Heart Disease in Survivors of Childhood Cancer

Summary

Background:

Cardiac and cardiovascular diseases are major causes of morbidity in the growing population of childhood cancer survivors, particularly as a result of treatment with anthracyclines and radiotherapy to the chest.

Methods:

The VersKiK study is based on record linkage between the German Childhood Cancer Registry and 13 statutory health insurance carriers (69% coverage of the German statutorily insured population). Registry data with information on primary cancers in childhood or adolescence were linked to data from the participating health insurance carriers. Cardiac diseases were analyzed in 11 863 patients with cancer diagnosed between 1991 and 2011 who survived at least 5 years after their diagnosis (up to January 2017). The health insurance data covered the years 2017–2021. The childhood cancer survivors were matched to a comparison group of 35 589 insurees without childhood cancer by age, sex, region of residence within Germany, and duration of insurance.

Results:

Cardiac diseases were identified in 6.4% of the 5-year survivors of childhood cancer and in 2.5% of the matched comparison group. Cardiomyopathy, valvular disease, and arrhythmia were the most common heart diseases in the survivors (1.9%, 2.7%, 1.5%). The largest difference from the comparison group was observed for cardiomyopathy (prevalence ratio = 14.0; 95% confidence interval: [10.9; 20.5]).

Conclusion:

Long-term survivors of childhood cancer in Germany are more likely to be diagnosed with cardiac disease than the general population of the same age. Systematic medical monitoring and individualized preventive measures for survivors are indicated. In the interpretation of these findings, it should be noted that the childhood cancer survivors underwent more intensive long-term medical followup than their counterparts in the comparison group.

The oncological treatment of children with cancer has improved considerably over recent decades, resulting in a 5-year survival rate above 80% in developed countries (1, 2). The long-term health effects of childhood cancer are therefore of major concern. Cardiovascular late effects lead to increased morbidity and early mortality among childhood cancer survivors (CCS) (3–12).

Cardiac disease can manifest as cardiomyopathy, valvular disease, arrhythmias, and pericardial disease, but also as ischemic coronary heart disease and myocardial infarction (13, 14). While cardiomyopathy, ranging from subclinical dysfunction to overt congestive heart failure, is the most commonly described cardiovascular sequela in CCS, ischemic heart disease is also relevant, particularly among older survivors. Furthermore, risk factors such as hypertension, dyslipidemia, and diabetes mellitus appear earlier in CCS than in their siblings or the general population (15, 16).

An elevated risk of cardiovascular health events has been reported among survivors of several childhood cancer types, in particular kidney cancer, bone tumors, lymphoma, and leukemia, owing to the cardiotoxicity of the oncological therapy (3, 17). The risk of cardiac disease is especially high in CCS exposed to high-dose anthracyclines (> 250 mg/m²) and/or chest-directed radiotherapy > 15 Gy (13, 18, 19, 20). Other possible risk factors include female sex, younger age at the time of treatment, and the presence of cardiovascular risk factors (15).

With improvements in contemporary cancer treatment, the number of long-term survivors of childhood cancer will increase. Currently, around 44 000 5-year survivors are being followed by the German Childhood Cancer Registry (GCCR) (21). The VersKiK study (Versorgung, Versorgungsbedarf und VersorgungsbedUrfnisse von Personen nach einer Krebserkrankung im Kindes- oder Jugendalter; Follow-up care and adherence to follow-up guidelines after cancer in childhood or adolescence) is based on a linkage between GCCR data and claims data from German statutory health insurance carriers (SHI). Here, we present frequencies of cardiac late effects among CCS in Germany and in a matched comparison group of insured persons without a history of childhood cancer.

Methods

This report was formulated in accordance with the STROBE guidelines (22).

Study design

VersKiK is a retrospective registry-based cohort study. It combines GCCR data on patients with childhood or adolescent cancer (diagnosed in the period 1991–2021) with claims data from 13 major SHI. The SHI data cover the years 2017–2021. A comparison population (ratio 1:3) was drawn from the data of the participating SHI and comprised persons matched to the GCCR cases by year of birth, sex, region of residence, and duration of insurance (23). Details of the goals and methodological implementation of the VersKiK study have been published elsewhere (23, 24, 25). An overview can be found in eBox 1 and eBox 2.

eBox 1. Aims of the VersKiK study.

• Quantification of the spectrum of late effects following childhood cancer: exploration of the rates of chronic disease (with a focus on cardiological, audiological, and psychological disorders) and of potential determinants (age at primary cancer diagnosis, sex, cancer type)

• Medical care and utilization of medical services: investigation of the type and frequency of diagnostic, therapeutic (inpatient and outpatient), and rehabilitative service and of potential determinants

• Adherence to guideline recommendations: evaluation of the recommended follow-up care for cardiological and audiological late effects (in association with anthracycline and/or platinum-containing cytostatics) after lymphatic leukemia, non-Hodgkin lymphoma, neuroblastoma or germ cell tumor in childhood; investigation of determinants of adherence

eBox 2. VersKiK study: inclusion criteria and time periods.

• Inclusion criteria for the GCCR base cohort

  • –
    Cancer diagnosis in childhood ^*1: 1 January 1991 to 31 December 2021
  • –
    Age < 15 years at primary diagnosis^*2
  • –
    Age < 18 years at primary diagnosis since 2009
  • –
    Survival and residence in Germany until at least 1 January 2017

• Assessment of outcomes in SHI claims data

  SHI data 1 January 2017 to 31 December 2021

  • –
    SHI-insured persons in the GCCR base cohort, identified by record linkage: combination of claims data from the 13 participating SHI with GCCR data on cancer
  • –
    Selection of a comparison population from the records of the participating SHI: persons without childhood cancer; persons born on 1 January 1976 or later, resident in Germany, who were alive at the time of data extraction; matching 1 : 3 by age, sex, region of residence, duration of insurance

Study population, available data, and analysis population

For the purposes of the VersKiK study, the GCCR defined a base cohort of registered patients with childhood cancer, as defined by the International Classification of Childhood Cancer (ICCC-3), diagnosed in the period 1991–2021. According to GCCR criteria, these persons were under 15 years of age at diagnosis (up to 2008) or under 18 years of age (from 2009 onward). Only those persons were eligible who were documented as alive on 1 January 2017 and were resident in Germany on 1 January 2017. A total of 49 624 patients met these criteria.

Stochastic record linkage was performed between the encrypted identity data of the GCCR base population and identically encrypted identity data of persons (birth year ≥ 1976) insured by the participating health insurance carriers. Overall, 26 127 former patients were identified in the SHI data (VersKiK/GCCR population) (24).

The data of the 26 127 former patients were linked with SHI data (diagnoses, hospitalizations, medication prescriptions) and with GCCR data on primary cancer diagnosis, year of diagnosis, age at diagnosis, sex, year of birth, vital status, recurrences, second malignancies, and region of residence.

For the analysis of cardiac late effects we formed a subpopulation of 5-year survivors. These were persons with cancer diagnosed between 1991 and 2011 who survived for at least 5 years after diagnosis until January 2017. A further eligibility criterion was uninterrupted insurance from January 2017 to December 2021 or until date of death (SHI observation period up to December 2021 or death, whichever occurred first). This group comprised 11 863 5-year survivors (eFigure). This analysis population was matched 1:3 with 35 589 SHI-insured persons for comparison.

eFigure. Selection of patients and formation of the analysis cohort:

GCCR, German Childhood Cancer Registry;

SHI, the 13 participating statutory health insurance carriers (modified from Aleshchenko et al. 2025 (25))

Definition of cardiac events

The events of interest according to the International Classification of Diseases (ICD-10) (26) documented in the SHI data between 01/2017 and 12/2021 were as follows:

• Congestive heart failure/cardiomyopathy (I42)

• Valvular heart disease (I34–I39)

• Cardiac arrhythmia (I47–I49)

• Pericardial disease (I31)

• Coronary heart disease/ischemic heart disease (I20–I25)

• Cardiac insufficiency (I50)

The presence of “any cardiac disease” was defined as observation of at least one of these specific cardiac diagnoses was recorded in at least two quarters during the SHI insurance period. The presence of “specific cardiac events” was defined accordingly.

Data analysis

To establish the absolute and relative frequencies of cardiac events, the prevalences in the cohort of 5-year survivors and in the the comparison group were calculated and stratified by sex and attained age (at 1 January 2017). The frequencies of cardiac late effects among 5-year survivors were also stratified according to primary cancer diagnosis and the duration of follow-up, i.e., the time between diagnosis and the start of the SHI insurance period on 1 January 2017.

To compare the frequencies of cardiac events in 5-year survivors and the comparison group, the ratios of the relative frequencies (prevalence ratio [PR] = RF_patients/ RF_{comparison group}) and their 95% confidence intervals were calculated using log-linear conditional regression models. In order to calculate PRs according to cancer diagnosis and follow-up period, comparison group members were selected according to the defined matching criteria.

Results

The VersKiK cohort of 5-year survivors included 11 863 patients with childhood cancer (diagnosed during the period 1991–2011) who survived until 1 January 2017. Survivors were more likely to be male than female. The most commonly occurring form of cancer was leukemia. At the beginning of the insurance period (January 2017) the median age was 21.5 years with a median follow-up time since cancer diagnosis of 14.5 years (interquartile range 9.5–19.5) (Table 1). A total of 11 681 patients (98.5%) were alive at the end of the SHI insurance period (December 2021).

Table 1. Description of the VersKiK cohort: 5-year survivors of childhood cancer in Germany (cancer diagnosis 1991–2011).

	Five-year survivors 1
Sex (N,%)
Female	5236 (44.1%)
Male	6627 (55.9%)
Primary cancer diagnosis (N, %)
Lymphoid leukemia	3459 (29.5%)
Acute myeloid leukemia	478 (4.0%)
Hodgkin lymphoma	669 (5.6%)
Non-Hodgkin lymphoma	825 (7.0%)
CNS tumor	2413 (20.3%)
Neuroblastoma	789 (6.7%)
Renal tumor	758 (6.4%)
Bone tumor	466 (4.0%)
Soft tissue sarcoma	618 (5.2%)
Intracranial germ cell tumor	120 (1.0%)
Other diagnoses	1268 (10.7%)
Vital status at 31 December 2021
Alive	11 681 (98.5%)
Deceased 2017–2021	182 (1.5%)
Median age at 1 January 2017* (years)	21.5 (IQR 15.5–26.5)
Median time from cancer diagnosis to at least January 2017 (years)	14.5 (IQR 9.5–19.5)
Total	11 863 (100%)

^*Start date of the insurance period: monitoring of health events (1 January 2017 to 31 December 2021)

CNS, Central nervous system; IQR, interquartile range

Cardiac events

During the SHI insurance period (2017–2021), cardiac events were identified in 6.4% of the group of 5-year survivors and in 2.5% of the comparison group. The PR between the two groups was 2.5 [2.3; 2.8]. The intergroup difference was highest in the age group 5–10 years (PR = 7.3 [4.6; 11.6]). The PRs for women and men were comparable (PR = 2.6 [2.3; 3.0] and PR = 2.5 [2.2; 2.9]) (Table 2).

Table 2. Frequency and prevalence ratios of cardiac events in the period 2017–2021 among 5-year survivors of childhood cancer and a matched^*1 comparison group in Germany.

	Frequency of cardiac disease*2, *3 N (%)
	Five-year survivors (N = 11 863)	Comparison group*1 (N = 35 589)	Prevalence ratio [95%CI]
Sex
Female	383 (7.3)	445 (2.8)	2.6 [2.3–3.0]
Male	374 (5.6)	449 (2.3)	2.5 [2.2–2.9]
Age at 1 January 2017 (years)
5 to < 10	61 (7.7)	25 (1.1)	7.3 [4.6–11.6]
10 to < 20	289 (6.5)	282 (2.1)	3.1 [2.7–3.7]
20 to < 30	294 (5.7)	392 (2.6)	2.2 [1.9–2.6]
≥ 30	113 (7.7)	195 (4.4)	1.8 [1.4–2.2]
Total	757 (6.4)	894 (2.5)	2.5 [2.3–2.8]

^*1Matching criteria: age, sex, region of residence, duration of insurance

^*2Defined according to the International Classification of Diseases (ICD-10): congestive heart failure/cardiomyopathy (I42); valvular heart disease (I34–I39); cardiac arrhythmia (I47–I49); pericardial disease (I31); coronary heart disease/ischemic heart disease (I20–I25); cardiac insufficiency (I50)

^*3At least one specific cardiac diagnosis in at least two quarters during the SHI insurance period 2017–2021

CI, Confidence interval; SHI, statutory health insurance carrier

A high prevalence of cardiac events was observed for 5-year survivors of the following cancers:

• Bone tumor (12.4%)

• Soft tissue sarcoma (9.5%)

• Renal tumor (8.7%)

• Acute myeloid leukemia (8.2%)

• Lymphoid leukemia (6.8%)

• Hodgkin lymphoma (7.6%)

• Non-Hodgkin lymphoma (5.9%)

The follow-up time is closely associated with the calendar period of cancer diagnosis. For a follow-up period of 5 to < 10 years after oncological therapy, thus recently treated children, the PRs ranged from 1.6 (CNS tumors) and 2.5 (Hodgkin lymphomas) to 9.6 (bone tumors) and 12.8 (neuroblastomas). For a follow-up period of 20 to < 30 years since oncological treatment, lower PRs were observed, ranging from 1.4 (CNS tumors) to 3.5 (renal tumors) (Table 3).

Table 3. Frequency of cardiac events between 2017–2021 in 5-year survivors of childhood cancer (diagnoses 1991–2011) stratified by follow-up duration and primary cancer diagnosis and in a matched^*1comparison group.

Follow-up period*2	Frequency of cardiac events N (%) *3 *4
	5 to < 10 years			10 to < 20 years			20 to < 30 years			All
Median age at 1 January 2017	13.5 (IQR = 8)			20.5 (IQR = 8)			28.5 (IQR = 8)			21.5 (IQR = 11)
Childhood cancer type	Survivors	Comparison group	PR [95% Cl]	Survivors	Comparison group	PR [95% Cl]	Survivors	Comparison group	PR [95% Cl]	Survivors	Comparison group	PR [95% Cl]
Lymphoid leukemia	85 (9.2)	56 (2.0)	4.6 [3.3; 6.3]	98 (5.9)	127 (2.6)	2.3 [1.8; 3.0]	53 (6.0)	88 (3.3)	1.8 [1.3; 2.5]	236 (6.8)	271 (2.6)	2.6 [2.2; 3.1]
Acute myeloid leukemia	13 (9.2)	8 (1.9)	4.9 [2.1; 11.5]	17 (7.5)	16 (2.3)	3.2 [1.6; 6.2]	9 (8.2)	14 (4.2)	1.9 [0.9; 4.3]	39 (8.2)	38 (2.6)	3.1 [2.0; 4.8]
Hodgkin lymphoma	9 (4.7)	11 (1.9)	2.5 [1.0; 5.8]	17 (6.1)	27 (3.2)	1.8 [1.0; 3.4]	25 (12.8)	35 (6.0)	2.1 [1.3; 3.5]	51 (7.6)	73 (3.6)	2.1 [1.5; 3.0]
Non-Hodgkin lymphoma	13 (5.5)	13 (18)	3.0 [1.4; 6.4]	24 (6.6)	25 (2.3)	2.9 [1.7; 5.0]	12 (5.1)	22 (3.2)	1.6 [0.8; 3.3]	49 (5.9)	60 (2.4)	2.5 [1.7; 3.5]
CNS tumor	29 (3.7)	54 (2.3)	1.6 [1.0; 2.5]	56 (4.8)	91 (2.6)	1.9 [1.3; 2.6]	20 (4.2)	44 (3.1)	1.4 [0.8; 2.3]	105 (4.4)	189 (2.6)	1.7 [1.3; 2.1]
Neuroblastoma	17 (8.8)	4 (0.7)	12.8 [4.3; 37.4]	16 (3.9)	21 (1.7)	2.3 [1.2; 4.3]	8 (4.3)	13 (2.3)	1.8 [0.8; 4.4]	41 (5.2)	38 (16)	3.2 [2.1; 5.0]
Renal tumor	22 (11.6)	11 (1.9)	6.0 [3.0; 12.1]	23 (6.0)	18 (16)	3.8 [2.1; 7.0]	21 (113)	18 (3.2)	3.5 [1.9; 6.4]	66 (8.7)	47 (2.1)	4.2 [2.9; 6.1]
Bone tumor	16 (12.4)	5 (13)	9.6 [3.6; 25.7]	27 (13.6)	16 (2.7)	5.1 [2.8; 9.2]	15 (10.8)	21 (5.0)	2.1 [1.1; 4.0]	58 (12.4)	42 (3.0)	4.1 [2.8; 6.1]
Soft tissue sarcoma	1 (8.8)	10 (18)	4.8 [2.2; 10.4]	32 (10.8)	21 (2.4)	4.6 [2.7; 7.8]	11 (7.8)	14 (3.3)	2.4 [1.1; 5.1]	59 (9.5)	45 (2.4)	3.9 [2.7; 5.7]
Intracranial germ cell tumor	0	0	/	1 (16)	12 (6.5)	0.3 [0.0; 1.9]	3 (10.7)	3 (3.6)	3.0 [0.6; 14.0]	4 (3.3)	15 (4.2)	0.8 [0.3; 2.4]
Other	16 (3.7)	25 (2.0)	1.9 [1.0; 3.6]	23 (3.7)	35 (1.9)	2.0 [1.2; 3.3]	10 (4.5)	16 (2.4)	1.9 [0.9; 4.1]	49 (3.9)	76 (2.0)	1.9 [1.4; 2.8]

^*1Matching criteria: age, sex, region of residence, duration of insurance

^*2Follow-up period: time from cancer diagnosis until 1 January 2017

^*3Defined according to the International Classification of Diseases (ICD-10): congestive heart failure/cardiomyopathy (I42), valvular heart disease (I34-I39), cardiac arrhythmias (I47-I49), pericardial disease (131), coronary heart disease/ ischemic heart disease (I20-I25), cardiac insufficiency (I50)

^*4At least one recorded cardiac diagnosis in at least two different quarters during the SHI insurance period 2017–2021

Cl, Confidence interval; CNS, central nervous system; IQR, interquartile range; N, absolute numbers; PR, prevalence ratio; SHI, statutory health insurance carrier; Survivors, 5-year survivors after childhood cancer

Specific cardiac events

A single cardiac condition was recorded in 5.1% of the 5-year survivors (n = 609), two cardiac conditions in 0.8% (n = 97), and three or more cardiac conditions in 0.4% (n = 51) of the patients with cancer in childhood or adolescence.

Cardiomyopathy was identified in 1.9% of 5-year survivors and in 0.1% of persons in the the comparison group and was associated with the highest group difference (PR = 14.9 [10.9; 20.5]) among the specific cardiac events (eTable). The younger study participants, with attained ages of 5 to < 10 years and 10 to < 20 years, showed the highest PRs, at PR = 32.0 [9.8; 104.2] and PR = 38.1 [19.3; 75.1], compared with study participants who were over 20 years of age on 1 January 2017. The group difference between former patients and the controls was higher in women than in men (PR = 21.0 [12.0; 36.8] versus PR = 12.3 [8.4; 18.1]).

eTable. Frequency of cardiac events between 2017–2021 in 5-year survivors of childhood cancer (diagnoses 1991–2011) and a matched^*1 comparison group, stratified by sex and age.

	Frequency (N, %)
	Survivors (N = 11 863)	Comparison group*1(N = 35 589)	Prevalence ratio 1 [95% CI]
Cardiomyopathy*2
Sex
Female	98 (1.9%)	14 (0.1%)	21.0 [12.0; 36.8]
Male	131 (2.0%)	32 (0.2%)	12.3 [8.4; 18.1]
Attained age at 1 January 2017 (years)
5 to < 10	32 (4.0%)	3 (0.1%)	32.0 [9.8; 104.2]
10 to < 20	113 (2.5%)	9 (0.1%)	38.1 [19.3; 75.1]
20 to < 30	65 (1.3%)	25 (0.2%)	7.7 [4.9; 12.2]
≥ 30	19 (1.3%)	9 (0.2%)	6.4 [2.9; 14.1]
Total	229 (1.9%)	46 (0.1%)	14.9 [10.9; 20.5]
Valvular heart disease*3
Sex
Female	161 (3.1%)	159 (1.0%)	3.0 [2.5; 3.8]
Male	162 (2.4%)	139 (0.7%)	3.5 [2.8; 4.4]
Attained age at 1 January 2017 (years)
5 to < 10	18 (2.3%)	11 (0.5%)	4.9 [2.3; 10.4]
10 to < 20	120 (2.7%)	122 (0.9%)	3.0 [2.3; 3.8]
20 to < 30	132 (2.6%)	118 (0.8%)	3.3 [2.6; 4.2]
≥ 30	53 (3.6%)	47 (1.1%)	3.4 [2.3; 5.0]
Total	323 (2.7%)	298 (0.8%)	3.3 [2.8; 3.8]
Cardiac arrhythmia*4
Sex
Female	105 (2.0%)	203 (1.3%)	1.6 [1.2; 2.0]
Male	75 (1.1%)	205 (1.0%)	1.1 [0.8; 1.4]
Attained age at 1 January 2017 (years)
5 to < 10	8 (1.0%)	11 (0.5%)	2.2 [0.9; 5.4]
10 to < 20	59 (1.3%)	120 (0.9%)	1.5 [1.1; 2.0]
20 to < 30	86 (1.7%)	186 (1.2%)	1.4 [1.1; 1.8]
≥ 30	27 (1.9%)	91 (2.1%)	0.9 [0.6; 1.4]
Total	180 (1.5%)	408 (1.2%)	1.3 [1.1; 1.6]
Pericardial disease*5
Sex
Female	14 (0.3%)	13 (0.1%)	3.2 [1.5; 6.9]
Male	14 (0.2%)	17 (0.1%)	2.5 [1.2; 5.0]
Attained age at 1 January 2017 (years)
5 to < 10	0 (0.0%)	0 (0.0%)	0
10 to < 20	6 (0.1%)	8 (0.1%)	2.3 [0.8; 6.6]
20 to < 30	14 (0.3%)	17 (0.1%)	2.4 [1.2; 5.0]
≥ 30	8 (0.6%)	5 (0.1%)	4.8 [1.6; 14.8]
Total	28 (0.2%)	30 (0.1%)	2.8 [1.7; 4.7]
Coronary heart disease*6
Sex
Female	20 (0.4%)	23 (0.2%)	2.6 [1.4; 4.8]
Male	25 (0.4%)	44 (0.2%)	1.7 [1.0; 2.8]
Attained age at 1 January 2017 (years)
5 to < 10	0 (0.0%)	1 (0.0%)	0
10 to < 20	6 (0.1%)	6 (0.0%)	3.0 [1.0; 9.4]
20 to < 30	22 (0.4%)	29 (0.2%)	2.3 [1.3; 3.9]
≥ 30	17 (1.2%)	31 (0.7%)	1.7 [0.9; 3.0]
Total	45 (0.4%)	67 (0.2%)	2.0 [1.4; 2.9]
Cardiac insufficiency*7
Sex
Female	59 (1.1%)	24 (0.2%)	7.4 [4.6; 11.9]
Male	58 (0.9%)	41 (0.2%)	4.2 [2.9; 6.3]
Attained age at 1 January 2017 (years)
5 to < 10	5 (0.6%)	1 (0.0%)	15.0 [1.8; 128.2]
10 to < 20	21 (0.5%)	15 (0.1%)	4.3 [2.2; 8.2]
20 to < 30	62 (1.2%)	32 (0.2%)	5.7 [3.8; 8.8]
≥ 30	29 (2.0%)	17 (0.4%)	5.2 [2.8; 9.4]
Total	117 (1.0%)	65 (0.2%)	5.4 [4.0; 7.3]

^*1Matching criteria: age, sex, region of residence, duration of insurance

^*2International Classification of Diseases (ICD-10): Heart failure/cardiomyopathy (I42); at least one recorded diagnosis in at least two different quarters during the SHI insurance period 2017–2021

^*3International Classification of Diseases (ICD-10): Valvular heart disease (I34-I39); at least one recorded diagnosis in at least two different quarters during the SHI insurance period 2017–2021

^*4International Classification of Diseases (ICD-10): Cardiac arrhythmia (I47-I49); at least one recorded diagnosis in at least two different quarters during the SHI insurance period 2017–2021

^*5International Classification of Diseases (ICD-10): Pericardial disease (I31); at least one recorded diagnosis in at least two different quarters during the SHI insurance period 2017–2021

^*6According to the International Classification of Diseases (ICD-10): coronary heart disease (I20-I25); at least one recorded diagnosis in at least two different quarters during the SHI insurance period 2017–2021

^*7According to the International Classification of Diseases (ICD-10): cardiac insufficiency (I50); at least one recorded diagnosis in at least two different quarters during the SHI insurance period 2017–2021

CI, Confidence interval; SHI, statutory health insurance carrier; Survivors, 5-year survivors after childhood cancer

The prevalence of valvular heart disease in 5-year survivors was 2.7%, compared with 0.8% among the matched controls (PR = 3.3 [2.8; 3.8]). The frequency of valvular heart disease increased with age in both groups. The younger 5-year survivors (< 10 years), with recent oncological therapy, showed a distinct difference from their controls in this regard (PR = 4.9 (2.3; 10.4]). Altogether, 1.5% of 5-year survivors were diagnosed with cardiac arrhythmia, against 1.2% in the comparison population (PR = 1.3 [1.1; 1.6]). Prevalence increased with attained age in both groups with the greatest group difference again found in the youngest age group (PR = 2.2 [0.9; 5.4]). Finally, the prevalence of pericardial disease was low in our study population, both in the 5-year survivors (0.2%) and in the comparison group (0.1%).

In an aging population of CCS, further cardiac sequelae may impair cardiac function. Coronary heart disease was present in 0.4% of 5-year survivors and 0.2% of comparison group members (PR = 2.0 [1.4; 2.9]). The prevalence of cardiac insufficiency was 1.0% among 5-year survivors and 0.2% in the comparison group (PR = 5.4 [4.0; 7.3] (eTable).

Discussion

In this registry-based cohort cardiac events were observed over a 5-year period in 6.4% of 11 863 5-year childhood cancer survivors and in 2.5% of a matched comparison group. A high prevalence of cardiac events was found in CCS with a history of bone tumor, soft tissue sarcoma, renal tumor, leukemia, or lymphoma. Cardiomyopathy, valvular heart disease, and arrhythmia were the most frequently occurring events among 5-year survivors (1.9%, 2.7%, 1.5%). The highest group difference between CCS and the comparison population was observed for cardiomyopathy (PR = 14.0 [10.9; 20.5].

Cardiac events

The Childhood Cancer Survivor Study (CCSS) is a multicenter cohort study with 31 participating centers in the US and Canada. The CCSS investigated the late effects of cancer and cancer treatments among 25 658 long-term survivors of pediatric cancer (diagnosed in 1970–1999; median age at follow-up/death: 38 years) (27, 28). The prevalence of self-reported major cardiovascular events was higher among CCS (n = 1780; 6.4%) than in 5051 siblings (n = 91; 1.8%) (11). The VersKiK study reproduced these findings, with a prevalence of 6.4% for cardiac events among German 5-year survivors.

We observed higher prevalence ratios in the follow-up period of 5 to < 10 years after cancer diagnosis than in the follow-up period of 20 to < 30 years, especially in patients with the following diagnoses: lymphoid leukemia, neuroblastoma, and bone tumor. These differences between CCS and the population-based comparison group, particularly earlier in the follow-up period may be due to an increased awareness on the part of the treating physicians and pediatricians and to targeted screening measures. Furthermore, both oncological therapy effects in recently diagnosed patients and changes in treatment regimes over time must be taken into consideration. Changes in therapy over recent decades have resulted in reduced radiation exposures, but more patients now receive cardiotoxic chemotherapy (29, 15).

Specific cardiac events

In a cross-sectional analysis of the St. Jude Lifetime (SJLIFE) cohort, cardiomyopathy was present in 7.4% of CCS (7). In contrast, a lower prevalence of 1.9% was observed among VersKiK 5-year survivors. Valvular heart disease occurred in 28% of the participants in the SJLIFE study (7) and in 2.7% of the CCS in the VersKiK study. Heart rhythm disorders had a prevalence of 4.6% in the SJLIFE study (7), but only 1.5% of the VersKiK CCS were diagnosed with cardiac arrhythmia.

Asymptomatic participants in the SJLIFE study underwent active medical examination. Hence, the high prevalence of specific cardiac events compared with VersKiK could be due to the diagnostic procedures used. In contrast, the assessment of cardiac events in VersKiK participants was based on SHI claims data, which refer only to clinically symptomatic diseases. Furthermore, the SJLIFE participants were older (median age 31 years versus 21.5 years in VersKiK). As might be expected we observed only a small number of cases of coronary heart disease in the younger VersKiK cohort (0.4%), while in the SJLIFE cohort coronary artery disease was present in 3.8% of asymptomatic CCS 22.6 years (median) after cardiotoxic treatment (7).

Limitations and strengths

The VersKiK study design has limitations. The considered insurance period (2017–2021) allows only a cross-sectional perspective and the estimation of prevalence for a short period of time. Furthermore, our study enrolled only CCS who had lived until at least 2017. This is associated with a survivorship bias: CCS who died before 2017 were not eligible for assessment of late effects. Despite the large sample size, the absolute number of participants with specific cardiac events in defined subgroups (stratification by attained age, primary cancer diagnosis, and duration of follow-up) is small, resulting in low precision of the estimates of late effects. A total of 13 SHI supported the VersKiK project, but private health insurance funds (about 10% of the population) were not part of the study. Finally, no information was available about cardiac diseases or genetic defects among CCS that predated the oncological treatment. It should also be taken into consideration that the presented comparison is influenced by the intensive medical monitoring of the CCS, especially in the first 5–10 years after cancer diagnosis—a level of care that is not offered to age-equivalent members of the general population, including the VersKiK comparison group.

The linkage of the GCCR data with the records of 13 SHI that cover 69% of the German statutorily insured population offers a good basis for quantitative analyses of late sequelae in CCS and their follow-up care in Germany. The VersKiK cohort comprised a large sample of 11 863 5-year survivors and a comparison group with 35 589 matched insured persons without childhood cancer. The recruitment of this sizeable population was not influenced by response bias. Both the 5-year survivors and the controls were insured by SHI, which minimized confounding by socio-economic status. Furthermore, the analysis of health events was based on claims data and was thus influenced neither by self-reporting nor by information bias. Hence, claims data are a good basis for non-selective quantification of cardiac late effects in 5-year survivors of childhood cancer.