Asymmetric dimethylarginine (ADMA) as a risk marker for stroke and TIA in a Swedish population

Atherosclerosis 185 (2006) 271–277

Asymmetric dimethylarginine (ADMA) as a risk marker for stroke and TIA in a Swedish population P. Wanby a,b,∗∗ , T. Teerlink c , L. Brudin d , L. Brattstr¨om a,b , I. Nilsson e , P. Palmqvist a,b , M. Carlsson a,b,e,∗ a

Department of Internal Medicine, County Hospital of Kalmar, SE-391 85 Kalmar, Sweden b Faculty of Health Sciences, Linkoping University, Linkoping, Sweden c Department of Clinical Chemistry, VU University Medical Center, Amsterdam, Netherlands d Department of Clinical Physiology, County Hospital of Kalmar, Sweden e Department of Clinical Chemistry, County Hospital of Kalmar, SE-391 85 Kalmar, Sweden Received 13 April 2005; received in revised form 8 June 2005; accepted 21 June 2005 Available online 1 August 2005

Abstract Asymmetric dimethylarginine (ADMA), an endogenous inhibitor of nitric oxide synthase, has been shown to be involved in the pathogenesis of atherosclerosis. The present study was initiated to investigate the role of ADMA as a risk marker of acute cerebrovascular disease (CVD). We examined 363 CVD patients and 48 controls. The ADMA concentration (mean ± S.D., ␮mol/L) in controls was 0.50 ± 0.06. Compared to controls, increased concentrations of ADMA were observed in cardio-embolic infarction (0.55 ± 0.08; p < 0.001; n = 71), and TIA (0.54 ± 0.05; p < 0.001; n = 31), but not in non-cardio-embolic infarction (0.51 ± 0.07; p = 0.56; n = 239) and haemorrhagic stroke (0.51 ± 0.11; p = 0.77; n = 22). In multivariate logistic regression models, CVD increased across quartiles of ADMA in all subgroups, but this association was only significant in the TIA group (odds ratio for highest versus lowest quartile 13.1; 95% CI: 2.9–58.6; p trend 0.001) A decreased arginine/ADMA ratio was significantly associated with CVD in the entire study population (p < 0.01). Our results indicate that ADMA is a weak independent marker for acute stroke and a strong marker for TIA and that relative arginine deficiency, measured as the l-arginine/ADMA ratio, is present in acute CVD. © 2005 Elsevier Ireland Ltd. All rights reserved. Keywords: Asymmetric dimethylarginine; Arginine; Atherosclerosis; Stroke; TIA

1. Introduction Longitudinal studies have shown that hypertension, insulin resistance, hypertriglyceridaemia, and low HDLcholesterol are strongly associated with ischemic stroke [1]. In each of these disorders there is an impairment of endothelial vasodilatory function due to diminished nitric oxide (NO) synthesis [2,3]. Endothelium derived NO is a potent vasodilator and antiatherogenic agent [4]. The reduction in NO bioavailability associated with atherosclerosis and its ∗

Corresponding author. Tel.: +46 480 81456; fax: +46 480 81025. Corresponding author. Tel.: +46 480 448204; fax +46 480 81998. E-mail addresses: [email protected] (P. Wanby), [email protected] (M. Carlsson). ∗∗

0021-9150/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.atherosclerosis.2005.06.033

risk factors appears to be influenced by the elevation in plasma of NG ,NG -dimethyl-l-arginine (asymmetric dimethylarginine, ADMA). ADMA is derived from the catabolism of proteins containing methylated arginine residues [5] and ADMA is an endogenous competitive inhibitor of NO synthase [6]. Increased levels of ADMA have been correlated to carotid intima-media thickness [3] and other investigators have observed a relationship between high ADMA levels and coronary heart disease [7,8]. In a study from South Korea on 52 patients with ischemic stroke [9], plasma levels of ADMA were elevated, but there have been no further reports on the association between ADMA and stroke. The present study was initiated to investigate the role of ADMA as a risk marker in acute cerebrovascular disease (CVD), also examining ADMA levels in subgroups of CVD in a

272

P. Wanby et al. / Atherosclerosis 185 (2006) 271–277

Swedish Caucasian population. Since l-arginine and symmetric dimethylarginine (SDMA) are related to ADMA and of interest in the NO synthesis, we also examined these amino acids.

2. Materials and methods

≤0.9 mmol/L for males and ≤1.0 mmol/L for women. Characteristics of the control group are given in Table 1. One hundred and seventy-one of the 363 patients with acute CVD, but none of the control subjects, were investigated using ultrasound of the carotids. All participants gave informed consent to participate in the study, which received the approval of the Ethical Committee of Link¨oping University, Sweden.

2.1. Subjects 2.2. Phenotypic characterisation Between January 2001 and December 2003, 442 patients initially diagnosed with acute cerebrovascular disease, i.e. stroke or transient ischemic attack (TIA), were recruited upon their admission to the acute stroke unit of the County Hospital of Kalmar, Sweden. According to the World Health Organisation criteria [10] stroke was defined as a neurological deficit observed by a physician and persisting for more than 24 h, without other disease explaining the symptoms and a TIA was diagnosed if the symptoms lasted less than 24 h. The medical history of all patients concerning previous stroke/TIA, diabetes, hypertension and current cigarette smoking was obtained from the patient or a relative following a written nurse-administrated standardised questionnaire. A history of hypertension was defined as the use of antihypertensive drugs or blood pressure more than 140/90 on at least two separate occasions and diabetes mellitus as the use of anti-diabetic drugs or diet treatment only or a fasting plasma glucose value ≥7.0 mmol/L. If the patient had an ongoing atrial fibrillation on admission, a history of atrial fibrillation or cardio-valvular disease, the stroke was considered to be due to a cardiac embolism unless the CT scan showed haemorrhage. Patients with subarachnoid haemorrhage were not included in the study. A thorough examination of medical records further identified patients with possible underlying causes for ischemic stroke and patients with a possible lipid altering treatment. Patients from these two categories were excluded from the study. These causes or treatments consisted of patients with a presence of systemic malignancy (n = 8), SLE (n = 1), a history of migraine (n = 4), patients with a kidney transplantation (n = 2), and patients with oestrogen (n = 1), thyroxine (n = 17) and lipid-lowering treatment (n = 25). In addition patients who retrospectively were found to have been misdiagnosed as patients with stroke were also excluded from the study (n = 21). Therefore, 363 patients (239 with a non-cardio-embolic infarction, 71 with a cardio-embolic infarction, 22 with an intracerebral haemorrhage and 31 with TIA) were evaluated. Forty-eight patients, 36 undergoing elective prosthetic surgery and 12 from the influenza vaccination unit of our hospital, all subjects without cardiovascular disease and without any feature of the metabolic syndrome (modified from the WHO classification [11]) were invited to participate in the study as controls. None of the controls had diabetes mellitus or a fasting plasma glucose ≥6.1 mmol/L, BMI > 30 kg/m2 , systolic blood pressure ≥160 mmHg or diastolic blood pressure ≥90 mmHg, fasting triglycerides ≥1.7 mmol/L or an HDL-cholesterol,

At approximately 7 a.m. in the morning during the first days after admission to hospital fasting venous blood samples were drawn for the measurements of glucose, HbA1c, creatinine, triglycerides, total-cholesterol, LDL- and HDLcholesterol, free fatty acids (FFA), ADMA, symmetric dimethylarginine (SDMA), and l-arginine. Blood pressure was measured with a mercury sphygmometer with the subject in the supine position. Weight and height were measured with the subject in light clothing without shoes and body mass index (BMI) was calculated as kilograms per square meter. In addition, diagnostic evaluation including a 12-lead electrocardiogram and brain computed tomography (CT) were performed in all patients. In a subset of patients, the presence of internal carotid artery (ICA) stenosis was non-invasively determined using a Duplex sonography combining B-mode imaging, colour flow and pulsed Doppler spectrum analysis (Sonos 5000 HPTM or Acuson SequoiaTM 512). A consultant clinical physiologist (MD) performed examinations using an 8 MHz probe with the patients in supine position. The Doppler angle was chosen as close to 60◦ as possible. Maximum systolic peak velocities above the normal limit 1.05 m/s [12] were classified as stenosis of the ICA (ranging from 50 to 100%; degrees of stenosis less than 50% are not possible to detect with this method). Both the right and left carotid in each subject were investigated and the highest ICA-stenosis value of the two was used when the subjects were classified as normal or stenotic. 2.3. Assays HbA1c was measured using a high performance liquid chromatography (HPLC) method, basically as described in [13]. Fasting plasma glucose, triglycerides, plasma total cholesterol, and HDL-cholesterol concentrations were analysed with commercially available kits using Cobas Integra 700 (Roche). Fasting plasma LDL-cholesterol was calculated with the use of Friedewald’s formula. Samples for analysis of FFA were obtained in tubes without additives or EDTA as anticoagulant and kept on ice for 30 min prior to centrifugation. The samples were thereafter kept at −70 ◦ C until analysis. FFA in serum were measured using an acyl-CoA oxidase-based colorimetric kit (Wako Chemicals Inc., Richmond, VA, USA), on a Cobas Mira (Roche). Plasma concentrations of ADMA, SDMA and l-arginine were assessed using HPLC as described previously [14].

P. Wanby et al. / Atherosclerosis 185 (2006) 271–277

273

Table 1 Clinical characteristics of subjects

Gender (females/males) Age (years) Smokers (%) History of hypertension (%) History of stroke/TIA (%) Prevalence of diabetes (%) SBP (mmHg) DBP (mmHg) BMI (kg/m2 ) Glucose (mmol/L) HbA1c (%) Creatinine (␮mol/L) Triglycerides (mmol/L) Total cholesterol (mmol/L) LDL-cholesterol (mmol/L) HDL-cholesterol (mmol/L) FFA (mmol/L) ADMA (␮mol/L) SDMA (␮mol/L) l-Arginine (␮mol/L) l-Arginine/ADMA

Non-cardio-embolic infarction, n = 239

Cardio-embolic infarction, n = 71

Haemorrhagic stroke, n = 22

TIA, n = 31

Controls, n = 48

p-Value

94/145 69.9 ± 11.2 19 38 18 25 160 ± 27 86 ± 13 26.3 ± 4.3 6.5 ± 3.1 4.8 (3.1–11.4) 89 ± 26 1.4 (0.4–5.0) 6.0 ± 1.4 3.8 ± 1.2 1.4 ± 0.5 0.55 (0.06–1.92) 0.51 ± 0.07 (0.25–0.81) 0.56 (0.34–2.03) 90 (11–145) 160 ± 60

32/39 77.1 ± 8.0 7 42 22 25 154 ± 24 84 ± 13 25.5 ± 5 6.4 ± 1.9 4.8 (3.7–13.1) 95 ± 26 1.2 (0.5–4.8) 5.5 ± 1.5 3.4 ± 1.3 1.4 ± 0.4 0.60 (0.28–1.2) 0.55 ± 0.08 (0.39–0.78) 0.66 (0.33–2.00) 77 (38–123) 140 ± 34

9/13 72.8 ± 12.0 9 50 27 23 170 ± 17 89 ± 10 24.9 ± 3.3 6.4 ± 1.6 4.5 (4.0–7.7) 91 ± 21 1.2 (0.6–3.1) 5.4 ± 1.1 3.4 ± 0.9 1.6 ± 0.7 0.54 (0.3–0.89) 0.51 ± 0.11 (0.36–0.78) 0.59 (0.31–1.06) 70 (55–107) 140 ± 31

11/20 70.6 ± 9.9 19 39 26 10 151 ± 26 85 ± 11 26.4 ± 3.5 5.4 ± 0.7 4.6 (3.2–5.4) 88 ± 19 1.3 (0.5–2.5) 5.8 ± 1.0 3.8 ± 0.9 1.4 ± 0.4 0.51(0.19–0.94) 0.54 ± 0.05 (0.46–0.67) 0.58 (0.47–1.07) 81.8 (47–108) 150 ± 30

25/23 71.2 ± 9.1 2 11 0 0 142 ± 14 81 ± 7 25.2 ± 2.9 4.8 ± 1.2 4.5 (3.7–6.2) 76 ± 16 1.1 (0.1–1.7) 5.7 ± 1.0 3.7 ± 0.9 1.7 ± 0.4 0.48 (0.13–0.93) 0.50 ± 0.06 (0.35–0.63) 0.56 (0.40–0.87) 86 (53–123) 173 ± 34

0.71 <0.001 <0.01 <0.01 <0.01 <0.01 <0.001 0.05 0.29 <0.001 <0.001 <0.01 <0.01 0.15 0.97 <0.01 0.08 <0.001* <0.001† 0.04‡ <0.01

Data presented as mean ± S.D. or as median and range. Since P-ADMA was nearly normally distributed it is presented as mean ± S.D., but the range is also given. SBP and DBP are systolic and diastolic blood pressure, respectively. p-Values (right column) show differences between any of the groups. * Based on ANOVA for log ADMA (the log-transformations of ADMA values did not significantly alter the outcome of the analysis). † Based on ANOVA for log SDMA. ‡ Based on ANOVA for log l-arginine.

Lower limits of quantification were 0.08 ␮mol/L for arginine and 0.01 ␮mol/L for ADMA and SDMA. The intra-assay and inter-assay coefficients of variation for all analytes were <1.2 and <3%, respectively. Since ADMA acts as a competitive inhibitor of NO synthase with l-arginine as its substrate, the ratio of plasma l-arginine/ADMA has been proposed as a determinant of endothelium-dependent dilation. We therefore also calculated the l-arginine/ADMA ratio. 2.4. Statistical analysis All statistical analyses were performed with STATISTICA (version 6.0, StatSoft® , Tulsa, USA). Variables not normally distributed are given as median (range). The results for continuous normally distributed variables are given as mean ± S.D. and for categorical variables as percentages. For categorical variables the Chi-squared test was applied for comparison between groups. The log values of ADMA and SDMA were normally distributed and ANOVA was, as for other continuous variables, applied, followed by Duncan’s test in case of significance. For variables not normally distributed (HbA1c, creatinine, triglycerides and FFA) group differences were analysed using Kruskal–Wallis nonparametric test, followed by Mann–Whitney’s test in case of significance. Pearson’s correlation coefficient was calculated to examine a possible correlation between ADMA and SDMA on the one hand and other continuous variables on

the other (using log values for log-normally distributed variables). The relationships between ADMA and other possible risk factors and CVD and its subgroups were analysed by multivariate logistic regression. The following independent variables (possible risk factors) were included in the model: gender (male/female), age (categorized in quartiles), systolic blood pressure (SBP) and HbA1c (categorized in two equally sized groups using the median value, < or ≥160 mmHg for SBP and < or ≥4.7% for HbA1c), as well as serum concentrations of creatinine, triglycerides, cholesterol, and ADMA (all categorized in quartiles). Similar models for multivariate analysis investigating SDMA, l-arginine/ADMA ratio and l-arginine (all categorized in quartiles) as independent variables were used. Since almost all control patients were non-smokers, it was not possible to include this variable in the multivariate analysis model. p-Values ≤0.05 were considered significant.

3. Results 3.1. Clinical characteristics and ADMA concentrations Of the patients included (n = 363, age 71.6 ± 10.6 years), 66% were classified as having a non-cardio-embolic infarction, 20% had cardio-embolic infarction, 6% had haemorrhagic stroke, and 8% had a TIA. Of all the patients 16%

P. Wanby et al. / Atherosclerosis 185 (2006) 271–277

274

were current smokers, 39% had a history of hypertension, 20% had suffered a previous stroke or TIA and 24% had diabetes (Table 1). There were no differences between the CVD subgroups regarding gender, smoking habits, earlier hypertension, earlier stroke, presence of diabetes or BMI. Patients with cardioembolic infarction were older (p < 0.001), and patients with haemorrhagic stroke had higher systolic blood pressure than the patients in other subgroups (p = 0.01). ADMA concentrations in the overall group of patients with acute CVD were increased compared with controls (p < 0.01). In the subgroup analysis, ADMA was increased in cardio-embolic infarction (p < 0.001) and TIA (p < 0.001) compared with controls but not in non-cardio-embolic infarction (p = 0.56) or in haemorrhagic stroke (p = 0.77). In the whole group of CVD patients, ADMA was not associated with previous stroke/TIA (p = 0.08), but when patients with haemorrhagic stroke were excluded from the CVD patients such an association was seen (p = 0.01). Compared with controls, there were significantly increased plasma levels of SDMA in the entire CVD group (p = 0.01) and in the subgroup of cardio-embolic infarction (<0.001). SDMA was associated with previous stroke/TIA (p = 0.01). 3.2. Univariate associations of plasma ADMA and SDMA in acute CVD In the patients with CVD, plasma concentrations of ADMA were positively associated with age, renal function (creatinine), SDMA and l-arginine, but not with traditional CVD risk factors such as blood pressure, BMI, glucose or

Table 2 Correlation of plasma ADMA and SDMA levels respectively in acute CVD (n = 363) with stroke risk factors in simple linear regression ADMA

Age (years) SBP (mmHg) DBP (mmHg) BMI (kg/m2 ) Glucose (mmol/L) Hba1c (%) Creatinine Triglycerides (mmol/L) Total Cholesterol (mmol/L) LDL-cholesterol (mmol/L) HDL-cholesterol (mmol/L) FFA (mmol/L) ADMA (␮mol/L) SDMA (␮mol/L) l-Arginine (␮mol/L) l-Arginine/ADMA

SDMA

r

p-Value

r

p-Value

0.26 0.06 0.02 −0.03 −0.02 −0.12 0.19 −0.007 −0.04 −0.004 −0.06 −0.11 – 0.52 0.21 –

<0.001 0.29 0.76 0.51 0.70 0.04 <0.01 0.90 0.56 0.94 0.28 0.07 – <0.001 <0.001 –

0.37 0.10 0.05 −0.08 0.01 −0.04 0.26 −0.003 −0.03 −0.02 −0.03 0.04 0.52 – −0.04 −0.26

<0.001 0.05 0.31 0.11 0.84 0.43 <0.001 0.96 0.60 0.73 0.59 0.54 <0.001 – 0.44 <0.001

lipids (Table 2). There was a slight negative correlation between ADMA and HbA1c. SDMA was positively associated with creatinine concentration (r = 0.26) but also with age, systolic blood pressure and ADMA. SDMA was negatively associated with the l-arginine/ADMA ratio (Table 2). 3.3. Odds ratios of CVD across quartiles of ADMA concentrations The results of the multivariate logistic regression analysis are given in Table 3. In the entire group of patients,

Table 3 Results of the separate logistic regression models for ADMA, l-arginine, l-arginine/ADMA ratio and SDMA, in acute CVD and in each of its subgroups (the multivariate models were corrected for confounding by age, HbA1c, triglycerides and systolic blood pressure) CVD (n = 363)

Non-cardio-embolic infarction (n = 239)

Haemorrhagic stroke (n = 22)

OR (95% CI)

p-Value

OR (95% CI)

0.09

1.00 1.08 (0.79–1.50) 1.18 (0.62–2.24) 1.28 (0.49–3.35)

0.61

1.00 1.48 (0.96–2.28) 2.19 (0.92–5.20) 3.23 (0.88–11.9)

0.08

1.00 1.13 (0.65–1.95) 1.27 (0.42–3.80) 1.43 (0.27–7.42)

1.00 0.67 (0.49–0.91) 0.45 (0.24–0.83) 0.30 (0.12–0.75)

0.03

1.00 0.72 (0.52–0.99) 0.51 (0.27–0.98) 0.37 (0.14–0.97)

0.04

1.00 0.53 (0.35–0.80) 0.28 (0.12–0.64) 0.15 (0.04–0.51)

<0.01

l-Arginine/ADMA <129 129–158 158–182 >182

1.00 0.65 (0.47–0.89) 0.43 (0.23–0.81) 0.28 (0.11–0.72)

<0.01

1.00 0.77 (0.56–1.07) 0.60 (0.31–1.15) 0.46 (0.17–1.23)

0.12

1.00 0.48 (0.31–0.75) 0.23 (0.10–0.57) 0.11 (0.03–0.43)

SDMA <0.50 ␮mol/L 0.50–0.58 ␮mol/L 0.58–0.69 ␮mol/L >0.69 ␮mol/L

1.00 1.28 (0.94–1.75) 1.65 (0.89–3.05) 2.12 (0.84–5.34)

0.11

1.00 1.08 (0.77–1.50) 1.16 (0.59–2.26) 1.25 (0.46–3.40)

0.66

1.00 2.10 (1.40–3.15) 4.39 (1.94–9.90) 9.21 (2.71–31.1)

OR (95% CI) ADMA <0.48 ␮mol/L 0.48–0.52 ␮mol/L 0.52–0.56 ␮mol/L >0.56 ␮mol/L

1.00 1.29 (0.96–1.74) 1.68 (0.92–3.06) 2.17 (0.88–5.35)

l-Arginine <66.9 ␮mol/L 66.9–80.5 ␮mol/L 80.5–92.5 ␮mol/L >92.5 ␮mol/L

p-Value

Cardio-embolic infarction (n = 71) p-Value

TIA (n = 31) OR (95% CI)

p-Value

0.66

1.00 2.35 (1.43–3.88) 5.54 (2.04–15.1) 13.1 (2.91–58.6)

0.001

1.00 0.31 (0.14–0.69) 0.09 (0.02–0.47) 0.03 (0.002–0.33)

<0.01

1.00 0.82 (0.49–1.37) 0.67 (0.24–1.88) 0.55 (0.12–2.58)

0.45

<0.01

1.00 0.40 (0.19–0.83) 0.16 (0.04–0.68) 0.06 (0.007–0.56)

0.01

1.00 0.56 (0.33–0.95) 0.31 (0.11–0.91) 0.17 (0.04–0.86)

0.03

<0.001

1.00 1.22 (0.69–2.18) 1.50 (0.47–4.76) 1.83 (0.32–10.4)

0.49

1.00 1.25 (0.76–2.08) 1.57 (0.57–4.31) 1.97 (0.43–8.94)

0.38

OR (95% CI)

p-Value

Inclusion of gender, creatinine and cholesterol levels did not significantly alter the outcome of the analysis. OR is odds ratio with 95% confidence interval within parenthesis.

P. Wanby et al. / Atherosclerosis 185 (2006) 271–277

increased systolic blood pressure and increased HbAlc and triglyceride levels were independent risk factors for CVD, whereas plasma levels of ADMA did not reach statistical significance. In the CVD subgroups, increased age, systolic blood pressure, and HbAlc were independent risk factors for cardio-embolic infarction. Systolic blood pressure, HbA1c and triglyceride levels were independent risk factors for noncardio-embolic infarction. In the haemorrhagic stroke group systolic blood pressure was the only independent variable that contributed to the model. In these subgroups the odds ratios increased across quartiles of plasma ADMA concentrations, but this association was weak and non-significant in the noncardio-embolic infarction and haemorrhagic stroke groups and borderline significant (p = 0.08) in the cardio-embolic infarction group. In contrast, in the TIA group, ADMA was the only independent variable that significantly contributed to the model, with an odds ratio of 2.35/quartile (p = 0.001). Decreased l-arginine concentrations were significantly associated with acute CVD (p = 0.03), cardio-embolic infarction (p < 0.01), haemorrhagic stroke (p < 0.01), and non-cardioembolic infarction (p = 0.04), but not with TIA (p = 0.45). A decreased l-arginine/ADMA ratio was significantly associated with acute CVD (p < 0.01), cardio-embolic infarction (p < 0.01), haemorrhagic stroke (p = 0.01) and TIA (p = 0.03), but not with non-cardio-embolic infarction (p = 0.12). Elevated SDMA concentrations were not significantly associated with overall CVD (p = 0.11). In the subgroup analysis, there was a significant positive multivariate relationship between SDMA and cardio-embolic infarction (p < 0.001), but not between SDMA and non-cardio-embolic infarction (p = 0.66), haemorrhagic stroke (p = 0.49), or TIA (p = 0.38). 3.4. ICA stenosis in CVD patients Of the patients with acute CVD, 171 had undergone investigation with ultrasound. Of these patients, 159 were found not to suffer from ICA stenosis and only 12 patients were diagnosed with an ICA stenosis of more than 50%. Plasma ADMA concentrations in the group of patients without ICA stenosis (0.52 ± 0.08 (0.25–0.78)) did not differ significantly from the group with ICA stenosis (0.54 ± 0.09 (0.37–0.70)) (p = 0.53). In addition, SDMA (p = 0.63) and l-arginine (p = 0.10) concentrations as well as the l-arginine/ADMA ratio (p = 0.10) were not different between patients with and without ICA stenosis.

4. Discussion This is the first study investigating plasma levels of ADMA in different subtypes of acute CVD. A previous study [9], conducted in an Asian population, demonstrated elevated ADMA levels in patients with ischemic stroke. In the present study we observed slightly increased ADMA levels in patients with earlier ischemic stroke or TIA. In a multiple logistic regression analysis, increased ADMA concentration was strongly

275

and independently associated with TIA. A tendency towards increased risk associated with elevated concentrations of ADMA was seen also in cardio-embolic infarction and in the group of CVD as a whole but did not reach statistical significance. Data from experimental studies suggest that ADMA inhibits vascular NO production at concentrations found in pathophysiological conditions. In addition to being a vasodilator, NO also inhibits platelet adhesion and aggregation, attenuates monocyte adhesion and infiltration, suppresses myointimal hyperplasia and reduces oxidative stress [15]. A decrease in NO availability could therefore enhance atherosclerosis and local inflammation of the vessel wall, which may play a critical role in plaque rupture [16]. In all, there are several arguments that ADMA could play an important role in the pathogenesis of vascular disease [17]. While elevated ADMA levels in patients with renal failure could be attributed, at least in part, to reduced renal excretion, the precise cause of elevated ADMA levels in other diseases has not been identified yet. One cause could be a reduced activity of the enzyme dimethylarginine dimethylaminohydrolase (DDAH), which metabolises ADMA. The activity of DDAH seems to be critical in regulating ADMA levels [5]. DDAH activity may be inhibited by elevated concentrations of homocysteine [18], cholesterol [19] and hyperglycemia [20]. Our finding of a stronger correlation between SDMA and creatinine than between ADMA and creatinine, supports the notion that ADMA is not only excreted by the kidney like SDMA, but is also metabolised in the liver where DDAH is highly expressed [21]. Non-cardio-embolic infarction and TIA are mainly due to atherosclerosis. Cardio-embolic infarction in this study is equivalent to atrial fibrillation, which often occurs on the background of other diseases. In elderly subjects such as stroke patients, ischemic heart disease is one of the main causes of atrial fibrillation. Risk factors for CVD such as diabetes [22], hypertension [23], and hyperlipidemia [17] as well as atherosclerosis [24] have been positively associated with elevated ADMA. According to these earlier observations we would therefore expect an increased ADMA concentration in CVD and its subgroups compared to healthy controls. We did observe an association between ADMA and TIA and in the subgroup of patients with ischemic stroke, between ADMA and previous stroke/TIA, but we could however not replicate any significant relationships between ADMA and known CVD risk factors such as diabetes mellitus, hypertension and dyslipidaemia. In many respects TIA can be regarded a transitional form to non-cardio-embolic infarction. It is, therefore, unclear why a strong association between risk and elevated ADMA concentrations was only observed in TIA and not in noncardio-embolic infarction. We speculate that the TIA group represents a more homogenous group of CVD with mainly large vessel disease [25], in contrast to the subgroup of CVD with non-cardio-embolic infarctions where small vessel disease contributes to a substantial proportion of this condition.

276

P. Wanby et al. / Atherosclerosis 185 (2006) 271–277

This could perhaps also explain why we found a tendency towards increased risk associated with elevated ADMA concentrations in cardio-embolic stroke, to which large vessel disease also may contribute. An association between plasma ADMA levels and large vessel disease has previously been demonstrated in peripheral artery disease [24] and in coronary heart disease [7]. Whether ADMA levels actually are increased in large vessel stroke diseases in comparison to lacunar infarction (small vessel disease) warrants further studies. SDMA is an inactive structural isomer of ADMA and does not inhibit NO synthase [26]. However, as l-arginine, ADMA, and SDMA share a common pathway for entry into the cell, a high plasma concentration of SDMA may, therefore, indirectly reduce NO production by competing with larginine for cellular uptake [6]. This may explain our findings of increased plasma levels of SDMA in patients with cardioembolic infarction. Reduced plasma l-arginine concentration has previously been reported in patients with hypercholesterolemia [27] and was in the present study observed in patients with acute CVD, cardio-embolic and haemorrhagic stroke and a reduced l-arginine/ADMA ratio was seen in these groups as in patients with TIA. A reduced level of larginine as well as a decreased l-arginine/ADMA ratio is compatible with ADMA acting as a competitive inhibitor of NO synthase with l-arginine as its substrate, resulting in a reduced NO synthesis. To minimize the influence of certain established clinical conditions associated with an increased risk of developing stroke, patients with a history of migraine, SLE, systemic malignancies, etc. were excluded. Patients on lipid-lowering therapy may have altered ADMA levels [28] and patients on thyroxine therapy may through an altered lipid profile, theoretically also have altered ADMA concentrations. Patients from these two categories were therefore also excluded. Since ADMA has been associated with characteristics of the metabolic syndrome only subjects without those characteristics were included as controls. The stroke and TIA patients were thus compared with a group of patients likely to be relatively free from atherosclerosis. The absolute concentration differences of ADMA between subgroups of CVD and controls may seem small. However, the HPLC method used for quantification is characterized by a very low imprecision, allowing reliable determination of these small concentration differences. In addition, the biological variation was also very small, as can be seen from the narrow distribution of ADMA concentrations in the control group, with an inter-individual coefficient of variation of approximately 10%. In the study by Valkonen et al. it was clearly shown that even slightly elevated ADMA concentrations in plasma were associated with a strongly increased risk for acute coronary events [7]. In addition, both the generation of NO and the formation of ADMA, by the sequence of protein methylation and proteolysis, are intracellular processes. ADMA in plasma probably originates from cellular spillover and the concentration of ADMA in plasma

only weakly mirrors its intracellular concentration. A limited increase in the plasma concentration of ADMA may thus reflect a much larger increase of its intracellular concentration, which may be sufficient to substantially inhibit NO production. In contrast to what would have been expected from previous reports, we did not find any associations between high ADMA levels and diabetes [22], high blood pressure [23] or dyslipidemia [17,19], nor between ADMA and the presence of ICA stenosis [3]. Although the group of patients with ICA stenosis was very small, the group of patients with CVD was quite large. The results of the present study therefore raise some doubts to the applicability of ADMA as a marker of risk factors in the metabolic syndrome. However, ADMA impairs the NO-dependent functions of the endothelium and the primary pathophysiological mechanism of ADMA is therefore different from other known risk factors for cardiovascular disease. Accordingly, it can be expected that the effect of ADMA on CVD are independent of other risk factors. In conclusion, in Swedish patients with cerebrovascular disease, ADMA was not associated with traditional cardiovascular risk factors besides nephropathy. Our results indicate that ADMA is a weak independent marker for acute stroke and a strong marker for TIA. We also conclude that relative arginine deficiency, measured as the l-arginine/ADMA ratio, is present in acute CVD and several of its subgroups apart from non-cardio-embolic infarction. Whether circulating ADMA levels plays a part in the pathogenesis of CVD or solely is an indicator of such disease remains to be further investigated. Acknowledgements This work was supported by a grant from the Medical Research Council of Southeast Sweden (FORSS). We are also indebted to Ms. Sigrid de Jong, Ms. Ingvor Gardtman, Ms. Ingrid Djukanovic, and Ms. Inger Gustafsson for excellent technical assistance. References [1] Elkind MS, Sacco RL. Stroke risk factors and stroke prevention. Sem Neurol 1998;18:429–40. [2] Higashi Y, Oshima T, Sasaki N, et al. Relationship between insulin resistance and endothelium dependent vascular relaxation in patients with essential hypertension. Hypertension 1997;29:280–5. [3] Miyazaki H, Matsuoka H, Cooke JP. Endogenous nitric oxide synthase inhibitor. A novel marker of atherosclerosis. Circulation 1999;99:1141–6. [4] Cooke JP, Dzau VJ. Nitric oxide synthase: role in the genesis of vascular disease. Annu Rev Med 1997;48:489–509. [5] McAllister RJ, Fickling SA, Whitley GSJ, Vallence. Metabolism of methylarginines by human vasculature: implications for the regulation of nitric oxide synthesis. Br J Pharmacol 1992;112:43–8. [6] Vallance P, Leone A, Calver A, Collier J, Moncada S. Endogenous dimethylarginine as an inhibitor of nitric oxide synthesis. J Cardiovasc Pharmacol 1992;20:S60–2.

P. Wanby et al. / Atherosclerosis 185 (2006) 271–277 [7] Valkonen VP, P¨aiv¨a H, Salonen JT, et al. Risk of acute coronary events and serum concentration of asymmetrical dimethylarginine. Lancet 2001;358:2127–8. [8] Lu TM, Ding YA, Lin SJ, Lee WS, Tai HC. Plasma levels of asymmetrical dimethylarginine and adverse cardiovascular events after percutaneous coronary intervention. Eur Heart J 2003;24:1912–9. [9] Yoo J-H, Lee S-C. Elevated levels of plasma homocysteine and asymmetric dimethylarginine in elderly patients with stroke. Atherosclerosis 2001;158:425–30. [10] Aho K, Harmsen P, Hatano S, Marquardsen J, Smirnov VE, Strasser T. Cerebrovascular disease in the community: results of a WHO collaboration study. Bull World Health Organ 1980;58:113–30. [11] Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 1997;20:1183–97. [12] Bluth EI, Wetzner SM, Stavros AT, Aufrichtig D, Marish KW, Baker JD. Carotid duplex sonography: a multicenter recommendation for standardized imaging and Doppler criteria. RadioGraphics 1988;8:487–506. Fig. 22. [13] Jeppsson JO, Jerntorp P, Sundkvist G, Englund H, Nylund V. Measurement of hemoglobin A1c by a new liquid-chromatography assay; methodology, clinical utility, and relation to tolerance evaluated. Clin Chem 1986;32:1867–72. [14] Teerlink T, Nijveldt RJ, de Jong S, van Leeuwen PAM. Determination of arginine, asymmetric dimethylarginine in human plasma and other biological samples by high-performance liquid chromatography. Anal Biochem 2002;303:131–7. [15] B¨oger RH. The emerging role of asymmetric dimethylarginine as a novel cardiovascular risk factor. Cardiovasc Res 2003;59:824–33. [16] van der Wal AC, Becker AE, van der Loos CM, Das PK. Site of intimal rupture or erosion of thrombosed coronary atherosclerotic plaques is characterized by an inflammatory process irrespective of the dominant plaque morphology. Circulation 1994;89:36– 44. [17] Lundman P, Eriksson MJ, St˝uhlinger M, Cooke JP, Hamsten A, Tornvall P. Mild-to-moderate hypertriglyceridemia in young men is associated with endothelial dysfunction and increased plasma concentrations of asymmetric dimethylarginine. J Am Coll Cardiol 2001;38:111–6.

277

[18] St˝uhlinger MC, Oka RK, Graf EE, et al. Endothelial dysfunction induced hyperhomocysteinemia: role of asymmetric dimethylarginine. Circulation 2003;108:933–8. [19] B¨oger RH, Bode-B¨oger SM, Szuba A, et al. Asymmetric dimethylarginine (ADMA): a novel risk factor for endothelial dysfunction: its role in hypercholesterolemia. Circulation 1998;98:1842–7. [20] Lin KY, Ito A, Asagami T, et al. Impaired nitric oxide synthase pathway in diabetes mellitus: role of asymmetric dimethylarginine and dimethylarginine dimethylaminohydrolase. Circulation 2002;106:987–92. [21] Nijveldt RJ, Teerlink T, Siroen MPC, van Lambalgen AA, Rauwerda JA, van Leeuwen PA. The liver is an important organ in the metabolism of asymmetrical dimethylarginine (ADMA). Clin Nutr 2003;22:17–20. [22] Abbasi F, Asagmi T, Coke JP, et al. Plasma concentrations of asymmetric dimethylarginine are increased in patients with type 2 diabetes mellitus. Am J Cardiol 2001;88:1201–3. [23] Surdacki A, Nowicki M, Sandmann J. Reduced urinary excretion of nitric oxide metabolites and increased plasma levels of asymmetrical dimethylarginine in men with essential hypertension. J Cardiovasc Pharmacol 1999;33:652–8. [24] B¨oger RH, Bode-B¨oger SM, Thiele W, Junker W, Alexander K, Fr¨olich JC. Biochemical evidence for impaired nitric oxide synthesis in patients with peripheral arterial occlusive disease. Circulation 1997;95:2068–74. [25] Purroy F, Montaner J, Rovira A, Delgado P, Quintana M, AlvarezSabin J. Higher risk of further vascular events among transient ischemic attack patients with diffusion-weighted imaging acute ischemic lesions. Stroke 2004;35:2313–9. [26] Chan NN, Chan JC. Asymmetric dimethylarginine (ADMA): a potential link between endothelial dysfunction and cardiovascular diseases in insulin resistance syndrome? Diabetologia 2002;45:1609–16. [27] Jeserich M, M˝unzel T, Just H, Drexler H. Reduced plasma l-arginine in hyper-cholesterolemia. Lancet 1992;339:561. [28] Lu TM, Ding YA, Leu HB, Yin WH, Sheu WH, Chu KM. Effect of rozuvastatin on plasma levels of asymmetric dimethylarginine in patients with hypercholesterolemia. Am J Cardiol 2004;94:157–61.