Thoracic aortic calcification is associated with incident stroke in the general population in addition to established risk factors

Transcription

1 European Heart Journal Cardiovascular Imaging (2015) 16, doi: /ehjci/jeu293 Thoracic aortic calcification is associated with incident stroke in the general population in addition to established risk factors Dirk M. Hermann 1 *, Nils Lehmann 2, Janine Gronewold 1, Marcus Bauer 3, Amir A. Mahabadi 3, Christian Weimar 1, Klaus Berger 4, Susanne Moebus 2, Karl-Heinz Jöckel 2, Raimund Erbel 3, and Hagen Kälsch 3, on behalf of the Heinz Nixdorf Recall Study Investigative Group 1 Department of Neurology, University Hospital Essen, Hufelandstr. 55, D Essen, Germany; 2 Institute of Medical Informatics, Biometry and Epidemiology, University of Duisburg-Essen Essen, Germany; 3 Department of Cardiology, University Hospital Essen, Essen, Germany; and 4 Institute of Epidemiology and Social Medicine, University of Münster, Münster, Germany Received 25 September 2014; accepted after revision 26 November 2014; online publish-ahead-of-print 30 December 2014 Aims The aorta is a major source of cerebral thromboembolism, but its role in stroke pathogenesis is not well understood due to its poor accessibility for non-invasive imaging. We examined whether thoracic aortic calcification (TAC), a marker of aortic plaque load, is associated with stroke in addition to established risk factors.... Methods A total of 3930 subjects from the population-based Heinz Nixdorf Recall study (45 75 years; 47.1% men) without and results previous stroke, coronary heart disease, or myocardial infarction were evaluated for incident stroke events over months. Cox proportional hazards regressions were used to examine associations with stroke of TAC in addition to established risk factors (age, sex, systolic blood pressure, LDL, HDL, diabetes, and smoking) and coronary artery calcification (CAC). 101 incident strokes occurred during the follow-up period. Subjects suffering a stroke had significantly higher TAC values at baseline than the remaining subjects (median ¼ 83.1 [Q1;Q3 ¼ 4.7;472.9] vs [0.0;117.1]; P, 0.001). In a multivariable Cox proportional hazards regression, log(tac + 1) (hazards ratio [HR] ¼ 1.09 [95% confidence interval ¼ ]; P ¼ 0.044) was associated with stroke in addition to established risk factors. Further analyses revealed that log(dtac + 1), i.e. calcification of the descending aorta (1.11 [ ]; P ¼ 0.016), but not log(atac + 1), i.e. calcification of the ascending aorta (1.02 [ ]; P ¼ 0.713), was associated with stroke. The HR for log(tac + 1) decreased to 1.06 ( ; P ¼ 0.202), when log(cac + 1) was also inserted into multivariable analyses.... Conclusion Calcification of the thoracic aorta, more specifically its descending segment, is associated with incident stroke in addition to established risk factors. CAC outperforms aortic calcification as a stroke predictor Keywords Aortic plaque Risk stratification Stroke prediction Subclinical atherosclerosis Introduction The detection of subclinical atherosclerosis is an important challenge in vascular medicine, which aims at the identification of subjects at risk for vascular events that might benefit from individualized disease prevention strategies. In addition to ultrasound techniques, which allow to evaluate atherosclerosis of the carotid arteries, 1 3 the heart, 4,5 and the most proximal part of the ascending aorta, 6,7 computed tomography (CT) has more recently been used for assessing subclinical atherosclerosis. Thus, population-based studies showed that coronary artery calcification (CAC) predicts the risk of incident myocardial infarction 8,9 and stroke 10 in addition to established risk factors. * Corresponding author. Tel: ; Fax: , dirk.hermann@uk-essen.de D.M.H. and N.L. equally contributed. Published on behalf of the European Society of Cardiology. All rights reserved. & The Author For permissions please journals.permissions@oup.com.

2 Aortic calcification is associated with stroke 685 The aorta is a major source of cerebral thromboembolisms due to its huge surface that is particularly prone to vulnerable plaques. 7 Aortic atherosclerosis is highly prevalent in stroke patients: in transoesophageal echocardiography (TOE),.40% of subjects exhibit aortic plaques. 11,12 In subjects, who have already suffered a stroke, aortic atheroma measuring 4 mm in diameter, which is found in 20 30% of stroke patients, 12 confers a particularly high stroke recurrence risk of 11.9% per year. 6 The aorta is poorly accessible for noninvasive imaging. TOE is unable to inspect the aorta over its entire length. Thus, the role of aortic plaques in stroke pathogenesis is not well understood. In the general population, information about health risks associated with subclinical aortic atherosclerosis is lacking. CT has recently been employed for evaluating aortic calcification, which is a highly sensitive, investigator-independent marker of subclinical atherosclerosis that offers itself for epidemiological studies. In a clinical cohort of 2303 asymptomatic subjects (mean age: 56 years) followed up over 4.4 years, CAC, but not thoracic aortic calcification (TAC), predicted coronary heart disease (CHD) after Framingham risk score (FRS) adjustment. 13 In the Multi-Ethnic Study of Atherosclerosis (MESA), a population-based cohort of 6807 subjects (mean age: 62 years) followed up over 4.5 years, TAC did not predict CHD in the whole cohort after adjustment for traditional risk factors. 14 In subanalyses, a predictive role of TAC for CHD was found in women, but not in men. 14 In the Heinz Nixdorf Recall (HNR) cohort, a population-based study on 4814 subjects aged years that evaluates the role of risk factors and subclinical atherosclerosis markers in the development of overt vascular diseases, TAC predicted CHD in addition to FRS variables. 15 This predictive role disappeared, when CAC was also included into multivariable analyses. 15 Whether TAC predicts stroke risk has hitherto not been examined. The aorta is a source of embolism to the brain but except for its most proximal part not the heart, which suggests a causal role of TAC in stroke, but not in CHD. To clarify how TAC affects stroke risk, we examined the association between TAC and incident stroke events in the HNR cohort. Based on its age profile and specific focus on vascular pathologies CT was routinely performed on occasion of the baseline examination the HNR study is well suited for analysing stroke predictors. 10 Methods Study cohort The HNR cohort is a random sample of men and women aged years who were prospectively enrolled via mandatory citizen registries in Essen, Bochum, and Mülheim/Ruhr, three cities of the industrialized Ruhr area, between December 2000 and August All subjects gave informed consent. Exclusion criteria were inability or unwillingness to give informed consent, conditions (medical or other) precluding follow-up over 5 years, severe psychiatric disorders or illegal substance abuse, and pregnancy in women. Computer-assisted interviews and questionnaires were used to document medical history. The study was approved by the institutional review board. Of 4814 subjects included, 4356 subjects had a negative history for previous stroke, CHD, and myocardial infarcts. Of these, 3930 subjects obtained TAC and CAC measurements in addition to FRS assessments. Subjects were followed up over months. During that time, stroke events (both ischaemic and haemorrhagic), defined as focal neurological deficits of presumed cerebrovascular origin that persisted overa period of 24 h, were assessed in annual questionnaires and a follow-up visit after 5 years. Stroke events were validated by an internal and external endpoint committee [K.B., Martin Dichgans (Ludwig Maximilians University Munich, Germany), and C.W.] that provided consensus decisions in case of disagreements. Stroke events were allocated to the date of stroke diagnosis, and non-stroke cases were censored at the date of last contact when the person was still alive or date of death. Trial of Org in Acute Stroke Treatment (TOAST) classifications were also provided. 16 Established risk factors Systemic blood pressure was measured with an automated oscillometric blood pressure device (Omron 705-CP, Omron, Mannheim, Germany), taking the mean of the second and third of three measurements with a minimum of 3 min between the measurements. In some cases, automated blood pressure values were missing. Then, measurements from a random-zero blood pressure device (Mark II, HawksleyTechn., Lancing, UK) were used. Hypertension was classified according to Joint National Committee (JNC)-7 thresholds. 17 Participants were considered diabetic in cases of physician-diagnosed diabetes, having a blood glucose of.200 mg/dl or fasting glucose of.126 mg/dl or taking anti-diabetic medication. For evaluating consequences of nicotine abuse, only current smoking was considered. Total, LDL and HDL cholesterol, and triglycerides were measured with standardized enzymatic methods. Antihypertensive, lipid-lowering, anti-diabetic, and anti-platelet medications were noted. With the data obtained, the FRS was determined. 18 Standardized height and weight measurements were used for calculating the body mass index. Thoracic aortic calcification and coronary artery calcification Non-enhanced CT scans were performed with a C-150 and C-100 electron-beam CT scanner (GE Imatron, South San Francisco, CA, USA), which imposes a radiation dose of msv/examination. Prospective electrocardiogram triggering was done at 80% of the RR interval. Contiguous 3-mm-thick slices were obtained from the level of the pulmonary artery bifurcation to the apex of the heart at an image acquisition time of 100 ms. Within this volume, TAC and CAC were defined as a focus of at least four contiguous pixels with a CT density of 130 Hounsfield units. Agatston scores were computed by summing weighted TAC and CAC scores of all foci in the regions of interest, 19,20 which in case of TAC were also evaluated for the ascending (ATAC) and descending (DTAC) aortic segments. The TAC and CAC scores were communicated neither to the participants nor to their physicians. Statistical analysis Continuous data are presented as mean + SD (normally distributed data) or median (Q1;Q3; non-normally distributed data), and categorical data as counts (%). Normally distributed data were analysed by unpaired t-tests, and non-normally distributed data by Mann Whitney tests. Chi-squared and Fisher s exact tests, as appropriate, were used for comparison of categorical variables. Cox proportional hazards regressions, c-statistics (Harrell s c), category-free net reclassification improvement (NRI), and integrated discrimination improvement (IDI) for time-toevent data were calculated, in which vascular risk factors alone and in combination with TAC and CAC were evaluated. Log-rank tests of trend within Kaplan Meier survival analysis were used to evaluate the effect of TAC categories in defined participant groups. Incidence rates and hazards ratios (HRs) were computed with their 95% confidence

3 686 D.M. Hermann et al. intervals (CIs). P-values of,0.05 indicate statistical significance. Analyses were performed using SAS 9.2 (SAS Institute, Cary, NC, USA), and Harrell s c-statistics were evaluated by STATA/IC Results Baseline characteristics The baseline characteristics of the 3930 HNR participants receiving TAC and CAC measurements in addition to FRS assessments are summarized in Table 1. A total of 101 subjects (60 men and 41 women) developed a stroke during the follow-up (92 ischaemic and 9 haemorrhagic). Of the ischaemic strokes, 12 were of presumable macroangiopathic, 12 of presumable microangiopathic, 26 of presumable cardioembolic, and 41 of unknown aetiology. Participants experiencing a stroke were older, more often revealed arterial hypertension, had higher FRS, and higher TAC and CAC scores than subjects not suffering a stroke. The median age of patients experiencing a stroke event during the follow-up period was 65 years. All acquired risk factors were more prevalent in subjects.65 than 65 years. The percentage of current smokers was lower in subjects.65 than,65 years, whereas the TAC and CAC scores were higher (not shown). Anti-hypertensives, lipid-lowering drugs, anti-diabetics, and platelet inhibitors were more frequently prescribed in old than young subjects (not shown). Aortic calcification is associated with stroke in addition to established risk factors Both TAC and CAC revealed a skew distribution with a high shoulder at the 0 value and decreasing frequencies towards higher values (median TAC ¼ 22.0 [Q1 ¼ 0.0;Q3 ¼ 147.5] in men; 12.6 [0.0;95.4] in women). Of 3930 participants (37.2%), 1463 had no TAC (27.1%) and 426 (11.4%) participants revealed TAC 100 and 400, respectively. To understand how TAC influences stroke risk, we first examined stroke incidence rates for TAC categories. In participants without TAC, stroke incidence was low (0.75 per 1000 person-years in men and 0.98 per 1000 person-years in women). Stroke incidence increased with detection of TAC in men (1.55, 2.03, and 5.16 per 1000 person-years for TAC ¼ 1 99, , and 400, respectively) and in women (0.85, 1.29, and 2.71 per 1000 person-years for TAC ¼ 1 99, , and 400, respectively). To elucidate how TAC affects stroke risk in addition to established risk factors, we next performed Cox proportional hazards regression analyses, in which age, sex, systolic blood pressure, LDL, HDL, diabetes, smoking status, and log(tac + 1) were included (Table 2). In an unadjusted regression, a regression adjusted for age and sex, and in a regression fully adjusted for age, sex, and all other FRS variables, log(tac + 1) was associated with stroke (fully adjusted HR ¼ 1.09 [ ]; P ¼ 0.044). In analyses, in which calcification of the ascending or descending aorta was evaluated in addition to age, sex, and FRS variables, log(dtac + 1), i.e. calcification of the descending aorta (fully adjusted HR ¼ 1.11 [ ]; P ¼ 0.016), but not log(atac + 1), i.e. calcification of the ascending aorta (fully adjusted HR ¼ 1.02 [ ]; P ¼ 0.713), was associated with stroke (Table 2). Additional analyses revealed that log(tac + 1) was associated with stroke in men (fully adjusted HR ¼ 1.14 [ ]; P ¼ 0.025), but not in women (fully adjusted HR ¼ 1.03 [ ]; P ¼ 0.707; see Supplementary data online, Table S1). Furthermore, log(tac + 1) was associated with stroke in subjects 65 years (fully adjusted HR ¼ 1.20 [ ]; P ¼ 0.002), but not in subjects.65 years (fully adjusted HR ¼ 1.02 [ ]; P ¼ 0.709; see Supplementary data online, Table S2). Association of aortic calcification with stroke is attenuated in the presence of CAC Spearman rank correlations revealed that TAC and CAC values moderately correlated with each other (r ¼ 0.37). In a linear model, log(cac + 1) explained 14.8% of the variance of log(tac + 1), suggesting 85.2% of unexplained variance. To investigate how log(tac + 1) influences stroke risk when information about CAC is available, we prepared multivariable regressions, in which all FRS variables (as above), log(tac + 1), and log(cac + 1) were included. Insertion of log(cac + 1) into statistical analyses reduced the HR of log(tac + 1) to 1.06 ( ; P ¼ 0.202). Aortic calcification is a moderate stroke predictor To further elucidate the predictive value of TAC and CAC, we calculated Harrell s c-statistics (AUC(t)), category-free NRI, and IDI for time-to-event data to analyse the benefit of adding TAC and CAC to the model based on FRS variables (age, sex, systolic blood pressure, LDL, HDL, diabetes, and smoking). This analysis revealed that the model including FRS variables and log(tac + 1) (AUC(t) ¼ 0.735) predicted stroke slightly better than the model including FRS variables only (0.729), as did the model including FRS variables, log(tac + 1), and log(cac + 1) (0.744) than the model including FRS variables only. This increase by (95% CI to 0.020) and (95% CI to 0.037) was not significant (P ¼ and 0.190, respectively). IDI for the model including TAC in addition to FRS variables was (95% CI to 0.005; P ¼ 0.222), and IDI for the model including TAC and CAC in addition to FRS variables was (95% CI to 0.010; P ¼ 0.262). Category-free NRI was 11.65% (95%CI 28.03to31.33; P ¼ 0.248)forthe model including TAC and 26.74% (95% CI ; P ¼ 0.008) for the model including TAC and CAC. These data indicated that the value of TAC for risk classification is limited in the general population. Aortic calcification discriminates stroke risk in subjects with a low and intermediate vascular risk profile To elucidate how TAC modifies stroke risk in the presence of vascular risk factors, we formed compound risk groups based on TAC and FRS categories. Inserting these groups into Cox regression revealed that, compared with subjects belonging to the low FRS (,10%) and lowest TAC (0) category, subjects of the high FRS (.20%) and highest TAC ( 400) category carried a 9.21-fold stroke risk (Figure 1). Log-rank tests for trend showed that, in subjects belonging to the low (,10%) or intermediate (10 20%) FRS category, stroke risk increased with increasing TAC category (P, and 0.035,

4 Aortic calcification is associated with stroke 687 Table 1 Baseline characteristics of the study population Men (n )... Women (n )... Stroke (n 5 60) No stroke (n ) P-value Stroke (n 5 41) No stroke (n ) P-value... Age (years) , ,0.001 BMI (kg/m 2 ) Systolic BP (mmhg) , ,0.001 Diastolic BP (mmhg) Hypertension (JNC 7) Normal or prehypertension (%) 1 (1.7) 253 (14.1) 6 (14.6) 727 (35.7) Stage 1 (%) 40 (66.6) 1277 (71.3), (63.4) 1150 (56.4),0.001 Stage 2 (%) 19 (31.7) 261 (14.6) 9 (22.0) 161 (7.9) Anti-hypertensive drugs (%) 21 (35.0) 531 (29.7) (58.5) 654 (32.1),0.001 Diabetes (%) 14 (23.3) 274 (15.3) (12.2) 186 (9.1) Anti-diabetic drugs (%) 8 (14.8) 98 (5.9) (7.7) 78 (4.1) Total cholesterol (mg/dl) LDL cholesterol (mg/dl) HDL cholesterol (mg/dl) Lipid-lowering drugs (%) 4 (6.7) 143 (7.9) (4.9) 199 (9.8) Anti-platelet drugs (%) 6 (11.1) 112 (6.7) (7.7) 95 (4.9) Smoking status (%) Never smoked (%) 15 (25.0) 534 (29.8) 22 (53.6) 1147 (56.2) Former smoking (%) 28 (46.7) 804 (44.9) (24.4) 464 (22.8) Current smoking (%) 17 (28.3) 453 (25.3) 9 (22.0) 427 (21.0) FRS (median [Q1;Q3]) 18.0 (14.0;24.5) 14.0 (9.0;22.0), (7.0;15.0) 7.0 (4.0;11.0) 0.003,10% 6 (10.0) 574 (32.1) 25 (60.9) 1505 (73.9) 10 20% 28 (46.7) 769 (42.9), (34.2) 482 (23.6) % 26 (43.3) 448 (25.0) 2 (4.9) 51 (2.5) TAC score (median [Q1;Q3]) 99.7 (17.6;569.8) 20.9 (0.0;140.3), (0.0;269.7) 12.5 (0.0;91.5) ATAC score (median [Q1;Q3]) 14.3 (0.0;141.7) 0.0 (0.0;31.7) (0.0;35.1) 0.0 (0.0;16.8) DTAC score (median [Q1;Q3]) 72.6 (0.9;334.9) 6.7 (0.0;64.0), (0.0;160.7) 5.2 (0.0;48.0) Unless otherwise indicated, data are mean + SD for continuous data and n (%) for categorical data. ATAC, calcification of the ascending aorta; BMI, body mass index; BP, blood pressure; DTAC, calcification of the descending aorta; FRS, Framingham risk score; HDL, high-density lipoprotein; JNC, Joint National Committee; LDL, low-density lipoprotein. Table 2 Cox proportional hazards regression analyses for total cohort Risk factors Unadjusted HR P-value HR adjusted for age P-value Fully adjusted P-value (95% CI) and sex (95% CI) HR (95% CI)... Age (per 5 years) 1.53 ( ),0.001 Sex (male vs. female) 1.67 ( ) Systolic BP (per 10 mmhg) 1.36 ( ), ( ),0.001 LDL (per 10 mg/dl) 1.03 ( ) ( ) HDL (per 5 mg/dl) 0.97 ( ) ( ) Diabetes (yes vs. no) 1.78 ( ) ( ) Smoking (yes vs. no) 1.21 ( ) ( ) log(tac + 1) 1.26 ( ), ( ) ( ) log(atac + 1) 1.17 ( ), ( ) ( ) log(dtac + 1) 1.26 ( ), ( ), ( ) HR, hazards ratio; CI, confidence interval; other abbreviations as in Table 1.

5 688 D.M. Hermann et al. respectively). Thus, TAC detected subjects with high stroke incidence in this low or intermediate FRS category. In the high FRS category, TAC did not discriminate stroke risk (P ¼ 0.222). Figure 1 Stroke riskin subjects belonging tothehnr studystratified on FRS and TAC categories. HRs of stroke events in the different combinations of FRS and TAC categories are shown, with the lowest TAC and FRS category as a reference. For the low and intermediate FRS category, log-rank tests for trends revealed significant differences between TACcategories, indicating that TACdiscriminates incident stroke insubjects at a low and intermediate vascular risk. Numbers of participants in each group: TAC ¼ 0/FRS, 10%: 955 participants; TAC ¼ 0/ 10 FRS 20%: 399 participants; TAC ¼ 0/FRS. 20%: 109 participants. 0, TAC, 100/FRS, 10%: 796 participants; 0, TAC, 100/10 FRS 20%: 466 participants; 0, TAC, 100/ FRS. 20%: 139 participants. 100 TAC, 400/FRS, 10%: 247 participants; 100 TAC, 400/10 FRS 20%: 254 participants; 100 TAC, 400/FRS. 20%: 119 participants. TAC 400/FRS, 10%: 112 participants; TAC 400/10 FRS 20%: 174 participants; TAC 400/FRS. 20%: 160 participants. Table 3 Aortic calcification is associated with ischaemic stroke of microangiopathic, cardioembolic, and unknown aetiology To further characterize the effect of TAC on incident stroke, we calculated incidence rates for various stroke aetiologies (Table 3). Log-rank tests for trend revealed that TAC discriminated the risk of ischaemic stroke of microangiopathic (P ¼ 0.001), cardioembolic (P, 0.001), and unknown (P ¼ 0.041) origin and the risk of haemorrhagic stroke (P ¼ 0.007), but not the risk of ischaemic stroke of macroangiopathic (P ¼ 0.334) origin. Discussion Incidence rates for different stroke aetiologies depending on TAC categories Using a sample of 3930 subjects aged years, we for the first time show that TAC is independently associated with incident stroke in the general population in addition to established risk factors that are part of the FRS. Interestingly, DTAC, i.e. calcification of the descending aorta, was more closely associated with stroke than ATAC, i.e. calcification of the ascending aorta. The association between TAC and stroke became weaker, when CAC was inserted into multivariable analyses, suggesting that TAC is a moderate stroke predictor. TAC was associated with stroke specifically in young subjects and subjects with a low or intermediate vascular risk profile. The association of TAC with incident CHD, but not with incident stroke, has previously been examined. In a primary care-based cohort of 2303 asymptomatic subjects (mean age: 56 years), in which 41 CHD events were detected during a follow-up of 4.4 years, CAC but not TAC was associated with CHD after FRS adjustment. 13 Similarly, in 6807 subjects belonging to the population-based MESA (mean age: 62 years), in which 232 CHD events were noted during 4.5-year follow-up, TAC was not associated with CHD after adjustment for established risk factors. 14 In subanalyses, TAC was found to be associated with CHD in women, but not in men. 14 In the HNR TAC 5 0 TAC TAC TAC Ischaemic, all 1.65 ( ) [22/13 309] Ischaemic, presumed macroangiopathic aetiology 0.30 ( ) [4/13 309] Ischaemic, presumed microangiopathic aetiology 0.15 ( ) [2/13 309] Ischaemic, presumed cardioembolic aetiology 0.38 ( ) [5/13 309] Ischaemic, unknown aetiology 0.83 ( ) [11/13 309] Haemorrhagic 0.08 ( ) [1/13 309] 2.09 ( ) [27/12 923] 0.31 ( ) [4/12 923] 0.15 ( ) [2/12 923] 0.54 ( ) [7/12 923] 1.08 ( ) [14/12 923] 0.31 ( ) [4/12 923] 3.13 ( ) [17/5428] 0.18 ( ) [1/5428] 0.55 ( ) [3/5428] 0.74 ( ) [4/5428] 1.66 ( ) [9/5428] 0.18 ( ) [1/5428] 6.78 ( ) [25/3685] 0.81 ( ) [3/3685] 1.36 ( ) [5/3685] 2.71 ( ) [10/3685] 1.90 ( ) [7/3685] 1.09 ( ) [4/3685] Ischaemic strokes were further classified according to the TOAST criteria. Data are incidence rates with 95% confidence intervals calculated as the number of stroke cases per totalperson-timeat risk inyearsmultiplied with1000 (strokesper1000person-years; absolutestrokeevents andperson-years areshowninsquare brackets). Log-rank tests for trend revealed that TAC discriminated the risk of ischaemic stroke (P, 0.001), the risk of ischaemic stroke of microangiopathic (P ¼ 0.001), cardioembolic (P, 0.001), and unknown (P ¼ 0.041) aetiology, as well as the risk of haemorrhagic stroke (P ¼ 0.007), but not the risk of ischaemic stroke of macroangiopathic aetiology (P ¼ 0.334). Abbreviations as in Tables 1 and 2.

6 Aortic calcification is associated with stroke 689 cohort, a population-based study, TAC was associated with CHD in addition to FRS variables in 4093 subjects aged years. 15 While the size of the first study was probably too small, the latter two studies were adequately powered for evaluating associations of TAC and CHD. Except for its initial segment, the aorta is a brainsupplying, but not a heart-supplying artery. Thus, thromboembolisms arising from the aorta may directly contribute to stroke. Associations of TAC with stroke have previously been shown in cross-sectional analyses within the Rotterdam study, 21 but not in longitudinal studies. The finding that DTAC, but not ATAC, was associated with stroke may have two different reasons: first, the descending aorta typically exhibits a higher plaque load than the ascending aorta, which has previously been reported both in MESA 22 and in an earlier paper from the HNR study. 20 Owing to the higher plaque load, DTAC may be a more sensitive surrogate marker of generalized atherosclerosis than ATAC. Secondly, in view of the backflow of blood during diastole, thrombi attached to descending thoracic aortic plaques might embolize into cerebral blood vessels. During diastole, blood flows back 30% of the distance moved during systole with a velocity of up to 25 cm/s, 23 which is due to the regurgitation of blood back to the heart before aortic valves close. Thus, we speculate that DTAC may not only represent a surrogate marker of stroke risk related to large-sized artery atherosclerosis, but represent a potential stroke origin. 24 Neurologists should be aware of the possible role of the descending aorta in stroke pathogenesis. Insertion of CAC into multivariable Cox proportional hazards regression analyses reduced the association of TAC and stroke, which is in line with previous observations with respect to CHD in the HNR cohort, where CAC outperformed TAC when both atherosclerosis markers were combined in multivariable analyses. 15 In reclassification analyses, the insertion of TAC into models containing FRS variables increased AUC(t), IDI, and NRI only modestly in our present study. In log-rank tests for trend, TAC discriminated incident stroke specifically in subjects belonging to the low or intermediate, but not the high FRS category. In this respect, TAC resembles CAC, which also distinguished incident stroke in subjects belonging to the low or intermediate, but not in the high FRS risk category, 10 and it differs from intima-media thickness 3 and ankle-brachial index, 25 which discriminated incident stroke in the high FRS risk category and intermediate and high FRS risk category, respectively. Hence, TAC and CAC appear todistinguishstrokeincidenceparticularlyinsubjectswithalowrisk profile, which may explain why TAC predicted stroke in young, but not in old, subjects. From these observations, it may be concluded that CAC and TAC provide information about endogenous (potentially genetic) atherosclerosis propensity, which gets masked with the development of clinically overt vascular risk factors. In log-rank tests for trend, TAC discriminated ischaemic stroke of presumable microangiopathic, cardioembolic, and unknown origin. This observation on the one hand suggests an association of aortic atherosclerosis with cerebral small vessel disease and CHD. It on the other hand indicates that TAC might provide hints for a possible aortic stroke origin in subjects, in which stroke aetiology is unknown. In view that the aorta can poorly be inspected by echocardiography, CT is an attractive tool for evaluating aortic atherosclerosis. The big strength of CT is that TAC and CAC can simultaneously be evaluated in the same scan, offering comprehensive information about subclinical atherosclerosis in different vascular territories. The drawback is that TAC and CAC do not provide information about plaque morphology and the presence of non-calcified plaques that are particularly prone for thrombus formation. In the clinical setting, non-calcified plaques may also elegantly be evaluated by on CT-based grounds. Based on the here presented data, the evaluation of TAC and CAC might be considered in young stroke patients belonging to the low or intermediate FRS category, in which stroke aetiology remains unknown based on standard examinations. In these patients, the presence of TAC and CAC might argue in favour of a cardiac or aortic stroke origin. In head-to-head comparison with CAC, TAC provides little added information. This limitation reduces the clinical utility of TAC as a risk predictor. Supplementary data Supplementary data are available at European Heart Journal Cardiovascular Imaging online. Acknowledgements We thank the Investigative Group, the staff, and all participants. Apart from the authors, the following persons are essentially involved in the HNR: D. Baumgart, H. Hirche, U. Slomiany (all Essen), J. Siegrist (Düsseldorf), R. Peter (Ulm), A. Schmermund (Frankfurt), and A. Stang (Halle). Conflict of interest: none declared. Funding This research was supported by the Heinz Nixdorf Foundation. References 1. del Sol AI, Moons KG, Hollander M, Hofman A, Koudstaal PJ, Grobbee DE et al. Is carotid intima-media thickness useful in cardiovascular disease risk assessment? The Rotterdam Study. Stroke 2001;32: Ohira T, Shahar E, Iso H, Chambless LE, Rosamond WD, Sharrett AR et al. Carotid artery wall thickness and risk of stroke subtypes: the atherosclerosis risk in communities study. Stroke 2011;42: Hermann DM, Gronewold J, Lehmann N, Seidel UK, Möhlenkamp S, Weimar C et al. Intima-media thickness predicts stroke risk in the Heinz Nixdorf Recall study in association with vascular riskfactors, age andgender. Atherosclerosis 2012;224: Kizer JR, Wiebers DO, Whisnant JP, Galloway JM, Welty TK, Lee ET et al. Mitral annularcalcification, aortic valve sclerosis, and incident stroke in adults free of clinical cardiovascular disease: the Strong Heart Study. Stroke 2005;36: De Marco M, Gerdts E, Casalnuovo G, Migliore T, Wachtell K, Boman K et al. Mitral annular calcification and incident ischemic stroke in treated hypertensive patients: the LIFE study. Am J Hypertens 2013;26: The French Study of Aortic Plaques in Stroke Group. Atherosclerotic disease of the aortic arch as a risk factor for recurrent ischemic stroke. The French Study of Aortic Plaques in Stroke Group. N Engl J Med 1996;334: Macleod MR, Amarenco P, Davis SM, Donnan GA. Atheroma of the aortic arch: an important and poorly recognised factor in the aetiology of stroke. Lancet Neurol 2004;3: Erbel R, Möhlenkamp S, Moebus S, Schmermund A, Lehmann N, Stang A et al. Coronary risk stratification, discrimination, and reclassification improvement based on quantification of subclinical coronary atherosclerosis: the Heinz Nixdorf Recall study. J Am Coll Cardiol 2010;56: Blaha MJ, Budoff MJ, DeFilippis AP, Blankstein R, Rivera JJ, Agatston A et al. Associations between C-reactive protein, coronary artery calcium, and cardiovascular events: implications for the JUPITER population from MESA, a population-based cohort study. Lancet 2011;378: Hermann DM, Gronewold J, Lehmann N, Moebus S, Jöckel KH, Bauer M et al. Coronary artery calcification is an independent stroke predictor in the general population. Stroke 2013;44: Toyoda K, Yasaka M, Nagata S, Yamaguchi T. Aortogenic embolic stroke: a transesophageal echocardiographic approach. Stroke 1992;23:

7 690 D.M. Hermann et al. 12. Di Tullio MR, Sacco RL, Savoia MT, Sciacca RR, Homma S. Aortic atheroma morphology and the risk of ischemic stroke in a multiethnic population. Am Heart J 2000; 139: WongND, Gransar H, Shaw L, Polk D, Moon JH, Miranda-Peats R et al. Thoracicaortic calcium versus coronary artery calcium for the prediction of coronary heart disease and cardiovascular disease events. JACC Cardiovasc Imaging 2009;2: Budoff MJ, Nasir K, Katz R, Takasu J, Carr JJ, Wong ND et al. Thoracic aortic calcification and coronary heart disease events: the multi-ethnic study of atherosclerosis (MESA). Atherosclerosis 2011;215: Kälsch H, Lehmann N, Berg MH, Mahabadi AA, Mergen P, Möhlenkamp S et al. Coronaryartery calcificationoutperforms thoracic aortic calcification for the prediction of myocardial infarction and all-cause mortality: the Heinz Nixdorf Recall study. Eur J Prev Cardiol 2014;21: Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL et al. Classification of subtype of acute ischemic stroke. Definitions for use in a multicenter clinical trial. TOAST. Trial of Org in Acute Stroke Treatment. Stroke 1993;24: Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr et al. The seventh report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA 2003;289: Wilson PW, D Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97: IMAGE FOCUS 19. Schmermund A, Möhlenkamp S, Berenbein S, Pump H, Moebus S, Roggenbuck U et al. Population-based assessment of subclinical coronary atherosclerosis using electron-beam computed tomography. Atherosclerosis 2006;185: Kälsch H, Lehmann N, Möhlenkamp S, Hammer C, Mahabadi AA, Moebus S et al. Prevalence of thoracic aortic calcification and its relationship to cardiovascular risk factors and coronary calcification in an unselected population-based cohort: the Heinz Nixdorf Recall Study. Int J Cardiovasc Imaging 2013;29: Elias-Smale SE, Odink AE, Wieberdink RG, Hofman A, Hunink MG, Krestin GP et al. Carotid, aortic arch and coronary calcification are related to history of stroke: the Rotterdam study. Atherosclerosis 2010;212: TakasuJ, BudoffMJ, O Brien KD, ShavelleDM, ProbstfieldJL, CarrJJet al. Relationship between coronary artery and descending thoracic aortic calcification as detected by computed tomography: the Multi-Ethnic Study of Atherosclerosis. Atherosclerosis 2009;204: McDonaldDA. Thevelocityofblood flowintherabbitaortastudiedwithhigh-speed cinematography. J Physiol 1952;118: Harloff A, Simon J, Brendecke S, Assefa D, Helbing T, Frydrychowicz A et al. Complex plaques in the proximal descending aorta: an underestimated embolic source of stroke. Stroke 2010;41: Gronewold J, Hermann DM, Lehmann N, Kröger K, Lauterbach K, Berger K et al. Ankle-brachial index predicts stroke in the general population in addition to classical risk factors. Atherosclerosis 2014;233: doi: /ehjci/jev026 Online publish-ahead-of-print 6 March Extensive myocardial calcification after acute myocarditis Felipe Díez-delhoyo 1 *, Eduardo Zatarain-Nicolás 1, Esther Pérez-David 1, Maria Luisa Sánchez-Alegre 2, and Francisco Fernández-Avilés 2 1 Department of Cardiology, Hospital General Universitario Gregorio Marañón, Instituto de Investigación Sanitaria del Hospital Gregorio Marañón, Complutense University, School of Medicine, C/Dr. Esquerdo 46, Madrid 28007, Spain and 2 Department of Radiology,Hospital General Universitario Gregorio Marañón, Instituto de Investigación Sanitaria del Hospital Gregorio Marañón, Madrid, Spain * Corresponding author. Tel: ; Fax: , felipediezdelhoyo@hotmail.com A 32-year-old male was admitted into the hospital due to acute heart failure. He hada recently diagnosed HIV infection with high viral load, primary haemophagocytic syndrome under immunosuppressive therapy, and CMV infection. Cardiac biomarkers were increased (high-sensitive T-Troponin 1350ng/L; NTproBNP ng/l). The 2D-echocardiogram showed global hypokinaesia and depressed left ventricular ejection fraction (LVEF). A computed tomography (CT) observed normal heart density (Panel A). Cardiac magnetic resonance (CMR) showed mid-wall late gadolinium enhancement (LGE) at the mid-distal interventricular septum (Panels B and C); thus, acute myocarditis was diagnosed. At discharge, full recovery of LVEF was confirmed. Two months later, due to a cryptogenic organizing pneumonia, a new CT demonstrated an extensive myocardial calcification in the middistal septum (Panel D), correlating with the area of LGE in CMR. A new CMR revealed persistence of LGE without myocardial oedema in T2-Stir sequences (Panel E). Transthoracic echocardiogram confirmed the appearance of a septal hyperintensity in the mentioned location (Panel F) and normal LVEF. Patient s follow-up 1 year later has been uneventful, and the cardiac calcification persists. Myocardial calcifications have been related to acute myocarditis in animal models, histological samples, and case reports, though it is infrequent. Viral infections and immunosuppressive status have been proposed as risk factors; nevertheless, history and prognosis are unknown. Progressive resorption of calcium has been described, suggesting that the deposit may be related to myocardial inflammation. Interestingly, calcification has not disappeared during follow-up in our patient despite good clinical course. Published on behalf of the European Society of Cardiology. All rights reserved. & The Author For permissions please journals.permissions@oup.com.