Patients with acute leukemia (AL) have a higher rate of congestive heart failure than patients with other cancers. AL may predispose to cardiac dysfunction before chemotherapy because of high cytokine release or direct leukemic myocardial infiltration. The aims of this study were to evaluate whether AL is associated with abnormalities of myocardial structure and function before chemotherapy and to identify possible risk factors associated with these myocardial changes.
Using an echocardiographic database, 76 patients with AL and 76 patients without cancer matched for age, gender, hypertension, and the presence of diabetes were retrospectively selected. Subsequently, to assess the effect of a nonhematologic malignancy, 28 women in each group were matched with women with breast cancer. Left ventricular (LV) mass, volumes, ejection fraction, and global longitudinal strain (GLS) were measured before chemotherapy.
The patients were predominantly male (63%), with a median age of 51 years, and had low prevalence of cardiovascular risk factors. Despite similar LV ejection fractions, patients with AL had higher LV mass and volumes and lower GLS (−19.3 ± 2.7% vs −20.9 ± 1.9%, P < .001) than patients without cancer. Similarly, GLS was lower in women with AL compared with women with breast cancer or without cancer. Among patients with AL, high body mass index, low LV ejection fraction, and a small number of circulating lymphocytes were all independently associated with low GLS.
Patients with AL had higher LV volumes and lower GLS than patients without cancer and lower GLS than patients with breast cancer, suggesting that AL by itself may be associated with these cardiac alterations.
The authors assessed whether AL is associated with cardiac alterations before chemotherapy.
Patients with AL had larger LV mass and volumes than patients without cancer.
Patients with AL had lower GLS than patients with breast cancer or without cancer.
Among patients with AL, BMI, LVEF, and lymphocyte count were correlated with GLS.
This finding suggests that AL by itself is associated with these cardiac alterations.
Although the overall incidence rate of leukemia is increasing at about 1.3% per year, leukemia mortality has declined by approximately 1.0% per year in recent decades. The larger number of acute leukemia (AL) survivors has raised the awareness of associated comorbidities, in particular cardiovascular disease. Recently, higher rates of symptomatic heart failure and cardiac death after anthracycline chemotherapy (13%) have been reported in patients with AL compared with other type of cancer.
Patients with AL typically have aggressive treatment regimens, containing high doses of anthracyclines administered over several consecutive days, which may explain their higher risk for congestive heart failure. However, other AL-related factors may also contribute to the observed worse clinical outcomes, such as prechemotherapy cardiac damage due to a robust systemic inflammatory response and/or direct myocardial infiltration by leukemic cells.
We hypothesized that AL per se may be associated with cardiac abnormalities before chemotherapy. Therefore, we sought to investigate whether AL is associated with abnormalities of myocardial structure and function before chemotherapy using echocardiography and to identify possible clinical and laboratory characteristics associated with these myocardial changes.
Using an echocardiographic database from January 2002 to October 2012, we retrospectively selected 102 consecutive patients with leukemia (≥18 years of age) and with available two-dimensional echocardiographic images obtained at Massachusetts General Hospital before anthracycline-based chemotherapy. All studies were performed routinely to evaluate the cardiac function of leukemia patients before anthracycline-based chemotherapy. We excluded patients with diagnoses of chronic leukemia ( n = 9), histories of known cardiomyopathy or congestive heart failure ( n = 7), prior treatment with anthracycline-based chemotherapy ( n = 5), or poor image quality ( n = 5). None had sepsis or congenital or valve heart disease. None received any chemotherapeutic agent or radiotherapy (chest or neck) before undergoing transthoracic echocardiography. Seventy-six patients with AL and 76 patients without cancer and with no known cardiac disease were matched (1:1) for age, gender, hypertension, and diabetes. To assess the effect of a nonhematologic malignancy, 28 women from each group (1:1:1) were subsequently matched for the same variables with women with breast cancer ( Figure 1 ). The institutional review board of Massachusetts General Hospital approved the study protocol.
Echocardiography and Strain Analysis
Transthoracic echocardiograms were acquired using commercially available machines (Vivid 7 or E9 [GE Healthcare, Milwaukee, WI]; iE33 [Philips Medical Systems, Andover, MA]). The images were recorded digitally with grayscale harmonic imaging, saved in compressed Digital Imaging and Communications in Medicine format, and analyzed offline. Apical four-, three-, and two-chamber views were stored in cine-loop format. Images were obtained between 40 and 80 frames/sec; when transformed into Digital Imaging and Communications in Medicine format, however, they were compressed to 30 frames/sec. An observer blinded to the different groups analyzed the echocardiographic images. Left ventricular (LV) end-diastolic volume, LV end-systolic volume, and LV ejection fraction (LVEF) were calculated from two-dimensional images, in two- and four-chamber views, using a modified Simpson biplane method. LV mass was calculated from the parasternal short-axis view using the cube formula. LV volumes and mass were analyzed as absolute values and were also indexed to body surface area. Global longitudinal strain (GLS) by speckle-tracking two-dimensional echocardiography was measured using an offline independent analysis program (2D Cardiac Performance Analysis; TomTec Imaging Systems, Munich, Germany) on Digital Imaging and Communications in Medicine stored images. LV endocardial border was traced at end-diastole from the three standard apical views. The quality of tracking was reviewed and manually adjusted for optimal tracking by visual assessment. The endocardium tracking of the three-chamber view was suboptimal, even after multiple attempts to adjust the tracing, in 18 patients. We decided to exclude all three-chamber view from our strain analysis as previously performed. Thus, GLS was calculated by averaging the values from the two- and four-chamber views.
Continuous variables are presented as mean ± SD or as median and interquartile range, and categorical variables are presented as percentages. The χ 2 test or Fisher exact test was used to compare categorical variables. Differences in continuous variables between patients with AL and those without cancer were assessed using Student’s t test or the Wilcoxon rank sum tests. Differences in continuous variables among the three groups of women with AL, with breast cancer, and without cancer were assessed by using one-way analysis of variance followed by Bonferroni correction. Correlations between GLS and clinical, laboratory, and echocardiographic variables were estimated using Pearson correlation coefficients or Spearman ρ test for nonparametric values. A linear regression was also performed to determine independent clinical and echocardiographic parameters associated with GLS. All variables reaching P values of <.25 on univariate analysis were included in the multivariate model. Intraobserver and interobserver variability of GLS was determined by using Bland-Altman analysis (including mean bias and 95% limits of agreement) in a random sample of 20 patients. All analyses were performed using Stata release 13 (StataCorp, College Station, TX). P values < .05 were considered to indicate statistical significance, and all reported P values were two tailed.
Patients with AL versus Patients without Cancer
The clinical, demographic, and echocardiographic characteristics are summarized in Table 1 . The median age of the patients was 51 years, 63% were men, and there was a low prevalence of cardiovascular risk factors. Patients without cancer had a higher rate of dyslipidemia and were more likely to smoke and be on statins than patients with AL.
|( N = 152)||( n = 76)||( n = 76)|
|Age (y)||50 (43–60)||51 (44–59)||51 (38–59)||.38|
|Female sex||56 (37)||28 (37)||28 (37)||≥.999|
|BMI (kg/m 2 )||27 (24-31)||26 (24-29)||27 (24-31)||.39|
|Cardiac risk factors|
|Diabetes||10 (7)||5 (7)||5 (7)||≥.999|
|Hypertension||10 (7)||5 (7)||5 (7)||≥.999|
|Dyslipidemia||28 (18)||24 (32)||4 (5)||<.001|
|Current smoking||18 (12)||13 (17)||5 (7)||.045|
|Prior CAD||7 (5)||4 (5)||3 (4)||≥.999|
|β-blockers||19 (13)||12 (16)||7 (9)||.22|
|ACE inhibitors/ARBs||6 (4)||3 (4)||3 (4)||≥.999|
|Statins||31 (20)||22 (29)||9 (12)||.009|
|Relative wall thickness||0.4 ± 0.1||0.4 ± 0.1||0.4 ± 0.1||.716|
|LV mass (g)||149 ± 44||141 ± 33||158 ± 51||.014|
|LV mass index (g/m 2 )||79 ± 22||74 ± 14||84 ± 27||.008|
|LVEDV (mL)||108 ± 28||101 ± 29||116 ± 32||.001|
|LVEDVI (mL/m 2 )||58 ± 16||53 ± 10||62 ± 19||.001|
|LVESV (mL)||41 ± 13||38 ± 10||45 ± 14||.001|
|LVESVI (mL/m 2 )||22 ± 7||20 ± 5||24 ± 8||.002|
|LVEF (%)||62 ± 5||62 ± 5||62 ± 6||.34|
|GLS (%)||−20.1 ± 2.4||−20.9 ± 1.9||−19.3 ± 2.7||<.001|
LV mass and volumes were slightly higher in patients with AL ( Table 1 ). Relative wall thickness was similar between the two groups. Although LVEF was similar between patients with AL and those without cancer (62 ± 6% vs 62 ± 5%, P = .34), GLS was lower in patients with AL compared with patients without cancer (−19.3 ± 2.7% vs −20.9 ± 1.9%, P < .001; Figure 2 ). Ten patients with AL and three control subjects had LVEFs < 55%. The LVEF range was 49% to 76% in patients with AL and 52% to 73% in control subjects. After excluding patients with LVEFs < 55%, GLS was still reduced in patients with AL compared with those without cancer (−19.5 ± 3.3% vs −20.8 ± 2.1%, P = .001). Sixteen patients with AL (21%) and two control subjects had GLS < −17.5% in our study, and only patients with AL had GLS < −16% (10 patients). Patients with acute myeloid leukemia were more likely to have GLS < −17.5% than patients with acute lymphocytic leukemia (12 [67%] vs six [33%], P = .023). Small pericardial effusions were present in two patients with AL. GLS was −18.7% in one patient and −15.7% in the other patient. In the univariate analysis, GLS was significantly associated with body mass index (BMI; β = 0.078, P = .049), LV mass indexed to body surface area (β = 0.018, P = .038), LVEF (β = −0.090, P = .014), and leukemia (β = 1.527, P < .001). After adjustment for potential confounders including gender, BMI, β-blocker use, LV mass index, LVEF, and leukemia, only LVEF (β = −0.075, P = .035) and leukemia (β = 1.36, P < .001) remained significantly and independently associated with GLS.
Patients with AL versus Women with Breast Cancer and without Cancer
Patient characteristics are summarized in Table 2 . Except for women without cancer, who had significantly more dyslipidemia than women with AL and with breast cancer, there was no difference in the clinical characteristics of the three groups.
|Variable||Women without cancer||Women with breast cancer||Women with AL||P|
|( n = 28)||( n = 28)||( n = 28)|
|Age (y)||51 (45–59)||52 (45–60)||51 (44–60)||.94|
|BMI (kg/m 2 )||25 (23–29)||26 (23–31)||27 (24–31)||.46|
|Cardiac risk factors|
|Diabetes||2 (7)||2 (7)||2 (7)||≥.999|
|Hypertension||4 (14)||4 (14)||4 (14)||≥.999|
|Dyslipidemia||12 (41)||5 (17)||1 (3)||.002|
|Current smoking||1 (4)||1 (4)||2 (7)||.77|
|Prior CAD||2 (7)||2 (7)||2 (7)||≥.999|
|β-blockers||5 (18)||5 (18)||2 (7)||.42|
|ACE inhibitors/ARBs||1 (4)||1 (4)||1 (4)||≥.999|
|Statins||9 (32)||2 (7)||5 (18)||.057|
|Relative wall thickness||0.4 ± 0.1||0.4 ± 0.1||0.4 ± 0.1||.40|
|LV mass (g)||117 ± 20||125 ± 25||126 ± 36||.42|
|LV mass indexed to BSA (g/m 2 )||68 ± 10||67 ± 20||67 ± 22||.97|
|LVEDV (mL)||87 ± 18||90 ± 13||96 ± 13||.22|
|LVEDVI (mL/m 2 )||51 ± 9||48 ± 9||51 ± 15||.58|
|LVESV (mL)||33 ± 7||35 ± 9||36 ± 12||.61|
|LVESVI (mL/m 2 )||20 ± 4||19 ± 5||19 ± 7||.77|
|LVEF (%)||61 ± 4||62 ± 6||63 ± 7||.81|
|GLS (%)||−21.5 ± 2.1||−21.2 ± 2.2||−19.4 ± 3.0||.005|