PDF hosted at the Radboud Repository of the Radboud University
Nijmegen
The following full text is a publisher's version.
For additional information about this publication click this link.
http://hdl.handle.net/2066/165755
Please be advised that this information was generated on 2022-05-31 and may be subject to
change.
JOURNAL OF CLINICAL ONCOLOGY
ORIGINAL REPORT
Frederika A. van Nimwegen, Michael
Schaapveld, Michael Hauptmann, Karen
Kooijman, Berthe M.P. Aleman, and Flora
E. van Leeuwen, Netherlands Cancer
Institute, Amsterdam; Michael
Schaapveld, Netherlands Comprehensive
Cancer Organization; Judith Roesink,
University Medical Center Utrecht,
Utrecht; C
`
ecile P.M. Janus, Erasmus MC
Cancer Institute, Rotterdam; Augustinus
D.G. Krol, Leiden University Medical
Center, Leiden; Richard van der Maazen,
Radboud University Medical Center,
Nijmegen, Netherlands; David J. Cutter,
and Sarah C. Darby, University of Oxford;
and David J. Cutter, Oxford University
Hospitals NHS Trust, Oxford, United
Kingdom.
Published online ahead of print at
www.jco.org on November 16, 2015.
Supported by Grant No. NKI 2010-4720
from the Dutch Cancer Society and by
core funding to the Oxford University
Clinical Trial Services Unit.
Berthe M.P. Aleman and Flora E. van
Leeuwen are joint last authors.
Presented in part at the European
Symposium of Late Effects in Childhood
Cancer Survivors, Amsterdam,
Netherlands, September 30, 2011.
Presented at the International Conference
on Late Effects in Childhood Cancer
Survivors, Arlington, VA, June 12, 2015.
Presented at the 13th International
Conference on Malignant Lymphoma,
Lugano, Switzerland, June 17, 2015.
Authors’ disclosures of potential conflicts
of interest are found in the article online at
www.jco.org. Author contributions are
found at the end of this article.
Corresponding author: Flora E. van
Leeuwen, PhD, Department of
Epidemiology and Biostatistics,
Netherlands Cancer Institute,
Plesmanlaan 121, 1066 CX Amsterdam,
Netherlands; e-mail: f.v.leeuwen@nki.nl.
© 2015 by American Society of Clinical
Oncology
0732-183X/16/3403w-235w/$20.00
DOI: 10.1200/JCO.2015.63.4444
Radiation Dose-Response Relationship for Risk of Coronary
Heart Disease in Survivors of Hodgkin Lymphoma
Frederika A. van Nimwegen, Michael Schaapveld, David J. Cutter, C`ecile P.M. Janus, Augustinus D.G. Krol,
Michael Hauptmann, Karen Kooijman, Judith Roesink, Richard van der Maazen, Sarah C. Darby,
Berthe M.P. Aleman, and Flora E. van Leeuwen
See accompanying editorial on page 208
ABSTRACT
Purpose
Cardiovascular diseases are increasingly recognized as late effects of Hodgkin lymphoma (HL) treat-
ment. The purpose of this study was to identify the risk factors for coronary heart disease (CHD) and to
quantify the effects of radiation dose to the heart, chemotherapy, and other cardiovascular risk factors.
Patients and Methods
We conducted a nested case-control study in a cohort of 2,617 5-year HL survivors, treated between
1965 and 1995. Cases were patients diagnosed with CHD as their first cardiovascular event after HL.
Detailed treatment information was collected from medical records of 325 cases and 1,204 matched
controls. Radiation charts and simulation radiographs were used to estimate in-field heart volume
and mean heart dose (MHD). A risk factor questionnaire was sent to patients still alive.
Results
The median interval between HL and CHD was 19.0 years. Risk of CHD increased linearly with
increasing MHD (excess relative risk [ERR]) per Gray, 7.4%; 95% CI, 3.3% to 14.8%). This results in a
2.5-fold increased risk of CHD for patients receiving a MHD of 20 Gy from mediastinal radiotherapy,
compared with patients not treated with mediastinal radiotherapy. ERRs seemed to decrease with each
tertile of age at treatment (ERR/Gy
,27.5years
, 20.0%; ERR/Gy
27.5-36.4years
, 8.8%; ERR/Gy
36.5-50.9yea rs
,
4.2%; P
interaction
= .149). Having $ 1 classic CHD risk factor (diabetes mellitus, hypertension, or
hypercholesterolemia) independently increased CHD risk (rate ratio, 1.5; 95% CI, 1.1 to 2.1). A high level
of physical activity was associated with decreased CHD risk (rate ratio, 0.5; 95% CI, 0.3 to 0.8).
Conclusion
The linear radiation dose-response relationship identified can be used to predict CHD risk for future
HL patients and survivors. Appropriate early management of CHD risk factors and stimulation of
physical activity may reduce CHD risk in HL survivors.
J Clin Oncol 34:235-243. © 2015 by American Society of Clinical Oncology
INTRODUCTION
Hodgkin lymphoma (HL) tr eatment has impr ov ed
over recent decades, leading to a 10-year survival rate
of more than 80%.
1
Howev er , radiotherapy and
chemotherapy are associated with increased car-
diovascular morbidity and mortality in long-term
survivors.
2-7
Although radiation doses and target
volumes have been reduc ed over the past decades,
mediastinal radiotherapy is still indicated for a
substantial proportion of patie nts,
8,9
which may
result in considerable radiation exposure of the heart.
Few studies have examined the dose-response
relationship for cardiac radiation and risk of cor-
onary heart disease (CHD) after radiotherapy. A
recent study by Darby et al
10
showed a linear dose-
response relationship between radiation dose to
the heart and CHD risk in breast cancer survivors
for a relatively low range of mean heart dose
(MHD) (range, 0.03-27.7 Gy; average, 5 Gy). The
shape of the dose-response relationship has not
been studied in HL patients, who generally receive
much higher MHDs and are usually younger at
diagnosis than breast cancer patients.
Schellong et al
11
and Mulrooney et al
12
ob-
served an association between cardiovascular dis-
eases and prescribed mediastinal radiation dose
and MHD among childhood HL (and other
cancer) survivors; however, the shape of the
radiation dose-response relationship and excess
relative risks (ERRs) were not described.
© 2015 by American Society of Clinical Oncology 235
VOLUME 34
•
NUMBER 3
•
JANUARY 20, 2016
Downloaded from ascopubs.org by Radboud University Nijmegen on January 27, 2020 from 131.174.248.154
Copyright © 2020 American Society of Clinical Oncology. All rights reserved.
In addition to the shape of the dose-response relationship, the
roles of established cardiovascular disease risk factors and lifestyle
on CHD risk have rarely been studied among HL survivors.
4,5,13
Therefore, the aim of this study was to assess the shape of the dose-
response curve for cardiac radiation dose and the risk of CHD in
adolescent and adult HL survivors and to investigate the role of
chemotherapy, lifestyle, and other established cardiovascular dis-
ease risk factors.
PATIENTS AND METHODS
Study Population
We conducted a nested case-control study in an existing cohort (N =
2,617) of HL survivors treated in the Netherlands between 1965 and 1995.
The cohort was derived from hospital-based cancer registries of four large
university hospitals and one cancer center. Details on patient selection and
data collection have been published previously.
2,7,12,14-17
Patients were
eligible for this study if (1) they survived $ 5 years after HL diagnosis; (2)
they were diagnosed with HL before the age of 51 years; (3) HL was their
first primary malignancy, except for nonmelanoma skin cancer or car-
cinoma in situ of the cervix uteri or the breast; and (4) radiotherapy for HL
was the only radiotherapy given to the neck or trunk before the cutoff date,
which was defined as the date of CHD for the cases or the date of HL
diagnosis plus a time interval equal to the interval from the date of HL
diagnosis to the date of CHD diagnosis of the corresponding case for
matched controls.
Cases and Controls
Cases (n = 325) were patients who developed CHD in the form of
either symptomatic myocardial infarction or angina pectoris requiring
intervention (Common Terminology Criteria for Adverse Events, version
4.0, grade $ 2; Appendix Text A1, online only)
18
as their first clinically
significant heart disease. Cases were identified from medical records or
postal questionnaires completed by their general practitioners. Follow-up
was complete up to October 2013. For each case with CHD, we attempted
to select four controls from the cohort, individually matched on sex, age at
HL diagnosis (# 1 year), and date of HL diagnosis (# 3 years). Controls
had to be free of any cardiac disease grade $ 2 at the cutoff date. In total,
1,204 controls were matched to the cases.
Data Collection
Detailed information on treatment (including radiation doses and
fields and cumulative doses of cytotoxic drugs), medical history, medi-
cation use, smoking, and established cardiovascular risk factors at both
diagnosis of HL and dur ing follow-up was collected from medical records
and radiation charts. In addition, a questionnaire on established car-
diovascular risk factors and lifestyle was mailed to all patients still alive in
2013 (n = 475) in three of the five centers (response rate, 70%). Patients
were defined as having a risk factor when the risk factor concerned was
diagnosed before CHD or the cutoff date. The ethics review board of the
Netherlands Cancer Institute approved this study.
Mean Heart Dose
The MHD was assessed using the percentage of cardiac volume within
field (%CVWF) method
19
and converted to equivalent dose in 2-Gy
fractions (EQD2).
20
We recently showed that this method gives reliable
MHD estimates for our patient population and compares well with MHD
based on computed tomog raphy (CT)-based dosimetry.
19
We outlined the
cardiac contour on the HL simulation radiographs to obtain the %CVWF.
Additional details can be found in Appendix Text A2.
When or iginal radiotherapy charts were unavailable, information
about radiotherapy, including dates, anatomic areas, dose, fractionation,
and treatment energy, was abstracted from clinical notes. We assigned an
average %CVWF to radiation-treated patients for whom no simulation
radiographs were available and an average prescribed dose to patients for
whom no prescribed dose was available (n = 473, including 105 cases), on
the basis of hospital, treatment period, and sex.
Statistical Analysis
Odds ratios for CHD for different levels of each factor were calculated
using conditional logistic regression on sets of individual cases and their
matched controls, and were interpreted as rate ratios (RRs). The Wald
method was used to calculate 95% CIs for factors with two levels. The
amount of information in each category, including the reference category
(so-called floating absolute risks), was used to calculate 95% CIs for factors
with more than two levels.
21
Multivariable regression was used to assess
and control for confounding and to evaluate interactions between radi-
ation dose and other factors.
The dose-response relationship was estimated by modeling the CHD
rate as K
m
(1 + bd), where K
m
is a constant specific to each matched set, b is
the ERR of CHD per unit increase in dose, and d is the MHD of an
individual patient. Nonlinearity was evaluated by including an exponential
term: K
m
[1 + bd$exp(dd)]. Goodness of fit was assessed by likelihood ratio
tests. Interactions were evaluated using interaction terms and likelihood
ratio tests. Approximate cumulative incidence of CHD for categories of
MHD, with other heart disease or death as a competing risk, was estimated
from CHD RRs together with the cumulative risk of CHD for the entire
cohort, assuming that the distribution of all individuals in the cohort
across the dose categories was equal to that for the control patients.
Significance tests were two-sided and P # 0.05 was considered to
indicate statistical significance. Analyses were performed using STATA
statistical software (version 13.0; STATA, College Station, TX) and Epicure
(version 1.8; Hiro Soft International Inc, Seattle WA).
RESULTS
Characteristics of the 325 cases and 1,204 controls are described in
Table 1. The median age of patients was 32.2 years (interquartile
range [IQR], 24.4 to 39.6) at the time of HL diagnosis, and the
median interval between HL and CHD was 19.0 years (IQR, 13.9 to
25.2). Myocardial infarction was diagnosed in 185 patients; angina
pectoris requiring intervention was diagnosed in 140 patients
(Appendix Table A1). In total, 169 of 325 cases died, 42.6% from a
cardiovascular disease, after a median follow-up period of 6.0 years
after their first CHD (Appendix Table A1). Thirty-one patients
died of their first CHD incident within a week.
Radiotherapy
Ninety-one percent of the cases had receiv ed mediastinal radi-
otherapy, given through parallel-opposed fields, c ompar ed with 79%
of the controls (Table 1). Mediastinal radiation therapy was associated
with a 2.63-fold increased risk of CHD (95% CI, 1.74 to 3.99; Table 2).
P ara-aortic radiotherapy, with or without splenic radiation, was not
associated with CHD risk (RR, 0.99; 95% CI, 0.76 to 1.28).
The av erage MHD was 22.0 Gy for cases and 20.4 Gy for controls
(Table 1). A linear radiation dose-response relationship best described
the data, and no significant deviation from linearity was observed
(P
exponential-term
= .356). The ERR for CHD increased by 7.4% per Gy
(95% CI, 3.3% to 14.8%; Fig 1), resulting in a 1.74-fold increased risk
at a MHD of 10 Gy (95% CI, 1.33 to 2.48) and a 2.48 -fold increased
risk at a MHD of 20 Gy (95% CI, 1.66 to 3.96). The approximate 25-
year cumulative CHD incidence was 4.1% for patients with a MHD of
236 © 2015 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY
van Nimwegen et al
Downloaded from ascopubs.org by Radboud University Nijmegen on January 27, 2020 from 131.174.248.154
Copyright © 2020 American Society of Clinical Oncology. All rights reserved.
0Gy,9.4%forpatientswithaMHDof15to20Gy,and12.6%for
patients with a MHD of $ 25 Gy (Fig 2). Results were similar when we
only included patients for whom the MHD was known (Appendix
Table A2). Cases had a median %CVWF of 66%, compared with 64%
for controls. Variation in %CVWF was limited, with interquartile
ranges of 57% to 71% and 55% to 70%, respectively (Tab le 1).
Other Treatment-Relate d Risk Facto rs
Chemotherapy was not associated with CHD r isk (RR,
0.87; 95% CI, 0.67 to 1.13), nor were anthracycline-containing
chemotherapy (RR, 1.11, 95% CI, 0.76 to 1.62) or vincristine-
cont aining chemotherapy (RR, 0.86 ; 95% C I, 0.66 to 1.13), after
accounting f or mediastinal radiotherapy. Splenectomy also
Table 1. Patient Demographics and Clinical Characteristics
Characteristic No. of Cases* % No. of Controls* %
Total 325 100 1,208 100
Sex
Men 236 72.6 889 73.9
Women 89 27.4 319 26.4
Age at diagnosis, years (median, IQR) 32.3 24.5-39.4 32.2 24.4-39.6
26 98 30.2 370 30.6
26-32 73 22.5 271 22.4
33-39 80 24.6 292 24.2
40-50 74 22.8 275 22.8
Year of diagnosis
1965-1974 116 35.7 429 35.5
1975-1984 124 38.2 513 32.5
1985-1995 85 26.2 266 22.0
Time to IHD/cutoff (median, IQR) 19.0 13.9-25.2 19.2 13.9-25.2
Smoking
Smoked at HL diagnosis 195 61.1 645 55.1
Smoking at end of follow-up 93 31.1 321 30.0
Ever smoked 236 74.0 820 69.7
Recent smoker at time of cutoff (, 5 years) 109 34.0 350 29.0
Unknown time of quitting smoking 77 23.7 251 20.8
Classic risk factors
Diabetes mellitus diagnosed before CHD/cutoff date 11 3.4 38 3.2
Hypercholesterolemia diagnosed before CHD/cutoff date 31 9.5 89 7.4
Hypertension diagnosed before CHD/cutoff date 54 16.6 122 10.1
Obesity at HL 16 5.2 33 3.0
Obesity at end of follow-up (BMI $ 30 kg/m
2
) 106 36.3 288 27.3
At least one of the above risk factors 80 24.6 213 17.6
Treatment of HL†
Radiotherapy 315 96.9 1097 90.8
Subdiaphragmatic radiotherapy 157 48.3 573 47.4
Mediastinal radiotherapy 296 91.1 957 79.2
Chemotherapy 200 61.5 805 66.6
Alkylating CT 167 84.3 686 87.1
Procarbazine 139 42.9 614 50.9
Vincristine 135 41.7 585 48.5
Anthracyclines 68 21.0 226 18.7
Splenectomy 103 32.0 384 32.5
Prescribed mediastinal dose, Gy (median, IQR)‡ 33 29-37 33 29-38
0 29 8.9 251 20.8
15-24 5 1.5 28 2.3
25-34 26 8.0 98 8.1
35-39 156 48.0 537 44.5
40-45 109 33.5 294 24.3
Mean heart dose, Gy (median, IQR) 21.7 18.4-25.7 20.2 17.5-24.8
0 17 5.2 160 13.3
1-5 12 3.7 93 7.7
5-14 19 5.9 80 6.6
15-1y 71 21.8 242 20.0
20-24 102 31.4 332 27.5
25-34 99 30.5 280 23.2
35-45 5 1.5 21 1.7
Percent cardiac contour within field (median, IQR) 64 55-70 66 57-71
NOTE. All patients were treated with parallel-opposed fields.
Abbreviations: BMI, body mass index; CHD, coronary heart disease; CT, computed tomography; HL, Hodgkin lymphoma; IHD, ischemic heart disease; IQR, interquartile
range.
*Two-hundred fifty-six cases had four controls, 48 cases had three controls, 15 cases had two controls, and six cases had only one control.
†Treatment variables are not mutually exclusive.
‡Prescribed dose was missing for five cases and 19 controls, simulation radiographs were missing for 84 cases and 271 controls, and both were missing for 16 cases
and 78 controls. Imputation was on the basis of hospital, sex, and treatment period.
www.jco.org © 2015 by American Society of Clinical Oncology 237
Dose-Response for Coronary Heart Disease After HL
Downloaded from ascopubs.org by Radboud University Nijmegen on January 27, 2020 from 131.174.248.154
Copyright © 2020 American Society of Clinical Oncology. All rights reserved.
did not affect CHD risk (RR, 0.91; 95% CI, 0.68 to 1.22;
Table 2).
Patient-Related Risk Factors
Twenty-five percent of cases had at least one classic car-
diovascular risk factor (diabetes mellitus, hypercholesterolemia, or
hypertension) diagnosed before the diagnosis of CHD (Table 1).
Only hypertension (RR, 1.85; 95% CI, 1.28 to 2.66) and the
presence of at least one risk factor (RR, 1.59; 95% CI, 1.17 to 2.19)
were associated with an increased risk of CHD (Table 3). When risk
factors were taken into account in a less conservative manner, that
is, by also including risk factors that were diag nosed around the
time of CHD diagnosis or cutoff date, not only hypertension but
also diabetes mellitus and hypercholesterolemia were associated
with a significantly increased risk of developing CHD (Appendix
Table A3). Obesity at the time of CHD diagnosis or the cutoff date
was associated with an increased risk of CHD as well (RR, 1.64;
95% CI, 1.24 to 2.16). Whereas ever smoking was not associated
with CHD risk, smoking within 5 years before a diagnosis of CHD
or the cutoff date was associated with an increased risk of CHD
(RR, 1.56; 95% CI, 1.13 to 2.15; Table 3). Patients with a high level
of physical activity at the time of the follow-up questionnaire (. 3
Table 2. Risk of Coronary Heart Disease According to Hodgkin Lymphoma Treatment
Treatment Factor
No. of Cases
(n = 325)
No. of Controls
(n = 1,204)
Crude
P
Adjusted*
PRR 95% CI RR 95% CI
Radiotherapy
No 10 111 1 ref 1 ref
Yes 315 1093 3.16 1.63 to 6.14 .001 2.99 1.52 to 5.85 .001
Mediastinal radiotherapy
No 29 251 1 ref 1 ref
Yes 296 953 2.71 1.79 to 4.08 , .001 2.63 1.74 to 3.99 , .001
Para-aortic radiotherapy
No 168 633 1 ref 1 ref
Yes 157 571 1.05 0.81 to 1.35 0.712 .99 0.76 to 1.28 .927
Splenic radiotherapy
No 229 878 1 ref 1 ref
Yes 96 326 1.13 0.86 to 1.50 .372 1.07 0.80 to 1.42 .656
Prescribed mediastinal dose, Gy
0 (no medias tinal radiotherapy) 29 251 1.00 0.68 to 1.48 1.00 0.67 to 1.48
15-24 5 28 1.49 0.57 to 3.89 1.51 0.58 to 3.96
25-34 26 97 2.31 1.50 to 3.57 2.30 1.49 to 3.56
35-39 156 535 2.56 2.12 to 3.08 2.52 2.10 to 3.03
40-45 109 293 3.12 2.50 to 3.90 , .001† 3.03 2.41 to 3.82 , .001†
Mean heart dose, Gy
0 17 160 1.00 0.60 to 1.66 1.00 0.60 to 1.67
1-5 12 93 1.19 0.65 to 2.19 1.14 0.62 to 2.10
5-14 19 80 2.16 1.30 to 3.60 2.14 1.28 to 3.58
15-19 71 239 2.83 2.16 to 3.71 2.76 2.10 to 3.59
20-24 102 332 2.90 2.32 to 3.63 2.79 2.23 to 3.49
25-34 99 279 3.35 2.64 to 4.26 3.21 2.52 to 4.09
35-45 5 21 2.62 0.99 to 6.90 , .001† 2.54 0.96 to 6.69 , .001†
Chemotherapy
No 125 402 1.00 ref 1 ref
Yes 200 802 0.79 0.61 to 1.02 .069 0.87 0.67 to 1.13 .298
Alkylating chemotherapy
No 159 532 1 ref 1 ref
Yes 166 673 0.80 0.62 to 1.04 .101 0.92 0.70 to 1.20 .519
Procarbazine
No 185 591 1 ref 1 ref
Yes 139 612 0.70
0.54 to 0.91 .008 0.82 0.63 to 1.07 .148
Vincristine
No 189 620 1 ref 1 ref
Yes 165 583 0.73 0.56 to 0.95 .020 0.86 0.66 to 1.13 .294
Anthracyclines
No 256 978 1 ref 1 ref
Yes 68 224 1.08 0.75 to 1.57 .670 1.11 0.76 to 1.62 .593
Splenectomy‡
No 219 794 1 ref 1 ref
Yes 103 384 1.00 0.75 to 1.33 .990 0.91 0.68 to 1.22 .521
NOTE. All patients were treated with parallel-opposed field. Boldface indicates statistically significant RRs.
Abbreviations: ref, reference category; RR, rate ratio.
*Radiation-related factors are adjusted for any chemotherapy. Chemotherapy factors are adjusted for mediastinal radiotherapy. Splenectomy was adjusted for
mediastinal radiotherapy and any chemotherapy.
† P for trend.
‡Total numbers of cases and controls may vary because of missing values or inclusion of only patients treated with mediastinal radiotherapy.
238 © 2015 by American Society of Clinical Oncology JOURNAL OF CLINICAL ONCOLOGY
van Nimwegen et al
Downloaded from ascopubs.org by Radboud University Nijmegen on January 27, 2020 from 131.174.248.154
Copyright © 2020 American Society of Clinical Oncology. All rights reserved.
