Large variation in assessment and outcome definitions to describe the burden of long-term morbidity in childhood cancer survivors: A systematic review.

We systematically reviewed outcome assessment methods, outcome classification, and severity grading of reported outcomes in studies investigating the burden of physical long-term morbidity in childhood cancer survivors (CCS). A MEDLINE and EMBASE search identified 56 studies reporting on three or more types of health conditions in 5-year CCS, for which information was extracted on outcome types and classification, methods of outcome ascertainment, and severity grading. There was substantial variability in classification and types of health conditions reported and in methods of outcome ascertainment. Only 59% of the included studies applied severity grading, mainly the common terminology criteria of adverse events. This large variation in assessment and definition of the burden of physical long-term morbidity in CCS challenges interpretation, comparison, and pooling data across studies. Global collaboration is needed to standardize assessments and harmonize definitions of long-term physical morbidity and associated outcomes in childhood cancer survivorship research.

Risk of subsequent primary leukaemias among 69,460 five-year survivors of childhood cancer diagnosed from 1940 to 2008 in Europe: A cohort study within PanCareSurFup.

BACKGROUND: Survivors of childhood cancers are at risk of developing subsequent primary leukaemias (SPLs), but the long-term risks beyond 20 years of treatment are still unclear. We investigated the risk of SPLs in five-year childhood cancer survivors using a large-scale pan-European (PanCareSurFup) cohort and evaluated variations in the risk by cancer and demographic factors. METHODS: This largest-ever assembled cohort comprises 69,460 five-year childhood cancer survivors from 12 European countries. Standardised incidence ratios (SIRs) and absolute excess risks (AERs) were calculated. RESULTS: One hundred fifteen survivors developed an SPL including 86 myeloid leukaemias (subsequent primary myeloid leukaemias [SPMLs]), 17 lymphoid leukaemias and 12 other types of leukaemias; of these SPLs, 31 (27%) occurred beyond 20 years from the first childhood cancer diagnosis. Compared with the general population, childhood cancer survivors had a fourfold increased risk (SIR = 3.7, 95% confidence interval [CI]: 3.1 to 4.5) of developing leukaemia, and eight leukaemias per 100,000 person-years (AER = 7.5, 95% CI: 6.0 to 9.2) occurred in excess of that expected. The risks remained significantly elevated beyond 20 years from the first primary malignancy (SIR = 2.4, 95% CI: 1.6 to 3.4). Overall, the risk ratio for SPML (SIR = 5.8, 95% CI: 4.6 to 7.1) was higher than that for other SPLs. CONCLUSIONS: We demonstrate that beyond 20 years after childhood cancer diagnosis, survivors experience an increased risk for SPLs compared with that expected from the general population. Our findings highlight the need for awareness by survivors and their healthcare providers for potential risk related to SPL.

Risk and associated risk factors of hospitalization for specific health problems over time in childhood cancer survivors: a medical record linkage study.

Childhood cancer survivors (CCS) experience higher hospitalization rates compared to the general population for neoplasms, circulatory diseases, endocrine/nutritional/metabolic diseases and eye disorders. We studied trends in hospitalization rates and associated patient and treatment-specific risk factors for diagnosis subgroups among these four diseases. We performed medical record linkage of a ≥5-year CCS cohort with national registers, and obtained a random reference sample matched on age, gender and calendar year per CCS. For each diagnosis subgroup we compared hospitalization rates and trends over time in CCS and the reference population. Further, we analyzed risk factors for hospitalizations within the four CCS diagnosis groups. We used multivariate Poisson regression for all models. We retrieved hospitalization data from 1382 CCS and 26,583 reference persons. CCS had increased hospitalization rates for almost all diagnosis subgroups examined. Hospitalization rates for endocrine/nutritional/metabolic diseases appeared to increase with longer time since primary cancer diagnosis up to 30 years after primary cancer diagnosis. Survivors initially treated with radiotherapy had increased hospitalization rates for neoplasms (P < 0.001), those initially treated with anthracyclines (2.5 [1.1-5.5]) and radiotherapy to thorax and/or abdomen (9.3 [2.4-36.6]) had increased hospitalization rates for diseases of the circulatory system, and those initially treated with radiotherapy to head and/or neck had increased hospitalization rates for endocrine/nutritional/metabolic diseases (6.7 [3.5-12.7]) and diseases of the eye (3.6 [1.5-8.9]). Our study highlights that long-term health problems resulting in hospitalizations are still clinically relevant later in life of CCS. The identified treatment-related risk factors associated with hospitalizations support targeted follow-up care for these risk groups of CCS.

Social outcomes in adult survivors of childhood cancer compared to the general population: linkage of a cohort with population registers.

OBJECTIVE: Self-reported data show differences in social outcomes (not being married/having a registered partnership; not living independently; using social benefits) for childhood cancer survivors compared with their peers. We aimed to determine differences in these social outcomes between survivors and the general population using national register data and explored associated risk factors. METHODS: We performed medical record linkage between a single-centre cohort of 1768 ≥ 5-year survivors of childhood cancer (diagnosed 1966-2001) and two national registers (1999-2011) and obtained a random reference sample matched on gender and year of birth per survivor. We used multivariable logistic regression to calculate in adult survivors of childhood cancer (born before 1990) the odds of the specified social outcomes at the end of follow-up in both groups. Within the survivors, we analysed risk factors for the social outcomes. RESULTS: We retrieved data from 1283 adult childhood cancer survivors and 25 082 reference persons. Survivors had higher odds (odds ratio; 95% confidence interval) of not being married (1.2; 1.07-1.42), not living independently (1.7; 1.41-2.00) and using social benefits (2.3; 1.98-2.69) compared with reference persons. Radiotherapy to head and/or neck, and an original central nervous system tumour diagnosis negatively influenced all social outcomes examined in childhood cancer survivors. CONCLUSIONS: National register data show differences between social outcomes in childhood cancer survivors and the general population, especially for survivors treated with radiotherapy to head and/or neck and those originally diagnosed with central nervous system tumours. Development and implementation of effective targeted support strategies to improve social outcomes of childhood cancer survivors needs consideration. Copyright © 2015 John Wiley & Sons, Ltd.

A new method to facilitate valid and consistent grading cardiac events in childhood cancer survivors using medical records.

BACKGROUND: Cardiac events (CEs) are among the most serious late effects following childhood cancer treatment. To establish accurate risk estimates for the occurrence of CEs it is essential that they are graded in a valid and consistent manner, especially for international studies. We therefore developed a data-extraction form and a set of flowcharts to grade CEs and tested the validity and consistency of this approach in a series of patients. METHODS: The Common Terminology Criteria for Adverse Events version 3.0 and 4.0 were used to define the CEs. Forty patients were randomly selected from a cohort of 72 subjects with known CEs that had been graded by a physician for an earlier study. To establish whether the new method was valid for appropriate grading, a non-physician graded the CEs by using the new method. To evaluate consistency of the grading, the same charts were graded again by two other non-physicians, one with receiving brief introduction and one with receiving extensive training on the new method. We calculated weighted Kappa statistics to quantify inter-observer agreement. RESULTS: The inter-observer agreement was 0.92 (95% CI 0.80-1.00) for validity, and 0.88 (0.79-0.98) and 0.99 (0.96-1.00) for consistency with the outcome assessors who had the brief introduction and the extensive training, respectively. CONCLUSIONS: The newly developed standardized method to grade CEs using data from medical records has shown excellent validity and consistency. The study showed that the method can be correctly applied by researchers without a medical background, provided that they receive adequate training.