Assessment of Sarcopenia Measures, Survival, and Disability in Older Adults Before and After Diagnosis With Cancer

Key Points Question What is the association of a cancer diagnosis with measures of sarcopenia? Findings In this matched cohort study of 515 older patients who developed cancer from the Health, Aging, and Body Composition (Health ABC) study, a steeper decline in gait speed prior to a cancer diagnosis and accelerated declines in appendicular lean mass after a cancer diagnosis were found, which were most striking in patients with metastatic disease. Meaning Accelerated losses in differing sarcopenia measures exist both prior to and after a cancer diagnosis and may present opportunities for targeted interventions to improve outcomes.


Introduction
Age-related loss in muscle mass and strength, known as sarcopenia, is detected as early as the 4th decade of life and shows a linear age-related decline. 1,2 Age-related sarcopenia may be further complicated by the development of cancer and related cancer treatments. 3 Low skeletal muscle mass is highly prevalent in older adults with cancer and strongly associated with a higher risk of adverse events, such as chemotherapy-related toxic effects, surgical complications, and poorer overall survival. [4][5][6][7] Cancer cachexia is a multifactorial syndrome caused by an elevated inflammatory response combined with reduced food intake that may contribute to additional muscle loss in patients with cancer. [8][9][10] However, the contribution of cancer to the trajectory of sarcopenia both preceding and after a cancer diagnosis remains unknown. Furthermore, the association of sarcopenia at cancer diagnosis with subsequent disability in older adults with cancer also remains largely unknown. 3 To fill these gaps, we examined the rate of decline of sarcopenia indices (ie, appendicular lean mass [ALM], muscle strength, and physical performance) in older adults with cancer before and after cancer diagnosis compared with the trajectory of a population without cancer. We also assessed the association of these sarcopenia indices with overall survival and major disability in patients with cancer.

Sample
The Health, Aging, and Body Composition (Health ABC) study 11 was a prospective longitudinal observational study designed to assess the association of body composition with physical function and disability in older adults. The study enrolled 3075 well-functioning community-dwelling older adults aged 70 to 79 years at enrollment from a random sample of white Medicare beneficiaries and all eligible black residents in designated zip code areas in and around Pittsburgh, Pennsylvania, and Memphis, Tennessee, between March 1997 and July 1998. At enrollment, participants reported no difficulty performing activities of daily living, walking 0.25 miles, or climbing 10 steps without resting.
Each local institutional review board approved the protocol, and all participants provided written informed consent. The current secondary analyses were considered exempt under category 4 of federal regulations. Follow-up clinical examinations or home-based interviews were performed annually, and short telephone interviews were performed between each annual interview at 6 months. A report of a new cancer diagnosis during follow-up assessments prompted an adjudication protocol with procurement of pathology and cytology reports as well as supportive radiologic and laboratory data (nonmelanoma skin cancer was counted as control). Date of diagnosis, primary cancer site, and presence of metastases were documented. The study period includes cancer reported and adjudicated through December 2004 and follow-up for mortality and major disability through December 2013. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. 12

Sarcopenia Indices
Appendicular lean mass (in kilograms) was assessed by summing lean tissue (fan-beam dual-energy radiography absorptiometry 4500A version 8.20a [Hologic]) across all extremities. Low ALM was defined as less than 19.75 kg for men and less than 15.02 kg for women. 13 Hand grip strength was assessed using the JAMAR hydraulic handheld dynamometer (MESM) and computed as an average from 2 trials of each hand. A hand grip strength less than 26 kg for men and less than 16 kg for women was used to define low hand grip strength. 13 Gait speed was evaluated over a 20-m course and examined as a continuous variable, with a value of 0.8 m/s or less considered abnormal. 14,15 Covariates Demographic characteristics included age, sex, race/ethnicity, education, and study site. Cancerrelated information included cancer site and cancer stage (limited vs metastatic).

Outcomes
Sarcopenia indices constituted the primary outcome variables. Secondary outcomes included overall survival and major disability after cancer diagnosis; these analyses were restricted to the cancer cohort. Major disability was defined as requiring a cane or walker for ambulation, inability to walk 0.25 miles or climb 10 steps or more, or requiring assistance with activities of daily living, including transferring, dressing, or bathing.

Statistical Analyses
Given sex differences between those with and without cancer, we selected a 2:1 match of participants without cancer to participants with cancer based on sex, race/ethnicity, education, and age at enrollment. In addition, we separately performed a 5:1 match for propensity scores to develop controls without cancer for patients with prostate, colorectal, lung, breast, and other cancers, as well as those with metastatic disease. Controls without cancer were assigned an age at pseudodiagnosis, based on average age at diagnosis for the patients with cancer. Greedy matching technique was used with a caliper width equal to 0.2 in the matching. 16 Using the matched sets, we used linear mixedeffect models with patient-level random intercepts to compare the slopes in each of the 3 sarcopenia indices (ALM, hand grip strength, and gait speed) between individuals who did and did not develop cancer, further adjusting for age at diagnosis, race/ethnicity, sex, education, enrollment site, and longitudinal comorbidities, including diabetes, cardiovascular disease, and arthritis at each measurement. Given concern for multiple comparisons, we chose a stricter significance threshold of a 2-tailed P value less than .01. The slopes were examined for the available period before cancer diagnosis (censoring at 1 month before cancer diagnosis) and for the period from cancer diagnosis to the last evaluation in patients with cancer. For determining the association between overall survival and sarcopenia measures, we used multivariable Cox regression; the at-risk period included time from cancer diagnosis to death or last contact (if alive). For determining the association between major disability and sarcopenia measures, we used multivariable proportional subdistribution hazards regression (Fine-Gray method) using death as a competing risk. Dichotomized thresholds of impairment for each sarcopenia measure were treated as time-varying covariates. Missing data (minimal) were treated as missing at random with no imputations; all analyses used available data.
SAS statistical software version 9.4 (SAS Institute) was used for analyses.

Results
diagnosed as having cancer during the first 7 years of the study. Participants were observed for a median (range) of 3.2 (0 to 6.9) years prior to cancer diagnosis and 0.5 (0 to 5.3) years after cancer diagnosis. The cancer subcohort was observed for a median (range) of 4.2 (0 to 17) years and 0.5 (0 to 17) years after cancer diagnosis for mortality and major disability analyses, respectively. Most  Table 2).

Discussion
Accelerated losses in differing sarcopenia indices were found among older adults before and after cancer diagnosis compared with a comparison group without cancer. Decline in gait speed was observed prior to cancer diagnosis, and accelerated decline in ALM was seen after cancer diagnosis, particularly in those with metastatic disease. Only low gait speed was associated with increased mortality and disability. The Health ABC study offered a rare opportunity to examine changes in  Accelerated declines in gait speed leading up to cancer diagnosis may be caused by symptoms, hospitalizations, and/or surgery associated with a cancer diagnosis. Although gait speed is a relatively simple measurement to obtain, walking is a complex task involving multiple systems that must work in concert. 21 Based on the Figure, it appears that the losses in gait speed preceding cancer diagnosis are largely recovered after cancer diagnosis, potentially suggesting a more acute decline, as would be anticipated from a hospitalization. However, it was surprising to find significant declines in gait speed before cancer diagnosis specifically within individuals with prostate cancers, as many of these individuals likely had asymptomatic and localized disease. The association of gait speed with mortality and disability is consistent with recent results in oncology and again highlights the importance of gait speed as a potential marker of frailty and as a critical tool in risk prediction in older adults with cancer. [21][22][23] The lack of association of low ALM with mortality is in contrast with previous reports in oncology; however, the vast majority of previous literature within cancer have relied on single-slice computed tomography for measuring body composition and did not use dual-energy radiography absorptiometry to assess muscle mass. 24 Additionally, our findings suggest potential discrepancies in the accelerated losses in sarcopenia measures. Although these measures all represent different facets of sarcopenia, the relationship between muscle mass, muscle strength, and physical performance is not linear, and prior research has shown only modest correlation. [25][26][27] Notably, in a study of older adults with cancer, there was no association between the skeletal muscle mass and physical performance impairments. 28 Muscle strength and physical performance do not depend solely on muscle mass, and many other factors complicate this relationship, including muscle composition. 29 Although the terms sarcopenia and cachexia are often used interchangeably within the context of cancer, these represent distinct entities. 30 Similar to sarcopenia, cachexia is associated with a decrease in muscle mass, muscle strength, and function and is also associated with losses in body weight as well as an increase in basal metabolic rate and markers of inflammation. 8

Limitations
This study should be considered in the context of its limitations. The older participants within the Health ABC study were recruited from only 2 US cities (Pittsburgh and Memphis) and were required to be well functioning at baseline, thus limiting the generalizability of our findings. Although the Health ABC study identified and confirmed new cancer diagnoses, no cancer treatment information was obtained, and we were unable to examine the association of specific treatments with trajectories of specific sarcopenia indices. Although the Health ABC study was a large study including more than 3000 participants, only 16.7% of participants developed cancer, and many of the subgroup analyses within specific cancer types were likely underpowered. Furthermore, the development of sarcopenia is often multifactorial and a multitude of events and conditions can influence this process, and we were not able to account for all events (eg, hospitalizations, weight loss, comorbid conditions) that may have contributed to the decline.

Conclusions
These limitations notwithstanding, this study affords a rare glimpse of sarcopenia indices both before and after cancer diagnosis compared with a population without cancer. Accelerated loses in gait speed prior to cancer diagnosis and accelerated loses in ALM after cancer diagnosis suggest that a cancer diagnosis does affect age-related losses in some sarcopenia indices. However, many unanswered questions remain. Are these losses the result of cancer cachexia or cancer treatment?
Do these accelerated losses represent a true increase in the rate of age-related decline or rather a phase-shift due to acute effects of cancer and cancer treatment? What are the best interventions to mitigate the decline in sarcopenia indices? Further work is needed to provide a deeper understanding