Journal of the Neurological Sciences 352 (2015) 68–73
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Serum levels of procalcitonin and high sensitivity C-reactive protein are associated with long-term mortality in acute ischemic stroke You-Mei Li a, Xue-Yuan Liu b,⁎ a b
Department of Neurology, Yangpu Hospital, Tongji University School of Medicine, Shanghai 20090, China Department of Neurology, Shanghai Tenth People's Hospital of Tongji University, Shanghai 200072, China
a r t i c l e
i n f o
Article history: Received 10 December 2014 Received in revised form 17 March 2015 Accepted 18 March 2015 Available online 27 March 2015 Keywords: Procalcitonin High sensitivity C-reactive protein Acute ischemic stroke Mortality Chinese
a b s t r a c t Objective: The aim of this study is to assess the prognostic value of systemic inflammation, as measured by the inflammatory biomarkers PCT and Hs-CRP, to predict the long-term mortality in ischemic stroke patients. Methods: We prospectively studied 374 patients with ischemic stroke who were admitted within 24 h after the onset of symptoms. Serum levels of PCT, Hs-CRP and NIH stroke scale (NIHSS) were measured at the time of admission. Clinical follow-up was performed at 1 year. The prognostic value of PCT to predict the mortality within one year was compared with Hs-CRP, NIHSS and with other known outcome predictors. Results: In the 64 non-survival patients, serum PCT levels were significantly (P b 0.0001) higher compared with those in survival patients. Multivariate COX regression analysis showed that log-transformed PCT and Hs-CRP were independent mortality predictors with adjusted hazard ratio of 4.24 (95% confidence interval [CI], 2.42–6.30) and 15.37 (95% confidence interval [CI], 3.25–41.08). The area under the receiver operating characteristic curve of PCT and Hs-CRP were 0.89 (95% CI, 0.85–0.93) and 0.68 (95% CI, 0.59–0.77) for mortality, respectively. Conclusion: Serum levels of PCT and HS-CRP at admission were independent predictor of long-term mortality after ischemic stroke in a Chinese sample. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Ischemic stroke is the third leading cause of mortality in most countries in the world and China has 2.5 million new stroke cases each year and 7.5 million stroke survivors [1]. Mortality after 1 year ranges between 21% and 27%; 15% to 30% of survivors are permanently disabled [2]. An early risk assessment with estimate of the severity of disease and prognosis is pivotal for optimized care and allocation of healthcare resources to improve outcome [3]. Inflammatory processes have fundamental roles in stroke in both the etiology of ischemic cerebrovascular disease and the pathophysiology of cerebral ischemia [4]. Inflammatory markers (fibrinogen and CRP) predicted the stroke severity and outcome [5]. It has been reported that it is possible to use the increase in the concentration of acute phase reactants and especially the high sensitivity C-reactive protein (Hs-CRP) to help predict future cerebrovascular mortality [6]. Procalcitonin (PCT) was a protein of 116 amino-acids with a molecular weight of 13 kDa. The probable site of PCT production during inflammation is the neuroendocrine cells in the lungs or intestine [7]. Mimoz et al. [8] found that an early and transient release of PCT into the circulation was observed after severe ⁎ Corresponding author at: No. 301, RanchangZhong Road, Shanghai 200072, China. Tel./fax: +86 21 66300588. E-mail address: [email protected] (X.-Y. Liu).
http://dx.doi.org/10.1016/j.jns.2015.03.032 0022-510X/© 2015 Elsevier B.V. All rights reserved.
trauma and the amount of circulating PCT seemed proportional to the severity of tissue injury and hypovolemia, yet unrelated to infection, indicating an inflammation-related induction of PCT. Data from large-sample studies in China about the relationship between inflammatory biomarkers and mortality in stroke patients are rare, and evidences in statistically strong power are needed. Thus, the primary aim of our prospective cohort study was to assess the prognostic value of systemic inflammation, as measured by the inflammatory biomarkers PCT and Hs-CRP, to predict the long-term mortality in ischemic stroke patients. 2. Method 2.1. Patients and study design We conducted a prospective cohort study at the Shanghai Tenth People's Hospital. From September 2010 to October 2013, all patients with an acute ischemic stroke event were included. Patients were eligible for inclusion if they were admitted to the emergency department with an acute ischemic stroke defined according to the World Health Organization criteria [9] and with symptom onset within 24 h. we excluded patients with intracranial hemorrhage, a history of recent surgery or trauma during the preceding 2 months, renal insufficiency (creatinine N1.5 mg/dl), malignancy, febrile disorders, acute or chronic inflammatory
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
disease at study enrollment (which were defined by physical examination, past medical history and white blood count test), autoimmune diseases, severe edema or a prior myocardial infarction onset b 3 months, as well as those with a history of valvular heart disease, or intracardiac thrombus on echocardiograph. In addition, patients with no evidence of infarction on CT or MRI within 24 h after symptom onset were excluded from the study. The control cases (N = 200) were of similar age and gender distribution to the AIS patients. They had no known diseases and were not using any medication. The median age of controls included in this study was 69 (IQR, 62–79) years and 45.0% were women. A detailed medical history was taken and clinical and laboratory examinations were performed on all participants in both groups. The study was approved by the local ethics committee of Shanghai Tenth People's Hospital of Tongji University. The patients or their relatives gave written informed consent prior to entering the study. 2.2. Clinical variables The following clinical and demographical data were taken: age, gender, stroke etiology, blood pressure, leukocyte count, and presence of risk factors (ie, age; sex; smoking history; hypercholesterolemia; history of hypertension, diabetes mellitus, previous ischemic stroke, or transient ischemic attack, respectively; positive family history for myocardial infarction, stroke, or transient ischemic attack). Stroke cause was determined according to the criteria of the TOAST (Trial of Org 10172 in Acute Stroke Treatment) classification [10], which distinguishes largeartery arteriosclerosis cardio embolism, small-artery occlusion, other causative factor, and undetermined causative factor. The clinical stroke syndrome was determined by applying the criteria of the Oxfordshire Community Stroke Project: total anterior circulation syndrome (TACS); partial anterior circulation syndrome (PACS); lacunar syndrome (LACS); and posterior circulation syndrome (POCS) [11]. The National Institutes of Health Stroke Scale (NIHSS) score was assessed on admission (with greater scores indicating increasing severity) [12]. 2.3. Neuroimaging Diagnosis of stroke was based on the results of strict neuroradiological examination (brain computer tomography, magnetic resonance imaging (MRI), or both) according to the International Classification of Diseases, ninth revision. MRI with diffusion-weighted imaging (DWI) was available in 221 stroke patients (59.1%). In those patients, DWI lesion volumes were determined by an experienced neurologist (Liu XY) who was unaware of the clinical and laboratory results. The infarct volume was calculated by using the formula 0.5 × a × b × c (where a is the maximal longitudinal diameter, b is the maximal transverse diameter perpendicular to a, and c is the number of 10-mm slices containing infarct) [13]. 2.4. End points and follow-up Our study finished 1-year follow-up. The end point of this study was death from any cause within a 1-year follow-up. The death was considered as an outcome variable. 1-year mortality was defined as long-term mortality. Follow-up was performed by two trained medical students with a structured follow-up telephone interview with the patient or, if not possible, with the closest relative or family physician. 2.5. Blood collection and quantification For the purpose of this study, blood samples of patients who were admitted to hospital were prospectively drawn from the antecubital vein. After centrifugation, serum of the samples were immediately stored at −80 °C before assay. White blood cell (WBC) count, Hs-CRP, PCT and other biochemical measurements were done using standard
69
laboratory methods. PCT levels were measured using Enzyme-Linked Fluorescent Assay by VIDAS B.R.A.H.M.S. (Biomerieux, Durham, USA), and Hs-CRP was analyzed by the Roche Cobas Integra 800 analyzer (Roche Diagnostic, Indianapolis, IN, USA). The lower detection limit was 0 ng/ml and the line range was 0–200 ng/ml. The intra-assay coefficient of variation [CV] and inter-assay CV were 3.54–7.07% and 4.15%–9.86%, respectively. 2.6. Statistical analysis Discrete variables are summarized as counts (percentage), and continuous variables as medians and interquartile ranges (IQRs). Twogroup comparison of not normally distributed data was performed using Mann–Whitney U test, and a Kruskal–Wallis one-way analysis of variance was used for multi group comparisons. Correlations among continuous variables were assessed by the Spearman rank-correlation coefficient. In addition, associations between PCT and NIHSS score and infarct volume were also assessed using ordered logistic regression models in multivariable adjustment with possible confounders. Cox regression analysis was assessed by univariate and multivariate analysis to identify independent predictors of mortality and report hazard ratio (HR). Therefore, common logarithmic transformation (ie, log) was performed to obtain normal distribution for skewed variables (ie, PCT and Hs-CRP concentrations). Second, we compared different prognostic risk scores from different predictive models by calculating receiver operating characteristic analysis using ROCR package. Thereby the area under the receiver operating characteristic curve (AUC) is a summary measure over criteria and cut-point choices. Finally, to study the ability of PCT for mortality prediction, we calculated Kaplan–Meier survival curves and stratified patients by PCT quarters. Finally, new reclassification metrics have been shown to provide information about the usefulness of the serum PCT and Hs-CRP. We calculated reclassification model (PCT + Hs-CRP + NHISS) to further investigate the benefit of PCT and Hs-CRP levels as compared with the NIHSS score alone, and results are reported as net reclassification improvement for mortality risk categories. All testing was two tailed, and p values less than 0.05 were considered to indicate statistical significance. All calculations were performed using SPSS for Windows, version 17.0 (SPSS Inc., Chicago, IL, USA) and the ROCR package (version 1.0-2), which is available from CRAN repository (http://cran.r-project.org/). 3. Results 3.1. Patient characteristics In our study, from 528 screened patients, the study cohort consisted of 454 patients with AIS at baseline. By the time of follow-up, at 1 year post-stroke, 48 declined the invitation to participate and 32 lost follow-up, leaving 374 individuals. The median age of patients included in this study was 69 (IQR, 63–79) years and 44.9% were women. The median NIHSS score on admission was 10 points (IQR, 6–15). The median time from stroke onset to inclusion in the study was 5.9 (IQR, 2.8–11.2) hours. Sixty-four patients died, thus the mortality rate was 17.1%. In addition, the number of tissue plasminogen activator-treated patients was 112 (29.9%). Basal characteristics of patients with acute ischemic stroke are provided in Table 1. 3.2. Main findings The results indicated that the serum PCT levels were significantly higher in stroke patients as compared to normal controls [0.78 (IQR, 0.12–1.68) ng/ml vs. 0.01 (IQR, 0.00–0.02)ng/ml; P b 0.0001]. PCT levels increased with increasing severity of stroke as defined by the NIHSS score. There was a positive correlation between levels of PCT and NIHSS (r = 0.225, P b 0.0001; Fig. 1A). There was still a significant positive correction (P = 0.006) between PCT serum levels and NIHSS score,
70
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
Table 1 Baseline characteristics of stroke patients. Demographic characteristics
Patients
Non survival
Survival
Pa
N Age (years), median (IQR) Female sex, no. (%) Stroke severity, median NIHSS score (IQR) Admission to hospital (h), median (IQR) DWI lesion (ml, IQR; n = 221) Hospital stay (days), median (IQR) Vascular risk factors, no. (%) Hypertension Diabetes mellitus Atrial fibrillation Hypercholesterolemia Coronary heart disease Family history for stroke Active smokers Clinical findings (median, IQR) Systolic blood pressure (mm Hg) Diastolic blood pressure (mmH g) BMI (kg m−2) Temperature (°C) Heart rate (beats min−1) Laboratory findings (IQR) PCT (ng ml−1) Hs-CRP (mg dl−1) Glucose (mmol l−1) White blood cell count HCY (μmol l−1) Stroke syndrome (%) TACS LACS PACS POCS Stroke etiology (%) Cardioembolic Small-vessel occlusive Large-vessel occlusive Other Unknown
374 69 (63–79) 168 (44.9) 10 (6–15) 5.9 (2.8–11.2) 26 (12–48) 42 (26–59)
64 77 (65–88) 31 (48.4) 20 (12–38) 5.5 (2.5–10.8) 38 (19–76) 28 (16–52)
310 61 (50–68) 137 (44.2) 6 (3–10) 6.1 (3.0 –11.3) 17 (7–33) 44 (27–66)
– b0.001 NS b0.0001 NS b0.0001 b0.05
258 (69.0) 115 (30.7) 76 (20.3) 127 (34.0) 117 (31.3) 82 (21.9) 90 (24.1)
41 (64.1) 19 (29.7) 13 (20.3) 21 (32.8) 21 (32.8) 15 (23.4) 16 (25.0)
217 (70.0) 96 (31.0) 63 (20.3) 106 (34.2) 96 (31.0) 67 (21.6) 74 (23.9)
NS NS NS NS NS NS NS
155 (147–168) 85 (80–90) 25.9 (23.6–27.8) 36.9 (36.5–37.4) 88 (76–99)
157 (149–172) 87 (82–94) 25.4 (23.4–27.4) 37.0 (36.5–37.7) 92 (79–103)
152 (146–169) 84 (77–89) 25.9 (23.7–27.9) 36.9 (36.4–37.3) 86 (75–98)
NS NS NS NS NS
0.78 (0.12–1.68) 0.66 (0.35–2.12) 6.20 (5.63–7.20) 8.3 (6.6–9.9) 14.2 (10.8–17.9)
3.05 (1.74–4.12) 1.28 (0.66–3.75) 6.42 (6.02–7.54) 8.3 (6.5–10.2) 15.9 (11.4–19.9)
0.51 (0.08–1.06) 0.41 (0.21–1.47) 5.96 (5.33–6.99) 8.2 (6.6–9.6) 11.5 (10.6–16.7)
b0.0001 b0.001 b0.01 NS 0.004
76 (20.3) 78 (20.9) 112 (29.9) 108 (28.9)
22 (34.4) 12 (18.8) 16 (25.0) 14 (21.8)
54 (17.4) 66 (21.3) 96 (30.9) 94 (30.4)
b0.01 NS NS NS
134 (35.8) 74 (19.8) 77 (20.6) 39 (10.4) 50 (13.4)
19 (29.7) 14 (21.9) 15 (23.4) 7 (10.9) 9 (14.1)
115 (37.1) 60 (19.4) 62 (20.0) 32 (10.3) 41 (13.2)
NS NS NS NS NS
IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; DWI, diffusion weighted imaging; Pct, procalcitonin; Hs-CRP, high sensitivity C-reactive protein; HCY, homocysteine; NS, not significant; LACS, lacunar syndrome; PACS, partial anterior circulation syndrome; POCS, posterior circulation syndrome; TACS, total anterior circulation syndrome; BMI, body mass index; TPA-T:Tissue plasminogen activator-treated. a P value was assessed using Mann–Whitney U test or Chi-square test.
using ordered logistic regression after multivariable adjustment for possible confounders, i.e., age, gender, Infarct volume, time from onset to admission, stroke syndrome, stroke etiology, vascular risk factors and serum levels of glucose, homocysteine, high-density lipoprotein, low-
density lipoprotein, cholesterol, triglycerides and Hs-CRP. Similarly, Hs-CRP was also correlated with NIHSS score (r = 0.196, P = 0.001). In the subgroup of patients (n = 221) in whom MRI was available, PCT levels paralleled lesion size. There was a positive
Fig. 1. Correlation between the serum PCT levels and other factors. (A) Correlation between the serum PCT levels and National Institutes of Health Stroke Scale (NIHSS); (B) correlation between the serum PCT levels and infarct volume.
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
correlation between levels of PCT and the infarct volume (r = 0.446, P b 0.0001; Fig. 2B.). There was still a significant positive correction (P = 0.002) between PCT serum levels and infarct volume, after multivariable adjustment for possible confounders, i.e., age, gender, NIHSS, time from onset to admission, stroke syndrome, stroke etiology, vascular risk factors and serum levels of glucose, homocysteine, high-density lipoprotein, low-density lipoprotein, cholesterol, triglycerides and Hs-CRP. There was a correlation between levels of PCT and Hs-CRP (r = 0.266, P b 0.0001). There was no correlation between levels of serum PCT levels and sex, age, blood levels of glucose, white blood cells count, homocysteine, high-density lipoprotein, low-density lipoprotein, cholesterol and triglycerides (P N 0.05, respectively). In addition, the relationship between PCT and smoking was also tested. There was a significant, albeit weak, positive correlation (r = 0.1662, P = 0.012). In the 64 non-survival patients, serum PCT levels were significantly (P b 0.0001) higher compared with those in survival patients (Table 1 and Fig. 2). In univariate cox regression analysis, log-transformed PCT and Hs-CRP levels as compared with the NIHSS score and other risk factors were presented in Table 2. After adjusting for all other significant outcome predictors, such as, age, sex, smoking, glucose, HCY, NIHSS, TPA-T, infarct volume and TACS, PCT and Hs-CRP remained independent mortality predictors with an adjusted HR of 4.24 (95% CI, 2.42–6.30) and 15.37 (95% CI, 3.25–41.08), respectively (Table 2). In the subgroup of patients (n = 221) in whom MRI evaluations were performed, PCT was an independent mortality predictor with an HR of 3.64 (95% CI, 1.54–5.88; P b 0.001) after adjustment for lesion size (HR, 1.15; 95% CI, 1.06–1.29; P = 0.009), the NIHSS score (HR, 1.11; 95% CI, 1.05–1.16; P b 0.001) and other predictors (Table 2). Similarly, Hs-CRP also was an independent mortality predictor with an HR of 12.33 (95% CI, 2.44–37.66; P b 0.001) after adjustment for lesion size, the NIHSS score and other predictors (Table 2). Based on the ROC curve, the optimal cutoff value of serum PCT levels as an mortality indicator was estimated to be 1.85 ng/ml, which yielded a sensitivity of 81.5% and a specificity of 84.7%, with the area under the curve at 0.887 (95%CI, 0.848–0.930; Fig. 3). With an AUC of 0.89, PCT showed a significantly greater discriminatory ability as compared with age and Hs-CRP, and the combined model (PCT and Hs-CRP) improved the NIHSS score and the markers alone (Table 3). In addition, a model containing known risk factors plus PCT compared with a model containing known risk factors without PCT showed a greater discriminatory ability (P b 0.01; Table 3; Fig. 3). Time to death was analysed by Kaplan–Meier curves based on PCT quartiles. Patients in the lower two quartile (PCT b 0.15 ng/ml and
Fig. 2. Serum levels of PCT in survivor and non-survivor. All data are medians and interquartile ranges (IQR). P values refer to Mann–Whitney U tests for differences between groups.
71
PCT between 0.12 and 0.78 ng/ml) had a minor risk of death compared to patients with PCT levels in the higher two quartile (PCT N 1.68 ng/ml and PCT between 0.78 and 1.68 ng/ml, P b 0.0001; HR = 3.13). 3.3. Reclassification In-sample reclassification behavior was further calculated. Seven patients who died were classified in higher risk categories and one patient who died was classified in lower risk categories using the combined model (PCT + Hs-CRP + NHISS) as compared with the NHISS as the only predictor variable. In addition, one hundred and sixty-five patients who survived were classified in lower risk categories and twenty-nine who died was classified in higher risk categories using both the combined model (PCT + Hs-CRP + NHISS) and the NHISS. Thus, the estimated net reclassification improvement for functional outcome was 0.2314. 4. Discussion To our knowledge this was the first study to show that PCT was independently associated 1-year mortality in Chinese patients with ischemic stroke, after adjusting for potential confounders. Specifically, PCT showed a significantly greater discriminatory ability as compared with Hs-CRP. Interestingly, Simon et al. [14] found that the diagnostic accuracy of PCT markers was higher than that of CRP markers among patients hospitalized for suspected bacterial infections. In addition, we demonstrated that PCT levels increased with lesion size, neurological deficit (assessed by the NIHSS), and reflect the severity of the stroke. Previous studies have demonstrated that PCT was selected to better discriminate infections from general inflammation [15] and has recently gained popularity as an early marker for sepsis [16]. Schiopu et al. [17] demonstrated a positive association between plasma PCT levels and cardiovascular risk in subjects with no previous history of acute cardiovascular events, while Katan et al. [18] reported that higher levels of PCT, a marker of infection, were independently associated with ischemic stroke risk in this multiethnic, urban cohort. In this study, we found that the serum level of PCT was substantially higher in stroke patients than in normal control subjects. This finding supported the finding from a previous study and suggested that PCT also contributes to the inflammatory process in the clinical setting of AIS [18]. Systemic markers of inflammation have been shown to be risk markers of stroke. In epidemiological studies, the leukocyte count was associated with the risk of first-time myocardial infarction and ischemic stroke, an effect that was independent of smoking and other vascular risk factors in a meta-analysis [19]. Huang et al. [20] found that elevated plasma Hs-CRP independently predicted risk of all-cause death within 3 months after acute stroke in Chinese patients, while Elkind et al. [21] reported that Hs-CRP levels were associated with increased mortality after ischemic stroke during approximately 4 years of follow-up. Consistent with those finding, we suggested a strong relationship between serum Hs-CRP levels at admission and the mortality within the 1 year after stroke. In our study, we found that PCT was a better mortality indicator than Hs-CRP, and combined model (PCT and Hs-CRP) improved the NIHSS score and those markers alone. Schuetz et al. [22] suggested that the prognostic accuracy of a single PCT value was limited. Luyt et al. [23] reported that PCT could be a prognostic marker of outcome during ventilatorassociated pneumonia, and Fluri et al. [24] found that the combination of established inflammatory markers (WBC, CRP) combined with biomarker of bacterial infection (PCT) improves prediction of infection after stroke compared to the strongest prognostic marker alone. Whether higher serum PCT levels are a cause of or merely a marker for mortality outcome in AIS patients remains uncertain. Levels of PCT and Hs-CRP, an acute-phase reactant, were strongly associated with stroke severity in our sample. Because stroke severity is also strongly
72
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
Table 2 Univariate and multivariate Cox regression analysis for mortality. Parameter
Univariate analysis
Predictor: death PCT b Age Male sex Smoking Glucose Hs-CRP b HCY NIHSS TPA-T Infarct volume c PCTb, c Hs-CRP b, c Temperature Systolic blood pressure Hypertension Atrial fibrillation Hypercholesterolemia Coronary heart disease Small-vessel occlusive Large-vessel occlusive Cardioembolic TACS PACS LACS POCS Predictor (PCT quartiles) Second vs. first quartiles Third vs. first quartiles Fourth vs. first quartiles
Multivariate analysis
HR a
95% CI a
P
HR a
95% CI a
P
6.34 1.22 1.65 1.53 1.08 28.15 1.06 1.44 0.34 1.22
3.43–11.83 1.09–1.41 1.22–2.47 1.37–1.75 1.02–1.33 7.08–115.43 1.02–1.11 1.18–1.64 0.17–0.69 1.11–1.35
b0.0001 b0.001 0.012 0.003 0.037 b0.001 0.004 b0.001 0.003 0.003
2.42–6.30 1.02–1.17 0.85–2.55 1.03–1.48 1.01–1.45 3.25–41.08 1.01–1.09 1.04–1.20 0.25–0.90 1.06–1.29 1.54–5.88 2.44–37.66
b0.0001 b0.001 0.225 0.033 0.044 b0.001 0.009 b0.001 0.023 0.009 b0.0001 b0.0001
1.03 1.01 0.75 1.63 1.76 3.18 0.62 1.07 1.13 3.64 1.78 0.45 0.68
0.89–2.42 0.54–2.42 0.55–2.00 1.00–2.70 0.75–4.09 0.92–6.15 0.21–1.80 0.68–1.62 0.74–1.69 2.04–5.65 1.21–3.99 0.13–1.07 0.27–1.49
0.405 0.903 0.611 0.068 0.212 0.13 0.354 0.805 0.776 b0.01 0.705 0.065 0.311
4.24 1.09 1.44 1.25 1.06 15.37 1.03 1.11 0.47 1.15 3.64 12.33 – – – – – – – – – 1.76 – –
0.76–3.42
0.213
2.04 3.63 4.72
0.18–22.94 1.29–10.16 2.91–7.66
0.562 0.014 b0.0001
–
HR, hazard ratio; CI, confidence interval; Pct, procalcitonin; Hs-CRP, high-sensitivity-C-reactive protein; HCY, homocysteine; TPA-T, tissue plasminogen activator-treated; NIHSS, National Institutes of Health Stroke Scale; LACS, lacunar syndrome; PACS, partial anterior circulation syndrome; POCS, posterior circulation syndrome; TACS, total anterior circulation syndrome. a Note that the hazard ratio corresponds to a unit increase in the explanatory variable. b Log-transformed to achieve normal distribution. Note that the hazard ratio corresponds to a log-unit increase in the explanatory variable. c In the subgroup of patients (n = 221) in whom MRI was available.
associated with mortality after stroke [25,26], it is not surprising that PCT and Hs-CRP are also associated with mortality. Because those markers remained independently associated with mortality even after adjusting for stroke severity, however, it seems that those markers may provide additional general prognostic information. Some role could be considered. Firstly, PCT and CRP levels are related to the severity of organ dysfunction [27]. In addition, inflammatory mediators produced
during critical illness (for example, tumor necrosis factor-a, interleukin1 and PCT) initiate a systemic cascade of endothelial damage, thrombin formation, and microvascular compromise [28].
Table 3 Receiver operating characteristics curve analysis. Parameter
PCT NHISS Hs-CRP HCY Age Sex TPA-T TACS Combined model A b Combined model B c Combined model C
Mortality AUC
95%CI
Pa
0.89 0.77 0.68 0.63 0.70 0.58 0.49 0.64 0.91 0.87 0.95
0.85–0.93 0.71–0.84 0.59–0.77 0.54–0.69 0.62–0.79 0.52–0.63 0.40–0.58 0.57–0.72 0.87–0.95 0.82–0.90 0.90–0.98
b0.01 b0.001 b0.0001 b0.001 b0.0001 b0.0001 b0.0001 0.048 0.653 b0.01
AUC: area under the curve; CI: confidence interval; PCT, procalcitonin; Hs-CRP, high sensitivity C reactive protein; HCY, homocysteine; NIHSS, National Institutes of Health Stroke Scale; TACS, total anterior circulation syndrome. Combined A including PCT/Hs-CRP. Combined B including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/TACS. Combined C Including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/TACS/PCT. a P value when other parameter was compared with PCT. b P value b 0.01 when combined model A is compared with model C. c P value b 0.01 when combined model B is compared with model C.
Fig. 3. Receiver operator characteristic curve demonstrating sensitivity as a function of 1-specificity for predicting the mortality within 1 year based on the different models. Model a including PCT + Hs-CRP; model b including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/ TACS; and model c including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/TACS/PCT.
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
Our research was a prospective study with integrated and detailed baseline, outcome, and blood sample data. The strengths of this study are studying serum PCT and Hs-CRP in Chinese patients with AIS simultaneously and its robust outcome measures, both of which increase the relevance of our findings. Genetic factors may be influencing the concentrations of serum biomarkers. Study from different population is more likely to draw accurate conclusions. Proof of this hypothesis awaits confirmation from other studies. A number of issues have to be taken into account when interpreting the results of the present study. Concentrations of inflammatory markers at the site of infarction may only partially be reflected by the inflammatory response in the peripheral blood. In addition, although patients with significant acute medical illnesses known to be associated with an inflammatory burden were excluded from this study, included participants were likely experiencing some degree of systemic inflammation prior to the stroke. As a result, levels of the inflammatory markers detected here were not solely due to stroke. Secondly, the level of PCT was measured only once for each patient (at admission). Serial measurements of PCT were not available because this was not considered during the planning of the epidemiological cohort that this study arises from. Therefore, our measurement was not able to reflect potentially dynamic changes of PCT level. Thirdly, in our study, all patients who died were included. The cause of death was not known. Thus, we could not obtain the relation between PCT and specific diseases, which caused of death. Further study according the cause of death should be considered. Lastly, the samples were also geographically limited, potentially limiting the generalizability of our results. MRI testing was not available for all patients which may have biased results. Larger studies are needed to confirm our results and elucidate the underlying mechanisms. 5. Conclusions Serum levels of PCT and Hs-CRP at admission were independent predictor of long-term mortality after ischemic stroke in Chinese sample. The predictive value of PCT was better than Hs-CRP, and combined model (PCT and Hs-CRP) may provide additional general prognostic information. Further studies are necessary to confirm this association. Conflict of interest statement The authors have no relevant potential conflicts of interest to declare. Moreover, the content has not been published or submitted for publication elsewhere. Acknowledgments All authors have contributed significantly, and all authors are in agreement with the content of the manuscript. We also express our gratitude to all the patients who participated in this study, and thereby made this work possible. References [1] Bonita R, Mendis S, Truelsen T, Bogousslavsky J, Toole J, Yatsu F. The global stroke initiative. Lancet Neurol 2004;3:391–3. [2] Slot KB, Berge E, Dorman P, Lewis S, Dennis M, Sandercock P. Impact of functional status at six months on long term survival in patients with ischaemic stroke: prospective cohort studies. BMJ 2008;336:376–9.
73
[3] Qi A, Li Y, Liu Q, et al. Thioredoxin is a novel diagnostic and prognostic marker in patients with ischemic stroke. Free Radic Biol Med 2015;80:129–35. [4] Tu WJ, Zhao SJ, Liu TG, et al. Combination of high-sensitivity C-reactive protein and homocysteine predicts the short-term outcomes of Chinese patients with acute ischemic stroke. Neurol Res 2013;35(9):912–21. [5] Vibo R, Kõrv J, Roose M, Kampus P, Muda P, Zilmer K, Zilmer M. Free Radic Res. Acute phase proteins and oxidised low-density lipoprotein in association with ischemic stroke subtype, severity and outcome. Free Radic Res 2007;41(3):282–7. [6] Yoldas T, Gonen M, Godekmerdan A, Ilhan F, Bayram E. The serum high-sensitive C reactive protein and homocysteine levels to evaluate the prognosis of acute ischemic stroke. Mediators Inflamm 2007;2007:15929. [7] Maruna P, Nedelnikova K, Gurlich R. Physiology and genetics of procalcitonin. Physiol Res 2000;49:S57–62. [8] Mimoz O, Edouard AR, Samii K, et al. Procalcitonin and C-reactive protein during the early posttraumatic systemic inflammatory response syndrome. Intensive Care Med 1998;24(2):185–8. [9] Hatano S. Experience from a multicentre stroke register: a preliminary report. Bull World Health Organ 1976;54:541–53. [10] Adams Jr HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh III EE. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993;24: 35–41. [11] Bamford J, Sandercock P, Dennis M, et al. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet 1991;337:1521–6. [12] Brott T, Marler JR, Olinger CP, Adams Jr HP, Tomsick T, Barsan WG, et al. Measurements of acute cerebral infarction: lesion size by computed tomography. Stroke 1989;20:871–5. [13] Sims JR, Gharai LR, Schaefer PW, Vangel M, Rosenthal ES. ABC/2 for rapid clinical estimate of infarct, perfusion, and mismatch volumes. Neurology 2009;72:2104–10. [14] Simon L, Gauvin F, Amre DK, et al. Serum procalcitonin and C-reactive protein levels as markers of bacterial infection: a systematic review and meta-analysis. Clin Infect Dis 2004;39(2):206–17. [15] Sakr Y, Sponholz C, Tuche F, Brunkhorst F, Reinhart K. The role of procalcitonin in febrile neutropenic patients: review of the literature. Infection 2008;36:396–407. [16] McGrane S, Girard TD, Thompson JL, et al. Procalcitonin and C-reactive protein levels at admission as predictors of duration of acute brain dysfunction in critically ill patients. Crit Care 2011;15(2):R78. [17] Schiopu A, Hedblad B, Engström G, et al. Plasma procalcitonin and the risk of cardiovascular events and death: a prospective population‐based study. J Intern Med 2012;272(5):484–91. [18] Katan M, Moon YP, DeRosa J, et al. Procalcitonin, copeptin and midregional pro-atrial natriuretic peptide as markers of ischemic stroke risk: the Northern Manhattan Study. Stroke 2014;45(Suppl. 1):A54-A54. [19] Danesh J, Collins R, Appleby P, Peto R. Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA 1998;279:1477–82. [20] Huang Y, Jing J, Zhao XQ, Wang CX, Wang YL, Liu GF, Wang CJ, Liu LP, Yang XM, Jiao Y, Jiao Y, Wang LS, Wang YJ, Gu WK. High-sensitivity C-reactive protein is a strong risk factor for death after acute ischemic stroke among Chinese. CNS Neurosci Ther 2012;18(3):261–6. [21] Elkind MS, Tai W, Coates K, Paik MC, Sacco RL. High-sensitivity C-reactive protein, lipoprotein-associated phospholipase A2, and outcome after ischemic stroke. Arch Intern Med 2006;166(19):2073–80. [22] Schuetz P, Suter-Widmer I, Chaudri A, Christ-Crain M, Zimmerli W, et al. Prognostic value of procalcitonin in community-acquired pneumonia. Eur Respir J 2011;37: 384–92. [23] Luyt CE, Guérin V, Combes A, et al. Procalcitonin kinetics as a prognostic marker of ventilator-associated pneumonia. Am J Respir Crit Care Med 2005;171(1):48–53. [24] Fluri F, Morgenthaler NG, Mueller B, Christ-Crain M, Katan M. Copeptin, procalcitonin and routine inflammatory markers—predictors of infection after stroke. PLoS One 2012;7(10):e48309. [25] Sacco RL, Boden-Albala B, Chen X, Lin IF, Kargman DE, Paik MC. Relationship of 6-month functional outcome and stroke severity: implications for acute stroke trials from the Northern Manhattan Stroke Study [abstract]. Neurology 1998;50(Suppl. 4): A327. [26] Adams Jr HP, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Neurology 1999;53:126–31. [27] Castelli GP, Pognani C, Meisner M, et al. Procalcitonin and C-reactive protein during systemic inflammatory response syndrome, sepsis and organ dysfunction. Crit Care 2004;8(4):R234. [28] Wheeler AP, Bernard GR. Treating patients with severe sepsis. N Engl J Med 1999; 340:207–14.
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Serum levels of procalcitonin and high sensitivity C-reactive protein are associated with long-term mortality in acute ischemic stroke You-Mei Li a, Xue-Yuan Liu b,⁎ a b
Department of Neurology, Yangpu Hospital, Tongji University School of Medicine, Shanghai 20090, China Department of Neurology, Shanghai Tenth People's Hospital of Tongji University, Shanghai 200072, China
a r t i c l e
i n f o
Article history: Received 10 December 2014 Received in revised form 17 March 2015 Accepted 18 March 2015 Available online 27 March 2015 Keywords: Procalcitonin High sensitivity C-reactive protein Acute ischemic stroke Mortality Chinese
a b s t r a c t Objective: The aim of this study is to assess the prognostic value of systemic inflammation, as measured by the inflammatory biomarkers PCT and Hs-CRP, to predict the long-term mortality in ischemic stroke patients. Methods: We prospectively studied 374 patients with ischemic stroke who were admitted within 24 h after the onset of symptoms. Serum levels of PCT, Hs-CRP and NIH stroke scale (NIHSS) were measured at the time of admission. Clinical follow-up was performed at 1 year. The prognostic value of PCT to predict the mortality within one year was compared with Hs-CRP, NIHSS and with other known outcome predictors. Results: In the 64 non-survival patients, serum PCT levels were significantly (P b 0.0001) higher compared with those in survival patients. Multivariate COX regression analysis showed that log-transformed PCT and Hs-CRP were independent mortality predictors with adjusted hazard ratio of 4.24 (95% confidence interval [CI], 2.42–6.30) and 15.37 (95% confidence interval [CI], 3.25–41.08). The area under the receiver operating characteristic curve of PCT and Hs-CRP were 0.89 (95% CI, 0.85–0.93) and 0.68 (95% CI, 0.59–0.77) for mortality, respectively. Conclusion: Serum levels of PCT and HS-CRP at admission were independent predictor of long-term mortality after ischemic stroke in a Chinese sample. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Ischemic stroke is the third leading cause of mortality in most countries in the world and China has 2.5 million new stroke cases each year and 7.5 million stroke survivors [1]. Mortality after 1 year ranges between 21% and 27%; 15% to 30% of survivors are permanently disabled [2]. An early risk assessment with estimate of the severity of disease and prognosis is pivotal for optimized care and allocation of healthcare resources to improve outcome [3]. Inflammatory processes have fundamental roles in stroke in both the etiology of ischemic cerebrovascular disease and the pathophysiology of cerebral ischemia [4]. Inflammatory markers (fibrinogen and CRP) predicted the stroke severity and outcome [5]. It has been reported that it is possible to use the increase in the concentration of acute phase reactants and especially the high sensitivity C-reactive protein (Hs-CRP) to help predict future cerebrovascular mortality [6]. Procalcitonin (PCT) was a protein of 116 amino-acids with a molecular weight of 13 kDa. The probable site of PCT production during inflammation is the neuroendocrine cells in the lungs or intestine [7]. Mimoz et al. [8] found that an early and transient release of PCT into the circulation was observed after severe ⁎ Corresponding author at: No. 301, RanchangZhong Road, Shanghai 200072, China. Tel./fax: +86 21 66300588. E-mail address: [email protected] (X.-Y. Liu).
http://dx.doi.org/10.1016/j.jns.2015.03.032 0022-510X/© 2015 Elsevier B.V. All rights reserved.
trauma and the amount of circulating PCT seemed proportional to the severity of tissue injury and hypovolemia, yet unrelated to infection, indicating an inflammation-related induction of PCT. Data from large-sample studies in China about the relationship between inflammatory biomarkers and mortality in stroke patients are rare, and evidences in statistically strong power are needed. Thus, the primary aim of our prospective cohort study was to assess the prognostic value of systemic inflammation, as measured by the inflammatory biomarkers PCT and Hs-CRP, to predict the long-term mortality in ischemic stroke patients. 2. Method 2.1. Patients and study design We conducted a prospective cohort study at the Shanghai Tenth People's Hospital. From September 2010 to October 2013, all patients with an acute ischemic stroke event were included. Patients were eligible for inclusion if they were admitted to the emergency department with an acute ischemic stroke defined according to the World Health Organization criteria [9] and with symptom onset within 24 h. we excluded patients with intracranial hemorrhage, a history of recent surgery or trauma during the preceding 2 months, renal insufficiency (creatinine N1.5 mg/dl), malignancy, febrile disorders, acute or chronic inflammatory
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
disease at study enrollment (which were defined by physical examination, past medical history and white blood count test), autoimmune diseases, severe edema or a prior myocardial infarction onset b 3 months, as well as those with a history of valvular heart disease, or intracardiac thrombus on echocardiograph. In addition, patients with no evidence of infarction on CT or MRI within 24 h after symptom onset were excluded from the study. The control cases (N = 200) were of similar age and gender distribution to the AIS patients. They had no known diseases and were not using any medication. The median age of controls included in this study was 69 (IQR, 62–79) years and 45.0% were women. A detailed medical history was taken and clinical and laboratory examinations were performed on all participants in both groups. The study was approved by the local ethics committee of Shanghai Tenth People's Hospital of Tongji University. The patients or their relatives gave written informed consent prior to entering the study. 2.2. Clinical variables The following clinical and demographical data were taken: age, gender, stroke etiology, blood pressure, leukocyte count, and presence of risk factors (ie, age; sex; smoking history; hypercholesterolemia; history of hypertension, diabetes mellitus, previous ischemic stroke, or transient ischemic attack, respectively; positive family history for myocardial infarction, stroke, or transient ischemic attack). Stroke cause was determined according to the criteria of the TOAST (Trial of Org 10172 in Acute Stroke Treatment) classification [10], which distinguishes largeartery arteriosclerosis cardio embolism, small-artery occlusion, other causative factor, and undetermined causative factor. The clinical stroke syndrome was determined by applying the criteria of the Oxfordshire Community Stroke Project: total anterior circulation syndrome (TACS); partial anterior circulation syndrome (PACS); lacunar syndrome (LACS); and posterior circulation syndrome (POCS) [11]. The National Institutes of Health Stroke Scale (NIHSS) score was assessed on admission (with greater scores indicating increasing severity) [12]. 2.3. Neuroimaging Diagnosis of stroke was based on the results of strict neuroradiological examination (brain computer tomography, magnetic resonance imaging (MRI), or both) according to the International Classification of Diseases, ninth revision. MRI with diffusion-weighted imaging (DWI) was available in 221 stroke patients (59.1%). In those patients, DWI lesion volumes were determined by an experienced neurologist (Liu XY) who was unaware of the clinical and laboratory results. The infarct volume was calculated by using the formula 0.5 × a × b × c (where a is the maximal longitudinal diameter, b is the maximal transverse diameter perpendicular to a, and c is the number of 10-mm slices containing infarct) [13]. 2.4. End points and follow-up Our study finished 1-year follow-up. The end point of this study was death from any cause within a 1-year follow-up. The death was considered as an outcome variable. 1-year mortality was defined as long-term mortality. Follow-up was performed by two trained medical students with a structured follow-up telephone interview with the patient or, if not possible, with the closest relative or family physician. 2.5. Blood collection and quantification For the purpose of this study, blood samples of patients who were admitted to hospital were prospectively drawn from the antecubital vein. After centrifugation, serum of the samples were immediately stored at −80 °C before assay. White blood cell (WBC) count, Hs-CRP, PCT and other biochemical measurements were done using standard
69
laboratory methods. PCT levels were measured using Enzyme-Linked Fluorescent Assay by VIDAS B.R.A.H.M.S. (Biomerieux, Durham, USA), and Hs-CRP was analyzed by the Roche Cobas Integra 800 analyzer (Roche Diagnostic, Indianapolis, IN, USA). The lower detection limit was 0 ng/ml and the line range was 0–200 ng/ml. The intra-assay coefficient of variation [CV] and inter-assay CV were 3.54–7.07% and 4.15%–9.86%, respectively. 2.6. Statistical analysis Discrete variables are summarized as counts (percentage), and continuous variables as medians and interquartile ranges (IQRs). Twogroup comparison of not normally distributed data was performed using Mann–Whitney U test, and a Kruskal–Wallis one-way analysis of variance was used for multi group comparisons. Correlations among continuous variables were assessed by the Spearman rank-correlation coefficient. In addition, associations between PCT and NIHSS score and infarct volume were also assessed using ordered logistic regression models in multivariable adjustment with possible confounders. Cox regression analysis was assessed by univariate and multivariate analysis to identify independent predictors of mortality and report hazard ratio (HR). Therefore, common logarithmic transformation (ie, log) was performed to obtain normal distribution for skewed variables (ie, PCT and Hs-CRP concentrations). Second, we compared different prognostic risk scores from different predictive models by calculating receiver operating characteristic analysis using ROCR package. Thereby the area under the receiver operating characteristic curve (AUC) is a summary measure over criteria and cut-point choices. Finally, to study the ability of PCT for mortality prediction, we calculated Kaplan–Meier survival curves and stratified patients by PCT quarters. Finally, new reclassification metrics have been shown to provide information about the usefulness of the serum PCT and Hs-CRP. We calculated reclassification model (PCT + Hs-CRP + NHISS) to further investigate the benefit of PCT and Hs-CRP levels as compared with the NIHSS score alone, and results are reported as net reclassification improvement for mortality risk categories. All testing was two tailed, and p values less than 0.05 were considered to indicate statistical significance. All calculations were performed using SPSS for Windows, version 17.0 (SPSS Inc., Chicago, IL, USA) and the ROCR package (version 1.0-2), which is available from CRAN repository (http://cran.r-project.org/). 3. Results 3.1. Patient characteristics In our study, from 528 screened patients, the study cohort consisted of 454 patients with AIS at baseline. By the time of follow-up, at 1 year post-stroke, 48 declined the invitation to participate and 32 lost follow-up, leaving 374 individuals. The median age of patients included in this study was 69 (IQR, 63–79) years and 44.9% were women. The median NIHSS score on admission was 10 points (IQR, 6–15). The median time from stroke onset to inclusion in the study was 5.9 (IQR, 2.8–11.2) hours. Sixty-four patients died, thus the mortality rate was 17.1%. In addition, the number of tissue plasminogen activator-treated patients was 112 (29.9%). Basal characteristics of patients with acute ischemic stroke are provided in Table 1. 3.2. Main findings The results indicated that the serum PCT levels were significantly higher in stroke patients as compared to normal controls [0.78 (IQR, 0.12–1.68) ng/ml vs. 0.01 (IQR, 0.00–0.02)ng/ml; P b 0.0001]. PCT levels increased with increasing severity of stroke as defined by the NIHSS score. There was a positive correlation between levels of PCT and NIHSS (r = 0.225, P b 0.0001; Fig. 1A). There was still a significant positive correction (P = 0.006) between PCT serum levels and NIHSS score,
70
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
Table 1 Baseline characteristics of stroke patients. Demographic characteristics
Patients
Non survival
Survival
Pa
N Age (years), median (IQR) Female sex, no. (%) Stroke severity, median NIHSS score (IQR) Admission to hospital (h), median (IQR) DWI lesion (ml, IQR; n = 221) Hospital stay (days), median (IQR) Vascular risk factors, no. (%) Hypertension Diabetes mellitus Atrial fibrillation Hypercholesterolemia Coronary heart disease Family history for stroke Active smokers Clinical findings (median, IQR) Systolic blood pressure (mm Hg) Diastolic blood pressure (mmH g) BMI (kg m−2) Temperature (°C) Heart rate (beats min−1) Laboratory findings (IQR) PCT (ng ml−1) Hs-CRP (mg dl−1) Glucose (mmol l−1) White blood cell count HCY (μmol l−1) Stroke syndrome (%) TACS LACS PACS POCS Stroke etiology (%) Cardioembolic Small-vessel occlusive Large-vessel occlusive Other Unknown
374 69 (63–79) 168 (44.9) 10 (6–15) 5.9 (2.8–11.2) 26 (12–48) 42 (26–59)
64 77 (65–88) 31 (48.4) 20 (12–38) 5.5 (2.5–10.8) 38 (19–76) 28 (16–52)
310 61 (50–68) 137 (44.2) 6 (3–10) 6.1 (3.0 –11.3) 17 (7–33) 44 (27–66)
– b0.001 NS b0.0001 NS b0.0001 b0.05
258 (69.0) 115 (30.7) 76 (20.3) 127 (34.0) 117 (31.3) 82 (21.9) 90 (24.1)
41 (64.1) 19 (29.7) 13 (20.3) 21 (32.8) 21 (32.8) 15 (23.4) 16 (25.0)
217 (70.0) 96 (31.0) 63 (20.3) 106 (34.2) 96 (31.0) 67 (21.6) 74 (23.9)
NS NS NS NS NS NS NS
155 (147–168) 85 (80–90) 25.9 (23.6–27.8) 36.9 (36.5–37.4) 88 (76–99)
157 (149–172) 87 (82–94) 25.4 (23.4–27.4) 37.0 (36.5–37.7) 92 (79–103)
152 (146–169) 84 (77–89) 25.9 (23.7–27.9) 36.9 (36.4–37.3) 86 (75–98)
NS NS NS NS NS
0.78 (0.12–1.68) 0.66 (0.35–2.12) 6.20 (5.63–7.20) 8.3 (6.6–9.9) 14.2 (10.8–17.9)
3.05 (1.74–4.12) 1.28 (0.66–3.75) 6.42 (6.02–7.54) 8.3 (6.5–10.2) 15.9 (11.4–19.9)
0.51 (0.08–1.06) 0.41 (0.21–1.47) 5.96 (5.33–6.99) 8.2 (6.6–9.6) 11.5 (10.6–16.7)
b0.0001 b0.001 b0.01 NS 0.004
76 (20.3) 78 (20.9) 112 (29.9) 108 (28.9)
22 (34.4) 12 (18.8) 16 (25.0) 14 (21.8)
54 (17.4) 66 (21.3) 96 (30.9) 94 (30.4)
b0.01 NS NS NS
134 (35.8) 74 (19.8) 77 (20.6) 39 (10.4) 50 (13.4)
19 (29.7) 14 (21.9) 15 (23.4) 7 (10.9) 9 (14.1)
115 (37.1) 60 (19.4) 62 (20.0) 32 (10.3) 41 (13.2)
NS NS NS NS NS
IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; DWI, diffusion weighted imaging; Pct, procalcitonin; Hs-CRP, high sensitivity C-reactive protein; HCY, homocysteine; NS, not significant; LACS, lacunar syndrome; PACS, partial anterior circulation syndrome; POCS, posterior circulation syndrome; TACS, total anterior circulation syndrome; BMI, body mass index; TPA-T:Tissue plasminogen activator-treated. a P value was assessed using Mann–Whitney U test or Chi-square test.
using ordered logistic regression after multivariable adjustment for possible confounders, i.e., age, gender, Infarct volume, time from onset to admission, stroke syndrome, stroke etiology, vascular risk factors and serum levels of glucose, homocysteine, high-density lipoprotein, low-
density lipoprotein, cholesterol, triglycerides and Hs-CRP. Similarly, Hs-CRP was also correlated with NIHSS score (r = 0.196, P = 0.001). In the subgroup of patients (n = 221) in whom MRI was available, PCT levels paralleled lesion size. There was a positive
Fig. 1. Correlation between the serum PCT levels and other factors. (A) Correlation between the serum PCT levels and National Institutes of Health Stroke Scale (NIHSS); (B) correlation between the serum PCT levels and infarct volume.
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
correlation between levels of PCT and the infarct volume (r = 0.446, P b 0.0001; Fig. 2B.). There was still a significant positive correction (P = 0.002) between PCT serum levels and infarct volume, after multivariable adjustment for possible confounders, i.e., age, gender, NIHSS, time from onset to admission, stroke syndrome, stroke etiology, vascular risk factors and serum levels of glucose, homocysteine, high-density lipoprotein, low-density lipoprotein, cholesterol, triglycerides and Hs-CRP. There was a correlation between levels of PCT and Hs-CRP (r = 0.266, P b 0.0001). There was no correlation between levels of serum PCT levels and sex, age, blood levels of glucose, white blood cells count, homocysteine, high-density lipoprotein, low-density lipoprotein, cholesterol and triglycerides (P N 0.05, respectively). In addition, the relationship between PCT and smoking was also tested. There was a significant, albeit weak, positive correlation (r = 0.1662, P = 0.012). In the 64 non-survival patients, serum PCT levels were significantly (P b 0.0001) higher compared with those in survival patients (Table 1 and Fig. 2). In univariate cox regression analysis, log-transformed PCT and Hs-CRP levels as compared with the NIHSS score and other risk factors were presented in Table 2. After adjusting for all other significant outcome predictors, such as, age, sex, smoking, glucose, HCY, NIHSS, TPA-T, infarct volume and TACS, PCT and Hs-CRP remained independent mortality predictors with an adjusted HR of 4.24 (95% CI, 2.42–6.30) and 15.37 (95% CI, 3.25–41.08), respectively (Table 2). In the subgroup of patients (n = 221) in whom MRI evaluations were performed, PCT was an independent mortality predictor with an HR of 3.64 (95% CI, 1.54–5.88; P b 0.001) after adjustment for lesion size (HR, 1.15; 95% CI, 1.06–1.29; P = 0.009), the NIHSS score (HR, 1.11; 95% CI, 1.05–1.16; P b 0.001) and other predictors (Table 2). Similarly, Hs-CRP also was an independent mortality predictor with an HR of 12.33 (95% CI, 2.44–37.66; P b 0.001) after adjustment for lesion size, the NIHSS score and other predictors (Table 2). Based on the ROC curve, the optimal cutoff value of serum PCT levels as an mortality indicator was estimated to be 1.85 ng/ml, which yielded a sensitivity of 81.5% and a specificity of 84.7%, with the area under the curve at 0.887 (95%CI, 0.848–0.930; Fig. 3). With an AUC of 0.89, PCT showed a significantly greater discriminatory ability as compared with age and Hs-CRP, and the combined model (PCT and Hs-CRP) improved the NIHSS score and the markers alone (Table 3). In addition, a model containing known risk factors plus PCT compared with a model containing known risk factors without PCT showed a greater discriminatory ability (P b 0.01; Table 3; Fig. 3). Time to death was analysed by Kaplan–Meier curves based on PCT quartiles. Patients in the lower two quartile (PCT b 0.15 ng/ml and
Fig. 2. Serum levels of PCT in survivor and non-survivor. All data are medians and interquartile ranges (IQR). P values refer to Mann–Whitney U tests for differences between groups.
71
PCT between 0.12 and 0.78 ng/ml) had a minor risk of death compared to patients with PCT levels in the higher two quartile (PCT N 1.68 ng/ml and PCT between 0.78 and 1.68 ng/ml, P b 0.0001; HR = 3.13). 3.3. Reclassification In-sample reclassification behavior was further calculated. Seven patients who died were classified in higher risk categories and one patient who died was classified in lower risk categories using the combined model (PCT + Hs-CRP + NHISS) as compared with the NHISS as the only predictor variable. In addition, one hundred and sixty-five patients who survived were classified in lower risk categories and twenty-nine who died was classified in higher risk categories using both the combined model (PCT + Hs-CRP + NHISS) and the NHISS. Thus, the estimated net reclassification improvement for functional outcome was 0.2314. 4. Discussion To our knowledge this was the first study to show that PCT was independently associated 1-year mortality in Chinese patients with ischemic stroke, after adjusting for potential confounders. Specifically, PCT showed a significantly greater discriminatory ability as compared with Hs-CRP. Interestingly, Simon et al. [14] found that the diagnostic accuracy of PCT markers was higher than that of CRP markers among patients hospitalized for suspected bacterial infections. In addition, we demonstrated that PCT levels increased with lesion size, neurological deficit (assessed by the NIHSS), and reflect the severity of the stroke. Previous studies have demonstrated that PCT was selected to better discriminate infections from general inflammation [15] and has recently gained popularity as an early marker for sepsis [16]. Schiopu et al. [17] demonstrated a positive association between plasma PCT levels and cardiovascular risk in subjects with no previous history of acute cardiovascular events, while Katan et al. [18] reported that higher levels of PCT, a marker of infection, were independently associated with ischemic stroke risk in this multiethnic, urban cohort. In this study, we found that the serum level of PCT was substantially higher in stroke patients than in normal control subjects. This finding supported the finding from a previous study and suggested that PCT also contributes to the inflammatory process in the clinical setting of AIS [18]. Systemic markers of inflammation have been shown to be risk markers of stroke. In epidemiological studies, the leukocyte count was associated with the risk of first-time myocardial infarction and ischemic stroke, an effect that was independent of smoking and other vascular risk factors in a meta-analysis [19]. Huang et al. [20] found that elevated plasma Hs-CRP independently predicted risk of all-cause death within 3 months after acute stroke in Chinese patients, while Elkind et al. [21] reported that Hs-CRP levels were associated with increased mortality after ischemic stroke during approximately 4 years of follow-up. Consistent with those finding, we suggested a strong relationship between serum Hs-CRP levels at admission and the mortality within the 1 year after stroke. In our study, we found that PCT was a better mortality indicator than Hs-CRP, and combined model (PCT and Hs-CRP) improved the NIHSS score and those markers alone. Schuetz et al. [22] suggested that the prognostic accuracy of a single PCT value was limited. Luyt et al. [23] reported that PCT could be a prognostic marker of outcome during ventilatorassociated pneumonia, and Fluri et al. [24] found that the combination of established inflammatory markers (WBC, CRP) combined with biomarker of bacterial infection (PCT) improves prediction of infection after stroke compared to the strongest prognostic marker alone. Whether higher serum PCT levels are a cause of or merely a marker for mortality outcome in AIS patients remains uncertain. Levels of PCT and Hs-CRP, an acute-phase reactant, were strongly associated with stroke severity in our sample. Because stroke severity is also strongly
72
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
Table 2 Univariate and multivariate Cox regression analysis for mortality. Parameter
Univariate analysis
Predictor: death PCT b Age Male sex Smoking Glucose Hs-CRP b HCY NIHSS TPA-T Infarct volume c PCTb, c Hs-CRP b, c Temperature Systolic blood pressure Hypertension Atrial fibrillation Hypercholesterolemia Coronary heart disease Small-vessel occlusive Large-vessel occlusive Cardioembolic TACS PACS LACS POCS Predictor (PCT quartiles) Second vs. first quartiles Third vs. first quartiles Fourth vs. first quartiles
Multivariate analysis
HR a
95% CI a
P
HR a
95% CI a
P
6.34 1.22 1.65 1.53 1.08 28.15 1.06 1.44 0.34 1.22
3.43–11.83 1.09–1.41 1.22–2.47 1.37–1.75 1.02–1.33 7.08–115.43 1.02–1.11 1.18–1.64 0.17–0.69 1.11–1.35
b0.0001 b0.001 0.012 0.003 0.037 b0.001 0.004 b0.001 0.003 0.003
2.42–6.30 1.02–1.17 0.85–2.55 1.03–1.48 1.01–1.45 3.25–41.08 1.01–1.09 1.04–1.20 0.25–0.90 1.06–1.29 1.54–5.88 2.44–37.66
b0.0001 b0.001 0.225 0.033 0.044 b0.001 0.009 b0.001 0.023 0.009 b0.0001 b0.0001
1.03 1.01 0.75 1.63 1.76 3.18 0.62 1.07 1.13 3.64 1.78 0.45 0.68
0.89–2.42 0.54–2.42 0.55–2.00 1.00–2.70 0.75–4.09 0.92–6.15 0.21–1.80 0.68–1.62 0.74–1.69 2.04–5.65 1.21–3.99 0.13–1.07 0.27–1.49
0.405 0.903 0.611 0.068 0.212 0.13 0.354 0.805 0.776 b0.01 0.705 0.065 0.311
4.24 1.09 1.44 1.25 1.06 15.37 1.03 1.11 0.47 1.15 3.64 12.33 – – – – – – – – – 1.76 – –
0.76–3.42
0.213
2.04 3.63 4.72
0.18–22.94 1.29–10.16 2.91–7.66
0.562 0.014 b0.0001
–
HR, hazard ratio; CI, confidence interval; Pct, procalcitonin; Hs-CRP, high-sensitivity-C-reactive protein; HCY, homocysteine; TPA-T, tissue plasminogen activator-treated; NIHSS, National Institutes of Health Stroke Scale; LACS, lacunar syndrome; PACS, partial anterior circulation syndrome; POCS, posterior circulation syndrome; TACS, total anterior circulation syndrome. a Note that the hazard ratio corresponds to a unit increase in the explanatory variable. b Log-transformed to achieve normal distribution. Note that the hazard ratio corresponds to a log-unit increase in the explanatory variable. c In the subgroup of patients (n = 221) in whom MRI was available.
associated with mortality after stroke [25,26], it is not surprising that PCT and Hs-CRP are also associated with mortality. Because those markers remained independently associated with mortality even after adjusting for stroke severity, however, it seems that those markers may provide additional general prognostic information. Some role could be considered. Firstly, PCT and CRP levels are related to the severity of organ dysfunction [27]. In addition, inflammatory mediators produced
during critical illness (for example, tumor necrosis factor-a, interleukin1 and PCT) initiate a systemic cascade of endothelial damage, thrombin formation, and microvascular compromise [28].
Table 3 Receiver operating characteristics curve analysis. Parameter
PCT NHISS Hs-CRP HCY Age Sex TPA-T TACS Combined model A b Combined model B c Combined model C
Mortality AUC
95%CI
Pa
0.89 0.77 0.68 0.63 0.70 0.58 0.49 0.64 0.91 0.87 0.95
0.85–0.93 0.71–0.84 0.59–0.77 0.54–0.69 0.62–0.79 0.52–0.63 0.40–0.58 0.57–0.72 0.87–0.95 0.82–0.90 0.90–0.98
b0.01 b0.001 b0.0001 b0.001 b0.0001 b0.0001 b0.0001 0.048 0.653 b0.01
AUC: area under the curve; CI: confidence interval; PCT, procalcitonin; Hs-CRP, high sensitivity C reactive protein; HCY, homocysteine; NIHSS, National Institutes of Health Stroke Scale; TACS, total anterior circulation syndrome. Combined A including PCT/Hs-CRP. Combined B including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/TACS. Combined C Including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/TACS/PCT. a P value when other parameter was compared with PCT. b P value b 0.01 when combined model A is compared with model C. c P value b 0.01 when combined model B is compared with model C.
Fig. 3. Receiver operator characteristic curve demonstrating sensitivity as a function of 1-specificity for predicting the mortality within 1 year based on the different models. Model a including PCT + Hs-CRP; model b including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/ TACS; and model c including NIHSS/age/Hs-CRP/HCY/sex/TPA-T/TACS/PCT.
Y.-M. Li, X.-Y. Liu / Journal of the Neurological Sciences 352 (2015) 68–73
Our research was a prospective study with integrated and detailed baseline, outcome, and blood sample data. The strengths of this study are studying serum PCT and Hs-CRP in Chinese patients with AIS simultaneously and its robust outcome measures, both of which increase the relevance of our findings. Genetic factors may be influencing the concentrations of serum biomarkers. Study from different population is more likely to draw accurate conclusions. Proof of this hypothesis awaits confirmation from other studies. A number of issues have to be taken into account when interpreting the results of the present study. Concentrations of inflammatory markers at the site of infarction may only partially be reflected by the inflammatory response in the peripheral blood. In addition, although patients with significant acute medical illnesses known to be associated with an inflammatory burden were excluded from this study, included participants were likely experiencing some degree of systemic inflammation prior to the stroke. As a result, levels of the inflammatory markers detected here were not solely due to stroke. Secondly, the level of PCT was measured only once for each patient (at admission). Serial measurements of PCT were not available because this was not considered during the planning of the epidemiological cohort that this study arises from. Therefore, our measurement was not able to reflect potentially dynamic changes of PCT level. Thirdly, in our study, all patients who died were included. The cause of death was not known. Thus, we could not obtain the relation between PCT and specific diseases, which caused of death. Further study according the cause of death should be considered. Lastly, the samples were also geographically limited, potentially limiting the generalizability of our results. MRI testing was not available for all patients which may have biased results. Larger studies are needed to confirm our results and elucidate the underlying mechanisms. 5. Conclusions Serum levels of PCT and Hs-CRP at admission were independent predictor of long-term mortality after ischemic stroke in Chinese sample. The predictive value of PCT was better than Hs-CRP, and combined model (PCT and Hs-CRP) may provide additional general prognostic information. Further studies are necessary to confirm this association. Conflict of interest statement The authors have no relevant potential conflicts of interest to declare. Moreover, the content has not been published or submitted for publication elsewhere. Acknowledgments All authors have contributed significantly, and all authors are in agreement with the content of the manuscript. We also express our gratitude to all the patients who participated in this study, and thereby made this work possible. References [1] Bonita R, Mendis S, Truelsen T, Bogousslavsky J, Toole J, Yatsu F. The global stroke initiative. Lancet Neurol 2004;3:391–3. [2] Slot KB, Berge E, Dorman P, Lewis S, Dennis M, Sandercock P. Impact of functional status at six months on long term survival in patients with ischaemic stroke: prospective cohort studies. BMJ 2008;336:376–9.
73
[3] Qi A, Li Y, Liu Q, et al. Thioredoxin is a novel diagnostic and prognostic marker in patients with ischemic stroke. Free Radic Biol Med 2015;80:129–35. [4] Tu WJ, Zhao SJ, Liu TG, et al. Combination of high-sensitivity C-reactive protein and homocysteine predicts the short-term outcomes of Chinese patients with acute ischemic stroke. Neurol Res 2013;35(9):912–21. [5] Vibo R, Kõrv J, Roose M, Kampus P, Muda P, Zilmer K, Zilmer M. Free Radic Res. Acute phase proteins and oxidised low-density lipoprotein in association with ischemic stroke subtype, severity and outcome. Free Radic Res 2007;41(3):282–7. [6] Yoldas T, Gonen M, Godekmerdan A, Ilhan F, Bayram E. The serum high-sensitive C reactive protein and homocysteine levels to evaluate the prognosis of acute ischemic stroke. Mediators Inflamm 2007;2007:15929. [7] Maruna P, Nedelnikova K, Gurlich R. Physiology and genetics of procalcitonin. Physiol Res 2000;49:S57–62. [8] Mimoz O, Edouard AR, Samii K, et al. Procalcitonin and C-reactive protein during the early posttraumatic systemic inflammatory response syndrome. Intensive Care Med 1998;24(2):185–8. [9] Hatano S. Experience from a multicentre stroke register: a preliminary report. Bull World Health Organ 1976;54:541–53. [10] Adams Jr HP, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh III EE. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org 10172 in Acute Stroke Treatment. Stroke 1993;24: 35–41. [11] Bamford J, Sandercock P, Dennis M, et al. Classification and natural history of clinically identifiable subtypes of cerebral infarction. Lancet 1991;337:1521–6. [12] Brott T, Marler JR, Olinger CP, Adams Jr HP, Tomsick T, Barsan WG, et al. Measurements of acute cerebral infarction: lesion size by computed tomography. Stroke 1989;20:871–5. [13] Sims JR, Gharai LR, Schaefer PW, Vangel M, Rosenthal ES. ABC/2 for rapid clinical estimate of infarct, perfusion, and mismatch volumes. Neurology 2009;72:2104–10. [14] Simon L, Gauvin F, Amre DK, et al. Serum procalcitonin and C-reactive protein levels as markers of bacterial infection: a systematic review and meta-analysis. Clin Infect Dis 2004;39(2):206–17. [15] Sakr Y, Sponholz C, Tuche F, Brunkhorst F, Reinhart K. The role of procalcitonin in febrile neutropenic patients: review of the literature. Infection 2008;36:396–407. [16] McGrane S, Girard TD, Thompson JL, et al. Procalcitonin and C-reactive protein levels at admission as predictors of duration of acute brain dysfunction in critically ill patients. Crit Care 2011;15(2):R78. [17] Schiopu A, Hedblad B, Engström G, et al. Plasma procalcitonin and the risk of cardiovascular events and death: a prospective population‐based study. J Intern Med 2012;272(5):484–91. [18] Katan M, Moon YP, DeRosa J, et al. Procalcitonin, copeptin and midregional pro-atrial natriuretic peptide as markers of ischemic stroke risk: the Northern Manhattan Study. Stroke 2014;45(Suppl. 1):A54-A54. [19] Danesh J, Collins R, Appleby P, Peto R. Association of fibrinogen, C-reactive protein, albumin, or leukocyte count with coronary heart disease: meta-analyses of prospective studies. JAMA 1998;279:1477–82. [20] Huang Y, Jing J, Zhao XQ, Wang CX, Wang YL, Liu GF, Wang CJ, Liu LP, Yang XM, Jiao Y, Jiao Y, Wang LS, Wang YJ, Gu WK. High-sensitivity C-reactive protein is a strong risk factor for death after acute ischemic stroke among Chinese. CNS Neurosci Ther 2012;18(3):261–6. [21] Elkind MS, Tai W, Coates K, Paik MC, Sacco RL. High-sensitivity C-reactive protein, lipoprotein-associated phospholipase A2, and outcome after ischemic stroke. Arch Intern Med 2006;166(19):2073–80. [22] Schuetz P, Suter-Widmer I, Chaudri A, Christ-Crain M, Zimmerli W, et al. Prognostic value of procalcitonin in community-acquired pneumonia. Eur Respir J 2011;37: 384–92. [23] Luyt CE, Guérin V, Combes A, et al. Procalcitonin kinetics as a prognostic marker of ventilator-associated pneumonia. Am J Respir Crit Care Med 2005;171(1):48–53. [24] Fluri F, Morgenthaler NG, Mueller B, Christ-Crain M, Katan M. Copeptin, procalcitonin and routine inflammatory markers—predictors of infection after stroke. PLoS One 2012;7(10):e48309. [25] Sacco RL, Boden-Albala B, Chen X, Lin IF, Kargman DE, Paik MC. Relationship of 6-month functional outcome and stroke severity: implications for acute stroke trials from the Northern Manhattan Stroke Study [abstract]. Neurology 1998;50(Suppl. 4): A327. [26] Adams Jr HP, Davis PH, Leira EC, et al. Baseline NIH Stroke Scale score strongly predicts outcome after stroke: a report of the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Neurology 1999;53:126–31. [27] Castelli GP, Pognani C, Meisner M, et al. Procalcitonin and C-reactive protein during systemic inflammatory response syndrome, sepsis and organ dysfunction. Crit Care 2004;8(4):R234. [28] Wheeler AP, Bernard GR. Treating patients with severe sepsis. N Engl J Med 1999; 340:207–14.
