Corticosteroid and Immunosuppressant Therapy for Cardiac Sarcoidosis: A Systematic Review.

Background Corticosteroid therapy for the treatment of clinically manifest cardiac sarcoidosis is generally recommended. Our group previously systematically reviewed the data in 2013; since then, there has been increasing quality and quantity of data and also interest in nonsteroid agents. Methods and Results Studies were identified from MEDLINE, EMBASE, Cochrane Controlled Trials Register, Cochrane Database of Systematic Reviews, and the National Institutes of Health ClinicalTrials.gov database. The quality of included articles was rated using Scottish Intercollegiate Guidelines Network 50. Outcomes examined were atrioventricular conduction, left ventricular function, ventricular arrhythmias, and mortality. A total of 3527 references were retrieved, and 34 publications met the inclusion criteria. There were no randomized trials, and only 2 studies were rated good quality. In the 34 reports (total of 1297 patients), 1125 patients received corticosteroids, 235 received additional or other immunosuppressant therapy, and 97 patients received no therapy. There were 178 patients treated for atrioventricular conduction disease, with 76/178 (42.7%) improving. In contrast, 21 patients were not treated with corticosteroids and/or immunosuppressant therapy, and none of them improved. Therapy was associated with the prevention of deterioration in left ventricular function. A total of 8 publications reported on ventricular arrhythmia burden, and 19 reported on mortality; the data quality was too limited to draw conclusions for the latter 2 outcomes. Conclusions The best quality data relate to atrioventricular nodal conduction and left ventricular function recovery. In both situations, therapy with corticosteroids and/or immunosuppressant therapy were sometimes associated with positive outcomes. The data quality is too limited to draw conclusions for ventricular arrhythmias and mortality.

cardiac sarcoidosis

F‐18 fluorodeoxyglucose

immunosuppressant therapy

Clinical Perspective

What Is New?

A total of 34 publications met the inclusion criteria of the study, 1125 patients received corticosteroids, 235 received additional or other immunosuppressant therapy, and 97 patients received no therapy.

What Are the Clinical Implications?

There were 178 patients with conduction disease treated with corticosteroid and/or immunosuppression therapy, with 76/178 (42.7%) improving. In contrast 21 patients were not treated with such therapy and none of them improved.

Therapy was associated with the prevention of deterioration in left ventricle function; the data quality was too limited to draw conclusions regarding the other 2 outcomes: ventricular arrythmia burden and mortality.

Sarcoidosis is a multisystem, granulomatous disease of unknown etiology. Pulmonary and lymph node involvement are most common, but heart, liver, spleen, skin, eyes, phalangeal bones, parotid gland, or other organs can be affected. Approximately 5% of patients with sarcoidosis have clinically manifest cardiac involvement, and another 20% to 25% of patients have asymptomatic (ie, clinically silent) cardiac involvement (based on autopsy studies and recent data using late gadolinium‐enhanced cardiac magnetic resonance technology). Studies suggest an increasing prevalence of cardiac sarcoidosis (CS); however, this is most likely the result of enhanced imaging technology and/or more in‐depth investigation. ,

Corticosteroids for the treatment of clinically manifest CS is advocated by experts and guidelines based on very modest quality data. , Our group previously systematically reviewed the data in 2013 ; since then, there has been an increasing quality and quantity of data and increasing interest in other nonsteroid agents. Hence the objectives of this study were to systematically review the published data from studies on the outcomes after corticosteroid and other immunosuppressant therapies (ISTs) in patients diagnosed with CS. Specifically, we focused on atrioventricular conduction recovery, left ventricular (LV) function, ventricular arrhythmia, and mortality.

METHODS

The authors declare that all supporting data are available within the article (and its online supplementary files). A literature search of electronic databases was performed by a medical librarian. The initial search was created in Medline (Ovid MEDLINE ALL from 1946 to September 11, 2018) using a combination of subject headings and keywords (Table S1). The search strategy was peer reviewed by another medical librarian using the Peer Review of Electronic Search Strategies Checklist guidelines and then translated to the following databases: Embase Classic+Embase 1947 to September 11, 2018 (Ovid); Cochrane Central Register of Controlled Trials August 2018 (Ovid). The search strategy is shown in Table S1. The initial searches were run on September 12, 2018. To ensure retrieval of the most up‐to‐date studies before submission of the article, a second search using the same databases and search strategies was performed on October 13, 2020. A gray literature search was performed by 1 investigator (S.F.) using the search terms specified in the Table S1 and through the following databases: clinicaltrials.gov, National Institute for Health and Care Excellent, Canadian Agency for Drugs and Technologies in Health, and OpenGrey.

Inclusion criteria were (1) publications in English or translations, (2) studies reporting original outcomes in patients with CS diagnosed based on either (A) Heart Rhythm Society Expert Consensus Recommendations on Criteria for Diagnosis of CS or (B) Japanese Ministry of Health and Welfare Criteria, (3) patients receiving corticosteroids and/or ISTs for the treatment of CS. Exclusion criteria were studies including data from <5 patients and with <3 months of follow‐up.

A total of 2 investigators (S.F., M.S.) independently reviewed all publications with regard to the inclusion and exclusion criteria. Full text of articles potentially meeting these criteria were retrieved and reviewed in detail by 2 investigators independently (S.F., M.S.), and a third investigator (D.B.) settled disagreements. The following data were retrieved from included studies: baseline patient characteristics, sample size, study design, choice of immunosuppressive agent used, and outcomes of interest. Outcomes of interest included atrioventricular conduction, LV systolic function, ventricular arrhythmia, and mortality.

Quality Rating

The Scottish Intercollegiate Guidelines Network 50 was used to assess the quality of each article regarding design quality and bias. Each study was assigned an overall assessment score (good, fair, poor) and given a numerical score (of 14) based on the Scottish Intercollegiate Guidelines Network 50 questionnaire.

RESULTS

Systematic Review

A total of 3527 references were retrieved (Figure); 34 published studies fulfilled the criteria for this review. There were no randomized controlled trials. Study design and inclusion and exclusion criteria are summarized in Table 1. , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , ,

Figure 1

Literature search strategy.

 

Table 1

Study Characteristics and Inclusion Criteria

First Author	Year	Number of Centers	Sample Size	Study Design	Inclusion Criteria	Criteria for Diagnosis of CS	Exclusion Criteria	SIGN 50 Score (of 14)	SIGN 50 Overall Assessment
Okamoto 11	1999	Single center	5	Retrospective	Cutaneous sarcoidosis patients who developed cardiac manifestations	JMHW	None	3	−
Yazaki 12	2001	Multicenter	95	Retrospective	CS diagnosis	JMHW	None	5	+
Kato 13	2003	Single center	20	Retrospective	Patients with CS presenting with AVB	JMHW	LVEF <50%	6	+
Chapelon‐Abric 14	2004	Multicenter	41	Retrospective	CS diagnosis	JMHW	None	6	+
Chiu 15	2005	Single center	43	Retrospective	CS diagnosis and treatment with steroids	JMHW	CAD, no regular follow‐up	8	+
Futamatsu 16	2006	Multicenter	21	Retrospective	CS diagnosis and treatment with steroids	JMHW	None	8	+
Banba 17	2007	Single center	15	Retrospective	CS diagnosis and presenting with PVCs	JMHW	None	5	+
Kudoh 18	2010	Single center	10	Prospective	CS diagnosis	JMHW	Patients already treated with steroids	3	−
Yodogawa 19	2011	Single center	31	Retrospective	CS diagnosis and presenting with PVCs	JMHW	CAD, other cardiac disease	5	+
Kandolin 20	2011	Single center	18	Retrospective	Age 18–55 y, presenting with AVB and diagnosed with CS or GCM	JMHW	None	7	+
Yodogawa 21	2013	Multicenter	15	Retrospective	Patients with CS presenting with AVB	JMHW	CAD, other cardiac disease	8	+
Nagai 22	2014	Single center	17	Prospective	Patients with CS treated with immunosuppressants	JMHW	None	11	++
Ise 23	2014	Single center	43	Retrospective	Patients with CS who had undergone CMR	JMHW	CAD	10	+
Takaya 24	2014	Single center	30	Retrospective	Patients with CS with positive myocardial uptake of 67 Ga on 18F‐FDG‐PET	JMHW	None	6	+
Takaya 25	2015	Single center	53	Retrospective	Patients with CS with initial presentation with either AVB, CHF, or VT	JMHW	None	10	+
Kandolin 3	2015	Multicenter	110	Retrospective	CS diagnosis	JMHW	None	6	+
Nagai 26	2015	Single center	83	Retrospective	CS diagnosis	JMHW	CAD	6	+
Orii 27	2015	Single center	32	Retrospective	CS diagnosis	JMHW	CAD, other cardiac disease	6	+
Nagai 28	2016	Single center	61	Retrospective	CS diagnosis and treatment with steroids	JMHW	CAD, failure to follow‐up >5 y	7	+
Segawa 29	2016	Single center	68	Retrospective	CS diagnosis	JMHW	None	4	−
Padala 30	2017	Single center	30	Retrospective	CS diagnosis	HRS	None	5	+
Ahmadian 31	2017	Single center	17	Retrospective	Patients with CS who had undergone serial FDG‐PET	JMHW	None	5	+
Yalagudri 32	2017	Single center	18	Retrospective	Patients with CS presenting with VT	HRS	None	8	+
Kaida 33	2018	Single center	15	Retrospective	Patients with CS presenting with AVB	JMHW	None	7	+
Muser 34	2018	Single center	20	Retrospective	Patients with CS presenting with refractory VT referred for catheter ablation	HRS	No FDG‐PET	8	+
Fussner 35	2018	Two centers	91	Retrospective	CS diagnosis	JMHW	None	8	+
Ballul 36	2019	Single center	36	Retrospective	Histologically proven CS	HRS	Asymptomatic patients	10	+
Chiba 37	2020	Single center	91	Retrospective	CS diagnosis	JMHW	None	9	+
Harper 38	2019	Single center	36	Retrospective	CS diagnosis and infliximab initiation	JMHW	Prior infliximab	9	+
Rosenthal 39	2019	Single center	28	Retrospective	CS diagnosis and immunosuppression, at least 2 PET studies	JMHW	<6 mo follow‐up, prior immunosuppression	5	+
Koyanagawa 40	2019	Single center	38	Retrospective	Patients with CS with PET CT and SPECT studies	JMHW	Prior corticosteroid use	7	+
Orii 41	2020	Single center	8	Retrospective	Patients with CS presenting with AVB	HRS	No CMR on file	8	+
Medor 42	2020	Single center	20	Prospective	Patients with CS undergoing ICD implantation	HRS	Normal FDG‐PET	11	++
Gilotra 43	2020	Single center	38	Retrospective	Patients with CS treated with TNF‐α inhibitor	HRS		9	+
Total			1297

SIGN 50 overall assessment: “++”=good, “+”=fair, “−”=poor. 18F‐FDG indicates 18fluourine‐flourodeoxyglucose; AVB, atrioventricular block; CAD, coronary artery disease; CHF, congestive heart failure; CMR, cardiac magnetic resonance imaging; CS, cardiac sarcoidosis; CT, computed tomography; FDG, F18‐fluorodeoxyglugose; GCM, giant cell myocarditis; HRS, Heart Rhythm Society Expert Consensus Recommendations on Criteria for Diagnosis of CS; ICD, implantable cardioverter defibrillator; JMHW, Japanese Ministry of Health and Welfare Criteria; LVEF, left ventricular ejection fraction; PET, positron emission tomography; PVC, premature ventricular contraction; SIGN, Scottish Intercollegiate Guidelines Network; SPECT, single‐photon emission computed tomography; TNF, tumor necrosis factor; and VT, ventricular tachycardia.

Quality rating using the Scottish Intercollegiate Guidelines Network 50 checklist for cohort studies is summarized in Table 1. Only 2 studies were deemed to be good in quality, with 29 studies rated as fair and 3 rated as poor. The average rating was 7.0, and the highest score was 11.

Patient Demographics

The number of patients (treated and not treated), follow‐up, average age, sex, and presentation are summarized in Table 2. In total, the studies included 1297 patients (1125 treated with corticosteroids, 235 treated with other ISTs, and 97 patients not treated with either). Pulmonary involvement was present in 57.9% of patients.

Table 2

Patient Demographics From Selected Studies

First Author	Number of Patients	Steroids and/or Other ISTs			Men: Women	Pulmonary Involvement, N (%)	Race/Ethnicity	Mean Follow‐Up, mo*	Average Age, y*	HF, N (%)	AVB, N (%)	Ventricular Arrhythmia, N (%)
		Corticosteroids	Other ISTs	None
Okamoto 11	5	5	0	0	1:4	3 (60)	Japanese	>24	62.2±9.3	1 (20)	3 (60)	0
Yazaki 12	95	75	0	20	34:61	56 (75)	Japanese	68±42	51±13, 57±15	43 (47.8)	28 (37.3)	17 (18.9)
Kato 13	20	7	0	13	1:19	19 (95)	Japanese	79.4±39.9	67.2±8.2	0	20 (100)	0
Chapelon‐Abric 14	41	39	0	2	23:18	37 (90)	73% White, 27% non‐specified	58 (7–312)	38 (18–66)	17 (41.5)	9 (22.0)	1 (2.4)
Chiu 15	43	43	0	0	16:27	11 (26)	Japanese	88±48	48±14	N/A	N/A	N/A
Futamatsu 16	21	21	0	0	6:15	11 (52)	Japanese	48.8±38.7	56.0±11.7	N/A	N/A	14 (66.7)
Banba 17	15	9	0	6	8:7	9 (60)	Japanese	6	53±13	N/A	10 (66.7)	7 (56.7)
Kudoh 18	10	10	0	0	2:8	5 (50)	Japanese	6	62.8±0.5	N/A	4 (40)	3 (30)
Yodogawa 19	31	31	0	0	6:25	21 (68)	Japanese	7.3±5.9	60±9	N/A	12 (38.7)	14 (45.3)
Kandolin 20	18	17	0	1	2:16	6 (33)	White	38±32	48 (36–55)	N/A	18 (100)	N/A
Yodogawa 21	15	15	0	0	2:13	12 (80)	Japanese	85.2±63.6	59.9±9.7	5 (33.3)	15 (100)	1 (6.67)
Nagai 22	17	7	10	0	3:14	17 (100)	Japanese	87.4±73.1	70.1±5.9	4 (24)	13 (76)	N/A
Ise 23	43	43	0	0	15:28	9 (21)	Japanese	39±19	59±10	N/A	N/A	N/A
Takaya 24	30	30	0	0	1:20	18 (60)	Japanese	12	61±12	N/A	13 (43)	12 (40)
Takaya 25	53	42	0	11	20:33	34 (64)	Japanese	34 (1–149)	60±13	31 (58.5)	22 (41.5)	31 (58.5)
Kandolin 3	110	110	62	0	33:77	11 (10)	White	79 (12–303)	51±9	65 (59)	48 (45)	36 (33)
Nagai 26	83	67	2	16	24:59	50 (60.2)	Japanese	91.2±52.8	60±12	11 (13.2)	33 (39.8)	24 (29)
Orii 27	32	6	0	0	8:24	19 (59.4)	Japanese	26±6	64±9	N/A	15 (46.9)	8 (25)
Nagai 28	61	61	1	0	17:44	35 (57)	Japanese	118.8 (94.8–156)	59 (52–67)	9 (15)	18 (30)	22 (36)
Segawa 29	69	69	0	0	18:50	52 (75.4)	Japanese	66	57±11.2	N/A	29 (42)	17 (66.7)
Padala 30	30	27	0	3	16:14	30 (100)	…	19 (7–66)	58±10	14 (47)	5 (17)	14 (47)
Ahmadian 31	17	17	0	0	8:9	15 (88)	70% White, 30% Black	22	58.2±12	N/A	7 (41)	N/A
Yalagudri 32	18	14	14	4	12:6	…	…	38 (10–75)	38±14	4 (22)	0	18 (100)
Kaida 33	15	15	0	0	4:11	6 (40)	Japanese	N/A	61.3±11	2 (13)	15 (100)	3 (20)
Muser 34	20	20	6	0	14:6	…	…	35 (20–66)	51±9	…	…	20 (100)
Fussner 35	91	41	29	21	65:26	38 (92.7)	74% White	44 (20–77)	51 (44–61)	47 (51.6)	31 (34.1)	22 (24.2)
Ballul 36	36	36	12	0	20:16	36 (100)	72% Black	43 (12–182)	48.5 (22–76)	13 (38.9)	12 (33.3)	9 (25)
Chiba 37	91	91	0	0	25:66	62 (68.1)	Japanese	84	57±11	35 (38.5)	31 (34.1)	33 (36.2)
Harper 38	36	32	36	0	26:10	26 (72)	78% White, 22% Black	12	46±11	6 (16.6)	7 (19.4)	8 (22.2)
Rosenthal 39	28	27	25	0	16:12	21 (72.4)	82% White, 11% Black	149.2±18	52.2	11 (39.3)	11 (39.3)	18 (64.3)
Koyanagawa 40	38	38	0	0	10:28	…	Japanese	34.5 (5–51.8)	63 (51–68)	N/A	N/A	N/A
Orii 41	8	8	0	0	0:8	…	Japanese	28±6	65±5	N/A	8 (100)	N/A
Medor 42	20	20	0	0	8:12	9 (45)	95% White	13.8±11	59.7±7.7	12 (60)	14 (70)	2 (10)
Gilotra 43	38	33	38	0	22:16	29 (62)	43% Black, 38% White	40.4	49.9±5.9	13 (27.7)	5 (10.6)	13 (27.7)
Totals	1297	1125	235	97

AVB indicates atrioventricular block; HF, heart failure; IST, immunosuppressant therapy; and N/A, not available.

Data are presented as mean±SD or median (interquartile range)

Corticosteroid and Other Immunosuppression Therapies

The number of patients receiving steroids, other ISTs, and no therapy is summarized in Table 2. The therapy protocols, doses, duration, and so on are shown in Table S2.

Outcomes: Atrioventricular Conduction Recovery

A total of 13 studies reported outcomes with respect to recovery of high‐grade atrioventricular block (Table 3). In these studies, 178 patients received corticosteroids and/or other ISTs, whereas 21 did not. Of 178 patients receiving immunosuppression, 76 had atrioventricular conduction recovery (42.6%). Of the 21 patients who were not treated with immunosuppressants, none of them had atrioventricular conduction recovery (0%).

Table 3

Studies Investigating the Effect of Immunosuppression on Atrioventricular Recovery in Patients With CS Presenting With AVB

First Author	Number of Patients on Steroids and/or Other ISTs	Atrioventricular Recovery, N (%)	Number of Patients on No Immunosuppressants	Atrioventricular Recovery, N (%)
Okamoto 11	3	3 (100)	0	…
Kato 13	7	4 (57.1)	13	0 (0)
Chapelon‐Abric 14	9	7 (77.7)	0	…
Banba 17	9	5 (55.6)	2	0 (0)
Yodogawa 19	12	4 (33.3)	0	…
Kandolin 20	17	4 (23.5)	1	0 (0)
Yodogawa 21	15	7 (46.7)
Takaya 25	17	7 (41.1)
Kandolin 3	35	7 (20)
Orii 27	10	6 (60)	5	0/5 (0)
Padala 30	5	2 (40)
Chiba 37	31	12 (38.7)	0
Orii 41	8	5 (62.5)	0
Total	178	76 (42.6)	21	0 (0)

AVB indicates atrioventricular block; and CS, cardiac sarcoidosis.

Outcomes: LV Function

The effect of corticosteroids and/or ISTs on LV function was reported in 18 studies, and the data are summarized in Table 4. Table 4 shows the results for patients with normal LV function. There was a total of 194 treated patients in 9 studies treated with immunosuppression, and in all studies, there was no significant change in LV ejection fraction (LVEF). In contrast, in the 1 study with a nontreatment group, the LVEF fell from 60.5±6.4% to 37.6±17.3% in 13 patients (mean±SD).

Table 4

Studies Investigating the Effect of Immunosuppression on LV Function in Patients With CS With Initially Normal LV Function

First Author	Steroids and/or Other ISTs			No Immunosuppressants			Comments
	Number of Patients	LVEF Before Treatment	LVEF After Treatment	Number of Patients	LVEF Before Treatment	LVEF After Treatment
Kato 13	7	66.7±6.5%	62.1±4.4%	13	60.5±6.4%	37.6±17.3%
Chiu 15	22	69±7%	69±5%	0	…	…	No difference (P=0.277)
Yodogawa 19	17	52.4±13.2%	55.1±12.2%	0	…	…	Trend to improvement (P=0.06)
Nagai 22	7	52.3±6.07%	45.7±15.5%	0	…	…	P=ns
Kandolin 3	44	56.8±5.6	54.9±7.6	0	…	…	No difference (P=0.145)
Padala 30	14	56%	54%	0			No difference (P=0.26)
Ahmadian 31	17	53.18±20	54.6±14	0			No difference (P=0.682)
Rosenthal 39	28	53.4±12.3%	48.7±11.4%	0			No difference (P=0.14)
Koyanagawa 40	38	64.7 (58.4–72.1) %	67.0 (58.2–71.1%)	0			No difference (P=0.65)
Total	194			13

All continuous data are presented as mean±SD or median (interquartile range). CS indicates cardiac sarcoidosis; IST, immunosuppressant therapy; LV, left ventricular; LVEF, left ventricular ejection fraction; and ns, nonsignificant.

Table 5 shows the results for patients with mild‐moderate dysfunction. There was a total of 324 treated patients in 11 studies, and in 4 studies the mean LVEF improved with treatment and in 5 studies it did not, and on 1 study there was a slight deterioration (45.2±13.6 to 40.0±12.0%; P=0.038). The other article presented results stratified by the extent of baseline late gadolinium enhancement and found that patients with extensive late gadolinium enhancement did not improve and vice versa. In the 1 study with a nontreatment group, LVEF decreased from 32.5% to 18.5%; there was no P value quoted.

Table 5

Studies Investigating the Effect of Immunosuppression on LV Function in Patients With CS With Initially Mild‐Moderate LV Dysfunction

First Author	Steroids and/or Other ISTs			No Immunosuppressants			Comments
	Number of Patients	LVEF Before Treatment	LVEF After Treatment	Number of Patients	LVEF Before Treatment	LVEF After Treatment
Padala 30	9	25%	46%	2 (3)*	41%	37%	P<0.001 for steroids
Chiu 15	10	40±10%	51±12%	0	…	…	Increase in EF (P=0.008)
Kudoh 18	10	34.6±12%	48.8±18.6%	0	…	…	Increase in EF (P<0.001)
Nagai 22	10	49.7±6.9%	53.6±13.3%	0	…	…	P=ns
Ise, 23 small‐extent LGE	21	45±11%	50±10%	0	…	…	Increase in EF (P<0.001)
Ise, 23 large‐extent LGE	22	36±6%	35±8%	0	…	…	No difference (P=0.213)
Takaya 24	30	43±15%	47±16%	0	…	…	No difference (P=0.367)
Kandolin 3	36	40.9±4.1	40.8±7.6	0	…	…	No difference (P=0.979)
Nagai 26	67	35.9±14.4	43.8% *	16	35.2±15.8%	18.5% *	P=0.03 for difference between groups
Kaida 33	15	46.3±14.3	47.9±10.6	0			No difference
Harper 38	36	41% (32–55)	41% (32–54)	0			No difference (P=0.43)
Medor 42	20	45.2±13.6%	40.0±12.0%	0			Decrease in EF (P=0.038)
Gilotra 43	38	45.0±16.5%	47.0±15.0%	0			No difference (P=0.10)
Total	324			18

All continuous data are presented as mean±SD or median (interquartile range). CS indicates cardiac sarcoidosis; EF, ejection fraction; IST, immunosuppressant therapy; LGE, late gadolinium enhancement; LV, left ventricular; LVEF, left ventricular ejection fraction; and ns, nonsignificant.

A total of 2 patients did not receive steroids, and 3 patients had delayed steroids.

Extracted from data provided in the article.

Table 6 shows the results for patients with severe dysfunction. There was a total of 54 treated patients in 4 studies; in 2 articles, there was no change in LVEF, and in 2 articles there was a significant improvement in mean LVEF. There were data on untreated patients in 2/4 articles. In 16 patients, the mean LVEF declined from 35.2% to 18.5%. In the other report, there were 2 patients who did not receive steroids. Their data is grouped with 3 other patients who received late steroids, and in the 5 patients the mean LVEF declined from 41% to 27%.

Table 6

Studies Investigating the Effect of Immunosuppression on LV Function in Patients With CS With Initially Severe LV Dysfunction

First Author	Steroids and/or Other ISTs			No Immunosuppressants			Comments
Chiu 15 EF <30%	11	22±7%	19±5%	0	…	…	No difference (P=0.082)
Yodogawa 19 EF <30%	14	26.2±5.5%	26.5±7.5%	0	…	…	No difference (P=0.77)
Kandolin 3 EF <35%	22	27.9±4.1	34.1±8.3	0	…	…	Increase in EF (P=0.005)
Kaida 33 EF <30%	7	20%	43%	0	…	…	Increase in EF (P<0.001)
Total	54			0

All continuous data are presented as mean±SD or median (interquartile range). CS indicates cardiac sarcoidosis; EF, ejection fraction; IST, immunosuppressant therapy; and LV, left ventricular.

Chiu

Yodogawa

Kaida

Outcomes: Ventricular Tachyarrhythmia

The effect of corticosteroids and/or ISTs on ventricular arrhythmia recurrence was examined in 8 articles (5 articles reported on sustained arrhythmia, and 3 articles reported on premature ventricular contraction [PVC] burden and nonsustained ventricular tachyarrhythmia (VT), as outlined in Tables 7 and 8, respectively , ). The data quality is too limited to draw conclusions.

Table 7

Studies Evaluating the Effect of Immunosuppression on Recurrence of Sustained Ventricular Arrhythmia in Patients With CS

Year	First Author	Follow‐Up, mo	End Points	Number of Patients Treated With Corticosteroids and/or Other ISTs	Arrhythmia Recurrence, N (%)	Number of Patients Not Treated With Immunosuppressants	Arrhythmia Recurrence	Comments
2006	Futamatsu 16	48.8±38.7	Sustained VT/VF	7	1/7 (14)	0		5 of 6 patients with no recurrence were also started in amiodarone
2016	Segawa 29	66	Sustained VT/VF	17	12/17 (71)	0		Presteroid VT was an independent predictor—7.64 (3.05–19.14)—of poststeroid VT
2017	Padala 30	19	Sustained VT/VF	11	3/11 (33)	3	3/3
2017	Yalagudri 32	38	Sustained VT/VF	14†	4/14 (36)	4	0/4 (all treated with ablation)
2018	Muser 34	35 (20–66)	Sustained VT/VF	20†	12 (60)	0		All patients also had ablation; patients stratified to PET responders and nonresponders. Responders: 2/9 (22%) had ventricular arrhythmia recurrence. Nonresponders: 10/11 (91%) had ventricular arrhythmia recurrence.
Total				69		7

All continuous data are presented as mean±SD or median (interquartile range). CS indicates cardiac sarcoidosis; IST, immunosuppressant therapy; PET, positron emission tomography; VF, ventricular fibrillation; and VT, ventricular tachycardia.

All studies used corticosteroid monotherapy for immunosuppression except Yalagudri et al, where all 14 patients were treated with methotrexate and corticosteroid combinations. In Muser et al, 14 patients were treated with corticosteroid monotherapy and 6 patients received both corticosteroids and methotrexate.

Table 8

Studies Evaluating the Effect of Immunosuppression on PVC Burden and NSVT in Patients With CS

Year	First Author	Follow‐Up, mo	End Points	Number of Patients Treated With Corticosteroids and/or Other ISTs	Arrhythmia Recurrence	Number of Patients Not Treated With Immunosuppressants	Arrhythmia Recurrence	Comments
2007	Banba 17		PVC burden	9	No change in PVC burden before and after steroids	9		All patients had positive gallium scan before initiation of steroids
2011	Yodogawa 19	7.3±5.9	NSVT or PVC burden	31	No change in PVC burden or NSVT prevalence	31
2020	Medor 42	13.1±11	NSVT or PVC burden	20	See comments	20	See comment	Significant increase in both end points after corticosteroids (for NSVT P=0.017, for PVC P=0.008)
Total				60		60

All continuous data are presented as mean±SD or median (interquartile range). CS indicates cardiac sarcoidosis; IST, immunosuppressant therapy; NSVT, nonsustained ventricular tachycardia; and PVC, premature ventricular contraction.

Outcomes: Mortality

The mortality data are summarized in Table 9. A total of 19 studies had data on mortality (713 patients treated with corticosteroids and/or ISTs, and 61 patients did not receive treatment). The data quality is too limited to draw conclusions.

Table 9

Studies Investigating the Effect of Immunosuppression on Mortality in Patients With CS

Year	First Author	Follow‐Up, mo	Steroids and/or Other ISTs		No Immunosuppressants
			Number of Patients	Mortality, N (%)	Number of Patients	Mortality, N (%)
1999	Okamoto 11	>24	5	0 (0)	0	…
2001	Yazaki 12	68±42	75	29 (39)	20	20 (100)
2003	Kato 13	79.4±39.9	7	0 (0)	13	2 (15.4)
2005	Chiu 15	88±48	43	7 (16)	0	…
2006	Futamatsu 16	48.8±38.7	21	0 (0)	0	…
2007	Banba 17	6	9	0 (0)	6	N/A
2010	Kudoh 18	6	10	0 (0)	0	…
2011	Kandolin 20	48 (1–123)	17	2 (11.8)	1	0 (0)
2011	Yodogawa 19	7.3±5.9	31	0 (0)	0	…
2013	Yodogawa 21	85.2±63.6	15	1 (6.7)	0	…
2014	Ise 23	39±19	43	6 (14)	0	…
2015	Kandolin 3	79.2	102	14 (13.7)	0
2015	Takaya 25	34 (1–149)	42	20 (47.6)	0
2018	Muser 34	35 (20–66)	20	1 (5)	0
2018	Fussner 35	44 (20–77)	70	5 (6.8)	21	1 (4.8)
2019	Ballul 36	43 (12–182.4)	36	3 (8.3)	0
2020	Chiba 37	84	91	4 (4.4)	0
2019	Koyanagawa 40	34.6 (5.0–51.8)	38	3 (7.9)	0
2020	Gilotra 43	40.4	38	0 (0)	0
Total			713		61

All continuous data are presented as mean±SD or median (interquartile range). CS indicates cardiac sarcoidosis; IST, immunosuppressant therapy; and N/A, not available.

DISCUSSION

This review synthesizes data from 1297 patients from 34 articles and extends the findings of our previous work published in 2013, which included 303 patients from 10 articles. Our main observations are that the best‐quality data relates to atrioventricular nodal conduction and LV function recovery. In both situations, therapy with corticosteroids and/or ISTs were sometimes associated with positive outcomes. The data quality is too limited to draw conclusions for ventricular arrhythmias and mortality.

There were 178 patients with treated atrioventricular conduction disease, with 76 (42.6%) recovering from atrioventricular conduction; in contrast, 21 patients were not treated, and none improved. However, atrioventricular nodal recovery is unpredictable and can be transient and hence our findings support the recommendations in the 2014 Heart Rhythm Society document on the management of arrhythmias associated with CS. Specifically, all patients with CS and advanced conduction system disease should have a pacemaker implanted. Other accumulating data have shown that these patients, even with normal left ventricular function at presentation, have a significant risk of sudden cardiac death in follow‐up, and hence guidelines and expert opinion recommend implantable cardioverter defibrillator implantation. , ,

The data on LV function suggest therapy with steroids and/or ISTs is associated with preservation of LV function in patients with normal or near normal function at presentation. This is based on data from 194 patients; however, there were only 13 untreated patients, all from the same study. In these patients, the mean LVEF declined without treatment, from 60.5±6.4% to 37.6±17.3%. The data on LV function recovery in patients with initial dysfunction are not conclusive. There was a total of 324 treated patients in 11 studies, and in 4 patients the mean LVEF improved with treatment, in 5 patients it did not, and in 1 patient there was a slight deterioration (45.2±13.6% to 40.0±12.0%; P=0.038). In the final article, the data showed that patients with less late gadolinium enhancement did improve. In the subgroup of patients with severe baseline LV dysfunction, in 2 articles there was no change in LVEF and in 2 articles there was significant improvement in mean LVEF. It should also be noted that there were data on untreated patients in only 2 articles. In 16 patients, the mean LVEF declined from 35.2±15.8% to 18.5%. In the other report, there were 2 patients who did not receive steroids. Their data is grouped with 3 other patients who received late steroids, and in the 5 patients the mean LVEF declined from 41% to 27%.

The interpretation of these observations on LV function is complicated as most patients were simultaneously treated with standard heart failure medications and with cardiac resynchronization therapy in some patients. Also, in many articles the findings were not interpreted in the context of extent of active inflammation and scarring, and the timing of reevaluation of ventricular function was rarely guided by imaging, demonstrating suppression of inflammation. The sum of the data suggest that therapy is associated with prevention of deterioration of LV function. There are insufficient data to answer the additional clinical question of whether immunosuppression improves ventricular function. Both are important goals; however, a better understanding would help inform clinician and patient shared decision making and expectations, and higher quality data are required to investigate this question.

The data related to ventricular arrhythmia are too limited to conclude whether steroids and ISTs are helpful in the management of ventricular arrhythmia. We found the following 2 main issues with the current literature: (1) most patients were treated simultaneously with a combination of corticosteroids and/or ISTs and antiarrhythmic drugs and sometimes ablation and (2) few studies examined outcomes related to the disease activity with gallium or F‐18 fluorodeoxyglucose (FDG) positron emission tomography scanning. It is likely that macro reentry phenomena around areas of fibrosis is the most frequent mechanism of ventricular arrhythmia in CS. , It is also possible that inflammation may play a role in initiating reentry with ventricular ectopy in patients with CS or by slowing conduction in diseased tissue.

Several studies have reported on the recurrence of sustained VT after therapy. Futamatsu et al reported on 7 patients with VT (6 sustained and 1 nonsustained). Following corticosteroid therapy, 6 patients had no recurrence of VT; however, 5 of the 7 patients were also started on amiodarone therapy. Padala et al investigated 14 patients with VT; early initiation of corticosteroid therapy was associated with no VT recurrence in 8/11 patients, and in contrast 3 patients with VT who did not receive early corticosteroid therapy all had recurrent arrhythmia. Segawa et al examined 17 patients, and 12 of 17 patients experienced recurrent VT after initiation of corticosteroid therapy. Finally, Muser et al published findings on 20 patients who underwent ablation for VT after immunosuppression. There was a higher recurrence rate post‐VT ablation in patients who did not have a reduction in FDG uptake on serial positron emission tomography scans.

There are 3 publications that looked at PVC burden before and after steroids. Banba et al studied 9 patients with CS and positive gallium scans and all patients had 24‐hour Holter monitors done before and after treatment. The average PVC burden increased from 3582±5456/day to 4673±6167/day (no P value stated). Also, PVC Lown grade was unchanged in 5 patients but worsened in 4 patients (from grade 0 to grade 1 in 2 patients and from grade 1 to grade 2 in 2 patients).

Yodogawa et al reported on a cohort of 31 patients who also had 24‐hour Holter monitoring before and after therapy and found no change in PVC (from 3098±5902 to 3024±8081; P=0.89) or nonsustained VT burden. The third article used implantable cardioverter defibrillator diagnostics to assess the relationship between nonsustained VT and premature ventricular complex PVC burden in patients with newly diagnosed clinically manifest CS. All 20 patients were corticosteroid responders based on serial FDG uptake. Patients with untreated CS had an average of 496.4±879.1 PVCs per day. After treatment with corticosteroids, the average PVC count increased to 1590.1±2362.2 per day (P=0.008). There was also a statistically significant increase in episodes of nonsustained VT before and after treatment with corticosteroids (P=0.017). Overall, 18 of 20 patients (90%) had an increase in PVC burden after corticosteroid initiation.

The Heart Rhythm Society consensus document recommended a stepwise approach to ventricular arrhythmia management. The first suggested step is treatment with immunosuppression if there is evidence of active inflammation. Antiarrhythmic medications are often started at the same time, with catheter ablation considered if VT cannot be controlled. Yalagudri et al examined this approach and found that patients with VT in the scar phase (ie, with no inflammation) responded well to antiarrhythmic drugs and ablation. A total of 14 patients had abnormal cardiac FDG uptake on positron emission tomography and were initially treated with immunosuppression therapy. In the inflammatory group, 4 patients had a recurrence of VT during follow‐up, and 3 were found to have disease reactivation. Intensified immunosuppression suppressed VT in all 3 patients, and in the other patient VT recurrence was found to be scar related and was successfully treated with ablation. These data are encouraging but require validation in other centers with larger numbers of patients.

The quantity and quality of the data on mortality has improved since the last systematic review; however, it is still too limited to comment on whether therapy improves prognosis; indeed, the question is unlikely to be answered. The most important conclusion is the generally improved prognosis in more recent studies. This is based on the information from the articles in this review and others (which did not report on steroids or ISTs). For example, Yazaki et al reported that patients with depressed LVEF had a 10‐year survival rate of 27%. Similarly, Chiu et al found the survival rate was 91% after 1 year, 57% after 5 years, and 19% after 10 years in patients with severe dysfunction (LVEF <30%). However, it should be noted that these data were published in 2001 and 2005. In the current era of heart failure therapy, including heart transplantation and mechanical circulatory support, deaths from heart failure have become rare, and the majority of fatalities are attributed to ventricular arrhythmias. In a recent Finnish nationwide study, 10‐year cardiac survival was 92.5% in 102 patients. Other recent studies have also shown a much‐improved prognosis in the current era. ,

Our review found 97 patients treated with nonsteroid therapy (ie, other ISTs); the outcomes were generally not stratified by therapy, and hence it is not possible to draw many conclusions. Methotrexate was the most frequently used, and cyclophosphamide, cyclosporin, hydroxychloroquine, and azathioprine were also used. There has been recent interest in tumor necrosis factor inhibitors (infliximab, adalimumab) for second‐line or third‐line therapy for CS, and 3 publications with a total 93 treated patients met the criteria for inclusion in this systematic review. , , The studies reported very similar findings; there were significant reductions in steroid dose and FDG uptake, but no change in other end points, including LVEF, , , atrioventricular nodal function, and VT burden. A total of 2 studies found an important risk of serious infection: 14% and 21%. These studies highlight the need for more information specifically related to the use of the tumor necrosis factor inhibitors in the treatment of CS and in general as to the goals of care in CS therapy (eg, should we aim for clinical stability or clinical stability+complete suppression of all cardiac FDG uptake).

Limitations

There are significant limitations to the data. There were no randomized controlled trials, and all of the studies were rated as being poor or fair quality. Most important, there were a very limited number of nontreated patients. Treatment regimens were also variable, and outcomes were not stratified by therapy subtype. Some studies had a short follow‐up duration of <1 year, making it difficult to comment on the durability of corticosteroid response.

Conclusions

Our systematic review identified 34 articles reporting outcomes following steroid therapy and IST. There were no randomized trials. Treatment protocols and outcomes were not standardized, and there were a limited number of nontreated patients. The best data relate to atrioventricular conduction and LV systolic function, and therapy appears to be beneficial in some patients. In both situations, therapy with corticosteroids and/or ISTs were sometimes associated with positive outcomes. The data quality is too limited to draw conclusions for ventricular arrhythmias and mortality. However, recent data suggest that the prognosis is generally much improved compared with older studies. There is a need for large, multicenter prospective registries and trials in this patient population. Higher quality evidence is needed and should be forthcoming from ongoing studies, including the CHASM CS‐RCT (Canadian/Japanese/US/European Cardiac Sarcoidosis multicenter randomized controlled trial).

Sources of Funding

None.

Disclosures

None.

Tables S1–S2

References 3, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43

Click here for additional data file.