Clinical neuroscience 39
Elevated lipoprotein (a) levels predict deep vein thrombosis in acute ischemic stroke patients Dongliang Yina,*, Peng Shaoc,* and Yunling Liub Lipoprotein (a) [Lp(a)] plays a crucial role in the pathogenesis of deep vein thrombosis (DVT). The purpose of this study was to investigate whether Lp(a) serum levels at admission could be a risk factor for DVT in Chinese patients with acute ischemic stroke (AIS). A total of 232 patients with AIS were included in the study. The patients were assessed for DVT using colour Doppler ultrasonography. We performed colour Doppler ultrasonography 15 days after the stroke and whenever clinically requested. The value of Lp(a) to predict the DVT was analyzed using logistic regression analysis after adjusting for the possible confounders. In our study, 44 out of the 232 patients (19.0%) were diagnosed with DVT at 15-day follow-up. Serum Lp(a) levels were higher in AIS with DVT than in those patients without DVT [656 (interquartile range, 521–898) mg/l vs. 253 (interquartile range, 143–440) mg/l; P < 0.0001]. Increased risk of DVT associated with Lp(a) levels greater than or equal to 300 mg/l was found in the multivariate analysis [odds ratio 12.14, 95% confidence interval (CI): 3.08–42.09; P < 0.0001]. Visible by the receiver operating characteristic, the optimal cutoff value of serum Lp(a) levels for predicting DVT was projected to be 420 mg/l, yielding a sensitivity of 88.5% and
a specificity of 75.4%. With an area under the curve (AUC) of 0.89 (95% CI, 0.84–0.94), Lp(a) exhibited greater discrimination in predicting DVT compared with Hs-CRP (AUC, 0.77; 95% CI, 0.69–0.85; P < 0.01), HCY (AUC, 0.76; 95% CI, 0.68–0.84; P < 0.01), and NIHSS score (AUC, 0.74; 95% CI, 0.66–0.82; P < 0.001). Elevated serum Lp(a) levels were independent predictors of DVT in AIS patients in China, revealing the critical role played by Lp(a) in the pathogenesis of DVT. NeuroReport 27:39–44 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Introduction
and recurrent disease [7]. However, Lippi et al. [8] failed to demonstrate a convincing association between Lp(a) and thrombosis in the venous district. Similarly, Vormittag et al. [9] found no association between Lp(a) and VTE.
Lipoprotein (a) [Lp(a)] consists of an LDL-like particle and the specific apolipoprotein (a) [apo(a)], which is covalently bound to the apoB of the LDL-like particle. Numerous epidemiologic studies have identified Lp(a) as a risk factor for atherosclerotic diseases such as coronary heart disease and stroke [1,2]. Lp(a) elevation predicts higher odds of poor functional outcomes for patients with stroke compared with patients with normal levels of Lp(a) [3,4]. In addition, elevated Lp(a) is an independent risk factor for ischemic stroke, and may be especially relevant for young stroke patients [5]. Recently, Lp(a) has been suggested as a risk factor for deep vein thrombosis (DVT). Serum levels of Lp(a) are determined largely by genetic variation in the gene encoding for apo(a), the specific protein component of Lp(a) that is very homologous to plasminogen. NowakGöttl et al. [6] found that Lp(a) greater than 300 mg/l was a risk factor for venous thromboembolism (VTE) during childhood. Another study showed that Lp(a) was an independent risk factor for VTE in adults, suggesting that it may be involved in the pathogenesis of idiopathic 0959-4965 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
NeuroReport 2016, 27:39–44 Keywords: acute ischemic stroke, deep vein thrombosis, lipoprotein (a), predictor a Department of Neurology, Jinan 3rd Peoples Hospital, bDepartment of Endocrinology, Qianfoshan Hospital of Shandong Province, Jinan and c Department of Rehabilitation Medicine, Yantaishan Hospital, Yantai, China
Correspondence to Dongliang Yin, No. 1 Wangsheren North Road, Jinan 250101, China Tel: + 86 135 0641 2392; fax: + 86 0531 85853111; e-mail: [email protected] *Dongliang Yin and Peng Shao contributed equally to the writing of this article. Received 30 August 2015 accepted 19 October 2015
There have been many reports describing the association of DVT with stroke [10]. Without VTE prophylaxis, up to 75% of patients with hemiplegia after stroke develop DVT [11]. DVT frequently occurs in the course of ischemic stroke and ∼ 5% of early deaths following stroke are attributed to pulmonary embolism (PE) [12]. However, little is known about the nature of the relationship between Lp(a) and DVT in patients with ischemic stroke. Thus, the purpose of this study was to investigate whether Lp(a) serum levels at admission could be a risk factor for DVT in Chinese patients with acute ischemic stroke (AIS).
Materials and methods The Institutional Review Committee on Human Research of the Jinan third People’s Hospital approved the study protocol. All patients received written DOI: 10.1097/WNR.0000000000000496
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
40 NeuroReport 2016, Vol 27 No 1
information concerning the background and procedures of the study, and the patients or their relatives gave written informed consent before participating in the study. From January 2013 to December 2014, the study population comprised 232 consecutive patients with an AIS diagnosis who had been referred to the Jinan third People’s Hospital. Patients were diagnosed according to the WHO criteria and with symptom onset within 48 h. Brain imaging (either computed tomography or MRI) was performed routinely within 24 h after admission. Exclusion criteria were extradural or subdural hematomas; hemorrhagic transformation of ischemic infarct; a DVT within the previous 3 months; formerly known abnormalities of any coagulation factor; use of unfractionated heparin, low-molecular-weight heparin, or any form of oral anticoagulant in the past month; life expectancy less than 3 months; malignant tumor, surgery, or trauma during the preceding 2 months; severe edema; serious infections at study enrollment; blood not drawn within 24 h after hospital admission; presence of diverse medical illnesses or current medications that influence serum Lp (a) levels, and incomplete workups for cerebrovascular status; inaccessible to follow-up; or failure to give consent. The control cases (N = 100) were of similar age and sex distribution as the AIS patients. They had no known diseases and were not using any medication. At baseline, demographic data (age and sex), BMI, and history of risk factors (hypertension, diabetes mellitus, atrial fibrillation, hyperlipidemia, smoking habit, and alcohol abuse) were obtained. For each patient, the time from stroke onset to admission was also recorded. At admission, the neurological deficit was quantified using the National Institutes of Health Stroke Scale (NIHSS) [13]. AIS was classified according to the TOAST system [14], which distinguishes large-artery arteriosclerosis, cardioembolism, small-artery occlusion, other causative factor, and undetermined causative factor. MRI was performed within 24 h after admission to assess the site, cause, and size of the brain infarct. MRI was performed using a stroke protocol, including T1-weighted, T2-weighted, and diffusion-weighted imaging (DWI) sequences, and a magnetic resonance angiography. DWI lesion volumes were 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) [15]. All patients received treatment according to current guidelines. The included patients were assessed for DVT using colour Doppler ultrasonography, regardless of whether they were symptomatic or asymptomatic. We performed colour Doppler ultrasonography 15 days after AIS and whenever clinically requested. Ultrasonography, using real-time B-mode with compression, was carried out with
a 7.5 or a 5.0 MHz transducer. Two areas of the leg were examined: the common femoral vein at the inguinal ligament and the popliteal vein at the knee-joint line traced down to the point of the trifurcation of the calf veins. Veins were only scanned in the transverse plane. We used Vivid 7 Dimension (GE, Miami, Florida, USA) with the 7–10 Hz linear probe. The diagnosis of DVT was based either on the presence of a noncompressible segment [compression ultrasound test (CUS)] or the flow impairment on color Doppler imaging. This method is recognized as sufficiently sensitive and specific, especially for proximal DVT detection [16]. Fasting venous blood samples were collected from all participants in vacutainer tubes, and were quickly centrifuged to avoid glycolysis. Serum samples were kept at − 80°C until assay. Lp(a) levels were measured using immunoturbidimetric method by Olympus 2700 (Olympus, Tokyo, Japan). The lower detection limit was 30 mg/l and the line range was 30–1200 mg/l. The intraassay coefficient of variation (CV) and interassay CV were 1.4–2.2% and 1.6–2.5%, respectively. The median serum Lp(a) level in 100 healthy controls was 148 mg/l and the 97.5 percentile was 375 mg/l. Routine blood biomarkers – for instance, fibrinogen, prothrombin time, activated partial thromboplastin time, thrombin time, glucose, high-sensitivity C-reactive protein (Hs-CRP), and homocysteine (HCY) were tested using standard detection methods. Data were presented as the percentage (number) or frequencies for discrete variables, and medians [interquartile range (IQR)] for continuous variables. Values of the measured parameters were checked for normal distribution by using the Kolmogorov–Smirnov test before statistical analysis. Mann–Whitney U-test was carried out for the comparison between groups for continuous variables; in addition, χ2-test was used for the comparison of proportions between groups. Spearman’s rank correlation assessed the correlations among laboratory parameters. To investigate whether Lp(a) allowed predicting of DVT in AIS, different statistical methods were used. First, whether a high level of Lp(a) was a risk factor for DVT was confirmed by calculating the odds ratio (OR) and the 95% confidence interval (CI). We used 300 mg/l as the cutoff point. To adjust for possible confounders – for example, age, sex, BMI, conventional vascular risk factors, history of vein thrombosis, and other blood biomarkers – we used a logistic regression model. Second, for further comparison of the prognostic values of different scores from different predictive models, receiver operating characteristic was graphed to determine the optimal cutoff values, combined with the calculation of area under the curve (AUC) variables. SPSS 20.0 (SPSS Inc., Chicago, Illinois, USA) was used for all the statistical analyses. P less than 0.05 denoted statistical significance.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
The relationship between Lp(a) and VDT Yin et al. 41
Results Patient characteristics and clinical variations
A total of 232 patients with AIS participated in this study, and completed the 15-day follow-up. The median age of participants was 59 (IQR, 52–71) years and 49.6% of them were women. In total, 42 (18.1%) patients had a history of vein thrombosis. The median time from injury onset to inclusion in the study was 7.2 (IQR, 4.2–15.5) h. Baseline characteristics of patients with AIS are provided in Table 1. The results indicated that the serum Lp(a) levels were higher in AIS patients as compared with normal controls [313 (IQR, 172–553) and 148 (IQR, 57–242) mg/l, respectively, P < 0.0001]. There was a correlation between levels of serum Lp(a) and NIHSS score (r = 0.456, P < 0.0001) and lesion volume (r = 0.427, P < 0.0001). Positive trends between serum Lp(a) and NIHSS score (P < 0.001) and lesion volume (P = 0.001), using ordered logistic regression after multivariate adjustment for possible confounders, were observed. In addition, there was a modest correlation between levels of serum Lp(a) levels and Hs-CRP (r = 0.240, P < 0.001). Statistical analyses revealed no influence of conventional Table 1
vascular risk factors, history of vein thrombosis, sex, WBC, age, prothrombin time, activated partial thromboplastin time, thrombin time, D-dimer, fibrinogen, glucose, and HCY on Lp(a) levels in AIS patients (P > 0.05, respectively).
Lp(a) and DVT
In the present study, 44 patients were diagnosed with DVT at 15-day follow-up, thus the rate of 19.0%. Serum Lp(a) levels were higher in AIS with DVT than in those patients without DVT [656 (IQR, 521–898) vs. 253 (IQR, 143–440) mg/l; P < 0.0001; Fig. 1]. As illustrated in Table 2, in the univariate model, Lp(a) was detected to be continuously associated with an increased risk of DVT (OR 1.007, 95% CI: 1.005–1.009; P < 0.001). Furthermore, increased risk of DVT was found to be correlated with Lp(a) levels greater than or equal to 300 mg/l (unadjusted OR 19.27, 95% CI: 5.76–65.48). Increased risk for DVT associated with Lp(a) levels greater than or equal to 300 mg/l was also found in the multivariate analysis (OR 12.14, 95% CI: 3.08–42.09; P < 0.0001). In addition, history of vein thrombosis, age,
Basal characteristic of patients with acute ischemic stroke Patients with DVT N = 232
Characteristic Female sex (n) Age [median (IQR)] (years) BMI (IQR) (kg/m2) NIHSS score (IQR) Lesion volumes (IQR) (ml) Cigarette smoking (%) Alcohol drinking (%) Hypertension (%) Diabetes at baseline (%) Hypercholesterolemia (%) Atrial fibrillation (%) Coronary heart disease (%) Family history of stroke (%) History of vein thrombosis (%) TOAST classification Large artery (%) Small artery (%) Cardioembolism (%) Other cause (%) Unknown (%) Laboratory findings [median (IQR)] Glucose (mM) White cell count (×109/l) Fibrinogen (g/l) D-Dimer (μg/l) Hs-CRP (mg/dl) PT (s) APTT (s) TT (s) HCY (μM) Lp(a) (mg/l)
59 25.5 7 14
115 (52–71) (24.3–27.0) (4–10) (7–26) 33.6 19.4 74.1 32.3 44.0 17.6 29.3 19.4 18.1
Yes (N = 44) 64 25.8 9 18
18.1 16.8 32.3 12.1 20.7 5.98 7.28 3.45 299 0.93 12.0 26.8 19.6 17.3 313
(5.45–6.32) (6.13–8.77) (2.75–4.69) (204–363) (0.39–1.76) (11.2–12.5) (24.2–31.6) (17.8–22.9) (14.2–19.9) (172–553)
23 (55–76) (24.4–27.5) (6–14) (9–32) 34.1 20.5 72.7 34.1 47.7 22.7 34.1 20.5 25.0
No (N = 188) 56 25.3 6 12
13.6 15.6 40.9 11.4 20.5 5.95 7.31 3.98 442 1.19 12.5 27.2 20.0 18.5 656
(5.44–6.33) (6.10–8.75) (2.92–6.77) (243–674) (0.47–1.98) (11.4–12.9) (24.4–33.2) (18.4–23.2) (15.5–21.4) (521–898)
92 (49–68)* (24.2–26.4) (3–9)* (6–25)* 33.5 19.1 74.5 31.9 43.1 16.5 28.2* 19.1 16.5* 19.1* 17.5 30.3* 12.3 20.8
6.00 7.27 3.28 276 0.76 11.8 26.6 19.1 15.0 253
(5.46–6.32) (6.15–8.80) (2.56–6.15)** (187–332)*** (0.30–1.57)** (11.0–12.3) (24.0–31.2) (17.3–22.4) (13.2–17.6)** (143–440)***
APTT, activated partial thromboplastin time; DVT, deep vein thrombosis; HCY, homocysteine; Hs-CRP, high-sensitivity C-reactive protein; IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; PT, prothrombin time; TT, thrombin time. P values were compared by Mann–Whitney U-test or χ2-test as appropriate. *P < 0.05. **P < 0.01. ***P < 0.001.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
42 NeuroReport 2016, Vol 27 No 1
Fig. 1
Fig. 2
ROC curve
1.0
1200.00 0.8
1000.00 800.00 Sensitivity
Serum levels of LP(a) (mg/l)
1400.00
Z = 9.372, P < 0.0001
600.00 400.00 200.00
0.6
0.4 NIHSS score Hs-CRP LP(a) HCY Reference line
0.00 AIS without DVT (N = 188)
0.2
AIS with DVT (N = 44)
The serum levels of Lp(a) between AIS patients with DVT and without DVT. The horizontal lines indicate medians levels and interquartile ranges (IQRs). P values refer to Mann–Whitney U-tests for differences between groups. The small circles represent the outliers. AIS, acute ischemic stroke; DVT, deep vein thrombosis; Lp(a), lipoprotein (a).
0.0 0.0
0.2
0.4 0.6 1−Specificity
0.8
1.0
HCY, and Hs-CRP can also be seen as DVT predictors (Table 2).
Receiver operator characteristic (ROC) curve demonstrating sensitivity as a function of 1 − specificity for predicting the DVT on the basis of the serum levels of Lp(a) and other markers. DVT, deep vein thrombosis; HCY, homocysteine; Hs-CRP, high-sensitivity-C-reactive protein; Lp(a), lipoprotein (a); NIHSS score, National Institutes of Health Stroke Scale.
Visible by the receiver operating characteristic, the optimal cutoff value of serum Lp(a) levels for predicting DVT was projected to be 420 mg/l, yielding a sensitivity of 88.5% and a specificity of 75.4%, with the AUC of 0.890 (95%CI: 0.844–0.936; P < 0.001; Fig. 2). With an AUC of 0.89 (95% CI: 0.84–0.94), Lp(a) exhibited greater discrimination in predicting DVT compared with HsCRP (AUC, 0.77; 95% CI: 0.69–0.85; P < 0.01), HCY
(AUC, 0.76; 95% CI: 0.68–0.84; P < 0.01), and NIHSS score (AUC, 0.74; 95% CI: 0.66–0.82; P < 0.001). Interestingly, the combined model improved the ability of these biomarkers in predicting DVT risk when compared with these alone (AUC of the combined model, 0.95; 95% CI: 0.92–0.98; P
Elevated lipoprotein (a) levels predict deep vein thrombosis in acute ischemic stroke patients Dongliang Yina,*, Peng Shaoc,* and Yunling Liub Lipoprotein (a) [Lp(a)] plays a crucial role in the pathogenesis of deep vein thrombosis (DVT). The purpose of this study was to investigate whether Lp(a) serum levels at admission could be a risk factor for DVT in Chinese patients with acute ischemic stroke (AIS). A total of 232 patients with AIS were included in the study. The patients were assessed for DVT using colour Doppler ultrasonography. We performed colour Doppler ultrasonography 15 days after the stroke and whenever clinically requested. The value of Lp(a) to predict the DVT was analyzed using logistic regression analysis after adjusting for the possible confounders. In our study, 44 out of the 232 patients (19.0%) were diagnosed with DVT at 15-day follow-up. Serum Lp(a) levels were higher in AIS with DVT than in those patients without DVT [656 (interquartile range, 521–898) mg/l vs. 253 (interquartile range, 143–440) mg/l; P < 0.0001]. Increased risk of DVT associated with Lp(a) levels greater than or equal to 300 mg/l was found in the multivariate analysis [odds ratio 12.14, 95% confidence interval (CI): 3.08–42.09; P < 0.0001]. Visible by the receiver operating characteristic, the optimal cutoff value of serum Lp(a) levels for predicting DVT was projected to be 420 mg/l, yielding a sensitivity of 88.5% and
a specificity of 75.4%. With an area under the curve (AUC) of 0.89 (95% CI, 0.84–0.94), Lp(a) exhibited greater discrimination in predicting DVT compared with Hs-CRP (AUC, 0.77; 95% CI, 0.69–0.85; P < 0.01), HCY (AUC, 0.76; 95% CI, 0.68–0.84; P < 0.01), and NIHSS score (AUC, 0.74; 95% CI, 0.66–0.82; P < 0.001). Elevated serum Lp(a) levels were independent predictors of DVT in AIS patients in China, revealing the critical role played by Lp(a) in the pathogenesis of DVT. NeuroReport 27:39–44 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
Introduction
and recurrent disease [7]. However, Lippi et al. [8] failed to demonstrate a convincing association between Lp(a) and thrombosis in the venous district. Similarly, Vormittag et al. [9] found no association between Lp(a) and VTE.
Lipoprotein (a) [Lp(a)] consists of an LDL-like particle and the specific apolipoprotein (a) [apo(a)], which is covalently bound to the apoB of the LDL-like particle. Numerous epidemiologic studies have identified Lp(a) as a risk factor for atherosclerotic diseases such as coronary heart disease and stroke [1,2]. Lp(a) elevation predicts higher odds of poor functional outcomes for patients with stroke compared with patients with normal levels of Lp(a) [3,4]. In addition, elevated Lp(a) is an independent risk factor for ischemic stroke, and may be especially relevant for young stroke patients [5]. Recently, Lp(a) has been suggested as a risk factor for deep vein thrombosis (DVT). Serum levels of Lp(a) are determined largely by genetic variation in the gene encoding for apo(a), the specific protein component of Lp(a) that is very homologous to plasminogen. NowakGöttl et al. [6] found that Lp(a) greater than 300 mg/l was a risk factor for venous thromboembolism (VTE) during childhood. Another study showed that Lp(a) was an independent risk factor for VTE in adults, suggesting that it may be involved in the pathogenesis of idiopathic 0959-4965 Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.
NeuroReport 2016, 27:39–44 Keywords: acute ischemic stroke, deep vein thrombosis, lipoprotein (a), predictor a Department of Neurology, Jinan 3rd Peoples Hospital, bDepartment of Endocrinology, Qianfoshan Hospital of Shandong Province, Jinan and c Department of Rehabilitation Medicine, Yantaishan Hospital, Yantai, China
Correspondence to Dongliang Yin, No. 1 Wangsheren North Road, Jinan 250101, China Tel: + 86 135 0641 2392; fax: + 86 0531 85853111; e-mail: [email protected] *Dongliang Yin and Peng Shao contributed equally to the writing of this article. Received 30 August 2015 accepted 19 October 2015
There have been many reports describing the association of DVT with stroke [10]. Without VTE prophylaxis, up to 75% of patients with hemiplegia after stroke develop DVT [11]. DVT frequently occurs in the course of ischemic stroke and ∼ 5% of early deaths following stroke are attributed to pulmonary embolism (PE) [12]. However, little is known about the nature of the relationship between Lp(a) and DVT in patients with ischemic stroke. Thus, the purpose of this study was to investigate whether Lp(a) serum levels at admission could be a risk factor for DVT in Chinese patients with acute ischemic stroke (AIS).
Materials and methods The Institutional Review Committee on Human Research of the Jinan third People’s Hospital approved the study protocol. All patients received written DOI: 10.1097/WNR.0000000000000496
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
40 NeuroReport 2016, Vol 27 No 1
information concerning the background and procedures of the study, and the patients or their relatives gave written informed consent before participating in the study. From January 2013 to December 2014, the study population comprised 232 consecutive patients with an AIS diagnosis who had been referred to the Jinan third People’s Hospital. Patients were diagnosed according to the WHO criteria and with symptom onset within 48 h. Brain imaging (either computed tomography or MRI) was performed routinely within 24 h after admission. Exclusion criteria were extradural or subdural hematomas; hemorrhagic transformation of ischemic infarct; a DVT within the previous 3 months; formerly known abnormalities of any coagulation factor; use of unfractionated heparin, low-molecular-weight heparin, or any form of oral anticoagulant in the past month; life expectancy less than 3 months; malignant tumor, surgery, or trauma during the preceding 2 months; severe edema; serious infections at study enrollment; blood not drawn within 24 h after hospital admission; presence of diverse medical illnesses or current medications that influence serum Lp (a) levels, and incomplete workups for cerebrovascular status; inaccessible to follow-up; or failure to give consent. The control cases (N = 100) were of similar age and sex distribution as the AIS patients. They had no known diseases and were not using any medication. At baseline, demographic data (age and sex), BMI, and history of risk factors (hypertension, diabetes mellitus, atrial fibrillation, hyperlipidemia, smoking habit, and alcohol abuse) were obtained. For each patient, the time from stroke onset to admission was also recorded. At admission, the neurological deficit was quantified using the National Institutes of Health Stroke Scale (NIHSS) [13]. AIS was classified according to the TOAST system [14], which distinguishes large-artery arteriosclerosis, cardioembolism, small-artery occlusion, other causative factor, and undetermined causative factor. MRI was performed within 24 h after admission to assess the site, cause, and size of the brain infarct. MRI was performed using a stroke protocol, including T1-weighted, T2-weighted, and diffusion-weighted imaging (DWI) sequences, and a magnetic resonance angiography. DWI lesion volumes were 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) [15]. All patients received treatment according to current guidelines. The included patients were assessed for DVT using colour Doppler ultrasonography, regardless of whether they were symptomatic or asymptomatic. We performed colour Doppler ultrasonography 15 days after AIS and whenever clinically requested. Ultrasonography, using real-time B-mode with compression, was carried out with
a 7.5 or a 5.0 MHz transducer. Two areas of the leg were examined: the common femoral vein at the inguinal ligament and the popliteal vein at the knee-joint line traced down to the point of the trifurcation of the calf veins. Veins were only scanned in the transverse plane. We used Vivid 7 Dimension (GE, Miami, Florida, USA) with the 7–10 Hz linear probe. The diagnosis of DVT was based either on the presence of a noncompressible segment [compression ultrasound test (CUS)] or the flow impairment on color Doppler imaging. This method is recognized as sufficiently sensitive and specific, especially for proximal DVT detection [16]. Fasting venous blood samples were collected from all participants in vacutainer tubes, and were quickly centrifuged to avoid glycolysis. Serum samples were kept at − 80°C until assay. Lp(a) levels were measured using immunoturbidimetric method by Olympus 2700 (Olympus, Tokyo, Japan). The lower detection limit was 30 mg/l and the line range was 30–1200 mg/l. The intraassay coefficient of variation (CV) and interassay CV were 1.4–2.2% and 1.6–2.5%, respectively. The median serum Lp(a) level in 100 healthy controls was 148 mg/l and the 97.5 percentile was 375 mg/l. Routine blood biomarkers – for instance, fibrinogen, prothrombin time, activated partial thromboplastin time, thrombin time, glucose, high-sensitivity C-reactive protein (Hs-CRP), and homocysteine (HCY) were tested using standard detection methods. Data were presented as the percentage (number) or frequencies for discrete variables, and medians [interquartile range (IQR)] for continuous variables. Values of the measured parameters were checked for normal distribution by using the Kolmogorov–Smirnov test before statistical analysis. Mann–Whitney U-test was carried out for the comparison between groups for continuous variables; in addition, χ2-test was used for the comparison of proportions between groups. Spearman’s rank correlation assessed the correlations among laboratory parameters. To investigate whether Lp(a) allowed predicting of DVT in AIS, different statistical methods were used. First, whether a high level of Lp(a) was a risk factor for DVT was confirmed by calculating the odds ratio (OR) and the 95% confidence interval (CI). We used 300 mg/l as the cutoff point. To adjust for possible confounders – for example, age, sex, BMI, conventional vascular risk factors, history of vein thrombosis, and other blood biomarkers – we used a logistic regression model. Second, for further comparison of the prognostic values of different scores from different predictive models, receiver operating characteristic was graphed to determine the optimal cutoff values, combined with the calculation of area under the curve (AUC) variables. SPSS 20.0 (SPSS Inc., Chicago, Illinois, USA) was used for all the statistical analyses. P less than 0.05 denoted statistical significance.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
The relationship between Lp(a) and VDT Yin et al. 41
Results Patient characteristics and clinical variations
A total of 232 patients with AIS participated in this study, and completed the 15-day follow-up. The median age of participants was 59 (IQR, 52–71) years and 49.6% of them were women. In total, 42 (18.1%) patients had a history of vein thrombosis. The median time from injury onset to inclusion in the study was 7.2 (IQR, 4.2–15.5) h. Baseline characteristics of patients with AIS are provided in Table 1. The results indicated that the serum Lp(a) levels were higher in AIS patients as compared with normal controls [313 (IQR, 172–553) and 148 (IQR, 57–242) mg/l, respectively, P < 0.0001]. There was a correlation between levels of serum Lp(a) and NIHSS score (r = 0.456, P < 0.0001) and lesion volume (r = 0.427, P < 0.0001). Positive trends between serum Lp(a) and NIHSS score (P < 0.001) and lesion volume (P = 0.001), using ordered logistic regression after multivariate adjustment for possible confounders, were observed. In addition, there was a modest correlation between levels of serum Lp(a) levels and Hs-CRP (r = 0.240, P < 0.001). Statistical analyses revealed no influence of conventional Table 1
vascular risk factors, history of vein thrombosis, sex, WBC, age, prothrombin time, activated partial thromboplastin time, thrombin time, D-dimer, fibrinogen, glucose, and HCY on Lp(a) levels in AIS patients (P > 0.05, respectively).
Lp(a) and DVT
In the present study, 44 patients were diagnosed with DVT at 15-day follow-up, thus the rate of 19.0%. Serum Lp(a) levels were higher in AIS with DVT than in those patients without DVT [656 (IQR, 521–898) vs. 253 (IQR, 143–440) mg/l; P < 0.0001; Fig. 1]. As illustrated in Table 2, in the univariate model, Lp(a) was detected to be continuously associated with an increased risk of DVT (OR 1.007, 95% CI: 1.005–1.009; P < 0.001). Furthermore, increased risk of DVT was found to be correlated with Lp(a) levels greater than or equal to 300 mg/l (unadjusted OR 19.27, 95% CI: 5.76–65.48). Increased risk for DVT associated with Lp(a) levels greater than or equal to 300 mg/l was also found in the multivariate analysis (OR 12.14, 95% CI: 3.08–42.09; P < 0.0001). In addition, history of vein thrombosis, age,
Basal characteristic of patients with acute ischemic stroke Patients with DVT N = 232
Characteristic Female sex (n) Age [median (IQR)] (years) BMI (IQR) (kg/m2) NIHSS score (IQR) Lesion volumes (IQR) (ml) Cigarette smoking (%) Alcohol drinking (%) Hypertension (%) Diabetes at baseline (%) Hypercholesterolemia (%) Atrial fibrillation (%) Coronary heart disease (%) Family history of stroke (%) History of vein thrombosis (%) TOAST classification Large artery (%) Small artery (%) Cardioembolism (%) Other cause (%) Unknown (%) Laboratory findings [median (IQR)] Glucose (mM) White cell count (×109/l) Fibrinogen (g/l) D-Dimer (μg/l) Hs-CRP (mg/dl) PT (s) APTT (s) TT (s) HCY (μM) Lp(a) (mg/l)
59 25.5 7 14
115 (52–71) (24.3–27.0) (4–10) (7–26) 33.6 19.4 74.1 32.3 44.0 17.6 29.3 19.4 18.1
Yes (N = 44) 64 25.8 9 18
18.1 16.8 32.3 12.1 20.7 5.98 7.28 3.45 299 0.93 12.0 26.8 19.6 17.3 313
(5.45–6.32) (6.13–8.77) (2.75–4.69) (204–363) (0.39–1.76) (11.2–12.5) (24.2–31.6) (17.8–22.9) (14.2–19.9) (172–553)
23 (55–76) (24.4–27.5) (6–14) (9–32) 34.1 20.5 72.7 34.1 47.7 22.7 34.1 20.5 25.0
No (N = 188) 56 25.3 6 12
13.6 15.6 40.9 11.4 20.5 5.95 7.31 3.98 442 1.19 12.5 27.2 20.0 18.5 656
(5.44–6.33) (6.10–8.75) (2.92–6.77) (243–674) (0.47–1.98) (11.4–12.9) (24.4–33.2) (18.4–23.2) (15.5–21.4) (521–898)
92 (49–68)* (24.2–26.4) (3–9)* (6–25)* 33.5 19.1 74.5 31.9 43.1 16.5 28.2* 19.1 16.5* 19.1* 17.5 30.3* 12.3 20.8
6.00 7.27 3.28 276 0.76 11.8 26.6 19.1 15.0 253
(5.46–6.32) (6.15–8.80) (2.56–6.15)** (187–332)*** (0.30–1.57)** (11.0–12.3) (24.0–31.2) (17.3–22.4) (13.2–17.6)** (143–440)***
APTT, activated partial thromboplastin time; DVT, deep vein thrombosis; HCY, homocysteine; Hs-CRP, high-sensitivity C-reactive protein; IQR, interquartile range; NIHSS, National Institutes of Health Stroke Scale; PT, prothrombin time; TT, thrombin time. P values were compared by Mann–Whitney U-test or χ2-test as appropriate. *P < 0.05. **P < 0.01. ***P < 0.001.
Copyright r 2016 Wolters Kluwer Health, Inc. All rights reserved.
42 NeuroReport 2016, Vol 27 No 1
Fig. 1
Fig. 2
ROC curve
1.0
1200.00 0.8
1000.00 800.00 Sensitivity
Serum levels of LP(a) (mg/l)
1400.00
Z = 9.372, P < 0.0001
600.00 400.00 200.00
0.6
0.4 NIHSS score Hs-CRP LP(a) HCY Reference line
0.00 AIS without DVT (N = 188)
0.2
AIS with DVT (N = 44)
The serum levels of Lp(a) between AIS patients with DVT and without DVT. The horizontal lines indicate medians levels and interquartile ranges (IQRs). P values refer to Mann–Whitney U-tests for differences between groups. The small circles represent the outliers. AIS, acute ischemic stroke; DVT, deep vein thrombosis; Lp(a), lipoprotein (a).
0.0 0.0
0.2
0.4 0.6 1−Specificity
0.8
1.0
HCY, and Hs-CRP can also be seen as DVT predictors (Table 2).
Receiver operator characteristic (ROC) curve demonstrating sensitivity as a function of 1 − specificity for predicting the DVT on the basis of the serum levels of Lp(a) and other markers. DVT, deep vein thrombosis; HCY, homocysteine; Hs-CRP, high-sensitivity-C-reactive protein; Lp(a), lipoprotein (a); NIHSS score, National Institutes of Health Stroke Scale.
Visible by the receiver operating characteristic, the optimal cutoff value of serum Lp(a) levels for predicting DVT was projected to be 420 mg/l, yielding a sensitivity of 88.5% and a specificity of 75.4%, with the AUC of 0.890 (95%CI: 0.844–0.936; P < 0.001; Fig. 2). With an AUC of 0.89 (95% CI: 0.84–0.94), Lp(a) exhibited greater discrimination in predicting DVT compared with HsCRP (AUC, 0.77; 95% CI: 0.69–0.85; P < 0.01), HCY
(AUC, 0.76; 95% CI: 0.68–0.84; P < 0.01), and NIHSS score (AUC, 0.74; 95% CI: 0.66–0.82; P < 0.001). Interestingly, the combined model improved the ability of these biomarkers in predicting DVT risk when compared with these alone (AUC of the combined model, 0.95; 95% CI: 0.92–0.98; P
