Measuring childhood cancer late effects: evidence of a healthy survivor effect

Introduction

Given considerable focus on health outcomes among childhood cancer survivors, we aimed to explore whether survivor bias is apparent during long-term follow-up of childhood cancer survivors.

Methods

We identified all 1-year survivors of cancer diagnosed before 20 years of age in Denmark, Finland, Iceland, and Sweden. From the general population, we randomly sampled a comparison cohort. Study individuals were followed for hospitalizations for diseases of the gastroenterological tract, endocrine system, cardiovascular system, or urinary tract from the start of the cancer registries to 2010. We estimated cumulative incidence with death as competing risk and used threshold regression to compare the hazards of the diseases of interest at ages 20, 40, 60, and 75 years.

Results

Our study included 27,007 one-year survivors of childhood cancer and 165,620 individuals from the general population. The cumulative incidence of all four outcomes was higher for childhood cancer survivors during early adulthood, but for three outcomes, the cumulative incidence was higher for the general population after age 55 years. The hazard ratios (HRs) decreased for all outcomes with increasing age, and for two of the outcomes, the hazards were higher for the general population at older ages (endocrine diseases: age-specific HRs = 3.0, 1.4, 1.0, 0.87; Cardiovascular diseases: age-specific HRs = 4.1, 1.4, 0.97, 0.84).

Conclusions

Our findings provide empirical evidence that survivor bias attenuates measures of association when comparing survivors with the general population. The design and analysis of studies among childhood cancer survivors, particularly as this population attains older ages, should account for survivor bias to avoid misinterpreting estimates of disease burden.

     

Cancer survival in Estonian migrants to Sweden.

OBJECTIVE: To quantify the eventual extra loss of life incurred to cancer patients in Estonia compared with those in Sweden that was possibly attributable to differences in society. DESIGN: Population based survival of cancer patients in Estonia was compared with that of Estonian immigrants to Sweden and that of all cancer patients in Sweden. The cancer sites studied were female breast and ovary, male lung and prostate, and male and female stomach and colon. SETTING: Data on incident cases of cancer were obtained from the population based Swedish and Estonian cancer registries. PARTICIPANTS: Data from Estonian patients in Sweden, Estonian patients in Estonia, and patients from the total Swedish population were included in the study. MAIN RESULTS: Differences in survival among the three populations, controlling for follow-up time and age at diagnosis, were observed in breast, colon, lung, ovarian, and prostate cancers. The survival rates of Estonians living in Sweden and the total population of Sweden were better than that of the Estonians living in Estonia. For cancers of the breast and prostate, the excess mortality in the older age group (75 and above) was much greater in Estonia than in the other populations. CONCLUSIONS: Most differences in cancer survival between Estonian and Swedish populations studied could be attributed to a longer delay in diagnosis, and also to inferior treatment (including access to treatment) in Estonia compared with Sweden. Estonia's lag in socioeconomic development, particularly in its public health organisation and funding, is probably the main source of the differences observed.     

Parental age as a risk factor of childhood leukemia and brain cancer in offspring

We use here the Swedish Family-Cancer Database to analyze the time trends in childhood leukemia and brain cancer between 1960 and 1994 and the effect of parental age on childhood leukemia and brain cancer of some 1500 cases each. The database includes all persons born in Sweden after 1940 with their biological parents, over 6 million individuals, whose cancers were retrieved from the Swedish Cancer Registry from years 1958-1994. Incidence in cancer increased from 1960 to 1994; low grade astrocytoma accounted for most of the increase, whereas high grade astrocytoma has not increased in incidence. There has been a moderate increase in leukemia to about 1980. We found a parental age effect for both leukemia and brain cancer, with the former (of about 50% excess in those over 35 years) being mediated by maternal age and the latter (of about 25% excess) by paternal age. Accumulation of chromosomal aberrations and mutations during the maturation of germ cells is a likely mechanism for these findings. They can help to explain partially the secular trends of these malignancies and the excess risks in offspring of the well educated.     

Mortality among 723,948 foreign- and native-born Swedes 1970-1999

BACKGROUND: Mortality in a population is regarded as an accurate and valid measure of the population's health. There are a few international studies, predominantly cross-sectional, of mortality among all foreign-born compared with an indigenous population, and the results have varied. No Swedish longitudinal study describing and analysing mortality data was found in a literature review. METHODS: This study describes and analyses the differences in mortality between foreign-born persons and native Swedes during the period 1970-1999, based on data from Statistics Sweden and the National Board of Health and Welfare. The database consisted of 723,948 persons, 361 974 foreign-born living in Sweden in 1970, aged > or = 16 years, and 361 974 Swedish controls matched for age, sex, occupation and type of employment, living in the same county in 1970. RESULTS: The results showed increased mortality for foreign-born persons compared with the Swedish controls [odds ratio (OR) 1.08; 95% confidence interval (CI) 1.07-1.08]. Persons who had migrated 'late' (1941-1970) to Sweden were 2.5 years younger at time of death than controls. In relation to country of birth, the highest risk odds were for men born in Finland (OR 1.21), Denmark (OR 1.11) and Norway/Iceland (OR 1.074). Age cohorts of foreign-born persons born between 1901 and 1920 had higher mortality at age 55-69 years than cohorts born between 1921 and 1944. CONCLUSIONS: Migrants had higher mortality than the native population, and migration may be a risk factor for health; therefore, this seems to be an important factor to consider when studying mortality and health.     

Sensitivity analysis to investigate the impact of a missing covariate on survival analyses using cancer registry data

Having substantial missing data is a common problem in administrative and cancer registry data. We propose a sensitivity analysis to evaluate the impact of a covariate that is potentially missing not at random in survival analyses using Weibull proportional hazards regressions. We apply the method to an investigation of the impact of missing grade on post‐surgical mortality outcomes in individuals with metastatic kidney cancer. Data came from the Surveillance Epidemiology and End Results (SEER) registry which provides population‐based information on those undergoing cytoreductive nephrectomy. Tumor grade is an important component of risk stratification for patients with both localized and metastatic kidney cancer. Many individuals in SEER with metastatic kidney cancer are missing tumor grade information. We found that surgery was protective, but that the magnitude of the effect depended on assumptions about the relationship of grade with missingness. Copyright © 2009 John Wiley & Sons, Ltd.     

Stockholm Birth Cohort Study 1953-2003: a new tool for life-course studies

AIMS: To create a new tool for life-course studies of health outcomes as well as social outcomes. METHODS: Two anonymous data sets, one a local birth cohort and the other a nationwide registry, covering information from early and middle life, respectively, were matched using a "key for probability matching" based on a large number of variables, common to both data sets. The first data set provides social and health information from birth, childhood, and adolescence on boys and girls, born in Stockholm in 1953. The second data set provides information on income, work, and education as well as any inpatient visits and any mortality from mid-life for the entire Swedish population. RESULTS: For 96% of the original cohort it was possible to add data from mid-life. Thus, a new database has been created, referred to as the Stockholm Birth Cohort Study, which provides rich and unique life-course data from birth to age 50 for 14,294 individuals: 7,305 men and 6,989 women. Comparison of matched and unmatched cases in the original cohort suggests that those individuals that could not be matched had slightly more favourable social and intellectual circumstances and had often moved away from Sweden in the 1980s. CONCLUSION: The new database provides excellent opportunities for life-course studies on health and social outcomes. It allows for studies that have not previously been possible in Sweden or elsewhere. Further, it provides an opportunity for collaborative work with similar databases in Copenhagen and Aberdeen.     

Mortalité tardive après cancer des enfants survivants à cinq ans dans la région Rhône-Alpes

BackgroundThe population of survivors of childhood cancer is currently growing. Studies from other countries have shown an increased risk of late mortality. In order to measure this risk within a French cohort, the mortality of children who had survived five years from a cancer diagnosis were compared to the mortality of the general population, according to follow-up interval and cancer and treatment characteristics.MethodsThe study population consisted of 635 children diagnosed with cancer before the age of 15 who had survived at least five years, and were registered in the Rhone-Alpes region cancer registry from 1987 to 1992. Mortality was compared with general population rates of the Rhone-Alpes region to assess age and sex standardized mortality ratio (SMR) and absolute excess risk of death.ResultsThe median follow-up of children was 14.0 years. Among the 42 observed deaths, 71.4% were attributed to a recurrence of the original cancer, 9.5% to a second cancer. The 15-year cumulative risk of death, all causes, was 7.1%. The overall mortality of the cohort was 20.7 fold greater than the general population (95% CI: 14.9–27.9), and the absolute excess risk of 6.9 per 1000 persons-years. The long term excess-mortality was higher in case of recurrence of original cancer (SMR = 99.9, 95% CI: 67.9–141.9, absolute excess risk 35.4 per 1000 persons-years); it was raised during the five to nine years follow-up interval after diagnosis (SMR = 33.8, 95% CI: 23.2–47.3) mainly due to the primary malignancy, and decreased after (10–14 years follow-up interval SMR = 6.5, 95% IC 2.4–14.2).ConclusionThe late mortality of childhood cancer is significantly increased during the five to nine years following diagnosis and decreases after, but the cohort follow-up has to be extended in order to assess outcome beyond 15 years after diagnosis.     

Childhood cancers in the Netherlands (1989-1997)

In a recent report by the Society of Integral Cancer Centres in the Netherlands, attention was devoted to the incidence of cancer in children and the mortality arising from this. In recent years the growing diagnostic and therapeutic possibilities have changed the perspective of childhood cancer enormously. Based on a careful classification and clinical staging, national and international investigations have resulted in new and successful therapeutical strategies. Overall prognoses of childhood cancer have improved dramatically from a 5-year survival rate in the 1960s and 1970s of less than 30%, to an 8-year survival rate of more than 70% between 1989 and 1997. However, this success means that more investigations into the long-term effects of childhood cancer and its treatment are needed. In a study at the Amsterdam University Hospital, 700 adult survivors of childhood cancer were reinvestigated of whom over 75% experienced one or more clinically relevant long-term effects. A continuous survey for long-term effects is needed for the development of new therapeutic strategies, which allow children treated for cancer to develop with the same possibilities in life as their healthy peers.     

Employment and insurance outcomes and factors associated with employment among long-term thyroid cancer survivors: a population-based study from the PROFILES registry

Purpose

To obtain insight into employment and insurance outcomes of thyroid cancer survivors and to examine the association between not having employment and other factors including quality of life.

Methods

In this cross-sectional population-based study, long-term thyroid cancer survivors from the Netherlands participated. Clinical data were collected from the cancer registry. Information on employment, insurance, socio-demographic characteristics, long-term side effects, and quality of life was collected with questionnaires.

Results

Of the 223 cancer survivors (response rate 87 %), 71 % were employed. Of the cancer survivors who tried to obtain insurance, 6 % reported problems with obtaining health care insurance, 62 % with life insurance, and 16 % with a mortgage. In a multivariate logistic regression analysis, higher age (OR 1.07, CI 1.02–1.11), higher level of fatigue (OR 1.07, CI 1.01–1.14), and lower educational level (OR 3.22, CI 1.46–7.09) were associated with not having employment. Employment was associated with higher quality of life.

Conclusions

Many thyroid cancer survivors face problems when obtaining a life insurance, and older, fatigued, and lower educated thyroid cancer survivors may be at risk for not having employment.

     

The PanCareSurFup cohort of 83,333 five-year survivors of childhood cancer: a cohort from 12 European countries

Childhood cancer survivors face risks from a variety of late effects, including cardiac events, second cancers, and late mortality. The aim of the pan-European PanCare Childhood and Adolescent Cancer Survivor Care and Follow-Up Studies (PanCareSurFup) Consortium was to collect data on incidence and risk factors for these late effects among childhood cancer survivors in Europe. This paper describes the methodology of the data collection for the overall PanCareSurFup cohort and the outcome-related cohorts. In PanCareSurFup 13 data providers from 12 countries delivered data to the data centre in Mainz. Data providers used a single variable list that covered all three outcomes. After validity and plausibility checks data was provided to the outcome-specific working groups. In total, we collected data on 115,596 patients diagnosed with cancer from 1940 to 2011, of whom 83,333 had survived 5 years or more. Due to the eligibility criteria and other requirements different numbers of survivors were eligible for the analysis of each of the outcomes. Thus, 1014 patients with at least one cardiac event were identified from a cohort of 39,152 5-year survivors; for second cancers 3995 survivors developed at least one second cancer from a cohort of 71,494 individuals, and from the late mortality cohort of 79,441 who had survived at least 5 years, 9247 died subsequently. Through the close cooperation of many European countries and the establishment of one central data collection and harmonising centre, the project succeeded in generating the largest cohort of children with cancer to date.     

The Childhood Cancer Survivor Study: a resource for research of long-term outcomes among adult survivors of childhood cancer

Improvements in the treatment of the cancers occurring among children and adolescents have resulted in a large number of patients achieving long-term survival. Treatment-related factors have been shown to have an impact on subsequent health status and quality of life, although there are limited data on survivors who are now 2 or more decades past treatment. The Childhood Cancer Survivor Study (CCSS) was established as a resource for investigating the long-term outcomes of a cohort of 5-year survivors of pediatric and adolescent cancer, who were diagnosed between 1970 and 1986. The CCSS consists of more than 14,000 active participants, including survivors of leukemia, brain tumors, Hodgkin's disease, non-Hodgkin's lymphoma, Wilms' tumor, neuroblastoma, soft-tissue sarcoma, and bone tumors, who have provided self-reported sociodemographicand health-related information. The survivor population has been found to be at increased risk of a broad spectrum of adverse outcomes, such as late mortality, second cancers, pulmonary complications, pregnancy loss, giving birth to offspring with low birth weights, and decreased educational attainment. The ongoing evaluation of large and diverse cohorts of cancer survivors, through resources such as the CCSS, will aid in further identifying high-risk individuals who should be the target of innovative intervention strategies.     

Occupations with increased risk of pancreatic cancer in the Swedish population

Aims: To identify occupations with increased risk of pancreatic cancer in the Swedish population gainfully employed in 1970 over the period 1971–89.

Methods: The base population was made up of Swedish men (1 779 646) and Swedish women (1 101 669) gainfully employed at the time of the 1970 census and were still alive and over age 24 on 1 January 1971. Information was drawn from two data sets: the Swedish cancer environment register and a background population register. After 19 years of follow up, 4420 men and 2143 women were diagnosed with histologically confirmed incident pancreatic adenocarcinoma. Log linear Poisson models were fitted, allowing for geographical area and town size. Risk estimators were also calculated for workers reporting the same occupation in 1960 and 1970.

Results: Among women, a statistically significant risk excess of pancreatic cancer was observed for "educational methods advisors", "librarian, archivist, curator", "motor vehicle driver", "typographer, lithographer", "purser, steward, stewardess", "other housekeeping and related workers", and the groups of occupations of "electrical, electronic, and related" and "glass, pottery, and tile workers". Men showed a higher incidence of pancreatic cancer among "technical assistants", "travelling agents", "other metal processing workers", "baker and pastry cook", "docker and freight handler", and "waiters".

Conclusions: This study does not indicate that occupational factors play an important role in the aetiology of pancreatic cancer in Sweden. Few occupations were at increased risk of pancreatic cancer in both men and women, and the associations observed are in accordance with some previous studies from Western countries.

     

General practitioners and completeness of cancer registry.

STUDY OBJECTIVE--The aim of the study was to evaluate the role of the general practitioner as a source of information for a cancer registry. DESIGN--The study involved a random sample of respondents to a letter inviting participation which was sent to all general practitioners in a specific area. Participating doctors were visited to maximise cooperation. Data collection consisted of setting up a retrospective (prevalence) registry of cancer patients diagnosed over a 20 month period, and a prospective (incidence) register over a subsequent 12 month period. The general practitioner cancer file was then linked to the total cancer registry data base to estimate missed cases. SETTING--The study took place in the catchment area of the Comprehensive Cancer Centre Middle-Netherlands (IKMN). PARTICIPANTS--Of 448 general practitioners in the IKMN region, 152 were willing to participate and of these 52 were randomly selected to take part (11% of all general practitioners in the region). MEASUREMENTS AND MAIN RESULTS--A total of 1637 tumours were identified from the general practitioners, of which 252 (15.4%) were not included in the cancer registry. Of these, only 22 (1.3%) were not included in the registry because they had tumours diagnosed clinically in outpatients and therefore had been omitted from the usual sources of information available to the cancer registry (pathology reports, hospital discharge letters). The missed cases were mostly older patients with digestive tract tumours. CONCLUSIONS--On cost-benefit grounds it was not considered feasible to initiate an active cancer registration system among general practitioners, provided that notification of pathological examinations to the registry is complete. Limited under-registration will occur when death certificates cannot be used as an additional source of information.     

The social patterning of child mortality: the importance of social class, gender, family structure, immigrant status and population density

Which structural dimensions are most important for life chances have been the subject of debate. In this study, which uses data from the Swedish Census-linked Deaths registry for the period 1980–86, structural differences in child mortality are analysed. The structural dimensions focused on are social class, gender, family structure, immigrant status and population density. When all structural dimensions were controlled for simultaneously, only social class and gender influenced disease mortality. Of these two structural dimensions, social class seemed to be the most important in terms of producing the greatest differences. For mortality due to accidents, on the other hand, all dimensions proved to be important. The mortality risk was higher for children in single-parent households, for children whose parents were born outside Sweden, and for children living in the least densely populated areas. The greatest differences in mortality from accidents, however, were attributable to gender, followed by social class. Furthermore, mortality differences by the father's, the mother's and the family's class position, determined by the dominance method of class assignment, were compared. The class position of both parents was independently associated with child mortality but the family's position seemed to be the better discriminator in terms of class variation. The implications of these findings are that death risks during the whole childhood period are socially structured in Sweden and that social class is still a structural dimension of great importance.     

The lifetime cost of hepatocellular carcinoma

Background

Hepatocellular carcinoma (HCC) is the second most common cancer in Taiwan. For males in Taiwan, it is the most dangerous cancer, with both the highest incidence and mortality rate.

Objective

To determine cancer-related medical care costs for long-term survivors of HCC.

Methods

The estimation of the lifetime cost was based on the insurer perspective and adopted an incidence-based approach. Data was sourced from the 1999–2002 cancer registry statistics of patients with HCC and the claims data of Taipei Veterans General Hospital (TVGH). In total there were 2873 HCC patients at TVGH. In addition to this data, the research used population National Health Insurance claims data from the National Health Research Institutes (1996–2002) as the comparison group. The probabilities of survival, dying of cancer or dying of other causes were estimated using cancer registry statistics. To estimate lifetime (10-year) cost, we divided the disease process into three phases: initial, continuing and terminal. The cost of HCC was calculated as the sum of the average cost of each phase. The expected lifetime cost for treatment of an HCC patient was estimated by incorporating the phase-specific costs with the survival and mortality rates.

Results

The results showed that 895 patients survived <1 year, and treatment for each of these patients cost on average New Taiwan dollars ($NT) 206 573 ($US1 = $NT33, year 2002 value) over this period. For those who survived ≥1 year, the terminal phase of treatment resulted in the highest costs, $NT237 032. On average, for each patient, the initial phase cost was $NT140 403 and the monthly cost for the continuing phase was $NT8687. For the average HCC patient, the 10-year lifetime cost was $NT418 554 (in nominal $NT).

Conclusion

Our study showed that the terminal phase cost the most out of the three treatment phases. The aggregate lifetime cost of HCC is useful for health policy making and clinical decision making.     

Testicular cancer among Swedish pulp and paper workers

BACKGROUND: The incidence of testicular cancer has increased in recent decades. The aims of the present study were to elucidate whether Swedish paper and pulp mill workers had an increased incidence of testicular cancer, and to investigate whether certain occupational groups within the pulp and paper mill workforce were at increased risk. METHODS: The study was based on the Swedish Cancer Environment Register, which links the incidence of cancer for the period 1971-1990 with 1960 and 1970 National Census data on specific industries and occupations for all employed subjects in Sweden. RESULTS: Among maintenance workers employed both in 1960 and in 1970 in paper mills there was an increased risk for testicular cancer (standardized incidence ratio (SIR) 7.4, 95% confidence intervals (CI) 1.5-22), especially for seminomas (SIR 10.1, 95% CI 2.1-29). Maintenance workers were also at increased risk when analyzing workers employed in 1960, 1970, or both years. CONCLUSIONS: This study indicates an increased risk for testicular cancer, especially seminomas, among maintenance workers, but not among process workers, in Swedish paper mills.     

Lung cancer and mesothelioma in the pleura and peritoneum among Swedish insulation workers

OBJECTIVES: To estimate the risk of cancer and death in Swedish insulation workers some years after their exposure to asbestos had stopped. One hypothesis was that the risk of lung cancer would tend to decrease some years after the exposure had ended. METHODS: In a cohort study the cancer morbidity and cause of death was investigated in 248 insulation workers and compared with the corresponding morbidity and mortality in the general population. Due to stringent regulations, exposure to asbestos of all types had almost ended in Sweden in the mid- 1970s. Through a questionnaire, surviving insulation workers were asked about their exposure to asbestos and their smoking habits. RESULTS: Between 1970 and 1994 there were 86 deaths compared with the 46.0 expected (standardised incidence ratio (SIR) 1.9; 95% confidence interval (95% CI) 1.5 to 2.3), the increase was mainly due to an increased cancer mortality. The morbidity was increased for lung cancer (11 cases v 2.5 expected (SIR 4.4; 95% CI 2.2 to 7.9)), peritoneal mesothelioma (seven cases; no expected incidence could be calculated as the occurrence is too rare in the general population), cancer in pancreas (five cases v 0.7 expected (SIR 7.1; 95% CI 2.3 to 16.7)). No cases of pleural mesothelioma were found. The risk of lung cancer did not tend to approach that of the general population after the exposure to asbestos decreased. CONCLUSIONS: In the 1980s and the early 1990s, Swedish insulation workers still have a highly increased risk of diseases related to asbestos. The attributable risk for death and cancer was about 50%. The study also confirms the previous finding that mesothelioma in insulation workers seems to be situated in the peritoneum more often than in the pleura.     

Care for survivors of childhood cancer in the Netherlands

Survival of childhood cancer has improved resulting in an increasing number of survivors who are at high risk of developing treatment-related health problems. The authors emphasize the need for specialized care for survivors of childhood cancer by describing three patients who all developed late effects of treatment. The first patient, a 32-year-old female, who had several late effects caused by treatment for nephroblastoma; the second a 39-year-old female, who developed breast cancer after thoracic irradiation for metastatic nephroblastoma; the third a 45-year-old female diagnosed with a meningioma caused by cranial irradiation for acute lymphoblastic leukaemia. In the Netherlands medical care for survivors is clustered at special outpatient clinics (in Dutch: Langetermijneffecten na kinderkanker (Long-term effects after childhood cancer; LATER)-outpatient clinics). In 2010 a guideline was published with recommendations for optimal follow-up and care for survivors of childhood cancer.     

Completeness of the cancer registry

In 1985 the regional cancer registry in the IKMN area started. Pathology reports are used as the primary source of information for the registry. In order to evaluate the completeness of the registry, the registry database was compared with the discharge diagnosis register of the participating hospitals (LMR). In 1986 a total of 3546 tumours were registered by the registry, using the pathology reports as the primary source. The comparison with the LMR revealed that 616 patients were included in the cancer registry but not in the LMR. On the other hand 376 patients, in 180 of whom the diagnosis was confirmed by pathological examination while in 196 the diagnosis was based on clinical grounds only, were missing in the registry. Specifically, tumours of the central nervous system, liver, gallbladder, bile ducts and pancreas were missed in a high percentage. It is concluded that--even with an optimal pathology notification system--the LMR check is necessary to avoid underregistration. Omitting this check results in a 5% underregistration.     

The evolution of weak standards: the case of the Swedish rheumatology quality registry

Research in sociology suggests that the effects of standards are not nearly as straightforward or as homogenising as they first appear. The present study extends these insights by demonstrating how even standards designed simply to collect data can produce extensive and unanticipated effects in medical fields as their uses evolve across actors and contexts. We draw on an embedded case study exploring the multifaceted consequences of the use of a practice‐driven voluntary documentation standard: the Swedish rheumatology quality registry from 1995–2014. Data collection included document analysis; 100 interviews with specialists, patients and stakeholders in the field; fieldwork; and observations of physician‐patient encounters. Our findings show that the scope and influence of the registry increased over time, and that this standard and its evolution contributed to changes in rheumatologist clinical practice, research practice, and governmental practice. These findings suggest that even initially ‘weak’, voluntary forms of standardisation can generate far‐reaching and unpredictable consequences for the performance and delivery of care as well as for the development of a medical field. Future work about how standards can contribute both to uniformity and diversity is warranted.