European Journal of Heart Failure (2014) doi:10.1002/ejhf.205
Subclinical left ventricular dysfunction by echocardiographic speckle-tracking strain analysis relates to outcome in sarcoidosis Emer Joyce1,2, Maarten K. Ninaber3, Spyridon Katsanos1, Philippe Debonnaire1,4, Vasilis Kamperidis1, Jeroen J. Bax1, Christian Taube3, Victoria Delgado1, and Nina Ajmone Marsan1* 1 Department
of Cardiology, Leiden University Medical Center, Leiden, The Netherlands; 2 Department of Advanced Heart Disease, Brigham and Womens Hospital, Boston, MA, USA; 3 Department of Pulmonology, Leiden University Medical Center, Leiden, The Netherlands; and 4 Department of Cardiology, Sint-Jan Hospital Bruges, Bruges, Belgium Received 10 August 2014; revised 18 September 2014; accepted 24 October 2014
Aims
Limited data exist on the risk of developing cardiac sarcoidosis (CS) and/or adverse events in sarcoidosis patients. Using LV global longitudinal strain (GLS), an emerging sensitive parameter of LV function, we evaluated the prevalence of subclinical cardiac dysfunction in sarcoidosis and investigated whether LVGLS predicts adverse outcomes in this population. ..................................................................................................................................................................... Methods A total of 130 patients with proven sarcoidosis undergoing echocardiography at our referral centre were identified. and results Following exclusion of those with evidence of CS (n = 14) or other pre-existing structural heart disease (n = 16), 100 patients (55 ± 13 years, 48% male, 90% pulmonary involvement) and 100 age- and gender-matched controls were included. LVGLS was measured by speckle-tracking analysis. The primary endpoint was a composite of all-cause mortality, heart failure hospitalization, device implantation, new arrhythmias, or future development of CS on advanced cardiac imaging modalities. LVGLS was significantly impaired in sarcoidosis patients compared with controls (–17.3 ± 2.5 vs. –20.0 ± 1.6%, P < 0.001). Overall, 27 patients (27%) reached the endpoint during a median follow-up of 35 months. On Cox proportional hazards model analysis, abnormal 24-h Holter, larger LV end-diastolic diameters, and more impaired LVGLS were significantly associated with the endpoint; however, only LVGLS remained independently associated on multivariate analysis [hazard ratio (HR) 1.4, 95% confidence interval (CI) 1.1–1.7, P = 0.006]. Patients with LVGLS less than –17.3% were significantly more likely to be free of the primary endpoint (log-rank P = 0.01). ..................................................................................................................................................................... Conclusion LVGLS is impaired in sarcoidosis patients, suggesting subclinical cardiac dysfunction despite the absence of conventional evidence of cardiac disease, and is independently associated with occurrence of cardiac events and/or development of CS.
.......................................................................................................... Cardiac sarcoidosis • cardiomyopathy
Speckle-tracking analysis •
Introduction Sarcoidosis is a multisystem granulomatous disease of unknown aetiology primarily involving the pulmonary system and lymph nodes but which can also affect the heart. The reported prevalence
.............
Keywords
Longitudinal strain •
Prognosis •
Sarcoid
of cardiac sarcoidosis (CS) varies from 20% up to 60% according to the study population (highest prevalence in Japanese patients) and to the type of study performed (clinical or autopsy report).1 – 3 Antemortem diagnosis of CS is challenging, mainly due to the focal extension of the granulomatous involvement, with only 50% of
*Corresponding author. Department of Cardiology, Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands. Tel: +31 71 526 2020, Fax: +31 71 526 6809, Email: [email protected]
© 2014 The Authors European Journal of Heart Failure © 2014 European Society of Cardiology
2
METHODS Patient population One-hundred and thirty consecutive patients with sarcoidosis attending our tertiary referral centre were identified from the departmental Cardiology Information System (EPD-Vision®, Leiden University Medical Center). The diagnosis of sarcoidosis required a compatible clinical picture with or without histological evidence of non-caseating granulomas and the absence of another disease process capable of producing a similar picture.13 Patients were required to have undergone 2DTTE around the time of, or following, their diagnosis. Patients were excluded if they had known structural heart disease (n = 16) or
..................................................................................................................................................................................
patients diagnosed post-mortem having clinical evidence of myocardial disease and only up to 5% of patients reporting symptoms.3 Cardiac involvement in sarcoidosis is also well known to portend a poor prognosis,3,4 and sudden cardiac death (due to either complete heart block or ventricular arrhythmias) may be its first manifestation, prior to clinical presentation or detection of structural abnormalities by traditional parameters or basic cardiac testing. However, if CS is identified at an early stage, corticosteroid therapy and/or cardiac device therapy may reduce the risk of sudden death and/or progression of LV dysfunction.4,5 Therefore, a highly sensitive and non-invasive method for an earlier diagnosis of CS has the potential for significant clinical value in identifying patients at risk for adverse outcomes. Standard two-dimensional transthoracic echocardiography (2DTTE) is recommended in many algorithms as a first-line screening tool, together with 12-lead ECG and 24-h Holter monitoring, for detection of CS in sarcoidosis patients without manifest cardiac disease.6 However, 2DTTE lacks sensitivity to detect early, mild, or focal cardiac involvement.6,7 Advanced imaging modalities, including 99m Technetium (99m Tc)–tetrofosmin single photon emission computed tomography (SPECT), [18 F]fluorodeoxyglucose positron emission tomography (18 F-FDG PET), or delayed contrast enhancement cardiac magnetic resonance imaging (DE-CMR) have demonstrated increased sensitivity and may facilitate earlier detection of CS.6,8,9 Consequently, abnormalities consistent with CS on these modalities have been invoked as diagnostic confirmation of CS in several recent studies2,10 alongside conventional diagnostic criteria proposed by the Japanese Ministry of Health and Welfare (JMHW).11 However, these modalities are expensive, not universally available, and may not be suitable for all patients. More recently, speckle-tracking analysis, a novel echocardiographic technique which assesses the intrinsic deformation (strain) of the myocardium, has been proposed as a sensitive tool to detect myocardial dysfunction before a decrease in LVEF occurs.12 The primary aim of this study was therefore to evaluate the prevalence of subclinical LV dysfunction in sarcoidosis patients without clinical or standard diagnostic evidence of cardiac involvement using LV global longitudinal strain (LVGLS) compared with a control population. Secondly, given the lack of outcome studies to determine the risk of cardiac events in sarcoidosis patients, including the risk of future development of CS, we sought to investigate whether impaired LVGLS may be associated with greater risk of adverse outcomes.
E. Joyce et al.
definite CS (n = 5) at the time of the first 2DTTE study (date of study enrolment). Regarding those excluded for known structural heart disease, the majority (n = 10) had significant ischaemic heart disease, as defined by >50% stenosis in ≥1 epicardial vessel and/or prior acute coronary syndrome and/or prior revascularization procedure, three due to congenital heart disease, two due to known dilated cardiomyopathy from another cause, and one due to significant valve disease, defined as grade ≥2 regurgitation or moderate or more stenosis. The presence of CS was defined by fulfilment of the histological or clinical diagnosis group criteria of the revised JMHW guidelines.11 Those who fulfilled one major criterion of the updated guidelines at study enrolment (n = 9) were also excluded, as this was considered to indicate high likelihood of future CS diagnosis. Consecutive controls were identified from the departmental echocardiographic database (EchoPac) using a specific echocardiographic search code identifying the absence of any structural cardiac abnormalities. Initial referral for echocardiography included chest pain, dyspnoea, syncope, palpitations, murmur on auscultation, or cardiovascular risk stratification, in the absence of any prior history of structural heart disease. The absence of structural cardiac disease was defined as a ‘normal’ echocardiogram according to American Society of Echocardiography guidelines.14 – 17 An age- and gender-matched comparator group was selected in a 1:1 ratio according to published comparability principles, in that they were representative of the same base experience as cases, but without any prior or current history of sarcoidosis.18 The Institutional Review Board of the Leiden University Medical Center approved the current retrospective evaluation of clinically acquired data, and waived the need for patient written informed consent. The study conforms with the principles outlined in the Declaration of Helsinki.
Study protocol Demographic and clinical data were recorded together with the results of chest radiography, pulmonary function testing, and serum ACE and lysozyme levels. Results of standard cardiac screening testing as currently recommended for sarcoidosis, including 2DTTE, 12-lead ECG, and ambulatory 24-h Holter monitoring, were also noted.6 LVGLS was prospectively measured by 2D speckle-tracking analysis on baseline 2DTTE in order to detect subclinical LV dysfunction. Advanced cardiac imaging, such as DE-CMR, myocardial SPECT, 67 gallium (67 Ga) scintigraphy, and 18 F-FDG PET,6,8 was carried out during follow-up at the discretion of the treating physician. The presence of an abnormality consistent with CS on ≥1 advanced imaging modality was recorded.6,11 During the follow-up period after the first 2DTTE evaluation, survival of all patients was checked by querying the survival statistics of the municipal civil registries of The Netherlands, and the occurrence of other clinical events up to the time of the most recent hospitalization or outpatient clinic visit was recorded. The primary endpoint was a composite of all-cause mortality, heart failure-related hospitalization, cardiac device implantation—permanent pacemaker, implantable cardioverter defibrillator (ICD), and/or biventricular pacemaker—new arrhythmias (sustained ventricular tachycardia not associated with ICD implantation and/or new atrial arrhythmia), or accepted evidence of development of CS on advanced cardiac imaging study.
Clinical testing Pulmonary function tests were recorded as the percentage of predicted values measured according to the American Thoracic © 2014 The Authors European Journal of Heart Failure © 2014 European Society of Cardiology
3
Table 1 Abnormalities on basic clinical cardiac testing considered possible evidence of cardiac sarcoidosis 2,7,17,20
Test Abnormalities ................................................................ 12-Lead ECG
24-h Holter monitoring
Two-dimensional echocardiography
Minor criteria abnormalitiesa Non-minor criteria abnormalities First degree or Mobitz type 1 atrioventricular blockb, c ST–T abnormalities (T-wave inversion ≥2 leads considered significant) Left ventricular hypertrophyd Right ventricular hypertrophy Supraventricular tachyarrhythmia (including atrial fibrillation or flutter) Sinus tachycardiae Prolonged corrected QT interval Mobitz type 1 atrioventricular blockb, c Significant premature ventricular contractions (>1% per 24 h) Non-sustained ventricular tachycardia >3 beatsc Sustained ventricular tachycardiac Supraventricular tachycardia >3 beatsc Regional wall motion abnormalities in ≥2 segments Right ventricular systolic dysfunction in the absence of pulmonary hypertensionf Diastolic dysfunction inappropriate for age
Adapted from Mehta et al.2 a Ventricular arrhythmias (ventricular tachycardia, multifocal or frequent premature ventricular contractions), complete right bundle branch block, axis deviation, or abnormal Q wave).11 b Higher degrees of atrioventricular block were an exclusion criterion at baseline (major criterion). c For Cox proportional hazards model, these abnormalities were only considered in patients who did not have a ‘similar’ arrhythmic or device event included in the primary endpoint at follow-up. d Only considered positive in the absence of hypertension. e In the absence of another obvious cause. f Right ventricular systolic dysfunction was defined by tricuspid annular plane systolic excursion 40 fps to ensure reliable analysis by the software. The LV endocardial border was traced at end-systole in all three apical views and the automatically created region of interest was manually adjusted to the thickness of the myocardium. Tracking quality was assessed in all segments and, if tracking remained of suboptimal quality following manual correction, segments were discarded. If >2 segments needed to be discarded, these patients were excluded from the subsequent analyses using LVGLS. For the remaining patients, LVGLS was then provided by the software as the average value of the peak systolic longitudinal strain of the three apical views, using a 17-segment model, in a bull’s eye plot (Figure 1).
Other cardiac imaging modalities Cardiac magnetic resonance (using a 1.5-T Gyroscan ACS-NT/Intera scanner; Philips Medical Systems, Eindhoven, The Netherlands), myocardial SPECT (using 99m Tc–tetrofosmin, 500 MBq injected at rest), 18 F-FDG imaging (using either PET or dual-isotope SPECT), and 67 Ga scintigraphy were all performed according to standard protocols. On CMR, the presence of delayed enhancement was considered evidence of CS (scarring).6,11 A perfusion defect or defects on myocardial SPECT imaging, defined as regions containing 8 years. Speckle-tracking analysis revealed an LVGLS of –18.3%. (B) A patient with biopsy-confirmed extracardiac sarcoidosis who presented >1 year after baseline 2D echocardiography evaluation with complete heart block requiring permanent pacemaker implantation (primary endpoint). LVGLS was significantly reduced (less negative) at –15.0%. AVC, aortic valve closure; APLAX, apical long axis; 2C, two-chamber; 4C, four-chamber.
© 2014 The Authors European Journal of Heart Failure © 2014 European Society of Cardiology
5
testing parameters, and LVGLS to the primary endpoint. The number of covariables entered into the multivariate model was limited due to the number of endpoint events; however, all variables with a P-value 1 abnormality on basic cardiac testing was significantly more common in the primary endpoint group (73% vs. 48%, P = 0.03). Baseline LVGLS was significantly more impaired (–15.8 ± 2.2 vs. –17.9 ± 2.3%, P < 0.001) in sarcoidosis patients who experienced the primary endpoint compared with those who did not (example in Figure 1).
Univariate and multivariate predictors of the primary endpoint As shown in Table 4, on univariate Cox proportional hazards model analysis, abnormal 24-h Holter ECG, higher LV end-diastolic diameter, and more impaired (more positive) LVGLS were significantly associated with the primary endpoint. Only LVGLS remained significantly and independently associated with the primary endpoint on multivariate analysis [hazard ratio (HR) 1.4, 95% confidence interval (CI) 1.1–1.7, P = 0.006]. The area under the curve for LVGLS to predict the primary endpoint was 0.73 (95% CI 0.62–0.85, P = 0.001) (Figure 2). A cut-off value of –17.3% (also corresponding to the median value) had a 72% sensitivity and 60% specificity to predict the occurrence of the primary endpoint. Those with an LVGLS less than –17.3% were significantly more likely to be free of the primary endpoint as
6
E. Joyce et al.
Table 2 Baseline demographic, clinical, and echocardiographic characteristics of the total patient population and the control population Variable Controls (n = 100) All patients (n = 100) P-value ........................................................................................................................................... Demographic Age, years Male gender, n (%) Hypertension, n (%) Diabetes, n (%) Hyperlipidaemia, n (%) Smoking (ever), n (%) Clinical (sarcoidosis patients only) Histological + clinical diagnosis vs. clinical diagnosis alone, n (%) Organ involvement, n (%) Pulmonary Skin Joint Eye Renal ± hypercalcaemia Neurological system Liver ± gallbladder Parotid gland Other Chest radiograph stage 0 1 2 3 4 Serum ACE, nmol/min/mL Lysozyme, % FVC predicted, % DLCO predicted, % Abnormal 12-lead ECG (overall), n (%) Abnormal 12-lead ECG (JMHW criteria), n (%) Abnormal 24-h Holter monitoring, n (%) NYHA class ≥1, n (%) Immunomodulatory medications at baseline, n (%) Steroids Azathioprine Methotrexate Echocardiographic LVEDD, cm LVESD, cm IVSd, cm PWd, cm Relative wall thickness, cm LVESV, mL LVEDV, mL LVEF, % WMSI MR grade ≥2 LAVI, mL/m2 E, cm/s A, cm/s E/A DT, ms Septal E’, m/s
55 ± 13 48 (48%) 27 (27%) 8 (8%) 21 (21%) 36 (36%)
55 ± 13 48 (48%) 24 (24%) 10 (10%) 26 (26%) 26 (26%)
0.92 1.0 0.63 0.62 0.43 0.06
66 (66%) 90 (90%) 36 (36%) 27 (27%) 17 (17%) 12 (12%) 8 (8%) 6 (6%) 4 (4%) 4 (4%) 10 (10%) 39 (39%) 30 (30%) 11 (11%) 10 (10%) 64 ± 36 238 ± 160 90 ± 23 75 ± 26 42 (42%) 22 (22%) 17 (17%) 31 (31)% 21 (21%) 18 (18%) 4 (4%) 3 (3%) 5.0 ± 0.52 3.2 ± 0.40 0.92 ± 0.19 0.91 ± 0.17 0.36 ± 0.08 38 ± 10 101 ± 20 63 ± 6 1.0 ± 0.00 2 (2%) 25 ± 8 76 ± 19 70 ± 18 1.0 (0.88, 1.3) 199 ± 40 7.1 ± 2.1
5.0 ± 0.53 3.2 ± 0.50 0.96 ± 0.18 0.92 ± 0.19 0.38 ± 0.08 46 ± 15 105 ± 31 57 ± 5 1.0 ± 0.03 5 (5%) 24 ± 7 76 ± 17 70 ± 16 1.1 (0.90, 1.3) 199 ± 41 6.6 ± 2.1
0.36 0.51 0.08 0.47 0.30
Subclinical left ventricular dysfunction by echocardiographic speckle-tracking strain analysis relates to outcome in sarcoidosis Emer Joyce1,2, Maarten K. Ninaber3, Spyridon Katsanos1, Philippe Debonnaire1,4, Vasilis Kamperidis1, Jeroen J. Bax1, Christian Taube3, Victoria Delgado1, and Nina Ajmone Marsan1* 1 Department
of Cardiology, Leiden University Medical Center, Leiden, The Netherlands; 2 Department of Advanced Heart Disease, Brigham and Womens Hospital, Boston, MA, USA; 3 Department of Pulmonology, Leiden University Medical Center, Leiden, The Netherlands; and 4 Department of Cardiology, Sint-Jan Hospital Bruges, Bruges, Belgium Received 10 August 2014; revised 18 September 2014; accepted 24 October 2014
Aims
Limited data exist on the risk of developing cardiac sarcoidosis (CS) and/or adverse events in sarcoidosis patients. Using LV global longitudinal strain (GLS), an emerging sensitive parameter of LV function, we evaluated the prevalence of subclinical cardiac dysfunction in sarcoidosis and investigated whether LVGLS predicts adverse outcomes in this population. ..................................................................................................................................................................... Methods A total of 130 patients with proven sarcoidosis undergoing echocardiography at our referral centre were identified. and results Following exclusion of those with evidence of CS (n = 14) or other pre-existing structural heart disease (n = 16), 100 patients (55 ± 13 years, 48% male, 90% pulmonary involvement) and 100 age- and gender-matched controls were included. LVGLS was measured by speckle-tracking analysis. The primary endpoint was a composite of all-cause mortality, heart failure hospitalization, device implantation, new arrhythmias, or future development of CS on advanced cardiac imaging modalities. LVGLS was significantly impaired in sarcoidosis patients compared with controls (–17.3 ± 2.5 vs. –20.0 ± 1.6%, P < 0.001). Overall, 27 patients (27%) reached the endpoint during a median follow-up of 35 months. On Cox proportional hazards model analysis, abnormal 24-h Holter, larger LV end-diastolic diameters, and more impaired LVGLS were significantly associated with the endpoint; however, only LVGLS remained independently associated on multivariate analysis [hazard ratio (HR) 1.4, 95% confidence interval (CI) 1.1–1.7, P = 0.006]. Patients with LVGLS less than –17.3% were significantly more likely to be free of the primary endpoint (log-rank P = 0.01). ..................................................................................................................................................................... Conclusion LVGLS is impaired in sarcoidosis patients, suggesting subclinical cardiac dysfunction despite the absence of conventional evidence of cardiac disease, and is independently associated with occurrence of cardiac events and/or development of CS.
.......................................................................................................... Cardiac sarcoidosis • cardiomyopathy
Speckle-tracking analysis •
Introduction Sarcoidosis is a multisystem granulomatous disease of unknown aetiology primarily involving the pulmonary system and lymph nodes but which can also affect the heart. The reported prevalence
.............
Keywords
Longitudinal strain •
Prognosis •
Sarcoid
of cardiac sarcoidosis (CS) varies from 20% up to 60% according to the study population (highest prevalence in Japanese patients) and to the type of study performed (clinical or autopsy report).1 – 3 Antemortem diagnosis of CS is challenging, mainly due to the focal extension of the granulomatous involvement, with only 50% of
*Corresponding author. Department of Cardiology, Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands. Tel: +31 71 526 2020, Fax: +31 71 526 6809, Email: [email protected]
© 2014 The Authors European Journal of Heart Failure © 2014 European Society of Cardiology
2
METHODS Patient population One-hundred and thirty consecutive patients with sarcoidosis attending our tertiary referral centre were identified from the departmental Cardiology Information System (EPD-Vision®, Leiden University Medical Center). The diagnosis of sarcoidosis required a compatible clinical picture with or without histological evidence of non-caseating granulomas and the absence of another disease process capable of producing a similar picture.13 Patients were required to have undergone 2DTTE around the time of, or following, their diagnosis. Patients were excluded if they had known structural heart disease (n = 16) or
..................................................................................................................................................................................
patients diagnosed post-mortem having clinical evidence of myocardial disease and only up to 5% of patients reporting symptoms.3 Cardiac involvement in sarcoidosis is also well known to portend a poor prognosis,3,4 and sudden cardiac death (due to either complete heart block or ventricular arrhythmias) may be its first manifestation, prior to clinical presentation or detection of structural abnormalities by traditional parameters or basic cardiac testing. However, if CS is identified at an early stage, corticosteroid therapy and/or cardiac device therapy may reduce the risk of sudden death and/or progression of LV dysfunction.4,5 Therefore, a highly sensitive and non-invasive method for an earlier diagnosis of CS has the potential for significant clinical value in identifying patients at risk for adverse outcomes. Standard two-dimensional transthoracic echocardiography (2DTTE) is recommended in many algorithms as a first-line screening tool, together with 12-lead ECG and 24-h Holter monitoring, for detection of CS in sarcoidosis patients without manifest cardiac disease.6 However, 2DTTE lacks sensitivity to detect early, mild, or focal cardiac involvement.6,7 Advanced imaging modalities, including 99m Technetium (99m Tc)–tetrofosmin single photon emission computed tomography (SPECT), [18 F]fluorodeoxyglucose positron emission tomography (18 F-FDG PET), or delayed contrast enhancement cardiac magnetic resonance imaging (DE-CMR) have demonstrated increased sensitivity and may facilitate earlier detection of CS.6,8,9 Consequently, abnormalities consistent with CS on these modalities have been invoked as diagnostic confirmation of CS in several recent studies2,10 alongside conventional diagnostic criteria proposed by the Japanese Ministry of Health and Welfare (JMHW).11 However, these modalities are expensive, not universally available, and may not be suitable for all patients. More recently, speckle-tracking analysis, a novel echocardiographic technique which assesses the intrinsic deformation (strain) of the myocardium, has been proposed as a sensitive tool to detect myocardial dysfunction before a decrease in LVEF occurs.12 The primary aim of this study was therefore to evaluate the prevalence of subclinical LV dysfunction in sarcoidosis patients without clinical or standard diagnostic evidence of cardiac involvement using LV global longitudinal strain (LVGLS) compared with a control population. Secondly, given the lack of outcome studies to determine the risk of cardiac events in sarcoidosis patients, including the risk of future development of CS, we sought to investigate whether impaired LVGLS may be associated with greater risk of adverse outcomes.
E. Joyce et al.
definite CS (n = 5) at the time of the first 2DTTE study (date of study enrolment). Regarding those excluded for known structural heart disease, the majority (n = 10) had significant ischaemic heart disease, as defined by >50% stenosis in ≥1 epicardial vessel and/or prior acute coronary syndrome and/or prior revascularization procedure, three due to congenital heart disease, two due to known dilated cardiomyopathy from another cause, and one due to significant valve disease, defined as grade ≥2 regurgitation or moderate or more stenosis. The presence of CS was defined by fulfilment of the histological or clinical diagnosis group criteria of the revised JMHW guidelines.11 Those who fulfilled one major criterion of the updated guidelines at study enrolment (n = 9) were also excluded, as this was considered to indicate high likelihood of future CS diagnosis. Consecutive controls were identified from the departmental echocardiographic database (EchoPac) using a specific echocardiographic search code identifying the absence of any structural cardiac abnormalities. Initial referral for echocardiography included chest pain, dyspnoea, syncope, palpitations, murmur on auscultation, or cardiovascular risk stratification, in the absence of any prior history of structural heart disease. The absence of structural cardiac disease was defined as a ‘normal’ echocardiogram according to American Society of Echocardiography guidelines.14 – 17 An age- and gender-matched comparator group was selected in a 1:1 ratio according to published comparability principles, in that they were representative of the same base experience as cases, but without any prior or current history of sarcoidosis.18 The Institutional Review Board of the Leiden University Medical Center approved the current retrospective evaluation of clinically acquired data, and waived the need for patient written informed consent. The study conforms with the principles outlined in the Declaration of Helsinki.
Study protocol Demographic and clinical data were recorded together with the results of chest radiography, pulmonary function testing, and serum ACE and lysozyme levels. Results of standard cardiac screening testing as currently recommended for sarcoidosis, including 2DTTE, 12-lead ECG, and ambulatory 24-h Holter monitoring, were also noted.6 LVGLS was prospectively measured by 2D speckle-tracking analysis on baseline 2DTTE in order to detect subclinical LV dysfunction. Advanced cardiac imaging, such as DE-CMR, myocardial SPECT, 67 gallium (67 Ga) scintigraphy, and 18 F-FDG PET,6,8 was carried out during follow-up at the discretion of the treating physician. The presence of an abnormality consistent with CS on ≥1 advanced imaging modality was recorded.6,11 During the follow-up period after the first 2DTTE evaluation, survival of all patients was checked by querying the survival statistics of the municipal civil registries of The Netherlands, and the occurrence of other clinical events up to the time of the most recent hospitalization or outpatient clinic visit was recorded. The primary endpoint was a composite of all-cause mortality, heart failure-related hospitalization, cardiac device implantation—permanent pacemaker, implantable cardioverter defibrillator (ICD), and/or biventricular pacemaker—new arrhythmias (sustained ventricular tachycardia not associated with ICD implantation and/or new atrial arrhythmia), or accepted evidence of development of CS on advanced cardiac imaging study.
Clinical testing Pulmonary function tests were recorded as the percentage of predicted values measured according to the American Thoracic © 2014 The Authors European Journal of Heart Failure © 2014 European Society of Cardiology
3
Table 1 Abnormalities on basic clinical cardiac testing considered possible evidence of cardiac sarcoidosis 2,7,17,20
Test Abnormalities ................................................................ 12-Lead ECG
24-h Holter monitoring
Two-dimensional echocardiography
Minor criteria abnormalitiesa Non-minor criteria abnormalities First degree or Mobitz type 1 atrioventricular blockb, c ST–T abnormalities (T-wave inversion ≥2 leads considered significant) Left ventricular hypertrophyd Right ventricular hypertrophy Supraventricular tachyarrhythmia (including atrial fibrillation or flutter) Sinus tachycardiae Prolonged corrected QT interval Mobitz type 1 atrioventricular blockb, c Significant premature ventricular contractions (>1% per 24 h) Non-sustained ventricular tachycardia >3 beatsc Sustained ventricular tachycardiac Supraventricular tachycardia >3 beatsc Regional wall motion abnormalities in ≥2 segments Right ventricular systolic dysfunction in the absence of pulmonary hypertensionf Diastolic dysfunction inappropriate for age
Adapted from Mehta et al.2 a Ventricular arrhythmias (ventricular tachycardia, multifocal or frequent premature ventricular contractions), complete right bundle branch block, axis deviation, or abnormal Q wave).11 b Higher degrees of atrioventricular block were an exclusion criterion at baseline (major criterion). c For Cox proportional hazards model, these abnormalities were only considered in patients who did not have a ‘similar’ arrhythmic or device event included in the primary endpoint at follow-up. d Only considered positive in the absence of hypertension. e In the absence of another obvious cause. f Right ventricular systolic dysfunction was defined by tricuspid annular plane systolic excursion 40 fps to ensure reliable analysis by the software. The LV endocardial border was traced at end-systole in all three apical views and the automatically created region of interest was manually adjusted to the thickness of the myocardium. Tracking quality was assessed in all segments and, if tracking remained of suboptimal quality following manual correction, segments were discarded. If >2 segments needed to be discarded, these patients were excluded from the subsequent analyses using LVGLS. For the remaining patients, LVGLS was then provided by the software as the average value of the peak systolic longitudinal strain of the three apical views, using a 17-segment model, in a bull’s eye plot (Figure 1).
Other cardiac imaging modalities Cardiac magnetic resonance (using a 1.5-T Gyroscan ACS-NT/Intera scanner; Philips Medical Systems, Eindhoven, The Netherlands), myocardial SPECT (using 99m Tc–tetrofosmin, 500 MBq injected at rest), 18 F-FDG imaging (using either PET or dual-isotope SPECT), and 67 Ga scintigraphy were all performed according to standard protocols. On CMR, the presence of delayed enhancement was considered evidence of CS (scarring).6,11 A perfusion defect or defects on myocardial SPECT imaging, defined as regions containing 8 years. Speckle-tracking analysis revealed an LVGLS of –18.3%. (B) A patient with biopsy-confirmed extracardiac sarcoidosis who presented >1 year after baseline 2D echocardiography evaluation with complete heart block requiring permanent pacemaker implantation (primary endpoint). LVGLS was significantly reduced (less negative) at –15.0%. AVC, aortic valve closure; APLAX, apical long axis; 2C, two-chamber; 4C, four-chamber.
© 2014 The Authors European Journal of Heart Failure © 2014 European Society of Cardiology
5
testing parameters, and LVGLS to the primary endpoint. The number of covariables entered into the multivariate model was limited due to the number of endpoint events; however, all variables with a P-value 1 abnormality on basic cardiac testing was significantly more common in the primary endpoint group (73% vs. 48%, P = 0.03). Baseline LVGLS was significantly more impaired (–15.8 ± 2.2 vs. –17.9 ± 2.3%, P < 0.001) in sarcoidosis patients who experienced the primary endpoint compared with those who did not (example in Figure 1).
Univariate and multivariate predictors of the primary endpoint As shown in Table 4, on univariate Cox proportional hazards model analysis, abnormal 24-h Holter ECG, higher LV end-diastolic diameter, and more impaired (more positive) LVGLS were significantly associated with the primary endpoint. Only LVGLS remained significantly and independently associated with the primary endpoint on multivariate analysis [hazard ratio (HR) 1.4, 95% confidence interval (CI) 1.1–1.7, P = 0.006]. The area under the curve for LVGLS to predict the primary endpoint was 0.73 (95% CI 0.62–0.85, P = 0.001) (Figure 2). A cut-off value of –17.3% (also corresponding to the median value) had a 72% sensitivity and 60% specificity to predict the occurrence of the primary endpoint. Those with an LVGLS less than –17.3% were significantly more likely to be free of the primary endpoint as
6
E. Joyce et al.
Table 2 Baseline demographic, clinical, and echocardiographic characteristics of the total patient population and the control population Variable Controls (n = 100) All patients (n = 100) P-value ........................................................................................................................................... Demographic Age, years Male gender, n (%) Hypertension, n (%) Diabetes, n (%) Hyperlipidaemia, n (%) Smoking (ever), n (%) Clinical (sarcoidosis patients only) Histological + clinical diagnosis vs. clinical diagnosis alone, n (%) Organ involvement, n (%) Pulmonary Skin Joint Eye Renal ± hypercalcaemia Neurological system Liver ± gallbladder Parotid gland Other Chest radiograph stage 0 1 2 3 4 Serum ACE, nmol/min/mL Lysozyme, % FVC predicted, % DLCO predicted, % Abnormal 12-lead ECG (overall), n (%) Abnormal 12-lead ECG (JMHW criteria), n (%) Abnormal 24-h Holter monitoring, n (%) NYHA class ≥1, n (%) Immunomodulatory medications at baseline, n (%) Steroids Azathioprine Methotrexate Echocardiographic LVEDD, cm LVESD, cm IVSd, cm PWd, cm Relative wall thickness, cm LVESV, mL LVEDV, mL LVEF, % WMSI MR grade ≥2 LAVI, mL/m2 E, cm/s A, cm/s E/A DT, ms Septal E’, m/s
55 ± 13 48 (48%) 27 (27%) 8 (8%) 21 (21%) 36 (36%)
55 ± 13 48 (48%) 24 (24%) 10 (10%) 26 (26%) 26 (26%)
0.92 1.0 0.63 0.62 0.43 0.06
66 (66%) 90 (90%) 36 (36%) 27 (27%) 17 (17%) 12 (12%) 8 (8%) 6 (6%) 4 (4%) 4 (4%) 10 (10%) 39 (39%) 30 (30%) 11 (11%) 10 (10%) 64 ± 36 238 ± 160 90 ± 23 75 ± 26 42 (42%) 22 (22%) 17 (17%) 31 (31)% 21 (21%) 18 (18%) 4 (4%) 3 (3%) 5.0 ± 0.52 3.2 ± 0.40 0.92 ± 0.19 0.91 ± 0.17 0.36 ± 0.08 38 ± 10 101 ± 20 63 ± 6 1.0 ± 0.00 2 (2%) 25 ± 8 76 ± 19 70 ± 18 1.0 (0.88, 1.3) 199 ± 40 7.1 ± 2.1
5.0 ± 0.53 3.2 ± 0.50 0.96 ± 0.18 0.92 ± 0.19 0.38 ± 0.08 46 ± 15 105 ± 31 57 ± 5 1.0 ± 0.03 5 (5%) 24 ± 7 76 ± 17 70 ± 16 1.1 (0.90, 1.3) 199 ± 41 6.6 ± 2.1
0.36 0.51 0.08 0.47 0.30
