Associations between cardiopulmonary fitness and cardiovascular events in survivors of childhood cancer: A report from the St. Jude Lifetime Cohort

Abstract

PURPOSE:

Childhood cancer survivors are at increased risk of premature cardiovascular events compared to peers. Increased cardiopulmonary fitness reduces the risk of cardiovascular morbidity/mortality within the general population but are poorly described in cancer survivors. We examined the associations between fitness and cardiovascular events in childhood cancer survivors.

METHODS:

Participants (n=2,433) completed a baseline, cardiopulmonary exercise test (CPET) to assess peak maximal oxygen consumption (VO_2peak). Metabolic equivalents (METs) were calculated by dividing VO_2peak by 3.5 ml·kg¹·min and peak METs achieved on CPET was used to document cardiopulmonary fitness. Additionally, we categorized participants (based on age- and sex-matched controls) as low (<50^th percentile ofachieved METs) and normal (≥50^th percentile). Subsequent cardiovascular disease was graded with the Common Terminology Criteria for Adverse Events v. 4.03. Associations between peak METs and subsequent cardiovascular disease in survivors were evaluated with multivariable Cox-proportional hazard regression, adjusted for cancer treatment, lifestyle, baseline cardiovascular disease, and cardiovascular risk factors. Additionally, a univariate analysis was conducted to examine the peak METs achieved on the CPET in survivors who died from a cardiovascular event and those who did not.

RESULTS:

Each 1 MET increase on the survivor’s CPET performance decreased the risk of incident cardiovascular disease (Hazard Ratio [HR] 0.80, 95% Confidence interval [CI] 0.72, 0.90). Among survivors with low baseline cardiopulmonary fitness, those who achieved 1 MET higher value on their CPET had lower risk of incident cardiovascular disease (HR:0.78, 95% CI 0.65, 0.96). The average peak METs achieved was lower (5.9 ± 2.17) among survivors who died from cardiovascular disease compared to those who did not (7.6 ± 2.5).

CONCLUSION:

Higher cardiopulmonary fitness was associated with lower risk for incident cardiovascular disease. Early identification of survivors with low cardiopulmonary fitness provides opportunities for risk mitigation through promotion of regular physical activity.

Introduction

Over the past four decades, progress in therapy for pediatric malignancies has resulted in dramatic improvement in survival.(1) With current five-year survival rates exceeding 85%, the number of long-term childhood cancer survivors residing in the United States (U.S.) is estimated to be over 522,000.(1) Unfortunately, cancer therapy is associated with an increased risk for early onset morbidity and mortality.(2, 3)

Chest radiation and anthracycline-based chemotherapy increase the risk of cardiac disease, one of the leading causes of non-neoplastic death in this population.(4, 5) While survivors cannot change their treatment history, it may be possible to reduce their risk for cardiac morbidity by engaging in a lifestyle that improves their cardiopulmonary fitness. For example, survivors traditionally exposed to cardiotoxic agents had a lower risk of developing serious cardiac conditions (Risk Ratio [RR]: 0.45, 95% CI: 0.26, 0.80) if they reported participating in 9 to 12 metabolic equivalent hours of exercise per week compared to survivors who reported no weekly exercise.(6, 7)

While associations of self-reported physical activity (PA) with cardiac morbidity and mortality have been documented in adult survivors of childhood cancer, this is prone to bias related to responder understanding and recall accuracy and does not provide all the information needed for effective intervention to this population. (8–10) Measured cardiopulmonary fitness, which is associated with self-reported PA, is an important physiologic metric that has been shown to be independently associated with cardiac morbidity and mortality throughout adulthood.(11, 12) Direct measurements of VO₂ peak for cardiopulmonary fitness improve the accuracy of cardiopulmonary disease prediction models and can be used as a biomarker of response to PA when prescribed as an intervention. Because physiological response to PA varies across persons and populations, and childhood cancer survivors may have a blunted physiological response to PA perturbation, assessing whether cardiopulmonary fitness is associated with cardiopulmonary morbidity and mortality in this population is important.(13–15) An objective measure of cardiopulmonary fitness may be used as a potential biomarker to enhance screening and risk prediction as well as evaluate the efficacy of PA interventions. Thus, the purpose of this study was to evaluate associations between clinically ascertained cardiopulmonary fitness and the risk of developing subsequent cardiac disease.

Methods

Study Population

Participants of this prospective cohort study included members of the St. Jude Lifetime Cohort (SJLIFE), a study designed to assess late health outcomes among aging survivors of childhood cancer.(16–18) For these analyses, we included adult survivors (≥18 years of age) of any childhood cancer treated at St. Jude Children’s Research Hospital (SJCRH) between 1962 and 2012, were <18 years when treated, and who completed both self-reported questionnaires and a symptom-limited maximal cardiopulmonary exercise test (CPET) as part of an on-campus assessment. Study measures and documents were approved by the SJCRH Institutional Review Board. Participants provided written informed consent before study activities.

Cardiopulmonary Fitness

Cardiopulmonary fitness was assessed with a CPET using breath by breath analysis (Ultima CardiO2; MGC Diagnostics, St Paul, MN) to determine peak ventilatory oxygen consumption (VO₂ peak) in milliliters per kilogram per minute.(19) The Bruce protocol treadmill test was used for most patients.(20) However, if a participant had amputation or poor balance, an arm or cycle ergometer with a ramping protocol was used.(21) Heart rhythm, using continuous 12-lead ECG, and blood pressure, using an automatic sphygmomanometer, were monitored for safety during testing and not used as outcomes or covariates in analysis. Resting blood pressure and heart rate (HR) were measured and recorded prior to the start of the test after five minutes of quiet sitting. Participants were then asked to exercise until perceived maximal exertion or they reached peak oxygen uptake (VO₂ peak, determined by plateau of VO₂ with increasing workloads, respiratory exchange ratio of >1.10, or 90% of maximum heart rate (220 beats per minute minus the participants age)).(22, 23) Tests were terminated for clinical concerns (N=196) such as electrocardiographic evidence of ischemia (≥ 2mm of ST depression) and hypertensive blood pressure (BP) response (>220/110mm Hg, or chest pain).(23) After a one-minute cooldown (walking at 1.5 mph and no incline, or unloaded cycling), participants sat in a chair quietly for five minutes where HR and blood pressure were monitored. Peak metabolic equivalents (METs) were calculated by dividing the participant’s peak VO₂ by 3.5. Additionally, age- and sex-specific percentiles were calculated from the survivors peak METs achieved on the CPET. Participants were then categorized into low (< 50^th percentile) or normal cardiopulmonary fitness (≥ 50^th percentile).(24)

Chronic cardiovascular conditions

Seven cardiac conditions/events were ascertained using data from medical records, self-report questionnaires, and in-person comprehensive medical evaluations (laboratory data, imaging, and clinical assessment). Frequent contact with participants queried subsequent events. Cardiomyopathy (including congestive heart failure), coronary artery disease (including myocardial infarction), left ventricular systolic dysfunction, vascular disease, cerebrovascular disease, and cerebrovascular accidents were graded using a modification of the National Cancer Institute Common Terminology Criteria for Adverse Events version 4.03 (CTCAE).(25) Conditions were graded as moderate (grade 2), severe/disabling (grade 3), life-threatening (grade 4), or fatal (grade 5). Cardiac conditions present prior to or determined concomitant to the CPET visit were considered prevalent at baseline. Conditions occurring more than 60 days after the CPET visit and not prevalent at baseline and those prevalent at baseline but that progressed to a higher grade were considered incident events. Conditions with potential for recurrence (myocardial infarction, cerebrovascular accident) were always considered incident events if they occurred 60 days after the baseline assessment, even if the participant already had a baseline myocardial infarction or cerebrovascular accident. Participants were followed through April 30, 2020, and those with no events were censored on this date.

Cardiovascular risk factors

Obesity, hypertension, and abnormal glucose metabolism were ascertained at the CPET visit using data from in-person comprehensive medical evaluations (laboratory data, and clinical assessment). Each condition was assigned a CTCAE grade as described above.(25) Prevalent hypertension and abnormal glucose metabolism were additionally categorized into treated, if the assigned grade was based on medication use, and untreated, if the grade assigned was based on abnormal laboratory values.

Other measures

Cancer diagnoses and treatment exposures were obtained from medical records by trained abstractors. For these analyses, anthracyclines, corticosteroids, alkylating agents, heavy metals, and bleomycin chemotherapy exposures were included as dichotomous variables. Radiation therapy to the pelvis, chest, brain, abdomen, or neck were also included as dichotomous variables. Demographic information including age at assessment, age at diagnosis, race, smoking status, and educational attainment were obtained via questionnaire completed before the baseline CPET visit.(26) Smoking status was categorized into past smokers (reported smoking at least 100 cigarettes in their lifetime), current smokers, and non-smokers. Current physical activity levels were determined using six items from the National Health and Nutrition Examination Survey.(27) The questions asked if the participant spent at least ten minutes doing vigorous PA, on how many days per week, and for how many total minutes per day they performed these activities. Identical questions were asked for moderate PA. To calculate the weekly METs of both vigorous and moderate activities, weekly minutes of vigorous activity were multiplied by six, weekly minutes of moderate activities were multiplied by three.(28) Weekly accumulations of vigorous and moderate METs were then summed to get MET minutes per week. For analysis, PA was dichotomized into meeting (≥450 MET minutes per week) or not meeting the 2018 Center for Disease Control and Prevention (CDC) guidelines.(29)

Statistical Analysis

To evaluate selection bias, the characteristics of participants (eligible adults who agreed to be in the study) and non-participants (eligible adults who declined to be in the study) were compared using chi-squared tests for categorical variables, and two sample t-tests or Mann-Whitney U tests for continuous variables. We compared age at assessment, age at diagnosis, sex, race, primary neoplasm, and primary treatment exposure (alkylating agents, heavy metals, anthracycline, corticosteroids, bleomycin and chest radiation) between participants and non-participants. A Kaplan Meier curve was used to examine the association between cardiopulmonary fitness (categorized as low and normal) and incident cardiac event (defined as grade 2–5) and compared using log-rank test. Then, multivariable Cox-proportional hazard regression was used to examine the associations between cardiopulmonary fitness, used as both a categorical variable (low and normal), and as a continuous variable (peak METs), and incident cardiac events. Time from CPET to incident event was used as the timescale and participants were censored at date of cardiac event or the last follow-up date (April 30^th, 2020) if they were alive without cardiac event. Models were adjusted for baseline cardiac condition, cardiac risk factors, PA status, age at baseline, age at diagnosis, and smoking status. The multivariable analysis involving the continuous cardiopulmonary fitness (peak METs) was performed in all survivors, as well as in survivors with low and normal cardiopulmonary fitness, separately. Due to a small number of events, a univariate analysis was performed to evaluate the mean METs achieved on the CPET in survivors who died from a cardiac event and in survivors who did not die from a cardiac event. Data were analyzed with SAS version 9.4 (SAS Institute, Cary, NC).

Results

Characteristics of participants

There were 2,837 survivors of childhood cancer potentially eligible for this analysis. Of these, 2,433 (85.7%) completed a CPET. The remaining survivors declined participation (n=17), completed a survey only (n=293), did not attend their CPET appointment (n=16), or could not complete their CPET due to time limitations (n=11), equipment malfunction (n=2), or scheduling limitations (n=56). Among the participants, 237 survivors were not tested due to medical contraindications to exercise (Figure 1).

Figure 1.

CONSORT diagram

Compared to non-participants, participants were younger at CPET assessment/campus visit, and older at diagnosis. There were no differences between treatments received or primary diagnoses in participants compared to non-participants (Table 1). Compared to those with low cardiopulmonary fitness, participants who had normal peak MET value (n=1,099, 47.3%), were more likely to meet PA guidelines and less likely to have hypertension, obesity, or abnormal glucose metabolism. (Table 2). The median follow-up time for survivors was 2.4 years (Interquartile range [IQR]: 0.15, 8.1 years).

Table 1.

Characteristics of survivors of childhood cancer participating and non-participants in the St Jude Lifetime Cohort.

	Participants n = 2433		Non-participants n= 404		p
Age in years, median (range)a	27.9	(18.0 – 64.8)	32.3	(18.0 – 63.3)	<0.01
Age at diagnosis in years, median (range)a	9.2	(0.1 – 24.5)	8.8	(0.1 – 23.5)	<0.01
Sex, n (%)
Male	1260	(51.8)	191	(47.3)	0.09
Female	1173	(48.2)	213	(52.7)
Race, n (%)b
White	1,999	(82.2)	345	(85.4)	<0.01
Black	368	(15.1)	57	(13.5)
Other	65	(2.7)	1	(0.25)
Unknown	1	(0.04)	1	(0.25)
Primary neoplasm, n (%)
Acute lymphoblastic leukemia	658	(27.0)	102	(25.3)	0.51
Acute myeloid leukemia	89	(3.7)	22	(5.5)
Other leukemia	17	(0.7)	4	(0.9)
Hodgkin lymphoma	340	(14.0)	51	(12.6)
Non-Hodgkin lymphoma	152	(6.3)	30	(7.4)
Central nervous system	366	(15.0)	62	(15.3)
Wilms tumor	139	(5.7)	27	(6.7)
Neuroblastoma	103	(4.2)	9	(2.2)
Bone tumor	188	(7.7)	27	(6.7)
Retinoblastoma	74	(3.0)	14	(3.5)
Soft tissue sarcoma	152	(6.3)	29	(7.2)
Other	155	(6.4)	27	(6.7)
Alkylating agent, n (%)
Yes	1,452	(59.7)	233	(57.7)	0.38
No	981	(40.3)	171	(42.3)
Heavy metal, n (%)
Yes	368	(15.1)	60	(14.8)	0.88
No	2,065	(84.9)	344	(85.2)
Anthracycline, n (%)
Yes	1,400	(57.5)	226	(55.9)	0.55
No	1,033	(42.5)	178	(44.1)
Corticosteroids, n (%)
Yes	1,055	(43.4)	158	(39.1)	0.11
No	1,378	(56.6)	246	(60.9)
Bleomycin, n (%)
Yes	180	(7.4)	24	(5.9)	0.29
No	2,253	(92.6)	380	(94.1)
Chest radiation, n (%)
Yes	695	(28.5)	110	(27.2)	0.58
No	1,738	(71.5)	294	(72.8)

Chi-square or t-test unless otherwise noted

^aMann-Whitney U test was used

SD = standard deviation;

Table 2.

Characteristics of survivors with low and normal cardiopulmonary fitness levels based on if they are above or below the median of percent predicted VO₂

	Normal cardiopulmonary fitness		Low cardiopulmonary fitness		p
	n = 1,099		n = 1,221
Age in years, mean median (range)a	27.9	(18.0 – 64.5)	27.7	(18.0 – 64.8)	0.48
Age at diagnosis in years, median (range)a	9.6	(0.1 – 24.4)	9.1	(0.1 – 23.5)	0.15
Sex, n (%)
Male	565	(51.4)	633	(55.4)	0.59
Female	534	(48.9)	588	(44.6)
Race, n (%)b
White	948	(86.3)	961	(78.7)	<0.01
Black	112	(10.2)	233	(2.2)
Other	39	(3.5)	27	(19.1)
Peak METS achieved on CPET, mean (SD)	9.1	(2.3)	6.1	(1.7)	<0.01
Met physical activity guidelinesc, n (%)
Yes	707	(64.3)	478	(39.1)	<0.01
No	392	(35.7)	743	(60.9)
Smoker, n (%)
Past	149	(13.6)	146	(11.9)	0.38
Current	189	(17.2)	199	(16.3)
Never	761	(69.2)	876	(71.8)
Cardiovascular disease at baseline, n (%)
Yes	112	(10.2)	248	(20.3)	<0.01
No	987	(89.8)	973	(79.7)
Obesity at baseline, n (%)
Yes	722	(65.7)	848	(69.5)	0.05
No	377	(34.3)	373	(30.5)
Hypertension at baseline, n (%)
Yes	211	(19.2)	344	(28.2)	<0.01
No	888	(80.8)	877	(71.8)
Abnormal glucose metabolism at baseline, n (%)
Yes	56	(5.1)	137	(11.2)	<0.01
No	1,043	(94.9)	1,084	(88.8)

Chi-square or t-test unless otherwise noted

^aMann-Whitney U test was used

^bFisher’s exact test used

^cPhysical activity guidelines based off the Center for Disease Control’s 2008 physical activity recommendations of 450 metabolic equivalent hours per week⁵⁶

SD = standard deviation; METS = metabolic equivalent, CPET = cardiopulmonary exercise test, SD= standard deviation

Any incident cardiac event

There were 65 (5.3%) incident cardiac events among survivors who had low cardiopulmonary fitness and 31 (2.8%) incident cardiac events among survivors with normal cardiopulmonary fitness (log-rank p: <0.01 [Figure 2]). After adjusting for the variables discussed above, survivors with low cardiopulmonary fitness had a 1.82 (95% CI: 1.17, 2.83) increased risk of developing a future cardiac event compared to those with normal cardiopulmonary fitness. Among all participants, each 1 MET increase in baseline CPET performance was associated with a 20% (HR: 0.80, 95% CI: 0.72, 0.90) decrease in risk of an incident cardiac event (Table 3). Among survivors with low cardiopulmonary fitness, each 1 MET increase in baseline CPET performance was associated with a 22% (HR: 0.78, 95% CI: 0.65, 0.96) reduction in risk for future cardiac events (Table 4). Among survivors with normal cardiopulmonary fitness, there was no statistically significant association between higher MET on baseline CPET and future risk of cardiac events. (Table 4).

Figure 2.

Kaplan-Meier survival curve of incident cardiovascular disease in survivors low and normal fitness

Table 3.

Multivariable cox regression analysis for risk of incident cardiovascular disease in all childhood cancer survivors

	Incident cardiovascular event
	HR	95% CI
Cardiopulmonary fitness
Increase of one MET at baseline	0.80	0.72, 0.90
Cardiovascular disease at baseline
Yes	2.08	1.31, 3.29
No	1.0	1.0
Obesity at baseline
Yes	0.73	0.44, 1.18
No	1.0	1.0
Impaired glucose metabolism at baseline
Yes	1.42	0.75, 2.69
No	1.0	1.0
Hypertension at baseline
Yes	0.72	0.43, 1.20
No	1.0	1.0
Anthracyclines
Yes	2.27	1.44, 3.56
No	1.0	1.0
Chest radiation
Yes	1.58	1.05, 2.39
No	1.0	1.0

Model additionally adjusted for sex, age at assessment, age at diagnosis, meeting 2018 CDC guidelines for physical activity, smoking

Note: 237 participants missing their MET at baseline due to acute injury (1), cardiovascular disease (112), uncontrolled diabetes (26), gastrointestinal disease (1), kidney disease (16), liver failure (2), musculoskeletal disease (40) paralysis (4) pulmonary disease (15), recent stroke (9), surgical restrictions (1)

MET = metabolic equivalent

Table 4:

Multivariable cox regression analysis for the association of increasing cardiopulmonary fitness by one MET and subsequent cardiovascular disease in survivors with lower cardiopulmonary fitness levels and survivors with higher cardiopulmonary fitness levels.

	Subsequent cardiovascular disease
	Low cardiopulmonary fitness group		Normal cardiopulmonary fitness group
	HR	95% CI	HR	95% CI
Cardiopulmonary fitness
Increase of one MET	0.78	0.65, 0.96	0.90	0.70, 1.15
Cardiovascular disease at baseline
Yes	2.4	1.41, 4.1	1.34	0.51, 3.60
No	1.0	1.0	1.0	1.0
Obesity at baseline
Yes	0.71	0.39, 1.31	0.99	0.39, 2.52
No	1.0	1.0	1.0	1.0
Impaired glucose metabolism at baseline
Yes	1.81	0.90, 3.65	0.62	0.08, 4.88
No	1.0	1.0	1.0	1.0
Hypertension at baseline
Yes	0.81	0.44, 1.48	0.46	0.15, 1.41
No	1.0	1.0	1.0	1.0
Anthracyclines
Yes	1.78	1.05, 3.02	4.23	1.53, 11.644
No	1.0	1.0	1.0	1.0
Chest radiation
Yes	1.25	0.76, 2.08	1.88	0.84, 4.19
No	1.0	1.0	1.0	1.0

Categories were made from the distribution of peak METs the survivors achieved on the CPET, separated by age and sex

Low/Below average (< 50%), above average/high (≥ 50%)

Additionally adjusted for age, age at diagnosis, meeting 2018 CDC physical activity guidelines, smoking and gender

237 survivors were missing their CPET and could not be classified as low or below average. The reasons included acute injury (1), cardiovascular (112), diabetes (26), gastrointestinal (1), kidney disease (16), liver failure (2) musculoskeletal impairment (40), neurological impairment (10), paralysis (4), pulmonary impairment (15), cerebrovascular impairment (9), recent surgery (1). MET = Metabolic equivalent; HR = hazard ratio; CI = confidence interval

Death from a cardiovascular event

When evaluating death from a cardiovascular event, the median follow-up time for survivors was 2.4 years (IQR: 1.3, 6.2 years). There were 8 (0.03%) deaths from serious cardiac disease events among all participants. The average peak METs achieved on the CPET in those who died from a cardiovascular event was 5.9 ± 2.2. Survivors of childhood cancer who did not die from a cardiovascular event achieved 7.6 ± 2.5 METs on their CPET, on average.

Discussion

This comprehensive evaluation of a large cohort of childhood cancer survivors demonstrates that cardiopulmonary fitness is associated with a reduced risk for future moderate to severe cardiac events, independent of past cardiac disease. Individuals who died from a cardiovascular cause during the study follow-up had lower baseline cardiorespiratory fitness (by ~2 METs) than those who did not die from a cardiovascular event. Thus, improving/increasing cardiopulmonary fitness may reduce risk of cardiovascular-related death in childhood cancer survivors, especially among those with low cardiopulmonary fitness. This study is unique because exposure data included detailed diagnosis, treatment, demographic, lifestyle, and prevalent chronic disease variables, and objectively measured cardiopulmonary fitness. These findings are important because poor cardiopulmonary fitness is actionable, even among people with clinically apparent cardiovascular-related disease. Addressing cardiopulmonary fitness, especially in survivors with already low cardiopulmonary fitness, may have potential to mitigate the risk of cardiovascular disease.

Our study is the first to examine associations between direct cardiopulmonary fitness and incident cardiac events in a large cohort of childhood cancer survivors with a broad spectrum of primary cancer diagnoses. We observed similar associations between objectively measured cardiopulmonary fitness and any incident cardiovascular disease to those reported in the general population.(30–33) While using self-reported PA is convenient in large cohorts, it becomes problematic when used as a surrogate for cardiopulmonary fitness. Cardiopulmonary fitness is influenced not only by behavior, but also biology, environment, and genetics.(34–36) Self-reported PA does not account for physiologic and/or genetic determinants of cardiopulmonary fitness, nor environmental factors such as toxicities from childhood cancer treatment. Additionally, measurement error, seasonal variation, response and recall bias are common when persons are asked about their physical activity, further impacting accuracy when used as a substitute for cardiopulmonary fitness.(37–39) Correlation coefficients between the commonly used PA questionnaires, such as the Stanford 7-day, Stanford Usual, Godin Leisure, and the International Physical Activity Questionnaire and objectively measured cardiopulmonary fitness range from 0.07 to 0.63, with the average coefficient being weak (r=0.29),(40, 41) indicating that accurate screening for low cardiopulmonary fitness should use objective measures.

Our findings are particularly relevant for childhood cancer survivors, whose treatment exposures confer higher risks of obesity, hypertension, diabetes, and/or early onset of cardiovascular disease.(2, 42, 43) These results also indicate that the risk of future cardiac events was most impacted by cardiopulmonary fitness among those survivors with a low cardiopulmonary fitness level (below the age and sex matched median). Lower cardiopulmonary fitness is common in childhood cancer survivors,(42) and our study identifies potentially effective intervention opportunities to improve a modifiable risk factor in a vulnerable population. However, due to their treatment, adult survivors of childhood cancer have a lower physiological reserve compared to cancer-free controls,(44) meaning their cardiopulmonary system does not adapt to repeated bouts of perturbation as effectively as their peers. While survivors do benefit from adopting higher PA levels,(45) the benefits from PA may be less when performing the same amount of work as their cancer-free peers.(46) Therefore, more research is needed to determine the most effective intervention strategies to improve cardiopulmonary fitness and reduce risk of cardiovascular disease in this group.

While our study did not have enough events to perform a multivariable analysis of the association between caridopulmonary fitness and specific cardiac causes of deaths, our results are consistent with previous literature. In cancer-free men, low cardiopulmonary fitness (peak VO₂ of <27.6 ml·kg¹·min⁻¹) increased the risk of cardiac related mortality 3.09 (95% CI: 1.10, 9.56) times compared to men who had a peak VO₂ of ≥27.6 ml·kg¹·min⁻¹.(47) Additionally, a 1-MET increase was associated with a 16.1% reduction of cardiac related mortality in both men and women.(48) Our study showed a similar difference in the cardiopulmonary fitness between survivors who died from a cardiac event and those who did not. Survivors who died from a cardiac event achieved an average of peak MET of 5.9 on the CPET (equivalent to 20.7 ml·kg¹·min⁻¹) compared to an average of peak MET of 7.6 (equivalent to 26.6 ml·kg¹·min⁻¹) in survivors who did not die from a cardiac event. While our study had fewer events and shorter follow-up, our results indicate that low cardiopulmonary fitness could be contributing to cardiac mortality in adult survivors of childhood cancer.

When examining associations between cardiopulmonary fitness and cardiovascular disease it is important to include survivors with existing cardiovascular disease. In doing so, our study increases understanding of the effects of cardiopulmonary fitness on childhood cancer survivors’ health. Survivors tend to develop cardiac conditions earlier in life compared to the general population.(49) Some of the increased risk of cardiac events is unavoidable due to cardiotoxic therapies, or the childhood cancer itself. Additionally, having a preexisting cardiac condition increases the risk of having a future cardiac event compared to those who do not have preexisting conditions.(50–52) Our results demonstrate that increased levels of cardiopulmonary fitness are protective against a future cardiac event, independent of prior cardiovascular disease in survivors of childhood cancer. All survivors, even those who have already developed cardiac disease, can benefit from developing healthier lifestyle habits to increase their cardiopulmonary fitness.

Our results should be interpreted in the context of limitations. First, although participation rates were high, not all eligible survivors were evaluated. Non-participants were significantly older at baseline, younger at treatment; both of these are risk factors for poor cardiopulmonary fitness, as well as early mortality.(53, 54) Thus, our cohort may have been healthier (better cardiopulmonary fitness than the older survivor population), and thus at lower risk for a subsequent event and/or cardiac-related death. This would underestimate the associations between cardiopulmonary fitness and cardiac outcomes, biasing our results towards the null. Second, as indicated above, median follow-up in our study was only 2.5 years. Survivors who will subsequently develop overt cardiovascular disease may have been censored too early, which would further underestimate our associations. Third, our participants’ cardiopulmonary fitness trajectory prior to evaluation is unknown. It is possible that their poor cardiopulmonary fitness predates the onset of the cardiovascular disease, and thus, is a potential risk factor for their baseline cardiovascular disease. Our models would then be adjusting for a causal intermediate (baseline cardiac event) which would likely bias our results towards the null. This is a possible reason for the lack of association between cardiopulmonary fitness and serious cardiac events. Future studies should focus on gathering both cardiovascular disease status and cardiopulmonary fitness earlier in survivorship to examine the temporal association between cardiovascular disease and cardiopulmonary fitness as this population transitions to long term survivorship. Fourth, we did not have enough power to perform a multivariate statistical analysis to evaluate the association between cardiopulmonary fitness and death from cardiovascular disease. While our results are consistent with past literature, confounders were not included. Future studies with longer follow-up time and more events are needed to compare results that are independent of known confounding variables. Finally, our sample was treated at one institution (SJCRH) between 1962 and 2012; our data may not be generalizable to all childhood cancer survivors, especially those treated more recently or at other institutions with different protocols.

Despite these potential limitations, this study provides important information about the negative consequences of poor cardiopulmonary fitness on the risk of a subsequent cardiovascular event in childhood cancer survivors. Higher cardiopulmonary fitness is protective against future cardiac events, and exercise is safe and effective in improving cardiopulmonary fitness. Early identification of survivors with or at risk for low cardiopulmonary fitness should be optimized to provide opportunities for risk mitigation through promotion of a lifestyle with regular physical activity.