Measuring left ventricular (LV) global longitudinal strain (GLS) is recommended in screening of long-term cancer survivors for cardiotoxicity. However, there are limited data on GLS in this setting, in particular in survivors with apparently normal LV function without risk factors of impaired GLS. In the present study, we measured GLS in 191 adult survivors of childhood lymphoma or acute lymphoblastic leukemia, with normal LV ejection fraction and fractional shortening (FS) and without known hypertension, diabetes mellitus, myocardial infarction, or stroke. We compared GLS in the survivors with 180 controls. Mean GLS was −19.0 ± 2.2% in the survivor group and −21.4 ± 2.0% in the controls (p <0.001). Impaired GLS, defined as mean – 1.96 SDs in the control group, occurred in 53 of 191 survivors (28%). We included survivors with impaired LV ejection fraction and/or FS or traditional risk factors (n = 231 in all) in multiple regression analyses to explore associations with previous cancer treatment. Survivors treated with mediastinal radiotherapy had an odds ratio of impaired GLS of 5.2 (95% confidence interval 2.2 to 12) compared with other survivors. Survivors treated with cumulative anthracycline doses >300 mg/m 2 had an odds ratio of 4.8 (95% confidence interval 1.7 to 14) of impaired GLS. In conclusion, this study demonstrates a high proportion of LV dysfunction assessed by GLS in apparently healthy adult survivors of childhood cancer. Impaired GLS was associated with previous exposure to mediastinal radiotherapy and high doses of anthracyclines. The prognostic role of measuring GLS in this specific patient population should be examined in prospective studies.
More than 80% of patients diagnosed with cancer in childhood currently survive >5 years, but the survivors face an increased risk of late mortality and morbidity with cardiovascular disease being the most frequent noncancer cause of mortality. Early detection and treatment of cardiotoxicity may improve outcomes, but cardiotoxicity often is asymptomatic and may not be detected by conventional echocardiographic screening. Left ventricular (LV) strain analyses are currently recommended in the echocardiographic follow-up of cancer survivors, and LV global longitudinal strain (GLS) is the preferred parameter. There is more evidence for its use during and shortly after cancer therapy, than in long-term survivors. In survivors of childhood cancer, small studies (from 19 to 111 patients) have shown lower LV strain in survivors than controls. Recently, a large study (1,820 patients) demonstrated a high prevalence of impaired GLS in adult survivors of childhood cancer, compared with normative values. This study included survivors with impaired LV ejection fraction (EF), diagnosed cardiomyopathy, hypertension, or high fasting glucose. Lower GLS has previously been observed in patients with uncomplicated diabetes and in hypertensive patients without LV hypertrophy. There are but little data available on the diagnostic yield of measuring GLS in survivors with apparently normal LV function and no other risk factors of developing impaired GLS than cardiotoxic cancer treatment. We have previously reported data on LV function in long-term adult survivors of childhood lymphoma and acute lymphoblastic leukemia. In the present study, our objectives were first to test the hypothesis that GLS is useful in detecting LV dysfunction in survivors with apparently normal LV systolic function and second to evaluate the relation between GLS and risk factors in the survivors, including previous cardiotoxic treatment.
Methods
We invited long-term survivors of childhood lymphoma or acute lymphoblastic leukemia to participate in 2 cross-sectional studies of health status including late cardiac effects, as previously reported. Eligible survivors were >18 years, diagnosed >5 years earlier with acute lymphoblastic leukemia before the age of 16 years or Hodgkin or non-Hodgkin lymphoma before 18 years. Survivors of leukemia and non-Hodgkin lymphoma were from the southeastern part of Norway (about half the population), whereas survivors of Hodgkin lymphoma were from all parts of Norway. The study participants answered a questionnaire including self-reported data on hypertension, diabetes, or cardiovascular disease. They then attended a 2-day visit at the Oslo University Hospital, Rikshospitalet, Oslo, Norway, where a clinical examination was performed, including a comprehensive echocardiographic examination. All survivors attending echocardiography were assessed for the present study. Detailed data on previous cancer treatment, including cumulative anthracycline doses and mediastinal radiotherapy (RT) doses, were obtained from medical records. Survivors with a history of secondary cancer treated with potentially cardiotoxic treatment were excluded from the present study. We randomly selected a comparison group from the Nord-Trøndelag Health Study. In this Norwegian population-based study, echocardiography was performed in 1,296 subjects without known hypertension, diabetes, or cardiovascular disease. Controls were matched on group level for age, gender, body weight, body surface area, and systolic blood pressure. All study participants provided written informed consent. The Regional Committee for Medical and Health Research Ethics approved the studies.
We performed echocardiography with assessment of LV function according to international guidelines. All studies in survivors were obtained using Vivid 7 or E9 scanners and in controls using Vivid 7 scanners from GE Healthcare (Horten, Norway). We used Simpson’s biplane method to measure LV 2-dimensional EF from apical 4- and 2-chamber views and M-mode recordings to measure LV fractional shortening (FS). The most recently published guidelines define impaired EF by a value <54% in women and <52% in men. Because of clinical and local practice, we chose to define EF <50% as impaired. We defined impaired FS by a value <27% in women and <25% in men. A single investigator (JRC) performed all LV strain analyses off-line using semi-automatic software (EchoPAC SW only, version 112) from GE Healthcare. We measured LV longitudinal strain by 2-dimensional speckle-tracking echocardiography in the 3 standard apical views. The peak systolic GLS was calculated as an average of 17 LV segments by the automated function imaging algorithm in EchoPAC. We manually defined the basal points of the LV myocardium and the LV apex. The endocardial border was then automatically detected. We visually controlled the endocardial border, with manual adjustment when needed. We adjusted the region of interest to exclude the pericardium. Tracking of the LV myocardium was visually controlled and approved only if all LV segments were acceptable. Aortic valve closure was determined automatically and controlled visually in the apical long-axis view. Strain measurements were performed on survivors and controls in random order and without knowledge to the survivors’ status. We report differences in GLS in absolute values, according to recommendations. To assess the reliability of GLS measurements, another investigator (R.M.) independently measured GLS in 101 study participants. The interobserver repeatability was good, with an intraclass correlation coefficient of 0.888.
We report normally distributed data as mean ± SD and skewed data as median and range. We used independent-samples t tests to compare means, Mann-Whitney U tests to compare distributions, and chi-square tests to compare categorical data. Normal values of GLS depend on equipment and software used. We defined a lower limit of normal of absolute GLS by calculating mean – 1.96 SDs in the control group. To explore associations between impaired GLS and other survivor characteristics, we performed logistic regression analyses with impaired GLS as dependent variable. First, we analyzed the survivor group without impaired EF, FS, or history of diabetes, hypertension, coronary artery disease, or stroke (primary study group). Thereafter, we analyzed all the survivors in whom GLS could be measured, in particular to explore associations between GLS and previous cardiotoxic treatment (secondary study group). To explore the effect of high anthracycline doses, we compared survivors treated with ≤300 mg/m 2 with those treated with >300 mg/m 2 . To explore the effect of high mediastinal RT doses, we divided the survivor group into a low-dose group (≤30 Gy) and a high-dose group (>30 Gy). A 2-sided p value <0.05 was considered significant. We used IBM SPSS Statistics, version 21 (IBM, New York, New York) for all statistical analyses.
Results
We identified 259 survivors with a complete echocardiographic examination, of whom 191 were available for the primary study group and 231 for the secondary study group ( Figure 1 ). Table 1 lists characteristics of the survivors and controls. We have previously reported LV dimensions and diastolic function in the survivors. Briefly, both survivors (reported data are from the primary study group) and controls had normal mean LV internal dimensions (5.0 ± 0.4 vs 5.1 ± 0.5 cm, p = 0.200) and mean LV wall thickness (0.8 ± 0.1 vs 0.8 ± 0.1 cm, p = 0.037). Peak early diastolic velocities of the septal and lateral insertions of the mitral annulus were lower in survivors than controls (13 ± 3 vs 14 ± 3 cm/s, p <0.001).
| Variable | Primary Study Group ∗ | Secondary Study Group † | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Acute Lymphoblastic Leukemia (n = 108) | Hodgkin Lymphoma (n = 52) | Non -Hodgkin Lymphoma (n = 31) | All survivors (n = 191) | Controls (n = 180) | P value (survivors vs. controls) | Acute Lymphoblastic Leukemia (n = 128) | Hodgkin Lymphoma (n = 65) | Non- Hodgkin Lymphoma (n = 38) | All survivors (n = 231) | P value (survivors vs. controls) | |
| Age at exam (years) | 29.5 ± 7.0 | 33.5 ± 8.1 | 30.9 ± 6.9 | 30.8 ± 7.5 | 32.4 ± 8.3 | 0.058 | 29.4 ± 7.1 | 34.2 ± 8.6 | 31.6 ± 6.8 | 31.1 ± 7.8 | 0.123 |
| Age at diagnosis (years) | 6.3 ± 4.1 | 13.8 ± 3.3 | 11.4 ± 4.2 | 9.2 ± 5.2 | – | 6.3 ± 4.1 | 13.8 ± 3.3 | 11.3 ± 4.1 | 9.3 ± 5.1 | ||
| Follow-up time (years) | 23.3 ± 7.5 | 19.6 ± 8.2 | 19.5 ± 7.5 | 21.6 ± 7.9 | – | 23.1 ± 7.6 | 20.4 ± 8.8 | 20.3 ± 7.6 | 21.9 ± 8.0 | ||
| Female gender | 54 (50%) | 30 (58%) | 10 (32%) | 94 (48%) | 88 (49%) | 0.950 | 63 (49%) | 36 (55%) | 14 (37%) | 113 (49%) | 0.995 |
| Body weight (kg) | 76.3 ± 18.4 | 71.6 ± 15.2 | 77.0 ± 11.2 | 75.1 ± 16.6 | 75.3 ± 13.7 | 0.930 | 76.4 ± 18.2 | 72.1 ± 14.3 | 76.4 ± 10.9 | 75.2 ± 16.2 | 0.932 |
| Body surface area (m 2 ) | 1.91 ± 0.25 | 1.82 ± 0.21 | 1.92 ± 0.16 | 1.89 ± 0.23 | 1.90 ± 0.19 | 0.691 | 1.91 ± 0.25 | 1.84 ± 0.21 | 1.92 ± 0.16 | 1.89 ± 0.23 | 0.780 |
| Systolic blood pressure (mmHg) | 126 ± 13 | 127 ± 13 | 131 ± 14 | 127 ± 14 | 126 ± 12 | 0.362 | 125 ± 13 | 129 ± 15 | 130 ± 16 | 127 ± 14 | 0.300 |
| Heart rate (bpm) | 66 ± 10 | 69 ± 9 | 67 ± 13 | 67 ± 10 | 65 ± 11 | 0.058 | 66 ± 11 | 71 ± 11 | 67 ± 13 | 68 ± 11 | 0.006 |
| Daily smoker | 19 (18%) | 7 (14%) | 6 (19%) | 33 (17%) | 26 (14%) | 0.559 | 24 (19%) | 8 (12%) | 8 (21%) | 40 (17%) | 0.447 |
| Anthracycline exposure | 80 (74%) | 36 (69%) | 29 (94%) | 145 (76%) | – | 98 (77%) | 44 (68%) | 35 (92%) | 177 (77%) | ||
| Cumulative isotoxic doxorubicin dose (mg/m 2 ) | 120 (40, 485) | 160 (50, 400) | 135 (50, 400) | 135 (40, 485) | – | 120 (40, 485) | 160 (50, 400) | 150 (50, 460) | 150 (40, 485) | ||
| Mediastinal RT exposure | 0 | 36 (69%) | 6 (19%) | 42 (22%) | – | 0 | 45 (69%) | 7 (18%) | 52 (23%) | ||
| Cumulative RT dose (Gy) | – | 40 (18, 44) | 13 (13, 40) | 36 (13, 44) | – | 40 (18, 44) | 13 (13, 40) | 40 (13, 44) | |||
| Global longitudinal left ventricular strain | -19.3 ± 2.3 | -18.1 ± 2.0 | -19.1 ± 2.2 | -19.0 ± 2.2 | -21.4 ± 2.0 | <0.001 | 19.0 ± 2.3 | 17.8 ± 2.5 | 18.8 ± 2.4 | 18.6 ± 2.4 | <0.001 |
| Impaired global longitudinal strain | 26 (24%) | 21 (40%) | 6 (19%) | 53 (28%) | – | 35 (27%) | 29 (45%) | 10 (26%) | 74 (32%) | ||
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
