N.J. Krnavek1, S. Ajasin2, E.C. Arreola2, M. Zahiri1, M. Noun1, P.J. Lupo2, B. Najafi1, M.M. Gramatges2
1. Baylor College of Medicine, Department of Surgery, Division of Vascular Surgery and Endovascular Therapy, USA; 2. Baylor College of Medicine, Department of Pediatrics, Section of Hematology and Oncology, USA
Corresponding author: Maria Monica Gramatges, Baylor College of Medicine, Department of Pediatrics, Feigin Center, 1102 Bates St, Suite 1200, Houston, Texas, 77030, (p) 832-824-4678; (f) 832-825-4651, email@example.com
J Frailty Aging 2020;in press
Published online December 23, 2020, http://dx.doi.org/10.14283/jfa.2020.71
Background: Survivors of childhood cancer (CCS) are at risk for early aging and frailty. Frailty in CCS has been assessed with established clinical criteria, a time-intensive approach requiring specialized training. There is an unmet need for cost-effective, rapid methods for assessing frailty in at-risk adolescent and young adult (AYA) CCS, which are scalable to large populations. Objectives: To validate a sensor-based frailty assessment tool in AYA CCS, compare frailty status between CCS and controls, and assess the correlation between frailty and number of CCS comorbidities. Design, Setting, and Participants: Mean frailty index (MFI) was assessed by a frailty wrist sensor in 32 AYA CCS who were ≥1 year off therapy and in remission. Results were compared with 32 AYA controls without cancer or chronic disease. Measurements: Frailty assessments with and without a simultaneous cognitive task were performed to obtain MFI. Results were compared between cases and controls using a Student t test, and the number of pre-frail/frail subjects by Chi Square test. The contribution of radiation therapy (RT) exposure to MFI was assessed in a sub-analysis, and the correlation between the number of comorbidities and MFI was measured using the Pearson method. Results: MFI was strongly correlated with gait speed in AYA CCS. CCS were more likely to be pre-frail than controls without cancer history (p=0.032), and CCS treated with RT were more likely to be pre-frail than CCS not treated with RT (p<0.001). The number of comorbidities was strongly correlated with MFI (ρ=0.65), with a 0.028 increase in MFI for each added condition (p<0.001). Conclusions: Results from this study support higher risk for frailty among CCS, especially those with multiple comorbidities or who were treated with RT. A wrist-worn sensor-based method is feasible for application in AYA CCS, and provides an opportunity for cost-effective, rapid screening of at-risk AYA CCS who may benefit from early interventions.
Key words: Survivorship, frailty, slowness, weakness.
Over 80% of children diagnosed with cancer survive their disease (1), but survivors of childhood cancer (CCS) are at risk for late effects including early onset aging and frailty (2-10). An increase in frailty accompanies physiologic aging, affecting ~9% of persons >65 years (11) and 25-40% of persons >80 years (12). Frailty identifies individuals more vulnerable to adverse health outcomes (13) and predicts risk for early mortality (14, 15). Compared with sibling controls, CCS are more likely to report poor general health, functional impairment, and activity limitations with a prevalence that increases with age (16, 4, 3). Among 1,922 adult CCS with a mean age of 33.6 +/- 8.1 years, up to 13.1% were frail and up to 31.5% were pre-frail when assessed by the Fried clinical criteria, similar to the frailty prevalence in the general population that is at least three decades older (5). Frailty in CCS has been associated with exposure to cranial radiation and higher doses of abdominal or pelvic radiation (17).
In addition to the Fried clinical criteria, a number of methods assess frailty by clinical or survey-based assessment of factors such as weight loss, exhaustion, poor grip strength, slow gait speed, low physical activity, chronic health conditions, functional impairments, clinical symptoms, and behavior/psychological factors such as poor sleep quality or mood (18). These methods were largely developed for the frail older patient and validated in geriatric populations. Given the high prevalence of frailty in adolescent and young adult (AYA) CCSs, there is an unmet need for efficient, reliable, and low-cost methods that assess frailty across the lifespan.
We conducted a pilot study in AYA CCS leveraging a sensor-based upper extremity frailty assessment tool that has a strong correlation with well-established clinical and survey-based methods of frailty assessment in adults (sensitivity and specificity of 85% and 93% for predicting pre-frailty and 100% sensitivity and specificity for predicting frailty compared with the Fried criteria, sensitivity and specificity of 78% and 82% compared with a validated modified Rockwood questionnaire, and good correlation (ρ = 0.78) with the 6 minute walk test) (19-22). The Fried criteria and Rockwood questionnaire are considered gold standards for measuring frailty in elderly populations, but include components that are not appropriate for application to AYA populations. Therefore, our first objective was to validate the frailty assessment tool by determining correlation between mean frailty index and gait speed, a component of the Fried criteria, in AYA CCS. We then 1) tested feasibility of applying the frailty assessment tool in AYA CCS in an outpatient setting, 2) compared mean frailty index and frailty status between CCS and controls and CCS with or without high-risk treatment exposures (i.e. radiation), and 3) assessed the correlation between mean frailty index and the number of comorbidities present at the time of enrollment.
For the frailty meter validation, eligible cases were recruited from the Texas Children’s Cancer and Hematology Centers (TCCC), and included CCS who were ≥15 years old, English-speaking, ≥1 year off therapy, and in remission.
For the feasibility assessment and comparisons between cases and controls, cases were recruited from CCS who met the same above-described eligibility criteria, but were non-overlapping with subjects recruited to the validation step. Controls were ≥15 years old, without cancer history, and recruited from routine well visit patients at a Texas Children’s Pediatric Clinic or Baylor Clinic. For CCS, comorbidities present at the time of enrollment were abstracted from the electronic medical record (EMR). Comorbidities were defined as chronic health conditions listed as an active problem in the EMR Problem List, present for at least one year (determined by problem start date) and requiring ongoing medical care (determined by scheduled subspecialty clinic visits). All cases and controls provided informed consent, as well as assent when applicable, for participation and were enrolled to an IRB-approved research protocol. This research was conducted in accordance with the ethical standards of the Baylor College of Medicine Institutional Review Board (H-38994) and with the 1964 Helsinki declaration and its later amendments.
Mean frailty index (MFI) was determined for each arm by trained personnel using a wrist-worn frailty meter (Figure 1) (26). A wearable sensor is placed on the patient’s wrist, and, while seated, the participant performs a repetitive elbow flexion and extension task as quick and steadily as possible for 20 seconds (single task assessment). The dual task assessment adds a cognitive load to the single task, and in older adults leads to worse performance and higher MFI scores, an effect that is most pronounced among adults with cognitive impairment (27). During the dual task assessment, the subject counts down from a random number provided by the test administrator while performing the task (27). The sensor contains a tri-axial gyroscope that estimates three-dimensional angular velocity of the upper arm and forearm segments. Outcomes measures representing the kinematics and kinetics of elbow flexion, including speed, rise time, and flexion number per 20-second interval (indicators of slowness); flexibility (indicator of rigidity); power and moment (indicators of weakness); speed variability (indicator of steadiness); and speed reduction (indicator of exhaustion) are derived from the angular velocity, anthropometric data, and sex, and captured wirelessly through a tablet device (BioSensics, LLC, Newton, MA, USA). MFI is derived from these outcome measures using an optimized linear regression model previously described by Lee et al., with a numeric output on a continuous scale between 0 and 1 (26). In addition to the MFI, these outcome measures quantify slowness, weakness, and exhaustion (20). Participants completed the single and dual task assessments for both the dominant and non-dominant upper extremities in an average time of just under 5 minutes.
Gait Speed Assessment
For the validation step, gait speed was determined by the Timed Up and Go (TUG) test (23), which is both reliable and reproducible in AYAs and for which normative values in this age range are available (24, 25). Briefly, the time (in seconds) required for a subject to rise from a standard armchair, walk a distance of 3 meters, turn, walk back to the chair, and sit down again is measured for each subject.
Quality of Life Assessment
Given the well-described inverse relationship between frailty and quality of life (28), and evidence for impact of frailty on quality of life among CCS (29), we included the PROMIS measure to assess this outcome in conjunction with frailty assessment in our case-control comparisons. Each participant completed the PROMIS questionnaire, a reliable measure of patient-reported health status for mental and physical well-being that has been validated in both children and adults (30). Metrics included in this study were the PROMIS Global Physical Health, Global Mental Health, and Global Health.
For the validation step, the relationship between MFI and TUG time was determined by the Pearson correlation test. For the subsequent comparisons, participants were classified based on pre-defined MFI thresholds obtained from the dominant arm as robust (MFI <0.20), pre-frail (0.20≤ MFI <0.35), or frail (MFI ≥0.35), using the benchmarks proposed by Rockwood et al (31). After confirming that the data were normally distributed, the mean MFI and MFI subcomponents were compared between cases and controls using a Student t test, and the number of pre-frail/frail subjects by a Chi Square test. For CCS, the relationship between the number of comorbidities and MFI was determined by linear regression, and correlation was tested using the Pearson method. Factors previously associated with frailty include any exposure to cranial radiation in both sexes and abdominal/pelvic radiation in excess of 34Gy/40Gy, respectively, in males) (17). Therefore, we conducted a sub-analysis to compare outcomes between CCS who met criteria for at-risk RT exposure to CCS who did not meet criteria for at-risk RT. PROMIS data were categorized by T score and established cut points for excellent, very good, good, fair, and poor, and then analyzed by the Chi Square test (32).
Seven CCS were recruited for the validation step (two females and five males), mean age of 22.8 years (15-30 years). The mean for the Timed Up and Go (TUG) was 7.87 seconds (range, 5.23-10.12 seconds). Results were comparable to age-normative values for all but two subjects, who each exceeded the upper bound of the 95% confidence interval for the mean by 0.68 seconds and by 1.15 seconds. MFI was 0.19, range 0.13-0.25, and a strong correlation was observed between MFI and TUG time (ρ =0.83, p=0.02).
There were 48 potentially eligible CCS seen for an office visit in the TCCC Late Effects Clinic between June 29 and July 31, 2019. Subjects were recruited four days per week, so that 34 CCS were approached for participation, of which 32 consented and two declined (94% participation rate). All controls who were approached consented to study, so that there were 32 CCS and 32 age-comparable controls that were enrolled. The distribution of primary diagnoses among CCS were as follows: leukemia/lymphoblastic lymphoma, 20; germ cell tumor, 1; Hodgkin/non-Hodgkin lymphoma, 2; rhabdomyosarcoma, 2; Wilm’s tumor, 2; central nervous system tumor, 3; retinoblastoma, 1; ovarian cancer, 1. Participants were a mean of 9.8 years since completion of cancer treatment (median 8.0 years, range 1-36 years). Table 1 shows the distribution of baseline characteristics among CCS and controls.
The mean MFI for the dominant arm, both single and dual tasks, was higher among CCS than controls (p=0.002 and p<0.001, respectively, Table 2). The difference in MFI was primarily driven by the weakness and slowness components of MFI, rather than exhaustion. For both the single and dual task assessments in the dominant arm, CCS were more likely to be pre-frail than controls (p=0.032 and p=0.003, respectively). None of the participants met the pre-determined MFI cutoff criteria for frailty.
* MFI thresholds for pre-frail and frail were pre-determined from the benchmarks proposed by Rockwood et al, (31) i.e. robust: MFI <0.20, pre-frail: 0.20≤ MFI <0.35, and frail: MFI ≥0.35. All results are displayed as a mean score ± SD
Out of 32 CCS, 13 had at-risk RT exposures: 12 were treated with cranial radiation (12-55.8 Gy), and one male was treated with pelvic RT (50.4 Gy). No females were treated with abdominal or pelvic RT. In the single task assessment, CCS with at-risk RT exposures had a higher MFI (p<0.001) and were more likely to be pre-frail (p<0.001) than CCS without RT exposure (Table 2). As expected, the 12 CCS whose treatment included CRT had a significantly higher MFI than CCS not exposed to CRT, and though there was some evidence of dose-dependence, this difference was not statistically significant (p=0.15, Table 3). No substantial difference in MFI was noted after the addition of a cognitive load (dual task assessment) in controls and cases, regardless of CRT exposure (Tables 2 and 3).
All results are displayed as mean score ± SD
Eleven CCS had no comorbidities, 10 had one condition, 5 had two conditions, 4 had three conditions, 1 had four conditions, and 1 had six conditions, described in more detail in Figure 2. The number of comorbidities was correlated with MFI (ρ = 0.65), with a 0.028 increase in MFI for each added condition (p<0.001).
No significant differences in the proportion of subjects reporting ‘poor’ or ‘fair’ vs. ‘good,’ very good,’ or ‘excellent’ global health, global mental, or global physical health status were observed between controls and CCS.
Cardiometabolic: congestive heart failure (1), hypertension (2), obesity (6); Endocrine: hypogonadism (1), hypothyroidism (2), primary ovarian failure (2), growth hormone deficiency (3); Vision/hearing: cataracts (2), severe visual impairment (2), hearing loss or tinnitus (3); Peripheral/Central nervous system: peripheral neuropathy (1), hemiparesis (1), chronic migraine (1), epilepsy (2); Musculoskeletal: arthritis (1), osteopenia/osteonecrosis (4); Neuropsychological: generalized anxiety disorder (2) major depressive disorder (2); Pulmonary: chronic lung disease (1), obstructive sleep apnea (1); Renal: chronic kidney disease (1); Hematological: chronic pancytopenia (1)
There is considerable need for methods that rapidly, effectively, and inexpensively screen CCS for evidence of pre-frailty or frailty conferred by cancer treatment. To date, studies conducted in CCS have used the clinical Fried criteria, which is both time-consuming and requires training to administer. In elderly persons, the wrist-worn sensor-based method for frailty status determination has a strong correlation with frailty status determined by both the Fried criteria and the Rockwood Frailty Index (19-22), and is a strong predictor of adverse outcomes such as prolonged hospitalizations and prospective falls (35, 36, 27, 37). Given that this tool has largely been applied in older adult populations, we first validated its use in AYA CCS by demonstrating correlation between frailty status determined by wrist-worn sensor and gait speed.
Our study suggests that this method is feasible for application in the outpatient setting. In our pilot study, CCS were more likely to have a higher MFI and be pre-frail than controls, and CCS with at-risk RT exposures were more likely to have a higher MFI and be pre-frail than CCS without a history of at-risk RT, in line with prior reports (5). Of note, the mean age of CCS in this study was 20.5 years, so it is not surprising that no CCS met thresholds for frailty. CRT-exposed CCS showed no discernible difference in MFI with the addition of a cognitive load (dual task) compared with the single task, and compared with non-CRT-exposed CCS. The absence of an effect on MFI observed with addition of a cognitive load suggests that the dual task may be of less importance when assessing frailty in CCS. MFI in CCS was strongly correlated with the number of comorbidities, but the study was not powered to detect associations with diagnosis, treatment dose, or modality other than RT exposure.
Detection of pre-frailty is important, because it offers an opportunity for early intervention in an anticipated trajectory of continued physical decline. The frailty assessment tool described here is a safe, low cost, and time-efficient method, requiring less than 5 minutes to complete compared with the 10-20 minutes required for the Fried method (33, 34). Moreover, it is a stand-alone tool that requires minimal training to use, and does not require additional equipment such as a dynamometer, stopwatch, or 15 foot floor tape. Our pilot data suggest higher MFI in CCS, especially CCS treated with RT compared with controls, and support prospective application of this method to predict risk for morbidity and mortality in CCS, correlated with objective functional and biological measures.
Acknowledgements: The authors would like to thank the patients who participated in this study as well as their families.
Funding: This work was supported by the Texas Children’s Cancer and Hematology Centers.
Conflicts of Interest: None declared by the Authors.
Ethical Standards: This study was approved by the Baylor College of Medicine institutional review board, and conducted in accordance with the Helsinki Declaration of 1975, as revised in 2000.
1. Howlader N, Noone AM, Krapcho N, Neyman R, Aminou S, Altekruse C et al. SEER cancer statistics review, 1975-2009 (vintage 2009 populations). Bethesda, MD: National Cancer Institute; 2012.
2. Hudson MM, Oeffinger KC, Jones K, Brinkman TM, Krull KR, Mulrooney DA et al. Age-dependent changes in health status in the Childhood Cancer Survivor cohort. J Clin Oncol. 2015;33(5):479-91.
3. Ness KK, Hudson MM, Ginsberg JP, Nagarajan R, Kaste SC, Marina N et al. Physical performance limitations in the Childhood Cancer Survivor Study cohort. J Clin Oncol. 2009;27(14):2382-9.
4. Ness KK, Mertens AC, Hudson MM, Wall MM, Leisenring WM, Oeffinger KC et al. Limitations on physical performance and daily activities among long-term survivors of childhood cancer. Ann Intern Med. 2005;143(9):639-47.
5. Ness KK, Krull KR, Jones KE, Mulrooney DA, Armstrong GT, Green DM et al. Physiologic frailty as a sign of accelerated aging among adult survivors of childhood cancer: a report from the St Jude Lifetime cohort study. J Clin Oncol. 2013;31(36):4496-503.
6. Rueegg CS, Michel G, Wengenroth L, von der Weid NX, Bergstraesser E, Kuehni CE et al. Physical performance limitations in adolescent and adult survivors of childhood cancer and their siblings. PLoS One. 2012;7(10):e47944.
7. Armstrong GT, Kawashima T, Leisenring W, Stratton K, Stovall M, Hudson MM et al. Aging and risk of severe, disabling, life-threatening, and fatal events in the childhood cancer survivor study. J Clin Oncol. 2014;32(12):1218-27.
8. Armstrong GT, Reddick WE, Petersen RC, Santucci A, Zhang N, Srivastava D et al. Evaluation of memory impairment in aging adult survivors of childhood acute lymphoblastic leukemia treated with cranial radiotherapy. J Natl Cancer Inst. 2013;105(12):899-907.
9. Henderson TO, Ness KK, Cohen HJ. Accelerated aging among cancer survivors: from pediatrics to geriatrics. American Society of Clinical Oncology educational book / ASCO American Society of Clinical Oncology Meeting. 2014:e423-30.
10. Ness KK, Kirkland JL, Gramatges MM, Wang Z, Kundu M, McCastlain K et al. Premature Physiologic Aging as a Paradigm for Understanding Increased Risk of Adverse Health Across the Lifespan of Survivors of Childhood Cancer. J Clin Oncol. 2018;36(21):2206-15.
11. Collard RM, Boter H, Schoevers RA, Oude Voshaar RC. Prevalence of frailty in community-dwelling older persons: a systematic review. J Am Geriatr Soc. 2012;60(8):1487-92.
12. Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56(3):M146-56.
13. Macklai NS, Spagnoli J, Junod J, Santos-Eggimann B. Prospective association of the SHARE-operationalized frailty phenotype with adverse health outcomes: evidence from 60+ community-dwelling Europeans living in 11 countries. BMC Geriatr. 2013;13:3.
14. Chang SF, Lin PL. Frail phenotype and mortality prediction: a systematic review and meta-analysis of prospective cohort studies. Int J Nurs Stud. 2015;52(8):1362-74.
15. Shamliyan T, Talley KM, Ramakrishnan R, Kane RL. Association of frailty with survival: a systematic literature review. Ageing Res Rev. 2013;12(2):719-36.
16. Ness KK, Morris EB, Nolan VG, Howell CR, Gilchrist LS, Stovall M et al. Physical performance limitations among adult survivors of childhood brain tumors. Cancer. 2010;116(12):3034-44.
17. Hayek S, Gibson TM, Leisenring WM, Guida JL, Gramatges MM, Lupo PJ et al. Prevalence and Predictors of Frailty in Childhood Cancer Survivors and Siblings: A Report From the Childhood Cancer Survivor Study. J Clin Oncol. 2019:JCO1901226.
18. Aguayo GA, Donneau AF, Vaillant MT, Schritz A, Franco OH, Stranges S et al. Agreement Between 35 Published Frailty Scores in the General Population. Am J Epidemiol. 2017;186(4):420-34.
19. Toosizadeh N, Wendel C, Hsu CH, Zamrini E, Mohler J. Frailty assessment in older adults using upper-extremity function: index development. BMC Geriatr. 2017;17(1):117.
20. Toosizadeh N, Mohler J, Najafi B. Assessing Upper Extremity Motion: An Innovative Method to Identify Frailty. J Am Geriatr Soc. 2015;63(6):1181-6.
21. Toosizadeh N, Berry C, Bime C, Najafi B, Kraft M, Mohler J. Assessing upper-extremity motion: An innovative method to quantify functional capacity in patients with chronic obstructive pulmonary disease. PLoS One. 2017;12(2):e0172766.
22. Toosizadeh N, Joseph B, Heusser MR, Orouji Jokar T, Mohler J, Phelan HA et al. Assessing Upper-Extremity Motion: An Innovative, Objective Method to Identify Frailty in Older Bed-Bound Trauma Patients. J Am Coll Surg. 2016;223(2):240-8.
23. Podsiadlo D, Richardson S. The timed “Up & Go”: a test of basic functional mobility for frail elderly persons. J Am Geriatr Soc. 1991;39(2):142-8.
24. Nicolini-Panisson RD, Donadio MV. Normative values for the Timed ‘Up and Go’ test in children and adolescents and validation for individuals with Down syndrome. Dev Med Child Neurol. 2014;56(5):490-7.
25. Kear BM, Guck TP, McGaha AL. Timed Up and Go (TUG) Test: Normative Reference Values for Ages 20 to 59 Years and Relationships With Physical and Mental Health Risk Factors. J Prim Care Community Health. 2017;8(1):9-13.
26. Lee H, Joseph B, Enriquez A, Najafi B. Toward Using a Smartwatch to Monitor Frailty in a Hospital Setting: Using a Single Wrist-Wearable Sensor to Assess Frailty in Bedbound Inpatients. Gerontology. 2018;64(4):389-400.
27. Toosizadeh N, Najafi B, Reiman EM, Mager RM, Veldhuizen JK, O’Connor K et al. Upper-Extremity Dual-Task Function: An Innovative Method to Assess Cognitive Impairment in Older Adults. Front Aging Neurosci. 2016;8:167.
28. Kojima G, Iliffe S, Jivraj S, Walters K. Association between frailty and quality of life among community-dwelling older people: a systematic review and meta-analysis. J Epidemiol Community Health. 2016;70(7):716-21.
29. Ness KK, Howell CR, Bjornard KL. Frailty and quality of life in adult survivors of childhood cancer. Expert Rev Qual Life Cancer Care. 2017;2(2):79-85.
30. Cella D, Yount S, Rothrock N, Gershon R, Cook K, Reeve B et al. The Patient-Reported Outcomes Measurement Information System (PROMIS): progress of an NIH Roadmap cooperative group during its first two years. Med Care. 2007;45(5 Suppl 1):S3-S11.
31. Rockwood K, Andrew M, Mitnitski A. A comparison of two approaches to measuring frailty in elderly people. J Gerontol A Biol Sci Med Sci. 2007;62(7):738-43.
32. Hays RD, Spritzer KL, Thompson WW, Cella D. U.S. General Population Estimate for “Excellent” to “Poor” Self-Rated Health Item. J Gen Intern Med. 2015;30(10):1511-6.
33. Kim H, Higgins PA, Canaday DH, Burant CJ, Hornick TR. Frailty assessment in the geriatric outpatient clinic. Geriatr Gerontol Int. 2014;14(1):78-83.
34. Dent E, Kowal P, Hoogendijk EO. Frailty measurement in research and clinical practice: A review. Eur J Intern Med. 2016;31:3-10.
35. Joseph B, Pandit V, Rhee P, Aziz H, Sadoun M, Wynne J et al. Predicting hospital discharge disposition in geriatric trauma patients: is frailty the answer? The journal of trauma and acute care surgery. 2014;76(1):196-200.
36. Joseph B, Pandit V, Zangbar B, Kulvatunyou N, Tang A, O’Keeffe T et al. Validating trauma-specific frailty index for geriatric trauma patients: a prospective analysis. J Am Coll Surg. 2014;219(1):10-7 e1.
37. Joseph B, Toosizadeh N, Orouji Jokar T, Heusser MR, Mohler J, Najafi B. Upper-Extremity Function Predicts Adverse Health Outcomes among Older Adults Hospitalized for Ground-Level Falls. Gerontology. 2017;63(4):299-307.