Congestive Endomyocardial biopsy plays a role in diagnosing patients with unexplained Hossein Ardehali, MD, PhD,a Atif Qasim, MD,b Thomas Cappola, MD,f David Howard,c Ralph Hruban, MD,d Joshua M. Hare, MD,a Kenneth L. Baughman, MD, FACC,a,e and Edward K. Kasper, MD, FACCa Baltimore, Md, Boston, Mass, and Philadelphia, Pa

Background The etiology of cardiomyopathy is usually inferred from clinical information and preliminary labora- tory studies. Patients with unexplained cardiomyopathy may be referred for endomyocardial biopsy (EMBx). It is unknown whether pathological information obtained from EMBx is beneficial or alters the diagnosis established clinically. This study was undertaken to evaluate the utility of EMBx in confirming or excluding a clinically suspected diagnosis. Methods We evaluated 845 patients with initially unexplained cardiomyopathy who underwent EMBx between 1982 and 1997 at The Johns Hopkins Hospital. For each patient, an initial clinical diagnosis, an EMBx diagnosis, and a final diagnosis prior to discharge based on all available data were established. Results The final diagnosis differed from the initial clinical diagnosis in 264 (31%) of these patients; EMBx made the diagnosis in 196 (75%) of these cases. Initial diagnoses most frequently altered were (34%) and idiopathic cardiomyopathy (25%). Initial diagnoses least likely to be altered were those in which biopsy was used to confirm or grade a previously documented illness, such as hemochromatosis (11%), (18%), or cardiomyopathy second- ary to doxorubicin toxicity (0%). EMBx was more sensitive than clinical diagnosis in detecting myocarditis and amyloi- dosis, and proved to be very specific in detecting ischemic cardiomyopathy, myocarditis, amyloidosis, and hemochroma- tosis. Conclusions In patients with unexplained cardiomyopathy after a standard evaluation, the clinical assessment of the etiology is inaccurate in 31% of patients. EMBx establishes the final diagnosis in 75% of these patients with a high degree of specificity. (Am Heart J 2004;147:919–23.)

and mortality,1 and accounts for a large proportion of See related Editorial on page 759. health care expenditures in developed countries.2 Cardiomyopathy is a common disorder resulting In the developed world, the most common cause of from a variety of causes that lead to impaired cardiac cardiomyopathy is coronary disease; however, performance and often results in congestive heart fail- many other causes have been characterized. It is im- ure. Despite recent advances, congestive heart failure portant to identify the cause of cardiomyopathy in or- remains a major cause of morbidity, hospitalizations der to devise an appropriate treatment strategy. The current guidelines recommend that patients with car- diomyopathy undergo a 12-lead EKG, a chest radio- graph, blood chemistry including thyroid and liver

From the aDivision of , Department of , The Johns Hopkins Hospi- function test, urine analysis, and cardiac imaging, usu- tal, bJohns Hopkins University School of Medicine, cDepartment of Epidemiology, Johns ally a transthoracic echocardiography.3 Patients with d Hopkins Bloomberg School of , Department of , The Johns risk factors for may be subjected to Hopkins Hospital, Baltimore, Md, eDivision of Cardiology, Department of Medicine, Brigham and Women’s Hospital, Boston, Mass, and fDivision of Cardiovascular Medi- coronary angiography. These studies may help to con- cine, Department of Medicine, Hospital of the University of Pennsylvania, Philadelphia, firm the clinical diagnosis and evaluate selected causes Pa. for the cardiomyopathy. Submitted May 23, 2003; accepted September 9, 2003. Reprint requests: Hossein Ardehali, MD, PhD, Johns Hopkins University, 844 Ross Despite evaluation, there remains a group of patients Building, 720 Rutland Ave, Baltimore, MD 21205. in whom the cause of the cardiomyopathy cannot be E-mail: [email protected] explained. These patients may be referred to a tertiary 0002-8703/$ - see front matter 4 © 2004, Elsevier Inc. All rights reserved. care center for endomyocardial biopsy (EMBx). Al- doi:10.1016/j.ahj.2003.09.020 though this procedure has become a common diagnos- American Heart Journal 920 Ardehali et al May 2004

tic modality in patients with unexplained cardiomyopa- minimum of 5 samples from the right ventricular septum was thy in some centers, it is unknown to what extent the obtained from each patient at the time of biopsy. histologic information obtained from EMBx improves A cardiac pathologist examined all specimens at a mini- diagnostic accuracy. In fact, the recent American Col- mum of four section levels stained with hematoxylin and eo- lege of Cardiology/American Heart Association (ACC/ sin. The Dallas criteria were used to establish a diagnosis of myocarditis.7 In cases of suspected myocarditis, the inflam- AHA) guidelines indicate that EMBx should not be per- matory cells were often further characterized using immuno- formed in the routine evaluation of cardiomyopathy, histochemical labeling for CD3, CD4, CD8, and CD68. The and that it should be performed only when there is “a diagnosis of idiopathic cardiomyopathy was established in strong reason to believe that the results will have a cases with significant myocyte hypertrophy and nuclear en- meaningful effect on subsequent therapeutic deci- largement, and no other explanation for the patient’s heart sions.”3 However, the guidelines also indicate that the failure. To identify the presence of hemochromatosis and overall utility of EMBx in patients with unexplained amyloidosis, Prussian blue and Congo red stains were used, cardiomyopathy is unclear.3,5 respectively. Doxorubicin cardiac toxicity was established In the present study, we evaluated 845 patients who and graded using ultrastructural examination of glutaralde- underwent EMBx at The Johns Hopkins Hospital to hyde fixed samples. Ischemic cardiomyopathy was defined by the presence of focal replacement fibrosis in a patient with determine how often EMBx results changed the initial . Replacement fibrosis is characterized clinical diagnosis. In addition, we assessed the sensitiv- by the presence of discrete foci of myocyte loss associated ity and specificity of EMBx compared with clinically with fibrosis and pigment-laden macrophages. established diagnoses as a means of evaluating their relative effectiveness as diagnostic tools. Diagnosis of the cause of cardiomyopathy All patients were given three diagnoses. The initial clinical Methods diagnosis was based on the clinical data at the time of refer- ral for biopsy, according to previously published defini- Selection of patients tions.6,8 A biopsy diagnosis was established according to the The study group consisted of 1230 patients who under- results of the EMBx. Final diagnosis was based on all available went EMBx for evaluation of unexplained cardiomyopathy at data (ie, clinical, laboratory, and EMBx data and “other inter- The Johns Hopkins Hospital between December 1982 and ventions”) prior to discharge. “Other interventions” refers to December 1997. Most of these patients were referred for bi- diagnostic means other than the ones already mentioned. For opsy after initial evaluation in a ’soffice or in an- example, in the case of cardiomyopathy due to amyloidosis, other hospital. Therefore, some common causes of heart fail- the diagnosis of systemic amyloidosis may have been made ure that are more easily identifiable (ie, ischemic heart based on abdominal fat pad biopsy, rectal biopsy, or biopsy disease, hypertension, etc) are underrepresented in this pa- of the tongue. Patients were assigned to one of the following tient group. Since we currently do not have the longitudinal categories for our analysis: idiopathic cardiomyopathy, myo- follow-up of deaths of patients who underwent EMBx after carditis, cardiomyopathy due to ischemic heart disease, drug- December 1997, they were not included in this study. related, infiltrative myocardial disease (ie, hemochromatosis, Patients whose cardiomyopathy had an identified cause amyloidosis), connective tissue disease, or neoplastic heart were excluded from our analysis. These etiologies included disease. peripartum cardiomyopathy, infection with human immuno- deficiency virus, substance abuse, valvular disease, hyperten- sion, endocrine diseases (thyroid disease and pheochromocy- Statistical analysis toma), neuromuscular disease, and congenital and familial For each etiology of cardiomyopathy, the sensitivity and cardiomyopathy. The remaining 845 patients were included specificity of clinical diagnosis and EMBx were computed in our analysis. This is a retrospective review analysis. using final diagnosis as gold standard. The ␬ statistic was cal- A history and physical examination were performed on all culated for both clinical diagnosis and EMBx to evaluate the of the patients, and selected laboratory studies such as thy- overall agreement between each of these and the final diag- roid function tests and antinuclear antibodies were also ob- nosis. Analyses were performed using SAS statistical software tained. All the subjects also underwent right heart catheter- (version 8, SAS Institute, Cary, NC). ization for hemodynamic measurements at the time of EMBx as described previously.6 Those subjects at high risk for hav- ing coronary artery disease underwent coronary angiography. Results The clinical characteristics of this cohort of patients have Role of EMBx in diagnosing patients with been reported previously.6 This study was approved by the Institutional Review Boards of the Joint Committee on Clini- unexplained cardiomyopathy cal Investigation at The Johns Hopkins Hospital. A specific final diagnosis was identified in 316 (37%) of 845 patients with initially unexplained cardiomyopa- Endomyocardial biopsy thy. Myocarditis was the most common final diagnosis EMBx and right heart catheterization were performed on in this population, followed by ischemic heart disease all patients as previously described.6 Intravenous access was and infiltrative disorders. The remaining 529 (63%) preferentially made through the right internal jugular vein. A patients who did not have a clear explanation for their American Heart Journal Ardehali et al 921 Volume 147, Number 5

Table I. Final diagnoses in 845 patients with initially unexplained cardiomyopathy and the method of diagnosis

Number of Number of cases cases Percentage diagnosed based diagnosed Number of of cases Total on clinical and based on cases where number laboratory histological diagnosed by diagnosis Type of of evaluation before evaluation of other was made cardiomyopathy cases EMBx EMBx interventions* by EMBx

Idiopathic 529 376 130 23 25 Myocarditis 104 69 35 0 34 Ischemic 81 36 17 28 21 Infiltrative Amyloidosis 33 27 6 0 18 Hemochromatosis 9 8 1 0 11 Sarcoidosis 12 9 2 1 17 Other 28 17 1 10 4 Drug related Adriamycin toxicity 14 14 0 0 0 Other drug 2 1 1 0 50 Connective tissue diseases Scleroderma 9 7 1 1 11 SLE 7 6 0 1 0 Polyarteritis nodosum 2 2 0 0 0 Dermato/polymyositis 1 1 0 0 0 Wegener’s 1 0 0 1 0 Other 8 4 1 3 13 Neoplastic heart disease 5 4 1 0 20 Total 845 581 196 68 11

*Please refer to the text for the definition of “other interventions.”

cardiomyopathy were classified as having idiopathic myopathy related to myocarditis, EMBx was more sen- cardiomyopathy (Table I, second column). sitive than clinical diagnosis (100% vs 66%, P Ͻ .001) In 264 (31%) of the 845 patients, the final diagnosis but was equally specific (86 vs 87%, respectively). For differed from the initial diagnosis that had been made ischemic cardiomyopathy, however, the clinical diag- prior to EMBx. In 196 (75%) of these 264 cases, EMBx nosis was slightly more sensitive than EMBx (44% vs led to the change in diagnosis, whereas other diagnos- 35%), but this difference was not statistically signifi- tic interventions changed the diagnosis in the remain- cant. For all of the common infiltrative disorders, ing 68 (25%) cases (Table I, columns 4 and 5). Initial EMBx was 100% specific, while clinical diagnosis diagnoses most likely to be altered by EMBx were showed lower specificity in most cases (range 94%– myocarditis (34%), idiopathic cardiomyopathy (25%), 100%). For amyloidosis, EMBx was also more sensitive and ischemic cardiomyopathy (21%). than clinical diagnosis in detection of disease (100% vs With regard to infiltrative disorders, EMBx estab- 81%). lished the diagnosis of cardiac amyloidosis in 18% of cases and confirmed this diagnosis in all other clini- Overall agreement of clinical diagnosis and EMBx cally suspected cases. EMBx also made the diagnosis in For clinical diagnosis, the ␬ value for overall agree- 1 case of hemochromatosis (11%) and 2 cases of sar- ment with the final diagnosis is 0.68 (95% CI 0.60– coidosis (16%). In addition to infiltrative disorders, 0.67), while the ␬ value for the overall agreement be- other diagnoses less likely to be altered were cardio- tween endomyocardial biopsy and final diagnosis is myopathies secondary to connective tissue disorders 0.6003 (95% CI 0.55–0.65). These values are signifi- (10% for all cases) and doxorubicin toxicity (0%). cantly different from 0 (chance agreement) with P Ͻ .001. Sensitivity and specificity of EMBx in diagnosing patients with unexplained cardiomyopathy Table II provides the sensitivity and specificity of Discussion both clinical diagnosis and biopsy diagnosis for various Assessment of the cause of cardiomyopathy can usu- forms of cardiomyopathy. For the diagnosis of cardio- ally be made based on the clinical presentation, his- American Heart Journal 922 Ardehali et al May 2004

Table II. Sensitivity and specificity of clinical diagnosis and endomyocardial biopsy in patients with unexplained cardiomyopathy

Sensitivity Specificity

Type of Clinical Endomyocardial Clinical Endomyocardial cardiomyopathy diagnosis biopsy diagnosis biopsy

Idiopathic 0.711 0.834 0.839 0.265 Myocarditis 0.663 1 0.872 0.857 Ischemic 0.444 0.346 0.992 0.986 Infiltrative Amyloidosis 0.818 1 0.946 1 Hemochromatosis 0.889 1 0.998 1 Sarcoidosis 0.750 0.750 0.982 1 Restrictive Other 0.607 0.036 0.980 1 Drug related Adriamycin toxicity 1 0.857 0.993 1 Other 0.500 0.500 1 1 Connective tissue disease Scleroderma 0.778 0.222 1 1 SLE 0.857 0 0.995 1 Dermatopolymyositis 1011 Polyarteritis nodosum 1 0.500 1 1 Wegners 0011 Other 0.625 0.125 0.992 1 Neoplastic heart disease 0.800 0.400 0.999 1

tory, physical examination, and selected laboratory sis in many of these cases. Only in ischemic cardiomy- studies. However, there are certain patients whose opathy did other methods besides biopsy prove to be diagnoses remain unclear despite extensive initial eval- more useful in diagnosing this condition. This is ex- uations. These patients may be referred to a tertiary pected as the histologic diagnosis of ischemic cardio- care center to undergo EMBx. Since its introduction myopathy relies on evidence of significant myocardial into clinical practice in 1963,9 EMBx has gradually be- injury and is dependent on sampling the area of injury. come an accepted diagnostic intervention. It is mainly EMBx is particularly useful under the following circum- utilized for monitoring myocardial rejection following stances: (1) diagnosing patients with infiltrative disor- , but it is also available for the ders (in whom the diagnosis may be suspected and evaluation of of uncertain etiology.10 when other organ system biopsies are negative), (2) However, it remains unclear what role EMBx plays in confirmation of doxorubicin toxicity when other the workup of patients with unexplained cardiomyopa- causes may also be responsible, and (3) diagnosing thy. In fact, some believe that because of the associ- idiopathic cardiomyopathy when other mores specific ated risks of major complications (such as ventricular etiologies are suspected. EMBX also has high utility in wall perforation and ) and its appar- patients with suspected myocarditis and a clinical ent poor diagnostic accuracy (possibly from high sam- syndrome compatible with giant cell or fulminant pling error), EMBx may not be an appropriate diagnos- myocarditis, when patients fail to respond to standard tic strategy in these patients.11 The recent ACC/AHA management for heart failure and a course of immuno- guidelines do not recommend EMBx in the routine suppressive or immunoabsorption is consid- evaluation of patients with cardiomyopathy. However, ered, or when there is the suggestion of chronic active the guidelines also indicate that the overall usefulness myocarditis and a restrictive as opposed to dilated of this procedure in patients with unexplained cardio- anatomy. Accurate diagnosis of myocarditis may have myopathy is unclear. therapeutic implications since recent reports suggest a In this report, we analyzed 845 patients with unex- long-term benefit of immunosuppressive therapy in plained cardiomyopathy who underwent EMBx to eval- patients with and HLA upregu- uate how often this procedure changed the clinically lation versus poor response when viral persistence established diagnoses in such patients. Our findings was present.12,13 suggest that the initial clinical assessment of the etiol- It should be noted that we excluded those with an ogy of new-onset cardiomyopathy is inaccurate in established diagnosis. This study applies only to that about one third of patients who have unexplained car- population that had no diagnosis made by history or diomyopathy. EMBx may establish the correct diagno- preliminary studies. The applicability of our results to American Heart Journal Ardehali et al 923 Volume 147, Number 5

the general population of patients with cardiomyopa- ment. Therefore, determination of an accurate diagno- thy is also limited by the referral nature of our pa- sis has major clinical implications for patients with tients. Furthermore, in view of our experience with new-onset cardiomyopathy. We conclude that EMBx the technique, patients at our institution may be more can play a major role in establishing an accurate diag- likely to undergo EMBx than the general population nosis in patients with unexplained cardiomyopathy. As with cardiomyopathy. The final diagnoses in our stud- with all diagnostic tools, the prior probabilities of the ies were heavily dependent on the histologic diag- diagnoses in question must be considered by the treat- noses, which could explain the high sensitivity and ing physician to determine whether EMBx is worth the specificity values obtained in our analysis. The absence additional risk to an individual patient. of imaging data (ie, magnetic resonance imaging, car- diac tomography, and echocardiography) on our pa- tient population is another limitation of our study. References The initial diagnosis was made based on available 1. Schocken DD, Arrieta MI, Leaverton PE, et al. Prevalence and mor- clinical data. In some instances, such as myocarditis, tality rate of congestive heart failure in the United States. J Am the clinical diagnosis was “clinical suspicion” of an Coll Cardiol 1992;20:301–6. 2. Cowie MR, Zaphiriou A. Management of chronic heart failure. etiology, since the accurate diagnosis could be made BMJ 2002;325:422–5. only by EMBx. However, in certain cases such as amy- 3. Hunt SA, Baker DW, Chin MH, et al. ACC/AHA guidelines for the loidosis, histologic data from other tissues were avail- evaluation and management of chronic heart failure in the adult: able prior to EMBx. By allowing a tissue specific-diag- executive summary. A report of the American College of Cardiology/ nosis to be included in the clinical diagnosis, it is American Heart Association task force on practice guidelines possible that a bias against the value of EMBx was in- (committee to revise the 1995 guidelines for the evaluation and troduced in our studies. management of heart failure). J Am Coll Cardiol 2001;38:2101– Of the 196 patients whose diagnoses were changed 13. by EMBx, 130 had a final diagnosis of idiopathic 4. Felker GM, Thompson RE, Hare JM, et al. Underlying causes and cardiomyopathy and 17 had a diagnosis of ischemic long-term survival in patients with initially unexplained cardiomy- cardiomyopathy. One may argue that these are not opathy. N Engl J Med 2000;342:1077–84. 5. Lie JT. Myocarditis and endomyocardial biopsy in unexplained “biopsy success” cases because idiopathic cardiomyop- heart failure: a diagnosis in search of a disease. Ann Intern Med athy is a diagnosis of exclusion and should be 1988;109:525–8. diagnosed with coronary angiography. Idiopathic car- 6. Felker GM, Hu W, Hare JM, et al. The spectrum of dilated cardio- diomyopathy is not entirely a diagnosis of exclusion myopathy: the Johns Hopkins experience with 1,278 patients. and histological changes of cardiomyopathy with hy- Medicine (Baltimore) 1999;78:270–83. pertrophy and fibrosis can be noted on biopsy. Addi- 7. Aretz HT, Billingham ME, Edwards WD, et al. Myocarditis: a his- tionally, the patients with idiopathic cardiomyopathy topathologic definition and classification. Am J Cardiovasc Pathol often had different prebiopsy diagnoses, and tissue 1987;1:3–14. analysis assisted in ruling out the illness that they did 8. Kasper EK, Agema WR, Hutchins GM, et al. The causes of dilated not have. Furthermore, ischemic cardiomyopathy was cardiomyopathy: a clinicopathologic review of 673 consecutive patients. J Am Coll Cardiol 1994;23:586–90. diagnosed initially by biopsy in patients who were not 9. Sekiguchi M, Konno S. Diagnosis and classification of primary good candidates for coronary angiography (due to myocardial disease with the aid of endomyocardial biopsy. Jpn their unstable clinical situation), had renal failure (thus Circ J 1971;35:737–54. making coronary angiography risky), or were initially 10. Veinot JP. Diagnostic endomyocardial biopsy pathology: second- diagnosed with myocarditis because of the presence of ary myocardial diseases and other clinical indications—a review. fever, rub, or lack of a history of coronary artery dis- Can J Cardiol 2002;18:287–96. ease. 11. Wu LA, Lapeyre AC III, Cooper LT. Current role of endomyocar- A recent report from our institution demonstrates dial biopsy in the management of dilated cardiomyopathy and that the prognosis of patients with cardiomyopathy myocarditis. Mayo Clin Proc 2001;76:1030–8. varies substantially depending on the underlying diag- 12. Wojnicz R, Nowalany-Kozielska E, Wojciechowska C, et al. Ran- nosis.4 In addition, specific treatment options are avail- domized, placebo-controlled study for immunosuppressive treat- ment of inflammatory dilated cardiomyopathy: two-year follow-up able for certain cardiomyopathies, such as myocarditis results. Circulation 2001;104:39–45. (fulminant and giant cell), valvular and ischemic heart 13. Frustaci A, Chimenti C, Calabrese F, et al. Immunosuppressive disease, heart dysfunction due to endocrinopathies, therapy for active lymphocytic myocarditis: virological and immu- connective tissue disease, drugs, hypertension, and nologic profile of responders versus nonresponders. Circulation infection or infiltrative disorders with cardiac involve- 2003;107:857–63.