|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 22, No 15 (August 1), 2004: pp. 3149-3155 © 2004 American Society of Clinical Oncology. DOI: 10.1200/JCO.2004.01.047 Using Health-Related Quality of Life Measures to Predict Cardiac Function in Survivors Exposed to AnthracyclinesFrom the Divisions of Pediatric Oncology, Biostatistics and Epidemiology, and Cardiology, Department of Pediatrics; and the Center for Outcomes Research, Department of Anesthesiology and Critical Care Medicine, The Children's Hospital of Philadelphia and the University of Pennsylvania School of Medicine, Philadelphia, PA; The Divisions of Pediatric Cardiology, Hematology/Oncology, Children's Hospital Medical Center, Cincinnati, OH; The Division of Pediatric Cardiology, New York University of Medicine; and the Department of Pediatric Cardiology, Memorial Sloan-Kettering Medical Center, New York, NY Address reprint requests to Jeffrey H. Silber, MD, PhD, the Center for Outcomes Research, the Children's Hospital of Philadelphia, 3535 Market St, Suite 1029, Philadelphia, PA 19104; e-mail: silber{at}email.chop.edu
PURPOSE: As the number of pediatric cancer survivors increases, so does the number of survivors previously exposed to anthracyclines as part of their cancer therapy. Because screening is costly, some have suggested that health-related quality of life (HRQL) measures might be useful in focusing screening tests on those patients with cases most likely to display positive findings. This study reports on the predictive ability of HRQL measures to detect patients with abnormalities on serial cardiac testing.
METHODS: Using 127 patients from the ACE-Inhibitor after Anthracycline (AAA) Trial, this study compared serial measures of the Short Form-36 (SF-36; for ages > 13 years) and Child Health Questionnaire-Child Form 87 (CHQ-CF87; for ages RESULTS: Generally, there was no clinically or statistically significant correlation between any HRQL measure and any cardiac function measure except between MCI and vitality and physical functioning. For each of these measures, the correlation between MCI was statistically significant (P < .006), but each HRQL subscale could explain no more than 7% of the variation in MCI. HRQL measures were not predictive of any other cardiac function measure. CONCLUSION: HRQL measures should not be used in isolation as a screen for cardiac function abnormalities in patients exposed to anthracylines who already have a mild degree of ventricular dysfunction. Patient history appears to be no substitute for cardiac testing in this cohort.
Currently, there are approximately 250,000 long-term survivors of pediatric cancer, and about half (125,000) of these survivors have been exposed to anthracyclines as part of their cancer therapy.1 Although anthracyclines have contributed significantly to the increased cure rate in pediatric oncology, anthracycline exposure has resulted in the development of long-term cardiac dysfunction in a substantial portion of long-term survivors. Half of children treated with anthracyclines will develop some abnormal changes on cardiac screening tests.2-4 It is estimated that approximately 5% will develop late congestive heart failure, and 40% will develop arrhythmias, with approximately 10% requiring medical interventions.2,5-7 The large and growing population of anthracycline-exposed survivors has created a major cardiac screening challenge. Current guidelines recommend the frequency of screening with an echocardiogram, electrocardiogram, and holter monitor for anthracycline-exposed children based on the age of child at time of exposure, the cumulative dose of anthracycline, and the presence or absence of chest irradiation. The aim of this study was to determine if a response on a well established health-related quality of life (HRQL) questionnaire could aid in determining who should be screened more intensively for cardiac changes and who may not need routine screening. The present report assessed whether any associations exist between HRQL responses on the Short Form-36 (SF-36) and Child Health Questionnaire-Child Form 87 (CHQ-CF87) and detailed cardiac function tests as part of a randomized trial that followed long-term survivors of pediatric cancer.8,9 The ACE Inhibitor After Anthracycline (AAA) trial was a multicenter, randomized, double-blind clinical trial comparing enalapril to placebo, to determine whether early treatment would prevent cardiac deterioration in pediatric cancer survivors who screened positive for anthracycline cardiotoxicity any time point after cancer treatment.8,9 Because long-term follow-up screening for anthracycline cardiotoxicity is expensive and time-consuming, we asked if HRQL data from the AAA trial would help to develop a more specific and efficient screening algorithm for cardiotoxicity determination. Specifically, this study determined whether changes in cardiac function tests were associated with responses on the SF-36 and CHQ-CF87 functional status measures. We hypothesized that the HRQL questionnaires could aid in determining when to screen exposed patients for anthracycline toxicity who have a history of mild ventricular dysfunction.
Patient Population A detailed description of the AAA trial methodology has been published elsewhere.8,9 Eligible patients included those diagnosed before age 20 years, who had been off treatment for at least 2 years, and who were at least 4 years from diagnosis. Patients had received anthracycline as part of their cancer treatment and had some decline in cardiac function at any time before enrolling onto the AAA study. There were 135 patients who were randomly assigned to either Enalapril or placebo in a double-blind manner. The study was conducted at four institutions, and institutional review board approval was obtained at each institution. The median follow-up time on the study was 2.8 years, and the longest follow-up was 7 years.8 Patients underwent baseline cardiac evaluations and received repeat cardiac evaluations every 6 months. Patients also were asked to fill out quality-of-life assessments, CHQ-CF87 for those 13 years old or younger or the SF-36 for those older than 13 years, each time they had a cardiac evaluation. Research nurses in the AAA study clinic usually administered these tests. Patients were eligible for inclusion in this study data set if they had at least one pair of observations (eg, HRQL and cardiac evaluation). Cardiac tests that did not include HRQL measurements within a 3-month window were not used, and HRQL measures not accompanied by a cardiac test within 3 months were also deleted. There were 127 of 135 patients who had both HRQL measurements and cardiac evaluations available for analysis.
Description of Cardiac Evaluations MCI (L/min/m2). All patients were exercised to maximum volition with a standard ramp cycle protocol.10 This protocol consisted of sitting on an electronically braked cycle ergometer for 3 minutes followed by 3 minutes of unloaded pedaling. The work rate was then increased continuously at a slope chosen to achieve the patient's predicted peak work rate in 10 to 12 minutes of cycling. Minute oxygen consumption, minute carbon dioxide production, minute ventilation, and respiratory exchange ratio were monitored continuously on a breath-by-breath basis throughout the study by use of a commercially available metabolic cart. Cardiac output was estimated from measurement of effective pulmonary blood flow by the helium acetylene inert gas rebreathing technique.11 Cardiac output was measured at rest, every 3 minutes of exercise and at peak exercise. MCI was obtained by dividing the maximal cardiac output by the patient's body-surface area. LVESWS (g/cm2). LVESWS was an echocardiographic measure that was obtained according to the methods of Colan et al12,13 The LVESWS was also used to determine the left ventricular contractile state (WSI), which relates LVESWS and rate-adjusted velocity of fiber shortening.12,13 ESF. ESF is the resting echocardiograph-derived measurement of left ventricular systolic function. ESF = (left ventricular end-diastolic dimension – left ventricular end-systolic dimension) ÷ (left ventricular end-diastolic dimension) x 100.14
Description of HRQL Measures
Statistical Analysis Because patients were followed serially for up to 7 years, some patients initially were younger than 13 years and by the end of the study were older than 13 years. Hence, patients might have initially started with the CHQ-CF87 and after time switched to the SF-36. A total of 85 CHQ-CF87 and 493 SF-36 questionnaires were completed and available for analysis. This approach uses the quality of life measure as a predictor, given that the hypothesis is that it can aid in screening patients for need of cardiac evaluation. To use the full data for each patient, even if the patient moved from completing the CHQ-CF87 to the SF-36, all data were included in the model. The basis for this approach is that the subscales used in this analysis were common to both the SF-36 and CHQ-CF87. The only exception is the vitality subscale, which is part of the SF-36 but does not have an equivalent on the CHQ-CF87 and Self Esteem, which is part of the CHQ-CF87 but does not have an equivalent on the SF-36. The analysis scheme used an indicator variable to adjust for differences, if any, between SF-36 and CHQ-CF87, when modeling included both patients evaluated on the two scales at two time periods. The correlations and associated 95% CIs in this study are correlations based on the mixed-effects linear model and incorporating the repeated measures involved.18 Their interpretation is similar to the usual Pearson correlation. This study has greater than 80% power to detect a correlation of 0.3 or greater between an HRQL subscale and a cardiac outcome, versus no correlation, while controlling for a Type-I error of 0.00625, allowing for eight multiple comparisons, based on the eight subscales of the SF-36.19
Characteristics of Randomly Assigned Patients by Treatment Overall, 135 patients were randomly assigned to the AAA study. Of these 135 patients, 127 patients were eligible for analysis of HRQL and cardiac function (74 males and 53 females) including 578 sets of observations linking the cardiac functioning with responses on the aforementioned questionnaires. Data were accrued over a 7-year period (Table 1). The median age at first available HRQL visit was 18.1 years of age.
HRQL Measures Tables 2 and 3 report baseline scores for the SF-36 and CHQ-CF87 respectively. These tables display descriptive statistics on each of the subscales for 99 patients above the age of 13 and 28 patients ages 13 and younger. The Cronbach's  was generally between 0.7 and 0.9 for both groups. There were some differences between similar subscales in the younger patients versus the older patients. For example, mean general health was 64.9 in the younger patients and 72.7 in the older patients.
We evaluated construct validity for the SF-36 and CHQ-CF87, respectively. As expected for the SF-36, general health was well correlated with Vitality (0.64; P < .0001), Role Function, Physical (0.55; P < .0001), Physical Functioning (0.43; P < .0001) and Mental Health (0.46; P < .0001), whereas Vitality was correlated with General Health (0.64; P < .0001), Mental Health (0.55; P < .0001) and role physical (0.44; P < .0001). In younger patients on the CHQ-CF87, the General Health Perceptions category was best correlated with Bodily Pain (0.50; P = .007), and Physical Function (0.41; P = .03).
HRQL As Predictor of Cardiac Function
Table 5 represents the results of linear mixed effects models with respect to estimating the four cardiac outcomes using the HRQL, the source of the subscale (SF-36 v CHQ-CF87), while adjusting for diagnosis, age at diagnosis, and sex. Again, the models have almost no ability to predict the cardiac outcomes except, as was displayed in Table 4, the prediction of MCI using either physical function or vitality. Vitality had the strongest association with MCI, but again, the association was clinically insignificant. The coefficient on vitality was 0.014, suggesting that a change of two standard deviations in vitality (Table 2; or a 17.8 x 2 = 35.6-point change) would imply only a 0.5 L/min/m2 change in MCI. Because the mean and SD of the MCI test is 8.3 ± 2.66,9 a decline of two standard deviations in vitality would correspond to a clinically trivial 0.18 standard deviation decline in MCI.
In a subset of 15 patients, both questionnaires were available, and significant declines in cardiac function were observed (a decline in ESF of at least 20% or a decline in MCI of at least 30%, as defined in the parent AAA trial). In these 15 patients, there were 17 events. A paired t test was used to compare predecline and at-decline HRQL subscales, and there were no differences in any of the subscales measured. Hence, despite a clinically significant decline in cardiac status, there was no concomitant change in HRQL in these patients.
In patients with cardiac disease, studies have demonstrated that self-reported HRQL measurements provide essential information beyond the measurement of clinical variables alone. For instance, performance measured by an exercise treadmill has been shown to be better predicted by self-perception of health than by conventional interpretation of functional capacity by physicians.20,21 Furthermore, in patients who had recently undergone elective percutaneous transluminal coronary angioplasty, the physical functioning scale of the SF-36 was shown to be more responsive to change with angioplasty than the Canadian Cardiovascular Society anginal class. The authors concluded that the SF-36 was "capable of depicting the burden of disease and the benefits of treatment across a number of highly valued health states."21,22 In asthma patients, investigators have found that the physiological assessment of respiratory function and the patient's subjective report of well-being overlapped, but overall the relationships between HRQL scores and physiological measures were weak.23 In adults with congestive heart failure, the relationship between the physiological assessment of cardiovascular function and the patient's report of functional status has recently been evaluated. Mitani et al24 examined the relationship between clinical indicators and the HRQL in congestive heart failure patients to identify the significance of HRQL assessment in the management of the disease. They performed a cross-sectional observational study of stable congestive heart failure patients with a left ventricular ejection fraction less than 40%. Ninety-one patients were given the SF-36 to evaluate their HRQL. Data from an echocardiogram and an electrocardiaogram were obtained within 6 months. The echocardiographic data (left ventricular ejection fraction and left ventricular systolic and diastolic diameters) were not significantly related to any of the HRQL subscales. Additionally, this group had right-heart catheterization data on 29 patients, and the cardiac index and pulmonary artery wedge pressure were not significantly correlated with any of the HRQL subscales.24 This study suggests, as ours does, that the evaluation of the HRQL through the patient's self-report should not replace the physiological assessment. Among children and young adults, Barber et al25 sought to assess how well the parents' assessment, or in the case of young adults, the patient's assessment of his or her own abilities, correlated with measurements made during a cardiopulmonary exercise test. The results of this study indicated that the majority of patients would not be correctly classified on ability based on subjective (questionnaire-derived) measures. The questionnaire data significantly overestimated exercise ability in 67% and underestimated it in 3% of the subjects.25 Hence, it appears that the relationship between clinical indicators and patient-derived outcomes measured with validated HRQL instruments has not yet been clarified. There is little guidance on the integration of these two different sets of outcome data. Our study suggests that health status measures such as the CHQ-CF87 and the SF-36 are not sensitive to changes in cardiac function in long-term survivors of pediatric cancer exposed to anthracyclines who have mild ventricular dysfunction. Use of HRQL questionnaires should not be considered as a substitute for cardiac function tests to screen for anthracycline cardiotoxicity, because the HRQL measure did not aid in the prediction of cardiac function or of its decline in this cohort. On non–effort dependent tests, there was no association between LVESWS, WSI, ESF, and physical function subscales on the SF-36 and CHQ-CF87 and the vitality subscale of the SF-36. Even on potentially effort dependent tests, such as MCI, the HRQL measures showed only a very slight association, with vitality explaining only a small part of the variation in cardiac function. We conclude that HRQL measures—specifically, the SF-36 and CHQ-CF87—do not appear to be good proxies for actual measurements of cardiac function and provide little aid for determining who and when to screen with cardiac evaluations. Thus, there appears to be little substitute for detailed cardiac evaluations in this population.
The authors indicated no potential conflicts of interest.
We thank Margaret R. Tartaglione, Patricia M. Timlin, Denise DePaul, Charles T. Heise, Judy K. Correll, Rinske Niermans, and Anna T. Meadows, for their assistance in conducting this trial.
Supported by the National Heart, Lung, and Blood Institute (grant R01 HL-50424), the National Center for Research Resources (grant M01-RR-00240), and the National Cancer Institute (grant CA-16520). Authors' disclosures of potential conflicts of interest are found at the end of this article.
1. Bleyer WA: The impact of childhood cancer on the United States and the world. CA Cancer J Clin 40:355-367, 1990[Medline]
2. Steinherz LJ, Steinherz PG, Tan CT, et al: Cardiac toxicity 4 to 20 years after completing anthracycline therapy. JAMA 266:1672-1677, 1991 3. Lipshultz SE, Colan SD, Gelber RD, et al: Late cardiac effects of doxorubicin therapy for acute lymphoblastic leukemia in childhood. N Engl J Med 324:808-815, 1991[Abstract] 4. Silber JH, Jakacki RI, Larsen RL, et al: Increased risk of cardiac dysfunction after anthracyclines in girls. Med Pediatr Oncol 21:477-479, 1993[Medline] 5. Krischer JP, Epstein S, Cuthbertson DD, et al: Clinical cardiotoxicity following anthracycline treatment for childhood cancer: The Pediatric Oncology Group experience. J Clin Oncol 15:1544-1552, 1997[Abstract]
6. 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 332:1738-1743, 1995
7. Steinherz LJ, Graham T, Hurwitz R, et al: Guidelines for cardiac monitoring of children during and after anthracycline therapy: Report of the Cardiology Committee of the Childrens Cancer Study Group. Pediatrics 89:942-949, 1992 8. Silber JH, Cnaan A, Clark BJ, et al: Design and baseline characteristics for the ACE Inhibitor After Anthracycline (AAA) study of cardiac dysfunction in long-term pediatric cancer survivors. Am Heart J 142:577-585, 2001[CrossRef][Medline]
9. Silber JH, Cnaan A, Clark BJ, et al: Enalapril to prevent cardiac function decline in long-term survivors of pediatric cancer exposed to anthracyclines. J Clin Oncol 22:820-828, 2004 10. Wasserman K, Hansen J, Sue D: Protocols for exercise testing, in Wasserman K (ed): Principles of Exercise Testing and Interpretation. Philadelphia, PA, Lea & Febiger, 1994, pp 95-111 11. Paridon SM: Exercise testing, in Garson A, Bricker JT, Fisher D, et al (eds): The Science and Practice of Pediatric Cardiology. Baltimore, MD, Williams & Wilkins, 1998, pp 875-888 12. Colan SD, Borow KM, Neumann A: Left ventricular end-systolic wall stress-velocity of fiber shortening relation: A load-independent index of myocardial contractility. J Am Coll Cardiol 4:715-724, 1984[Abstract] 13. Colan SD, Parness IA, Spevak PJ, et al: Developmental modulation of myocardial mechanics: Age- and growth-related alterations in afterload and contractility. J Am Coll Cardiol 19:619-629, 1992[Abstract] 14. Vargo TA: Cardiac catheterization-hemodynamic measurements, in Garson AJ, Bricker JT, McNamara DG (eds): The Science and Practice of Pediatric Cardiology. Philadelphia, PA, Lea & Febiger, 1990 15. Ware JE, Snow KK, Kosinski M, et al: SF-36 Health Survey Manual and Interpretation Guide. Lincoln, RI, QualityMetric Inc, 2000 16. Landgraf JM, Abetz L, Ware JE: The CHQ User's Manual. (second printing). Boston, MA, HealthAct, 1999, pp 27-38 17. Laird NM, Ware JH: Random effects models for longitudinal data. Biometrics 38:963-974, 1982[CrossRef][Medline] 18. Lipsitz SR, Leong T, Ibrahim J, et al: A partial correlation coefficient and coefficient of determination for multivariate normal repeated measures data. Statistician 50:87-95, 2001 19. Dixon W, Massey F: Introduction to Statistical Analysis (ed 4). New York, NY, McGraw-Hill, 1983, p 224 20. Permanyer-Miralda G, Alonso J, Anto JM, et al: Comparison of perceived health status and conventional functional evaluation in stable patients with coronary artery disease. J Clin Epidemiol 44:779-786, 1991[CrossRef][Medline]
21. Rumsfeld JS, MaWhinney S, McCarthy M, et al: Health-related quality of life as a predictor of mortality following coronary artery bypass graft surgery. JAMA 281:1298-1303, 1999 22. Krumholz HM, McHorney CA, Clark L, et al: Changes in health after elective percutaneous coronary revascularization: A comparison of generic and specific measures. Med Care 34:754-759, 1996[CrossRef][Medline]
23. Tsukino M, Nishimura K, Ikeda A, et al: Physiologic factors that determine the health-related quality of life in patients with COPD. Chest 110:896-903, 1996 24. Mitani H, Hashimoto H, Isshiki T, et al: Health-related quality of life of Japanese patients with chronic heart failure. Circ J 67:215-220, 2003[CrossRef][Medline] 25. Barber G, Heise CT: Subjective estimates of exercise ability: Comparison to objective measurements. Ped Exerc Sci 3:327-332, 1991 Submitted January 8, 2004; accepted May 8, 2004.
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2004 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
13 years) to serial cardiac performance tests including echocardiographic shortening fraction, left ventricular end systolic wall stress (LVESWS), LVESWS-index, and maximal cardiac index (MCI; a measure of cardiac output at peak exercise). 
