ORIGINAL ARTICLE

Extensive late gadolinium enhancementon cardiovascular magnetic resonance predictsadverse outcomes and lack of improvement in LVfunction after steroid therapy in cardiac sarcoidosisTakayuki Ise,1 Takuya Hasegawa,1 Yoshiaki Morita,2 Naoaki Yamada,2 Akira Funada,1

Hiroyuki Takahama,1 Makoto Amaki,1 Hideaki Kanzaki,1 Hideo Okamura,1

Shiro Kamakura,1 Wataru Shimizu,1 Toshihisa Anzai,1 Masafumi Kitakaze1,3

1Department of CardiovascularMedicine, National Cerebraland Cardiovascular Center,Osaka, Japan2Department of Radiology,National Cerebral andCardiovascular Center, Osaka,Japan3Department of ClinicalResearch and Development,National Cerebral andCardiovascular Center, Osaka,Japan

Correspondence toDr Takuya Hasegawa,Department of CardiovascularMedicine National Cerebraland Cardiovascular Center,Osaka, Japan 5-7-1 Fujishiro-dai, Suita, Osaka 565-8565,Japan; hasegawa@hsp.ncvc.go.jp

Received 31 October 2013Revised 17 April 2014Accepted 22 April 2014

To cite: Ise T, Hasegawa T,Morita Y, et al. HeartPublished Online First:[please include Day MonthYear] doi:10.1136/heartjnl-2013-305187

ABSTRACTBackground Gadolinium-enhanced cardiovascularmagnetic resonance is an emerging tool for thediagnosis of cardiac sarcoidosis (CS); however, thecorrelations between extent of late gadoliniumenhancement (LGE) and efficacy of steroid therapy andadverse outcomes in patients with CS remain unclear.Objective We aimed to clarify the prognostic impact ofextent of LGE in patients with CS.Methods Before the start of steroid therapy,43 consecutive LGE-positive patients with CS weredivided into two groups based on the extent of LGE by amedian value: small-extent LGE (LGE mass <20% of LVmass; n=21) and large-extent LGE (LGE mass ≥20% ofLV mass; n=22). We examined the correlations betweenextent of LGE and outcomes after steroid therapy.Results Among the 6 patients who died from heartdisorders, 11 patients who were hospitalised because ofheart failure and 6 patients who suffered life-threateningarrhythmia during the follow-up period, large-extent LGEpredicted higher incidences of cardiac mortality andhospitalisation for heart failure. Multivariate Coxregression analysis showed that large-extent LGE wasindependently associated with combined adverseoutcomes including cardiac death, hospitalisation forheart failure, and life-threatening arrhythmias. In thesmall-extent LGE group, LV end-diastolic volume indexsignificantly decreased and LVEF significantly increasedafter steroid therapy, whereas in the large-extent LGEgroup, neither LV volume nor LVEF changedsubstantially.Conclusions Large-extent LGE correlates with absenceof LV functional improvement and high incidence ofadverse outcomes in patients with CS after steroidtherapy.

INTRODUCTIONSarcoidosis is a multisystem granulomatous disorderof unknown cause with symptomatic myocardialinvolvement in up to 7% of affected patients.1–3

Although it is generally associated with a low mor-tality rate, concomitant cardiac involvement worsenits prognosis.4 5 Therefore, detection of myocardialinvolvement is critical for management of patientswith sarcoidosis. In patients with sarcoidosis,gadolinium-enhanced cardiovascular magnetic

resonance (CMR) is a useful diagnostic tool toqualitatively detect myocardial involvement.6–8 Inmost patients with cardiac sarcoidosis (CS), lategadolinium enhancement (LGE) is typically loca-lised in the basal and lateral segments of the LVwall or epicardium, which does not fit any specificcoronary perfusion area.9 LGE in CMR has beenreported to reflect myocardial fibrosis and granu-lomatous inflammation in patients with CS.7 It hasalso been reported that the presence of myocardialLGE can predict adverse events in patients with sys-temic sarcoidosis.7 10 However, the prognosticimpact of the extent of LGE has not been fullyinvestigated. In this study, we examined the correla-tions between the extent of LGE and adverse out-comes, as well as the efficacy of steroid therapy inpatients with CS.

METHODSStudy patientsMedical records were screened to identify patientsdiagnosed with CS in our institution from May2000 to May 2012. CS was diagnosed according tothe guidelines of the Specific Diffuse PulmonaryDisease Research Group, Sarcoidosis Division( Japanese Ministry of Health and Welfare).11 Inbrief, CS was diagnosed on the basis of histologicalfindings or clinical findings. Histological diagnosisof CS was confirmed when histological analysis ofendomyocardial biopsy specimens demonstratedepithelioid granuloma without caseating granu-lomas. Clinical diagnosis of CS was confirmed bythe presence of an electrocardiographic (ECG)abnormality suggesting myocardial injury, and atleast one of the following items: abnormal wallmotion, regional wall thinning, or dilatation of theLV; perfusion defect on thallium-201 myocardialscintigraphy or abnormal accumulation bygallium-67-citrate scintigraphy ortechnetium-99m-pyrophosphatemyocardial scintig-raphy; abnormal intracardiac pressure, low cardiacoutput, or depressed LVEF; and interstitial fibrosisor cellular infiltration over moderate grade even ifthe findings were non-specific. All patients under-went coronary angiography, and no significant cor-onary artery stenosis was noted. All baselinecharacteristics, including CMR data, were collected

Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187 1

Heart failure and cardiomyopathies Heart Online First, published on May 14, 2014 as 10.1136/heartjnl-2013-305187

Copyright Article author (or their employer) 2014. Produced by BMJ Publishing Group Ltd (& BCS) under licence.

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

within 1 month before steroid therapy initiation. Estimatedglomerular filtration rate was calculated using the Modificationof Diet in Renal Disease Study equation,12 with coefficientsmodified for Japanese patients,13 as follows: estimated glomeru-lar filtration rate (mL/min/1.73 m2) = 194×serum creatin-ine−1.094 × age−0.287×(0.739 if female).

As per the standard protocol,14 CS patients administered astarting dose of 60 mg prednisolone on alternate days for1 month, and this dose was tapered gradually to the final main-tenance dose of 10 mg on alternate days. All study patientsreceived the final maintenance dose of prednisolone after6 months. No patients enrolled in this study were receivingimmunosuppressant therapy. This study was approved by theethics committee of National Cerebral and CardiovascularCenter, and patients gave informed consent.

EchocardiographyAll patients underwent echocardiographic examinations withcommercially available ultrasonography systems before and6 months after initiation of steroid therapy. LV volumes andLVEF were measured by the modified Simpson’s method,according to the guidelines of the American Society ofEchocardiography.15

CMR imagingCMR imaging was performed using a 1.5-T MR system(Magnetom Sonata, Siemens, Erlangen, Germany) with a stan-dardised clinical protocol. All CMR images were electrocardio-graphically gated and obtained during repeated breath-holds.Cine images were acquired with a steady-state free precession(SSFP) with the following parameters: repetition time, 3.2 ms;echo time, 1.6 ms; flip angle, 55°; matrix, 190×190; field ofview, 340 mm; section thickness, 6 mm; section interval,10 mm; sensitivity encoding factor, 2. After localisation of theheart, cine images of 9–12 contiguous short-axis sectionsencompassing the entire LV and 2-, 3-, and 4-chamber long-axissections were collected. Then, gadopentetate meglumine(0.15 mmol/kg; Magnevist; Schering AG, Berlin, Germany) wasadministered at a rate of 3–4 mL/s using a power injector. LGEimages were acquired 10 min after the injection of gadopente-tate meglumine, with an inversion-recovery SSFP pulse sequencewith inversion time of 300 ms.16 17 The parameters used inSSFP for LGE were repetition time, 3.5 ms; echo time, 1.7 ms;flip angle, 60°; matrix, 256×129; field of view, 340 mm;section thickness, 8 mm; section interval, 10 mm; sensitivityencoding factor, 1. Among the 9–12 short-axis slices, weexcluded both ends of the apex and the base because the scansof these sections did not include the LV muscle or the bevelledmyocardium, which caused incorrect signal intensities. Then,seven adjacent slices in the middle of the remaining slices wereobtained by using localiser of LV long-axis.17

CMR data analysesCine images were analysed using ARGUS (Siemens, Germany)to calculate LV volumes, mass and function. LGE images wereanalysed using Ziostation 2 (Ziosoft, Tokyo, Japan). Regions ofLGE in seven slices of short-axis LGE imaging were automatic-ally defined as those exhibiting signal intensity above a predeter-mined threshold. We used a threshold of 5 SDs above the signalintensity of non-damaged myocardium, because LGE quantifica-tion with the threshold of 5SD demonstrated the best agreementwith visual assessment and best reproducibility among differenttechniques with different thresholds, as previously reported.18 19

The LGE mass was calculated using the LGE area obtained from

the seven LGE imaging slices. The extent of LGE was expressedas a percentage of LV mass according to the following equation:

LGEmass (%) ¼ {LGEmass (g)}=fLVmass ðgÞg � 100:

The patients were divided into two groups using the medianvalue for the extent of LGE: small-extent LGE group (LGE<20%) and large-extent LGE group (LGE ≥20%). Themethods used for assessing LGE and representative images forboth groups are shown in figure 1.

Clinical follow-upThe primary composite outcome was defined as cardiac death,hospitalisation for heart failure, and life-threatening arrhythmia.Life-threatening arrhythmia was defined as documented orappropriate implantable cardioverter defibrillator treatment fortermination of ventricular fibrillation or sustained ventriculartachycardia and successful cardiopulmonary resuscitation forcardiac arrest. Follow-up information was obtained by retro-spective chart review. The composite end-point included onlythe first event for each patient. If a patient was admitted to ourhospital due to heart failure and then died of heart failure, itcounted as one event. Additionally, we did not count hospitaladmissions due to non-cardiac causes as events of hospitaladmission. No patients were lost to follow-up.

Statistical analysesAll data are expressed as means±SD. Statistical analyses wereperformed using JMP10 (SAS Institute, Cary, North Carolina,USA). A p value less than 0.05 was considered statistically sig-nificant. Continuous variables were compared using paired orunpaired Student t test, as appropriate. Categorical variables ofthe two groups were compared using the χ2 test. Long-term sur-vival was estimated by Kaplan–Meier analysis, and differences insurvival were assessed using the log-rank test. Univariate andmultivariate Cox proportional hazards regression models wereconstructed to investigate the predictors of baseline data forcombined adverse outcomes. Echocardiographic measurementswere used for analysis of pretherapy and post-therapy LVvolumes and LVEF. Pearson’s correlation coefficient analysis wasused to assess the correlation between extent of LGE and LVEFchanges. Multivariate linear regression analysis was also per-formed with adjustment for LVEF. Receiver-operating character-istic (ROC) curve was used to examine the performancecharacteristic of %LGE mass. Area under the curve (AUC), and95% confidence of ROC curve, were calculated to provide ameasure of the accuracy of %LGE mass to predict combinedadverse outcomes. Variables with a p value <0.05 in the uni-variate models were included in the multivariate analysis.

RESULTSOverall, 71 patients were diagnosed with CS. Of these, 21 wereexcluded because they did not undergo CMR. Among the 50CS patients who underwent CMR, seven patients, including twopatients without LGE, were excluded because they did notreceive steroid therapy. Eventually, 43 patients met the inclusioncriteria for this study (15 men and 28 women; mean age, 59±10 years; age range, 29–73 years). The mean follow-up dur-ation was 39±19 months (range, 8–73 months). Twenty-twopatients were assigned to the large-extent LGE group, and 21patients were assigned to the small-extent LGE group, using themedian value for the extent of LGE.

2 Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

Baseline characteristicsPatient baseline characteristics are summarised in table 1.Demographic factors, New York Heart Association (NYHA)functional class, and organ involvement did not differ betweenthe two groups. The B-type natriuretic peptide (BNP) level inthe small-extent LGE group was significantly lower than that inthe large-extent LGE group. No significant differences in sys-temic inflammatory markers, indicating the activity of systemicsarcoidosis, were observed between the two groups. Before initi-ation of steroid therapy, there was also no difference betweenthe two groups in type of medication or implanted device,including permanent pacemaker, implantable cardioverter-defibrillator, and cardiac resynchronisation therapy.

Echocardiographic data and CMR analysisEchocardiographic and CMR data before steroid therapy arealso shown in table 1. CMR was performed 13±7 days beforethe initiation of steroid therapy. Lower LVEF and higher LV end-diastolic and end-systolic volume indices were observed in thelarge-extent LGE group before steroid therapy. Although thedifference between the two groups was not statistically signifi-cant, there was a trend toward larger LV mass in the large-extentLGE group.

Extent of LGE as a predictor of adverse eventsDuring the follow-up period, six patients died of heart disor-ders, including five patients with heart failure and one patient

with refractory ventricular arrhythmia. There were no non-cardiac deaths in this study. Additionally, 11 patients were hospi-talised for heart failure, and six suffered life-threateningarrhythmia. Among these six life-threatening arrhythmias, fivepatients suffered sustained ventricular tachycardia, and onepatient suffered ventricular fibrillation. Four of the five ventricu-lar tachycardia cases were identified by the implantable cardio-verter defibrillator electrogram; three cases were terminated byappropriate shock, and one case by antitachycardia pacing.Furthermore, one ventricular tachycardia case, with symptomsof palpitation and decreased blood pressure, was identified byECG and terminated by cardioversion. Implantable cardioverterdefibrillator was implanted in him after the event.

There were no cardiac deaths in the small-extent LGE groupduring the follow-up period. The survival rate in thelarge-extent LGE group was lower than that in the small-extentLGE group: 95% after 1 year, 77% after 3 years, and 72% after5 years. A log-rank test revealed a significant difference in com-bined adverse outcomes (log-rank: χ2=8.10, p=0.004), cardiacmortality (log-rank: χ2=6.36, p=0.012), and hospitalisation forheart failure (log-rank: χ2=8.60, p=0.003) (figure 2) betweenthe small-extent and large-extent LGE groups. On the otherhand, extent of LGE did not appear to be associated with futureoccurrences of life-threatening arrhythmias (log-rank: χ2=0.87,p=0.352). The univariate Cox proportional hazards modelshowed that the extent of LGE expressed as %LGE mass,NYHA functional class, BNP and LVEF were associated with

Figure 1 Quantification of thevolume of late gadoliniumenhancement (LGE) in CMR. (A, B) arerepresentative images from patientswith small-extent LGE and large-extentLGE, respectively. Left panels showoriginal LGE images, and right panelsindicate the automatically colouredLGE area according to the followingpredetermined setting: 5 SDs abovethe mean signal intensity of remotenon-diseased myocardium.

Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187 3

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

combined adverse outcomes (table 2). Additionally, multivariateanalysis revealed that extent of LGE was an independent pre-dictor of combined adverse outcomes (adjusted HR=1.10, 95%CI 1.01 to 1.21, p=0.021). In the present patient population,ROC curve analysis indicated that %LGE mass had the modestability to predict combined adverse outcomes (AUC=0.77, 95%CI 0.60 to 0.89). A cutoff level of %LGE ≥21.9% best pre-dicted combined adverse outcomes, with a sensitivity of 81.3%

and a specificity of 70.3%. The test’s positive and negative pre-dictive values were 61.9% and 86.3%, respectively.

Association of extent of LGE with improvement in LVfunction after steroid therapyChanges in LV end-diastolic volume index and LVEF from base-line to 6 months after steroid therapy are presented in figure 3.In the small-extent LGE group, LV end-diastolic volume index

Table 1 Patient characteristics

Overall (n=43) Small LGE (n=21) Large LGE (n=22)p ValueSmall vs large

Age (years) 59±10 59±12 60±9 0.798Male, n (%) 15 (35) 6 (29) 9 (41) 0.396Body Mass Index (kg/m2) 22.3±3.1 22.7±3.0 21.9±3.3 0.417Systolic blood pressure (mm Hg) 123±18 126±17 120±18 0.250Diastolic blood pressure (mm Hg) 72±12 71±11 73±13 0.642NYHA functional class 0.875

I, n (%) 7 (16) 4 (19) 3 (14)II, n (%) 27 (63) 13 (62) 14 (64)III, n (%) 9 (21) 4 (19) 5 (23)IV, n (%) 0 (0) 0 (0) 0 (0)

Organ involvementLung, n (%) 9 (21) 4 (19) 5 (23) 0.767Eye, n (%) 3 (7) 2 (10) 1 (5) 0.522Skin, n (%) 2 (5) 1 (5) 1 (5) 0.973

Laboratory results at the beginning of steroid therapyWhite blood cell count (/μL) 5701±1587 5428±1288 5961±1820 0.276Haemoglobin (g/dL) 13.4±1.5 13.3±1.6 13.4±1.3 0.709Serum creatinine (mg/dL) 0.8±0.2 0.8±0.2 0.8±0.2 0.845eGFR (mL/min/1.73m2) 55±15 54±14 56±16 0.640Albumin (g/dL) 4.1±0.3 4.0±0.3 4.1±0.3 0.573Calcium (mg/dL) 9.5±0.4 9.4±0.2 9.6±0.6 0.311BNP (pg/mL) 293±233 216±174 368±261 0.031ACE (U/L) 14.7±8.4 15.0±9.2 14.4±7.6 0.843Lysozyme (μg/mL) 8.0±4.1 7.2±3.2 8.9±5.0 0.222

Medications at the beginning of steroid therapyβ blockers, n (%) 18 (42) 8 (38) 10 (45) 0.625ACE inhibitors, n (%) 17 (40) 7 (33) 10 (45) 0.416ARB, n (%) 9 (21) 4 (19) 5 (23) 0.767Diuretics, n (%) 23 (53) 9 (43) 14 (64) 0.172Antiarrhythmic agents, n (%) 9 (21) 5 (24) 4 (18) 0.650

Implanted devices at the beginning of steroid therapyPacemaker, n (%) 7 (16) 5 (24) 2 (9) 0.191Implantable cardioverter defibrillator, n (%) 7 (16) 3 (14) 4 (18) 0.729Cardiac resynchronisation therapy, n (%) 2 (5) 0 (0) 2 (9) 0.157

Echocardiographic analysisLVEF (%) 41±10 45±11 36±6 0.001LV diastolic dimension (mm) 57±8 52±6 61±8 <0.001LV systolic dimension (mm) 44±9 39±8 49±7 <0.001LV end-diastolic volume index (mL/m2) 111±38 92±27 130±38 <0.001LV end-systolic volume index (mL/m2) 68±31 52±22 84±30 <0.001

CMR analysisLVEF (%) 38±12 43±12 32±10 0.002LV end-diastolic volume index (mL/m2) 112±42 92±27 132±45 0.001LV end-systolic volume index (mL/m2) 73±38 54±24 91±41 <0.001LV mass (g) 130±38 122±47 138±27 0.195LV mass index (g/m2) 82±23 78±28 87±17 0.219LGE mass (g) 26±16 14±11 38±11 <0.001% LGE mass (%) 19±10 11±5 28±5 <0.001

ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; BNP, B-type natriuretic peptide; CMR, cardiovascular magnetic resonance; eGFR, estimated glomerularfiltration rate; LGE, late gadolinium enhancement; NYHA, New York Heart Association.

4 Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

significantly decreased (92±27 and 85±20 mL/m2 before andafter steroid therapy, respectively; p=0.005) and LVEF signifi-cantly increased (45±11 and 50±10% before and after steroidtherapy, respectively; p<0.001) at 6 months after steroidtherapy. However, in the large-extent LGE group, neither LVvolumes nor LVEF changed substantially (LV end-diastolicvolume index: 130±38 and 131±40 mL/m2 before and aftersteroid therapy, respectively; p=0.731, LVEF: 36±6 and 35±8% before and after steroid therapy, respectively; p=0.213).Furthermore, regression analysis revealed a significant negativecorrelation between extent of LGE and changes in LVEF(r2=0.38, p<0.001) (figure 4A). We observed an associationbetween extent of LGE and changes in LVEF even after adjust-ment for LVEF before steroid therapy (r2=0.43, p<0.001). Onthe other hand, a significant association was not found betweenchanges in LVEF and pretherapy LVEF (p=0.145) (figure 4B).

DISCUSSIONLGE imaging by CMR has revolutionised the diagnosis of car-diomyopathies. Furthermore, in patients with sarcoidosis, thistechnique is also highly useful for the detection of myocardialinvolvement. Since the extent of LGE can now be quantified bycommercially available software, we attempted to clarify theprognostic impact of quantitative evaluation of LGE in CSpatients in the present study. We obtained evidence that theextent of LGE before steroid therapy was inversely correlatedwith improvement in LVEF at 6 months after steroid therapy,and that a larger extent of LGE predicted a higher incidence ofcardiac events, even after adjustment for LVEF.

Prediction of adverse outcomesIt has been reported that survival of CS patients with preservedLVEF is better than that of patients with low LVEF.14 20

Recently, the prognostic capability of LGE imaging for predict-ing outcomes has been assessed. It has been shown that the pres-ence of LGE predicts adverse outcomes in patients with dilatedcardiomyopathy21 22 and in patients with hypertrophic cardio-myopathy.23–25

In patients with systemic sarcoidosis, Greulich et al and Patelet al reported that the presence of myocardial LGE can predictadverse cardiac outcomes.7 10 However, in their study patients,systemic sarcoidosis was diagnosed on the basis of involvementof organs other than the heart; therefore, many patientswithout myocardial involvement were included. In patientswith CS, granulomatous inflammation and myocardialfibrosis tend to develop during the early stages of the disease.Since LGE reflects myocardial fibrosis and granulomatousinflammation,7 most CS patients may have a certain extent ofLGE when they were diagnosed with CS.6 Indeed, our studydemonstrated that, among the 50 patients who were diagnosedwith CS according to the guideline11 and underwent CMR inour hospital, 96 percent of the patients did have LGE, and thattheir baseline characteristics revealed relatively advanced myo-cardial impairment compared with those reported in previousstudies.7 10 We therefore hypothesised that quantitative evalu-ation of LGE would be useful for the prediction of cardiacevents in CS patients. The results of the present study demon-strated that CS patients with a large extent of LGE had a highincidence of adverse outcomes, including cardiac death andhospitalisation for heart failure.

Figure 2 Kaplan–Meier cumulative event curves for (A) combined adverse outcomes, (B) cardiovascular deaths, (C) hospitalisation for heart failure,and (D) life-threatening arrhythmia. The combined adverse outcomes include cardiovascular deaths, hospitalisation for heart failure, andlife-threatening arrhythmia. The log-rank p value for the difference between the Kaplan–Meier curves is shown.

Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187 5

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

In this study, approximately 25% of the CS patientsrequired hospitalisation for heart failure during the follow-upperiod. Furthermore, 45% of these patients died of refractoryheart failure. The extent of LGE indicates the severity of myo-cardial fibrosis, which is linked to the severity of LV dysfunc-tion.26 27 Therefore, the presence of large-extent LGE couldpredict refractory heart failure in CS patients. On the otherhand, life-threatening arrhythmias occurred in 14% of thesepatients with CS. The frequency of life-threatening arrhyth-mias was higher in the large-extent LGE group, although thedifference was not significant. Even localised lesions in themyocardium could be origins of re-entry arrhythmias, regard-less of the extent of LGE, that is, myocardial fibrosis. In fact,in this study population, two patients with small-extent LGEwho had a localised lesion in the LV septum or apex showedsymptomatic ventricular tachyarrhythmias. This may implythat even small amounts of LGE can precipitate arrhythmias,whereas heart failure tends to occur when LGE is moreextensive.

Prediction of improvement in LV functionAlthough little is known about the overall beneficial or adverseeffects of steroid therapy in patients with CS, it has beenreported that such patients with mildly to moderately reducedLVEF (30–54%) showed improvement in LV function aftersteroid therapy to a greater extent than that in patients with CSwith severely reduced LVEF,19 suggesting that adequate improve-ment in myocardial function cannot be expected in patientswith severe myocardial impairment. In patients with dilated orischaemic cardiomyopathy, a smaller extent of LGE, indicating alow degree of myocardial impairment, has been reported to bean indicator of recovery of LV systolic function.28 29 In thepresent study, extent of LGE predicted improvement in LVEF inpatients with CS after steroid therapy, being consistent with pre-vious reports.

Autopsy findings in patients with CS in a previous studyshowed that LGE represents fibrosis and granulomatous inflam-mation in the myocardium.7 Considering our results showing

Table 2 Univariate and multivariate Cox proportional hazards models for combined adverse outcomes including cardiac death, hospitalisationfor heart failure, and life-threatening arrhythmias

Univariate analysis Multivariate analysisVariable Unadjusted HR (95% CI) p Value Adjusted HR(95%CI) p Value

Age* 1.07 (0.71 to 1.74) 0.765 – –

Male 1.75 (0.63 to 4.87) 0.279 – –

NYHA functional class† 2.45 (1.08 to 5.73) 0.032 2.40 (1.04 to 5.80) 0.040eGFR‡ 1.00 (0.96 to 1.03) 0.852 – –

BNP§ 5.60 (1.19 to 30.9) 0.028 0.71 (0.10 to 5.43) 0.740CMR analysis

LVEF¶ 0.80 (0.62 to 1.00) 0.049 0.89 (0.67 to 1.19) 0.463LV end-diastolic volume index** 1.08 (0.96 to 1.21) 0.203 – –

LV end-systolic volume index** 1.13 (0.99 to 1.26) 0.065 – –

LV mass index†† 1.00 (0.81 to 1.21) 0.979 – –

% LGE mass¶ 1.71 (1.26 to 2.47) 0.001 1.64 (1.10 to 2.66) 0.015

Multivariable model included NYHA functional class, BNP, LVEF and %LGE mass.*HR reflects risk with an increase of 10 years.†With an increase of 1 functional class.‡With an increase of 1 mL/min/1.73 m2.

§With an increase of 1 logBNP.¶With an increase of 5%.**With an increase of 10 mL/m2.

††With an increase of 10 g/m2.BNP, B-type natriuretic peptide; CMR, cardiovascular magnetic resonance; eGFR, estimated glomerular filtration rate; LGE, late gadolinium enhancement. NYHA, New York HeartAssociation.

Figure 3 Alterations in LV end-diastolic dimensions (A) and LVEF (B)after steroid therapy.

6 Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

that patients with large-extent LGE before steroid therapy hadno improvement in LVEF, LGE appears to reflect the degree ofirreversible myocardial damage rather than granulomatousinflammation. In fact, in this study, some patients with severelyreduced LVEF and mild LGE showed a good response to steroidtherapy, suggesting that extent of LGE is a useful indicator ofthe expected improvement in LVEF, independent of LVEFbefore steroid therapy.

Study limitationsSeveral limitations to this study should be acknowledged. Thiswas a single-centre retrospective observational study, and onlypatients who underwent CMR imaging before steroid therapywere enrolled. The results of the present study are limited bythe small sample size and number of events. Some patientswith CS who required a mechanical device, such as a pace-maker or implantable cardioverter-defibrillator for urgent con-duction blocks or ventricular tachyarrhythmia at the initialvisit, could not undergo CMR. There is a possibility thatpatients prone to arrhythmia events have therefore beenexcluded.

CONCLUSIONSWe conclude that the presence of large-extent LGE in patientswith CS is linked to the absence of LV functional improvementand a high incidence of adverse outcomes after steroid therapy.We suggest that CMR imaging is useful for establishing the diag-nosis of CS and also for predicting adverse events and the effi-cacy of steroid therapy.

Key messages

What is known on this subject?Late gadolinium enhancement (LGE) on cardiac MRI is anemerging tool for the diagnosis of cardiac sarcoidosis (CS).

What might this study add?We suggest that cardiac MRI is useful for establishing thediagnosis of CS and also for predicting adverse events and theefficacy of steroid therapy.

How might this impact on clinical practice?The prognostic impact of LGE in patients with CS after steroidtherapy has not been well investigated. We demonstrated thatlarge-extent LGE in patients with CS is linked to the absence ofLV functional improvement and a high incidence of adverseoutcomes after steroid therapy.

Contributors The contribution of each author is as follows; conception and designor analysis or analysis and interpretation of data, TI, TH YM and NY. Drafting of themanuscript or revising it critically for important content, TI, TH, NY, AF, HT, MA, HK,HO, SK and WS. Final approval of the manuscript submitted, TA and MK.

Funding This work was supported by grants-in-aid from the Ministry of Health,Labor, and Welfare-Japan (H23-Nanchi-Ippan-22 to MK), grants-in-aid from theMinistry of Education, Culture, Sports, Science and Technology-Japan (21390251 toMK), grants from the Japan Heart Foundation (MK), and grants from the JapanCardiovascular Research Foundation (MK).

Competing interests None.

Patient consent Obtained.

Ethics approval The ethics committee of National Cerebral and CardiovascularCenter.

Provenance and peer review Not commissioned; externally peer reviewed.

REFERENCES1 James DG, Neville E, Carstairs LS. Bone and joint sarcoidosis. Semin Arthritis

Rheum 1976;6:53–81.2 Johns CJ, Michele TM. The clinical management of sarcoidosis. A 50-year

experience at the Johns Hopkins Hospital. Medicine 1999;78:65–111.3 Shammas RL, Movahed A. Sarcoidosis of the heart. Clin Cardiol 1993;16:462–72.4 Sharma OP, Maheshwari A, Thaker K. Myocardial sarcoidosis. Chest

1993;103:253–8.5 Perry A, Vuitch F. Causes of death in patients with sarcoidosis. A morphologic study

of 38 autopsies with clinicopathologic correlations. Arch Pathol Lab Med1995;119:167–72.

6 Smedema JP, Snoep G, van Kroonenburgh MP, et al. Evaluation of the accuracy ofgadolinium-enhanced cardiovascular magnetic resonance in the diagnosis of cardiacsarcoidosis. J Am Coll Cardiol 2005;45:1683–90.

7 Patel MR, Cawley PJ, Heitner JF, et al. Detection of myocardial damage in patientswith sarcoidosis. Circulation 2009;120:1969–77.

8 Smedema JP, Snoep G, van Kroonenburgh MP, et al. The additional value ofgadolinium-enhanced MRI to standard assessment for cardiac involvement inpatients with pulmonary sarcoidosis. Chest 2005;128:1629–37.

9 Mahrholdt H, Wagner A, Judd RM, et al. Delayed enhancement cardiovascularmagnetic resonance assessment of non-ischaemic cardiomyopathies. Eur Heart J2005;26:1461–74.

10 Greulich S, Deluigi CC, Gloekler S, et al. CMR imaging predicts death and otheradverse events in suspected cardiac sarcoidosis. JACC Cardiovasc Imaging2013;6:501–11.

11 Hiraga H, Yuwai K, Hiroe M. Guideline for the diagnosis of cardiac sarcoidosis:study report on diffuse pulmonary diseases (in Japanese). Tokyo, Japan: TheJapanese Ministry of Health and Welfare, 1993:23–4.

12 Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerularfiltration rate from serum creatinine: a new prediction equation. Modification of dietin renal disease study group. Ann Intern Med 1999;130:461–70.

13 Matsuo S, Imai E, Horio M, et al. Revised equations for estimated gfr from serumcreatinine in Japan. Am J Kidney Dis 2009;53:982–92.

Figure 4 Relationship between the extent of late gadoliniumenhancement (LGE) and changes in LVEF.

Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187 7

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

14 Yazaki Y, Isobe M, Hiroe M, et al. Prognostic determinants of long-term survival inJapanese patients with cardiac sarcoidosis treated with prednisone. Am J Cardiol2001;88:1006–10.

15 Lang RM, Bierig M, Devereux RB, et al. Recommendations for chamberquantification: a report from the American Society of Echocardiography’s Guidelinesand Standards Committee and the Chamber Quantification Writing Group,developed in conjunction with the European Association of Echocardiography, abranch of the European Society of Cardiology. J Am Soc Echocardiogr2005;18:1440–63.

16 Amano Y, Kumita S, Takayama M, et al. Comparison of contrast-enhanced MRIwith iodine-123 BMIPP for detection of myocardial damage in hypertrophiccardiomyopathy. AJR Am J Roentgenol 2005;185:312–18.

17 Kono AK, Yamada N, Higashi M, et al. Dynamic late gadolinium enhancementsimply quantified using myocardium to lumen signal ratio: normal range of ratioand diffuse abnormal enhancement of cardiac amyloidosis. J Magn Reson Imaging2011;34:50–5.

18 Bondarenko O, Beek AM, Hofman MB, et al. Standardizing the definition ofhyperenhancement in the quantitative assessment of infarct size and myocardialviability using delayed contrast-enhanced CMR. J Cardiovasc Magn Reson2005;7:481–5.

19 Gao P, Yee R, Gula L, et al. Prediction of arrhythmic events in ischemic and dilatedcardiomyopathy patients referred for implantable cardiac defibrillator: evaluation ofmultiple scar quantification measures for late gadolinium enhancement magneticresonance imaging. Circ Cardiovasc Imaging 2012;5:448–56.

20 Chiu CZ, Nakatani S, Zhang G, et al. Prevention of left ventricular remodeling bylong-term corticosteroid therapy in patients with cardiac sarcoidosis. Am J Cardiol2005;95:143–6.

21 Gulati A, Jabbour A, Ismail TF, et al. Association of fibrosis with mortality andsudden cardiac death in patients with nonischemic dilated cardiomyopathy. JAMA2013;309:896–908.

22 Assomull RG, Prasad SK, Lyne J, et al. Cardiovascular magnetic resonance, fibrosis,and prognosis in dilated cardiomyopathy. J Am Coll Cardiol 2006;48:1977–85.

23 O’Hanlon R, Grasso A, Roughton M, et al. Prognostic significance of myocardialfibrosis in hypertrophic cardiomyopathy. J Am Coll Cardiol 2010;56:867–74.

24 Bruder O, Wagner A, Jensen CJ, et al. Myocardial scar visualized by cardiovascularmagnetic resonance imaging predicts major adverse events in patients withhypertrophic cardiomyopathy. J Am Coll Cardiol 2010;56:875–87.

25 Rubinshtein R, Glockner JF, Ommen SR, et al. Characteristics and clinicalsignificance of late gadolinium enhancement by contrast-enhanced magneticresonance imaging in patients with hypertrophic cardiomyopathy. Circ Heart Fail2010;3:51–8.

26 Butler CR, Kumar A, Toma M, et al. Late gadolinium enhancement in cardiactransplant patients is associated with adverse ventricular functional parameters andclinical outcomes. Can J Cardiol 2013;29:1076–83.

27 Motoyasu M, Kurita T, Onishi K, et al. Correlation between late gadoliniumenhancement and diastolic function in hypertrophic cardiomyopathy assessed bymagnetic resonance imaging. Circ J 2008;72:378–83.

28 Leong DP, Chakrabarty A, Shipp N, et al. Effects of myocardial fibrosis andventricular dyssynchrony on response to therapy in new-presentation idiopathicdilated cardiomyopathy: insights from cardiovascular magnetic resonance andechocardiography. Eur Heart J 2012;33:640–8.

29 Kim RJ, Wu E, Rafael A, et al. The use of contrast-enhanced magnetic resonanceimaging to identify reversible myocardial dysfunction. N Engl J Med2000;343:1445–53.

8 Ise T, et al. Heart 2014;0:1–8. doi:10.1136/heartjnl-2013-305187

Heart failure and cardiomyopathies

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

doi: 10.1136/heartjnl-2013-305187 published online May 14, 2014Heart

 Takayuki Ise, Takuya Hasegawa, Yoshiaki Morita, et al. cardiac sarcoidosisin LV function after steroid therapy inadverse outcomes and lack of improvement

predictscardiovascular magnetic resonance Extensive late gadolinium enhancement on

http://heart.bmj.com/content/early/2014/05/14/heartjnl-2013-305187.full.htmlUpdated information and services can be found at:

These include:

References http://heart.bmj.com/content/early/2014/05/14/heartjnl-2013-305187.full.html#ref-list-1

This article cites 28 articles, 5 of which can be accessed free at:

P<P Published online May 14, 2014 in advance of the print journal.

serviceEmail alerting

the box at the top right corner of the online article.Receive free email alerts when new articles cite this article. Sign up in

CollectionsTopic

(3096 articles)Epidemiology   � (7620 articles)Drugs: cardiovascular system   �

 Articles on similar topics can be found in the following collections

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from

Notes

(DOIs) and date of initial publication. publication. Citations to Advance online articles must include the digital object identifier citable and establish publication priority; they are indexed by PubMed from initialtypeset, but have not not yet appeared in the paper journal. Advance online articles are Advance online articles have been peer reviewed, accepted for publication, edited and

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on May 19, 2014 - Published by heart.bmj.comDownloaded from