The 6-minute walk test is a good predictor of cardiorespiratory fitness in childhood cancer survivors when access to comprehensive testing is limited

Abstract

Cardiovascular disease is up to 10 times more likely among childhood cancer survivors compared with siblings. Low cardiorespiratory fitness is a modifiable risk-factor for cardiovascular diseases. Yet, cardiorespiratory fitness is not routinely screened in pediatric oncology, and healthy VO_2max cut-points are unavailable. We aimed to predict cardiorespiratory fitness by developing a simple algorithm and establish cut-points identifying survivors’ cardiovascular fitness health-risk zones. We recruited 262 childhood cancer survivors (8–18 years old, ≥1-year post-treatment). Participants completed gold-standard cardiorespiratory fitness assessment (Cardiopulmonary Exercise Test (CPET; VO_2max)) and 6-minute walk test (6MWT). Associations with VO_2max were included in a linear regression algorithm to predict VO_2max, which was then cross-validated. We used Bland-Altman’s limits of agreement and Receiver Operating Characteristic curves using FITNESSGRAM’s ‘Healthy Fitness Zones’ to identify cut-points for adequate cardiorespiratory fitness. 199 participants (aged 13·7±2·7 years, 8·5±3·5 years posttreatment) were included. We found a strong positive correlation between VO_2max and 6MWT distance (r=0·61, r²=0·37, p<0·001). Our regression algorithm included 6MWT distance, waist-to-height ratio, age and sex to predict VO_2max (r=0·79, r²=0·62, p<0·001). Forty percent of predicted VO_2max values were within ±3 ml/kg/min of measured VO_2max. The cut-point for FITNESSGRAM’s ‘health-risk’ fitness zone was 39·8 ml/kg/min (males: AUC=0·88), and 33·5 ml/kg/min (females: AUC=0·82). We present an algorithm to reasonably predict cardiorespiratory fitness for childhood cancer survivors, using inexpensive measures. This algorithm has useful clinical application, particularly when CPET is unavailable. Our algorithm has the potential to assist clinicians to identify survivors below the cut-points with increased cardiovascular disease-risk, to monitor and refer for tailored interventions with exercise specialists.

Introduction

Due to improvements in treatment protocols and supportive care in high-income countries, an increasing number of children are surviving cancer well into adulthood.[1] However, the cost of cure is high, with most survivors experiencing medical late-effects.[2] Young survivors face medical complications normally experienced by older adults, with survivors up to 10 times more likely to develop cardiovascular diseases (CVD) compared with their siblings before they turn 30.[3] CVD are generally related to cardio-toxic treatment including anthracyclines and chest radiation,[4] with survivors seven times more likely to die from CVD compared with age-matched peers.[5]

With high CVD morbidity, cardio-oncology has become increasingly important.[4] The ‘multiple-hit hypothesis’ was developed among adult breast cancer patients, highlighting the increased risk of developing CVD from direct (i.e. radiotherapy, chemotherapy) and indirect effects (i.e. lifestyle, body composition changes).[6] The hypothesis suggests that a child diagnosed with acute lymphoblastic leukaemia treated with anthracycline, methotrexate and radiotherapy, and not physically active is likely to have compounded CVD-risk.

Survivors are known to have low physical activity levels, with around three-quarters not achieving recommended guidelines.[7] Although it is reported that late-effects are discussed in up to 85% of consultations with adolescent survivors,[8] as few as 20% of oncologists provide lifestyle guidance in survivorship.[9] Chronic low physical activity, coupled with accumulating treatment late-effects can contribute to reduced cardiorespiratory fitness.[10] As survivors age, the fitness gap between survivors and the general population widens.[10]

There is a strong relationship between low fitness, CVD and mortality-risk.[11, 12] The gold-standard fitness measurement is a cardiopulmonary exercise test (CPET), conducted on a treadmill or cycle ergometer to determine aerobic capacity (VO_2max), which is the maximum amount of oxygen consumed during exercise.[13] Incremental stages are completed until maximal exertion.[13] Fitness can be predicted using indirect techniques including the 6-minute walk test (6MWT), which have demonstrated validity in healthy children,[14] do not require maximal exertion or expensive equipment, and are endorsed by the American Thoracic Society.[15]

There are currently no pediatric oncology guidelines recommending frequency and specificity of fitness assessments despite survivors’ known cardiovascular-risks.[16] This is important because cardiotoxicity is not always clinically symptomatic, therefore the impact can be underestimated. Recently, Clinical Oncology Society of Australia reaffirmed the attempt to promote physical activity among cancer survivors.[17] However, linking survivors directly to exercise professionals remains challenging [18] and CPET assessments requiring trained staff are not routine in most hospitals. Indirect measures including the 6MWT may be less precise, but are cost-effective, quick and less exertive, and may provide guidance on survivors’ cardiorespiratory fitness.[19] Submaximal fitness testing has displayed an acceptable, reliable and valid measurement in pediatric obesity [20] and an acceptable measurement in pulmonary hypertension populations.[21] However, the link between VO_2max and 6MWT performance has yet to be investigated among childhood cancer survivors.[15] Further, there are no cut-points to determine healthy fitness levels in childhood cancer survivors to flag clinicians to refer high-risk patients for intervention with an exercise specialist.

The primary aim was to develop a simple algorithm to predict cardiorespiratory fitness (VO_2max) using 6MWT performance among childhood cancer survivors. The secondary aim was to establish cut-points to determine predicted fitness levels (_predVO_2max) that differentiate between healthy and risk zones.

Methods

Participants

This cross-sectional bi-national study recruited a convenience sample of participants from the Kids Cancer Centre, Sydney Children’s Hospital, Australia and St. Jude Children’s Research Hospital, United States. Inclusion criteria were 8–18 years old children who had completed treatment for any malignancy ≥1 year prior, able to communicate in English and willingness to provide informed consent. We excluded potential participants with any medical contraindication to CPET (e.g. unstable heart condition, uncontrolled asthma or cognitive impairment),[13] at the recommendation of the treating consultant, or if the survivor was pregnant.

Procedures

We recruited participants between July 2017 and September 2018 through regular follow-up clinic visits. The researcher and nursing staff identified potential participants through clinical lists. Final eligibility and contraindications were confirmed by the treating consultant. We contacted parents of eligible participants by telephone prior to their clinic visit. We collected informed consent on the day of their clinic appointment prior to commencing any study procedures. The study was approved by Sydney Children’s Hospital Network Human Research Ethics Committee (LNR/16/SCHN/403) and St. Jude Children’s Research Hospital Institutional Review Board (#00000029, FWA00004775).

Physical assessment

Cardiorespiratory fitness

We assessed cardiorespiratory fitness using two exercise assessments, the 6MWT (submaximal intensity) and CPET (maximal intensity), separated by ≥30 minutes rest. For both exercise assessments, we evaluated exercise tolerance and physiological exercise response by monitoring heart rate, oxygen saturation (SpO₂) and self-rated perceived exertion (RPE).

The 6MWT was supervised by an accredited exercise physiologist, according to the American Thoracic Society recommendations.[15] Participants were encouraged every minute to walk as far as they could without running, taking rest breaks if required. We measured the total distance walked, heart rate and oxygen saturation using a portable pulse oximeter (Nonin Wristox 3150SK, Plymouth, USA). We collected RPE at the cessation of the assessment using Borg’s 1–10 RPE scale in Australia (or Borg’s 6–20 scale in the United States, which was converted to 1–10).[22] The 6MWT has demonstrated high test-retest reliability with an intraclass correlation coefficient of 0.99 in paediatric obese youth [20] and 0.98 in children and adolescents with an acquired brain injury.[23]

Participants performed spirometry (Forced Expiratory Volume [FEV₁], Forced Vital Capacity [FVC] and FEV₁/FVC ratio) to assess respiratory function and participant safety, according to the ATS/ERS guidelines.[24] The spirometer was connected to the metabolic cart using respiratory analysis software BreezeSuite™ (MGC Diagnostics, St Paul, USA).

VO_2max was assessed during a gold-standard CPET on a treadmill ergometer (Trackmaster TMX425CP, Newton, USA) using the Bruce Protocol.[13] Participants had their heart rate and cardiac response continuously monitored using 12-lead Electrocardiogram (ECG; Xscribe X12+, Mortara, Milwaukee, USA) before, during, and for 10 minutes after the medically-supervised test. Ventilatory gas exchange data were collected using breath-bybreath analysis, averaged every 10 seconds, to calculate ventilation, oxygen uptake (VO₂) and carbon dioxide output (VCO2), analysed using a metabolic cart (Ultima Cardio2, MGC Diagnostics, St Paul, USA). The metabolic cart and gas analysers were calibrated prior to each assessment to room air temperature, humidity, pressure and manufacture calibrated gases (O₂ and CO₂), as well as the pneumotach being calibrated to 10 samples using a Carefusion 3L syringe. SpO₂ was recorded using a pulse oximeter. The Bruce Protocol includes incremental 3-minute stages, commencing at 2·7 km/h and 10% gradient, increasing by 1·3 km/h and 2% gradient every stage. An exercise physiologist encouraged participants to walk or run for as long as they could. The test was terminated when the participant was unable to maintain the required output or if they requested to stop. For the test to be considered maximal, two of the following criteria needed to be achieved: heart rate ≥85% of age-predicted maximum (based on a maximum heart rate 200 [25]), respiratory exchange ratio (RER=VCO₂/VO₂) ≥1·0, or RPE ≥8/10.[26–28] Stopping criteria included abnormal ECG (e.g. ±2mm ST segment deviation), symptom development (e.g. dizziness, angina), or desaturation (SpO₂<88%).[13] VO_2max is presented as absolute (l/min) and relative to body weight (ml/kg/min).[13]

Clinical and demographic factors

We determined participant’s cancer history including diagnosis, treatments received and dates of diagnosis and treatment completion from hospital medical records. We classified treatment intensity according to the Intensity of Treatment Rating (ITR-3), based on tumour type, tumour stage/risk, and treatment modalities received.[29]

Data Management and Statistical Analyses

For the primary outcome, we calculated the association between the 6MWT (total distance) and CPET (VO_2max) using Pearson’s correlation coefficient, after testing for a linear relationship and normal distribution. We conducted a sensitivity analysis using Fisher’s r-to-z transformation, to identify the impact of institution, sex, walking distance, height and age (dichotomised by the median value) on the correlation between the two assessments. VO_2max (in children [30] and adolescents [31]) and 6MWT distance [32] were converted to age- and sex-specific percentiles. We conducted a linear regression to predict VO_2max, using Pearson’s correlation to identify the relationship of VO_2max with 6MWT distance and anthropometric variables. Variables with p<0·2 were entered into a multivariate linear regression. We used five-time hold-out cross-validation method to evaluate the accuracy of the calculated algorithm when 20% of the cases were randomly removed and compared against the other 80%.[33] Bland-Altman limits of agreement was calculated to compare the measured VO_2max with predicted VO_2max (_predVO_2max). Receiver operating characteristic (ROC) curves examined the diagnostic performance of VO_2max against the binary variable of being above/below the ‘health-risk fitness zone’ using the age/sex stratified FITNESSGRAM Healthy Fitness Zone (HFZ; The Cooper Institute [34]). FITNESSGRAM is a battery of physical assessments conducted in ≥10 million school-aged children, stratifying students into the ‘healthy fitness zone’, ‘needing improvement’ or ‘health-risk zone’. These zones are based on a 20-meter shuttle aerobic test to estimate VO2max from the number of performed laps using a prediction equation.[35] Using the ROC curve, we calculated sensitivity and specificity, generating cut-points for VO_2max using Youden’s Index (J) to identify the prevalence of low fitness. Sensitivity and specificity >80% were considered high and acceptable between 5080%.[36] Statistical analyses were performed using SPSS Version 25 (IBM, Armonk, NY).

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Results

Two-hundred and sixty-two participants were included. In Australia, 64 eligible participants were invited, 46 of whom were recruited (recruitment rate=72%). Four declined, eight had scheduling conflicts, and six parents did not return the voice messages. In the US, 221 eligible participants were invited (from the St. Jude Lifetime Cohort Study [37]), 212 of whom agreed to participate and had valid data (96%). Two declined and seven had scheduling conflicts. Participants were 13·7±2·7 years old (range: 8–18 years) and 8·6±3·6 years (range: 1–16 years) post-treatment completion (Table 1). The most common malignancies were acute lymphoblastic leukaemia (32%), retinoblastoma (15%), and central nervous system tumours (12%). Sixty participants did not achieve the criteria for maximality and were removed from the analysis, leaving 202 participants for the final analysis. Of those included 90% achieved ≥85% maximum heart rate, 81% ≥1.0 RER and 54% ≥8 RPE (100% of included participants achieved ≥2 criteria). There were no adverse events at either site. Participants from St. Jude were older at inclusion and diagnosis, and farther post-treatment than Australian participants (all p<0·05), with no difference in cancer treatment intensity (p=0·07). Australian and US participants were representative of the eligible survivor population from each institution as they did not differ with their corresponding non-respondents by age (p=0·29 and p=0·17) and sex (p=0·11 and p=0·91), respectively. Participants not achieving maximal CPET were younger during the study (p<0·001) and at diagnosis (p<0·014), completed treatment more recently (p=0·05), and had a lower distance (p<0·001) and maximum heart rate (p=0·02) during the 6MWT.

Table 1.

Baseline clinical characteristics of participants.

	Total sample (n=199)
	Mean (range)	S.D.
Age at study (years)	13·7 (8–18)	2·7
Age at diagnosis (years)	3·3 (0–15)	3·2
Years since treatment completion	8·5 (1–16)	3·6
	n	%
Sex, male	107	54
Institution
Sydney Children’s Hospital	46	23
St. Jude Children’s Research Hospital	153	77
Cancer type
Acute lymphoblastic leukaemia	68	34
Retinoblastoma	27	14
Central nervous system tumours	21	11
Wilms’ tumour	15	8
Hodgkin’s and non-Hodgkin’s lymphoma	16	8
Acute myeloid leukaemia	12	6
Neuroblastoma	10	5
Other (e.g. osteosarcoma, rhabdomyosarcoma)	30	15
Treatment
Surgery	184	93
Chemotherapy	177	89
Radiotherapy	53	27
Bone marrow transplant	24	12
Treatment intensity (ITR-3)
Level 1/2	134	67
Level 3/4	65	33

ITR = Intensity of Treatment Rating.

Sensitivity analysis

We found no difference in the correlation between CPET and 6MWT between institution (Sydney vs. St. Jude, z=1·16, p=0·25), sex (male vs. female, z=0·94, p=0·35), and median values of 6MWT performance (≤564m vs. >564m, z=−0·91, p=0·36), height (≤160cm vs. >160cm, z=0·64, p=0·52), and age (<14 vs. ≥14 years old, z=−0·76, p=0·45).

Relationship between VO_2max and the 6-minute walk test distance

Mean VO_2max was 36·4±10·2 ml/kg/min (Table 2; range: 14·7–67·6), equating to the 29^th±30 percentile (range: 1–99). Mean 6MWT distance was 571±106m (range: 308–860), equating to the 28^th±32 percentile (range: 1–99). There was a moderate positive association between maximum heart rate during the CPET and at the conclusion of the 6MWT (r=0.43, p<0.001).

Table 2.

Data from the final sample completing maximal cardiopulmonary exercise test, 6-minute walk test, body composition analysis and spirometry.

	Total sample (n=199)
	Mean (range)	S.D.
Cardiopulmonary exercise test
VO2max (relative; ml/kg/min)	36·4 (14·7–67·6)	10·2
VO2max (absolute; l/min)	2·1 (0·7–4·5)	0·8
Percentile	28·8 (1–99)	30·4
FITNESSGRAM Zone
‘Healthy fitness’ zone, n (%)	47	24%
‘Health risk’/’Needs improvement’ zone, n (%)	152	76%
Time to completion (minutes)	10:19 (5:00–16:45)	2:59
Max heart rate (bpm)	188·6 (141–219)	12·8
Max heart rate (% of predicted max)	94·3 (75–110)	6·4
Max RER	1·08 (0·88–1·43)	0·09
SpO2	96·9 (87–100)	2·3
RPE	7·8 (2–10)	2·1
6-minute walk test
Distance (m)	553 (308–860)	106·1
Percentile	28·2 (1–99)	32·1
Max heart rate	140·0 (84–191)	21·5
SpO2	98·0 (80–100)	1·8
RPE	1·7 (0–10)	2·2
Body composition
Body mass index (kg/m2)	22·7 (12·9–60·2)	6·2
Waist-to-height ratio	0·47 (0·35–0·93)	0·08
Body fat percentage (%)	26·4 (3·0–79·6)	12·6
Lung function (spirometry)
FVC (L)	3·3 (1·2–5·9)	1·1
FVC predicted (%)	96·4 (53–129)	12·8
FEV1 (L)	2·9 (1·1–5·0)	0·9
FEV1 predicted (%)	95·9 (47–126)	12·8

VO_2max = cardiorespiratory fitness, RER = respiratory exchange ratio, SpO₂ = oxygen saturation, RPE = Borg’s 1–10 Rate of Perceived Exertion scale, FVC = forced vital capacity, FEV₁ = forced expiratory volume within one second, bpm = beats per minute. Percentiles for VO_2max (children [30] and adolescents [31]) and 6MWT distance [32].

There was a moderate positive correlation between CPET (VO_2max; ml/kg/min) and 6MWT (m) (n=203, r=0·57, r²=0·31, p<0·001), which strengthened when four outliers were removed (Figure 1, n=199, r=0·61, r²=0·37, p<0·001; three participants with VO_2max >90^th percentile and 6MWT <5^th percentile, and one participant with a body mass index (BMI; kg/m²) >60).

Figure 1.

Relationship between VO_2max (ml/kg/min) and 6MWT distance (m) among 199 childhood cancer survivors after the removal of four outliers.

Predicting cardiorespiratory fitness in childhood cancer survivors

Sex (r=−0·37, p<0·001), BMI (r=−0·55, p<0·001), waist-to-height ratio (r=−0·57, p<0·001) and age (r=−0·09, p=0·24) significantly correlated with VO_2max (Supplementary Table1, online only). Waist-to-height ratio had a stronger relationship with VO_2max than BMI in our study, and has demonstrated a stronger link with CVD risk in children [38], hence was the preferred body composition measurement. The algorithm calculated from the multivariate analysis including 6MWT distance, waist-to-height ratio, age and sex significantly predicted VO_2max (_predVO_2max; r=0·79, r²=0·62, p<0·001). When cross-validating the algorithm by comparing _predVO_2max to measured VO_2max, the algorithm remained significant when validating the training set (80% sample; r=0·75, r²=0·56, p<0·001) against the testing set (20% sample; r=0·79, r²=0·62. p<0·001). The algorithm to predict VO_2max was:

$$\begin{array}{l}predVO2max(ml/kg/min)\\ =42\cdot 4+(6MWTdistancex0\cdot 042)\\ +(waisttoheightratiox-56\cdot 2)+(agex-0\cdot 05)\\ +(sex[male=0,female=1]x-5\cdot 8)\end{array}$$

Agreement between measured and predicted VO_2max

Using the Bland-Altman plot to compare VO_2max measured from CPET with _predVO_2max, the mean difference between measurements was −0·01±6·77 ml/kg/min. The upper and lower limits were 13·26 and −13·28 ml/kg/min, respectively. Ninety-four percent of data-points were within ±1·96 standard deviations of the mean difference (Figure 2). Forty percent of predVO_2max were within ±3 ml/kg/min of measured VO_2max. When testing the algorithm on the 63 participants who completed a submaximal or symptom-limited CPET, the mean _predVO_2max was 31·8±7·1 ml/kg/min, which was higher than when measured using CPET (25·7±6·4 ml/kg/min, p<0·001).

Figure 2.

Bland-Altman plot of the difference and mean of measured VO_2max (Cardiopulmonary Exercise Test) and _predVO_2max equation (using the 6-minute walk test distance, waist-to-height ratio, age and sex).

Identifying predicted VO_2max cut-points for the ‘health-risk’ zones

The total area under the ROC curve was 0·88 in males (Figure 3; p<0·001, 95% confidence interval=0·81–0·94) and 0·82 in females (p<0·001, 95% confidence interval=0·74–0·91), indicating good accuracy for our diagnostic test. The _predVO_2max cut-point for fitness above the ‘health-risk’ zone was 39·8 ml/kg/min in males (sensitivity=82%, specificity=79%, J=1·61) and 33·5 ml/kg/min in females (sensitivity=80%, specificity=69%, J=1·49).

Figure 3.

Receiver operating characteristic curve identifying the sensitivity and specificity of the pre_dVO_2max equation to determine whether fitness is above or below ‘health-risk’ zones using age/sex FITNESSGRAM Healthy Fitness Zone cut-points [34] in A) male and B) female participants. AUC = area under curve, CI = confidence interval.

Discussion

Gold-standard cardiorespiratory fitness assessments are not routinely conducted after pediatric cancer globally. Our study aimed to validate a simple clinically accessible algorithm to predict fitness. Our algorithm estimating cardiorespiratory fitness using 6MWT distance, waist-to-height ratio, age and sex was sensitive in survivors of childhood cancer. Using our algorithm, which could be both cost- and time-effective, we calculated cut-points to identify childhood cancer survivors with low cardiorespiratory fitness, potentially clinically useful when gold-standard facilities are unavailable. Our ROC curves calculated the _predVO_2max cut-point for FITNESSGRAM’s ‘health-risk’ fitness to determine accurately the prevalence of low fitness across survivors was 39·8 ml/kg/min in males and 33·5 ml/kg/min in females. Using an example, a 17-year old male survivor walking 481m in the 6MWT, with a waist-toheight ratio of 0·49 resulting in a _predVO_2max of 34·2 ml/kg/min is below the male cut-point, falling within the ‘health-risk’ zone. This _predVO_2max would flag to clinicians that he should be monitored, provided with education on how to safely increase physical activity levels, and supported to improve fitness to reduce the risk of developing CVD.

Cardiorespiratory fitness is a crucial physiological indicator, with each 3·5 ml/kg/min decrease in VO_2max equating to a 21% increase in heart failure [39] and 13% increase in mortality-risk from CVD in adults,[40] whilst in children low fitness increased CVD risk factors.[41] Our cohort had a mean measured VO_2max around the 29^th percentile, displaying similar deficits to other childhood cancer survivors of similar age (VO_2max 9–15% lower than healthy controls in these studies).[42–44] It is important to intervene during childhood as the fitness deficit (and subsequent CVD-risk and impact on quality of life) between survivors and those without history of cancer accelerates when entering adulthood (VO_2max 20–23% lower among survivors).[45–47]

We suggest that the 6MWT, a standardised clinical tool, represents a simple, cost-effective estimate of fitness when combined with routine anthropometric measures among childhood cancer survivors. This finding is important considering growing pressure on clinics with increasing numbers of survivors that do not have capacity to conduct thorough fitness assessments on every patient.[1] We must acknowledge that our algorithm appeared to be more sensitive for survivors with low-moderate fitness, and less accurate among those with higher fitness. Our findings regarding the concordance between CPET and 6MWT are echoed in studies with healthy children [14] and children with pulmonary hypertension.[21] Our cut-point to predict CVD-risk based on cardiorespiratory fitness was similar to previous research in healthy male (39·8 vs. 42·0 ml/kg/min) and female children (33·5 vs. 35·0 ml/kg/min).[12] In contrast, a validation study of adult cancer survivors reported the 6MWT in isolation underestimates VO_2max and should not replace CPET, reporting mean correlation of r=0·72 across four prediction equations.[48] Our model is refined by taking into account waist-toheight ratio, which has shown to be predictive of metabolic-risk among childhood cancer survivors.[49]

Childhood cancer survivors have an increasing late-effects burden, particularly affecting the cardiovascular system.[50] A proactive strategy would be to assess cardiorespiratory fitness at least annually after treatment. Even if clinics have CPET access, assessments can cost approximately AUD$300 [51] or USD$160/patient [52] for each physician-supervised assessment (one repeat assessment/100 patients could cost approximately AUD$60,000 or USD$32,000). These costs would be substantially lower for the 6MWT which requires reduced staff time and equipment. With a growing population of childhood cancer survivors requiring life-time surveillance and accumulation of late-effects as they age, this increasing cost would be unsustainable long-term, hence the need for cost-effective assessments. Further, calculating _predVO_2max may be useful for patients who cannot reach maximal exertion using CPET, which occurred in 60 (23%) participants in our cohort. A submaximal CPET might under-report true VO_2max, highlighting the usefulness of an indirect estimation for these survivors.

With a simple and time-effective algorithm, estimating _predVO_2max can be used as a “teachable moment” for unfit survivors to be motivated to increase their physical activity and fitness.[9] Teachable moments can have a powerful influence on behaviour change.[53] Educating parents and encouraging survivors to be physically active can increase fitness levels and may assist in reducing the risk of developing cardio-metabolic late-effects. Pre-clinical studies and clinical trials indicate that interventions promoting aerobic physical activity (e.g. brisk walking, swimming, jogging) can increase cardiorespiratory fitness, which has been demonstrated in CCS versus non-exercising controls (standard mean difference = 0.69), [54] and may provide cardio-protective benefits by attenuating treatment-induced effects.[54, 55]

This study was limited by the heterogeneous sample (i.e. varied cancer, treatment and pubertal stages (which we did not assess)), meaning direct comparisons across clinical characteristics were not possible. Our algorithm to predict VO_2max appeared weaker with higher measured VO_2max. Although this is a weakness of the algorithm, it is more clinically relevant to identify survivors with low to medium fitness who are at increased need for support, than identifying survivors with high fitness. Sixty participants performed submaximally and were removed from the algorithm, which may introduce bias. This study was not powered to investigate the effect of specific treatment types and doses in the algorithm, whilst age-specific VO_2max cut-points such as in the general population,[56] would provide more precision in identifying CVD-risk and should be investigated in the future, as we could only calculate sex-specific cut-points. The rest between tests (≥30 minutes) was adequate based on heart rate returning to resting values, but longer rest (or testing the following day) may provide complete physical recovery, which occurred in a small number of St. Jude participants due to clinic scheduling. We did not provide a familiarization attempt, nor did we assess validity and reliability of both aerobic exercise tests, which would strengthen the clinical utility of the 6MWT. Strengths of the study included a large sample from two diverse populations in Australia and the United States (which were representative of the general survivor population at each institution), which may increase the generalisability of the findings. Using the gold-standard CPET (applying the Bruce Protocol), our study is the first to develop cardiorespiratory fitness cut-points to identify survivors with modifiable CVD-risk.

Conclusion

We found that cardiorespiratory fitness can be predicted reasonably with inexpensive assessment and routine clinical measures in survivors of childhood cancer. Our algorithm combined the 6MWT, waist-to-height ratio, age and sex to provide a simple calculation to predict cardiorespiratory fitness, identifying screening _predVO_2max cut-points for low fitness of 39·8 ml/kg/min in males and 33·5 ml/kg/min in females. This simple method has the potential to change practice by identifying an important modifiable CVD-risk factor in large numbers of survivors earlier who otherwise may be overlooked until symptoms develop. Although CPET is more precise and should not be replaced if facilities and resources are available, our research can assist clinicians to refer survivors for individualised interventions with an exercise specialist by estimating fitness where CPET is unavailable or in survivors with otherwise symptom-limited CPETs, determining easily whether fitness is improving or regressing over time. Future studies should be conducted to confirm the findings of this study.

Novelty & Impact Statement:

This cross-sectional study of 199 childhood cancer survivors identified that cardiorespiratory fitness can be predicted using the 6-minute walk test and waist-to-height ratio. Predicted VO_2max <39.8 ml/kg/min (males) and <33.5 ml/kg/min (females) was associated with increased cardiovascular disease risk. Childhood cancer survivors experience high cardiovascular disease incidence. Early identification of modifiable risk factors including predicted cardiorespiratory fitness can be used in clinics when comprehensive assessment is unavailable to motivate lifestyle changes and mitigate risk.

Table of abbreviations

SCHN

Sydney Children’s Hospital Network

CPET

Cardiopulmonary Exercise Test

6MWT

6-minute walk test

CVD

Cardiovascular disease

SpO₂

Oxygen saturation

RPE

Rate of perceived exertion

FEV₁

Forced Expiratory Volume

FEV

Forced Vital Capacity

ECG

Electrocardiogram

RER

Respiratory exchange ratio

ITR-3

Intensity of Treatment Rating

ROC

Receiver operating characteristic

HFZ

Healthy Fitness Zone

BMI

Body mass index