Late new morbidity in survivors of adolescent and young-adulthood brain tumors in Finland: a registry-based study

Abstract

Background

Brain tumors (BTs) in adolescence and young adulthood (AYA) differ from those in childhood or late adulthood. However, research concerning late effects in this particular survivor group is limited. This study evaluates late morbidity of survivors diagnosed in AYAs.

Methods

We identified from the Finnish Cancer Registry all survivors diagnosed with BT at the ages 16–24 years between 1970 and 2004 (N = 315) and used data from the Hospital Discharge Registry to evaluate their late (≥5 y after diagnosis) morbidity requiring treatment in a specialized health care setting. A sibling cohort of BT patients diagnosed before the age of 25 years was used as a comparison cohort (N = 3615).

Results

The AYA BT survivors had an increased risk for late-appearing endocrine diseases (HR, 2.9; 95% CI, 1.1–8.0), psychiatric disorders (HR, 2.0; 95% CI, 1.2–3.2), diseases of the nervous system (HR, 9; 95% CI, 6.6–14.0), disorders of vision/hearing loss (HR, 3.6; 95% CI, 1.5–8.5), diseases of the circulatory system (HR, 4.9; 95% CI, 2.9–8.1), and diseases of the kidney (HR, 5.9; 95% CI, 2.5–14.1). Survivors with irradiation had an increased risk for diseases of the nervous system compared with non-irradiated survivors (HR, 3.3; 95% CI, 1.8–6.2). The cumulative prevalence for most of the diagnoses remained significantly increased for survivors even 20 years after cancer diagnosis.

Conclusions

The AYA BT survivors have an increased risk of morbidity for multiple new outcomes for ≥5 years after their primary diagnosis. This emphasizes the need for structured late-effect follow-up for this patient group.

Brain tumors (BTs) in adolescence and young adulthood (AYA) differ from those seen in childhood and later adulthood in their incidence, distribution of different histologies, treatment, and prognosis.^1,2 In a large European study, the 5-year survival for BTs of young adults (aged 15–24 years) was 62%.² The Institute of Medicine has given recommendations for improving the health care for survivors of adult cancer and has highlighted the need for a follow-up plan for each cancer survivor.³ However, to organize structured and effective follow-up, information about late morbidity is particularly necessary for this age group. Studies on the late morbidity of BT survivors diagnosed in AYA have so far been scarce, and there has been concern about the lack of knowledge needed to design a plan for late follow-up.⁴

A study on late morbidity of survivors of young adult cancer was recently published,⁵ reporting that survivors of young adulthood CNS tumors have a 2-fold risk for late morbidity leading to hospitalization. However, specific reasons for hospitalization were not reported for BT survivors separately. The Childhood Cancer Survivor Study (CCSS) group has comprehensive data on cancer survivors diagnosed before the age of 21 years and separate data on BT survivors as well.^6,7 They have reported that childhood BT survivors are at an increased risk for late-appearing endocrinopathies, neurosensory deficits such as hearing impairment and blindness, focal neurologic dysfunction, seizures, and strokes. Unfortunately, data on the adolescent population are not analyzed separately. Reports on the survivors of adult BTs have also shown an increased risk for endocrinopathies, mental disorders, and neurological dysfunction.^5,8–10

The aim of our study was to evaluate late morbidity of survivors diagnosed with a BT at age 16–24 years in a nationwide, registry-based study and to assess the need for late follow-up of this particular group. The Hospital Discharge Registry (HDR), which collects data on diagnoses made in specialized health care settings, was used to collect data on morbidity. To our knowledge, this study is the first to concentrate specifically on late morbidity of survivors of adolescent and young adult BTs, and thus it provides important information to guide the follow-up planning for these survivors.

Patients and Methods

This study is part of a research project assessing the late effects of childhood, adolescent, and young adulthood cancer in Finland, the methodology of which is previously described.¹¹ In the present study, all patients in Finland diagnosed with a primary neuroepithelial BT after their 16th birthday and before age 25 years, between 1970 and 2004, were included. Data were retrieved from the Finnish Cancer Registry (FCR). As a control group, we used a sibling cohort of BT patients diagnosed before age 25 years in Finland at the aforementioned time period. The sibling cohort was identified via their common parents from the Population Register Centre. Registry linkages were done using the unique personal identification numbers assigned to all residents of Finland.

In Finland, all neoplasms are registered into FCR, which provides data for epidemiological research. It contains almost complete coverage for malignant CNS tumors as well as 80.6% coverage for benign CNS tumors.¹² Data on the survivors gathered from the FCR included date of birth, sex, date of diagnosis, tumor morphology, primary site, primary treatment modality, possible emigration date, date of death, and underlying cause of death. Treatment data are available on a basic level (surgical treatment [yes/no] and radiation treatment [yes/no]), but chemotherapy details or doses of radiation are not systematically available. The FCR uses the ICD-O-3 system for coding topography and morphology. However, coding of astrocytomas and gliomas was revised from the original designations since there had been changes in classifying tumors. After code conversion, the BT cases were grouped as proposed by the WHO classification of CNS tumors¹³ to astrocytic tumors, other glial tumors (oligodendroglial and oligoastrocytic tumors), ependymal tumors, embryonal tumors, other specified neuroepithelial tumors (neuronal, mixed neuronal-glial tumors, and tumors of pineal region), and BTs not otherwise specified .

Data on principal diagnoses of all inpatients in Finland since 1975 and outpatients in specialized health care settings since 1994 are collected by The National Institute for Health and Welfare (Hospital Discharge Registry [HDR]). Data are recorded in the International Classification of Diseases (ICD) format. Prior to analyses, diagnoses coded according to ICD-8 or ICD-9 were converted to ICD-10 format. The completeness and accuracy of the HDR data on different diagnoses vary and have been evaluated to range from satisfactory to very good.¹⁴ FCR data were linked with the HDR data for morbidity information.

The Ethical Committee of the Hospital District of Southwest Finland has approved the present study (register number 18/180/2010), and the National Institute for Health and Welfare has given permission to use the registry data (register number THL/531/5.05.00/2010).

Statistical Analysis

The 5-year BT survivor group was compared with the sibling cohort on 7 main outcomes (Table 1) and the most common separate diagnoses within main outcome groups. Diagnoses that are primarily cared for in a primary health care setting were excluded. The effect of irradiation treatment (survivors with a history of irradiation treatment vs survivors with no irradiation treatment) on the main outcomes was evaluated separately. A birth year cohort (1959 or earlier, 1960–1969, 1970–1979, 1980–1989, 1990 or later), sex, and relationship between sex and cancer status (survivor/sibling) were used as predictors for main outcomes when sex-specific HRs were compared with siblings. Contrasts were used to compare treatment eras (1970–1979, 1980–1989, 1990–2004).

Table 1.

Diagnoses included in the 7 main outcome groups analyzed (modified from Gunn et al¹¹, used with permission)

Main Outcomes Studied	Diagnoses Included in the Categories
Endocrine diseases	Hypothyroidism (E01, E03, E89.0), insulin-dependent diabetes mellitus(E10), disorders of pituitary gland (E22.0–E23.7, E89.3), disorders of adrenal gland (E26.0–E27.9), and other endocrine disorder caused by dysfunction in CNS (E28, E30, E34.3, E34.4, E89.4).
Mental and behavioral disorders	Mental and behavioral disorders due to psychoactive substance use (F10.0–F19.9), schizopherenia/schizotypal/delusional disorders (F20.0–F29), mood disorders (F30.0–F39), neurotic/stress-related/somatoform disorders (F40.0–F48.9), and organic mental disorders/personality or behavioral/emotional disorders/eating disorders or other unspecified disorder (F06.0–F09, F50, F60.0–F69, F90.0–F99).
Diseases of nervous system	Demyelinating diseases (G35.0–G37.9), epilepsy (G40.0–G41.9), migraine or other headache syndromes (G43.0–G44.8), sleep apnea (G47.3), nerve, nerve root and plexus disorders (G50–G59.8), cerebral palsy and other paralytic syndromes (G80.0–G83.9), other disorders of the brain or nervous system (excluding hydrocephalus) (G93–G94, G97–G99) and intracranial injury (S06). Systemic atrophies affecting the CNS and extrapyramidal and movement disorders were not included because of very small number of cases.
Disorders of vision or hearing loss	Disorders of lens (H25.0–H28.8), disorders of retina (H31, H33.0–H36.8), disorders of optic nerve and visual pathways (H46.0–H48.8), strabismus (H49.0–H50.9), visual disturbances and blindness (H53.0–H54.9), and hearing loss (H90.0–H91.9).
Diseases of circulatory system	Ischemic heart diseases (I20.0–I25.9), cardiac arrhythmias (I47.0–I49.9), cardiomyopathy/heart failure (I42, I50), intracranial nontraumatic hemorrhage (I60.0–I62.9), and transient cerebral ischemic attacks/cerebral infarction/occlusion/stenosis of precerebral arteries and other cerebrovascular diseases including sequelae of cerebrovascular diseases(G45, I63, I65–I67, I69).
Diseases of musculoskeletal system and connective tissue	Arthrosis (M15.0–M19.9), deforming dorsopathies (M40.0–M43.9), spondylopathies and intervertebral disc disorders (M45.0–M51.9), disorders of bone density and structure (M80.0–M85.9), and other osteopathies and chondropathies (M86.0–94.9).
Diseases of the kidney	Glomerular diseases and renal tubule-interstitial diseases (excluding obstructive and reflux uropathy) (N00.0–N08.8, N10.0–N12, N14.0–N16.8) and renal failure (N17.0–N19).

Cox regression was used to calculate hazard ratios (HRs). Only diagnoses made after 5-year survival were analyzed for survivors. Follow-up began at the earliest on January 1, 1975, and at the age of 21 years. Only the first hospitalization for each main outcome was registered. Follow-up was completed on December 31, 2008, or earlier in case of death/emigration. The analyses were adjusted for sex and birth year cohorts (1959 or earlier, 1960–1969, 1970–1979, 1980–1989, 1990 or later).

The age-specific cumulative prevalence was analyzed at 5, 10, 20, and 30 years from cancer diagnosis as a percentage of survivors who had been diagnosed by the end of each period of the survivors who were alive. The mean age of survivors at the aforementioned time points was used as the evaluation age of the comparison cohort (eg, 25.6 years at 5 years from diagnosis, etc.). The Fisher exact test was used for comparisons.

A P value <.05 was considered as significant. SAS, version 9.3, was used for statistical analyses.

Results

Characteristics of the 5-year survivors are described in Table 2. There were 506 patients diagnosed with a neuroepithelial BT between the ages of 16 and 24 years, 315 of whom survived at least 5 years. The median age at the time of cancer diagnosis was 20.5 years, and the median age at the end of follow-up was 35.7 years (range, 21.7–62.9 y). The median survivor follow-up after 5 years was 11.5 years (range, 0.1–33.9 y). Surgery was performed in 87.3% of the survivors, and 40.0% had received irradiation. The comparison cohort included 3615 study subjects.

Table 2.

Characteristics of the 5-year survivors (N = 315)

	N	%
Sex
Male	171	54.3
Female	144	45.7
Year of diagnosis
1970–1979	65	20.6
1980–1989	99	31.4
1990–1999	109	34.6
2000–2004	42	13.3
Tumor classification
Ependymal tumors	17	5.4
Astrocytic tumors	175	55.6
Other gliomas	48	15.2
Embryonal tumor	11	3.5
Other specified	64	20.3
Treatment
Irradiation	126	40.0
Surgery	275	87.3
Surgery, no irradiation	163	51.8
Irradiation, no surgery	14	4.4
Surgery + irradiation	112	35.6
Not known	16	5.1

There was a significantly increased HR among the survivors compared with the sibling cohort for endocrine diseases, psychiatric disorders, diseases of the nervous system, disorders of vision/hearing loss, diseases of the circulatory system, and diseases of the kidney (Table 3). The HR for disorders of the musculoskeletal system/connective tissue, in general, was not significantly increased. However, a significant increase was found when disorders of bone were analyzed separately (HR, 2.0; 95% CI, 1.1–3.7). Psychiatric disorders of the survivors included 5 cases of psychoactive substance use (HR, 0.8; 95% CI, 0.3–1.9 ), 2 cases of schizophrenia/delusional disorders (HR, 1.1; 95% CI, 0.3–4.7), 9 cases of mood disorders (HR, 1.7; 95% CI, 0.8–3.4), and 4 cases of neurotic/somatoform disorders (HR, 1.3; 95% CI, 0.5–3.8). The most common diagnoses of the nervous system were epilepsy (n = 40; HR, 37.1; 95% CI, 20.7–66.6), mononeuropathy/nerve root compression (n = 9), other disorders of the brain (n = 6), hemiplegia (n = 4), sleep apnea (n = 2), and intracranial injury (n = 3). Cases of multiple sclerosis, other demyelinating disease,or other headache syndrome were single. The increased HR for diseases of the circulatory system was mostly explained by the increase in cerebrovascular diseases (HR, 9.3; 95% CI, 4.9–17.7). Among the survivors, there was one case of ischemic heart disease, 4 cases of arrhythmias, 2 cases of cardiomyopathies, 9 cases of intracranial nontraumatic hemorrhage, and 21 cases of cerebrovascular diseases reported. The survivors who had received irradiation treatment compared with the survivors with no irradiation showed a significantly increased HR only for diseases of the nervous system (Table 3). The HRs for diseases of the nervous system among the irradiated and non-irradiated survivors compared with the sibling cohort separately were 13.7 (95% CI, 9.0–20.8) and 4.6 (95% CI, 2.7–7.9), respectively. Before the 5-year survival, we did not find any significant associations between irradiation treatment and any of the main outcomes (data not shown).

Table 3.

Hazard ratios for different new diagnoses in 5-year survivors of brain tumors compared with the sibling cohort and the impact of irradiation treatment

Outcome	All Survivors Compared with Sibling Cohort			Irradiated Survivors Compared with Non-irradiated Survivors		Non-irradiated Survivors Compared with Sibling Cohort		Irradiated Survivors Compared with Sibling Cohort
	N	HR	(95% CI)	HR	(95% CI)	HR	(95% CI)	HR	(95% CI)
Endocrine diseases	6	2.9	(1.1–8.0)	0.8	(0.1–4.9)	3.6	(1.0, 12.3)	2.5	(0.6, 11.1)
Psychiatric disorders	20	2.0	(1.2–3.2)	1.4	(0.5–3.8)	1.6	(0.9, 3.1)	2.4	(1.2, 4.8)
Diseases of nervous system	47	9.6	(6.6–14.0)	3.3	(1.8–6.2)	4.6	(2.7, 7.9)	13.7	(9.0, 20.8)
Disorders of vision or hearing loss	7	3.6	(1.5–8.5)	0.7	(0.1–3.8)	4.6	(1.7, 12.3)	2.8	(0.6, 11.9)
Diseases of circulatory system	22	4.9	(2.9–8.1)	1.8	(0.8–4.3)	3.4	(1.7, 7.0)	7.1	(3.8, 13.4)
Disorders of musculoskeletal system or connective tissue	19	1.4	(0.9–2.3)	1.3	(0.5–3.4)	1.2	(0.6, 2.4)	1.8	(0.9, 3.5)
Diseases of the kidney	8	5.9	(2.5–14.1)	0.8	(0.2–3.3)	7.4	(2.7, 20.3)	5.2	(1.5, 18.4)

Analyses are adjusted with sex and birth year cohort.

The female survivors had a slightly more increased hazard (P = .049; HR, 2.7; 95% CI, 1.0–7.3) for disorders of the musculoskeletal system/connective tissue compared with their female siblings than the hazard for male survivors compared with their male siblings, when the interaction between sex and a cancer status (survivor/sibling) was analyzed. For other main outcomes, the hazard caused by cancer and its treatment was similar with both sexes. There was no significant difference between treatment eras in any main outcome studied.

Table 4 shows the age-specific cumulative prevalence of main outcomes for all BT survivors at time points 5, 10, 20, and 30 years after BT diagnosis. The prevalence of endocrine diseases and psychiatric disorders was higher among the survivors than the sibling cohort at each time point, but a significant difference was seen in endocrine diseases at 20 years and psychiatric disorders at 10 and 20 years. The survivors had significantly increased prevalence compared with the sibling cohort for diseases of the nervous system, disorders of vision/hearing loss, and diseases of the circulatory system at each time point evaluated. No differences in prevalence were found for disorders of the musculoskeletal system/connective tissue between the survivors and the sibling cohort. The prevalence for diseases of the kidney was similar among the survivors and the siblings during the first decade after the BT diagnosis, but this prevalence increased significantly after that time point. To further evaluate the reason for this, we investigated the frequency of survivors with diabetes registered in the HDR at 5 and 30 years from BT diagnosis (0% and 3%, respectively).

Table 4.

The age-specific cumulative prevalence of main outcomes for all brain tumor survivors at points of time 5, 10, 20, and 30 years after cancer diagnosis compared with the outcome of the sibling cohort

Outcome	Age-specific Cumulative Prevalence (%)
	5 Years			10 Years			20 Years			30 Years
	Survivors N = 315	Siblings N = 2207	P	Survivors N = 219	Siblings N = 1738	P	Survivors N = 111	Siblings N = 959	P	Survivors N = 31	Siblings N = 356	P
Endocrine diseases	1.9	0.8	.109	1.8	1.0	.283	4.5	1.4	.031	6.5	1.7	.126
Psychiatric disorders	6.7	4.1	.054	9.1	5.4	.032	12.6	6.6	.030	9.7	5.9	.426
Diseases of nervous system	25.4	1.7	<.001	28.8	2.5	<.001	31.5	3.2	<.001	25.8	3.9	<.001
Disorders of vision or hearing loss	5.4	1.8	<.001	8.2	1.8	<.001	6.3	1.5	.004	9.7	1.1	.013
Diseases of circulatory system	2.2	0.5	.003	3.7	0.7	<.001	5.4	1.4	.009	16.1	3.7	.010
Disorders of musculoskeletal system or connective tissue	5.4	4.2	.375	5.9	4.4	.301	8.1	5.8	.398	9.7	8.4	.739
Diseases of the kidney	0.0	0.5	.382	0.0	0.5	.609	2.7	0.5	.041	9.7	1.1	.013

The mean age of the survivors at each point of time is used as the age in which the sibling cohort is evaluated.

Discussion

AYA cancer survivors are a group much less studied than the survivors diagnosed in their childhood. Research concerning BT survivors in particular is still limited. There are differences in BTs and their late effects in this age group compared with the survivors diagnosed in childhood or later in adulthood.^1,2,4 People in this age group are also trying to build an independent life of their own, including family and career planning. Therefore, it is essential that their cancer and its late effects are treated carefully to ensure them as long and satisfying a life as possible.

Our results show that survivors of AYA BTs are clearly at an increased risk for multiple new outcomes 5 years after their primary cancer diagnosis. In particular, the risk for outcomes of the nervous system, cerebrovascular disease, and kidney-related morbidity need to be followed carefully even years after completion of cancer treatment. Endocrinopathies and disorders of vision/hearing loss might also need more structured screening in this particular age group. Irradiation has an independent effect on the risk for diseases of the nervous system but seems not to explain all of the excess risks for the survivors. The somewhat low number of irradiated survivors in our cohort may have had some effect on the results. On the other hand, the problems experienced by irradiated patients may have already appeared before the 5-year survival. The prevalence of diseases of the circulatory system and diseases of the kidney increased during the entire follow-up time, suggesting that some late effects may become significant even decades after the end of treatment (although normal aging also increases the risk). Changes in treatment protocols may have also had an impact on the prevalences. The techniques of irradiation ,as well as the total doses, have been modified in the recent decades. On the other hand, even though the role of chemotherapy may have also increased in the modern treatment of AYA BTs, there may still be several late-appearing late effects that are not yet recognized as such (eg, endothelial effects from chemotherapy).

With regard to endocrinological morbidity, our results showed a subtle increase in HR among the survivor cohort that was not impacted by irradiation. Another study assessing young adult cancer survivors showed no increase in hospitalization rates due to endocrinological causes.⁵ Childhood BT survivors are known to be at a highly elevated risk for endocrinological late effects⁶ and for hospital contacts due to endocrinological reasons.¹⁵ Multimodal treatment with surgery, irradiation, and chemotherapy together increases the risk the most.⁶ Also, adults with nonpituitary BTs, treated with cranial irradiation, are at increased risk for hypopituitarism; one study showed that 41% of survivors had at least one pituitary hormone deficiency.¹⁰ Nonetheless, endocrinological late effects in adulthood do not present as dramatically as those seen in childhood, and some are probably treated in the general health care setting. Our study concentrated on new diagnoses emerging after 5-year survival, while most endocrinological morbidity is probably diagnosed earlier. However, considering the low figures seen in our study, some survivors might be lost to follow-up or probable underdiagnosis (eg, adult growth hormone deficiency). Even if the risk for endocrinopathies among the AYA cancer survivors is not that striking, these survivors could still benefit from appropriate replacement therapies.^10,16

Cancer survivors in general face multiple physical and psychosocial issues that cause stress. Almost one-third of adults with a BT met the criteria for major depressive disorder in a study in which two-thirds of the patients were still receiving chemotherapy.⁸ A study assessing childhood BT survivors found that they had an increased risk for hospital contacts for mental disorders, with those youngest at the time of cancer diagnosis being at the greatest risk.¹⁷ A study analyzing young adult cancer survivors separately found no increase in psychiatric hospitalization,⁵ and another study concentrating on adults found BT survivors to be at a risk for hospitalization due to depression (but only during the first year after a BT diagnosis).⁹ The general young adult population in Finland has a lifetime prevalence of 18% for depressive disorders and 14% for substance abuse.¹⁸ The incidence of schizophrenia has varied between 30 and 56 per 100 000.¹⁹ However, since psychiatric morbidity does not necessarily lead to treatment contact or hospitalization, the figures cannot be compared directly.^18,19 Our results showed a slightly increased risk for late-appearing psychiatric disorders in BT survivors. The cumulative prevalence among the survivors was significantly increased at 10 and 20 years after treatment, emphasizing that this concern should also be kept in mind at follow-ups.

Patients with BTs are inevitably at risk for neurological consequences because of the tumor itself and its treatment, especially irradiation.⁷ The risk, compared with that of a general population, decreases markedly over time after cancer diagnosis and treatment; however, CCSS has shown childhood BT survivors to have a greatly increased risk for new coordination problems, motor problems, and seizure disorders more than 5 years after cancer diagnosis when compared with a sibling cohort.⁷ The vulnerability of the brain is greatest during childhood, but adolescents and adults also suffer from CNS damage that is visible with MRI.^20,21 A study concentrating on young adulthood cancer survivors showed only a 1.5-time risk for hospitalization due to diseases of the nervous system compared with a group randomly selected from a health insurance plan.⁵ Our study concentrating on the AYA BT survivors showed a clearly higher risk. Irradiated survivors also showed a significantly greater risk compared with the survivors who had not undergone irradiation treatment, as expected based on other studies.^7,21

Survivors with a history of CNS tumor at childhood and adolescence are at a great risk for auditory complications.²² Irradiation to the posterior fossa or the temporal lobe and chemotherapy with platinum compounds are known risk factors for ototoxicity.^7,22,23 The risk for ototoxicity diminishes greatly with increasing age at the time of a treatment.^22,23 Earlier studies have shown that cancer survivors, both children and adults, suffer from disorders of vision (eg, cataracts, blindness, and double vision).^7,24,25 Although the risk for disorders of vision/hearing loss shown in our study was increased compared with the sibling cohort, it is probable that the small total number of cases explains why irradiation showed no effect. Most of the cases are probably diagnosed less than 5 years after cancer diagnosis, and all subsequent new diagnoses might not end up being treated in a specialized health care setting.

The present study showed a significantly increased HR for diseases of the circulatory system that were, for the most part, explained with cerebrovascular morbidity. The cumulative prevalence for circulatory diseases remained significantly increased among the survivors at each time point assessed, and ageing of the survivors and the sibling cohorts did not diminish the difference. Supporting our results, the CCSS has found that childhood BT survivors have a relative risk of 29.0 for late-occurring strokes when compared with a sibling cohort. Young adult cancer survivors, for their part, have an approximately 2-fold risk for hospitalizations due to all circulatory diseases.⁵ There were only a few cases of cardiac morbidity found in the present study. Although the overall cardiac morbidity among childhood, adolescent, and young-adult cancer survivors has been increased in other studies, it must also be taken into consideration that the BT survivors in those studies had only a slightly increased risk when assessed separately.^26,27 Whether the persistent risk for circulatory diseases found in the present study is due to direct effect of cranial irradiation or caused by increased frequency of hypertension, obesity, and diabetes over time remains unclear.

Although, the number of kidney-related diseases among the survivors shown in our study was low, the HR compared with the sibling cohort was greatly increased. However, cumulative prevalence differed only after 20 years from tumor diagnosis of survivors. Decline in glomerular function among childhood cancer survivors with nephrotoxic therapy has been shown to be persistent and deepen over time.²⁸ Long-term sequelae observed in our study seem to appear late, when the survivors' own kidney function is also declining with normal ageing. Increased rates of diabetes did not seem to be the only explanation for the increase in kidney diseases. Unfortunately, we have no accurate data available on chemotherapy treatments, and, thus we are not able to evaluate which effect the change in treatment protocols has had on the results.

Our study has some limitations. Since it is based on national health care registries, it has no risk of bias from self-reporting. However, the registries in question have no information on survivors' chemotherapy or doses of received irradiation. The HDR includes data only on visits in a specialized health care setting. This is particularly problematic when evaluating adults whose follow-up might have been shifted to general health care. However, because the data from the sibling cohort were collected similarly, the HRs should be accurate. The cumulative prevalence figures may also underestimate the morbidity somewhat as the data from the HDR start in 1975, and some of the earliest diagnoses may be missing. Again, however, it should not have a major impact on the comparison, because the sibling cohort shares this concern. The sibling cohort was used as a comparison group to diminish the effect of genetic and environmental variation for certain diagnoses.

A Canadian study has shown that childhood cancer survivors diagnosed at the age of 15–19 years are less likely to contact a specialist about their health problems and more likely to contact a general practitioner than the survivors diagnosed at a younger age.²⁹ Primary health care may, however, in many cases lack the experience in dealing with various late effects in this survivor group.

The present study evaluated the late new morbidity needing treatment in a specialized health care setting among the insufficiently studied group of AYA BT survivors. The results highlight the need for systematic late follow-up for survivors of adolescent and young adult cancers.

Funding

Foundation of Nona and Kullervo Väre (to M.E.G.); Children′s Cancer Foundation Väre (to M.E.G.); Children′s Cancer Trust in Memoriam of Emilia Heinonen (to P.M.L); The National Graduate School of Clinical Investigation (to M.E.G.); The Foundation of Wilho Kyttä (to M.E.G.); and the Foundation for Pediatric Research (to M.E.G.). Study sponsors had no involvement in the study.