A 40 year-old patient was complaining of continued shortness of breath

The patient is a 58-year-old African American man with past medical history significant for a positive purified protein derivative (PPD) 2 years ago. No treatment was given. He was in his usual state of good health until 6 weeks before admission, when he became aware of breathlessness on mild exertion. Over the ensuing 4 weeks, he noted progressive fatigue; swelling of his lower extremities, particularly on standing; and increasing abdominal girth. His exercise tolerance decreased until he could no longer walk a single city block comfortably. He denied chest pain, palpitations, orthopnea, paroxysmal nocturnal dyspnea, and nocturia. He also denied fever and significant weight loss but did complain of occasional night sweats. The patient was treated with furosemide 20 mg/d without improvement.

His family history was remarkable for a father and a brother who died of unknown heart disease in their 50s. He was married with one healthy son. He immigrated to the United States from St Croix 10 years before admission and worked as a mailroom clerk. He had no significant travel history, was a lifetime nonsmoker, and drank alcohol only socially.

Physical examination revealed an obese African American man in mild respiratory distress. He was afebrile. The heart rate was 80 bpm and regular, and his blood pressure was 120/70 mm Hg. Respiratory rate was 20 breaths per minute. His head was normal except for ill-defined hyperpigmentation periorbitally and bitemporally. His neck was supple without lymphadenopathy. Carotid upstrokes were normal. Jugular venous pressure was estimated at 14 cm water, and there was a Kussmaul's sign; prominent x and y descents were noted. The thyroid gland was normal. There were no gynecomastia and no palpable lymph nodes anywhere. The lungs were clear. Cardiac examination showed a nondisplaced point of maximal intensity and normal S1 and S2. There was a soft S3 gallop and a soft early systolic murmur heard at the lower right sternal border. There was no rub. The abdomen was distended with shifting dullness. The liver span was percussed 18 cm in the midclavicular line, and a spleen tip was palpable. There was pitting edema extending bilaterally from the ankles to midthighs. Genitourinary and rectal examination were unremarkable. The neurological examination was significant for mildly decreased sensation to light touch and proprioception in the lower extremities; the motor examination and reflexes were normal.

The ECG showed normal sinus rhythm, borderline-low QRS voltage, and QS complexes in V1 through V3, suggestive of anteroseptal myocardial infarction. Chest roentgenogram showed mild cardiomegaly and no evidence of effusions or focal infiltrates. Admission laboratory data revealed a normal blood count, blood electrolytes, glucose, blood urea nitrogen, and prothrombin time. Liver function tests showed normal transaminases, alkaline phosphatase of 450 U/L, total bilirubin of 1.7 mg/dL (direct, 0.5 mg/dL), total protein of 6.5 g/dL, albumin of 4 g/dL, and lactic dehydrogenase of 457 U/mL. Ferritin was 81 ng/mL; amylase, 63 U/L; and erythrocyte sedimentation rate, 3 mm/h. Urinalysis showed 3+ protein and no casts or cells.

The patient was placed on a low-sodium diet and given intravenous furosemide. With a net diuresis of 1 L, his peripheral edema decreased by the next morning. A diagnostic paracentesis revealed 3.3 g/dL protein, 88 mg/dL glucose, 109 U/mL lactic dehydrogenase, 63 U/L amylase, and 23 nucleated cells per milliliter (mostly lymphocytes).

A transthoracic echocardiogram (Fig 1) revealed thickened left ventricular (LV) and right ventricular (RV) walls with preserved systolic function, mild left and right atrial dilatation, and moderate tricuspid and mitral regurgitation. A small pericardial effusion was present, and the pericardial reflections were described as echo dense. Mitral inflow pattern consisted of a rapid filling phase with short deceleration time and a diminished A wave (late filling wave corresponding to atrial systole). The pattern of ultrasonic reflection from the myocardium was fine and diffuse without discrete “speckling” or highly refractile pattern.

Cardiac catheterization revealed a pulmonary artery pressure of 60/20 mm Hg, RV pressure of 60/22 mm Hg, LV pressure of 123/24 mm Hg, and mean right atrial pressure of 23 mm Hg, with a tracing showing a prominent y descent. The diastolic pressure tracings in both ventricles exhibited a characteristic square root (dip and plateau) pattern. Left ventriculogram showed an increased ejection fraction and mild mitral regurgitation. Cardiac output was 3.2 L/min, with an index of 1.7 L·min−1·m−2. The coronary arteries were normal.

Serum protein electrophoresis was normal, and urine was negative for Bence Jones protein. A 24-hour urine sample revealed a protein of 141 mg/dL. The patient continued to be treated conservatively with diuretics, with marked weight loss and decreased dyspnea. A diagnostic procedure was performed.

Clinical Discussion (I. Kronzon)

This patient had ankle edema, ascites, and marked neck vein distention. These findings are characteristic of failure of the right side of the heart. In addition, there were symptoms of marked shortness of breath on exertion, which also suggests failure of the left side of the heart. The LV systolic function, which was evaluated by echocardiography and left ventriculography, was increased, with evidence of a hyperkinetic left ventricle. RV systolic function on echocardiography was normal. There were no physical, echocardiographic, or angiographic findings to suggest severe mitral or tricuspid regurgitation. I believe that the differential diagnosis is that of diastolic dysfunction. Indeed, the cardiac catheterization showed significantly elevated diastolic pressure in both ventricles, with equalization of all diastolic pressures. There also was a characteristic square root appearance of the diastolic ventricular pressures and a steep y descent noted on the right atrial pressure curve. Although LV diastolic dysfunction leading to symptoms of failure of the left side of the heart is quite common, this syndrome is frequently associated with hypertension or ischemic heart disease and usually is not associated with failure of the right side of the heart. In this patient, it appears that diastolic dysfunction affects both chambers equally.

Two pathophysiological mechanisms can lead to this clinical picture: restrictive cardiomyopathy and constrictive pericarditis. Differentiation between these two conditions is difficult. In many patients, a comprehensive workup, including noninvasive and invasive techniques, does not solve the clinical puzzle, and the diagnosis finally is made at surgery or sometimes at autopsy. Several clinical hints and a few findings, however, should help differentiate between the two. Table 1 summarizes these findings.

The hallmark of both syndromes is significant, frequently severe failure of the right side of the heart. Jugular venous distention, ankle edema, ascites, hepatomegaly, and elevation of liver enzymes as a result of liver engorgement may be associated with these disorders. In both disorders, most of the ventricular filling occurs in early diastole, with a rapid decrease in right atrial pressure immediately after the opening of the tricuspid valve; therefore, a prominent y descent is noted in the jugular veins. Kussmaul's sign, a paradoxical increase in jugular venous distention during inspiration, also can occur in both conditions. Cardiac examination may be useful. In constrictive pericarditis, the heart (which is encased within a thickened pericardium) is quite quiet. Therefore, the point of maximal intensity frequently is not palpable. By contrast, the point of maximal intensity is usually well detected in patients with restrictive cardiomyopathy. Although S1 and S2 are normal in both conditions, S3 does not occur in constrictive pericarditis. What one can hear is a higher-pitched pericardial knock, which occurs at the end of rapid ventricular filling. A lower pitched S3 is more likely to be present in restrictive cardiomyopathy. However, because the timings for S3 and pericardial knock are similar (approximately 180 ms after S2), the interpretation of the auscultatory findings may be misleading, and one may be mistaken for the other.

Chest roentgenogram and fluoroscopy may be helpful. Approximately 50% of patients with constrictive pericarditis have pericardial calcification. Dramatic cases of encasement of the heart within a calcified “eggshell” are sometimes seen. However, the absence of calcification does not rule out constrictive pericarditis, and pericardial calcification does not necessarily indicate constrictive physiology. Newer tomographic technologies such as CT scanning and magnetic resonance imaging of the heart may help in the differential diagnosis. Pericardial thickening can be detected in most patients with constrictive pericarditis, and it characteristically is absent in restrictive cardiomyopathy. In our patient, however, CT and magnetic resonance imaging of the heart were not performed.

Cardiac catheterization clearly demonstrates the characteristic patterns of diastolic dysfunction. In constrictive pericarditis, the whole heart characteristically is encased within a thickened, noncompliant pericardium that limits its diastolic filling. Filling occurs early in diastole and stops when the nondistensible pericardium is stretched to its limit. At this point, all diastolic pressures are high and equal. Equalization of elevated diastolic pressures is therefore the hallmark of pericardial constriction. In restrictive cardiomyopathy, cardiac catheterization also reveals signs of rapid ventricular filling that reaches a high diastolic pressure plateau (square root sign) as seen in constrictive pericarditis. However, the extent of the process may differ in different chambers; therefore, although both RV and LV diastolic pressures may be elevated, they frequently are not identical. Usually, the LV diastolic pressure is higher than the RV diastolic pressure. Unfortunately, in ≈30% of patients with restrictive cardiomyopathy studied by cardiac catheterization, the diastolic pressures in the left and right chambers are nearly equal. To evaluate these patients, changing of preload or afterload may be useful. Maneuvers such as infusion of 500 cm3 normal saline or exercise will separate right- and left-sided diastolic pressures in patients with restrictive cardiomyopathy. In contrast, the diastolic pressures will increase but remain equal in patients with constrictive pericarditis. We were not told whether such maneuvers were performed in the patient under discussion.

Echocardiography and, in particular, Doppler echocardiographic studies also can be useful in the differential diagnosis. Normal LV function or hyperkinesis is the rule in all cases of constrictive pericarditis and can be present in some cases of advanced restrictive cardiomyopathy in which diastolic dysfunction is dominant. Table 2 compares other characteristic Doppler echocardiographic features of the two disorders. Doppler echocardiography can clearly differentiate between the two disorders. In both disorders (constrictive and restrictive), the flow velocity pattern across an AV valve (best observed by transmitral flow studies) is similar: it stops abruptly after rapid ventricular filling, and thus the deceleration time of transvalvular flow velocity is quite rapid (usually >160 ms). However, the effect of the respiratory cycle on the transvalvular flow differs in the two disorders (see Table 2). An accurate determination of respiratory variation of flow velocity was not performed in this patient.

The patient had a positive PPD that had been noted 2 years earlier and was not treated. One of the most common causes of constrictive pericarditis is tuberculous pericarditis. The patient did complain of occasional severe night sweats, which also are frequently associated with active tuberculosis. Not infrequently, tuberculous pericarditis may be the only manifestation of tuberculosis. In this patient, however, active tuberculosis seems unlikely because of the lack of fever, normal sedimentation rate, and lack of other radiological or laboratory evidence of chronic infection.

The echocardiographic findings of RV and LV wall thickening are quite remarkable. They occur in the presence of relatively low voltage on the ECG. The combination of significant LV wall thickening, low voltage on the ECG, and failure of the right side of the heart is highly suggestive of restrictive cardiomyopathy secondary to an infiltrative disorder.

Hemodynamic studies showed severe diastolic dysfunction with equalization of all diastolic pressures. A hemodynamic challenge with preload change (by saline infusion or exercise) was not described in this patient. However, the pulmonary artery and RV systolic pressures were markedly elevated, to 60 mm Hg. This finding is unusual in patients with constrictive pericarditis, in which pulmonary artery pressures rarely exceed 40 mm Hg. Thus, the catheterization findings also support the diagnosis of restrictive cardiomyopathy.

Table 3 lists the conditions associated with restrictive cardiomyopathy. Some are congenital and present themselves early in life. Others can be ruled out by this patient's history. Hemochromatosis could be an attractive clinical diagnosis in this patient with vague skin discoloration; however, the normal ferritin level rules this diagnosis out. Sarcoidosis is another possibility; however, massive cardiac sarcoidosis usually presents itself with conduction abnormalities, and systolic dysfunction frequently precedes diastolic dysfunction clinically. There is no clinical or laboratory evidence of secondary spread of neoplasm into the myocardium. Thus, we are left with a common pathogenesis for restrictive cardiomyopathy, namely, cardiac amyloidosis.

Characteristically, cardiac amyloidosis can be associated with dramatic ventricular wall thickening that can be demonstrated by echocardiography and other imaging techniques. A characteristic echocardiographic finding of myocardial speckling has been described, but it is found in only 50% of patients with amyloidosis. Thus, the lack of speckling on this patient's echocardiogram does not rule out amyloidosis. Long-standing amyloidosis can lead to LV systolic dysfunction; however, diastolic dysfunction with well-preserved LV wall motion frequently is an early manifestation of the disorder. Another approach to diagnosing amyloid at this stage is radionuclide imaging. Technetium pyrophosphate characteristically is absorbed by myocardial amyloid, and this technique has been used to demonstrate and diagnose cardiac amyloidosis. This test was not done in this patient.

There are different forms of amyloidosis both clinically and biochemically. They include primary amyloidosis, in which there is no additional known disease; amyloidosis associated with multiple myeloma; secondary amyloidosis associated with conditions such as tuberculosis, osteomyelitis, or leprosy; and amyloidosis associated with heredofamilial disorders, which also is frequently associated with peripheral neuropathy and occasionally with familial Mediterranean fever. Amyloidosis associated with old age also has been described. In this patient, secondary amyloidosis is unlikely. This form of amyloidosis infrequently creates symptomatic heart disease. There is also no clinical evidence of a long-standing, active inflammatory process. The diagnosis of multiple myeloma is not supported because of the protein electrophoresis, absence of Bence Jones protein, and normal sedimentation rate. Amyloidosis associated with aging is unlikely in this 58-year-old patient. This entity also is frequently an incidental finding discovered at autopsy that usually does not cause the severe clinical manifestations observed in this patient. We therefore remain with a primary or, alternatively, a heredofamilial form of amyloidosis. The family history of this patient, with two relatives who died at a relatively young age of vaguely described heart disease, and his peripheral neuropathy may support the latter diagnosis. However, I believe that once the heart is affected by the infiltration of amyloid that is severe enough to cause severe congestive heart failure, the prognosis is poor, and from the cardiologist's clinical point of view, there is little difference. Because this case was diagnosed years ago, I assume that the diagnostic procedure performed was a cardiac biopsy. However, I believe that this procedure is not without danger, and amyloid can be diagnosed in most cases by such procedures as gingival biopsy or, better yet, biopsy of the subcutaneous abdominal fat. Cardiac biopsy should be reserved only for cases in which these biopsies are negative.

Pathological Findings (G. Gallo)

The patient had RV biopsy. The endomyocardial biopsy was stained with Congo red, which showed green birefringence of deposits under polarization microscopy, typical of amyloid (Fig 2A). Frozen sections were incubated with a panel of antibodies against IgG, IgA, IgM (heavy-chain specific), κ light chain, λ light chain (light-chain specific), transthyretin (TTR), amyloid A protein, and amyloid P component. Deposits stained for λ light chain and amyloid P component (Fig 2B) but were negative for the other proteins tested. The deposits were present diffusely in vessels and around myocardial cells.

Electron microscopy demonstrated randomly oriented fibrils typical of amyloid in the plasmalemmal sheath and interstitially between muscle cells (Fig 3).

Hematology Commentary and Follow-up (D.R. Jacobson)

This patient came to my attention after the diagnosis of amyloidosis was made, and I cared for this patient in the clinic after his discharge from the hospital. Of the comments made so far, the one statement with which I disagree is the comment that because the prognosis of symptomatic cardiac amyloidosis is poor, it makes little difference what type of amyloid is present. Considerable progress has been made recently in understanding amyloid on the molecular level, and this is translating into better approaches to treatment.

First, I would like to address the nomenclature of amyloid because the terms “primary” and “secondary” amyloid are used here. These terms originated long before amyloid was understood on a molecular level, and now that we can determine the specific protein deposited in most patients with amyloidosis, it is recommended that these terms be abandoned and that amyloid be referred to whenever possible by the chemical classification123 ; unfortunately, the terms “primary” and “secondary” persist in the literature, causing considerable confusion. Historically, secondary amyloidosis referred to the amyloid that accompanied chronic inflammatory processes such as tuberculosis and rheumatoid arthritis. Familial amyloidosis was recognized by the positive family history. All other types of amyloidosis, except that associated with the multiple myeloma, were called primary in the sense of idiopathic; this category included unrecognized inherited forms, secondary amyloidosis without an identified cause, and localized amyloidosis. This classification was based on the assumption that all forms of amyloidosis would consist of a single predominant protein.

We now know that all forms of amyloid consist of a minor glycoprotein, the P (pentagonal) component, which is identical in all types of amyloid, and the major fibrillar component, 15 of which have been identified in human amyloidosis. As a group, the amyloid precursor proteins are small, with molecular weights of 4000 to 25 000 D. Their tertiary structures are characterized by a substantial β-pleated sheet structure, which is thought to play a role in amyloid formation, but the precise mechanisms of fibril formation remain poorly understood. The major protein component defines the type of amyloidosis and determines the pathogenesis of the disease. Classification based on clinical syndromes is now avoided whenever chemical information is available. Thus, in what was called secondary amyloidosis, the amyloid fibrils consist of the amyloid A (AA) protein; these diseases are called AA amyloidosis. The amyloid material consisting of immunoglobulin light chains or light-chain fragments (AL amyloid) originates from a single clone of plasma cells. When this clone has expanded to the extent that the criteria for the diagnosis of multiple myeloma are fulfilled, the disease is called myeloma-associated amyloid; when the clone has a limited proliferative capacity and myeloma criteria are not met, the disease previously was called primary systemic amyloidosis. In such cases, the cause of the patient's disease is essentially the same as monoclonal gammopathy of undetermined significance, with the additional feature that the monoclonal protein happens to be amyloidogenic. From the standpoint of amyloidosis, myeloma-associated amyloidosis and primary systemic amyloidosis are identical processes, and both should instead be referred to as AL amyloid.

Of the various types of amyloid that form deposits in the cardiac ventricles and cause congestive heart failure, the most common are AL and TTR amyloid; cardiac AA occurs less often. TTR is a serum transport protein consisting of four identical subunits of 127 amino acids each. TTR transports thyroxine- and retinol-binding protein and is synthesized primarily in the liver, choroid plexus, and retina. Normal-sequence TTR has a low-grade inherent tendency to form amyloid, and a small amount of TTR amyloid in the cardiac ventricles is found incidentally at autopsy in >25% of people 80 years of age or older.4 This process, usually asymptomatic, has been called senile cardiac amyloidosis. Because other organs may be involved, the alternative name “senile systemic amyloidosis” also is used.

In some patients, the process of TTR amyloid deposition is accelerated, leading to congestive heart failure and/or arrhythmias. Occasionally, patients with large amounts of cardiac TTR amyloid have deposits consisting of normal-sequence TTR56 ; in these patients, the stimulus for accelerated deposition is not known. More commonly, the stimulus for symptomatic TTR amyloidosis is a TTR point mutation, changing the conformation of the molecule and leading to increased deposition. Nearly 50 different amyloidogenic TTR mutations are now known, most of which have been described in single kindreds or ethnic groups.7 The most severely affected organs are typically the heart, peripheral nervous system, eye, and gastrointestinal tract. The disease caused by variant TTR typically is called familial amyloid cardiomyopathy or familial amyloid polyneuropathy.1

One amyloidogenic TTR variant, TTR Ile 122, is carried by 4% of African Americans8 (D.R.J. and colleagues, unpublished data, 1995) and has been found in several patients with severe congestive heart failure.91011 The risk of TTR Ile 122 carriers developing symptomatic amyloidosis is not yet known, but it clearly is greater than for people with the normal TTR sequence. When I first saw this patient, immunohistochemistry had not yet been performed, so we did not know the type of amyloid. At the time, I thought that he may have had TTR Ile 122 amyloidosis because there reportedly was no evidence of a monoclonal protein in the serum or urine. While awaiting immunohistochemistry, we performed genetic studies12 and determined that the patient was indeed heterozygous for TTR Ile 122; however, immunohistochemistry demonstrated that the patient's true diagnosis was AL amyloid, and the genetic studies turned out to be a coincidental finding.

After AL amyloidosis was diagnosed, the patient was evaluated for multiple myeloma, as should all patients with AL amyloid. A bone marrow examination revealed 6% plasma cells, and no lytic lesions were seen on skeletal survey; thus, the patient did not have myeloma. Reportedly, only about half of patients with AL amyloid have a monoclonal immunoglobulin protein detectable in the serum or urine. This is really an issue of the practical limitation of the assays used clinically to detect monoclonal proteins. The amyloid precursor, the monoclonal immunoglobulin molecule, is synthesized by monoclonal plasma cells in the bone marrow and is deposited as amyloid in the heart; thus, this protein must travel through the bloodstream. So if a sensitive-enough assay is used, in theory, all patients with AL amyloid must have a monoclonal serum and/or urine protein (free immunoglobulin light chains are small enough to be filtered by the glomerulus and appear in the urine). When we repeated the urine protein immunoelectrophoresis on a concentrated specimen in a research laboratory, the monoclonal protein was detected, even though this test remained negative as reported by the routine clinical laboratory.

So does it matter if we know whether the patient has TTR or AL cardiac amyloid? A decade ago, perhaps not, but in 1996, yes. Chemotherapy is of value for multiple myeloma, so it would seem logical that the same chemotherapy might be of use for AL amyloid even if diagnostic criteria for myeloma are not fulfilled. For many years, reports suggested that chemotherapy is valuable for treating AL amyloid, and two recently randomized, controlled trials have demonstrated a survival advantage for patients receiving chemotherapy,131415 so this should now be standard therapy for patients with AL amyloid, even in the absence of myeloma.

On the other hand, patients with TTR amyloid will not benefit from chemotherapy. For patients with TTR amyloid, studies to determine whether a TTR variant is present can help guide management. Particularly in younger patients, if a TTR variant is present, liver transplantation can be performed as a means of replacing the source of variant TTR with normal-sequence TTR; patients who receive liver transplantations have gradual resolution of their amyloid and may achieve complete resolution of symptoms.1617 At present, there is no known effective treatment for amyloid consisting of normal-sequence TTR.

This patient's congestive heart failure improved with large doses of diuretics (furosemide and spironolactone), and he improved symptomatically. He also was treated with monthly melphalan and prednisone, and his heart disease appeared to stabilize or even improve. The median survival of patients with symptomatic congestive heart failure resulting from AL amyloidosis is months for patients not receiving or not responding to chemotherapy.18 In a subset of patients, however, chemotherapy leads to resorption of the amyloid, improvement in cardiac function, and longer survival.18 Chemotherapy was discontinued after nearly 2 years because of thrombocytopenia. Shortly thereafter, the patient moved out of town and has been lost to follow-up. The optimal duration of chemotherapy for patients who respond is not known. This patient's clinical stabilization and nearly symptom-free state 2 years after diagnosis clearly demonstrate the value of determining the specific type of amyloid present in each patient and instituting appropriate therapy.

Clinical Diagnosis

The clinical diagnosis was cardiac amyloidosis, probably heredofamilial.

Final Diagnosis

The final diagnosis is λ light-chain amyloidosis of the myocardium.

Figure 1. A, Two-dimensional echocardiogram. Apical view reveals left ventricular (LV) and right ventricular (RV) hypertrophy. B, Transmitral flow velocity (by pulsed Doppler echocardiography) shows short mitral deceleration time (128 ms). LA indicates left atrium; DT, deceleration time.

Figure 2. A, Congo red stain reveals deposits characteristic of amyloid. B, Incubation with anti-λ light-chain antibodies reveals characteristic diffuse amyloid deposits.

Figure 3. Electron microscopy shows randomly oriented fibrils typical of amyloid.

This work was supported in part by an Established Scientist Award, American Heart Association, New York City affiliate (Dr Jacobson).

Footnotes

References

• 1 Jacobson DR, Buxbaum JN. Genetic aspects of amyloidosis. Adv Hum Genet..1991; 20:69-123.MedlineGoogle Scholar
• 2 Husby G. Nomenclature and classification of amyloid and amyloidosis. J Intern Med..1992; 232:511-512.CrossrefMedlineGoogle Scholar
• 3 WHO-IUIS Nomenclature Sub-Committee. Nomenclature of amyloid and amyloidosis. Bull World Health Organ..1993; 71:105-108.MedlineGoogle Scholar
• 4 Smith TJ, Kyle RA, Lie JT. Clinical significance of histopathological patterns of cardiac amyloidosis. Mayo Clin Proc..1984; 59:547-555.CrossrefMedlineGoogle Scholar
• 5 Christmansson L, Betsholtz C, Gustavsson Å, Johansson B, Sletten K, Westermark P. The transthyretin cDNA sequence is normal in transthyretin-derived senile systemic amyloidosis. FEBS Lett..1991; 281:177-180.CrossrefMedlineGoogle Scholar
• 6 Gustavsson A, Jahr H, Tobiassen R, Jacobson DR, Sletten K, Westermark P. Amyloid fibril composition and transthyretin gene structure in senile systemic amyloidosis. Lab Invest..1995; 73:703-708.MedlineGoogle Scholar
• 7 Saraiva MJM. Transthyretin mutations in health and disease. Hum Mutat..1995; 5:191-196.CrossrefMedlineGoogle Scholar
• 8 Jacobson DR, Reveille JD, Buxbaum JN. Frequency and genetic background of the position 122 (Val-Ile) transthyretin gene in the black population. Am J Hum Genet..1991; 49:192-198.MedlineGoogle Scholar
• 9 Jacobson DR, Gorevic PD, Buxbaum JN. A homozygous transthyretin variant associated with senile systemic amyloidosis: evidence for a late-onset disease of genetic etiology. Am J Hum Genet..1990; 47:127-136.MedlineGoogle Scholar
• 10 Nichols WC, Liepnieks JJ, Snyder EL, Benson MD. Senile cardiac amyloidosis associated with homozygosity for a transthyretin variant (ILE-122). J Lab Clin Med..1991; 117:175-180.MedlineGoogle Scholar
• 11 Saraiva MJM, Sherman W, Marboe C, Figueira A, Costa P, De Freitas AF, Gawinowicz MA. Cardiac amyloidosis: report of a patient heterozygous for the transthyretin-isoleucine 122 variant. Scand J Immunol..1990; 32:341-346.CrossrefMedlineGoogle Scholar
• 12 Jacobson DR. A specific test for transthyretin 122 (Val-Ile) based on PCR-primer introduced restriction analysis (PCR-PIRA): confirmation of the gene frequency in blacks. Am J Hum Genet..1992; 50:195-198.MedlineGoogle Scholar
• 13 Kyle RA, Gertz MA, Garton JP, Greipp PR, Witzig TE, Lust JA. Primary systemic amyloidosis (AL): randomized trial of colchicine vs melphalan and prednisone vs melphalan, prednisone, and colchincine. In: Kisilevsky R, Benson MD, Frangione B, Gauldie J, Muckle TJ, Young ID, eds. Amyloid and Amyloidosis 1993. New York, NY: Parthenon Publishing; 1994:648-650.Google Scholar
• 14 Skinner M, Anderson J, Wang M, Simms R, Falk R, Jones LA, Cohen AS. Treatment of patients with primary amyloidosis. In: Kisilevsky R, Benson MD, Frangione B, Gauldie J, Muckle TJ, Young ID, eds. Amyloid and Amyloidosis 1993. New York, NY: Parthenon Publishing; 1994:232-234.Google Scholar
• 15 Gertz MA, Kyle RA. Amyloidosis: prognosis and treatment. Semin Arch Rheum..1994; 24:124-138.CrossrefMedlineGoogle Scholar
• 16 Holmgren G, Ericzon B-G, Groth C-G, Steen L, Suhr O, Andersen O, Wallin BG, Seymour A, Richardson S, Hawkins PN, Pepys MB. Clinical improvement and amyloid regression after liver transplantation in hereditary transthyretin amyloidosis. Lancet..1993; 341:1113-1116.CrossrefMedlineGoogle Scholar
• 17 Skinner M, Lewis WD, Jones LA, Kasirsky J, Kane K, Ju ST, Jenkins R, Falk RH, Simms RW, Cohen AS. Liver transplantation as a treatment for familial amyloidotic polyneuropathy. Ann Intern Med..1994; 120:133-134.CrossrefMedlineGoogle Scholar
• 18 Gertz MA, Kyle RA, Greipp PR. Response rates and survival in primary systemic amyloidosis. Blood..1991; 77:257-262.CrossrefMedlineGoogle Scholar

Page 2

On the basis of early studies, cardiac muscle shortening (ejection) was considered to have a predominantly negative effect on contractile performance compared with isometric (ISOV) contraction.1 This was ascribed to deactivation of shortening and viscous resistance.2 Recently, a more complicated picture has emerged as positive effects of shortening and ejection have been characterized in isolated heart muscle and in the intact ventricle. In isolated muscle preparations, shortening has been associated with a prolonged time course of contraction34 and an increase in the amplitude of the intracellular calcium transient.5 The latter effect is consistent with a length-dependent decrease in myofilament calcium affinity and subsequent displacement of calcium from the contractile proteins.56 The energetic consequences of shortening have varied in muscle preparations, with reports of both a decrease7 and an increase8 in myocardial energy consumption (V̇o2) compared with isometric contraction. In addition, a precise understanding of the interaction between shortening and alterations in the partitioning of energy input between that used for cross-bridge formation and E-C coupling is lacking. It is conceivable that shortening results in differential energetic effects, ie, an increase in energy consumption due to increased cycling of activator calcium and a decrease in energy consumption due to deactivation of cross bridges.

In the intact ventricle, several recent studies have demonstrated a positive effect of ejection on LV pressure generation. Over a range of contraction modes, including volume withdrawal by rapid flow pulses,19 volume ramp ejection,1011 and ejection controlled by an impedance afterload,12 a consistent increase in ES pressure of “ejecting” beats compared with ISOV beats at matched ESV has been observed. This positive effect has been ascribed to a beat-to-beat “memory” of greater length-dependent activation in the ejecting beat due to the larger EDV and average volume compared with an ISOV beat at the same ESV.

In contrast, one previous report1 demonstrated an ejection-related increase in contractile performance that was independent of ED loading. In that study, Yasumura and coworkers used rapid volume withdrawal (8 to 19 EDV/s) beginning at ES in isolated canine hearts and found an increase in Emax of the volume withdrawal contractions compared with ISOV contractions at matched EDV. The time course of this effect was not described. These results cannot be explained by length-dependent activation and suggest some other effect related to shortening. ESV withdrawal contractions were also found to have a significant decrease in total V̇o2 that was proportional to the speed of withdrawal. Because quick release of cardiac muscle during relaxation results in transient increases in intracellular calcium concentration, we hypothesized that the positive effect of rapid VR during relaxation in the intact ventricle is due to an increase in myocyte calcium available for activation. If this hypothesis is correct, such a shortening-related increase in contractility would be expected to be associated with an increase in nonmechanical V̇o2 (V̇o2 for basal metabolism and E-C coupling) because of the increased energetic requirements of calcium handling and a decrease in mechanical V̇o2 (V̇o2 for cross-bridge cycling) caused by cross-bridge deactivation. The effect on total V̇o2 would depend on the balance between these two effects.

In the present study, we characterized the positive effect of rapid volume withdrawal at ES in the isolated, red blood cell-perfused rabbit heart and tested this hypothesis by partitioning total V̇o2 into nonmechanical and mechanical components. To accomplish this, we used a recently described technique using the drug BDM to estimate nonmechanical V̇o2.1314 We found that rapid VR during relaxation resulted in a gradually appearing positive inotropic effect that was associated with substantial increases in nonmechanical V̇o2 and decreases in mechanical V̇o2.

Methods

Heart Preparation

We used an isolated rabbit heart preparation with retrograde perfusion of bovine red blood cells suspended in a KH buffer. The details of this preparation have been reported elsewhere.1415 Adult male, nonfasted New Zealand White rabbits (weight, 2.4 to 3.8 kg) were premedicated with fentanyl (0.044 mg/kg IM) and droperidol (2.2 mg/kg IM), anesthetized with ketamine (20 mg/kg IM) and xylazine (1 mg/kg IM), and ventilated through a tracheostomy with a respirator. After median thoracotomy, the rabbits were anticoagulated with heparin (1000 U/kg IV). The venae cavae and pulmonary veins were ligated, and a perfusion cannula was immediately connected to the aortic root 4 to 5 mm above the aortic valve. The heart was excised from the chest after initiation of retrograde perfusion. Coronary perfusion pressure was kept constant (80 to 85 mm Hg) by a pressurized flow chamber. The temperature of the isolated heart was maintained at 36°C to 37°C.

The LV was vented at the apex, the left atrium was opened, and the chordae tendineae were cut. A thin-walled latex balloon was placed in the LV and secured at the level of the mitral valve ring. LV pressure was measured inside the LV balloon by a high-fidelity micromanometer catheter (Millar Instruments), and LV volume was controlled by a computer-driven linear servomotor. The operating characteristics of the servomotor are described below. Coronary blood flow was collected by continuous drainage of the collapsed RV and was measured by timed collection in a graduated cylinder. Thebesian flow was ignored because it is negligible in rabbit heart.15 Coronary arteriovenous O2 content difference was measured continuously with an AVOX analyzer (AVOX Systems), which was calibrated at intervals throughout each experiment with a Lex-O2-Con O2 content analyzer (Lexington). Heart rate was controlled (160 to 210 bpm) by epicardial pacing from the LV.

Perfusate Preparation

The nonrecirculated perfusate consisted of bovine red blood cells suspended in KH buffer. Fresh, whole cow blood was obtained from a local slaughterhouse and prepared as described by Marshall.16 The details of initial separation and daily washing of the red blood cells have been described elsewhere.14 The red blood cells were used within 4 days of collection. The KH buffer used for washing and final red blood cell suspension had the following composition (in mmol/L): 108.0 NaCl, 4.0 KCl, 2.5 CaCl2, 1.4 KH2PO4, 25.0 NaHCO3, 11.0 dextrose, and 10.0 sodium pyruvate (all from Sigma Chemical Co). Gentamicin (3 mg/L) was added to the buffer to suppress growth of bacteria. Before each experiment, the red blood cells were suspended to a final hematocrit of 35% in a mixture of KH buffer and insulin (10 IU/L). The perfusate was equilibrated in a silicone elastomer tubing oxygenator with 98% O2/2% CO2 to achieve a Po2 of >100 mm Hg and a PCo2 of ≈40 mm Hg; pH was adjusted to 7.35 to 7.45 by addition of NaOH. The temperature of the perfusate was maintained between 35°C and 37°C with water jackets around the oxygenator and the pressurized chamber in the arterial line.

The average time between heart excision and the start of data collection was 20 minutes (range, 15 to 30 minutes). Data collection for either the VR or BDM protocol required an average of 60 minutes (range, 40 to 90 minutes). We have validated the mechanical and energetic stability of this preparation previously.13 To further ascertain stability, the LV developed pressure at maximum LV volume was measured again after data collection. On average, there was a decrement of <1% in developed pressure. Hearts with decreases >5% were excluded.

Servomotor Control System

Contraction mode was varied by use of a volume- and pressure-integrated servomotor directed by a computer control system.17 The servomotor system is composed of a linear motor, a piston-cylinder device (based on a ground-glass syringe), an LVDT, an analog position controller, and a high-current amplifier. The piston barrel is connected to one end of the metal motor coil shaft. The flared end of the cylinder is attached to the LV balloon and placed in the mitral annulus. The piston-cylinder device allows a maximal infusion or withdrawal of 1.5 mL from the initial EDV at a resolution of 0.01 mL. The LVDT (Trans-Tek, Inc) is mounted on the back of the linear motor. Its core is attached to the other end of the motor coil shaft. The LVDT has a frequency response of 1 kHz and resolution of ±0.015 mm. This allows precise measurement of piston position and, with proper calibration, instantaneous LV volume. Piston position is controlled by a classic analog PD controller. A volume command from the computer control system is used as reference input to the PD controller. Output voltage from the controller is sent to a high-current amplifier. An electric current is then delivered to the motor coil to change volume in the LV to match the volume command. Thus, the volume-servo system controls volume in the LV balloon according to the volume command signal from the computer control system. Similarly, a PD compensator controls pressure (by controlling volume) at any level commanded by the computer. With this system, LV pressure or volume can be altered over a wide range of rates at any point in the cardiac cycle.

The computer control system is based on an Intel 486 processor on an IBM-compatible computer and supervises the volume-servo system. An analog/digital convertor samples instantaneous LV pressure and volume (calculated from the LVDT position signal). In addition, a digital-to-analog convertor sends the volume command to the analog PD controller of the volume-servo system and another sends the pressure command to the analog PD controller. One digital input/output line is used to pace the heart under computer control and another is used to switch between analog pressure or volume control during the ejection cycle. Because the computer control system operates in real time with a 1-ms update cycle, the volume-servo system can be made to respond to any pressure or volume command signal almost instantaneously.

BDM Partitioning of V̇o2

We used a recently described method to partition V̇o2 in the intact heart using the negative inotropic drug BDM.1314 In previous whole-heart experiments, mechanical unloading has been used to partition energy consumption (V̇o2) into mechanical (cross-bridge cycling) and nonmechanical (basal metabolism plus E-C coupling) components (reviewed in Reference 18), ie, the mechanically unloaded V̇o2 has been considered to represent nonmechanical V̇o2. However, mechanical unloading cannot unequivocally distinguish mechanical from nonmechanical V̇o2 because the mechanically unloaded LV undergoes shape changes and pressure fluctuations in the negative range with each contraction.15 Therefore, under mechanically unloaded conditions, an uncertain amount of energy continues to be used for cross-bridge cycling. We have demonstrated previously13 that such cross bridge-related V̇o2 is a significant portion of mechanically unloaded V̇o2. Furthermore, use of mechanical unloading to study energy partitioning during ejecting beats obviously alters the loading conditions of interest. To determine the energetic effects of ejection on partitioning of V̇o2, it is mandatory to use a technique that does not require a change in contraction mode.

The BDM method was introduced by Alpert et al19 as a new approach to partitioning energy utilization in isolated rabbit heart muscle. The technique is based on the myofilament selective negative inotropic effect of low-concentration BDM (<6 mmol/L) in rabbit heart muscle.2021 In isometrically contracting rabbit heart muscle at constant length, perfusion with up to 6 mmol/L BDM results in a linear decline in the relation between initial heat and TTI. Extrapolation of this relation to the initial heat-axis intercept (ie, zero TTI) provides an estimate of tension-independent heat or the energy used for E-C coupling. The method has been validated in isolated muscle by two independent tests.19 We subsequently adapted this technique to partition V̇o2 in the red blood cell-perfused, isolated rabbit heart.1314 Similar to the initial heat-TTI relation in isolated rabbit heart muscle, we reported a linear relation between V̇o2 and FTI up to about 5 mmol/L perfusate (BDM) under ISOV contraction conditions with constant volume. By analogy, we consider extrapolation of the V̇o2-FTI relation to the V̇o2 axis to be an estimate of nonmechanical V̇o2, which includes V̇o2 for both E-C coupling and basal metabolism but not cross-bridge cycling. Several additional validations of this approach have been obtained in the isolated rabbit heart. The myofilament selectivity of low-concentration BDM was verified by the demonstration that <10 mmol/L BDM had no effect on the intracellular calcium transient measured by indo 1 fluorescence.13 Nonmechanical V̇o2 estimated by BDM changed appropriately with the administration of isoproterenol, an agent known to increase energy costs for E-C coupling (M.W.W., MD, et al, unpublished data, 1995, and Reference 22). Finally, the BDM estimate of nonmechanical V̇o2, corrected for basal metabolic V̇o2 and normalized for unit mass, was found to be very similar to estimates of E-C coupling energy consumption in isolated muscle. In contrast, a similar correction of the mechanically unloaded V̇o2 yielded values of E-C coupling energy consumption that were approximately twice those obtained in isolated heart muscle. Thus, the BDM estimate of nonmechanical V̇o2 seems much more reasonable than the mechanically unloaded V̇o2.

Experimental Protocol

LV pressure, LV volume, coronary blood flow, and arteriovenous O2 content difference were measured in seven hearts while contraction mode was varied randomly between ISOV and VR beginning at ES and EDV was kept constant. VR was performed at two rates, slow (9.0±1.7 mL/s) and rapid (50.0±8.3 mL/s). VR was begun at ES and continued until LV pressure was reduced to the control value of LVEDP. This resulted in an average EF of 50.2±7.6% and minimum LV pressure of 4.0±4.3 mm Hg during rapid VR and an average EF of 43.9±5.2% and minimum LV pressure of 6.4±4.2 mm Hg during slow VR. After each VR, the withdrawn volume was returned to the LV by no later than a time period equal to 90% of the cardiac cycle length. EDV was set at a value that resulted in an LVEDP of 8 to 12 mm Hg (mean, 9.3±5.5 mm Hg) during initial conditions. EDV was kept constant during VR and ISOV conditions. Mechanical and energetic variables were recorded during steady state, with a delay of at least 2 minutes between a change in contraction mode and data acquisition.

In eight additional hearts, BDM was infused at five or six incremental perfusate concentrations ≤5 mmol/L with an infusion pump (Baxter Health Care) while contraction mode was varied randomly between ISOV and rapid VR at the same EDV. Measurements were made during each contraction mode under control conditions and then after stable conditions were present at each BDM.

In three hearts, basal metabolism was measured after arrest with an infusion of 20% KCl solution as previously described.14 Coronary blood flow and arteriovenous O2 difference were measured >15 minutes after arrest as the LV volume mode was varied between ISOV and rapid VR, analogous to the above contraction modes.

At the end of each protocol, the remnant great vessels and atria were trimmed, the ventricles were separated, and the RV free wall and LV (including septum) were weighed. LV and RV weights were 4.75±0.74 g and 1.46±0.39 g, respectively.

Data Analysis

Mechanical Measurements

LV pressure, LV volume, and arteriovenous O2 difference were recorded on a strip chart and collected with a computer (Gateway 2000) at 5-ms sampling intervals and stored for off-line analysis. Each contraction mode was ISOV until the time of maximum LV pressure, which was considered to be ES. ED was considered to be LV pressure when dP/dt was 10% of maximum positive dP/dt. Maximum dP/dt was calculated from the digitized LV pressure. sFTI was calculated as the integrated total wall force between ED and ES with use of an assumed spherical geometry for the LV.14 Each mechanical data point was determined as the average of complete cardiac cycles over 4 seconds. The LV volume was calculated as the sum of the intraventricular balloon fluid volume plus the small volume of the balloon walls and connector extending into the LV. V0, the mechanically unloaded LV volume at which maximum systolic pressure is zero, was determined during ISOV contractions in each heart. PVA, a measure of the total mechanical energy output of the LV,218 was calculated as the sum of two areas: a triangle bounded by the ES, ED, and V0 points and a smaller area that reflected the curvilinear ED pressure-volume relation. For the purposes of the present study, we assumed the ESPVR was linear1523 and represented by the line that connected the peak pressure-volume coordinate and V0 for each contraction mode. Thus, the first area was calculated as (Pes−Ped)×(V−V0)/2, where Pes and Ped are ES and ED pressures and V is ventricular volume. The second area was approximated by Ped×(V−V0)/4.23 Emax, the slope of the ESPVR, was defined as LV developed pressure/V−V0.

V̇o2 Measurement

Total O2 consumption per minute was calculated as the product of coronary flow (mL/min) and arteriovenous O2 content difference (vol %) and was normalized for heart rate to yield total V̇o2 per beat (mL O2/beat). V̇o2 of the mechanically unloaded RV was considered to be constant and was calculated as the product of mechanically unloaded total V̇o2 multiplied by the ratio of RV weight to biventricular weight.1223 LV V̇o2 was calculated as total V̇o2 minus RV V̇o2 and was normalized for 100 g LV wet weight to give LV O2 consumption (mL O2·beat−1·100 g−1). Mechanical efficiency (%) was calculated as the ratio of PVA to total LV V̇o2 after both parameters were converted to standard energy units (joules).24

Statistics

Data are reported as mean±SD. For mechanical and energetic parameters, we tested for differences between contraction modes using repeated measures ANOVA followed by Student-Newman-Keuls test if the F test was positive. With the BDM method, nonmechanical V̇o2 is the V̇o2 intercept of the V̇o2-sFTI relation obtained under control conditions and as perfusate (BDM) is varied. In individual hearts, the V̇o2 intercept was estimated by linear regression analysis of this relation during both ISOV and VR contraction modes. We have described previously14 the use of a repeated measures ANOVA with multiple linear regression and dummy variables to detect differences in the slope and V̇o2 intercept of regression lines of pooled V̇o2-sFTI data obtained with the BDM method. This approach provides a rigorous test of differences because it accounts for between-subject variability. With this method, all V̇o2-sFTI data were pooled and fitted to the following regression model:

Results

Cardiac Mechanics

Representative recordings from a single heart under steady-state conditions during each contraction mode are shown in Fig 1A. Compared with ISOV contraction, VR at ES is associated with a steady-state increase in peak LV pressure and dP/dt. The positive effect on contractility is greater during rapid VR, with a 15% increase in developed pressure in this example. Note that the EDV is constant between contraction modes. Slow tracings obtained as the contraction mode was switched from ISOV to rapid VR are shown in Fig 1B. The positive effect due to VR had a gradual onset after the abrupt change in contraction mode, with >50 beats required to reach a new steady state.

Table 1 summarizes the group cardiac mechanical variables in seven hearts during control and VR contraction modes. The rate of rapid VR at ES was about five times greater on average than the rate of slow VR (26.8±5.1 versus 5.0±0.9 EDV/s). There was a trend toward a lower LVEDP during all VR that was not statistically significant. Compared with ISOV contractions, rapid VR was associated with a 15% increase in LV developed pressure (92±24 versus 106±28 mm Hg; P<.01), a 17% increase in maximum LV dP/dt (1223±401 versus 1435±505 mm Hg·s−1; P<.01), and a 13% increase in Emax (69±20 versus 79±23 mm Hg/mL; P<.01). Slow VR was associated with a smaller effect on contractile function, a 7% increase in LV developed pressure (P<.01), a 5% increase in Emax (P<.05), and a trend toward an increase in maximum dP/dt.

Cardiac Energetics

Group effects on cardiac energetic parameters due to alteration in contraction mode are summarized in Table 2. VR was associated with a modest increase (<12%) in coronary blood flow at constant coronary perfusion pressure that was not significant at the P<.05 level. Consistent with the increase in LV pressure at the same EDV, PVA increased 6% with slow VR and 14% with rapid VR compared with ISOV contraction (1548±406 versus 1640±461 versus 1765±492 mm Hg·mL−1·beat−1·100 g−1, respectively; each P<.01). Despite the increase in PVA, V̇o2 was unchanged with slow VR and decreased by 8% with rapid VR (0.0744±0.0195 versus 0.0683±0.0141 mL O2·beat−1·100 g−1; P<.05). Thus, rapid VR increased mechanical efficiency compared with ISOV contraction (14.0±2.5 versus 17.2±3.9%; P<.01).

Nonmechanical V̇o2

At constant EDV, there was a linear V̇o2-sFTI relation during infusion of incremental BDM ≤ 5 mmol/L during both contraction modes. The linear regression of V̇o2 on sFTI in individual hearts is summarized in Table 3. V̇o2-sFTI relations were highly linear for all hearts during each contraction mode (ISOV r=.91±.05, rapid VR r=.94±.04), and as depicted in Fig 2, nonmechanical V̇o2 determined from linear regression in individual hearts was uniformly larger for rapid VR than ISOV contractions. With use of the multiple linear regression model for pooled data, the V̇o2 axis intercept of the V̇o2-sFTI relation obtained with the BDM method was 0.0248±0.0021 mL O2·beat−1·100 g−1 for ISOV contraction and 0.0312±0.0022 mL O2·beat−1·100 g−1 for rapid VR contractions (P<.01), which amounted to an increase of 26% in nonmechanical V̇o2. The slope of the V̇o2-sFTI relation was significantly lower for the rapid VR contraction compared with ISOV contraction. The latter must be the case given the decrease in total V̇o2 for VR contraction compared with ISOV contraction during control conditions and the above increase in V̇o2 intercept of the rapid VR contractions. The fit of the group BDM data to the multiple linear regression model is summarized in Table 4. Finally, there was no significant difference in the V̇o2 for basal metabolism between ISOV and VR contraction mode. V̇o2 after KCl arrest was 0.616±0.130 mL O2·min−1·100 g−1 for ISOV LV volume and 0.614±0.122 mL O2·min−1·100 g−1 (P=.98) during rapid VR volume mode at the same EDV.

Discussion

The present study characterizes the mechanoenergetic effects of ejection (VR) at ES in the isolated rabbit ventricle. There were three major findings. First, VR results in a gradually appearing increase in systolic force generation. This effect cannot be ascribed to length-dependent activation, and its magnitude is proportional to the rate of VR. Second, nonmechanical V̇o2 increases during VR; this effect supports a significant alteration in activation due to ejection. Finally, total LV V̇o2 decreases during VR.

Mechanical Effects of Ejection

Several recent reports101225 have characterized positive aspects of ejection on the ES pressure of the intact ventricle. Positive effects have been demonstrated over a range of afterload histories with a variable relationship to EF1225 and a time course that ranges from the first beat1025 to a gradual appearance10 after the onset of ejection. In general, however, previous studies have focused on a comparison of ISOV and “ejecting” beats at matched ESV. Accordingly, the positive effects observed have been ascribed to greater length-dependent activation of ejecting beats compared with ISOV beats at the same ESV. Shortening deactivation has been invoked as an opposing effect, proportional to EF. Thus, the ES pressure of ejecting beats has been considered to reflect a balance between these two opposing influences during ejection.425 In contrast, the current study describes a positive effect of ejection on LV pressure that is independent of length-dependent activation. Such shortening activation may represent a third effect that mediates cardiac contractility during ejection.

Two previous studies described analogous effects. Yasumura and coworkers1 performed rapid LV volume withdrawal at ES and found an average 10% steady-state increase in Emax for ESV withdrawal contractions compared with ISOV beats at the same EDV. We used an identical contraction mode with entirely analogous mechanical results. At an average VR rate (13.5 EDV/s) that was intermediate between our slow and rapid VR rates (5 and 27 EDV/s), Yasumura et al1 reported an intermediate increase in Emax. The time course of the positive effect they observed was not reported. If one assumes that it was gradual, the similarity of these results between species would support a fundamental interaction between VR during relaxation and ventricular contractile function. Sugiura and coworkers10 used a volume ramp to simulate ejection in isolated dog hearts. They reported a first-beat increase in LV pressure for ejecting contractions with matched ESV and small stroke volume and a gradually appearing further increase that was independent of stroke volume. They speculated that the latter, gradually appearing effect was due to a long-term increase in calcium availability, which in turn was related to a length-dependent alteration in calcium handling. Our results show a very similar gradual increase in LV pressure at the same EDV. Thus, the present study suggests that an alternate mechanism specifically related to ejection, rather than a change in EDV, accounts for a gradually appearing increase in LV pressure after the initiation of ejection. The time course of the VR effect and the associated increase in nonmechanical V̇o2, discussed below, support an ejection-related increase in activation.

An alternative mechanism for the contractile effect of VR is an increase in myocardial turgor (Gregg effect) induced by VR at ES. However, this is unlikely for several reasons. The LV was vented such that VR rapidly eliminated afterload but did not create a suction force on the endocardium. Furthermore, VR did not decrease LV chamber compliance, because LVEDP tended to be lower with VR. Although coronary blood flow showed a trend toward a modest (<10%) increase during VR, we have reported previously23 that an adenosine-induced doubling of coronary blood flow in our isolated rabbit heart preparation is required to produce an LV pressure increase comparable to that associated with VR.

Increase in Nonmechanical V̇o2 With Ejection

In the present study, nonmechanical V̇o2 increased by 26% in association with rapid VR. Basal metabolism has been shown to be independent of contraction mode in isolated muscle.19 Our results in the arrested heart confirm that altered basal metabolism cannot explain the observed increase in nonmechanical V̇o2. Accordingly, the increase in nonmechanical V̇o2 can be attributed specifically to an increase in energy utilization associated with E-C coupling. Our results reflect an approximate 50% increase in V̇o2 for activation on the basis of prior reports14 that E-C coupling accounts for roughly 50% of total nonmechanical V̇o2 at moderate loading conditions. Because calcium reuptake by the SR calcium ATPase is the predominant energetic cost associated with activation, these results strongly support a link between muscle shortening and calcium cycled per beat. As discussed below, studies that used calcium transients56 suggested a similar link, but their methodology does not provide information regarding the total amount of calcium cycled.

As mentioned previously, one possible mechanism underlying an increase in calcium cycled per beat is that VR displaces calcium from troponin C, resulting in an increase in free calcium, and that the latter effect in some way has a positive influence on calcium reuptake and subsequent activation.4 There is evidence from cardiac muscle preparations to support such an interaction. Rapid shortening of isometrically contracting cardiac muscle, particularly during relaxation,6 results in a transient increase (spike) in the intracellular calcium signal.26 Myofilament calcium displacement from troponin C is thought to be the basis of the calcium signal increase.27 A mechanism to link this calcium displacement to a subsequent increase in contractile force has not been established, but we and others4 have speculated that the effect may be mediated by the SR. For example, a net shift in the time course of SR calcium uptake due to abrupt calcium displacement after ES could influence the kinetics of transfer of calcium from uptake to release pools. A recent study by DeTombe and Little4 supports this possibility; their study demonstrated altered inotropic effects of ejection in rat cardiac muscle strips when Sr2+, which is not handled by the SR, was substituted for calcium. Their results were consistent with a shortening-related myofilament displacement of calcium followed by sequestration by the SR. Alternatively, the source of the additional calcium cycled could be external and could be mediated via sarcolemmal calcium transport. For example, mechanosensitive sarcolemmal ion channels could be the source as a result of greater cyclic stretch and deformation that occurs with ejecting beats. Although these channels transport mainly monovalent cations, increased intracellular Na+ could in turn result in increased Ca2+ via the Na-Ca exchanger.2829 Furthermore, a mechanosensitive sarcolemmal calcium channel has been reported.30

The precise alteration in calcium cycling that underlies these results is beyond the scope of the current study. However, the present results demonstrate a link between ventricular ejection, increased activation, and a subsequent increase in contractile performance. Although the contraction mode used was chosen to isolate and maximize such an interaction, it is possible that this effect participates in the regulation of cardiac contractility during normal ejection.

Influence of Ejection on Total V̇o2

The final major finding of the present study is the decrease in total V̇o2 associated with VR. An important aspect of this result is the dependence of V̇o2 reduction on the rate of VR. We found no change in V̇o2 during VR at an average of 5 EDV/s and an 8% decrease in V̇o2 when the VR rate was 27 EDV/s. A similar result was reported by Yasumura et al1 in the dog ventricle. Thus, this effect on energy utilization appears to be dependent on the rate of VR and presumably muscle shortening. Previous work7 in cardiac muscle demonstrated a large decrease in V̇o2, with quick releases (2 to 4 muscle lengths/s) after peak isometric tension, which supports the concept that during isometric contraction, maintenance of tension after ES is an energy-consuming process, ie, cross-bridge formation and associated splitting of ATP continue during relaxation. Our results are consistent with an analogous time course of cross-bridge energy utilization during ISOV contraction in the beating heart. In addition, our results support the occurrence of a considerable magnitude of cross-bridge-related energy utilization after ES in ISOV contractions. When the increase in nonmechanical V̇o2 during rapid VR is combined with the decrease in total V̇o2, mechanical V̇o2 decreased by ≈25% compared with ISOV contraction. The magnitude of this energy saving is comparable to that observed during the aforementioned quick-release experiments. Although the precise mechanism responsible for the shortening-related decrease in total V̇o2 remains unclear, our results once again clearly indicate that it is a decrease in the mechanical component that is responsible.

In summary, our results support a novel effect of a positive interaction between myofilament shortening and activator calcium cycling. The physiological significance of interactions between shortening and calcium cycling remains to be determined. Additional experiments that use more normal ejection will be required to accomplish this. Furthermore, elucidation of mechanisms that underlie shortening-calcium cycling interactions may provide insight into elements of the pathophysiology of heart failure. Recently,31 isolated cardiac muscle obtained from myopathic human hearts was shown to display abnormalities in the calcium transient that were modest during isometric contractions but increased markedly when the specimens were allowed to shorten. Other reports32 showed that the protein level of SR calcium ATPase, which would be involved in the reuptake of a shortening-related increase in calcium in our hypothesis, is selectively reduced in human dilated cardiomyopathy. Thus, the current results may be relevant to a mechanism that is specifically impaired in heart failure. Additionally, there are clinical situations in which myocardial load is altered acutely either at ES or during relaxation itself. Examples include acute mitral regurgitation and ventricular septal defect, in which shortening continues past the time of aortic valve closure with continuing ejection into the left atrium or the right ventricle. Similarly, acute aortic regurgitation increases load after ES. It is possible that an acute alteration in the “baseline” level of shortening activation is involved in the response to these pathological situations.

Selected Abbreviations and Acronyms

Reprint requests to Matthew W. Watkins, MD, Cardiology Unit, Medical Center Hospital of Vermont-McClure 1, Burlington, VT 05401.

Figure 1. Top, Data from a single heart during each contraction mode are depicted. Compared with ISOV, LV volume (LVV) reduction at ES is associated with a steady-state increase in LV pressure (LVP) and dP/dt. EDV is constant. The magnitude of the contractility effect increases when the rate of VR is increased from 10 mL/s (slow VR) to 56 mL/s (rapid VR). Bottom, The increase in LVP after an abrupt change from ISOV contraction to rapid VR requires >50 beats to attain steady state. Return to the initial ISOV condition is associated with a gradual pressure decline to the initial LVP.

Figure 2. The increase in nonmechanical V̇o2 in individual hearts is shown for each heart (mean, 26%) associated with rapid VR compared with ISOV at the same EDV.

This study was supported by American Heart Association Grant-in-Aid No. 93006340 and NHLBI grant No. HL-51201. We thank Stephen P. Bell and Judit Fabian for excellent technical assistance and Dr Bryan K. Slinker for help in statistical analysis.

References

• 1 Yasumura Y, Nozawa T, Futaki S, Tanaka N, Suga H. Time-invariant oxygen cost of mechanical energy in dog left ventricle: consistency and inconsistency of time-varying elastance model with myocardial energetics. Circ Res.1989; 64:764-778.CrossrefMedlineGoogle Scholar
• 2 Suga H, Sagawa K, Demer L. Determinants of instantaneous pressure in canine left ventricle. Circ Res.1980; 46:256-263.CrossrefMedlineGoogle Scholar
• 3 Hisano R, Cooper G IV. Correlation of force-length area with oxygen consumption in ferret papillary muscle. Circ Res.1987; 61:318-325.CrossrefMedlineGoogle Scholar
• 4 DeTombe PP, Little WC. Inotropic effects of ejection are myocardial properties. Am J Physiol. 1994;266(Heart Circ Physiol 35):H1202-H1213.Google Scholar
• 5 Allen DG, Kurihara S. The effects of muscle length on intracellular calcium transients in mammalian cardiac muscle. J Physiol (Lond).1982; 327:79-94.CrossrefGoogle Scholar
• 6 Allen DG, Nichols CG, Smith GL. The effects of changes in muscle length during diastole on the calcium transient in ferret ventricular muscle. J Physiol (Lond).1988; 406:359-370.CrossrefGoogle Scholar
• 7 Cooper G. Myocardial energetics during isometric twitch contractions of cat papillary muscle. Am J Physiol.1979; 236:H244-H253.CrossrefMedlineGoogle Scholar
• 8 Alpert NR, Mulieri LA. Determinants of energy utilization in the activated myocardium. Fed Proc.1986; 45:2597-2600.MedlineGoogle Scholar
• 9 Hunter WC, Janicki JS, Weber KT, Noordergraaf A. Systolic mechanical properties of the left ventricle. Circ Res.1983; 52:319-327.CrossrefMedlineGoogle Scholar
• 10 Sugiura S, Hunter WC, Sagawa K. Long-term versus intrabeat history of ejection as determinants of canine ventricular end-systolic pressure. Circ Res.1989; 64:255-264.CrossrefMedlineGoogle Scholar
• 11 Igarashi Y, Goto Y, Yamada O, Tatsuo I, Suga H. Transient vs steady-state end-systolic pressure-volume relation in dog left ventricle. Am J Physiol.1987; 252:H998-H1004.MedlineGoogle Scholar
• 12 Burkhoff D, DeTombe PP, Hunter WC, Kass DA. Contractile strength and mechanical efficiency of left ventricle are enhanced by physiological afterload. Am J Physiol. 1991;260(Heart Circ Physiol 29):H569-H578.Google Scholar
• 13 Yaku H, Slinker BK, Mochizuki T, Lorell BH, LeWinter MM. Use of 2,3-butanedione monoxime to estimate nonmechanical V̇o2 in rabbit hearts. Am J Physiol.1993; 265:H834-H842.CrossrefMedlineGoogle Scholar
• 14 Higashiyama A, Watkins MW, Chen Z, LeWinter MM. Preload does not influence nonmechanical oxygen consumption in the isolated rabbit heart. Am J Physiol.1994; 266:H1047-H1054.MedlineGoogle Scholar
• 15 Watkins MW, Slinker BK, Goto Y, LeWinter MM. 2,3-Butanedione monoxime increases the efficiency of chemomechanical energy transduction in rabbit ventricle. Am J Physiol.1992; 263:H1811-H1818.MedlineGoogle Scholar
• 16 Marshall RC. Correlation of contractile dysfunction with oxidative energy production and tissue high energy phosphate stores during partial coronary flow disruption in rabbit heart. J Clin Invest.1988; 82:86-95.CrossrefMedlineGoogle Scholar
• 17 Slinker BK, Schroff SG, Kirkpatrick RD, Campbell KB. Left ventricular function depends on previous beat ejection but not previous beat pressure load. Circ Res.1991; 69:1051-1057.CrossrefMedlineGoogle Scholar
• 18 Suga H. Ventricular energetics. Physiol Rev.1990; 70:247-277.CrossrefMedlineGoogle Scholar
• 19 Alpert NR, Blanchard EM, Mulieri LA. Tension independent heat in rabbit papillary muscle. J Physiol (Lond).1989; 414:433-453.CrossrefGoogle Scholar
• 20 Blanchard EM, Smith GL, Alpert NR. The effects of 2,3-butanedione monoxime on initial beat, tension, and aequorin light output of ferret papillary muscle. Pflugers Arch.1990; 416:219-221.CrossrefMedlineGoogle Scholar
• 21 Perreault CL, Mulieri LA, Alpert NR, Ransil BJ, Allen PD, Morgan JP. Cellular basis of 2,3-butanedione monoxime in human myocardium. Am J Physiol. 1992;263(Heart Circ Physiol 32):H503-H510.Google Scholar
• 22 Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. London, England: Kluwer Academic Publishers; 1991.Google Scholar
• 23 Goto Y, Slinker BK, LeWinter MM. Decreased contractile efficiency and increased nonmechanical energy cost of hyperthyroid rabbit heart. Circ Res.1990; 66:999-1011.CrossrefMedlineGoogle Scholar
• 24 Watkins MW, Higashiyama A, LeWinter MM. Rapid volume reduction during relaxation improves systolic performance. Circulation. 1993;88(suppl):I-86. Abstract.Google Scholar
• 25 Hunter WC. End-systolic pressure as a balance between opposing effects of ejection. Circ Res.1989; 64:265-275.CrossrefMedlineGoogle Scholar
• 26 Babu A, Sonnenblick E, Gulati J. Molecular basis for the influence of muscle length on myocardial performance. Science.1988; 240:74-76.CrossrefMedlineGoogle Scholar
• 27 Housmans PR, Lee NKM, Blinks JR. Active shortening retards the decline of the intracellular calcium transient in mammalian heart muscle. Science.1983; 221:159-161.CrossrefMedlineGoogle Scholar
• 28 Hansen DE, Borganelli M, Stacy GP Jr, Taylor LK. Dose-dependent inhibition of stretch-induced arrhythmias by gadolinium in isolated canine ventricles. Circ Res.1991; 69:820-831.CrossrefMedlineGoogle Scholar
• 29 Morris CE. Mechanosensitive ion channels. J Membr Biol.1990; 113:93-107.CrossrefMedlineGoogle Scholar
• 30 Ruknudin A, Sachs F, Bustamante JO. Stretch-activated ion channels in tissue-cultured chick heart. Am J Physiol. 1993;264(Heart Circ Physiol 33):H960-H972.Google Scholar
• 31 Vahl CF, Bonz A, Timek T, Hagl S. Intracellular calcium transient of working human myocardium of seven patients transplanted for congestive heart failure. Circ Res.1994; 74:952-958.CrossrefMedlineGoogle Scholar
• 32 Meyer M, Schillinger W, Pieske B, Holubarsch C, Heilmann C, Posival H, Kuwajima G, Mikoshiba K, Just H, Hasenfuss G. Alterations in sarcoplasmic reticulum proteins in failing human dilated cardiomyopathy. Circulation.1995; 92:778-784.CrossrefMedlineGoogle Scholar

Page 3

The patient is a 58-year-old African American man with past medical history significant for a positive purified protein derivative (PPD) 2 years ago. No treatment was given. He was in his usual state of good health until 6 weeks before admission, when he became aware of breathlessness on mild exertion. Over the ensuing 4 weeks, he noted progressive fatigue; swelling of his lower extremities, particularly on standing; and increasing abdominal girth. His exercise tolerance decreased until he could no longer walk a single city block comfortably. He denied chest pain, palpitations, orthopnea, paroxysmal nocturnal dyspnea, and nocturia. He also denied fever and significant weight loss but did complain of occasional night sweats. The patient was treated with furosemide 20 mg/d without improvement.

His family history was remarkable for a father and a brother who died of unknown heart disease in their 50s. He was married with one healthy son. He immigrated to the United States from St Croix 10 years before admission and worked as a mailroom clerk. He had no significant travel history, was a lifetime nonsmoker, and drank alcohol only socially.

Physical examination revealed an obese African American man in mild respiratory distress. He was afebrile. The heart rate was 80 bpm and regular, and his blood pressure was 120/70 mm Hg. Respiratory rate was 20 breaths per minute. His head was normal except for ill-defined hyperpigmentation periorbitally and bitemporally. His neck was supple without lymphadenopathy. Carotid upstrokes were normal. Jugular venous pressure was estimated at 14 cm water, and there was a Kussmaul's sign; prominent x and y descents were noted. The thyroid gland was normal. There were no gynecomastia and no palpable lymph nodes anywhere. The lungs were clear. Cardiac examination showed a nondisplaced point of maximal intensity and normal S1 and S2. There was a soft S3 gallop and a soft early systolic murmur heard at the lower right sternal border. There was no rub. The abdomen was distended with shifting dullness. The liver span was percussed 18 cm in the midclavicular line, and a spleen tip was palpable. There was pitting edema extending bilaterally from the ankles to midthighs. Genitourinary and rectal examination were unremarkable. The neurological examination was significant for mildly decreased sensation to light touch and proprioception in the lower extremities; the motor examination and reflexes were normal.

The ECG showed normal sinus rhythm, borderline-low QRS voltage, and QS complexes in V1 through V3, suggestive of anteroseptal myocardial infarction. Chest roentgenogram showed mild cardiomegaly and no evidence of effusions or focal infiltrates. Admission laboratory data revealed a normal blood count, blood electrolytes, glucose, blood urea nitrogen, and prothrombin time. Liver function tests showed normal transaminases, alkaline phosphatase of 450 U/L, total bilirubin of 1.7 mg/dL (direct, 0.5 mg/dL), total protein of 6.5 g/dL, albumin of 4 g/dL, and lactic dehydrogenase of 457 U/mL. Ferritin was 81 ng/mL; amylase, 63 U/L; and erythrocyte sedimentation rate, 3 mm/h. Urinalysis showed 3+ protein and no casts or cells.

The patient was placed on a low-sodium diet and given intravenous furosemide. With a net diuresis of 1 L, his peripheral edema decreased by the next morning. A diagnostic paracentesis revealed 3.3 g/dL protein, 88 mg/dL glucose, 109 U/mL lactic dehydrogenase, 63 U/L amylase, and 23 nucleated cells per milliliter (mostly lymphocytes).

A transthoracic echocardiogram (Fig 1) revealed thickened left ventricular (LV) and right ventricular (RV) walls with preserved systolic function, mild left and right atrial dilatation, and moderate tricuspid and mitral regurgitation. A small pericardial effusion was present, and the pericardial reflections were described as echo dense. Mitral inflow pattern consisted of a rapid filling phase with short deceleration time and a diminished A wave (late filling wave corresponding to atrial systole). The pattern of ultrasonic reflection from the myocardium was fine and diffuse without discrete “speckling” or highly refractile pattern.

Cardiac catheterization revealed a pulmonary artery pressure of 60/20 mm Hg, RV pressure of 60/22 mm Hg, LV pressure of 123/24 mm Hg, and mean right atrial pressure of 23 mm Hg, with a tracing showing a prominent y descent. The diastolic pressure tracings in both ventricles exhibited a characteristic square root (dip and plateau) pattern. Left ventriculogram showed an increased ejection fraction and mild mitral regurgitation. Cardiac output was 3.2 L/min, with an index of 1.7 L·min−1·m−2. The coronary arteries were normal.

Serum protein electrophoresis was normal, and urine was negative for Bence Jones protein. A 24-hour urine sample revealed a protein of 141 mg/dL. The patient continued to be treated conservatively with diuretics, with marked weight loss and decreased dyspnea. A diagnostic procedure was performed.

Clinical Discussion (I. Kronzon)

This patient had ankle edema, ascites, and marked neck vein distention. These findings are characteristic of failure of the right side of the heart. In addition, there were symptoms of marked shortness of breath on exertion, which also suggests failure of the left side of the heart. The LV systolic function, which was evaluated by echocardiography and left ventriculography, was increased, with evidence of a hyperkinetic left ventricle. RV systolic function on echocardiography was normal. There were no physical, echocardiographic, or angiographic findings to suggest severe mitral or tricuspid regurgitation. I believe that the differential diagnosis is that of diastolic dysfunction. Indeed, the cardiac catheterization showed significantly elevated diastolic pressure in both ventricles, with equalization of all diastolic pressures. There also was a characteristic square root appearance of the diastolic ventricular pressures and a steep y descent noted on the right atrial pressure curve. Although LV diastolic dysfunction leading to symptoms of failure of the left side of the heart is quite common, this syndrome is frequently associated with hypertension or ischemic heart disease and usually is not associated with failure of the right side of the heart. In this patient, it appears that diastolic dysfunction affects both chambers equally.

Two pathophysiological mechanisms can lead to this clinical picture: restrictive cardiomyopathy and constrictive pericarditis. Differentiation between these two conditions is difficult. In many patients, a comprehensive workup, including noninvasive and invasive techniques, does not solve the clinical puzzle, and the diagnosis finally is made at surgery or sometimes at autopsy. Several clinical hints and a few findings, however, should help differentiate between the two. Table 1 summarizes these findings.

The hallmark of both syndromes is significant, frequently severe failure of the right side of the heart. Jugular venous distention, ankle edema, ascites, hepatomegaly, and elevation of liver enzymes as a result of liver engorgement may be associated with these disorders. In both disorders, most of the ventricular filling occurs in early diastole, with a rapid decrease in right atrial pressure immediately after the opening of the tricuspid valve; therefore, a prominent y descent is noted in the jugular veins. Kussmaul's sign, a paradoxical increase in jugular venous distention during inspiration, also can occur in both conditions. Cardiac examination may be useful. In constrictive pericarditis, the heart (which is encased within a thickened pericardium) is quite quiet. Therefore, the point of maximal intensity frequently is not palpable. By contrast, the point of maximal intensity is usually well detected in patients with restrictive cardiomyopathy. Although S1 and S2 are normal in both conditions, S3 does not occur in constrictive pericarditis. What one can hear is a higher-pitched pericardial knock, which occurs at the end of rapid ventricular filling. A lower pitched S3 is more likely to be present in restrictive cardiomyopathy. However, because the timings for S3 and pericardial knock are similar (approximately 180 ms after S2), the interpretation of the auscultatory findings may be misleading, and one may be mistaken for the other.

Chest roentgenogram and fluoroscopy may be helpful. Approximately 50% of patients with constrictive pericarditis have pericardial calcification. Dramatic cases of encasement of the heart within a calcified “eggshell” are sometimes seen. However, the absence of calcification does not rule out constrictive pericarditis, and pericardial calcification does not necessarily indicate constrictive physiology. Newer tomographic technologies such as CT scanning and magnetic resonance imaging of the heart may help in the differential diagnosis. Pericardial thickening can be detected in most patients with constrictive pericarditis, and it characteristically is absent in restrictive cardiomyopathy. In our patient, however, CT and magnetic resonance imaging of the heart were not performed.

Cardiac catheterization clearly demonstrates the characteristic patterns of diastolic dysfunction. In constrictive pericarditis, the whole heart characteristically is encased within a thickened, noncompliant pericardium that limits its diastolic filling. Filling occurs early in diastole and stops when the nondistensible pericardium is stretched to its limit. At this point, all diastolic pressures are high and equal. Equalization of elevated diastolic pressures is therefore the hallmark of pericardial constriction. In restrictive cardiomyopathy, cardiac catheterization also reveals signs of rapid ventricular filling that reaches a high diastolic pressure plateau (square root sign) as seen in constrictive pericarditis. However, the extent of the process may differ in different chambers; therefore, although both RV and LV diastolic pressures may be elevated, they frequently are not identical. Usually, the LV diastolic pressure is higher than the RV diastolic pressure. Unfortunately, in ≈30% of patients with restrictive cardiomyopathy studied by cardiac catheterization, the diastolic pressures in the left and right chambers are nearly equal. To evaluate these patients, changing of preload or afterload may be useful. Maneuvers such as infusion of 500 cm3 normal saline or exercise will separate right- and left-sided diastolic pressures in patients with restrictive cardiomyopathy. In contrast, the diastolic pressures will increase but remain equal in patients with constrictive pericarditis. We were not told whether such maneuvers were performed in the patient under discussion.

Echocardiography and, in particular, Doppler echocardiographic studies also can be useful in the differential diagnosis. Normal LV function or hyperkinesis is the rule in all cases of constrictive pericarditis and can be present in some cases of advanced restrictive cardiomyopathy in which diastolic dysfunction is dominant. Table 2 compares other characteristic Doppler echocardiographic features of the two disorders. Doppler echocardiography can clearly differentiate between the two disorders. In both disorders (constrictive and restrictive), the flow velocity pattern across an AV valve (best observed by transmitral flow studies) is similar: it stops abruptly after rapid ventricular filling, and thus the deceleration time of transvalvular flow velocity is quite rapid (usually >160 ms). However, the effect of the respiratory cycle on the transvalvular flow differs in the two disorders (see Table 2). An accurate determination of respiratory variation of flow velocity was not performed in this patient.

The patient had a positive PPD that had been noted 2 years earlier and was not treated. One of the most common causes of constrictive pericarditis is tuberculous pericarditis. The patient did complain of occasional severe night sweats, which also are frequently associated with active tuberculosis. Not infrequently, tuberculous pericarditis may be the only manifestation of tuberculosis. In this patient, however, active tuberculosis seems unlikely because of the lack of fever, normal sedimentation rate, and lack of other radiological or laboratory evidence of chronic infection.

The echocardiographic findings of RV and LV wall thickening are quite remarkable. They occur in the presence of relatively low voltage on the ECG. The combination of significant LV wall thickening, low voltage on the ECG, and failure of the right side of the heart is highly suggestive of restrictive cardiomyopathy secondary to an infiltrative disorder.

Hemodynamic studies showed severe diastolic dysfunction with equalization of all diastolic pressures. A hemodynamic challenge with preload change (by saline infusion or exercise) was not described in this patient. However, the pulmonary artery and RV systolic pressures were markedly elevated, to 60 mm Hg. This finding is unusual in patients with constrictive pericarditis, in which pulmonary artery pressures rarely exceed 40 mm Hg. Thus, the catheterization findings also support the diagnosis of restrictive cardiomyopathy.

Table 3 lists the conditions associated with restrictive cardiomyopathy. Some are congenital and present themselves early in life. Others can be ruled out by this patient's history. Hemochromatosis could be an attractive clinical diagnosis in this patient with vague skin discoloration; however, the normal ferritin level rules this diagnosis out. Sarcoidosis is another possibility; however, massive cardiac sarcoidosis usually presents itself with conduction abnormalities, and systolic dysfunction frequently precedes diastolic dysfunction clinically. There is no clinical or laboratory evidence of secondary spread of neoplasm into the myocardium. Thus, we are left with a common pathogenesis for restrictive cardiomyopathy, namely, cardiac amyloidosis.

Characteristically, cardiac amyloidosis can be associated with dramatic ventricular wall thickening that can be demonstrated by echocardiography and other imaging techniques. A characteristic echocardiographic finding of myocardial speckling has been described, but it is found in only 50% of patients with amyloidosis. Thus, the lack of speckling on this patient's echocardiogram does not rule out amyloidosis. Long-standing amyloidosis can lead to LV systolic dysfunction; however, diastolic dysfunction with well-preserved LV wall motion frequently is an early manifestation of the disorder. Another approach to diagnosing amyloid at this stage is radionuclide imaging. Technetium pyrophosphate characteristically is absorbed by myocardial amyloid, and this technique has been used to demonstrate and diagnose cardiac amyloidosis. This test was not done in this patient.

There are different forms of amyloidosis both clinically and biochemically. They include primary amyloidosis, in which there is no additional known disease; amyloidosis associated with multiple myeloma; secondary amyloidosis associated with conditions such as tuberculosis, osteomyelitis, or leprosy; and amyloidosis associated with heredofamilial disorders, which also is frequently associated with peripheral neuropathy and occasionally with familial Mediterranean fever. Amyloidosis associated with old age also has been described. In this patient, secondary amyloidosis is unlikely. This form of amyloidosis infrequently creates symptomatic heart disease. There is also no clinical evidence of a long-standing, active inflammatory process. The diagnosis of multiple myeloma is not supported because of the protein electrophoresis, absence of Bence Jones protein, and normal sedimentation rate. Amyloidosis associated with aging is unlikely in this 58-year-old patient. This entity also is frequently an incidental finding discovered at autopsy that usually does not cause the severe clinical manifestations observed in this patient. We therefore remain with a primary or, alternatively, a heredofamilial form of amyloidosis. The family history of this patient, with two relatives who died at a relatively young age of vaguely described heart disease, and his peripheral neuropathy may support the latter diagnosis. However, I believe that once the heart is affected by the infiltration of amyloid that is severe enough to cause severe congestive heart failure, the prognosis is poor, and from the cardiologist's clinical point of view, there is little difference. Because this case was diagnosed years ago, I assume that the diagnostic procedure performed was a cardiac biopsy. However, I believe that this procedure is not without danger, and amyloid can be diagnosed in most cases by such procedures as gingival biopsy or, better yet, biopsy of the subcutaneous abdominal fat. Cardiac biopsy should be reserved only for cases in which these biopsies are negative.

Pathological Findings (G. Gallo)

The patient had RV biopsy. The endomyocardial biopsy was stained with Congo red, which showed green birefringence of deposits under polarization microscopy, typical of amyloid (Fig 2A). Frozen sections were incubated with a panel of antibodies against IgG, IgA, IgM (heavy-chain specific), κ light chain, λ light chain (light-chain specific), transthyretin (TTR), amyloid A protein, and amyloid P component. Deposits stained for λ light chain and amyloid P component (Fig 2B) but were negative for the other proteins tested. The deposits were present diffusely in vessels and around myocardial cells.

Electron microscopy demonstrated randomly oriented fibrils typical of amyloid in the plasmalemmal sheath and interstitially between muscle cells (Fig 3).

Hematology Commentary and Follow-up (D.R. Jacobson)

This patient came to my attention after the diagnosis of amyloidosis was made, and I cared for this patient in the clinic after his discharge from the hospital. Of the comments made so far, the one statement with which I disagree is the comment that because the prognosis of symptomatic cardiac amyloidosis is poor, it makes little difference what type of amyloid is present. Considerable progress has been made recently in understanding amyloid on the molecular level, and this is translating into better approaches to treatment.

First, I would like to address the nomenclature of amyloid because the terms “primary” and “secondary” amyloid are used here. These terms originated long before amyloid was understood on a molecular level, and now that we can determine the specific protein deposited in most patients with amyloidosis, it is recommended that these terms be abandoned and that amyloid be referred to whenever possible by the chemical classification123 ; unfortunately, the terms “primary” and “secondary” persist in the literature, causing considerable confusion. Historically, secondary amyloidosis referred to the amyloid that accompanied chronic inflammatory processes such as tuberculosis and rheumatoid arthritis. Familial amyloidosis was recognized by the positive family history. All other types of amyloidosis, except that associated with the multiple myeloma, were called primary in the sense of idiopathic; this category included unrecognized inherited forms, secondary amyloidosis without an identified cause, and localized amyloidosis. This classification was based on the assumption that all forms of amyloidosis would consist of a single predominant protein.

We now know that all forms of amyloid consist of a minor glycoprotein, the P (pentagonal) component, which is identical in all types of amyloid, and the major fibrillar component, 15 of which have been identified in human amyloidosis. As a group, the amyloid precursor proteins are small, with molecular weights of 4000 to 25 000 D. Their tertiary structures are characterized by a substantial β-pleated sheet structure, which is thought to play a role in amyloid formation, but the precise mechanisms of fibril formation remain poorly understood. The major protein component defines the type of amyloidosis and determines the pathogenesis of the disease. Classification based on clinical syndromes is now avoided whenever chemical information is available. Thus, in what was called secondary amyloidosis, the amyloid fibrils consist of the amyloid A (AA) protein; these diseases are called AA amyloidosis. The amyloid material consisting of immunoglobulin light chains or light-chain fragments (AL amyloid) originates from a single clone of plasma cells. When this clone has expanded to the extent that the criteria for the diagnosis of multiple myeloma are fulfilled, the disease is called myeloma-associated amyloid; when the clone has a limited proliferative capacity and myeloma criteria are not met, the disease previously was called primary systemic amyloidosis. In such cases, the cause of the patient's disease is essentially the same as monoclonal gammopathy of undetermined significance, with the additional feature that the monoclonal protein happens to be amyloidogenic. From the standpoint of amyloidosis, myeloma-associated amyloidosis and primary systemic amyloidosis are identical processes, and both should instead be referred to as AL amyloid.

Of the various types of amyloid that form deposits in the cardiac ventricles and cause congestive heart failure, the most common are AL and TTR amyloid; cardiac AA occurs less often. TTR is a serum transport protein consisting of four identical subunits of 127 amino acids each. TTR transports thyroxine- and retinol-binding protein and is synthesized primarily in the liver, choroid plexus, and retina. Normal-sequence TTR has a low-grade inherent tendency to form amyloid, and a small amount of TTR amyloid in the cardiac ventricles is found incidentally at autopsy in >25% of people 80 years of age or older.4 This process, usually asymptomatic, has been called senile cardiac amyloidosis. Because other organs may be involved, the alternative name “senile systemic amyloidosis” also is used.

In some patients, the process of TTR amyloid deposition is accelerated, leading to congestive heart failure and/or arrhythmias. Occasionally, patients with large amounts of cardiac TTR amyloid have deposits consisting of normal-sequence TTR56 ; in these patients, the stimulus for accelerated deposition is not known. More commonly, the stimulus for symptomatic TTR amyloidosis is a TTR point mutation, changing the conformation of the molecule and leading to increased deposition. Nearly 50 different amyloidogenic TTR mutations are now known, most of which have been described in single kindreds or ethnic groups.7 The most severely affected organs are typically the heart, peripheral nervous system, eye, and gastrointestinal tract. The disease caused by variant TTR typically is called familial amyloid cardiomyopathy or familial amyloid polyneuropathy.1

One amyloidogenic TTR variant, TTR Ile 122, is carried by 4% of African Americans8 (D.R.J. and colleagues, unpublished data, 1995) and has been found in several patients with severe congestive heart failure.91011 The risk of TTR Ile 122 carriers developing symptomatic amyloidosis is not yet known, but it clearly is greater than for people with the normal TTR sequence. When I first saw this patient, immunohistochemistry had not yet been performed, so we did not know the type of amyloid. At the time, I thought that he may have had TTR Ile 122 amyloidosis because there reportedly was no evidence of a monoclonal protein in the serum or urine. While awaiting immunohistochemistry, we performed genetic studies12 and determined that the patient was indeed heterozygous for TTR Ile 122; however, immunohistochemistry demonstrated that the patient's true diagnosis was AL amyloid, and the genetic studies turned out to be a coincidental finding.

After AL amyloidosis was diagnosed, the patient was evaluated for multiple myeloma, as should all patients with AL amyloid. A bone marrow examination revealed 6% plasma cells, and no lytic lesions were seen on skeletal survey; thus, the patient did not have myeloma. Reportedly, only about half of patients with AL amyloid have a monoclonal immunoglobulin protein detectable in the serum or urine. This is really an issue of the practical limitation of the assays used clinically to detect monoclonal proteins. The amyloid precursor, the monoclonal immunoglobulin molecule, is synthesized by monoclonal plasma cells in the bone marrow and is deposited as amyloid in the heart; thus, this protein must travel through the bloodstream. So if a sensitive-enough assay is used, in theory, all patients with AL amyloid must have a monoclonal serum and/or urine protein (free immunoglobulin light chains are small enough to be filtered by the glomerulus and appear in the urine). When we repeated the urine protein immunoelectrophoresis on a concentrated specimen in a research laboratory, the monoclonal protein was detected, even though this test remained negative as reported by the routine clinical laboratory.

So does it matter if we know whether the patient has TTR or AL cardiac amyloid? A decade ago, perhaps not, but in 1996, yes. Chemotherapy is of value for multiple myeloma, so it would seem logical that the same chemotherapy might be of use for AL amyloid even if diagnostic criteria for myeloma are not fulfilled. For many years, reports suggested that chemotherapy is valuable for treating AL amyloid, and two recently randomized, controlled trials have demonstrated a survival advantage for patients receiving chemotherapy,131415 so this should now be standard therapy for patients with AL amyloid, even in the absence of myeloma.

On the other hand, patients with TTR amyloid will not benefit from chemotherapy. For patients with TTR amyloid, studies to determine whether a TTR variant is present can help guide management. Particularly in younger patients, if a TTR variant is present, liver transplantation can be performed as a means of replacing the source of variant TTR with normal-sequence TTR; patients who receive liver transplantations have gradual resolution of their amyloid and may achieve complete resolution of symptoms.1617 At present, there is no known effective treatment for amyloid consisting of normal-sequence TTR.

This patient's congestive heart failure improved with large doses of diuretics (furosemide and spironolactone), and he improved symptomatically. He also was treated with monthly melphalan and prednisone, and his heart disease appeared to stabilize or even improve. The median survival of patients with symptomatic congestive heart failure resulting from AL amyloidosis is months for patients not receiving or not responding to chemotherapy.18 In a subset of patients, however, chemotherapy leads to resorption of the amyloid, improvement in cardiac function, and longer survival.18 Chemotherapy was discontinued after nearly 2 years because of thrombocytopenia. Shortly thereafter, the patient moved out of town and has been lost to follow-up. The optimal duration of chemotherapy for patients who respond is not known. This patient's clinical stabilization and nearly symptom-free state 2 years after diagnosis clearly demonstrate the value of determining the specific type of amyloid present in each patient and instituting appropriate therapy.

Clinical Diagnosis

The clinical diagnosis was cardiac amyloidosis, probably heredofamilial.

Final Diagnosis

The final diagnosis is λ light-chain amyloidosis of the myocardium.

Figure 1. A, Two-dimensional echocardiogram. Apical view reveals left ventricular (LV) and right ventricular (RV) hypertrophy. B, Transmitral flow velocity (by pulsed Doppler echocardiography) shows short mitral deceleration time (128 ms). LA indicates left atrium; DT, deceleration time.

Figure 2. A, Congo red stain reveals deposits characteristic of amyloid. B, Incubation with anti-λ light-chain antibodies reveals characteristic diffuse amyloid deposits.

Figure 3. Electron microscopy shows randomly oriented fibrils typical of amyloid.

This work was supported in part by an Established Scientist Award, American Heart Association, New York City affiliate (Dr Jacobson).

Footnotes

References

• 1 Jacobson DR, Buxbaum JN. Genetic aspects of amyloidosis. Adv Hum Genet..1991; 20:69-123.MedlineGoogle Scholar
• 2 Husby G. Nomenclature and classification of amyloid and amyloidosis. J Intern Med..1992; 232:511-512.CrossrefMedlineGoogle Scholar
• 3 WHO-IUIS Nomenclature Sub-Committee. Nomenclature of amyloid and amyloidosis. Bull World Health Organ..1993; 71:105-108.MedlineGoogle Scholar
• 4 Smith TJ, Kyle RA, Lie JT. Clinical significance of histopathological patterns of cardiac amyloidosis. Mayo Clin Proc..1984; 59:547-555.CrossrefMedlineGoogle Scholar
• 5 Christmansson L, Betsholtz C, Gustavsson Å, Johansson B, Sletten K, Westermark P. The transthyretin cDNA sequence is normal in transthyretin-derived senile systemic amyloidosis. FEBS Lett..1991; 281:177-180.CrossrefMedlineGoogle Scholar
• 6 Gustavsson A, Jahr H, Tobiassen R, Jacobson DR, Sletten K, Westermark P. Amyloid fibril composition and transthyretin gene structure in senile systemic amyloidosis. Lab Invest..1995; 73:703-708.MedlineGoogle Scholar
• 7 Saraiva MJM. Transthyretin mutations in health and disease. Hum Mutat..1995; 5:191-196.CrossrefMedlineGoogle Scholar
• 8 Jacobson DR, Reveille JD, Buxbaum JN. Frequency and genetic background of the position 122 (Val-Ile) transthyretin gene in the black population. Am J Hum Genet..1991; 49:192-198.MedlineGoogle Scholar
• 9 Jacobson DR, Gorevic PD, Buxbaum JN. A homozygous transthyretin variant associated with senile systemic amyloidosis: evidence for a late-onset disease of genetic etiology. Am J Hum Genet..1990; 47:127-136.MedlineGoogle Scholar
• 10 Nichols WC, Liepnieks JJ, Snyder EL, Benson MD. Senile cardiac amyloidosis associated with homozygosity for a transthyretin variant (ILE-122). J Lab Clin Med..1991; 117:175-180.MedlineGoogle Scholar
• 11 Saraiva MJM, Sherman W, Marboe C, Figueira A, Costa P, De Freitas AF, Gawinowicz MA. Cardiac amyloidosis: report of a patient heterozygous for the transthyretin-isoleucine 122 variant. Scand J Immunol..1990; 32:341-346.CrossrefMedlineGoogle Scholar
• 12 Jacobson DR. A specific test for transthyretin 122 (Val-Ile) based on PCR-primer introduced restriction analysis (PCR-PIRA): confirmation of the gene frequency in blacks. Am J Hum Genet..1992; 50:195-198.MedlineGoogle Scholar
• 13 Kyle RA, Gertz MA, Garton JP, Greipp PR, Witzig TE, Lust JA. Primary systemic amyloidosis (AL): randomized trial of colchicine vs melphalan and prednisone vs melphalan, prednisone, and colchincine. In: Kisilevsky R, Benson MD, Frangione B, Gauldie J, Muckle TJ, Young ID, eds. Amyloid and Amyloidosis 1993. New York, NY: Parthenon Publishing; 1994:648-650.Google Scholar
• 14 Skinner M, Anderson J, Wang M, Simms R, Falk R, Jones LA, Cohen AS. Treatment of patients with primary amyloidosis. In: Kisilevsky R, Benson MD, Frangione B, Gauldie J, Muckle TJ, Young ID, eds. Amyloid and Amyloidosis 1993. New York, NY: Parthenon Publishing; 1994:232-234.Google Scholar
• 15 Gertz MA, Kyle RA. Amyloidosis: prognosis and treatment. Semin Arch Rheum..1994; 24:124-138.CrossrefMedlineGoogle Scholar
• 16 Holmgren G, Ericzon B-G, Groth C-G, Steen L, Suhr O, Andersen O, Wallin BG, Seymour A, Richardson S, Hawkins PN, Pepys MB. Clinical improvement and amyloid regression after liver transplantation in hereditary transthyretin amyloidosis. Lancet..1993; 341:1113-1116.CrossrefMedlineGoogle Scholar
• 17 Skinner M, Lewis WD, Jones LA, Kasirsky J, Kane K, Ju ST, Jenkins R, Falk RH, Simms RW, Cohen AS. Liver transplantation as a treatment for familial amyloidotic polyneuropathy. Ann Intern Med..1994; 120:133-134.CrossrefMedlineGoogle Scholar
• 18 Gertz MA, Kyle RA, Greipp PR. Response rates and survival in primary systemic amyloidosis. Blood..1991; 77:257-262.CrossrefMedlineGoogle Scholar