Original Paper Cerebrovasc Dis 2014;37:444–450 DOI: 10.1159/000363279
Received: February 4, 2014 Accepted: April 28, 2014 Published online: July 23, 2014
Plasma Natriuretic Peptides and Incidence of Subtypes of Ischemic Stroke John Berntsson Elisabet Zia Yan Borné Olle Melander Bo Hedblad Gunnar Engström Department of Clinical Sciences in Malmö, Lund University, Lund, Sweden
Abstract Background and Purpose: Natriuretic peptides predict poor outcomes in cardiovascular disease. However, the knowledge of their relationship to stroke is limited and prospective studies from the general population are few. The purpose of this study was to explore the relationship between N-terminal pro-brain natriuretic peptide (NT-proBNP) and midregional pro-atrial natriuretic peptide (MR-proANP) plasma levels and the risk for ischemic stroke and its subtypes. Methods: NT-proBNP and MR-proANP were measured in fasting blood samples from 4,862 subjects (40.2% men, mean age 57.5 ± 6.0 years) without cardiovascular disease from the Malmö Diet and Cancer Study, a prospective, populationbased study in Sweden. Incidence of ischemic stroke was monitored over a mean follow-up of 14.9 ± 3.0 years. Stroke cases were etiologically classified according to the TOAST classification. Cox proportional-hazards regression was used to study the incidence of stroke in relationship to NT-proBNP and MR-proANP. Results: During follow-up, 227 had a firstever ischemic stroke (large-artery atherosclerosis, n = 35; cardioembolic stroke, n = 44; small-artery occlusion, n = 80; undetermined cause, n = 68). In the age- and sex-adjusted
© 2014 S. Karger AG, Basel 1015–9770/14/0376–0444$39.50/0 E-Mail [email protected] www.karger.com/ced
model, only NT-proBNP was associated with total ischemic stroke. This association was completely explained by an increased incidence of cardioembolic stroke. Adjusted for cardiovascular risk factors (age, sex, hypertension, diabetes, smoking, body mass index and low-density lipoprotein cholesterol), the hazard ratios (HRs, 95% confidence interval, 95% CI) for cardioembolic stroke were 1.00 (reference), 1.42 (0.34–6.00), 2.79 (0.77–10.12) and 5.64 (1.66–19.20), respectively, for the 1st, 2nd, 3rd and 4th quartiles of NT-proBNP. The corresponding HRs (95% CIs) for quartiles of MR-proANP were 1.00 (reference), 1.83 (0.55–6.14), 1.20 (0.33–4.34) and 3.96 (1.31–11.99), respectively. In total, 335 (6.9%) subjects were diagnosed with atrial fibrillation during follow-up. Among the cardioembolic stroke cases, 30% were diagnosed with atrial fibrillation before the stroke event and another 36% within 6 months after the stroke. Of the cardioembolic stroke cases with atrial fibrillation, 59% were in the top quartile of NT-proBNP, 69% in the top quartile of MR-proANP and 79% were either in the top quartile of NT-proBNP or in the top quartile of MR-proANP. Conclusion: High plasma levels of NT-proBNP and MR-proANP are associated with a substantially increased risk of cardioembolic stroke, but not with other subtypes of ischemic stroke. The results suggest that assessment of stroke risk, including electrocardiography, is warranted in subjects with high NT-proBNP or MR-proANP. © 2014 S. Karger AG, Basel
John Berntsson Department of Clinical Sciences in Malmö, Cardiovascular Epidemiology Skåne University Hospital, Lund University, CRC, House 60, Level 13 SE–20502 Malmö (Sweden) E-Mail john.berntsson @ med.lu.se
Downloaded by: Siriraj Medical Library, Mahidol University 202.28.191.34 - 3/11/2015 2:25:31 PM
Key Words Natriuretic peptides · Stroke · Epidemiology
Natriuretic peptides, such as atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP), are released from cardiomyocytes in response to atrial or ventricular wall stretch [1]. BNP and its N-terminal prohormone (NT-proBNP) are widely used for the screening and diagnosis of acute congestive heart failure and high levels of NT-proBNP are associated with a worse prognosis in heart failure patients [2]. Natriuretic peptides are often elevated in the acute situation of ischemic stroke and have been associated with a worse prognosis [3, 4]. NT-proBNP and midregional pro-atrial natriuretic peptide (MR-proANP) have also been associated with an increased risk of death and cardiovascular events, including stroke in studies from the general population [5–11]. The reason for the relationship between natriuretic peptides and the incidence of stroke is incompletely known. One prospective study of NT-proBNP, which included ischemic stroke subtypes, reported a relationship between NT-proBNP and cardioembolic stroke and other types of nonlacunar ischemic strokes [5]. To our knowledge, no previous study has explored the relationship between MR-proANP and the incidence of ischemic stroke subtypes. The purpose of this study was to explore the relationship between natriuretic peptide levels and the risk for ischemic stroke and if there was a difference between its subtypes.
Methods Study Population Between 1991 and 1996, citizens in Malmö born between 1926 and 1945 were invited to participate in the Malmö Diet and Cancer Study [12]. A total of 28,449 subjects (60% women) participated. Of them, a random sample of 6,103 participants who entered the Malmö Diet and Cancer Study between 1991 and 1994 were selected to participate in a study of the epidemiology of carotid artery diseases, of whom 5,543 subjects donated fasting blood samples [13]. For the present study, the sample was restricted to those with complete information on NT-proBNP, MR-proANP, blood pressure, diabetes, low-density lipoprotein (LDL) and body mass index (BMI; n = 4,937). Information about cystatin C was available for 4,839 subjects. Subjects with prior stroke [14] or hospitalization attributable to heart failure or atrial fibrillation, according to national registers, were excluded (n = 75). Previous validation studies of diagnoses from these registers have shown high case validity [15, 16]. A total of 4,862 subjects (40.2% men, 46–68 years old) remained after the exclusions. The ethics committee at Lund University approved the Malmö Diet and Cancer Study. All individuals provided informed consent.
Natriuretic Peptides and Stroke
Blood pressure was measured using a mercury-column sphygmomanometer after 10 min of rest in the supine position. Height and weight were measured as described [12, 13, 17]. Information about smoking, diabetes and blood pressure-lowering medications was obtained from a self-administered questionnaire [13]. Participants were categorized into smokers and nonsmokers. Subjects with missing information on smoking (n = 103, 2.1%) were coded in a separate category (dummy variable) to keep them in multivariate analysis. Diabetes mellitus was defined as self-reported physician diagnosis or treatment with antidiabetic medication or a fasting whole blood glucose level ≥6.1 mmol/l. Fasting cholesterol and glucose levels were analyzed on fresh blood samples according to standard procedures at the Department of Clinical Chemistry, University Hospital Malmö. LDL levels were calculated according to the Friedewald formula. Plasma levels of NT-proBNP, MR-proANP and cystatin C were measured from fasting plasma samples that had been frozen at –80 ° C immediately after collection. NT-proBNP was measured using the Dimension RxL N-BNP (Dade Behring, Germany) [18]. MR-proANP was measured using immunoluminometric sandwich assays (BRAHMS, Berlin, Germany) [19]. Cystatin C was measured using a particle-enhanced immunonephelometric assay (N Latex Cystatin, Siemens Diagnostics) [20]. The minimum detection limits for NT-proBNP and MR-proANP were 2.0 pg/ml and 6 pmol/l, respectively. Mean interassay coefficients of variation were 10% or less for MR-proANP, 2.7% for NT-proBNP and 4.3% for cystatin C. Cystatin C equation (CKD-EPI 2012) was used to calculate estimated glomerular filtration rate (GFR) [21].
Follow-Up All subjects were followed from the baseline examination until first-ever stroke event, emigration from Sweden, death or until 31 December 2008, by linkage to the Stroke Register of Malmö [14]. Stroke cases who moved out from the city of Malmö were identified from the Swedish Hospital Discharge Register [22]. Case retrieval and methods of defining ischemic stroke have been described before [17]. Ischemic stroke subtypes were classified by a specialized neurologist and a specialized research nurse according to Adams et al. [23]. Incident atrial fibrillation was ascertained by linkage to the Swedish Hospital Discharge Register and the Swedish Cause of Death Register. A validation study from the present cohort reported high case validity for atrial fibrillation [16]. Statistics NT-proBNP and MR-proANP were log-normalized due to skewed distributions. Subjects were categorized into quartiles, separately for men and women, based on NT-proBNP and MR-proANP concentrations. Cox proportional-hazards regression was used to compare the incidence of ischemic stroke and its subtypes in relation to NT-proBNP and MR-proANP and to calculate hazard ratios (HRs) adjusted for age, sex, smoking, systolic blood pressure, blood pressure medication, diabetes, LDL and BMI. In addition, we added GFR to the model in an analysis of subjects with information on cystatin C. We used log-log survival curves to evaluate the proportional-hazards assumption. All comparisons were two-sided and statistical significance was set at the level of 5%. Statistical analyses were performed with IBM SPSS Statistics 21.0.
Cerebrovasc Dis 2014;37:444–450 DOI: 10.1159/000363279
445
Downloaded by: Siriraj Medical Library, Mahidol University 202.28.191.34 - 3/11/2015 2:25:31 PM
Introduction
Table 1. Baseline characteristics
Baseline characteristics are presented in table 1. Median NT-proBNP (interquartile range) was 47.0 pg/ml in men (25.0–89.9) and 70.3 pg/ml in women (42.1–123.0). Median MR-proANP (interquartile range) was 59.8 pmol/l in men (46.6–78.4) and 70.7 pmol/l in women (55.0–89.3). The mean age at screening was 57.6 ± 6.0 years in men and 57.4 ± 5.9 years in women. During an average follow-up of 14.9 ± 3.0 years, a total of 227 had a first-ever ischemic stroke (large-artery atherosclerosis, n = 35; cardioembolic, n = 44; small-artery occlusion, n = 80; undetermined cause, including cases with two/more causes identified, negative evaluation or incomplete/missing evaluation, n = 68). The incidence of ischemic and cardioembolic stroke in relation to sex-specific quartiles of NT-proBNP and MR-proBNP is shown in tables 2, 3 and figures 1, 2. Adjusted for age and sex, log-normalized NT-proBNP was significantly associated with total ischemic stroke (HR: 1.17, 95% confidence interval, 95% CI: 1.02–1.34, per 1 standard deviation, SD; table 2). This relationship was completely explained by increased incidence of cardioembolic stroke, and NT-proBNP was unrelated to ischemic stroke after exclusion of cardioembolic cases. The age- and sex-adjusted HR for cardioembolic stroke was 2.12 (95% CI: 1.60–2.81) per 1 SD and the risk factoradjusted HR was 2.08 (95% CI: 1.55–2.78, p < 0.001). Additional adjustment for GFR did not change the results (HR: 2.15, 95% CI: 1.59–2.91, p < 0.001). Neither largeartery atherosclerosis (risk factor-adjusted HR per 1 SD: 1.11, 95% CI: 0.77–1.58, p = 0.58), small-artery occlusion (HR: 0.88, 95% CI: 0.70–1.11) nor undetermined ischemic stroke (HR: 1.04, 95% CI: 0.81–1.34) were associated with NT-proBNP. MR-proANP was associated with cardioembolic subtype only. The age- and sex-adjusted HR for cardioembolic stroke was 1.98 (95% CI: 1.51–2.59) per 1 SD and the risk factor-adjusted HR was 1.90 (95% CI: 1.45– 2.50, p < 0.001; table 3). Neither large-artery atherosclerosis (risk factor-adjusted HR per 1 SD: 1.05, 95% CI: 0.75–1.47, p = 0.79), small-artery occlusion (HR: 0.83, 95% CI: 0.66–1.04) nor undetermined ischemic stroke (HR: 1.00, 95% CI: 0.78–1.29) were associated with MRproANP. A total of 335 (6.9%) subjects were diagnosed with atrial fibrillation during follow-up. Among the cardioembolic stroke cases, 30% (13 of 44) were diagnosed with atrial fibrillation before the stroke event and another 36% (16 of 44) were diagnosed with 446
Cerebrovasc Dis 2014;37:444–450 DOI: 10.1159/000363279
Men (n = 1,955) Age, years Current smoking, % Systolic BP, mm Hg Diastolic BP, mm Hg BP-lowering medication, % Diabetes, % LDL, mmol/l BMI, kg/m2 Estimated GFR, ml/min NT-proBNP, pg/ml Median First quartile n Second quartile n Third quartile n Fourth quartile n MR-proANP, pmol/l Median First quartile n Second quartile n Third quartile n Fourth quartile n
Women (n = 2,907)
57.6±6.0 22.4 143.2±18.7 88.7±9.6 16.6 10.9 4.1±0.9 26.1±3.5 103.2±15.6 (n = 1,910)
57.4±5.9 21.1 139.9±19.1 85.6±9.0 14.9 5.7 4.2±1.0 25.5±4.2 99.4±14.2 (n = 2,854)
47.0 89.9 491
70.3 123.0 727
59.8 78.4 488
70.7 89.3 726
Values are presented as mean ± SD unless otherwise stated. BP = Blood pressure.
atrial fibrillation at the time of stroke (n = 9) or within 6 months after the stroke event (n = 7). A sensitivity analysis was performed, in which cases with incidence of atrial fibrillation was censored at the time of atrial fibrillation diagnosis. However, the results for cardioembolic stroke remained significant when adjusted for risk factors (HR: 2.03, 95% CI: 1.40–2.94 per 1 SD; HR: 5.51, 95% CI: 1.24–24.61 for 4th vs. 1st quartile of NT-proBNP and HR: 1.99, 95% CI: 1.40–2.83 per 1 SD; HR: 3.72, 95% CI: 1.03–13.45 for 4th vs. 1st quartile of MR-proANP). Notably, of the 29 cardioembolic stroke cases with atrial fibrillation before or within 6 months after stroke, 79% were in the top quartile of either NT-proBNP (59%) or MR-proANP (69%). Berntsson/Zia/Borné/Melander/Hedblad/ Engström
Downloaded by: Siriraj Medical Library, Mahidol University 202.28.191.34 - 3/11/2015 2:25:31 PM
Results
Table 2. Adjusted HRs and 95% CIs for incident ischemic stroke by NT-proBNP
Quartiles of NT-proBNP
Q1
Total ischemic stroke Incidence (n) 2.32 (43) Age and sex 1 Risk factors1 1 Total noncardioembolic stroke n 40 Age and sex 1 Risk factors1 1 Cardioembolic stroke n 3 Age and sex 1 Risk factors1 1 1 Adjusted
p trend
log NT-proBNP, per 1 SD
Q2
Q3
Q4
2.93 (53) 1.05 (0.70–1.57) 1.12 (0.74–1.68)
2.81 (51) 0.93 (0.62–1.40) 0.98 (0.65–1.49)
4.52 (80) 1.36 (0.93–1.98) 1.30 (0.88–1.91)
0.124 0.238
1.17 (1.02–1.34) 1.13 (0.99–1.30)
48 1.02 (0.67–1.56) 1.10 (0.72–1.68)
40 0.79 (0.51–1.23) 0.85 (0.54–1.32)
55 1.01 (0.67–1.54) 0.97 (0.63–1.48)
0.810 0.965
1.00 (0.86–1.17) 0.98 (0.84–1.14)
5 1.37 (0.33–5.77) 1.42 (0.34–6.00)
11 2.78 (0.77–10.05) 2.79 (0.77–10.12)
25 5.77 (1.72–19.38) 5.64 (1.66–19.20)
Received: February 4, 2014 Accepted: April 28, 2014 Published online: July 23, 2014
Plasma Natriuretic Peptides and Incidence of Subtypes of Ischemic Stroke John Berntsson Elisabet Zia Yan Borné Olle Melander Bo Hedblad Gunnar Engström Department of Clinical Sciences in Malmö, Lund University, Lund, Sweden
Abstract Background and Purpose: Natriuretic peptides predict poor outcomes in cardiovascular disease. However, the knowledge of their relationship to stroke is limited and prospective studies from the general population are few. The purpose of this study was to explore the relationship between N-terminal pro-brain natriuretic peptide (NT-proBNP) and midregional pro-atrial natriuretic peptide (MR-proANP) plasma levels and the risk for ischemic stroke and its subtypes. Methods: NT-proBNP and MR-proANP were measured in fasting blood samples from 4,862 subjects (40.2% men, mean age 57.5 ± 6.0 years) without cardiovascular disease from the Malmö Diet and Cancer Study, a prospective, populationbased study in Sweden. Incidence of ischemic stroke was monitored over a mean follow-up of 14.9 ± 3.0 years. Stroke cases were etiologically classified according to the TOAST classification. Cox proportional-hazards regression was used to study the incidence of stroke in relationship to NT-proBNP and MR-proANP. Results: During follow-up, 227 had a firstever ischemic stroke (large-artery atherosclerosis, n = 35; cardioembolic stroke, n = 44; small-artery occlusion, n = 80; undetermined cause, n = 68). In the age- and sex-adjusted
© 2014 S. Karger AG, Basel 1015–9770/14/0376–0444$39.50/0 E-Mail [email protected] www.karger.com/ced
model, only NT-proBNP was associated with total ischemic stroke. This association was completely explained by an increased incidence of cardioembolic stroke. Adjusted for cardiovascular risk factors (age, sex, hypertension, diabetes, smoking, body mass index and low-density lipoprotein cholesterol), the hazard ratios (HRs, 95% confidence interval, 95% CI) for cardioembolic stroke were 1.00 (reference), 1.42 (0.34–6.00), 2.79 (0.77–10.12) and 5.64 (1.66–19.20), respectively, for the 1st, 2nd, 3rd and 4th quartiles of NT-proBNP. The corresponding HRs (95% CIs) for quartiles of MR-proANP were 1.00 (reference), 1.83 (0.55–6.14), 1.20 (0.33–4.34) and 3.96 (1.31–11.99), respectively. In total, 335 (6.9%) subjects were diagnosed with atrial fibrillation during follow-up. Among the cardioembolic stroke cases, 30% were diagnosed with atrial fibrillation before the stroke event and another 36% within 6 months after the stroke. Of the cardioembolic stroke cases with atrial fibrillation, 59% were in the top quartile of NT-proBNP, 69% in the top quartile of MR-proANP and 79% were either in the top quartile of NT-proBNP or in the top quartile of MR-proANP. Conclusion: High plasma levels of NT-proBNP and MR-proANP are associated with a substantially increased risk of cardioembolic stroke, but not with other subtypes of ischemic stroke. The results suggest that assessment of stroke risk, including electrocardiography, is warranted in subjects with high NT-proBNP or MR-proANP. © 2014 S. Karger AG, Basel
John Berntsson Department of Clinical Sciences in Malmö, Cardiovascular Epidemiology Skåne University Hospital, Lund University, CRC, House 60, Level 13 SE–20502 Malmö (Sweden) E-Mail john.berntsson @ med.lu.se
Downloaded by: Siriraj Medical Library, Mahidol University 202.28.191.34 - 3/11/2015 2:25:31 PM
Key Words Natriuretic peptides · Stroke · Epidemiology
Natriuretic peptides, such as atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP), are released from cardiomyocytes in response to atrial or ventricular wall stretch [1]. BNP and its N-terminal prohormone (NT-proBNP) are widely used for the screening and diagnosis of acute congestive heart failure and high levels of NT-proBNP are associated with a worse prognosis in heart failure patients [2]. Natriuretic peptides are often elevated in the acute situation of ischemic stroke and have been associated with a worse prognosis [3, 4]. NT-proBNP and midregional pro-atrial natriuretic peptide (MR-proANP) have also been associated with an increased risk of death and cardiovascular events, including stroke in studies from the general population [5–11]. The reason for the relationship between natriuretic peptides and the incidence of stroke is incompletely known. One prospective study of NT-proBNP, which included ischemic stroke subtypes, reported a relationship between NT-proBNP and cardioembolic stroke and other types of nonlacunar ischemic strokes [5]. To our knowledge, no previous study has explored the relationship between MR-proANP and the incidence of ischemic stroke subtypes. The purpose of this study was to explore the relationship between natriuretic peptide levels and the risk for ischemic stroke and if there was a difference between its subtypes.
Methods Study Population Between 1991 and 1996, citizens in Malmö born between 1926 and 1945 were invited to participate in the Malmö Diet and Cancer Study [12]. A total of 28,449 subjects (60% women) participated. Of them, a random sample of 6,103 participants who entered the Malmö Diet and Cancer Study between 1991 and 1994 were selected to participate in a study of the epidemiology of carotid artery diseases, of whom 5,543 subjects donated fasting blood samples [13]. For the present study, the sample was restricted to those with complete information on NT-proBNP, MR-proANP, blood pressure, diabetes, low-density lipoprotein (LDL) and body mass index (BMI; n = 4,937). Information about cystatin C was available for 4,839 subjects. Subjects with prior stroke [14] or hospitalization attributable to heart failure or atrial fibrillation, according to national registers, were excluded (n = 75). Previous validation studies of diagnoses from these registers have shown high case validity [15, 16]. A total of 4,862 subjects (40.2% men, 46–68 years old) remained after the exclusions. The ethics committee at Lund University approved the Malmö Diet and Cancer Study. All individuals provided informed consent.
Natriuretic Peptides and Stroke
Blood pressure was measured using a mercury-column sphygmomanometer after 10 min of rest in the supine position. Height and weight were measured as described [12, 13, 17]. Information about smoking, diabetes and blood pressure-lowering medications was obtained from a self-administered questionnaire [13]. Participants were categorized into smokers and nonsmokers. Subjects with missing information on smoking (n = 103, 2.1%) were coded in a separate category (dummy variable) to keep them in multivariate analysis. Diabetes mellitus was defined as self-reported physician diagnosis or treatment with antidiabetic medication or a fasting whole blood glucose level ≥6.1 mmol/l. Fasting cholesterol and glucose levels were analyzed on fresh blood samples according to standard procedures at the Department of Clinical Chemistry, University Hospital Malmö. LDL levels were calculated according to the Friedewald formula. Plasma levels of NT-proBNP, MR-proANP and cystatin C were measured from fasting plasma samples that had been frozen at –80 ° C immediately after collection. NT-proBNP was measured using the Dimension RxL N-BNP (Dade Behring, Germany) [18]. MR-proANP was measured using immunoluminometric sandwich assays (BRAHMS, Berlin, Germany) [19]. Cystatin C was measured using a particle-enhanced immunonephelometric assay (N Latex Cystatin, Siemens Diagnostics) [20]. The minimum detection limits for NT-proBNP and MR-proANP were 2.0 pg/ml and 6 pmol/l, respectively. Mean interassay coefficients of variation were 10% or less for MR-proANP, 2.7% for NT-proBNP and 4.3% for cystatin C. Cystatin C equation (CKD-EPI 2012) was used to calculate estimated glomerular filtration rate (GFR) [21].
Follow-Up All subjects were followed from the baseline examination until first-ever stroke event, emigration from Sweden, death or until 31 December 2008, by linkage to the Stroke Register of Malmö [14]. Stroke cases who moved out from the city of Malmö were identified from the Swedish Hospital Discharge Register [22]. Case retrieval and methods of defining ischemic stroke have been described before [17]. Ischemic stroke subtypes were classified by a specialized neurologist and a specialized research nurse according to Adams et al. [23]. Incident atrial fibrillation was ascertained by linkage to the Swedish Hospital Discharge Register and the Swedish Cause of Death Register. A validation study from the present cohort reported high case validity for atrial fibrillation [16]. Statistics NT-proBNP and MR-proANP were log-normalized due to skewed distributions. Subjects were categorized into quartiles, separately for men and women, based on NT-proBNP and MR-proANP concentrations. Cox proportional-hazards regression was used to compare the incidence of ischemic stroke and its subtypes in relation to NT-proBNP and MR-proANP and to calculate hazard ratios (HRs) adjusted for age, sex, smoking, systolic blood pressure, blood pressure medication, diabetes, LDL and BMI. In addition, we added GFR to the model in an analysis of subjects with information on cystatin C. We used log-log survival curves to evaluate the proportional-hazards assumption. All comparisons were two-sided and statistical significance was set at the level of 5%. Statistical analyses were performed with IBM SPSS Statistics 21.0.
Cerebrovasc Dis 2014;37:444–450 DOI: 10.1159/000363279
445
Downloaded by: Siriraj Medical Library, Mahidol University 202.28.191.34 - 3/11/2015 2:25:31 PM
Introduction
Table 1. Baseline characteristics
Baseline characteristics are presented in table 1. Median NT-proBNP (interquartile range) was 47.0 pg/ml in men (25.0–89.9) and 70.3 pg/ml in women (42.1–123.0). Median MR-proANP (interquartile range) was 59.8 pmol/l in men (46.6–78.4) and 70.7 pmol/l in women (55.0–89.3). The mean age at screening was 57.6 ± 6.0 years in men and 57.4 ± 5.9 years in women. During an average follow-up of 14.9 ± 3.0 years, a total of 227 had a first-ever ischemic stroke (large-artery atherosclerosis, n = 35; cardioembolic, n = 44; small-artery occlusion, n = 80; undetermined cause, including cases with two/more causes identified, negative evaluation or incomplete/missing evaluation, n = 68). The incidence of ischemic and cardioembolic stroke in relation to sex-specific quartiles of NT-proBNP and MR-proBNP is shown in tables 2, 3 and figures 1, 2. Adjusted for age and sex, log-normalized NT-proBNP was significantly associated with total ischemic stroke (HR: 1.17, 95% confidence interval, 95% CI: 1.02–1.34, per 1 standard deviation, SD; table 2). This relationship was completely explained by increased incidence of cardioembolic stroke, and NT-proBNP was unrelated to ischemic stroke after exclusion of cardioembolic cases. The age- and sex-adjusted HR for cardioembolic stroke was 2.12 (95% CI: 1.60–2.81) per 1 SD and the risk factoradjusted HR was 2.08 (95% CI: 1.55–2.78, p < 0.001). Additional adjustment for GFR did not change the results (HR: 2.15, 95% CI: 1.59–2.91, p < 0.001). Neither largeartery atherosclerosis (risk factor-adjusted HR per 1 SD: 1.11, 95% CI: 0.77–1.58, p = 0.58), small-artery occlusion (HR: 0.88, 95% CI: 0.70–1.11) nor undetermined ischemic stroke (HR: 1.04, 95% CI: 0.81–1.34) were associated with NT-proBNP. MR-proANP was associated with cardioembolic subtype only. The age- and sex-adjusted HR for cardioembolic stroke was 1.98 (95% CI: 1.51–2.59) per 1 SD and the risk factor-adjusted HR was 1.90 (95% CI: 1.45– 2.50, p < 0.001; table 3). Neither large-artery atherosclerosis (risk factor-adjusted HR per 1 SD: 1.05, 95% CI: 0.75–1.47, p = 0.79), small-artery occlusion (HR: 0.83, 95% CI: 0.66–1.04) nor undetermined ischemic stroke (HR: 1.00, 95% CI: 0.78–1.29) were associated with MRproANP. A total of 335 (6.9%) subjects were diagnosed with atrial fibrillation during follow-up. Among the cardioembolic stroke cases, 30% (13 of 44) were diagnosed with atrial fibrillation before the stroke event and another 36% (16 of 44) were diagnosed with 446
Cerebrovasc Dis 2014;37:444–450 DOI: 10.1159/000363279
Men (n = 1,955) Age, years Current smoking, % Systolic BP, mm Hg Diastolic BP, mm Hg BP-lowering medication, % Diabetes, % LDL, mmol/l BMI, kg/m2 Estimated GFR, ml/min NT-proBNP, pg/ml Median First quartile n Second quartile n Third quartile n Fourth quartile n MR-proANP, pmol/l Median First quartile n Second quartile n Third quartile n Fourth quartile n
Women (n = 2,907)
57.6±6.0 22.4 143.2±18.7 88.7±9.6 16.6 10.9 4.1±0.9 26.1±3.5 103.2±15.6 (n = 1,910)
57.4±5.9 21.1 139.9±19.1 85.6±9.0 14.9 5.7 4.2±1.0 25.5±4.2 99.4±14.2 (n = 2,854)
47.0 89.9 491
70.3 123.0 727
59.8 78.4 488
70.7 89.3 726
Values are presented as mean ± SD unless otherwise stated. BP = Blood pressure.
atrial fibrillation at the time of stroke (n = 9) or within 6 months after the stroke event (n = 7). A sensitivity analysis was performed, in which cases with incidence of atrial fibrillation was censored at the time of atrial fibrillation diagnosis. However, the results for cardioembolic stroke remained significant when adjusted for risk factors (HR: 2.03, 95% CI: 1.40–2.94 per 1 SD; HR: 5.51, 95% CI: 1.24–24.61 for 4th vs. 1st quartile of NT-proBNP and HR: 1.99, 95% CI: 1.40–2.83 per 1 SD; HR: 3.72, 95% CI: 1.03–13.45 for 4th vs. 1st quartile of MR-proANP). Notably, of the 29 cardioembolic stroke cases with atrial fibrillation before or within 6 months after stroke, 79% were in the top quartile of either NT-proBNP (59%) or MR-proANP (69%). Berntsson/Zia/Borné/Melander/Hedblad/ Engström
Downloaded by: Siriraj Medical Library, Mahidol University 202.28.191.34 - 3/11/2015 2:25:31 PM
Results
Table 2. Adjusted HRs and 95% CIs for incident ischemic stroke by NT-proBNP
Quartiles of NT-proBNP
Q1
Total ischemic stroke Incidence (n) 2.32 (43) Age and sex 1 Risk factors1 1 Total noncardioembolic stroke n 40 Age and sex 1 Risk factors1 1 Cardioembolic stroke n 3 Age and sex 1 Risk factors1 1 1 Adjusted
p trend
log NT-proBNP, per 1 SD
Q2
Q3
Q4
2.93 (53) 1.05 (0.70–1.57) 1.12 (0.74–1.68)
2.81 (51) 0.93 (0.62–1.40) 0.98 (0.65–1.49)
4.52 (80) 1.36 (0.93–1.98) 1.30 (0.88–1.91)
0.124 0.238
1.17 (1.02–1.34) 1.13 (0.99–1.30)
48 1.02 (0.67–1.56) 1.10 (0.72–1.68)
40 0.79 (0.51–1.23) 0.85 (0.54–1.32)
55 1.01 (0.67–1.54) 0.97 (0.63–1.48)
0.810 0.965
1.00 (0.86–1.17) 0.98 (0.84–1.14)
5 1.37 (0.33–5.77) 1.42 (0.34–6.00)
11 2.78 (0.77–10.05) 2.79 (0.77–10.12)
25 5.77 (1.72–19.38) 5.64 (1.66–19.20)
