Abstract
A driamycin (ADR; doxorubicin HCl), is a potent anthracycline chemotherapeutic agent that is effective against a wide range of human malignancies, such as leukemias, lymphomas, and many solid tumors.1,2 For many malignancies (eg, breast cancer, lymphoma), it is the most effective single agent. 3 The use of ADR has been limited, however, by the development of a dose-dependent cardiomyopathy induced by this antineoplastic agent.1,4–7 The overall incidence of clinical cardiotoxicity for patients treated with ADR is approximately 2%, with the incidence of congestive heart failure increasing to 10% at 500 mg/m2 and to over 40% at 700 mg/m2.4,8–10 The cardiac-specific injury of ADR is usually irreversible, although aggressive support for a low cardiac output may reverse the dysfunction in some patients.6,11,12 Nevertheless, once ADR-induced heart failure has been diagnosed, the prognosis is poor; mortality varies from 30 to 60%.6,11 It is also noteworthy that various other commonly used antineoplastic agents, cyclophosphamide, 13 5-fluorouracil, 14 paclitaxel,15,16 streptozotocin, 13 bevacizumab (Avastin), 13 etoposide, 16 bleomycin, 16 tamoxifen, 17 and trastuzumab (Herceptin), 17 as well as radiation therapy, 17 which is commonly used as an adjuvant to lumpectomy, are also cardiotoxic, although different mechanisms may be involved. Therefore, the development of sensitive methods for early detection of drug- or radiation-induced cardiotoxicity in individual patients could greatly facilitate the management of a wide range of neoplastic diseases.
Development of ADR-induced cardiomyopathy is dose dependent and cumulative.4,8–10 To reduce the incidence of cardiotoxicity, limiting treatment with this agent to a total lifetime dose of ADR of 550 mg/m2 is recommended. 10 However, the individual response to ADR is highly variable, and bolus administration of up to 900 mg/m2 has been reported without evidence of heart failure 18 ; up to 1,500 mg/m2 has been administered by continuous infusion without evidence of cardiotoxicity. 2 In some patients, drug resistance, rather than toxicity, is dose limiting 2 ; in other instances, restriction of ADR therapy to the recommended guidelines may lead to the premature discontinuation of treatment in patients who might otherwise benefit from higher doses. Conversely, some patients may be highly sensitive to ADR, experiencing toxicity at doses that are lower than the recommended limit. Patients with cardiac disorders or with previous radiation therapy to the mediastinum are generally excluded from ADR treatment protocols. It has also been reported that in some pediatric cases, for which treatment with ADR was well within the recommended limit, congestive heart failure developed several years after treatment. 19 Consequently, the availability of a sensitive, noninvasive method for assessing the early onset of ADR cardiomyopathy in the individual patient would have a significant clinical impact.
At the present time, there does not appear to be any effective method for predicting which patients will develop ADR-induced heart failure 6 or for sensitively detecting its onset at a very early stage, when administration can effectively be discontinued or useful treatment for cardiomyopathy can be initiated. Current methods of assessing cardiotoxicity are limited to measurement of cardiac function by radionuclide ventriculography (RNV), echocardiography, and detection of tissue damage by percutaneous endomyocardial biopsy. 20 RNV is not particularly sensitive for detecting early damage as assessed by endomyocardial biopsy scores. In combination with exercise, RNV sensitivity improves, but at the expense of specificity. Artifactual errors of 5% or more in ejection fraction measurements are common. Therefore, it has been recommended that a combination of RNV or echocardiography and biopsy examinations be employed in the management of patients treated with ADR. Biopsy measurements are invasive, involve some risk, and can yield false-negative results. Fatal congestive heart failure has been encountered in at least one patient with a relatively normal biopsy score. 21 Isner and colleagues failed to detect histologic signs of ADR toxicity in 7 of 20 patients (35%) who had unequivocal clinical signs of dilated cardiomyopathy. 22 Owing to the irreversibility of ADR toxicity, by the time contractile dysfunction or tissue damage is apparent, it is often too late to intervene. In addition, the combined use of RNV and endocardial biopsy is expensive, even by magnetic resonance imaging (MRI) standards, costing $25,650/heart failure death prevented (in 1982 dollars). 20 A better understanding of the effects of ADR on the heart in vivo could lead to less expensive and more reliable noninvasive methods for monitoring the early onset of ADR cardiotoxicity.
In this issue, Gabrielson and colleagues introduce the use of cardiac single-photon emission computed tomography (SPECT) with 99mTc annexin V to detect early onset of ADR-induced cardiomyopathy. 23 The method is based on the movement of phosphatidylserine (PS) from the inner phospholipid envelope to the external envelope of the cell membrane during apoptosis. However, rupture of the plasma membrane during necrosis would also expose PS to annexin V even if it were situated on the inner surface of the phospholipid bilayer. Therefore, the method detects cell death and loss of membrane integrity rather than a specific form of cell death. This is not a problem since cell death, regardless of how it occurs, is the phenomenon of clinical interest. In this respect, the study clearly demonstrates the ability of SPECT with annexin V to detect myocardial cell death at a lower dose and therefore earlier stage of pathology than was detected by echocardiography. Furthermore, this method should be suitable for detection of numerous other cardiomyopathies, including those induced by other antineoplastic agents and by radiation.
A number of critical issues, however, remain to be resolved. It would clearly be desirable to detect ADR cardiomyopathy at an early stage before irreversible damage to the heart has occurred. Clearly, this is not being accomplished in this case since cell death is being detected. A better understanding of the mechanism underlying ADR cardiotoxicity might lead to the design of methods that could detect the onset of toxicity at an earlier reversible state. Thus, there has been considerable evidence that ADR cardiotoxicity resembles the effects of ischemia in the heart.24–28 Recent evidence points to the possibility that ADR is producing a reversible vasoconstrictive effect on the vascular endothelium16,29,30 or smooth muscle 31 in the heart (J.R. Forder, J.C. Chatham, and J.D. Glickson, unpublished, 1996), perhaps as a consequence of interaction with nitric oxide synthase.32,33 This suggests that monitoring of cardiac blood flow with 99mTc-sestamibi or by dynamic contrast-enhanced MRI or arterial spin labeling might detect cardiotoxicity at an earlier stage and that vasodilators may be effectively employed to avoid further progression of the myopathy. There is also evidence of an inflammatory mechanism of cardiotoxicity 13 that might be detected by assaying inflammatory cytokines and might also be treated by anti-inflammatory agents.
Another issue is the choice of animal model. The rat has traditionally been used as the model and was employed in the study by Gabrielson and colleagues. 23 However, normotensive rats often exhibit nephrotoxicity in addition to cardiotoxicity. The hypertensive rat develops a delayed nephrotoxicity, so, initially, cardiotoxicity is observed in the absence of nephrotoxicity. Alternatively, mice, dogs, and primates develop a pathology more closely resembling human ADR cardiotoxicity. Demonstration of onset of apoptosis in animals treated even with low doses of ADR supports the conclusion that cardiotoxicity was occurring in this instance, but it might have also been accompanied by nephrotoxicity or even hepatotoxicity that usually does not occur in the human. The human pathology often occurs up to 20 years after exposure to the drug, which suggests a mutagenic mechanism. In this study, the animals were examined only 10 days following treatment with the drug; this delay should be long enough to avoid confusion between acute and chronic toxicity, but it is not clear that it is long enough to produce the same long-term effects as observed in human patients. The best method for addressing these issues is to move the study as rapidly as possible into the clinic since precise replication of the human pathology is difficult to accomplish in any animal model.