- Email: [email protected]
Subsequent mortality experience in five-year survivors of childhood, adolescent and young adult cancer in Scotland: A population based, retrospective cohort study
European Journal of Cancer (2013) 49, 3274– 3283
Available at www.sciencedirect.com
journal homepage: www.ejcancer.com
Subsequent mortality experience in five-year survivors of childhood, adolescent and young adult cancer in Scotland: A population based, retrospective cohort study David H. Brewster a,b,⇑, David Clark a, Leanne Hopkins a, Jacqui Bauer a, Sarah H. Wild b, Angela B. Edgar c, W. Hamish Wallace c a
Information Services Division, NHS National Services Scotland, Edinburgh, Scotland, United Kingdom Centre for Population Health Sciences, University of Edinburgh, Edinburgh, Scotland, United Kingdom c Royal Hospital for Sick Children, Edinburgh, Scotland, United Kingdom b
Available online 8 June 2013
KEYWORDS Adolescent Child Cohort studies Mortality Neoplasms Prognosis Risk Survivors Young adult
Abstract Aim: To assess the risk of death in patients who survive at least 5 years after diagnosis of childhood, adolescent or young adult cancer. Patients and Methods: This was a population-based retrospective cohort study using linked national cancer registry and mortality records in Scotland. The study population consisted of 5229 individuals who were diagnosed with cancer before the age of 25 years between 1981 and 2003, and who survived at least 5 years after the date of diagnosis of their primary cancer. Indirect standardisation was used to calculate mortality ratios standardised for age and sex and absolute excess risks (AERs) compared to the general Scottish population. Results: During 58,358 person-years of follow-up, there were 359 deaths among the cohort of cancer survivors. The overall SMR was 6.1 (95% confidence interval (CI) 5.5–6.7) and AER 51 (45–58) per 10,000 person-years. Largely because of age- and sex-related differences in background mortality, SMRs were higher in patients diagnosed at 0–14 years (SMR 11.0, 95% CI 9.3–12.9) than 15–24 years (4.7, 4.1–5.3), and in females (9.2, 7.8–10.8) than males (4.8, 4.2– 5.5). SMRs and AERs varied substantially by primary cancer and by underlying cause of death. In general, SMRs were little altered by standardisation for an area-based indicator of socio-economic deprivation. Adjusted for age and sex, the risk of death was significantly lower in five-year survivors diagnosed during 1998–2003 compared to those diagnosed during 1981–1985 (Relative hazard ratio, 0.54, 95% CI 0.36–0.81).
⇑ Corresponding author: Address: Scottish Cancer Registry, Information Services Division, NHS National Services Scotland, Gyle Square, 1 South Gyle Crescent, Edinburgh EH12 9EB, Scotland, United Kingdom. Tel.: +44 131 275 6092; fax: +44 131 275 7511. E-mail address: [email protected] (D.H. Brewster).
0959-8049/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejca.2013.05.004
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
3275
Conclusion: Long-term survivors of cancer in childhood and young adulthood remain at higher risk of mortality than the general population, although the absolute risk of death is low and the excess risk has decreased over time. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Continuing advances in therapy mean that approximately 80% of children and young people with cancer can now expect to be alive 5 years after diagnosis. However, approximately two thirds of survivors experience at least one late effect, and approximately one third experience a late effect that is severe or life threatening.1 At least twelve publications, in some cases involving the same cohorts described at different time points, have reported that five-year survivors of childhood cancer have a 7–17-fold increased risk of mortality from all causes, compared to the general population.2–13 The excess mortality is not limited to cancer but also applies to non-cancer causes of death. Many studies have demonstrated an increased risk of second primary cancers arising in survivors of childhood cancer.14–21 Other long-term complications include effects on the endocrine, cardiac and pulmonary systems, renal impairment, gastrointestinal dysfunction, musculoskeletal sequelae, neurocognitive dysfunction, and psychosocial manifestations.1,22–24 Long term morbidity and mortality risks in childhood cancer survivors relate largely to treatment modality, although they may also be determined by individual host characteristics.1 The fundamental challenge among this group of patients is to further improve survival prospects while minimising the incidence and severity of treatment-induced late effects.25 Although the risk of second primary malignancies following childhood cancer has been extensively studied in the United Kingdom (UK) and elsewhere,14–21 less attention has been devoted to other late effects of therapy, such as cardiac disease. A major active follow-up study, the British Childhood Cancer Survivor Study,26 is currently underway, and although this is gathering very detailed data, it may not be entirely representative since it relies on patients consenting to participate. A single centre study from Sheffield, UK suggested that survivors of childhood cancer may have a tendency to over-report serious late effects, such as second primary cancers, compared to what is documented in hospital medical records.27 A recent report on cardiac outcomes from five-year survivors of selected childhood cancers in the United States (US) reported increased risks of selfreported congestive heart failure, myocardial infarction, pericardial disease and valvular abnormalities compared to siblings.23 The authors noted that the proportion of deaths among eligible participants who refused to participate or who were lost to follow-up was higher than
among study participants suggesting that the reported risks may be underestimates. On the other hand, studies from specialist centres tend to overestimate the prevalence of chronic disease,28 emphasising the need for studies to be population-based, if possible. Compared to follow-up of younger children, considerably less research has been carried out into late effects of therapy for cancer diagnosed during adolescence or young adulthood,4,29,30 although there is evidence that around half of survivors of adolescent cancer experience late effects of therapy.31 The aim of the present study was to describe patterns of mortality among a cohort of patients who have survived at least 5 years after the diagnosis of cancer in childhood, adolescence or young adulthood in Scotland. 2. Patients and methods The study population comprised patients registered with the Scottish Cancer Registry who survived at least 5 years after the diagnosis of cancer in childhood, adolescence, or young adulthood (age between 0 and 24 years). We decided to study this combined age group in the interests of greater statistical power, and because this age group is now managed in the context of a single national network of clinicians in Scotland (the ‘Managed Service Network for Children and Young People with Cancer’). In Scotland, cancer registration records are linked to mortality records by computerised probability matching, which is believed to be highly accurate.32 Emigrations of patients registered with cancer from Scotland to other UK countries are notified to the cancer registry by the National Health Service Central Register (NHSCR). We included data for individuals whose year of diagnosis was in 1981 (the first year of the linked database in Scotland)32 or subsequent years up to 2003 (to allow 5 years’ survival, and at least one further year’s follow-up). Follow-up was from 5 years after diagnosis to date of emigration from Scotland, date of death, or end of 2009, whichever occurred first. In terms of diagnosis of first cancer, the study population was re-classified according to the third edition of the International Classification of Childhood Cancer (ICCC-3) (0–14-year-olds)33 and the most up-to-date version of the diagnostic classification of cancer in adolescents and young adults (15–24-year-olds) developed by Birch et al.34 The data were then mapped to a common ‘study classification’. Briefly, this was based primarily on ICCC-3 but with separation of group XI (Other malignant epithelial neoplasms and malignant
3276
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
melanomas) into two categories: (1) Melanoma and skin carcinomas and (2) Other carcinomas, excluding renal, hepatic, gonadal, and skin. Mapping of cancers in adolescents and young adults was based initially on specific morphology codes (sometimes in combination with topography codes); thereafter the mapping was based on groups and sub-groups within the Birch et al. classification (see Appendix A). Individuals were assigned to fifths of Carstairs deprivation scores, an indicator of socio-economic position based on postcode sectors of residence at the time of original diagnosis, by applying 1981, 1991 and 2001 census-derived Carstairs scores to the periods of diagnosis 1981–1985, 1986–1995, 1996–2003, respectively. The Carstairs deprivation index is based on small area of residence, and is derived from four variables collected at each decennial census: social class, unemployment, overcrowding and car ownership.35 The characteristics of patients surviving beyond 5 years were compared to non-survivors, stratified by age at diagnosis, sex, socio-economic position at diagnosis and tumour type using the chi-squared test of associ-
ation. Both for all causes of death combined and for specific causes of death (see Appendix B), indirectly standardised mortality ratios (SMRs) were calculated, using the general population as an external comparison group to generate expected numbers of deaths (based on age-, sex-, deprivation category-, and calendar periodspecific rates of disease). Absolute excess risks (AERs) were calculated as the observed minus the expected number of deaths divided by the number of person-years at risk and expressed as the rate per 10,000 cancer survivors per year. The AER reflects the additional mortality burden beyond background mortality. 95% confidence intervals (CIs) around SMRs and AERs were calculated based on the assumption that the observed numbers of deaths followed a Poisson distribution. SMRs with 95% confidence intervals that did not include the value 1.0 were regarded as statistically significant. Cumulative mortality was estimated using the Kaplan–Meier method,36 and differences in cumulative mortality were assessed for statistical significance using the log rank test. Cox proportional hazards modelling was used to assess whether the hazard of death decreased over time.
Table 1 Characteristics of the original cohort including the study population (the five-year survivors). v2
Characteristic
Five-year survivors Number
%
Number
%
All patients combined
5229
75
1751
25
Age at diagnosis of first cancer (years) <1 1–4 5–9 10–14 15–19 20–24
188 802 555 631 1043 2010
65 74 72 72 72 80
102 280 216 241 408 504
35 26 28 28 28 20
63.7
0.000
Sex Male Female
2762 2467
73 77
1004 747
27 23
10.8
0.001
Carstairs deprivation fifth 1 – least deprived 2 3 4 5 – most deprived
1087 1005 1040 1055 1036
77 74 76 75 72
323 344 321 359 403
23 26 24 25 28
11.9
0.018
884 980 587 93 81 154 19 188 317 694 709 459 64
69 82 61 44 96 83 45 56 70 89 93 76 79
401 209 383 116 3 31 23 150 133 89 54 142 17
31 18 39 56 4 17 55 44 30 11 7 24 21
601.6
0.000
Type of first cancer Leukaemias Lymphomas CNS tumours Neuroblastoma Retinoblastoma Renal tumours Hepatic tumours Bone tumours Soft tissue sarcomas Germ cell tumours Melanoma and skin carcinomas Other carcinomas Other and unspecified neoplasms
Non-survivors
P-Value
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
3. Results From an original cohort of 6980 children and young people with cancer, the study population comprised 5229 individuals who had survived for at least 5 years after the date of diagnosis of their primary cancer. Only five patients were recorded as emigrating from Scotland within 5 years of the original diagnosis of cancer, and only 37 beyond 5 years after diagnosis. The proportion of survivors varied significantly according to type of primary cancer, and was significantly higher in those diagnosed at an older age than those diagnosed at a younger age, females compared to males, and the least compared to the most deprived fifth (Table 1). During 58,358 person-years of follow-up, there were 359 deaths (Table 2). The overall age-, sex-, and deprivation standardised mortality ratio (SMR) was 6.2 (95% CI 5.6–6.9). In general, SMRs were little altered by standardisation for deprivation. For example, the overall SMR, standardised for age and sex only was 6.1 (5.5– 6.7). For the sake of comparability with other studies, the SMRs presented hereafter are standardised for age and sex only. Compared to background rates, the relative risk of death was higher in patients who were diagnosed with cancer in the age range 0–14 years (SMR 11.0, 95% CI 9.3–12.9) than in those diagnosed between the ages of 15 and 24 years (4.7, 4.1–5.3). SMRs were higher in females (9.2, 7.8–10.8) than males (4.8, 4.2– 5.5), and were particularly high in patients with the following types of primary cancer: leukaemias (11.7, 9.0– 14.8), CNS tumours (14.7, 11.5–18.4), hepatic tumours (65, 18–167), and bone tumours (10.3, 6.5–15.6). The overall AER, standardised for age and sex only, was 51 (45–58) per 10,000 person-years. AERs varied substantially by primary cancer (Table 2). Neoplasms were identified as the cause of death for the majority (76%) of deaths, 65% due to the primary cancer and 11% due to subsequent cancers. Other causes of death included injury, poisoning, and other external factors (15%), and diseases of the circulatory system (4.5%). The SMR for death from cancer was higher (32, 28– 36) than for non-cancer deaths (1.7, 1.3–2.1). The SMR for death from a subsequent cancer (a subset of all cancer deaths) was 4.6 (3.2–6.2). Table 3 shows the numbers of deaths, SMRs and AERs for those major causes of death with statistically significantly increased SMRs. SMRs were significantly increased in respect of deaths from diseases of the nervous system (2.9, 1.2–6.0), diseases of the circulatory system (2.4, 1.4–3.9), diseases of the respiratory system (3.4, 1.2–7.4), and congenital malformations, deformations and chromosomal abnormalities (4.7, 1.3–11.9). In respect of the remaining 52 deaths (not included in Table 3), no significantly increased SMRs were found for any of the following causes of death: certain infectious and parasitic diseases;
3277
endocrine, nutritional and metabolic diseases; diseases of the digestive system; diseases of the genitourinary system; injury, poisoning and certain other consequences of external causes; external causes of mortality; or any other causes. Twelve deaths were attributed to suicide (SMR 1.1, 0.6–2.0). No deaths were recorded from interstitial pneumonitis and/or pulmonary fibrosis caused by radiation or drugs. AERs also varied considerably according to the underlying cause of death (Table 3). Cumulative mortality was significantly higher in patients diagnosed during 1981–1985, compared to patients diagnosed during 1986–1991, 1992–1997 and 1998–2003 (all P < 0.001) (Figure 1). However, other pair-wise comparisons did not show any statistically significant differences. Adjusted for age and sex, the risk of death was significantly lower in five-year survivors diagnosed during 1998–2003 compared to those diagnosed during 1981–1985 (Relative hazard ratio, 0.54, 95% CI 0.36–0.81).
4. Discussion Using routinely collected data, we have been able to estimate mortality of survivors of cancer at a young age relative to the whole population, after adjusting for age and sex. In spite of surviving at least 5 years after their primary cancer, this group of patients remains at significantly increased risk of death compared to their peers. However, the absolute risk of death is low, reflecting relatively low background risks of death among young people. Our cohort study has a number of strengths. Most importantly, it was population-based, and not, therefore, subject to some of the biases associated with single centre studies or surveys that depend on participation. The study has been performed against the background of a longstanding national health service funded primarily through taxation and free at the point of use. Scottish Cancer Registry data have been shown to be of comparatively high quality,37,38 and the linkage with official mortality records is believed to be highly accurate and complete.32 Unlike many other studies, we were able to standardise our analyses for socio-economic deprivation. Importantly, this made little difference to the results, although this observation may reflect limitations of the indicator of deprivation that we used, which was area-based (rather than based on individual characteristics) and was relevant at the time of diagnosis of cancer.35 It is complicated by the fact that the experience of cancer at a young age may impact negatively on adult socio-economic position and information on socio-economic position during follow-up was not available in this study.39 Ideally, the impact of socio-economic position on later mortality should be assessed in other
3278
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
Table 2 Numbers of person-years at risk and deaths, and standardised mortality ratios (SMRs) and absolute excess risks (AERs)* with 95% confidence intervals (CIs) among five-year survivors according to various characteristics. SMR
95% LCI**
95% UCI**
AER*
100.0
6.1
5.5
6.7
51
45
58
7 47 42 49 76 138 145 214
1.9 13.1 11.7 13.6 21.2 38.4 40.4 59.6
12.2 13.5 11.8 8.8 5.8 4.2 11.0 4.7
4.9 9.9 8.5 6.5 4.6 3.5 9.3 4.1
25.1 18.0 16.0 11.6 7.3 5.0 12.9 5.3
32 48 66 66 53 46 56 48
11 35 46 47 39 36 47 40
68 66 91 90 69 57 67 57
30,929.9 27,428.7
202 157
56.3 43.7
4.8 9.2
4.2 7.8
5.5 10.8
52 51
43 42
61 61
12,306.9 11,483.5 11,616.6 11,757.4 11,117.9
66 55 75 86 77
18.4 15.3 20.9 24.0 21.4
5.4 4.6 6.3 7.5 6.8
4.2 3.5 5.0 6.0 5.3
6.9 6.0 7.9 9.3 8.4
44 38 54 63 59
32 26 41 49 44
58 52 71 81 76
9238.4 11,423.9 6309.2 1010.2 1058.8 1764.0 177.4 1896.2 3609.0 8154.2 7511.2
67 73 75 3 3 6 4 22 24 26 25
18.7 20.3 20.9 0.8 0.8 1.7 1.1 6.1 6.7 7.2 7.0
11.7 5.2 14.7 7.3 8.0 7.1 65.2 10.3 8.2 2.1 3.1
9.0 4.1 11.5 1.5 1.7 2.6 17.8 6.5 5.2 1.4 2.0
14.8 6.5 18.4 21.3 23.4 15.5 167.0 15.6 12.1 3.1 4.6
66 52 111 26 25 29 222 105 58 17 23
50 38 85 2 2 8 57 62 35 6 11
86 68 141 82 80 69 561 165 91 32 39
5234.6 971.4
28 3
7.8 0.8
4.8 2.4
3.2 0.5
6.9 7.1
42 18
24 6
66 77
Characteristic
Person-years at risk
Numbers of deaths
%
All patients combined
58,358.6
359
Age at diagnosis of first cancer (years) <1 2012.2 1–4 9013.3 5–9 5871.2 10–14 6543.2 15–19 11,920.5 20–24 22,998.2 0–14 23,439.9 15–24 34,918.7 Sex Male Female Carstairs deprivation fifth 1 – least deprived 2 3 4 5 – most deprived Type of first cancer Leukaemias Lymphomas CNS tumours Neuroblastoma Retinoblastoma Renal tumours Hepatic tumours Bone tumours Soft tissue sarcomas Germ cell tumours Melanoma and skin carcinomas Other carcinomas Other and unspecified neoplasms * **
95% LCI**
95% UCI**
AER – absolute excess risk per 10,000 person years. LCI – lower confidence interval; UCI – upper confidence interval.
populations using alternative indicators which can be measured longitudinally. The main weakness of our study is the very limited information available on treatment. Thus, we can only
make inferences about treatment based on era of diagnosis. In relation to outcomes, causes of death were not validated, although this limitation also, of course, applies to the reference population used to calculate
Table 3 Numbers of deaths, standardised mortality ratios (SMRs) and absolute excess risks (AERs)* with 95% confidence intervals (CIs) for selected (significantly increased) causes of death among five-year survivors. AER*
95% LCI**
36.1 6.2
5.2 0.0
3.3 0.2
7.7 0.8
6.0 3.9 7.4 11.9
0.1 0.8 0.1 0.3
0.1 0.1 0.2 0.0
0.5 2.1 1.0 1.2
Primary cause of death
Numbers of deaths
SMR
95% LCI**
95% UCI**
Neoplasms (all cancers) – Neoplasms (subsequent cancers only) Diseases of the nervous system Diseases of the circulatory system Diseases of the respiratory system Congenital malformations, deformations and chromosomal abnormalities
274 39
32.0 4.6
28.4 3.2
7 16 6 4
2.9 2.4 3.4 4.7
1.2 1.4 1.2 1.3
* **
AER – absolute excess risk per 10,000 person years. LCI – lower confidence interval; UCI – upper confidence interval.
95% UCI**
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283 18%
16%
Percentage Cumulative Mortality
14%
12%
10%
8%
6%
4%
2%
0%
0
5
10
15
20
25
Time in Years 1981-1985
1986-1991
1992-1997
1998-2003
Fig. 1. Cumulative mortality among five-year survivors by period of diagnosis of first cancer.
expected numbers of deaths. However, it must be acknowledged that some causes of death may have been incorrectly assigned. For example, in a previous study from the Nordic countries, 12% (148 of 1208) of fatalities attributed to the first malignancy were re-classified as non-cancer deaths, and 15 cases officially recorded as death due to malformation were re-classified as deaths caused by primary cancer (almost all these cases being neurofibromatosis type I).12 Note, however, that neurofibromatosis was classified under the neoplasms chapter of the ninth revision of the International Classification of Diseases (ICD9); in Scotland, neurofibromatosis was only classified as a congenital malformation from the year 2000 when ICD10 coding of death certificates was introduced. Our study population is smaller than some other studies, which limits the capacity to carry out sub-group analyses. Inevitably, the length of follow-up available for the most recently diagnosed patients is necessarily limited. While it is possible that some members of the cohort may have emigrated from Scotland without this being recorded and are therefore lost to follow-up, we believe that the proportion of unrecorded emigrations from the cohort (at any time after diagnosis of the primary cancer) is low. Other publications have reported that five-year survivors of childhood cancer (variously defined) have
3279
a 7–17-fold increased risk of mortality from all causes, compared to the general population.2–13 In our study the overall SMR for survivors of childhood cancer was 11.0, which lies just below the middle of the range of estimates reported from other studies. A recently published study of five-year survivors of young adult (20– 24-year-olds) cancer from British Columbia reported an overall SMR of 5.9 (4.9–6.9) and AER of 5.3 per 1000 years at risk (or 53 per 10,000 years at risk), which is of a similar order of magnitude to our own findings in 15–24-year-olds (SMR 4.7, 4.1–5.3; AER 48, 40–57 per 10,000 years at risk).29 In comparing studies, it is important to note that there is some heterogeneity in terms of age group at diagnosis of primary cancer, calendar period of diagnosis, and period of follow-up. However, the finding of a higher SMR in females, reflecting lower background mortality in young adulthood compared to males, is consistent across studies. Lower background mortality in children compared to adolescents/young adults is also likely to explain the higher SMR seen in survivors of childhood cancer. Recent studies of longterm mortality in Nordic and British survivors of childhood cancer showed that excess mortality continued to occur even beyond 25 years after diagnosis.2,3 However, other studies have, like our own, shown a reduced risk of death among five-year survivors diagnosed with cancer in the more recent past compared to historical cohorts.2,6,12 Associations between congenital anomalies and/or genetic factors and cancer at a young age have been established in many studies.40 This might explain our observation of an increased risk of death (SMR 4.7) from congenital malformations, deformations and chromosomal abnormalities. However, as noted previously, it is also important to consider the possibility of misdiagnosis or misclassification of the underlying cause of death in some circumstances. For example, hydrocephalus acquired as a consequence of a brain tumour might be incorrectly diagnosed and/or coded as a congenital malformation. The Scottish Intercollegiate Guidelines Network has recently published an updated guideline on the late effects of the treatment of cancer in young people.41 The guideline development group has specifically stated that health care professionals should be aware that survivors of childhood cancer are at particular and lifelong increased risk of development of a subsequent primary cancer and that this may occur at any site on the body. The guideline group were not able to identify any evidence of either benefits or harms of specific screening programmes for survivors of childhood cancer, nor were any studies identified on outcomes for survivors of childhood cancer entering national screening programmes (e.g. mammography for breast cancer) at an earlier age than for general population groups.
3280
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
If patients, parents, carers and clinicians are to make fully informed decisions about treatment and follow-up, it is important to quantify the long-term outcomes of conditions. Our study has confirmed that long-term survivors of cancer in childhood and young adulthood remain at higher risk (albeit low absolute risk) of mortality compared to their peers. It has also demonstrated the utility of routinely collected, linked records for monitoring many fatal late effects at a relatively low cost compared to the costs of active follow-up. The results of such high level monitoring may identify specific effects that warrant more detailed investigation. In future, as increasing numbers of patients survive their cancer, high level monitoring of late effects, both fatal and non-fatal (for example, from hospitalisation records), should become as routine as monitoring five-year survival. In future, it may also be possible to capture more detailed treatment information through computerised radiotherapy planning systems and systemic anti-cancer therapy prescribing systems. This would enable risk estimates to be related to specific types of treatment. Conflict of interest statement None declared. Acknowledgements This study was supported by a grant from the Chief Scientist Office, Scottish Government Health and Social
Care Directorates (CZG/2/508). Beyond arranging review of the original grant application and providing funding, the Chief Scientist Office had no role in any other part of the study. Appendix A Study classification of primary neoplasm and mappings from ICCC-3 and teenagers and young adults with cancer (TYAC) classifications. Study classification of primary neoplasm. Study classification Leukaemias Lymphomas CNS tumours Neuroblastoma Retinoblastoma Renal tumours Hepatic tumours Bone tumours Soft tissue sarcomas Germ cell tumours Melanoma and skin carcinomas Other carcinomas* Other and unspecified neoplasms *Except
renal, hepatic, gonadal, and skin.
Mapping of Childhood Cancer Classification (ICCC-3) to study classification. ICCC-3 classification
Study classification
Group I – leukaemias, myeloproliferative diseases, and myelodysplastic diseases Group II – lymphomas and reticuloendothelial neoplasms Group III – CNS and miscellaneous intracranial and intraspinal neoplasms Group IV – neuroblastoma and other peripheral nervous cell tumours Group V – retinoblastoma Group VI – renal tumours Group VII – hepatic tumours Group VIII – malignant bone tumours Group IX – soft tissue and other extraosseous sarcomas Group X – germ cell tumours, trophoblastic tumours, and neoplasms of gonads Group XI – other malignant epithelial neoplasms and malignant melanomas
All Group I = leukaemias
Group XII – other and unspecified malignant neoplasms
All Group II = lymphomas All Group III = CNS tumours All Group IV = neuroblastoma All All All All All All
Group Group Group Group Group Group
V = retinoblastoma VI = renal tumours VII = hepatic tumours VIII = bone tumours IX = soft tissue sarcomas X = germ cell tumours
Category XId = melanoma and skin carcinomas Category XIe = melanoma and skin carcinomas Remainder = other carcinomas All Group XII = other and unspecified neoplasms
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
3281
Appendix A (continued) Mapping of TYAC classification to study classification *Any
cases with ICDO morphology codes 9510/3–9514/3 = retinoblastoma. cases with ICDO morphology codes 8970/3 = hepatic tumours. *Any cases with ICDO morphology codes 8964/3 = renal tumours. *Any cases with ICDO morphology codes 8963/3 + ICD9 189.0 or ICD10 C64 = renal tumours. *Any remaining cases with ICDO morphology codes 8963/3 = soft tissue sarcomas. *Any cases with ICDO morphology codes 9501–9504 + ICD9 191–192, 237.5, 237.6, 237.9 or ICD10 C70.0–C72.9, D32–D33, D42–D43 = CNS tumours. *Any remaining cases with ICDO morphology codes 9501/3–9504/3 = neuroblastoma. *Any cases with ICDO morphology codes 9505–9508 = CNS tumours. *Any cases with ICDO morphology codes 9520/3–9523/3 = neuroblastoma. *Any
Thereafter: TYAC classification
Study classification
Group 1 – leukaemia All Group 1 = leukaemias Group 2 – lymphoma All Group 2 = lymphomas Group 3 – central nervous system and other intracranial and All Group 3 = CNS tumours intraspinal neoplasms (CNS tumours) Group 4 – osseous and chondromatous neoplasms, Ewing tumour All Group 4 = bone tumours and other neoplasms of bone (bone tumours) Group 5 – soft tissue sarcomas (STS) All Group 5 = soft tissue sarcomas Group 6 – germ cell and trophoblastic neoplasms (germ cell tumours) All Group 6 = germ cell tumours Group 7 – melanoma and skin carcinoma All Group 7 = melanoma and skin carcinomas Group 8 – carcinomas (except of skin) Category 8.5.1 = renal tumours Category 8.5.3 = germ cell tumours Category 8.6.3 = hepatic tumours Remainder of Group 8 = other carcinomas Group 9 – miscellaneous specified neoplasms NEC (miscellaneous Category 9.1.1 = renal tumours specified) Category 9.1.2 = neuroblastoma Category 9.2.1 = neuroblastoma Category 9.2.2 = germ cell tumours Category 9.2.3 = lymphomas Remainder of Group 9 = other and unspecified neoplasms Group 10 – unspecified malignant neoplasms NEC (unspecified) All Group 10 = other and unspecified neoplasms Appendix B Disease-specific outcome codes from mortality records (primary underlying cause). Disease grouping
ICD9
ICD10
Certain infectious and parasitic diseases Neoplasms – the primary neoplasm* – subsequent neoplasm(s)* Endocrine, nutritional and metabolic diseases – Diabetes mellitus Diseases of the nervous system Diseases of the circulatory system – Coronary heart disease
001–139 140–239
A00–B99 C00–D48
240–279 250 320–359 390–459 410–414
E00–E90 E10–E14 G00–G99 excluding G45–G46 I00–I99, G45–G46 I20–I25 (continued on next page)
3282
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
Appendix B (continued) – Cardiomyopathy – Cerebrovascular disease Diseases of the respiratory system – Interstitial pneumonitis and/or pulmonary fibrosis caused by radiation or drugs Diseases of the digestive system Diseases of the genitourinary system Congenital malformations, deformations and chromosomal abnormalities Injury, poisoning and certain other consequences of external causes** External causes of morbidity and mortality** – Suicide (DSH and undetermined intent) – Accidental poisoning – All other accidents
425 430–438 460–519 508
I25.5, I41.2, I42, I43 I60–I69, G45–G46 J00–J99 J70
520–579 580–629 740–759
K00–K93 N00–N99 Q00–Q99
800–999
S00–T98
E800–E999 E950–E959, E980–E989 E850–E869 E800–E849, E870–E949
V01–Y98 X60–X84, Y87.0, Y10–Y34, Y87.2 X40–X49 V01–V99, W00–W99, X00–X39, X50– X59, Y40–Y59, Y85, Y86
Other causes by inspection of primary and subsequent ICD9/10 codes. **In general, the external cause codes take precedence over the injury etc codes in mortality data. *Classified
References 1. Bhatia S, Constine LS. Late morbidity after successful treatment of children with cancer. Cancer J 2009;15:174–80. 2. Garwicz S, Anderson H, Olsen JH, et al. Late and very late mortality in 5-year survivors of childhood cancer: changing pattern over four decades – experience from the Nordic countries. Int J Cancer 2012;131:1659–66. 3. Reulen RC, Winter DL, Frobisher C, et al. Long-term causespecific mortality among survivors of childhood cancer. JAMA 2010;304:172–9. 4. Prasad PK, Signorello LB, Friedman DL, Boice Jr JD, Pukkala E. Long-term non-cancer mortality in pediatric and young adult cancer survivors in Finland. Pediatr Blood Cancer 2012;58:421–7. 5. Wilson CL, Cohn RJ, Johnston KA, Ashton LJ. Late mortality and second cancers in an Australian cohort of childhood cancer survivors. Med J Aust 2010;193:258–61. 6. Armstrong GT, Pan Z, Ness KK, Srivastava D, Robison LL. Temporal trends in cause-specific late mortality among 5-year survivors of childhood cancer. J Clin Oncol 2010;28:1224–31. 7. Armstrong GT, Liu Q, Yasui Y, et al. Late mortality among 5year survivors of childhood cancer: a summary from the Childhood Cancer Survivor Study. J Clin Oncol 2009;27:2328–38. 8. Mertens AC, Liu Q, Neglia JP, et al. Cause-specific late mortality among 5-year survivors of childhood cancer: the Childhood Cancer Survivor Study. J Natl Cancer Inst 2008;100:1368–79. 9. Dama E, Pastore G, Mosso ML, et al. Late deaths among fiveyear survivors of childhood cancer. A population-based study in Piedmont Region, Italy. Haematologica 2006;91:1084–91. 10. Cardous-Ubbink MC, Heinen RC, Langeveld NE, et al. Longterm cause-specific mortality among five-year survivors of childhood cancer. Pediatr Blood Cancer 2004;42:563–73. 11. Mertens AC, Yasui Y, Neglia JP, et al. Late mortality experience in five-year survivors of childhood and adolescent cancer: the Childhood Cancer Survivor Study. J Clin Oncol 2001;19:3163–72. 12. Mo¨ller TR, Garwicz S, Barlow L, et al. Decreasing late mortality among five-year survivors of cancer in childhood and adolescence: a population-based study in the Nordic countries. J Clin Oncol 2001;19:3173–81.
13. Hudson MM, Jones D, Boyett J, Sharp GB, Pui CH. Late mortality of long-term survivors of childhood cancer. J Clin Oncol 1997;15:2205–13. 14. Reulen RC, Frobisher C, Winter DL, et al. Long-term risks of subsequent primary neoplasms among survivors of childhood cancer. JAMA 2011;305:2311–9. 15. Olsen JH, Moller T, Anderson H, et al. Lifelong cancer incidence in 47,697 patients treated for childhood cancer in the Nordic countries. J Natl Cancer Inst 2009;101:806–13. 16. Olsen JH, Garwicz S, Hertz H, et al. Second malignant neoplasms after cancer in childhood or adolescence. Nordic Society of Paediatric Haematology and Oncology Association of the Nordic Cancer Registries. BMJ 1993 Oct 23;307(6911):1030–6. 17. Garwicz S, Anderson H, Olsen JH, et al. Second malignant neoplasms after cancer in childhood and adolescence: a population-based case-control study in the 5 Nordic countries. The Nordic Society for Pediatric Hematology and Oncology. The Association of the Nordic Cancer Registries. Int J Cancer 2000;88:672–8. 18. Jenkinson HC, Hawkins MM, Stiller CA, Winter DL, Marsden HB, Stevens MC. Long-term population-based risks of second malignant neoplasms after childhood cancer in Britain. Br J Cancer 2004;91:1905–10. 19. de Vathaire F, Hawkins M, Campbell S, et al. Second malignant neoplasms after a first cancer in childhood: temporal pattern of risk according to type of treatment. Br J Cancer 1999;79:1884–93. 20. Kingston JE, Hawkins MM, Draper GJ, Marsden HB, Kinnier Wilson LM. Patterns of multiple primary tumours in patients treated for cancer during childhood. Br J Cancer 1987;56:331–8. 21. Hawkins MM, Draper GJ, Kingston JE. Incidence of second primary tumours among childhood cancer survivors. Br J Cancer 1987;56:339–47. 22. Edgar AB, Morris EM, Kelnar CJ, Wallace HB. Long-term follow-up of survivors of childhood cancer. Endocr Dev 2009;15:159–80. 23. Mulrooney DA, Yeazel MW, Kawashima T, et al. Cardiac outcomes in a cohort of adult survivors of childhood and adolescent cancer: retrospective analysis of the Childhood Cancer Survivor Study cohort. BMJ 2009;339:b4606.
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283 24. Recklitis CJ, Diller LR, Li X, Najita J, Robison LL, Zeltzer L. Suicide ideation in adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. J Clin Oncol 2010;28:655–61. 25. Scottish Intercollegiate Guidelines Network. Long term follow up of survivors of childhood cancer (SIGN 76). Edinburgh: SIGN; 2004. http://www.sign.ac.uk//pdf/sign76.pdf. 26. Hawkins MM, Lancashire ER, Winter DL, et al. The British Childhood Cancer Survivor Study: objectives, methods, population structure, response rates and initial descriptive information. Pediatr Blood Cancer 2008;50:1018–25. 27. Taylor N, Absolom K, Michel G, et al. Comparison of selfreported late effects with medical records among survivors of childhood cancer. Eur J Cancer 2010;46:1069–78. 28. Ness KK, Leisenring W, Goodman P, et al. Assessment of selection bias in clinic-based populations of childhood cancer survivors: a report from the childhood cancer survivor study. Pediatr Blood Cancer 2009;52:379–86. 29. Zhang Y, Goddard K, Spinelli JJ, Gotay C, McBride ML. Risk of late mortality and second malignant neoplasms among 5-year survivors of young adult cancer: a report of the childhood, adolescent, and young adult cancer survivors research program. J Cancer Epidemiol 2012;2012:103032. 30. Woodward E, Jessop M, Glaser A, Stark D. Late effects in survivors of teenage and young adult cancer: does age matter? Ann Oncol 2011;22:2561–8. 31. Pentheroudakis G, Pavlidis N. Late toxicity in survivors from adolescent cancers. Cancer Treat Rev 2007;33:656–63.
3283
32. Kendrick S, Clarke J. The Scottish record linkage system. Health Bull (Edinb) 1993;51:72–9. 33. Steliarova-Foucher E, Stiller C, Lacour B, Kaatsch P. International Classification of Childhood Cancer, third edition. Cancer 2005;103:1457–67. 34. Birch JM, Alston RD, Kelsey AM, Quinn MJ, Babb P, McNally RJQ. Classification and incidence of cancers in adolescents and young adults in England 1979–1997. Br J Cancer 2002;87:1267–74. 35. Morris R, Carstairs V. Which deprivation? A comparison of selected deprivation indexes. J Public Health Med 1991;13:318–26. 36. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81. 37. Brewster D, Crichton J, Harvey JC, Dawson G. Completeness of case ascertainment in a Scottish Regional Cancer Registry for the year 1992. Public Health 1997;111:339–43. 38. Brewster DH, Stockton D, Harvey J, Mackay M. Reliability of cancer registration data in Scotland, 1997. Eur J Cancer 2002;38:414–7. 39. Holmqvist AS, Wiebe T, Hjorth L, Lindgren A, Ora I, Moell C. Young age at diagnosis is a risk factor for negative late socioeconomic effects after acute lymphoblastic leukemia in childhood. Pediatr Blood Cancer 2010;55:698–707. 40. Little J. Epidemiology of Childhood Cancer. IARC Scientific Publications No. 149. Lyon: International Agency for Research on Cancer; 1999. 41. Scottish Intercollegiate Guidelines Network. Long term follow-up of survivors of childhood cancer (SIGN 132). Edinburgh: SIGN; 2013.
Available at www.sciencedirect.com
journal homepage: www.ejcancer.com
Subsequent mortality experience in five-year survivors of childhood, adolescent and young adult cancer in Scotland: A population based, retrospective cohort study David H. Brewster a,b,⇑, David Clark a, Leanne Hopkins a, Jacqui Bauer a, Sarah H. Wild b, Angela B. Edgar c, W. Hamish Wallace c a
Information Services Division, NHS National Services Scotland, Edinburgh, Scotland, United Kingdom Centre for Population Health Sciences, University of Edinburgh, Edinburgh, Scotland, United Kingdom c Royal Hospital for Sick Children, Edinburgh, Scotland, United Kingdom b
Available online 8 June 2013
KEYWORDS Adolescent Child Cohort studies Mortality Neoplasms Prognosis Risk Survivors Young adult
Abstract Aim: To assess the risk of death in patients who survive at least 5 years after diagnosis of childhood, adolescent or young adult cancer. Patients and Methods: This was a population-based retrospective cohort study using linked national cancer registry and mortality records in Scotland. The study population consisted of 5229 individuals who were diagnosed with cancer before the age of 25 years between 1981 and 2003, and who survived at least 5 years after the date of diagnosis of their primary cancer. Indirect standardisation was used to calculate mortality ratios standardised for age and sex and absolute excess risks (AERs) compared to the general Scottish population. Results: During 58,358 person-years of follow-up, there were 359 deaths among the cohort of cancer survivors. The overall SMR was 6.1 (95% confidence interval (CI) 5.5–6.7) and AER 51 (45–58) per 10,000 person-years. Largely because of age- and sex-related differences in background mortality, SMRs were higher in patients diagnosed at 0–14 years (SMR 11.0, 95% CI 9.3–12.9) than 15–24 years (4.7, 4.1–5.3), and in females (9.2, 7.8–10.8) than males (4.8, 4.2– 5.5). SMRs and AERs varied substantially by primary cancer and by underlying cause of death. In general, SMRs were little altered by standardisation for an area-based indicator of socio-economic deprivation. Adjusted for age and sex, the risk of death was significantly lower in five-year survivors diagnosed during 1998–2003 compared to those diagnosed during 1981–1985 (Relative hazard ratio, 0.54, 95% CI 0.36–0.81).
⇑ Corresponding author: Address: Scottish Cancer Registry, Information Services Division, NHS National Services Scotland, Gyle Square, 1 South Gyle Crescent, Edinburgh EH12 9EB, Scotland, United Kingdom. Tel.: +44 131 275 6092; fax: +44 131 275 7511. E-mail address: [email protected] (D.H. Brewster).
0959-8049/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejca.2013.05.004
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
3275
Conclusion: Long-term survivors of cancer in childhood and young adulthood remain at higher risk of mortality than the general population, although the absolute risk of death is low and the excess risk has decreased over time. Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction Continuing advances in therapy mean that approximately 80% of children and young people with cancer can now expect to be alive 5 years after diagnosis. However, approximately two thirds of survivors experience at least one late effect, and approximately one third experience a late effect that is severe or life threatening.1 At least twelve publications, in some cases involving the same cohorts described at different time points, have reported that five-year survivors of childhood cancer have a 7–17-fold increased risk of mortality from all causes, compared to the general population.2–13 The excess mortality is not limited to cancer but also applies to non-cancer causes of death. Many studies have demonstrated an increased risk of second primary cancers arising in survivors of childhood cancer.14–21 Other long-term complications include effects on the endocrine, cardiac and pulmonary systems, renal impairment, gastrointestinal dysfunction, musculoskeletal sequelae, neurocognitive dysfunction, and psychosocial manifestations.1,22–24 Long term morbidity and mortality risks in childhood cancer survivors relate largely to treatment modality, although they may also be determined by individual host characteristics.1 The fundamental challenge among this group of patients is to further improve survival prospects while minimising the incidence and severity of treatment-induced late effects.25 Although the risk of second primary malignancies following childhood cancer has been extensively studied in the United Kingdom (UK) and elsewhere,14–21 less attention has been devoted to other late effects of therapy, such as cardiac disease. A major active follow-up study, the British Childhood Cancer Survivor Study,26 is currently underway, and although this is gathering very detailed data, it may not be entirely representative since it relies on patients consenting to participate. A single centre study from Sheffield, UK suggested that survivors of childhood cancer may have a tendency to over-report serious late effects, such as second primary cancers, compared to what is documented in hospital medical records.27 A recent report on cardiac outcomes from five-year survivors of selected childhood cancers in the United States (US) reported increased risks of selfreported congestive heart failure, myocardial infarction, pericardial disease and valvular abnormalities compared to siblings.23 The authors noted that the proportion of deaths among eligible participants who refused to participate or who were lost to follow-up was higher than
among study participants suggesting that the reported risks may be underestimates. On the other hand, studies from specialist centres tend to overestimate the prevalence of chronic disease,28 emphasising the need for studies to be population-based, if possible. Compared to follow-up of younger children, considerably less research has been carried out into late effects of therapy for cancer diagnosed during adolescence or young adulthood,4,29,30 although there is evidence that around half of survivors of adolescent cancer experience late effects of therapy.31 The aim of the present study was to describe patterns of mortality among a cohort of patients who have survived at least 5 years after the diagnosis of cancer in childhood, adolescence or young adulthood in Scotland. 2. Patients and methods The study population comprised patients registered with the Scottish Cancer Registry who survived at least 5 years after the diagnosis of cancer in childhood, adolescence, or young adulthood (age between 0 and 24 years). We decided to study this combined age group in the interests of greater statistical power, and because this age group is now managed in the context of a single national network of clinicians in Scotland (the ‘Managed Service Network for Children and Young People with Cancer’). In Scotland, cancer registration records are linked to mortality records by computerised probability matching, which is believed to be highly accurate.32 Emigrations of patients registered with cancer from Scotland to other UK countries are notified to the cancer registry by the National Health Service Central Register (NHSCR). We included data for individuals whose year of diagnosis was in 1981 (the first year of the linked database in Scotland)32 or subsequent years up to 2003 (to allow 5 years’ survival, and at least one further year’s follow-up). Follow-up was from 5 years after diagnosis to date of emigration from Scotland, date of death, or end of 2009, whichever occurred first. In terms of diagnosis of first cancer, the study population was re-classified according to the third edition of the International Classification of Childhood Cancer (ICCC-3) (0–14-year-olds)33 and the most up-to-date version of the diagnostic classification of cancer in adolescents and young adults (15–24-year-olds) developed by Birch et al.34 The data were then mapped to a common ‘study classification’. Briefly, this was based primarily on ICCC-3 but with separation of group XI (Other malignant epithelial neoplasms and malignant
3276
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
melanomas) into two categories: (1) Melanoma and skin carcinomas and (2) Other carcinomas, excluding renal, hepatic, gonadal, and skin. Mapping of cancers in adolescents and young adults was based initially on specific morphology codes (sometimes in combination with topography codes); thereafter the mapping was based on groups and sub-groups within the Birch et al. classification (see Appendix A). Individuals were assigned to fifths of Carstairs deprivation scores, an indicator of socio-economic position based on postcode sectors of residence at the time of original diagnosis, by applying 1981, 1991 and 2001 census-derived Carstairs scores to the periods of diagnosis 1981–1985, 1986–1995, 1996–2003, respectively. The Carstairs deprivation index is based on small area of residence, and is derived from four variables collected at each decennial census: social class, unemployment, overcrowding and car ownership.35 The characteristics of patients surviving beyond 5 years were compared to non-survivors, stratified by age at diagnosis, sex, socio-economic position at diagnosis and tumour type using the chi-squared test of associ-
ation. Both for all causes of death combined and for specific causes of death (see Appendix B), indirectly standardised mortality ratios (SMRs) were calculated, using the general population as an external comparison group to generate expected numbers of deaths (based on age-, sex-, deprivation category-, and calendar periodspecific rates of disease). Absolute excess risks (AERs) were calculated as the observed minus the expected number of deaths divided by the number of person-years at risk and expressed as the rate per 10,000 cancer survivors per year. The AER reflects the additional mortality burden beyond background mortality. 95% confidence intervals (CIs) around SMRs and AERs were calculated based on the assumption that the observed numbers of deaths followed a Poisson distribution. SMRs with 95% confidence intervals that did not include the value 1.0 were regarded as statistically significant. Cumulative mortality was estimated using the Kaplan–Meier method,36 and differences in cumulative mortality were assessed for statistical significance using the log rank test. Cox proportional hazards modelling was used to assess whether the hazard of death decreased over time.
Table 1 Characteristics of the original cohort including the study population (the five-year survivors). v2
Characteristic
Five-year survivors Number
%
Number
%
All patients combined
5229
75
1751
25
Age at diagnosis of first cancer (years) <1 1–4 5–9 10–14 15–19 20–24
188 802 555 631 1043 2010
65 74 72 72 72 80
102 280 216 241 408 504
35 26 28 28 28 20
63.7
0.000
Sex Male Female
2762 2467
73 77
1004 747
27 23
10.8
0.001
Carstairs deprivation fifth 1 – least deprived 2 3 4 5 – most deprived
1087 1005 1040 1055 1036
77 74 76 75 72
323 344 321 359 403
23 26 24 25 28
11.9
0.018
884 980 587 93 81 154 19 188 317 694 709 459 64
69 82 61 44 96 83 45 56 70 89 93 76 79
401 209 383 116 3 31 23 150 133 89 54 142 17
31 18 39 56 4 17 55 44 30 11 7 24 21
601.6
0.000
Type of first cancer Leukaemias Lymphomas CNS tumours Neuroblastoma Retinoblastoma Renal tumours Hepatic tumours Bone tumours Soft tissue sarcomas Germ cell tumours Melanoma and skin carcinomas Other carcinomas Other and unspecified neoplasms
Non-survivors
P-Value
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
3. Results From an original cohort of 6980 children and young people with cancer, the study population comprised 5229 individuals who had survived for at least 5 years after the date of diagnosis of their primary cancer. Only five patients were recorded as emigrating from Scotland within 5 years of the original diagnosis of cancer, and only 37 beyond 5 years after diagnosis. The proportion of survivors varied significantly according to type of primary cancer, and was significantly higher in those diagnosed at an older age than those diagnosed at a younger age, females compared to males, and the least compared to the most deprived fifth (Table 1). During 58,358 person-years of follow-up, there were 359 deaths (Table 2). The overall age-, sex-, and deprivation standardised mortality ratio (SMR) was 6.2 (95% CI 5.6–6.9). In general, SMRs were little altered by standardisation for deprivation. For example, the overall SMR, standardised for age and sex only was 6.1 (5.5– 6.7). For the sake of comparability with other studies, the SMRs presented hereafter are standardised for age and sex only. Compared to background rates, the relative risk of death was higher in patients who were diagnosed with cancer in the age range 0–14 years (SMR 11.0, 95% CI 9.3–12.9) than in those diagnosed between the ages of 15 and 24 years (4.7, 4.1–5.3). SMRs were higher in females (9.2, 7.8–10.8) than males (4.8, 4.2– 5.5), and were particularly high in patients with the following types of primary cancer: leukaemias (11.7, 9.0– 14.8), CNS tumours (14.7, 11.5–18.4), hepatic tumours (65, 18–167), and bone tumours (10.3, 6.5–15.6). The overall AER, standardised for age and sex only, was 51 (45–58) per 10,000 person-years. AERs varied substantially by primary cancer (Table 2). Neoplasms were identified as the cause of death for the majority (76%) of deaths, 65% due to the primary cancer and 11% due to subsequent cancers. Other causes of death included injury, poisoning, and other external factors (15%), and diseases of the circulatory system (4.5%). The SMR for death from cancer was higher (32, 28– 36) than for non-cancer deaths (1.7, 1.3–2.1). The SMR for death from a subsequent cancer (a subset of all cancer deaths) was 4.6 (3.2–6.2). Table 3 shows the numbers of deaths, SMRs and AERs for those major causes of death with statistically significantly increased SMRs. SMRs were significantly increased in respect of deaths from diseases of the nervous system (2.9, 1.2–6.0), diseases of the circulatory system (2.4, 1.4–3.9), diseases of the respiratory system (3.4, 1.2–7.4), and congenital malformations, deformations and chromosomal abnormalities (4.7, 1.3–11.9). In respect of the remaining 52 deaths (not included in Table 3), no significantly increased SMRs were found for any of the following causes of death: certain infectious and parasitic diseases;
3277
endocrine, nutritional and metabolic diseases; diseases of the digestive system; diseases of the genitourinary system; injury, poisoning and certain other consequences of external causes; external causes of mortality; or any other causes. Twelve deaths were attributed to suicide (SMR 1.1, 0.6–2.0). No deaths were recorded from interstitial pneumonitis and/or pulmonary fibrosis caused by radiation or drugs. AERs also varied considerably according to the underlying cause of death (Table 3). Cumulative mortality was significantly higher in patients diagnosed during 1981–1985, compared to patients diagnosed during 1986–1991, 1992–1997 and 1998–2003 (all P < 0.001) (Figure 1). However, other pair-wise comparisons did not show any statistically significant differences. Adjusted for age and sex, the risk of death was significantly lower in five-year survivors diagnosed during 1998–2003 compared to those diagnosed during 1981–1985 (Relative hazard ratio, 0.54, 95% CI 0.36–0.81).
4. Discussion Using routinely collected data, we have been able to estimate mortality of survivors of cancer at a young age relative to the whole population, after adjusting for age and sex. In spite of surviving at least 5 years after their primary cancer, this group of patients remains at significantly increased risk of death compared to their peers. However, the absolute risk of death is low, reflecting relatively low background risks of death among young people. Our cohort study has a number of strengths. Most importantly, it was population-based, and not, therefore, subject to some of the biases associated with single centre studies or surveys that depend on participation. The study has been performed against the background of a longstanding national health service funded primarily through taxation and free at the point of use. Scottish Cancer Registry data have been shown to be of comparatively high quality,37,38 and the linkage with official mortality records is believed to be highly accurate and complete.32 Unlike many other studies, we were able to standardise our analyses for socio-economic deprivation. Importantly, this made little difference to the results, although this observation may reflect limitations of the indicator of deprivation that we used, which was area-based (rather than based on individual characteristics) and was relevant at the time of diagnosis of cancer.35 It is complicated by the fact that the experience of cancer at a young age may impact negatively on adult socio-economic position and information on socio-economic position during follow-up was not available in this study.39 Ideally, the impact of socio-economic position on later mortality should be assessed in other
3278
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
Table 2 Numbers of person-years at risk and deaths, and standardised mortality ratios (SMRs) and absolute excess risks (AERs)* with 95% confidence intervals (CIs) among five-year survivors according to various characteristics. SMR
95% LCI**
95% UCI**
AER*
100.0
6.1
5.5
6.7
51
45
58
7 47 42 49 76 138 145 214
1.9 13.1 11.7 13.6 21.2 38.4 40.4 59.6
12.2 13.5 11.8 8.8 5.8 4.2 11.0 4.7
4.9 9.9 8.5 6.5 4.6 3.5 9.3 4.1
25.1 18.0 16.0 11.6 7.3 5.0 12.9 5.3
32 48 66 66 53 46 56 48
11 35 46 47 39 36 47 40
68 66 91 90 69 57 67 57
30,929.9 27,428.7
202 157
56.3 43.7
4.8 9.2
4.2 7.8
5.5 10.8
52 51
43 42
61 61
12,306.9 11,483.5 11,616.6 11,757.4 11,117.9
66 55 75 86 77
18.4 15.3 20.9 24.0 21.4
5.4 4.6 6.3 7.5 6.8
4.2 3.5 5.0 6.0 5.3
6.9 6.0 7.9 9.3 8.4
44 38 54 63 59
32 26 41 49 44
58 52 71 81 76
9238.4 11,423.9 6309.2 1010.2 1058.8 1764.0 177.4 1896.2 3609.0 8154.2 7511.2
67 73 75 3 3 6 4 22 24 26 25
18.7 20.3 20.9 0.8 0.8 1.7 1.1 6.1 6.7 7.2 7.0
11.7 5.2 14.7 7.3 8.0 7.1 65.2 10.3 8.2 2.1 3.1
9.0 4.1 11.5 1.5 1.7 2.6 17.8 6.5 5.2 1.4 2.0
14.8 6.5 18.4 21.3 23.4 15.5 167.0 15.6 12.1 3.1 4.6
66 52 111 26 25 29 222 105 58 17 23
50 38 85 2 2 8 57 62 35 6 11
86 68 141 82 80 69 561 165 91 32 39
5234.6 971.4
28 3
7.8 0.8
4.8 2.4
3.2 0.5
6.9 7.1
42 18
24 6
66 77
Characteristic
Person-years at risk
Numbers of deaths
%
All patients combined
58,358.6
359
Age at diagnosis of first cancer (years) <1 2012.2 1–4 9013.3 5–9 5871.2 10–14 6543.2 15–19 11,920.5 20–24 22,998.2 0–14 23,439.9 15–24 34,918.7 Sex Male Female Carstairs deprivation fifth 1 – least deprived 2 3 4 5 – most deprived Type of first cancer Leukaemias Lymphomas CNS tumours Neuroblastoma Retinoblastoma Renal tumours Hepatic tumours Bone tumours Soft tissue sarcomas Germ cell tumours Melanoma and skin carcinomas Other carcinomas Other and unspecified neoplasms * **
95% LCI**
95% UCI**
AER – absolute excess risk per 10,000 person years. LCI – lower confidence interval; UCI – upper confidence interval.
populations using alternative indicators which can be measured longitudinally. The main weakness of our study is the very limited information available on treatment. Thus, we can only
make inferences about treatment based on era of diagnosis. In relation to outcomes, causes of death were not validated, although this limitation also, of course, applies to the reference population used to calculate
Table 3 Numbers of deaths, standardised mortality ratios (SMRs) and absolute excess risks (AERs)* with 95% confidence intervals (CIs) for selected (significantly increased) causes of death among five-year survivors. AER*
95% LCI**
36.1 6.2
5.2 0.0
3.3 0.2
7.7 0.8
6.0 3.9 7.4 11.9
0.1 0.8 0.1 0.3
0.1 0.1 0.2 0.0
0.5 2.1 1.0 1.2
Primary cause of death
Numbers of deaths
SMR
95% LCI**
95% UCI**
Neoplasms (all cancers) – Neoplasms (subsequent cancers only) Diseases of the nervous system Diseases of the circulatory system Diseases of the respiratory system Congenital malformations, deformations and chromosomal abnormalities
274 39
32.0 4.6
28.4 3.2
7 16 6 4
2.9 2.4 3.4 4.7
1.2 1.4 1.2 1.3
* **
AER – absolute excess risk per 10,000 person years. LCI – lower confidence interval; UCI – upper confidence interval.
95% UCI**
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283 18%
16%
Percentage Cumulative Mortality
14%
12%
10%
8%
6%
4%
2%
0%
0
5
10
15
20
25
Time in Years 1981-1985
1986-1991
1992-1997
1998-2003
Fig. 1. Cumulative mortality among five-year survivors by period of diagnosis of first cancer.
expected numbers of deaths. However, it must be acknowledged that some causes of death may have been incorrectly assigned. For example, in a previous study from the Nordic countries, 12% (148 of 1208) of fatalities attributed to the first malignancy were re-classified as non-cancer deaths, and 15 cases officially recorded as death due to malformation were re-classified as deaths caused by primary cancer (almost all these cases being neurofibromatosis type I).12 Note, however, that neurofibromatosis was classified under the neoplasms chapter of the ninth revision of the International Classification of Diseases (ICD9); in Scotland, neurofibromatosis was only classified as a congenital malformation from the year 2000 when ICD10 coding of death certificates was introduced. Our study population is smaller than some other studies, which limits the capacity to carry out sub-group analyses. Inevitably, the length of follow-up available for the most recently diagnosed patients is necessarily limited. While it is possible that some members of the cohort may have emigrated from Scotland without this being recorded and are therefore lost to follow-up, we believe that the proportion of unrecorded emigrations from the cohort (at any time after diagnosis of the primary cancer) is low. Other publications have reported that five-year survivors of childhood cancer (variously defined) have
3279
a 7–17-fold increased risk of mortality from all causes, compared to the general population.2–13 In our study the overall SMR for survivors of childhood cancer was 11.0, which lies just below the middle of the range of estimates reported from other studies. A recently published study of five-year survivors of young adult (20– 24-year-olds) cancer from British Columbia reported an overall SMR of 5.9 (4.9–6.9) and AER of 5.3 per 1000 years at risk (or 53 per 10,000 years at risk), which is of a similar order of magnitude to our own findings in 15–24-year-olds (SMR 4.7, 4.1–5.3; AER 48, 40–57 per 10,000 years at risk).29 In comparing studies, it is important to note that there is some heterogeneity in terms of age group at diagnosis of primary cancer, calendar period of diagnosis, and period of follow-up. However, the finding of a higher SMR in females, reflecting lower background mortality in young adulthood compared to males, is consistent across studies. Lower background mortality in children compared to adolescents/young adults is also likely to explain the higher SMR seen in survivors of childhood cancer. Recent studies of longterm mortality in Nordic and British survivors of childhood cancer showed that excess mortality continued to occur even beyond 25 years after diagnosis.2,3 However, other studies have, like our own, shown a reduced risk of death among five-year survivors diagnosed with cancer in the more recent past compared to historical cohorts.2,6,12 Associations between congenital anomalies and/or genetic factors and cancer at a young age have been established in many studies.40 This might explain our observation of an increased risk of death (SMR 4.7) from congenital malformations, deformations and chromosomal abnormalities. However, as noted previously, it is also important to consider the possibility of misdiagnosis or misclassification of the underlying cause of death in some circumstances. For example, hydrocephalus acquired as a consequence of a brain tumour might be incorrectly diagnosed and/or coded as a congenital malformation. The Scottish Intercollegiate Guidelines Network has recently published an updated guideline on the late effects of the treatment of cancer in young people.41 The guideline development group has specifically stated that health care professionals should be aware that survivors of childhood cancer are at particular and lifelong increased risk of development of a subsequent primary cancer and that this may occur at any site on the body. The guideline group were not able to identify any evidence of either benefits or harms of specific screening programmes for survivors of childhood cancer, nor were any studies identified on outcomes for survivors of childhood cancer entering national screening programmes (e.g. mammography for breast cancer) at an earlier age than for general population groups.
3280
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
If patients, parents, carers and clinicians are to make fully informed decisions about treatment and follow-up, it is important to quantify the long-term outcomes of conditions. Our study has confirmed that long-term survivors of cancer in childhood and young adulthood remain at higher risk (albeit low absolute risk) of mortality compared to their peers. It has also demonstrated the utility of routinely collected, linked records for monitoring many fatal late effects at a relatively low cost compared to the costs of active follow-up. The results of such high level monitoring may identify specific effects that warrant more detailed investigation. In future, as increasing numbers of patients survive their cancer, high level monitoring of late effects, both fatal and non-fatal (for example, from hospitalisation records), should become as routine as monitoring five-year survival. In future, it may also be possible to capture more detailed treatment information through computerised radiotherapy planning systems and systemic anti-cancer therapy prescribing systems. This would enable risk estimates to be related to specific types of treatment. Conflict of interest statement None declared. Acknowledgements This study was supported by a grant from the Chief Scientist Office, Scottish Government Health and Social
Care Directorates (CZG/2/508). Beyond arranging review of the original grant application and providing funding, the Chief Scientist Office had no role in any other part of the study. Appendix A Study classification of primary neoplasm and mappings from ICCC-3 and teenagers and young adults with cancer (TYAC) classifications. Study classification of primary neoplasm. Study classification Leukaemias Lymphomas CNS tumours Neuroblastoma Retinoblastoma Renal tumours Hepatic tumours Bone tumours Soft tissue sarcomas Germ cell tumours Melanoma and skin carcinomas Other carcinomas* Other and unspecified neoplasms *Except
renal, hepatic, gonadal, and skin.
Mapping of Childhood Cancer Classification (ICCC-3) to study classification. ICCC-3 classification
Study classification
Group I – leukaemias, myeloproliferative diseases, and myelodysplastic diseases Group II – lymphomas and reticuloendothelial neoplasms Group III – CNS and miscellaneous intracranial and intraspinal neoplasms Group IV – neuroblastoma and other peripheral nervous cell tumours Group V – retinoblastoma Group VI – renal tumours Group VII – hepatic tumours Group VIII – malignant bone tumours Group IX – soft tissue and other extraosseous sarcomas Group X – germ cell tumours, trophoblastic tumours, and neoplasms of gonads Group XI – other malignant epithelial neoplasms and malignant melanomas
All Group I = leukaemias
Group XII – other and unspecified malignant neoplasms
All Group II = lymphomas All Group III = CNS tumours All Group IV = neuroblastoma All All All All All All
Group Group Group Group Group Group
V = retinoblastoma VI = renal tumours VII = hepatic tumours VIII = bone tumours IX = soft tissue sarcomas X = germ cell tumours
Category XId = melanoma and skin carcinomas Category XIe = melanoma and skin carcinomas Remainder = other carcinomas All Group XII = other and unspecified neoplasms
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
3281
Appendix A (continued) Mapping of TYAC classification to study classification *Any
cases with ICDO morphology codes 9510/3–9514/3 = retinoblastoma. cases with ICDO morphology codes 8970/3 = hepatic tumours. *Any cases with ICDO morphology codes 8964/3 = renal tumours. *Any cases with ICDO morphology codes 8963/3 + ICD9 189.0 or ICD10 C64 = renal tumours. *Any remaining cases with ICDO morphology codes 8963/3 = soft tissue sarcomas. *Any cases with ICDO morphology codes 9501–9504 + ICD9 191–192, 237.5, 237.6, 237.9 or ICD10 C70.0–C72.9, D32–D33, D42–D43 = CNS tumours. *Any remaining cases with ICDO morphology codes 9501/3–9504/3 = neuroblastoma. *Any cases with ICDO morphology codes 9505–9508 = CNS tumours. *Any cases with ICDO morphology codes 9520/3–9523/3 = neuroblastoma. *Any
Thereafter: TYAC classification
Study classification
Group 1 – leukaemia All Group 1 = leukaemias Group 2 – lymphoma All Group 2 = lymphomas Group 3 – central nervous system and other intracranial and All Group 3 = CNS tumours intraspinal neoplasms (CNS tumours) Group 4 – osseous and chondromatous neoplasms, Ewing tumour All Group 4 = bone tumours and other neoplasms of bone (bone tumours) Group 5 – soft tissue sarcomas (STS) All Group 5 = soft tissue sarcomas Group 6 – germ cell and trophoblastic neoplasms (germ cell tumours) All Group 6 = germ cell tumours Group 7 – melanoma and skin carcinoma All Group 7 = melanoma and skin carcinomas Group 8 – carcinomas (except of skin) Category 8.5.1 = renal tumours Category 8.5.3 = germ cell tumours Category 8.6.3 = hepatic tumours Remainder of Group 8 = other carcinomas Group 9 – miscellaneous specified neoplasms NEC (miscellaneous Category 9.1.1 = renal tumours specified) Category 9.1.2 = neuroblastoma Category 9.2.1 = neuroblastoma Category 9.2.2 = germ cell tumours Category 9.2.3 = lymphomas Remainder of Group 9 = other and unspecified neoplasms Group 10 – unspecified malignant neoplasms NEC (unspecified) All Group 10 = other and unspecified neoplasms Appendix B Disease-specific outcome codes from mortality records (primary underlying cause). Disease grouping
ICD9
ICD10
Certain infectious and parasitic diseases Neoplasms – the primary neoplasm* – subsequent neoplasm(s)* Endocrine, nutritional and metabolic diseases – Diabetes mellitus Diseases of the nervous system Diseases of the circulatory system – Coronary heart disease
001–139 140–239
A00–B99 C00–D48
240–279 250 320–359 390–459 410–414
E00–E90 E10–E14 G00–G99 excluding G45–G46 I00–I99, G45–G46 I20–I25 (continued on next page)
3282
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283
Appendix B (continued) – Cardiomyopathy – Cerebrovascular disease Diseases of the respiratory system – Interstitial pneumonitis and/or pulmonary fibrosis caused by radiation or drugs Diseases of the digestive system Diseases of the genitourinary system Congenital malformations, deformations and chromosomal abnormalities Injury, poisoning and certain other consequences of external causes** External causes of morbidity and mortality** – Suicide (DSH and undetermined intent) – Accidental poisoning – All other accidents
425 430–438 460–519 508
I25.5, I41.2, I42, I43 I60–I69, G45–G46 J00–J99 J70
520–579 580–629 740–759
K00–K93 N00–N99 Q00–Q99
800–999
S00–T98
E800–E999 E950–E959, E980–E989 E850–E869 E800–E849, E870–E949
V01–Y98 X60–X84, Y87.0, Y10–Y34, Y87.2 X40–X49 V01–V99, W00–W99, X00–X39, X50– X59, Y40–Y59, Y85, Y86
Other causes by inspection of primary and subsequent ICD9/10 codes. **In general, the external cause codes take precedence over the injury etc codes in mortality data. *Classified
References 1. Bhatia S, Constine LS. Late morbidity after successful treatment of children with cancer. Cancer J 2009;15:174–80. 2. Garwicz S, Anderson H, Olsen JH, et al. Late and very late mortality in 5-year survivors of childhood cancer: changing pattern over four decades – experience from the Nordic countries. Int J Cancer 2012;131:1659–66. 3. Reulen RC, Winter DL, Frobisher C, et al. Long-term causespecific mortality among survivors of childhood cancer. JAMA 2010;304:172–9. 4. Prasad PK, Signorello LB, Friedman DL, Boice Jr JD, Pukkala E. Long-term non-cancer mortality in pediatric and young adult cancer survivors in Finland. Pediatr Blood Cancer 2012;58:421–7. 5. Wilson CL, Cohn RJ, Johnston KA, Ashton LJ. Late mortality and second cancers in an Australian cohort of childhood cancer survivors. Med J Aust 2010;193:258–61. 6. Armstrong GT, Pan Z, Ness KK, Srivastava D, Robison LL. Temporal trends in cause-specific late mortality among 5-year survivors of childhood cancer. J Clin Oncol 2010;28:1224–31. 7. Armstrong GT, Liu Q, Yasui Y, et al. Late mortality among 5year survivors of childhood cancer: a summary from the Childhood Cancer Survivor Study. J Clin Oncol 2009;27:2328–38. 8. Mertens AC, Liu Q, Neglia JP, et al. Cause-specific late mortality among 5-year survivors of childhood cancer: the Childhood Cancer Survivor Study. J Natl Cancer Inst 2008;100:1368–79. 9. Dama E, Pastore G, Mosso ML, et al. Late deaths among fiveyear survivors of childhood cancer. A population-based study in Piedmont Region, Italy. Haematologica 2006;91:1084–91. 10. Cardous-Ubbink MC, Heinen RC, Langeveld NE, et al. Longterm cause-specific mortality among five-year survivors of childhood cancer. Pediatr Blood Cancer 2004;42:563–73. 11. Mertens AC, Yasui Y, Neglia JP, et al. Late mortality experience in five-year survivors of childhood and adolescent cancer: the Childhood Cancer Survivor Study. J Clin Oncol 2001;19:3163–72. 12. Mo¨ller TR, Garwicz S, Barlow L, et al. Decreasing late mortality among five-year survivors of cancer in childhood and adolescence: a population-based study in the Nordic countries. J Clin Oncol 2001;19:3173–81.
13. Hudson MM, Jones D, Boyett J, Sharp GB, Pui CH. Late mortality of long-term survivors of childhood cancer. J Clin Oncol 1997;15:2205–13. 14. Reulen RC, Frobisher C, Winter DL, et al. Long-term risks of subsequent primary neoplasms among survivors of childhood cancer. JAMA 2011;305:2311–9. 15. Olsen JH, Moller T, Anderson H, et al. Lifelong cancer incidence in 47,697 patients treated for childhood cancer in the Nordic countries. J Natl Cancer Inst 2009;101:806–13. 16. Olsen JH, Garwicz S, Hertz H, et al. Second malignant neoplasms after cancer in childhood or adolescence. Nordic Society of Paediatric Haematology and Oncology Association of the Nordic Cancer Registries. BMJ 1993 Oct 23;307(6911):1030–6. 17. Garwicz S, Anderson H, Olsen JH, et al. Second malignant neoplasms after cancer in childhood and adolescence: a population-based case-control study in the 5 Nordic countries. The Nordic Society for Pediatric Hematology and Oncology. The Association of the Nordic Cancer Registries. Int J Cancer 2000;88:672–8. 18. Jenkinson HC, Hawkins MM, Stiller CA, Winter DL, Marsden HB, Stevens MC. Long-term population-based risks of second malignant neoplasms after childhood cancer in Britain. Br J Cancer 2004;91:1905–10. 19. de Vathaire F, Hawkins M, Campbell S, et al. Second malignant neoplasms after a first cancer in childhood: temporal pattern of risk according to type of treatment. Br J Cancer 1999;79:1884–93. 20. Kingston JE, Hawkins MM, Draper GJ, Marsden HB, Kinnier Wilson LM. Patterns of multiple primary tumours in patients treated for cancer during childhood. Br J Cancer 1987;56:331–8. 21. Hawkins MM, Draper GJ, Kingston JE. Incidence of second primary tumours among childhood cancer survivors. Br J Cancer 1987;56:339–47. 22. Edgar AB, Morris EM, Kelnar CJ, Wallace HB. Long-term follow-up of survivors of childhood cancer. Endocr Dev 2009;15:159–80. 23. Mulrooney DA, Yeazel MW, Kawashima T, et al. Cardiac outcomes in a cohort of adult survivors of childhood and adolescent cancer: retrospective analysis of the Childhood Cancer Survivor Study cohort. BMJ 2009;339:b4606.
D.H. Brewster et al. / European Journal of Cancer 49 (2013) 3274–3283 24. Recklitis CJ, Diller LR, Li X, Najita J, Robison LL, Zeltzer L. Suicide ideation in adult survivors of childhood cancer: a report from the Childhood Cancer Survivor Study. J Clin Oncol 2010;28:655–61. 25. Scottish Intercollegiate Guidelines Network. Long term follow up of survivors of childhood cancer (SIGN 76). Edinburgh: SIGN; 2004. http://www.sign.ac.uk//pdf/sign76.pdf. 26. Hawkins MM, Lancashire ER, Winter DL, et al. The British Childhood Cancer Survivor Study: objectives, methods, population structure, response rates and initial descriptive information. Pediatr Blood Cancer 2008;50:1018–25. 27. Taylor N, Absolom K, Michel G, et al. Comparison of selfreported late effects with medical records among survivors of childhood cancer. Eur J Cancer 2010;46:1069–78. 28. Ness KK, Leisenring W, Goodman P, et al. Assessment of selection bias in clinic-based populations of childhood cancer survivors: a report from the childhood cancer survivor study. Pediatr Blood Cancer 2009;52:379–86. 29. Zhang Y, Goddard K, Spinelli JJ, Gotay C, McBride ML. Risk of late mortality and second malignant neoplasms among 5-year survivors of young adult cancer: a report of the childhood, adolescent, and young adult cancer survivors research program. J Cancer Epidemiol 2012;2012:103032. 30. Woodward E, Jessop M, Glaser A, Stark D. Late effects in survivors of teenage and young adult cancer: does age matter? Ann Oncol 2011;22:2561–8. 31. Pentheroudakis G, Pavlidis N. Late toxicity in survivors from adolescent cancers. Cancer Treat Rev 2007;33:656–63.
3283
32. Kendrick S, Clarke J. The Scottish record linkage system. Health Bull (Edinb) 1993;51:72–9. 33. Steliarova-Foucher E, Stiller C, Lacour B, Kaatsch P. International Classification of Childhood Cancer, third edition. Cancer 2005;103:1457–67. 34. Birch JM, Alston RD, Kelsey AM, Quinn MJ, Babb P, McNally RJQ. Classification and incidence of cancers in adolescents and young adults in England 1979–1997. Br J Cancer 2002;87:1267–74. 35. Morris R, Carstairs V. Which deprivation? A comparison of selected deprivation indexes. J Public Health Med 1991;13:318–26. 36. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81. 37. Brewster D, Crichton J, Harvey JC, Dawson G. Completeness of case ascertainment in a Scottish Regional Cancer Registry for the year 1992. Public Health 1997;111:339–43. 38. Brewster DH, Stockton D, Harvey J, Mackay M. Reliability of cancer registration data in Scotland, 1997. Eur J Cancer 2002;38:414–7. 39. Holmqvist AS, Wiebe T, Hjorth L, Lindgren A, Ora I, Moell C. Young age at diagnosis is a risk factor for negative late socioeconomic effects after acute lymphoblastic leukemia in childhood. Pediatr Blood Cancer 2010;55:698–707. 40. Little J. Epidemiology of Childhood Cancer. IARC Scientific Publications No. 149. Lyon: International Agency for Research on Cancer; 1999. 41. Scottish Intercollegiate Guidelines Network. Long term follow-up of survivors of childhood cancer (SIGN 132). Edinburgh: SIGN; 2013.