Regarding the results of CMR study as a “gold standard” method, we calculated the sensitivity and specificity, as well negative and positive predictive values for symptoms and laboratory tests suggesting cardiac involvement (Table 4 for details).
Analysis of differences between CS-positive and -negative group suggested some features as potential factors increasing chance for CS. When analyzed separately (univariate analysis), the odds ratio (OR) for having cardiac involvement in men compared to women was 2.5. Symptoms suggesting CS were present in 61% of the patients (OR: 3.5). Symptom rates in CS-positive and negative groups are presented in Table 5.
Extrathoracic sarcoidosis also was a risk factor for CS (OR: 3.5). In 24 (49%) cases, CS was diagnosed contemporary with radiological progression in the lungs (OR: 3.0). Elevated serum NT-proBNP (>125) was also identified as increasing the chance for CS (OR: 3.8). Another factor associated with cardiac involvement was any changes in routine electrocardiography (OR: 5.4). Elevated C-reactive protein (CRP) and serum ACE were not associated with an increased risk of CS. The forest plot for mentioned risk factors with detailed numeric data is presented in Figure 1.
We performed logistic regression analysis in 176 patients in whom results of all tests were accessible. The regression equation had the form:
y = −3.75 + 1.53 × (sex: male = 1; female = 0) + 1.05 × (symptoms: yes = 1; no = 0) + 1.17 × (ECG changes: yes = 1; no = 0) + 1.00 × (extrathoracic sarcoidosis: yes = 1; no = 0) + 0.96 × (progressive disease: yes = 1; no = 0) + 1.56 × (NT-proBNP >125: yes = 1; no = 0).
Taking into account the risk factors for CS and analyzing their significance in multivariate analysis, we developed a scoring system (CS Risk Index [CSRI]) setting arbitrary points (but in some proportion to coefficients in regression equation) to chosen factors as presented in Table 6.
For example, a male patient seen when disease progressing in the lung with ECG abnormalities but without evidence of extrathoracic disease, without symptoms, with normal NT-proBNP level has the CSRI = 2 + 1 + 1 + 0 + 0 + 0 = 4. Factors like abnormalities in Holter ECG and echocardiography as well continuous variables (e.g., ACE, CRP, NT-proBNP) were not significant in the multivariate models.
Using the proposed scoring system, we calculated a score of CSRI for every patient and performed ROC curve analysis (Fig. 2). The mean value of CSRI was significantly different between CS-positive and negative groups (4.43 ± 1.52 vs. 2.5 ± 1.36. respectively, P < 0.0001).
The best accuracy was achieved at the level of 5 points with a sensitivity of only 50%, but specificity 94%. The negative predicted value was 0.84 and positive predicted value 0.74 with accuracy 0.82. Detailed data of sensitivity, specificity, and accuracy at given CS risk index levels are presented in eTable 1, http://links.lww.com/MD/B188 (see in 3 Supplemental Digital Content, http://links.lww.com/MD/B188). Table 7 shows the numbers of patients with positive and negative results of CMR study with likelihood ratio in chosen CSRI intervals.
None of 26 patients with a CS risk index of 0 or 1 had CS (all females). At the levels from 2 to 4, the post-test probability of CS (assuming 25% prevalence) was comparable to pretest probability. In the group of patients with the score ≥5, post-test probability increased almost 3 times reaching 70%. All patients with a CSRI 7 and 8 had CS. Only 8 patients without CS had CSRI at the level 5 to 6 and none of them had CSRI 7 or 8.
We also tested different versions of the index with lower numbers of components (e.g., in case of inaccessible NT-proBNP measurement), but these produced significantly lower AUC under ROC curve and accuracy.
This is one of the largest screening studies investigating the incidence of CS using CMRs. CMR with LGE imaging is regarded as a study of choice for diagnosis cardiac involvement in sarcoidosis.[22,27,28] All 49 diagnosed cardiac sarcoidosis patients fulfilled the criteria of active myocardial inflammation owing to sarcoidosis. Choosing this criterion, we wanted to avoid erroneous categorization in the case of fibrosis of undetermined etiology. Our results confirm active heart involvement in one-fourth of the patients, a similar percentage as previous autopsy findings. The incidence is much higher than we found earlier in a symptom-based cross-sectional study (4.7% in the cohort of 1375 sarcoidosis patients). This study confirms that the prevalence of CS is underestimated unless CMRs are done.
The diagnosis of CS remains difficult. Cardiac involvement may occur at any stage of sarcoidosis and significantly affects the outcome of the disease. An endomyocardial biopsy has a poor sensitivity (about 36%), probably because of the patchy nature of sarcoidosis.[12,30] That is why this invasive procedure should not be a part of routine evaluation in patients with suspected CS. The diagnostic criteria were first developed by the Japanese Ministry of Health and Welfare (JMHW) in 1993, modified in 2006. They proposed in the absence of endomyocardial biopsy to identify cardiac involvement by the abnormal function or imaging damages. The development of imaging techniques has enabled the identification of CS changes at an earlier, less advanced stage of the disease.[17,22,28,33–35] In 2014, (later than we started our study) experts of WASOG (World Association for Sarcoidosis and Other Granulomatous Disorders) and HRS (Heart Rhythm Association) proposed new criteria for the diagnosis of CS.[14,15] The diagnostic criteria of CS for our study were in line with these proposals.
Recently developed advanced imaging techniques (CMR, PET) are crucial in the recommended pathway for CS diagnosis. However, they may be difficult to obtain in many settings (e.g., routine screening because of their limited availability and cost).
In this screening study, we analyzed factors which may serve as CS predictors. Symptoms are a factor, but are not specific for CS.[36,37] CS was detected in 39% of our patients with symptoms, indicating that if any symptoms are present, they should not be ignored.[10,21,23,29,38,39] Unfortunately, the absence of cardiac-related symptoms, despite preserved left ventricular systolic function, does not exclude the diagnosis of CS.
Another interesting result of our study was that males had a higher risk for CS diagnosis than females. Some studies found a higher risks for sarcoidosis for women, whereas others did not.[3,11,13] A Case Control Etiologic Study of Sarcoidosis (ACCESS), analyzing the large group (of 736 cases) included only newly diagnosed sarcoidosis patients, in whom CS identified in only 17 (2.1%) with no sex predilection.[11,40] The study of Morimoto et al indicated that despite more frequent general disease in women, the incidence of heart involvement in both sexes was comparable. The result of our study supports our earlier retrospective investigation performed in a large cohort of 1375 sarcoidosis patients, where men had cardiac involvement significantly more frequently than women (OR 2.34).
The role of the ECG in the diagnosis of CS has been discussed for many years. On the one hand, results of this test could be normal, even in advanced heart involvement, (diagnosed histologically, postmortem). On the other hand, there is some evidence that ECG abnormalities often occur before the development of cardiac events.[39,41–43] Moreover, ECG is commonly available, an inexpensive tool for practitioners. In our study, any ECG abnormality was associated with a higher risk of CS. Furthermore, we have found that multiorgan and extrathoracic disease (understood as sarcoidosis detected in organs other than lungs, mediastinal lymph nodes and heart) is also predictor for CS. In the retrospective study of Chapelon-Abric et al, a high rate of neurosarcoidosis was reported. Nagai et al also noted tendency to more organ involvement in their LGE-positive group. We noted difference between CS-negative and positive patients in serum ACE activity (but small and without difference in distribution of abnormal results), whereas features of active disease, expressed as progression observed in chest X-ray pictures, were related with significantly higher risk of CS. So far, there is lack of an objective, randomized study confirming correlation between radiological pulmonary progression and CS detection; however, one study found an association of abnormal ECG and parenchymal lung infiltrates.
Another marker, which also seems to be a useful tool for identifying CS, is the plasma NT-proBNP level. In our study, NT-proBNP level was comparable in the CS-negative and positive groups; however, abnormal results (>125 pg/mL) were more frequently detected in CS-positive group (P = 0.0023)
Other investigations, performed in this study, like Holter ECG monitoring and ECHO could be also useful as predictors of cardiac involvement, with sensitivity 39% and 70% as well as specificity of 85% and 58%, respectively. In other studies of pulmonary sarcoidosis patients, ECHO abnormalities were described in 14% to 56% of cases.[45,46] Holter monitoring investigation is superior to conventional ECG for arrhythmias and conduction disturbances detection. In the study of Mehta et al, the presence of abnormal Holter monitoring findings was the most predictive attribute for CS.
We were looking for a simple tool for practitioners, especially working outside large clinical centers, to help in detecting CS. CSRI seems to be a simple tool without need of expensive diagnostic equipment. It may be used as a clinical algorithm in identifying patients that should be more extensively investigated for the presence of CS. The best accuracy of CSRI was achieved at the level of 5 points, where negative predicted value was 0.84. That means that approximately 5 from 6 patients with the scores <5 had a negative CMR investigation. The positive predicted value of 0.74 indicates that 3 of every 4 patients with the scores ≥5 had CS detected. Low CSRI (values 0–1) is possible only in females (as male sex gives 2 points) and is associated with very low probability of CS (none of our patients with such results has CS detected, all were women, 7 of them symptomatic, 4 with radiological progression, 2 with extrathoracic sarcoidosis, and 1 case with ECG changes). Value 2 given only by male sex was present in 31 cases (4 of them CS-positive). Values between 2 and 4 do not change the post-test probability significantly, so should be regarded as uncertain results and are indication for further investigations. CSRI of 7 and 8 was present only in CS-positive patients in investigated group. In our opinion, patients with such high level of CSRI should be urgently diagnosed for heart involvement. However, CSRI of zero does not fully exclude CS. The opinion (and intuition) of the clinician caring for the patient remains important. We encourage lung disease centers to test and evaluate the proposed index.
4.1 Limitations of the study
As only one person evaluated the CMR imaging study, interobserver and intraobserver agreement was not assessed (although, unclear, unusual images were assessed by another specialist). Second, we diagnosed CS by CMR study; however, it would be ideal if the diagnosis could be confirmed with CMR-guided endomyocardial biopsy. None of our patients had such a procedure. The proposed CSRI was based only on white population investigated in a single referral center and was not validated in other populations.
Active cardiac sarcoidosis was diagnosed in 24.4% of our patients with sarcoidosis, much more than in symptoms-based studies. Male sex, cardiac-related symptoms, ECG changes, serum NT-proBNP level, multiorgan involvement, and radiological pulmonary progression taken in a complex may serve as a component for cardiac sarcoidosis risk index. This index may be a good and cost-effective tool for preliminary assessment when access to specialized equipment is limited and helpful for determining the diagnostic urgency.
Acknowledgments for all colleagues from National TB & Lung Diseases Research Institute, Warsaw, Poland, who provided and cared for study patients (in alphabetical order): Janusz Burakowski MD PhD, Justyna Fijołek MD PhD, Dariusz Gawryluk MD, Jacek Grudny MD PhD, Jarosław Kober MD PhD, Marek Kram MSc, Katarzyna Modrzewska MD, Urszula Nowicka MD PhD, Mateusz Polaczek MD, Elżbieta Radzikowska MD PhD, Janusz Szopiński MD PhD, Anna Śniady MD, Jolanta Winek MD PhD, Jolanta Załęska MD PhD, and for Paul Enright MD from Global Health Institute, The University of Arizona for his help in language editing. MMB had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
1. Statement on sarcoidosis
. Joint Statement of the American Thoracic Society (ATS), the European Respiratory Society (ERS) and the World Association of Sarcoidosis
and Other Granulomatous Disorders (WASOG) adopted by the ATS Board of Directors and by the ERS Executive Committee, February 1999. Am J Respir Crit Care Med
2. Judson MA. The three tiers of screening for sarcoidosis
organ involvement. Respir Med
3. Rybicki BA, Major M, Popovich J Jr, et al. Racial differences in sarcoidosis
incidence: a 5-year study in a health maintenance organization. Am J Epidemiol
4. Rybicki BA, Iannuzzi MC. Epidemiology
: recent advances and future prospects. Semin Respir Crit Care Med
5. Song Z, Marzilli L, Greenlee BM, et al. Mycobacterial catalase-peroxidase is a tissue antigen and target of the adaptive immune response in systemic sarcoidosis
. J Exp Med
6. Tana C, Giamberardino MA, Di GM, et al. Immunopathogenesis of sarcoidosis
and risk of malignancy: a lost truth? Int J Immunopathol Pharmacol
7. Iannuzzi MC. Genetics of sarcoidosis
. Semin Respir Crit Care Med
8. Baughman RP, Lower EE, du Bois RM. Sarcoidosis
9. Silverman KJ, Hutchins GM, Bulkley BH. Cardiac sarcoid: a clinicopathologic study of 84 unselected patients with systemic sarcoidosis
10. Chapelon-Abric C, de ZD, Duhaut P, et al. Cardiac sarcoidosis
: a retrospective study of 41 cases. Medicine (Baltimore)
11. Baughman RP, Teirstein AS, Judson MA, et al. Clinical characteristics of patients in a case control study of sarcoidosis
. Am J Respir Crit Care Med
12. Kandolin R, Lehtonen J, Graner M, et al. Diagnosing isolated cardiac sarcoidosis
. J Intern Med
13. Morimoto T, Azuma A, Abe S, et al. Epidemiology
in Japan. Eur Respir J
14. Judson MA, Costabel U, Drent M, et al. The WASOG Sarcoidosis
Organ Assessment Instrument: An update of a previous clinical tool. Sarcoidosis Vasc Diffuse Lung Dis
15. Birnie DH, Sauer WH, Bogun F, et al. HRS expert consensus statement on the diagnosis and management of arrhythmias associated with cardiac sarcoidosis
. Heart Rhythm
16. Lynch JP III, Hwang J, Bradfield J, et al. Cardiac involvement in sarcoidosis
: evolving concepts in diagnosis and treatment. Semin Respir Crit Care Med
17. Mantini N, Williams B Jr, Stewart J, et al. Cardiac sarcoid: a clinician's review on how to approach the patient with cardiac sarcoid. Clin Cardiol
18. Yazaki Y, Isobe M, Hiroe M, et al. Prognostic determinants of long-term survival in Japanese patients with cardiac sarcoidosis
treated with prednisone. Am J Cardiol
19. Nagai T, Nagano N, Sugano Y, et al. Effect of corticosteroid therapy on long-term clinical outcome and left ventricular function in patients with cardiac sarcoidosis
. Circ J
20. Nagai T, Nagano N, Sugano Y, et al. Effect of discontinuation of prednisolone therapy on risk of cardiac mortality associated with worsening left ventricular dysfunction in cardiac sarcoidosis
. Am J Cardiol
21. Mehta D, Lubitz SA, Frankel Z, et al. Cardiac involvement in patients with sarcoidosis
: diagnostic and prognostic value of outpatient testing. Chest
22. Patel MR, Cawley PJ, Heitner JF, et al. Detection of myocardial damage in patients with sarcoidosis
23. Smedema JP, Snoep G, van Kroonenburgh MP, et al. Cardiac involvement in patients with pulmonary sarcoidosis
assessed at two university medical centers in the Netherlands. Chest
24. Nagai T, Kohsaka S, Okuda S, et al. Incidence and prognostic significance of myocardial late gadolinium enhancement in patients with sarcoidosis
without cardiac manifestation. Chest
25. Matoh F, Satoh H, Shiraki K, et al. The usefulness of delayed enhancement magnetic resonance imaging for diagnosis and evaluation of cardiac function in patients with cardiac sarcoidosis
. J Cardiol
26. Friedrich MG, Sechtem U, Schulz-Menger J, et al. Cardiovascular magnetic resonance in myocarditis: A JACC White Paper. J Am Coll Cardiol
27. Smedema JP, Snoep G, van Kroonenburgh MP, et al. Evaluation of the accuracy of gadolinium-enhanced cardiovascular magnetic resonance in the diagnosis of cardiac sarcoidosis
. J Am Coll Cardiol
28. Cheong BY, Muthupillai R, Nemeth M, et al. The utility of delayed-enhancement magnetic resonance imaging for identifying nonischemic myocardial fibrosis in asymptomatic patients with biopsy-proven systemic sarcoidosis
. Sarcoidosis Vasc Diffuse Lung Dis
29. Martusewicz-Boros MM, Boros PW, Wiatr E, et al. Cardiac sarcoidosis
: is it more common in men? Lung
30. Uemura A, Morimoto S, Hiramitsu S, et al. Histologic diagnostic rate of cardiac sarcoidosis
: evaluation of endomyocardial biopsies. Am Heart J
31. Kim JS, Judson MA, Donnino R, et al. Cardiac sarcoidosis
. Am Heart J
32. Hiraga H, Yuwai K, Hiroe M. Guidelines for the Diagnosis of Cardiac Sarcoidosis
: Study Report of Diffuse Pulmonary Diseases. 1993; Tokyo, Japan: Japanese Ministry of Health and Wellfare, 23–24.
33. Osman F, Foundon A, Leyva P, et al. Early diagnosis of cardiac sarcoidosis
using magnetic resonance imaging. Int J Cardiol
34. Manins V, Habersberger J, Pfluger H, et al. Cardiac magnetic resonance imaging in the evaluation of cardiac sarcoidosis
: an Australian single-centre experience. Intern Med J
35. Patel AR, Klein MR, Chandra S, et al. Myocardial damage in patients with sarcoidosis
and preserved left ventricular systolic function: an observational study. Eur J Heart Fail
36. Zipse MM, Sauer WH. Cardiac sarcoidosis
. Curr Cardiol Rep
37. Yeboah J, Lee C, Sharma OP. Cardiac sarcoidosis
: a review 2011. Curr Opin Pulm Med
38. Adamson P, Melton I, O’Donnell J, et al. Cardiac sarcoidosis
: the Christchurch experience. Intern Med J
39. Darlington P, Gabrielsen A, Sorensson P, et al. Cardiac involvement in Caucasian patients with pulmonary sarcoidosis
. Respir Res
40. Judson MA, Baughman RP, Teirstein AS, et al. Defining organ involvement in sarcoidosis
: the ACCESS proposed instrument. ACCESS Research Group. A Case Control Etiologic Study of Sarcoidosis
. Sarcoidosis Vasc Diffuse Lung Dis
41. Zipse MM, Sauer WH. Electrophysiologic manifestations of cardiac sarcoidosis
. Curr Opin Pulm Med
42. Nagao S, Watanabe H, Sobue Y, et al. Electrocardiographic abnormalities and risk of developing cardiac events in extracardiac sarcoidosis
. Int J Cardiol
43. Roukoz H, Shah M, Masilamani LJ, et al. fQRS as a marker of granulomatous disease in patients presenting with ventricular tachycardia and normal left ventricular ejection fraction. Indian Heart J
44. Handa T, Nagai S, Ueda S, et al. Significance of plasma NT-proBNP levels as a biomarker in the assessment of cardiac involvement and pulmonary hypertension in patients with sarcoidosis
. Sarcoidosis Vasc Diffuse Lung Dis
45. Fahy GJ, Marwick T, McCreery CJ, et al. Doppler echocardiographic detection of left ventricular diastolic dysfunction in patients with pulmonary sarcoidosis
46. Nunes H, Freynet O, Naggara N, et al. Cardiac sarcoidosis
. Semin Respir Crit Care Med
cardiac sarcoidosis; epidemiology; heart involvement; prevalence; risk factors; risk index; sarcoidosis
Supplemental Digital Content
Copyright © 2016 The Authors. Published by Wolters Kluwer Health, Inc. All rights reserved.