|
|
||
 |
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 24, No 16 (June 1), 2006: pp. 2527-2535 © 2006 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.03.9297 Quality of Life Among Long-Term Adolescent and Adult Survivors of Childhood Cancer
From the Unité de recherche en santé des populations, Centre de recherche du Centre Hospitalier Affilié Universitaire de Québec; Département de Médecine sociale et préventive, Université Laval, Québec; Centre for Chronic Disease Prevention and Control, Public Health Agency of Canada, Ottawa; Department of Psychology, Haematology/Oncology Program, Population Health Sciences, Research Institute, Hospital for Sick Children, University of Toronto, Toronto; Department of Pediatrics and Epidemiology and Biostatistics, University of Western Ontario and Children's Health Research Institute, London, Canada Address reprint requests to Elizabeth Maunsell, PhD, Unité de recherche en santé des populations, Hôpital du Saint-Sacrement, 1050 chemin Sainte-Foy, Québec, QC, Canada, G1S 4L8; e-mail: elizabeth.maunsell{at}uresp.ulaval.ca
PURPOSE: We assessed effects of childhood or adolescent cancer on quality of life among adolescent and adult cancer survivors, a group who are thought to be at particular risk for adverse late effects.
PATIENTS AND METHODS: We studied 1,334 survivors and 1,477 age- and sex-matched, general population controls from across Canada using a mailed questionnaire which included the Short Form-36 (SF-36) and measures of self-esteem, optimism, and life satisfaction. General linear models and logistic regression were used. Survivor-control differences corresponding to an effect size (ES)
RESULTS: Participants were age 15 years to 37 years. Most survivors (83.8%) were diagnosed CONCLUSION: Overall, a sizeable majority of adolescent and adult long-term survivors of childhood cancer in Canada appear to have adapted well.
Cancer is diagnosed in approximately 1,300 children and adolescents, up to 19 years of age, in Canada every year.1 Seventy-eight percent of children and adolescents now survive at least 5 years after their initial diagnosis.2 With growing numbers of survivors, it is increasingly important to understand the nature and extent of any late effects on their quality of life.3 Adolescent and adult survivors of childhood cancers are thought to be particularly at risk4,5 for poor physical, psychological, and social outcomes as a result of both their cancer and its treatment.6-15 Despite several large studies comparing adolescent and young adult survivors' quality of life with that of similar individuals without cancer,9,11,15 gaps in our knowledge remain. Few studies have been national or population-based in scope.9,11 While some studies have used population controls,10,12,16-18 others have compared published population norms19,20 with unknown age and sex distributions, or used sibling controls9,11,14 who may themselves have been affected by the survivor's cancer.21-23 Only two studies with controls, one American and one Dutch, examined clinical characteristics affecting survivors' long-term quality of life.9,15 Finally, findings from studies conducted in the United States could conceivably differ because of lack of universal health insurance. Lack of health insurance could negatively affect survivors both when they were children (if parents lacked health insurance or lost employment through which insurance was provided because of the child's cancer) or later when survivors were adults and could not obtain health insurance themselves. Thus, long-term quality of life effects of childhood cancers might differ in countries like Canada, where access to health care is universal. In this study we assessed the long-term effects of having had cancer during childhood or adolescence on several indicators of quality of life by comparing survivors with controls who have never had cancer. We also examined whether particular groups of these cancer survivors were at risk of poorer quality of life compared with their healthy peers.
Study Design and Patients Details of study methodology have been published.24 Briefly, study data were collected as part of the Canadian Childhood Cancer Surveillance and Control Program.25 Survivors were identified from patients at 12 pediatric oncology centers and provincial cancer registries from all 10 Canadian provinces. Eligible survivors had survived 5 years after diagnosis, between 1981 and 1990 inclusively, of an incident, primary cancer before age 20. The International Classification of Childhood Cancer was used.26 In each center, a random sample of potentially eligible patients stratified by diagnosis, age at diagnosis, and calendar year of diagnosis was selected. General population controls were selected by province using either provincial health insurance records or random digit dialing in provinces where access to health insurance records was not possible. Frequency matching was used to create a control group with a similar overall age and sex distribution as survivors. Individuals previously diagnosed with cancer were ineligible. Research ethics board approval was obtained from all participating centers.
Data Collection These analyses are based on 1,334 survivors and 1,477 controls. Specifically, among the 2,229 survivors who met the eligibility criteria, 664 could not be traced or did not return questionnaires, 231 declined participation, and 1,334 returned completed questionnaires giving a response rate of 60% (1,334 of 2,229). Including an estimation of potentially eligible control subjects residing at unanswered telephone numbers,24 we identified 3,323 eligible control subjects. Of these, 1,477 returned completed questionnaires resulting in a response rate of 44%.24
Measures Physical health problems were evaluated using some questions from the Childhood Cancer Survivor Study.31 Participants were asked whether they had ever had, or had ever been told by a doctor that they had, a number of different problems. Responses were grouped into cardiovascular, endocrine, hormonal, neurologic, renal, and pulmonary problems. We assumed that health problems occurring before diagnosis would be similar for survivors and similarly aged controls. Thus, if survivors reported more health problems we assumed these problems occurred after diagnosis and were related to cancer and its treatment.32-35 No information on recurrence or second cancers was collected. Three indicators of clinical outcome were used: treatment modality, number of treatment series, and organ dysfunction at treatment end. Treatment modality included the different combinations of chemotherapy, radiation therapy, and surgery; a treatment series could include any combination of these modalities. More than one treatment series occurred when the initial planned treatment was extended due to incomplete response or when a new treatment series was initiated after the survivor had relapsed. No other information about relapse subsequent to the end of treatment was available. Organ dysfunction refers to the number of organs or body systems documented in the medical record as impaired at the end of treatment. Information was obtained on several sociodemographic and personal characteristics including ethnicity, and experience of stressful life events and chronic stressors in the preceding 12-month period.36-38
Analysis
Because a number of small mean differences were statistically significant given the large sample size, we explored possible clinical significance of survivor-control differences by considering them in relation to effect size (ES).41 ES is the ratio of the survivor-control difference to the standard deviation among controls. ESs of 0.20 to 0.49 are generally considered to be small, 0.50 to 0.79 medium, and
Participants' characteristics are presented in Table 1. Most participants (88.2%) were younger than 30 years old at study. Just over half of the survivors were adolescents or moving into adolescence when diagnosed (55.0%), and most survivors were 10 or more years past diagnosis (83.5%).
Significantly fewer survivors than controls reported very good or excellent general health (62.3% and 71.2%, respectively; adjusted odds ratio, 0.61; 95% CI, 0.51 to 0.71). Survivors were more likely than controls to report that they had, or been told by a doctor that they had, one or more physical health problems, including endocrine, hormonal, cardiovascular, neurologic, and renal problems (Table 2).
In survivor-control comparisons stratified by sex, survivors of both genders reported significantly poorer scores in the PCS, general health, role-physical, and social function relative to their respective controls (Table 3). Although the effect of having had cancer during childhood was, or almost was, significantly greater for females than for males for the first three of these scores (interaction P values were .11, .048, and .08, respectively), differences comparing female survivors with controls on these outcomes were small (ES  –0.30) and those for male survivors and controls were negligible (ES –0.18). For the other outcomes, among both females and males, survivor-control differences were even smaller. Thus considered overall, there were no survivor-control sex differences that might be considered clinically important (ES for general health, –0.25; all others ranged from –0.03 to –0.19; data not shown).
Compared with controls, survivors of CNS and bone cancers had significantly poorer quality of life in several domains (Table 4). Among CNS survivors, general health, physical function, and role limitations due to physical health problems all showed relatively large absolute differences compared with controls, although ESs were small. CNS survivors also reported poorer quality of life in the psychosocial dimensions, but again ESs were small. For bone cancer survivors, negative and clinically important effects on quality of life compared with controls were found in the physical domain, with a large ES in physical function (ES, –1.02) and medium one for role limitations due to physical health problems (ES, –0.60). Twenty-three of 78 bone cancer survivors had an amputation.
Compared with controls, survivors with organ dysfunction in  two organs had consistent evidence of poorer quality of life (Table 5). Large effects were observed for general health and physical function (ES, –0.87 and –0.96 respectively), and medium effects for role-physical, social function, and life satisfaction (ES, –0.57 to –0.64). This was the only survivor group that had small to borderline clinically important effects in all other outcomes assessed including psychosocial ones (ES, –0.27 to –0.49). Survivors who had undergone two treatment series also reported poorer general health and role function related to physical health (ES, –0.53 for each; Table 5).
Considering treatment modality, survivor-control differences in mean PCS were generally largest and sometimes of borderline clinical importance among survivors who had all three treatment modalities (ESs, –0.31 to –0.49; Table 6). In absolute terms, the largest treatment ESs seen were in survivors who had cranial radiation alone or cranial radiation combined with other treatments (although the ES estimate for cranial radiation alone, based on only seven survivors, might be unstable). ESs for different treatment modalities on the MCS were small or negligible.
When the five important clinical characteristics were assessed together (CNS or bone cancer, more than one treatment series, two organs with dysfunction, all three treatment modalities and cranial radiation [CRT]) in relation to the PCS and MCS, the most important for predicting poorer quality of life among survivors were CNS or bone cancer, more than one treatment series, and two organs with dysfunction. Only these three variables remained independently associated with the PCS in the model containing all five clinical characteristics (data not shown). While only a minority of survivors (27%) had one of the three characteristics and few had two (7%) or three (1%) of them simultaneously, large ESs consistent with clinical importance were found for the PCS comparing survivors with two and three problems to those with none (–0.79 and –1.13, respectively).
This retrospective cohort study was conducted among adolescent and adult survivors of childhood and adolescent cancers from all 10 Canadian provinces, the majority of whom had been diagnosed and treated more than 10 years previously. Survivors in general reported more specific physical health problems than population controls with no cancer history. However, quality of life differences between survivors and controls were small, and for the most part probably not clinically important. More detailed analyses of clinical characteristics revealed some clinically important quality of life deficits, mainly in the physical domain, among bone and CNS survivors, those who had more than one treatment series (a proxy for relapse) and who had documented organ dysfunction at the end of treatment, and survivors who had all three modalities or CRT as part of treatment. CNS cancer survivors were the only group found to have consistently poorer physical and psychosocial functioning, but individually many effects were not considered clinically important. Although the case for less than 10% of survivors, clinically important deficits in the physical aspects of quality of life resulted from the clustering of two or three particular clinical characteristics simultaneously, namely diagnosis of CNS or bone cancer, more than one treatment series, and dysfunction in two organs when treatment was completed. Overall, our findings indicate that the quality of life of a sizeable majority of adolescent and adult long-term survivors of childhood cancer in Canada is comparable with those of their healthy peers. Analysis of the specific physical health problems indicates that the excess endocrine, hormonal, neurologic, cardiovascular, and renal problems we found is consistent with other reports.32-35 Our findings that a diagnosis of bone and CNS cancer and that the accumulation of all three treatment modalities are associated with poorer physical quality of life have been reported by others,6,9,14,43-45 as has the association of cranial radiation and surgery.9,35 To our knowledge, this is the first large population-based study to document the effects of the clustering of these somewhat correlated clinical characteristics. Finally the finding that, despite more health problems and poorer self-rated health, these long-term survivors were functioning as well in the psychosocial sphere as controls has been found in a number of studies,6,15,18,43,44 but not others.9,11 Discrepancies here could be related to methodologic differences between studies. As in a previous study,15 we found that the long-term effects of childhood cancer on quality of life and other psychosocial outcomes—effects measured by the differences between survivors and controls—did not differ meaningfully among female and male survivors. While female survivors had poorer outcomes than male survivors, these differences were also found between female and male controls in our and other studies,9,15 and indeed are usually found in the general healthy population. Thus, the sex difference observed in survivors was not an effect specific to having had childhood cancer. In fact, these findings provide further evidence of the extent to which most long-term survivors resemble their healthy peers. We have considered possible study limitations. Although participation was modest, particularly in controls, several observations about the similarity of survivors and controls who were theoretically targeted compared with those who participated in the study were reassuring about the likely absence of selection bias.24 First, the long-term survivors who participated in the study had clinical characteristics that were similar to those of survivors who either refused to participate or were lost to follow-up. Second, survivors and population controls had similar distributions of age, sex, and area of residence, and the vast majority of both groups were white. Third, the population controls recruited for our study closely resembled Canadians of the same age and sex based on comparisons with 1996 census information.24 Other biases are unlikely to have affected our findings. Validated and widely used, self-report instruments suitable for population surveys were used to measure outcomes in our survivors and controls. As well, overall and subgroup comparisons of survivors with controls incorporated the same potential confounding variables. Despite possible limitations, this study makes several important contributions. First, it is one of the first studies to be based on a large random sample of a diverse childhood cancer survivor population, and the first such study in Canada. Second, given that an important goal of cancer treatment is, after curing or controlling disease, to help those affected regain a quality of life that is, to the extent possible, comparable with that of similarly aged individuals who have not had to face this disease, the inclusion of a population-based control group allowed us to judge the success of treatment in these terms. For the majority of survivors, reported quality of life, self-esteem, optimism, and life satisfaction were comparable with those of their peers. Furthermore, the control group allowed us to see that the effect of childhood cancer on diverse aspects of quality of life was essentially the same in females and males. Third, we identified particular subgroups of survivors more likely to report poor physical quality of life, particularly in the minority of survivors where there was a clustering of clinical characteristics, all known at the end of treatment. This information will be useful for helping identify survivors who need and could benefit from ongoing long-term follow-up.6,45 Finally, our findings also support the relevance of using such information to tailor follow-up strategies to survivors' likelihood of poor quality of life and late effects.46 For survivors at lower risk, screening with postal questionnaires may be an effective and economic strategy for identifying those needing additional care.46 This latter approach might also avoid undermining the positive levels of psychosocial adaptation reported by the majority of long-term adolescent and adult survivors of childhood cancer.
We recognize the contribution of past members of the Canadian Childhood Cancer Surveillance and Control Program (CCCSCP) Late Effects Working Group to the planning and conduct of this study, past and present members of the CCCSCP Management Committee for overseeing the study, and the participating pediatric oncology centers and provincial agencies: British Columbia Cancer Agency, British Columbia Children's Hospital, Alberta Cancer Board, Alberta Children's Hospital, Cross Cancer Institute, Saskatchewan Cancer Agency, Allan Blair Cancer Centre, Saskatoon Cancer Centre, Manitoba Cancer Treatment and Research Foundation, Chedoke-McMaster Hospital, Hospital for Sick Children, Hôpital Ste-Justine, Montréal Children's Hospital, Centre hospitalier de l'Université Laval, Isaac Walton Killam Children's Hospital, Janeway Child Health Centre. Finally, we thank all participants who generously gave their time to participate in the study.
The authors indicated no potential conflicts of interest.
Supported by Health Canada. E.M. was a National Health Research and Development Program National Health Research Scholar and then an Investigator of the Canadian Institutes of Health Research when this study was planned, conducted, and analyzed. Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Related Editorial
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
