Subclinical Anthracycline Cardiotoxicity in Patients With Acute Promyelocytic Leukemia in Long-Term Remission After the AIDA Protocol
Abstract
Anthracycline chemotherapy remains a critical component of cancer treatment despite its established risk of cardiotoxicity. To investigate whether the AIDA protocol, which combines idarubicin, mitoxantrone, and all-trans retinoic acid (ATRA) for treatment of acute promyelocytic leukemia (APL) results in late cardiotoxicity, 34 APL patients in long-term remission were evaluated. The cumulative dose of idarubicin and mitoxantrone were 80 mg/m2 and 50 mg/m2, respectively. Median follow-up was 7 years. Segmental wall motion abnormalities (SWMAs) were detected in 11 AIDA patients who still presented with an ejection fraction (EF) within normal limits (EF 56% in the AIDA group vs 59% in the control group, P=.01). However, parameters of diastolic dysfunction were significantly impaired in the AIDA group (E/A ratio: 1.04 in the AIDA group vs 1.28 in the control group, P=.001; E/E’ lateral ratio: 10.04 in the AIDA group vs 5.79 in the control group, P≤.001) as well as left atrial volume (52 mL in the AIDA group vs 35 mL in the control group, P<.001). Cardiac toxicity due to anthracycline therapy is often frequent. Changes in diastolic function are helpful in the detection of subclinical anthracycline cardiotoxicity in long-term cardiac follow-up despite a preserved systolic ventricular function.
Anthracycline chemotherapy is an effective therapy for numerous types of malignant tumors, solid and hematological, and remains a critical component of treatment despite their well-established risk of cardiotoxicity.1 Congestive heart failure (HF) and high-grade arrhythmias, until sudden death, are clinically significant anthracycline cardiotoxicity effects that can occur some years after therapy. Late subclinical abnormalities of left ventricular (LV) function in long-term survivors of cancer treated with anthracyclines are common, and the most frequent method of assessing cardiotoxicity is an echocardiographic assessment of LV systolic function during treatment and during the years following its completion.2–5 Evidence suggests that up to 5% of patients treated with anthracycline will develop HF 15 years after treatment.6 Also, the impact of diastolic dysfunction on outcome and prognosis has been established for many different cardiovascular diseases: it is well known that significant diastolic dysfunction is related to progressive LV dilation, predicts cardiac death, and is associated with a 4-fold increase in mortality in patients with HF.6–10 Thus, clearly, LV systolic and diastolic functions both play an important and complementary role in the risk stratification of patients. Conventionally, E/A ratio is widely used to estimate a degree of diastolic dysfunction, but its curve can be influenced by preload, afterload, and pericardial-limiting factors. The use of tissue Doppler imaging increases the value of these assessments, and its role has been well validated.11–13 This study investigated late cardiac toxicity in long-term acute promyelocytic leukemia (APL) patients treated with the AIDA protocol that combines idarubicin (IDA), mitoxantrone (MTZ), and all-trans retinoic acid (ATRA) evaluating both aspects of the cardiac cycle.
Patients and Methods
Study Population
Between March 1993 and March 2003, 77 adult patients with newly diagnosed APL were treated at the Department of Cellular Biotechnology and Haematology of the University “Sapienza” of Rome. Based on the protocols in use in our Institution, 49 patients diagnosed before January 2000 were treated according to the AIDA 0493 regimen,14 whereas the remaining 28 patients were treated with a slightly modified protocol (AIDA 2000). The latter differed from the original AIDA 0493, for the introduction of a risk-adapted treatment approach in which patients defined as high risk according to Lo Coco and colleagues15 and Sanz and colleagues16 received more intensive chemotherapy during consolidation. However, the cumulative dose of IDA and MTZ did not differ in the AIDA 0493 and AIDA 2000 regimens and consisted in all patients, regardless of risk assessment, of a total dose of 80 mg/m2 and 50 mg/m2 of IDA and MTZ, respectively, corresponding to a cumulative dose of daunorobucin of 520 mg/m2 at an equivalence dose ratio of 1:4 for both IDA and MTZ.17 Median time of exposition to anthracyclines was 5 months (range 3–10 months). A total of 51 patients were in first hematological and molecular remission without clinical signs of heart dysfunction for a median time of 7 years (range 5–14 years) and they were prospectively followed-up. Of these, 34 patients gave informed consent to enter the study and underwent a complete cardiology evaluation (visit, electrocardiography, and echocardiography). There were 16 men and 18 women (aged between 27 and 60). Based on anamnestic records, none of the patients reported a history of heart disease or evidence of valvular disease prior to APL diagnosis and pre-therapy echocardiographic parameters were described as in the normal range in all cases. At the time of their first echocardiogram, no data regarding tissue Doppler imaging (TDI) were available. The results were compared with those obtained in the same study period in a group of 47 healthy controls matched for age and sex. The study was approved by the local institutional review board.
Control Group
The healthy control group for comparison of the standard and TDI echocardiography data consisted of 47 patients (aged 24–62 years) with no clinical history of cardiovascular disease or HF who were recently studied according to the same protocol (Table I shows their clinical characteristics). Patients were also considered healthy if they had history of high blood pressure while taking treatment for not longer than 2 years.
| Characteristics | AIDA Patients (n=34) | Controls (n=47) | P Value |
|---|---|---|---|
| Age (range), y | 48.5 (27–60) | 45.5 (24–62) | NS |
| Sex | |||
| Male | 16 | 25 | NS |
| Female | 18 | 22 | NS |
| Cardiovascular risk factors | |||
| Hypertension | 6 (17%) | 11 (23%) | NS |
| Smoker | 10 (29%) | 14 (30%) | NS |
| Hypercholesterolemia | 6 (17%) | 3 (6%) | NS |
| Diabetes mellitus | 1 (3%) | 1 (2%) | NS |
| Family history of CAD | 10 (30%) | 10 (21%) | NS |
| History of cardiovascular disease | None | None | – |
| Atrial fibrillation | None | None | – |
| Clinical signs of heart failure | None | None | – |
| Systolic blood pressure, mm Hg | 131±19 | 130±21 | NS |
| Heart rate, beats per min | 73±12 | 75±13 | NS |
| Creatinine, mg/dL | 1.1±0.2 | 1.0±0.2 | NS |
| Hemoglobin, g/dL | 13.9±1.9 | 14.1±2.0 | NS |
| Exposition time to chemotherapy, mo | 5 (3–10) | – | – |
| Median follow-up, y | 7 (5–14) | – | – |
- Abbreviations: CAD, coronary artery disease; NS, not significant.
Echocardiography
Echocardiography was performed according to the recommendations of the American Society of Echocardiography. The study was carried out with Toshiba’s Aplio CV (Tustin, CA). LV end-diastolic (EDV) and end-systolic volumes (ESV) were obtained from apical 4-chamber views according to Simpson’s rule. LV ejection fraction (EF) was derived from these volumes and considered abnormal if lower than 50%. Pulsed Doppler examination of the LV inflow was performed with the sample volume placed between the mitral leaflet tips. The following parameters were recorded: peak early (E wave) and atrial (A wave) flow velocities, their ratio E/A, and the E-wave deceleration time. Pulsed TDI was used to record the velocity profile at the lateral mitral annulus and the peak E’ velocity was recorded. The ratio between peak E and E’ lat was calculated. Diastolic function was classified as normal (E/A ratio >1, E/E’ <8), mild (impaired relaxation if E/A ratio was <1), moderate (if ratio between E and A was >1 with E/E’ ratio >15), and severe (restrictive if E/A >2 and E/E’ >15). All the parameters obtained with pulsed wave Doppler and TDI were measured by a single investigator from 3 cardiac cycles and then averaged. Three left atrial (LA) dimensions were obtained to calculate the LA volume. The first (SA1) was measured by 2-dimensional guided M-mode echocardiography obtained in the parasternal short-axis view at the base of the heart, and the second (SA2) and the third (SA3) were obtained measuring the short- and the long-axis dimensions in the apical 4-chamber view at ventricular end-systole. Volume was calculated by the formula π/6 (SA1×SA2×SA3).
Statistical Analysis
Statistical analysis was performed using SPSS software v17 (SPSS, IBM, Armonk, NY). Categoric data are presented as percentages; normally distributed continuous data as mean±standard deviation. The results in the two groups (AIDA patients and healthy controls) were compared using percentile distributions with a nonparametric Mann–Whitney U test. A receiver operating characteristic (ROC) curve analysis was constructed to analyze the diagnostic performance of the commonly utilized diastolic dysfunction tests. Cutoff points for the detection of any kind of diastolic dysfunction were chosen, and sensitivity and specificity were calculated.
Results
Our study included 34 patients with a median age of 48 (range 27–60) years. The baseline characteristics are shown in Table I. They had no known cardiovascular disease at the time of diagnosis of APL, and no differences were found between AIDA and control group in terms of sex, age, cardiovascular risk factors, systolic blood pressure, or heart rate. Clinical symptoms or signs of HF were not observed in any patient, and they were in New York Heart Association class I at the time of echocardiographic examination. All patients were in sinus rhythm at the time of this study.
The LVEDV and LVESV did not differ significantly between the patients treated with anthracycline and the control groups, but a significant difference was found regarding ejection fraction (EF; 56% in patients treated with AIDA protocol vs 59% in the control group, P=.01) (Table II). It should be noted that echocardiographic LV volumes and systolic function in the AIDA group were still within the normal range. In 11 AIDA group patients (32%), some regional wall motion abnormalities (RWMAs) were described by the echocardiographer, but, in 10 cases, the overall systolic function was described as normal and in only 1 case mildly reduced. To better define the pathogenesis of the RWMA, 9 of 11 patients underwent a stress test with myocardial perfusion imaging. A reduced myocardial perfusion was demonstrated in 4 cases. Of these, 2 patients underwent coronary angiography, which excluded the presence of significant epicardial coronary disease. Of the 2 remaining cases with SWM alteration, one already experienced anthracycline-related acute cardiotoxicity that resulted in a stable reduction of EF to 45%, whereas the other patient refused further examinations.
| Echocardiographic Parameters | AIDA (n=34) | Controls (n=47) | P Value |
|---|---|---|---|
| Systolic function | |||
| LVEDV, mL | 112.32 (29.68) | 106.47 (22.59) | .316 |
| LVESV, mL | 45.02 (14.76) | 43.57 (10.78) | .609 |
| EF, % | 56.53 (4.39) | 59.89 (4.24) | .01 |
| RWMA | |||
| Hypokinesia | 11 | 0 | <.001 |
| Akinesia/dyskinesia | 0 | 0 | NS |
| Diastolic dysfunction | |||
| E/A ratio | 1.04 (0.31) | 1.28 (0.29) | .001 |
| Left atrial volume, mL | 52.24 (17.10) | 35.91 (13.11) | <.001 |
| E deceleration time, ms | 218.24 (57.53) | 193.98 (38.89) | .74 |
| E/E’ ratio | 10.04 (4.01) | 5.79 (1.38) | <.001 |
| Mild | 18 | 0 | <.001 |
| Moderate | 0 | 0 | NS |
| Severe | 0 | 0 | NS |
- Abbreviations: EF, ejection fraction; LVEDV, left ventricular end-diastolic volume; LVESV, left ventricular end-systolic volume; NS, not significant; RWMA, regional wall motion abnormalities. Bold values indicate significance.
Overall, LV diastolic function parameters were significantly impaired in the ALP group. More than half of the population (52%) treated with the AIDA protocol presented with mild diastolic dysfunction compared with none in the normal group (P<.001). We observed a decreased E/A ratio (1.04 in the AIDA group vs 1.28 in the control group, P=.001) and an elevated E/E’ lateral ratio (10.04 in the AIDA group vs 5.79 in the control group, P<.001).
These patients also ended up with a significantly greater left atrial volume (LAV, 52 mL vs 35 mL, P=<.001).
E/e’ was the best index for the recognition of diastolic dysfunction in the long-term survivor group, with an area under the ROC curve of 0.81 (95% confidence interval [CI], 0.71–0.92). Considering the cutoff point of 8, an elevated E/e’ value has 62% sensitivity and 96% of specificity in recognizing signs of diastolic dysfunction in these patients. The LA volume has an area under the ROC curve of 0.77 (95% CI, 0.67–0.87), and given a cutoff point of 40.5 mL, sensitivity and specificity were 79% and 68%, respectively.
Discussion
To the best of our knowledge, the treatment of adult ALP long-term survivors who have had anthracycline-related cardiotoxicity has not been systematically studied, and these patients represent a population at increased risk of premature cardiovascular disease. Some of these patients, including long-term survivors of ALP, were reported to have undergone heart transplantation;18 however, a lower cumulative dose of anthracycline seems to maintain high cure rates, promoting the reduction of long-term cardiotoxicity.19 An interesting finding of our investigation is that, at the time of the study, clinical symptoms of HF were not observed in any patients and they still had an EF within the normal range. Although the median follow-up in our series (7 years) may still be considered relatively short to assess the very long-term cardiac toxicity, our findings suggest that the use of less cardiotoxic anthracyclines such as idarubicin and mitoxantrone in combination with ATRA may reduce the occurrence of late cardiac adverse events as compared with daunorubicin. Detection of subclinical cardiotoxicity has always been difficult for cardiologists, and a classic evaluation of these patients includes only an echocardiographic estimation of the LV systolic function. The incidence of cardiotoxicity in patients treated with anthracycline, as diagnosed by endomyocardial biopsy (EMB), is approximately 85%,20–22 and a good correlation exists between cumulative dose of anthracycline and EMB grade.23 However, given this invasive nature, the role of EMB in patients treated with these chemotherapeutic agents is reserved for only select cases.24 Numerous cross-sectional and longitudinal studies indicate that patients who received high-dose anthracycline, especially >300 mg/m2, are at risk for exhibiting subclinical cardiovascular dysfunction and clinically significant cardiomyopathy,25–27 and the incidence of severe echocardiographic abnormalities increases with the duration of follow-up.28,29 The results of the present study clearly show that subclinical cardiomyopathy characterized by systolic and diastolic dysfunction are frequently present also in patients receiving agents such as IDA and MTZ that are contained in the AIDA protocol, as indicated by the detection of RWMAs and/or impaired relaxation in 32% and 52% of our cases, respectively. Eleven patients presented with segmental kinetic alterations, and similar changes in regional cardiac function have been recently described even in children treated with anthracycline with normal EF, especially in the septal area,30 and considered the first sign of cardiotoxicity. Assessment of the diastolic function has a key role in our understanding of the physiologic damage caused by anthracyclines. To detect anthraycline-induced cardiotoxicity, assessment of diastolic function is being recommended in addition to systolic parameters, since the evaluation of the E/A ratio alone did not appear enough. Accordingly with other authors,31 the E/A ratio significantly decreases those treated with anthracycline, but it can happen in a lot of other clinical situations, such as hypertension, diabetes, or ischemic heart disease, and this parameter can also be influenced by preload, afterload, and pericardium effects. However, the use of TDI may be helpful in making an early diagnosis of chemotherapy-induced cardiomyopathy.32 Our study supports the following findings: late subclinical cardiotoxicity may be easily detected with TDI and the E/E’ ratio appears to be the best parameter to identify earlier abnormal filling pressures in patients exposed to anticancer therapy. An E/E’ ratio of 10 may not appear particularly elevated but, in select cardiovascular clinical scenarios, it seems to be associated with a worse outcome,33 also considering the relatively young age of the studied population. An enlargement of the left atrium reflects the history of chronically increased left atrial pressures.34 As expected, as effect of chronic exposure to left ventricular high filling pressures, a higher LA volume was also found in the AIDA group. Even if the E/E’ was the best index in the detection of advanced diastolic dysfunction, it has been thought that, because the LA volume reflects a “chronic” exposure to abnormal LV diastolic function, the period of our follow-up was not enough to permit the atrium to become extremely dilated. Some authors conclude that the LA volume provides a sensitive expression of the severity of diastolic function and that appears to be a useful index of cardiovascular risk.35 We can translate this to our group of patients.
Limitations
Our study is small and we examined relatively young people, wherein it is known that the incidence of diastolic dysfunction increases with age. A degree of diastolic dysfunction can be identified up to 40% in people older than 65 years without risk factors and it can make their evaluation more difficult.36 Potentially, new software such as tissue tracking37 could be helpful in this clinical scenario. It should also be noted that in our study, the LA volume was not indexed to body surface area. It would also be interesting in the future to study the effects of anthracycline on B-type natriuretic peptide or N-terminal pro–B-type natriuretic peptide measurements in this group of patients, which can also provide additional information for their risk stratification.38
Conclusions
Cardiac toxicity due to anthracycline therapy is often frequent and changes in diastolic function, when evaluated with a multiparameter approach, are helpful in the detection of subclinical anthracycline cardiotoxicity in long-term cardiac follow-up. The importance of this approach has to be underlined despite the preservation of systolic ventricular function, indicating that APL long-term survivors should be regularly monitored to identify those at risk for cardiac dysfunction and to ensure earlier therapeutic intervention for those with threatened HF.
