Although left ventricular (LV) dysfunction occurs not uncommonly in the course of cancer therapy, little is known about its natural history and prognostic impact on patients. To investigate the incidence, predictors, and impact on survival of LV systolic dysfunction and recovery during cancer therapy, we conducted a retrospective cohort observational study over 1 year at the University of Texas MD Anderson Cancer Center. We enrolled patients with a decrease in ejection fraction by echocardiography to <50% while undergoing cancer therapy from January 2009 to December 2009. We collected and analyzed their chart data. Of 7,648 patients with echocardiograms in 2009, 366 (4.8%) had ejection fraction <50% and 104 met study criteria. LV systolic dysfunction was associated with cardiotoxic therapy in 53 patients (51%). Recovery occurred in 57 patients (55%) and was independently predicted by younger age, smaller left atrial volume index, and lower B-type natriuretic peptide. At last follow-up, 69 patients (66%) were dead, and 35 (34%) were alive. There was a 20% advantage in 2-year survival among patients with LV systolic recovery compared with those without (95% confidence interval 4% to 41%, p = 0.02). In this retrospective study, LV systolic dysfunction recovery occurred in over half of the patients, appeared independent of cardiotoxic etiology, and associated with a 20% survival benefit at 2 years. Multivariable predictors of recovery are younger age, a small left atrial volume index, and lower B-type natriuretic peptide.
In the noncancer population, myocardial recovery has been frequently observed in select cardiomyopathies, such as tachyarrhythmia induced, endocrine, nutritional, viral, catecholamine induced, and in patients with heart failure (HF) treated with mechanical circulatory support. Multiple predictors of myocardial recovery have been identified in noncancer patients with new onset left ventricular (LV) systolic dysfunction, such as LV end-diastolic volume, baseline LV ejection fraction (EF), and systolic blood pressure. In contrast, LV systolic dysfunction in patients with cancer has mostly been studied from the prism of direct cytotoxic effects of cardiotoxic chemotherapy and thus, recovery has been shown to occur less often.
Published data suggest that angiotensin-converting enzyme inhibitors (ACEIs) and β blocker therapy and earlier intervention are associated with better chances of LV function recovery, although large-scale validation studies are lacking. The impact on survival of patients who recover from LV systolic dysfunction during cancer therapy has also not been fully studied. Herein, we sought to investigate the incidence, predictors, and impact on survival of recovery from LV systolic dysfunction in patients with cancer during cancer therapy.
Methods
With Institutional Review Board approval, we retrospectively queried the MD Anderson echocardiography laboratory database and identified sequential patients with echocardiograms performed during the year of 2009 whose EFs were <50%. We then excluded those without previous documentation of EF >50% before initiation of cancer therapy or subsequent follow-up echocardiograms. The date of LV systolic dysfunction diagnosis was defined as that of the first abnormal echocardiogram in our system, subsequent to any imaging modality or documented office note recording a normal EF, which in many cases occurred before 2009. Patients in the cohort were in- or outpatients, age ≥18 years with advanced cancer actively receiving cancer therapy. Patients were considered to have received cardiotoxic therapy if they had been treated with agents known to be associated with a >5% risk of LV dysfunction at currently employed doses (anthracyclines and trastuzumab).
All echocardiograms were reviewed by 2 independent investigators blinded to the sequence and dates of the echocardiograms, who re-measured all parameters according to the published guidelines of the American Society of Echocardiography. We measured LVEF using the biplane method of disks (modified Simpson’s). We included those with EF confirmed as <50% and with previous EF >50% by ≥5 percentage points. All subsequent echocardiograms were similarly measured. Any discordance between the readers was resolved by consensus. Recovery was based on the last echocardiogram.
Echocardiographic 2-dimensional parameters including LVEF, left atrial volume index (LAVI), LV hypertrophy, LV end-diastolic dimension and volume, LV mass and LV mass index, and the presence of valvular heart disease were measured according to the published recommendations and documented. Valvular heart disease was subclassified into mild, moderate, or severe whether regurgitant or stenotic lesions were present. Diastolic function was assessed using Doppler and tissue Doppler techniques as published by the American Society of Echocardiography and classified as grade 1 (normal), 2 (impaired relaxation), 3 (pseudonormal), and 4 (restrictive).
LV systolic dysfunction recovery was defined as an increase in EF of ≥10% points from the lowest documented EF and absent if EF did not increase by ≥10% points. We also separately analyzed recovery defined as EF that returned to a value ≥50%. We further subclassified recovery into 3 groups: (1) full recovery—defined as EF that increased ≥10% points and was ≥50% at the last echocardiogram; (2) partial recovery—defined as EF that increased by ≥10% points but remained <50%; and (3) no recovery—defined as <10% point increase in EF and EF <50%.
Clinical data were retrospectively collected from the time of diagnosis of LV systolic dysfunction, with special attention to New York Heart Association class and HF symptoms, presence of atrial fibrillation, hemodynamic shock (defined as low blood pressure requiring admission to intensive care unit and specific therapy for hypotension), history of any coronary artery disease, diabetes mellitus, hypertension, and tobacco use. Pulmonary hypertension was defined as mean pulmonary artery pressure >25 mm Hg at rest of >30 mm Hg at exercise. Cancer therapy at time of LV systolic dysfunction was also documented, with specific attention to use of chemotherapy agents and radiation therapy to the chest. Laboratory data were documented when obtained at the time of LV systolic dysfunction, specifically B-type natriuretic peptide (BNP) levels, serum troponin-I levels, and serum creatinine. Additionally, attempts at determination of etiology of HF were also investigated and documented.
HF therapy instituted at the time of LV systolic dysfunction was documented according to the major family of agents, that is ACEIs, β blockers, diuretics, aldosterone receptor antagonist, and so on. Patients were treated according to the published guidelines to the extent that hemodynamic status and concomitant cancer therapies permitted. Follow-up was included retrospectively from the date of LV systolic dysfunction diagnosis until November 2011. Date and cause of death were obtained from the medical records. For those alive by November 2011, clinical status was documented.
Categorical variables are expressed as number and percent and continuous variables as mean (SD or median [interquartile range]) when the distribution was not normal. Comparisons between patients with and without recovery were done with the Fisher’s exact test for categorical variables and the Wilcoxon rank-sum test for continuous variables. Potential predictors of recovery were those available at baseline. Univariable logistic regression analyses were performed to evaluate the association between a patient variable and a recovery. The association was expressed as odds ratio and its 95% confidence interval (CI). Variables with a p ≤0.2 in univariable analysis and some prespecified variables (baseline LVEF, LV end-diastolic dimension, and the use of cardiotoxic chemotherapy) were chosen for multivariable logistic regression analysis, where variables with a p <0.05 were considered independently associated with recovery. Survival was described with the Kaplan-Meier method, and differences between patients with and without recovery were tested with the log-rank test. To evaluate the discriminative ability of LAVI to differentiate between recovery and no recovery, we performed receiver-operating characteristic analysis and quantified the ability with the c-statistic. SAS 9.2 (SAS Institute Inc., Cary, North Carolina) was used for all statistical analyses.
Results
Throughout 2009, of 7,648 consecutive patients with cancer who underwent echocardiographic evaluation, 366 (4.8%) had EF <50%. Of these, 104 patients had previously documented normal EF and at least 1 subsequent echocardiogram, meeting inclusion criteria. Baseline characteristics are listed in Table 1 .
| Variable | Recovery (n = 57) | No Recovery (n = 47) | Univariable OR for Recovery (95% CI) ∗ | p Value |
|---|---|---|---|---|
| Age (yrs) | 52 (±16) | 58 (±16) | 0.9 (0.9–1.0) | 0.05 |
| Men | 26 (46) | 19 (40) | 1.2 (0.6–2.7) | 0.6 |
| Malignancy | ||||
| Leukemia/lymphoma | 30 (53) | 21 (45) | 1.4 (0.6–3.0) | 0.4 |
| Other | 27 (47) | 26 (55) | 1 | |
| NYHA class | ||||
| 0 | 22 (39) | 11 (23) | 1 | 0.4 |
| I | 17 (30) | 16 (34) | 0.5 (0.2–1.4) | 0.4 |
| II | 10 (17) | 10 (21) | 0.5 (0.2–1.6) | |
| III–IV | 8 (14) | 10 (21) | 0.4 (0.1–1.3) | |
| Atrial fibrillation | 18 (32) | 14 (30) | 1.1 (0.5–2.5) | 0.8 |
| Atrial flutter | ||||
| Current or past | 17 (30) | 12 (26) | 1.2 (0.5–2.9) | 0.7 |
| Unknown | 1 (2) | 2 (4) | — | |
| Sepsis | 18 (32) | 8 (17) | 2.2 (0.9–5.8) | 0.1 |
| Smoker | 17 (30) | 20 (43) | 0.6 (0.3–1.3) | 0.2 |
| Coronary artery disease | 12 (21) | 17 (36) | 0.5 (0.2–1.1) | 0.1 |
| Diabetes mellitus | 6 (11) | 9 (19) | 0.5 (0.2–1.5) | 0.2 |
| Hypertension | 26 (46) | 22 (47) | 1.0 (0.4–2.1) | 0.9 |
| Cardiotoxic chemotherapy | 42 (74) | 39 (83) | 0.6 (0.2–1.5) | 0.3 |
| Anthracyclines | 30 (53) | 23 (49) | 1.2 (0.5–2.5) | 0.7 |
| Trastuzumab | 5 (9) | 2 (4) | 2.2 (0.4–12) | 0.4 |
| Cyclophosphamide | 18 (32) | 18 (38) | 0.7 (0.3–1.7) | 0.5 |
| Chest radiation | 15 (26) | 10 (21) | 1.3 (0.5–3.3) | 0.6 |
| Pulmonary embolism | 11 (20) | 10 (21) | 0.9 (0.3–2.4) | 0.8 |
| Diastolic HF | ||||
| Grade 1 | 15 (26) | 13 (28) | 1 | |
| Grade 2 | 7 (12) | 9 (19) | 0.7 (0.2–2.3) | 0.9 |
| Grade 3 | 6 (11) | 7 (15) | 0.7 (0.2–2.7) | |
| Grade 4 | 17 (30) | 14 (30) | 1 (0.3–2.9) | |
| HF diagnosis | ||||
| Chemo-induced | 30 (53) | 23 (49) | 1.2 (0.5–2.7) | 0.9 |
| Other cause | 19 (33) | 17 (36) | 1 | |
| Unknown | 8 (14) | 7 (15) | — | |
| HF at follow-up † | 19 (33) | 31 (66) | 0.3 (0.1–0.6) | 0.001 |
| BNP (pg/ml) | 577 (176–1,653) | 1,332 (307–2,817) | 0.9 (0.9–0.9) ‡ | 0.04 |
| Troponin I (ng/dl) | 0.13 (0.03–0.51) | 0.12 (0.04–0.4) | 1.0 (0.9–1.0) § | 0.8 |
| Creatinine (mg/dl) | 0.8 (0.6–1.1) | 0.9 (0.7–1.2) | 0.6 (0.3–1.4) | 0.3 |
| β blocker | 49 (86) | 38 (81) | 1.5 (0.5–4.1) | 0.5 |
| ACEI/ARB | 34 (60) | 33 (70) | 0.6 (0.3–1.4) | 0.3 |
| Spironolactone | 4 (7) | 16 (34) | 0.1 (0.05–0.5) | 0.001 |
| Statins | 18 (32) | 18 (38) | 0.7 (0.3–1.7) | 0.5 |
∗ Per 1 unit increase of continuous variable.
† Not considered for multivariable analysis as it was measured after baseline.
The most common malignancies were leukemia (30%), lymphoma (19%), genitourinary cancer (11%), breast (10%), lung (10%), sarcomas (6%), and others (17%). In this cohort, 32% and 68% of patients had stage III and IV cancer, respectively.
Cancer therapy consisted of anthracyclines, 48%; cyclophosphamide, 35%; vascular endothelial growth factor inhibitors, 26%; and trastuzumab, 7%. In this cohort, 25 patients (24%) received radiation therapy to the chest. Cardiotoxic therapy had been prescribed in 60 patients (58%).
The mean documented EF before chemotherapy initiation was 61% ± 9%. Regional wall motion abnormalities were described in 41 patients (40%). Information on diastolic function was available in 91 patients (88%) and was abnormal in 73 patients (80%), with 22 patients (30%) presenting with Grade 4 diastolic dysfunction. Echocardiographic parameters at presentation of LV systolic dysfunction of patients with and without recovery are listed in Table 2 .
| Echocardiographic | EF Recovery | OR ∗ | CI ∗ | p Value | |
|---|---|---|---|---|---|
| Yes | No | ||||
| Ejection fraction at baseline | 34 ± 8.5 | 33 ± 8.5 | 1.01 | 0.97–1.06 | 0.6 |
| Left atrial volume index (mL/m 2 ) | 27 ± 8.5 | 33 ± 10 | 0.93 | 0.88–0.97 | 0.001 |
| LV end-diastolic volume (mm 3 ) | 105 ± 34 | 114.5 ± 39 | 0.99 | 0.98–1.00 | 0.2 |
| LV mass index (g/m 2 ) | 100 ± 28 | 109 ± 36.5 | 0.99 | 0.98–1.00 | 0.1 |
| LV hypertrophy | |||||
| No | 35 (71) | 30 (77) | 1.0 | ||
| Mild to moderate | 14 (29) | 9 (23) | 1.3 | 0.5–3.5 | 0.6 |
| Pulmonary hypertension | |||||
| No | 25 (45) | 16 (35) | 1.0 | ||
| Mild | 32 (38) | 13 (28) | 1.0 | 0.4–2.6 | 0.2 |
| Moderate-to-severe | 10 (18) | 17 (37) | 0.4 | 0.1–1.1 | |
LV systolic dysfunction was thought to be chemotherapy induced in 53 patients (51%), other causes such as sepsis and arrhythmia in 36 patients (35%), and unknown in 15 patients (14%). Evaluation of coronary artery disease was carried out in 38 patients (35%), none of whom had severe enough coronary artery disease to account for their cardiomyopathy.
At the last echo, LV systolic dysfunction recovery had occurred in 57 patients (55%). The mean EF of those with recovery versus no recovery was 51% ± 0.8% and 31% ± 0.9% (p <0.0001), respectively. In the recovery group, LVEF was >55% in 20 patients (35%), 50% to 54% in 13 patients (23%), 45% to 49% in 15 patients (26%), and 40% to 44% in 9 patients (16%).
Baseline characteristics of recovered versus not were similar ( Table 1 ), except for LAVI that was significantly lower in those who recovered (26 ± 8 vs 33 ± 10 ml/m 2 , p = 0.0002). HF treatment was administered to 102 patients (98%). Interestingly, there was a significantly higher use of spironolactone in those who did not recover compared with those who did (35% vs 7%, respectively, p = 0.0004). Finally, there was no difference in rates of recovery between those with exposure to cardiotoxic therapy and those with other etiologies of LV systolic dysfunction (61% vs 38.5%, respectively, odds ratio 1.4, 95% CI 0.64 to 3.06, p = 0.39).
When we analyzed recovery as return of LV function to EF ≥50%, 33 patients (32%) recovered and 71 patients (68%) did not. Further, when stratified into full (EF increase ≥10% points and EF ≥50%), partial (EF increase ≥10% points and EF <50%), and no recovery (EF increase <10% and points and EF <50%), there were 33 (32%), 24 (23%), and 47 (45%) patients, respectively.
We conducted uni- and multivariable analyses to determine baseline predictors of recovery. Results of the univariable analysis are listed in Tables 1 and 2 . Multivariable analysis confirmed age, LAVI, and BNP as independent predictors of recovery ( Table 3 ).
