Academic Medical Centre, Emma Kinder Ziekenhuis, University of Amsterdam, Departments of 1Pediatric Oncology, 2Medical Oncology and 3Pediatrics, Amsterdam, The Netherlands
Received 2 January 2002; accepted 16 January 2002.
Abstract
Background
The aim of this systematic review was to summarise and appraise the published evidence with regard to the frequency and risk factors of subclinical cardiotoxicity in apparently healthy survivors of childhood cancer after anthracycline therapy.
Patients and methods
A search was made in Medline for studies published between 1966 and May 2001 that included >50 children and reported on the frequency of measures of subclinical cardiotoxicity. Information about the studies was abstracted by two reviewers and a validity score was calculated for each study.
Results
The reported frequency of subclinical cardiotoxicity varied between 0% and 57% in the 25 studies included. Differences in outcome definitions of subclinical cardiotoxicity and differences in study patients with respect to the dose of anthracycline seemed to explain part of the wide variance of the frequency of subclinical cardiotoxicity. Fourteen of the 25 studies showed serious methodological limitations.
Conclusions
The reported frequency of subclinical cardiotoxicity shows a wide variation. Well designed studies with accurate and precise outcome measurements in well described groups of patients, after a sufficiently long follow-up period, are needed to obtain more insight into the frequency and importance of risk factors, and the clinical consequences of anthracycline-related subclinical cardiotoxicity.
Key words: anthracyclines, cardiotoxicity, children, heart, systematic review, systolic function
Introduction
The introduction of anthracyclines in the treatment of childhood cancer has improved the survival rate, but at the cost of myocardial damage during and after treatment. In addition to clinical cardiotoxicity, manifesting itself as clinical heart failure, studies have also reported subclinical cardiotoxicity, i.e. abnormalities measured by diagnostic techniques, in previously healthy survivors of childhood cancer [1]. Knowledge about the extent of subclinical cardiotoxicity is important for the development of new treatment protocols that aim at reducing cardiotoxicity, and for the follow-up of patients who have been treated with anthracyclines. In this systematic review an evaluation is made of the available evidence with regard to the frequency and risk factors of subclinical cardiotoxicity up to many years after chemotherapeutic treatment.
Patients and methods
Search for studies
The objective of the literature search was to identify all studies reporting on the frequency of subclinical cardiotoxicity, defined as an abnormal systolic function, after anthracycline treatment for childhood cancer. The selection process involved four steps. First, Medline was searched for potentially relevant articles published from 1966 to May 2001. The sensitive search strategy is presented in Table 1. Secondly, relevant articles that possibly met the following inclusion criteria, i.e. (i) original report, (ii) published in English, (iii) study population of >50 children who were treated with doxorubicin, daunorubicin or epirubicin, and (iv) reported subclinical cardiotoxicity as outcome, were selected on the basis of the title and abstract by two researchers (L.K., H.J.H.v.d.P.) and were retrieved for more detailed examination. Thirdly, the bibliographies of all relevant articles and reviews were checked for additional references to empirical studies. Finally, all retrieved articles were screened by the two researchers to ensure that they described original research and met the inclusion criteria. Inter-observer agreement was calculated for the second and fourth steps of the selection process. In cases of disagreement, the abstracts and articles were re-examined and discussed until consensus was achieved.
|
To determine the validity of the selected studies, two researchers (L.K., H.J.H.v.d.P.) assessed the design and execution of each study. The validity was assessed according to the criteria for prognostic studies described by Laupacis et al. [2]. A sample was considered to be well-defined if the mean, the median or the range of the cumulative anthracycline dose was mentioned. A sample of study patients was defined as representative for the underlying population if it consisted of >95% of the whole cohort from the start of therapy, or the cohort of survivors or a random sample of the cohort of patients with respect to the cumulative dose. It was also assessed whether the original cohort could be identified and whether the study patients entered the cohort at a similar well-defined point in the course of the disease. A minimal follow-up of >1 year after the end of chemotherapeutic treatment, or a median follow-up of 2 years, was considered to be adequate to determine the frequency of subclinical cardiotoxicity. Follow-up was considered to be complete if the outcome was assessed at the end date of the study for >95% of the study population. It was assessed whether objective and unbiased outcome criteria, defined as outcome measured by one observer who was blinded for possible risk factors, were applied. It was also assessed whether there was adjustment for important prognostic factors. Each study was graded on the basis of meeting or not meeting the criteria, and a validity score was calculated (range 08).
Definition and frequency of subclinical cardiotoxicity
The frequency of subclinical cardiotoxicity was calculated as the number of patients with an abnormal systolic function divided by the number in patients of the study group. The 95% confidence interval of the frequency of subclinical cardiotoxicity was calculated by CIA , a statistical program [3].
Risk factors for subclinical cardiotoxicity
The risk factors for subclinical cardiotoxicity, adjusted for other possible confounders in multivariate analysis, were extracted from the various studies.
Results
Selection of articles
The search strategy yielded 282 potentially relevant references from the Medline search (Table 1). Fifty-six articles were retrieved for more detailed examination. The inter-observer agreement for this first selection was 98%. Two additional article were retrieved by two reviewers. Twenty-five studies from the 58 retrieved articles met all the inclusion criteria, with an inter-observer agreement of 96% [428]. The other 33 articles were excluded for the following reasons: not original research, adult study population, impossible to differentiate between adults and children in the study group, <50 children or not published in English.
Description of the selected studies
The frequency and risk factors of an abnormal systolic function were evaluated in one longitudinal study until the end of treatment [17], and in 24 transversal studies in survivors only [416, 1828]. In the transversal studies, patients were investigated at different intervals after the end of treatment. Table 2 presents details of the studies included, length of follow-up and the frequency of abnormal systolic function. Thirteen studies investigated a group that included patients who also manifested signs of clinical heart failure [5, 6, 11, 1315, 1722, 23]. Eight studies described the frequency of subclinical cardiotoxicity after treatment with doxorubicin [4, 6, 1113, 15, 19, 22], two studies after treatment with daunorubicin [21, 26] and 15 studies after a combination of anthracycline therapy [5, 710, 14, 1618, 20, 2325, 27, 28]. Eleven studies described a study group of patients treated with a mean or median anthracycline dose <300 mg/m2 [7, 11, 12, 17, 20, 21, 2428]. Three studies only reported the range of anthracyclines [4, 8, 23].
|
|
Risk factors
Data on the risk factors included in multivariate analyses in 10 of the studies are presented in Table 4. A higher cumulative dose, longer follow-up time, mediastinal radiation, female sex, higher dose in 3 weeks, diagnosis, lower age (two studies) and older age (one study) at diagnosis are reported as independent risk factors for a decreased systolic function and for an increased afterload.
|
This review demonstrates that at follow-up of children treated for cancer, systolic cardiac dysfunction can be detected in previously asymptomatic patients. The studies that investigated subclinical cardiotoxicity >1 year after the end of therapy are studies with a transversal design and different lengths of follow-up. The reported frequency of subclinical cardiotoxicity, defined as abnormal systolic function and/or abnormal afterload varied between 0% and 57%. Variation in the definition of the outcome and variation in the study groups with respect to the cumulative dose seemed to explain a part of this wide range. The outcome was defined using different assessment methods of systolic function and afterload, and these measurements are based on different pathophysiological mechanisms. Six studies of the 15 studies that reported an abnormal SF in children treated with anthracyclines had an acceptable validity. In this subgroup of studies the range of the frequency of an abnormal SF seemed to be higher for patients treated with a mean or median cumulative dose >300 mg/m2 anthracyclines (15.527.8%) than for patients treated with a cumulative dose <300 mg/m2 (015.2%). The relation between subclinical cardiotoxicity and subsequent clinical heart failure is as yet unclear.
Most studies in this systematic review have methodological limitations. Most investigated the cardiac function in a non-representative subgroup of the original cohort, whereas in some of the studies the original cohort was not even reported. Selection of a non-random subgroup of patients probably leads to an overestimation of the frequency. Not all studies reported that the outcome was measured at the end of the study; therefore, cases of subclinical heart failure could have been missed. The objectivity of echocardiographic measurements is dependent on the reproducibility of the measurement, and measurements performed by one observer only can improve the reproducibility. Only five studies mentioned that all echocardiographic measurements were performed by one observer who was blinded for possible risk factors. There is debate about whether and which echocardiographic measurements should be adjusted for age or body surface.
Only 10 studies applied multivariate analysis to the risk factors of a lower systolic function or increased afterload at follow-up. A higher cumulative dose of anthracyclines, both older and younger age at diagnosis, longer follow-up period, radiation therapy, female sex, a higher dose intensity and type of cancer have been reported as risk factors for decreased systolic function or increased afterload. These results must be interpreted in the light of the validity of the study. For example, Steinherz et al. [5] reported that a longer follow-up period increased the risk of a lower SF. However, when it is not clear whether all patients were monitored until the end of the study, overestimation of the relationship between follow-up time and a decreased SF is possible if patients with a lower SF have been monitored for a longer period of time than patients with a normal SF. Moreover, as far as the authors know, no studies have described the course of subclinical cardiotoxicity related to the follow-up time in individual patients.
In conclusion, the reported frequency of subclinical cardiotoxicity after anthracycline therapy in studies including >50 children varied between 0% and 56%. Differences in outcome definition and study groups with respect to the cumulative dose of anthracyclines could explain a part of this wide range. Most studies showed serious methodological limitations. More cohort studies and randomized trials are needed to investigate the factors that increase the risk of subclinical or clinical cardiotoxicity after anthracycline therapy in children. These studies should focus on the cardiotoxic effects of different types of anthracyclines, cumulative dose, dose intensity and the cardiotoxic effects of co-treatment weighted against the benefit of anthracyclines on survival. These studies should be performed in well defined and representative patients groups from the start of treatment, or in survivors with a complete and sufficient length of follow-up period, and include precise and accurate outcome measures. Patients with subclinical cardiotoxicity should be monitored for a long period of time to obtain insight into the clinical consequences.
Footnotes
+ Correspondence to: Dr L. C. M. Kremer, Emma Kinderziekenhuis, Academic Medical Centre, University of Amsterdam, Department of Pediatrics, G8-262, Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands. Tel: +31-20-5662453; Fax: +31-20-6917735; E-mail: L.C.Kremer@amc.uva.nl
References
1. Grenier MA, Lipshultz SE. Epidemiology of anthracycline cardiotoxicity in children and adults. Semin Oncol 1998; 25: 7285.[ISI][Medline]
2. Laupacis A, Wells G, Richardson WS, Tugwell P. Users guides to the medical literature. V. How to use an article about prognosis. Evidence-Based Medicine Working Group. JAMA 1994; 272: 234237.[ISI][Medline]
3. Gardner MJ, Altman DG. Statistics with Confidence. BMJ Press 1989.
4. Hausdorf G, Morf G, Beron G et al. Long term doxorubicin cardiotoxicity in childhood: non-invasive evaluation of the contractile state and diastolic filling. Br Heart J 1988; 60: 309315.[Abstract]
5. Steinherz LJ, Steinherz PG, Tan CT et al. Cardiac toxicity 4 to 20 years after completing anthracycline therapy. JAMA 1991; 266: 16721677.[Abstract]
6. Lipshultz SE, Colan SD, Gelber RD et al. Late cardiac effects of doxorubicin therapy for acute lymphoblastic leukemia in childhood. N Engl J Med 1991; 324: 808815.[Abstract]
7. Larsen RL, Jakacki RI, Vetter VL et al. Electrocardiographic changes and arrhythmias after cancer therapy in children and young adults. Am J Cardiol 1992; 70: 7377.[ISI][Medline]
8. Schwartz CL, Hobbie WL, Truesdell S et al. Corrected QT interval prolongation in anthracycline-treated survivors of childhood cancer. J Clin Oncol 1993; 11: 19061910.[Abstract]
9. Silber JH, Jakacki RI, Larsen RL et al. Increased risk of cardiac dysfunction after anthracyclines in girls. Med Pediatr Oncol 1993; 21: 477479.[ISI][Medline]
10. Jakacki RI, Larsen RL, Barber G et al. Comparison of cardiac function tests after anthracycline therapy in childhood. Implications for screening. Cancer 1993; 72: 27392745.[ISI][Medline]
11. Hudson MM, Greenwald C, Thompson E et al. Efficacy and toxicity of multiagent chemotherapy and low-dose involved-field radiotherapy in children and adolescents with Hodgkins disease. J Clin Oncol 1993; 11: 100108.[Abstract]
12. Ewer MS, Ali MK, Gibbs HR et al. Cardiac diastolic function in pediatric patients receiving doxorubicin. Acta Oncol 1994; 33: 645649.[ISI][Medline]
13. Sorensen K, Levitt G, Sebag-Montefiore et al. Cardiac function in Wilms tumor survivors. J Clin Oncol 1995; 13: 15461556.[Abstract]
14. BuLock FA, Mott MG, Oakhill A, Martin RP. Left ventricular diastolic function after anthracycline chemotherapy in childhood: relation with systolic function, symptoms, and pathophysiology. Br Heart J 1995; 73: 340350.[Abstract]
15.
Lipshultz SE, Lipsitz SR, Mone SM et al. Female sex and drug dose as risk factors for late cardiotoxic effects of doxorubicin therapy for childhood cancer. N Engl J Med 1995; 332: 17381743.
16. Pihkala J, Saarinen UM, Lundstrom U et al. Myocardial function in children and adolescents after therapy with anthracyclines and chest irradiation. Eur J Cancer 1996; 32A: 97103.
17. BuLock FA, Mott MG, Oakhill A, Martin RP. Early identification of anthracycline cardiomyopathy: possibilities and implications. Arch Dis Child 1996; 75: 416422.[Abstract]
18. Godoy LY, Fukushige J, Igarashi H et al. Anthracycline-induced cardiotoxicity in children with malignancies. Acta Paediatr Jpn 1997; 39: 188193.[Medline]
19. Kakadekar AP, Sandor GG, Fryer C et al. Differences in dose scheduling as a factor in the etiology of anthracycline-induced cardiotoxicity in Ewing sarcoma patients. Med Pediatr Oncol 1997; 28: 2226.[ISI][Medline]
20. Sung RY, Huang GY, Shing MK et al. Echocardiographic evaluation of cardiac function in paediatric oncology patients treated with or without anthracycline. Int J Cardiol 1997; 60: 239248.[ISI][Medline]
21. Sorensen K, Levitt G, Bull C et al. Anthracycline dose in childhood acute lymphoblastic leukemia: issues of early survival versus late cardiotoxicity. J Clin Oncol 1997; 15: 6168.[Abstract]
22. Ewer MS, Jaffe N, Ried H et al. Doxorubicin cardiotoxicity in children: comparison of a consecutive divided daily dose administration schedule with single dose (rapid) infusion administration. Med Pediatr Oncol 1998; 31: 512515.[ISI][Medline]
23. Nysom K, Holm K, Lipsitz SR et al. Relationship between cumulative anthracycline dose and late cardiotoxicity in childhood acute lymphoblastic leukemia. J Clin Oncol 1998; 16: 545550.[Abstract]
24. Zalewska-Szewczyk B, Lipiec J, Bodalski J. Late cardiotoxicity of anthracyclines in children with acute leukemia. Klin Pediatr 1999; 211: 356359.
25. Lanzarini L, Bossi G, Laudisa ML et al. Lack of clinically significant cardiac dysfunction during intermediate dobutamine doses in long-term childhood cancer survivors exposed to anthracyclines. Am Heart J 2000; 140: 315323.[ISI][Medline]
26. Rammeloo LA, Postma A, Sobotka-Plojhar MA et al. Low-dose daunorubicin in induction treatment of childhood acute lymphoblastic leukemia: no long-term cardiac damage in a randomized study of the Dutch Childhood Leukemia Study Group. Med Pediatr Oncol 2000; 35: 1319.[ISI][Medline]
27. Kapusta L, Thijssen JM, Groot-Loonen J et al. Tissue Doppler imaging in detection of myocardial dysfunction in survivors of childhood cancer treated with anthracyclines. Ultrasound Med Biol 2000; 26: 10991108.[ISI][Medline]
28. Bossi G, Lanzarini L, Laudisa ML et al. Echocardiographic evaluation of patients cured of childhood cancer: a single center study of 117 subjects who received anthracyclines. Med Pediatr Oncol 2001; 36: 593600.[ISI][Medline]