.22 times as likely to develop SIE (95% CI, 1.36-7.62; P = .008). Those with admission GCS scores or 12 or lower were 2.57 times as likely to develop SIE (95% CI, 1.09-6.11; P = .032). In addition, patients with an NIHSS admission score of 12 or higher were 2.44 times as likely to develop SIE (95% CI, 1.07-5.57; P = .034). Patients with a D-dimer levels higher than 1000 ug/L were

Table 1 Demographic and clinical characteristics of the 196 study patients

2.11 times as likely to have SIE; however, this finding was not statistically significant (P = .064)

Multivariate logistic regression analysis of the factors identified by univariate analysis revealed that a BUN/Cr higher than 15 was an independent factor predicting SIE (Table 4). Patients with a BUN/Cr ratio higher than 15 were

Variables	Non-SIE (n = 166)	SIE (n = 30)	P
NIHSS	7.3 +- 6.6	9.6 +- 7.2	.093
b12	128 (77.6)	17 (58.6)	.030 ?
>=12	37 (22.4)	12 (41.4)
GCS	14.1 +- 1.9	13.1 +- 2.7	.012 ?
N12	139 (83.7)	20 (66.7)	.028 ?
<=12	27 (16.3)	10 (33.3)
Continuous data expressed as mean +- SD, and categorical data expressed as %. * Statistically significant at P b .05.

3.41 times more likely to develop SIE (95% CI, 1.38-8.46;

Table 2 NIHSS and GCS recorded on ED admission of the 196 study patients

P = .008).

Discussion

In many patients with ischemic stroke, neurologic symptoms progress rapidly during the initial hours [4]. Because early clinical deterioration results in increased mortality and functional disability, the goal of this study was to identify factors useful for predicting early deterioration of stroke patients in the ED.

Although we simultaneously tested many factors identi- fied in other studies, none passed all of our statistical tests under the conditions of our study. Based on our criteria, stroke severity scales (NIHSS and GCS) were significantly associated with early deterioration but were not found to be independent predictors of stroke. Patients with a D-dimer

Table 3 Univariate logistic regression analysis for possible factors associated with SIE

Variables	OR ?	95% CI	P
Age (per year)	1.02	0.98-1.06	.371
Sex
Female	Reference	-
Male	1.32	0.50-2.58	.768
d-Dimer
<=1	Reference	-
N1	2.11	0.96-4.62	.064
BUN/Cr
<=15	Reference	-
N15	3.22	1.36-7.62	.008 ?
GCS
N12	Reference	-
<=12	2.57	1.09-6.11	.032 ?
NIHSS
b12	Reference	-
>=12	2.44	1.07-5.57	.034 ?
Tested with univariate logistic regression analysis and presented with OR and 95% CI. * Statistically significant at P b .05.

level higher than 1000 ug/L were 2.11 times as likely to develop SIE; however, this finding was not statistically significant. Thus, although our results agree with past studies in terms of a general association between initial stroke severity and coagulation factors among patients with early deterioration after stroke, these factors were not found to be sufficiently predictive in our study design. Rather, we found that the BUN/Cr ratio at admission was an independent predictive factor for early deterioration after stroke, a finding not previously reported to our knowledge.

The basic mechanisms underlying early clinical worsen- ing in patients with ischemic stroke are not well understood. Although several biochemical factors have been associated with early neurologic deterioration, most of them have a low predictive value [5,15,16]. The role of cytokines [17] and IL- 6 [4] in the pathophysiology of acute brain ischemia, release of excitatory amino acids [18] and D-dimers [10], as well as oxygen free radicals, iron accumulation, and nitric oxide production have been all been suggested as mechanisms for Clinical worsening [19].

Table 4 Multivariate logistic regression analysis for possible factors associated with SIE

Variables	OR ?	95% CI	P
Age (per year)	1.02	0.98-1.07	.399
Sex
Female	Reference	-
Male	1.37	0.52-3.58	.520
BUN/Cr
<=15	Reference	-
N15	3.41	1.38-8.46	.008 ?
NIHSS
b12	Reference	-
>=12	1.29	0.50-3.37	.601
Tested with multivariate logistic regression analysis and presented with OR and 95% CI. * Statistically significant at P b .05.

Although the BUN/Cr ratio has not previously been identified as a predictor of early deterioration after stroke, it is well documented to prediCT worsening condition in patients with chronic heart failure [20,21]. Cardiovascular disease and stroke are often comorbid. In stroke-prone spontaneously hypertensive rats, dietary salt loading led to increased blood pressure, increased BUN, and an increased incidence of both cerebral infarction and Renal damage [22]. Thus, BUN levels may be broadly indicative of underlying cardiovascular-renal-cerebrovascular susceptibilities that predispose a patient to a more rapid decline after stroke. The association between elevated BUN/Cr ratio and SIE in our patients suggests that renal insufficiency may lead to neurologic worsening in stroke patients.

Another interpretation of the BUN/Cr ratio is that it indicates relative dehydration of the patients. Churchill et al

[23] observed that stroke patients with stable renal function and high BUN levels were often medicated with diuretics,

whereas Kelly et al [24] reported that dehydration in the days immediately after acute ischemic stroke was associated with an increased BUN/Cr ratio and a predisposition for venous thromboembolism. This indicates that stroke patients with a higher BUN/Cr ratio may be at risk of early deterioration specifically due to an increased risk of venous thromboem- bolism. Hence, an immediate intervention for such patients should be the maintenance of proper hydration [24].

We did not test the previously identified inflammation markers such as IL-6 and IL-10 in the present study. Interleukin-6 has been found to strongly correlate with other measures such as coagulation markers (fibrinogen) and early signs of stroke severity (infarct volume) [4], making such measurement somewhat redundant. In addition, IL-10 has been shown to only be significantly associated with early deterioration in patients with subcortical infarcts or lacunar stroke [12], excluding it as a universal measure to be used for the undiagnosed stroke patient in the ED.

Our results are limited by sample size and thus statistical power. A further larger scale study is needed to verify the findings described herein and determine whether the BUN/ Cr ratio is a predictor of early deterioration for different stroke subtypes. According to our calculations, if the sample size was increased to 800 and 12% of patients had SIE, the power would be 80% (assuming a 5% level of significance). A longitudinal study of changes in the day-to-day BUN/Cr ratio during hospitalization may clarify the relationship between the BUN/Cr ratio and the stage of stroke evolution. We also acknowledge that our choice of NIHSS and GCS measure cutoff scores may have influenced the findings [25]. In conclusion, our findings suggest that a BUN/Cr ratio higher than 15 may be associated with SIE. Monitoring of hydration status would also seem to be critical in these patients. Being able to accurately identify patients at risk for early deterioration after stroke will allow for the design of clinical trials of stroke intervention targeting patients

with SIE.

Acknowledgments

The authors acknowledge Tsong-Hai Lee for valuable discussions.

References

1. Davalos A, Castillo J. Progressing stroke. In: Fisher M, Bogousslavsky J, editors. Current review of cerebrovascular disease. Philadelphia: Current Medicine Inc; 1999. p. 149-60.
2. Silvestrelli G, Parnetti L, Paciaroni M, et al. Early admission to stroke unit influences clinical outcome. Eur J Neurol 2006;13:250-5.
3. Roquer J, Rodriguez-Campello A, Gomis M, et al. Acute stroke unit care and early Neurological deterioration in ischemic stroke. J Neurol 2008;255:1012-7.
4. Vila N, Castillo J, Davalos A, et al. Proinflammatory cytokines and early neurological worsening in ischemic stroke. Stroke 2000;31: 2325-9.
5. Toni D, Fiorelli M, Gentile M, et al. Progressing neurological deficit secondary to acute ischemic stroke. A study on predictability, pathogenesis, and prognosis. Arch Neurol 1995;52: 670-5.
6. Davalos A, Toni D, Iweins F, et al. Neurological deterioration in acute ischemic stroke: potential predictors and associated factors in the European cooperative acute stroke study (ECASS) I. Stroke 1999;30: 2631-6.
7. Tei H, Uchiyama S, Ohara K, et al. Deteriorating ischemic stroke in 4 Clinical categories classified by the Oxfordshire Community Stroke Project. Stroke 2000;31:2049-54.
8. Fisher M, Garcia JH. Evolving stroke and the ischemic penumbra. Neurology 1996;47:884-8.
9. Mayer SA, Lignelli A, Fink ME, et al. Perilesional blood flow and edema formation in acute intracerebral hemorrhage: a SPECT study. Stroke 1998;29:1791-8.
10. Barber M, Langhorne P, Rumley A, et al. Hemostatic function and progressing ischemic stroke: D-dimer predicts early clinical progres- sion. Stroke 2004;35:1421-5.
11. Barber M, Langhorne P, Rumley A, et al. D-Dimer predicts early clinical progression in ischemic stroke: confirmation using routine clinical assays. Stroke 2006;37:1113-5.
12. Vila N, Castillo J, Davalos A, et al. Levels of anti-inflammatory cytokines and neurological worsening in acute ischemic stroke. Stroke 2003;34:671-5.
13. Gonzalez-Martinez F, Navarro-Gutierrez S, de Leon-Belmar J. Initial emergency department blood pressure as predictor of survival after acute ischemic stroke. Neurology 2006;66:1609.
14. Stead LG, Gilmore RM, Decker WW, et al. Initial emergency department blood pressure as predictor of survival after acute ischemic stroke. Neurology 2005;65:1179-83.
15. Davalos A, Cendra E, Teruel J, et al. Deteriorating ischemic stroke: risk factors and prognosis. Neurology 1990;40:1865-9.
16. Davalos A, Castillo J, Pumar JM, et al. Body temperature and fibrinogen are related to early neurological deterioration in acute ischemic stroke. Cerebrovasc Dis 1997;7:64-9.
17. Clark WM. Cytokines and reperfusion injury. Neurology 1997;49: S10-4.
18. Castillo J, Davalos A, Noya M. Progression of ischaemic stroke and excitotoxic aminoacids. Lancet 1997;349:79-83.
19. Castillo J. Deteriorating stroke: diagnostic criteria, predictors, mechanisms and treatment. Cerebrovasc Dis 1999;9(Suppl 3): 1-8.
20. Yancy CW, Lopatin M, Stevenson LW, et al. Clinical presenta- tion, management, and in-hospital outcomes of patients admitted with acute decompensated heart failure with preserved systolic function: a report from the Acute Decompensated Heart Failure National Registry (ADHERE) Database. J Am Coll Cardiol 2006;47:76-84.
21. Cauthen CA, Lipinski MJ, Abbate A, et al. Relation of blood urea nitrogen to long-term mortality in patients with heart failure. Am J Cardiol 2008;101:1643-7.
22. Okada M, Kobayashi M, Maruyama H, et al. Effects of a selective endothelin A-receptor antagonist, BQ-123, in salt-loaded stroke-prone spontaneously hypertensive rats. Clin Exp Pharmacol Physiol 1995;22:763-8.
23. Churchill M, Grimm S, Reding M. Risks of diuretic usage following stroke. Neurorehabil Neural Repair 2004;18:161-5.
24. Kelly J, Hunt BJ, Lewis RR, et al. Dehydration and venous thromboembolism after acute stroke. QJM 2004;97:293-6.
25. DeGraba TJ, Hallenbeck JM, Pettigrew KD, et al. Progression in acute stroke: value of the initial NIH stroke scale score on patient stratification in future trials. Stroke 1999;30:1208-12.