The impact of low muscle mass on prognosis following neoadjuvant chemotherapy for resectable locally advanced rectal cancer: a systematic review and meta‐analysis

Sarcopenia is characterized by the progressive and generalized loss of muscle mass and function. There is an increasing body of evidence to suggest that cancer patients with pre‐existing sarcopenia are at a greater risk of both short‐ and long‐term clinical complications. The aim of this review is to examine the impact of low muscle mass on prognostic outcomes in patients with locally advanced rectal cancer (LARC) who undergo neoadjuvant chemoradiotherapy (nCRT) prior to surgery.


Introduction
Colorectal cancer (CRC) is the third leading cause of cancer deaths and fourth most commonly diagnosed cancer in the world. 1 In the UK, the largest proportion of bowel cancer cases occurs in the rectum, accounting for 31.5% and 23.1% of cases in males and females, respectively. 2 Surgery is the gold standard treatment for people with rectal cancer provided the tumour is deemed resectable. 3 In patients who have locally advanced rectal cancer (LARC), those who undergo neoadjuvant chemoradiotherapy (nCRT) have been found to have less local recurrence and better overall and disease-free survival (DFS). 3 Further advantages of nCRT include decreased tumour volume and enhanced probability of anal sphincter preservation. 4 Treatment with nCRT also improves the likelihood of a complete pathological response, with an R0 resection being recorded in up to 20% of cases. 5 Increasing age is one of the most important risk factors for developing cancer, with three quarters of all cancers in the UK being diagnosed in those aged over 60 and a third diagnosed in those aged over 75. 6 Progressive loss of muscle mass, strength, and physical function is one of the most distinctive features of ageing, 7 with muscle mass peaking around the middle of the third decade of life, followed by a slow rate of decline until the fifth decade. After this, the rate of muscle loss increases, ranging between 0.5% and 1.4% loss each year. 8 The implications of skeletal muscle loss are far reaching and can impact an individual's ability to perform activities of daily living (ADLs), as well as being a precipitating factor in a decline in physical function, increased risk of frailty, disability, and death in the elderly. 9 The term sarcopenia was first used when referring to the progressive loss of lean body mass seen in ageing. 10 The definition has since been expanded within guidelines published by the European Working Group on Sarcopenia in Older People (EWGSOP) to include the presence of lower muscle mass and weakness and/or impaired performance, with muscle strength now thought to be the most reliable measure of sarcopenia due to its superiority in predicting adverse outcomes. 11 The EWGSOP also describes two categories of sarcopenia: primary sarcopenia in which ageing appears to be the only cause and secondary sarcopenia in which conditions such as malignancy play a role. 11 It is estimated that the prevalence of sarcopenia in the older population ranges between 4% and 27%, depending on the gender and country. 12 Despite sharing many pathological features, sarcopenia is not to be confused with cachexia. 13 Cachexia is a multifactorial condition characterized by an energy imbalance, which results from decreased nutritional intake and increased energy expenditure caused by a hypermetabolic state. Commonly seen in diseases such as cancer, 14 cachexia results in skeletal muscle mass losses with or without loss of adipose tissue. 15 Unlike sarcopenia, the definition of cachexia is centred on weight loss rather than the mass and function of skeletal muscle. 16,17 It is therefore possible that a patient with cancer may suffer both cachexia and sarcopenia and progress from secondary sarcopenia to cachexia if sufficient weight loss occurs during their treatment. 13 It is becoming increasingly apparent that body composition and nutritional status play an important role in predicting both short-and long-term clinical outcomes in cancer patients. 18 For example, low skeletal muscle mass has been shown to be a negative prognostic indicator in malignant diseases of the lungs and gastrointestinal tract. 19,20 Specific to CRC, reduced skeletal muscle mass and increased visceral fat mass have each been shown to be negative prognostic factors. [21][22][23] Similarly, an association between low skeletal muscle mass and short-term outcomes, such as length of stay (LOS) and infection rate, has been shown in patients with CRC and those with colorectal liver metastasis. [24][25][26] Low skeletal muscle mass in cancer patients is multifactorial, resulting from both modifiable (such as nutritional status and physical activity levels) and non-modifiable (such as age and hormone concentrations) features. 9 In relation to the impact of cancer treatment, both chemo-and radiotherapy have well-known side effects, including nausea, vomiting, and anorexia, all of which can accelerate losses of skeletal muscle mass in cancer patients. 27 Further illustrating the detrimental impact of cancer treatment on muscle mass, a number of studies have shown a relationship between the magnitude of muscle loss and the duration of treatment, with the most pronounced reductions in patients with advanced or metastatic disease. 27,28 When looking at patients with resectable disease, Eriksson et al. 29 showed that skeletal muscle mass decreased during nCRT for CRC liver metastasis and that those who were sarcopenic preoperatively had worse overall survival (OS).
Despite the evidence outlined above, the current literature is lacking consolidated information relating to the impact that low muscle mass has on the prognosis of patients requiring nCRT prior to surgery for specific cancer types. Therefore, the aim of this review is to examine the relationship between pre-existing sarcopenia and prognostic outcomes in patients with locally advanced CRC who underwent nCRT prior to surgery. The primary outcome was OS, with DFS being a secondary outcome.

Review design
This systematic review was registered with PROSPERO (The International Prospective Register of Systematic Reviews) prior to the literature search (registration number CRD42019157313) and carried out in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) statement. 30 Any comparative study relating to the prognostic outcomes of sarcopenic versus non-sarcopenic patients diagnosed with LARC who received nCRT prior to surgery was included. The exclusion criteria included any study, which included (i) patients with irresectable or advanced/metastatic CRC; (ii) combined outcomes of different types of cancer or treatment modalities (i.e., preoperative and postoperative chemotherapy); or (iii) introduced interventions preventing muscle loss during the treatment pathway.

Literature search
A trained clinical research librarian performed the literature search using MEDLINE, PubMed, and Embase databases via the National Institute for Health and Care Excellence (NICE) Healthcare Databases Advanced Search. There was no language or date restriction, and the databases were searched from inception until October 2021. The Cochrane Library of Systematic Reviews was searched for previous reviews, with www.clinicaltrials.gov searched for any unpublished studies. Systematic reviews of similar topics were searched for any relevant previous studies specific to locally advanced CRC.
Medical subject headings (MeSH) included the terms 'COLORECTAL TUMOR', 'NEOADJUVANT THERAPY', 'TREAT-MENT OUTCOME', and 'PROGNOSIS'. Free-text words included 'chemoradiotherapy', 'sarcopenia', and 'muscle mass'. Details of the full search strategy can be found in the Supporting information Appendix S1. Abstracts were independently screened by two authors (JH and TS) using the Rayyan systematic review software (2016, Qatar Computing Research Institute, Doha, Qatar). 31 If either author deemed the abstract potentially relevant, it was considered for full-text review. Full-text versions were independently screened against the inclusion and exclusion criteria by two authors (JH and TS). All discrepancies were resolved by consensus.

Data extraction
Study characteristics including author, year of publication, mean/median patient age (years), percentage of male individuals, percentage defined as sarcopenic at baseline (%), follow-up duration (months), and gender-specific cut-off values used for defining sarcopenia (cm 2 /m 2 ) were extracted by one author (JH), with outcome data (OS and DFS) independently extracted by two authors (JH and TS). When outcome data were not shown in the correct format, the corresponding author of the study was contacted via email, and the data were requested. The Risk Of Bias In Non-randomized Studies of Interventions (ROBINS-I) tool was used to measure risk of bias for included studies independently by two authors (JH and EH), with any disagreement resolved by consensus of a third author (TS).

Statistical analysis
Meta-analysis on reported hazard ratios (HR) and 95% confidence intervals (CI) was performed using DerSimonian-Laird random-effects models. These were converted to logHR, and standard errors were calculated from 95% CI. Where possible, multivariate HR are reported. Despite intention to perform investigation of heterogeneity and publication bias, there were too few included studies. Statistical heterogeneity was assessed using the I 2 statistic, with values above 50% taken as evidence of statistical heterogenicity. Grading of Recommendations, Assessment, Development and Evaluation (GRADE) criteria was used to assess the certainty of evidence. All analyses were conducted on Stata Version 16.

Search results
The primary search strategy produced a total of 1236 results including 108 duplicate studies. Of these, 1014 were excluded as not being relevant, leaving 14 studies for full-text review. From this, nine studies were excluded for reasons including cumulating data for preoperative and postoperative chemotherapy, excluding patients based on age/other demographics, and failure to have baseline imaging pretreatment. This left five studies that were deemed suitable for inclusion and analysis (Figure 1).

Study characteristics
Characteristics of the included studies, 32-36 which comprised of a total of 598 patients, can be seen in Table 1. The earliest study was published in 2018, with the latest being 2020. All papers were published as journal articles in peer-reviewed journals, and all included studies were retrospective cohort studies.

Risk of bias
Risk of bias was assessed using the ROBINS-I tool 37 as shown in Table 2. Four studies [32][33][34]36 were considered at moderate risk of bias for confounding factors (patient age, cancer stage, and comorbidities), with one study 35 considered at high risk of bias as it failed to perform multivariable statistical adjustment due to the small number of patients. Four studies [32][33][34]36 were deemed at moderate risk of bias in missing data because they   excluded patients who, despite meeting the criteria for inclusion, did not have the appropriate imaging pre and post nCRT. All five studies 32-36 were judged to have moderate risk of bias in relation to measurement of outcomes because they were all retrospective, non-blinded studies. Two studies 35,36 were deemed to be at moderate risk of bias in relation to reporting results; one was for reporting results of subgroups of the total cohort, whereas the other divided the cohort into quartiles to report findings. Overall, one study 35 was deemed to be at severe risk of bias, with the remaining four studies [32][33][34]36 considered at moderate risk.

Data synthesis
Four studies reported OS, [32][33][34]36 with three studies reporting DFS. [32][33][34] Data relating to OS and DFS were obtained for one further study 35 following correspondence with the lead author.
A total of 598 patients from five studies were included in the analysis of HR for OS. HR were extracted from multivari-ate analysis in two studies and univariate analysis in three studies. Meta-regression analysis, assessing the relationship between pre-existing sarcopenia and worse OS, showed a significant association (pooled HR: 1.69, 95% CI: 1.15-2.48) (Figure 2). There was low statistical heterogeneity (I 2 = 19.15%).
A total of 505 patients from four studies were included for analysis of HR for DFS. HR were extracted from multivariate analysis in one study and univariate analysis in three studies. Meta-regression analysis, assessing the relationship between pre-existing sarcopenia and shorter DFS, appeared to show an association; however, this was not statistically significant (pooled: HR 1.07, 95% CI: 0.63-1.82) (Figure 3). Heterogeneity between studies was moderate to high (I 2 = 56.05%).

Discussion
Over the past decade, interest in the association between sarcopenia and prognostic outcomes in multiple cancer types has intensified. [38][39][40] When focusing on CRC, often colon and  Muscle mass and prognosis after chemotherapy for rectal cancer rectal cancers are treated as the same entity, despite having different disease profiles and treatment options. 41 To our knowledge, this review is the first to evaluate the relationship between pre-existing sarcopenia and prognostic outcomes of patients with LARC who underwent both nCRT and resectional surgery.
The predominant finding of this review is that of a significant association between pre-existing sarcopenia and worse OS. This finding is in keeping with that of other studies and reviews, which have looked at patients who underwent curative resection of CRC and colorectal liver metastasis, 24,42,43 as well as patients with various solid cancer types across different disease stages. 44 Collectively, this indicates that a patient's body composition, specifically their muscle mass, at the time of diagnosis is an important indicator for oncological outcomes.
Based on the studies eligible for inclusion in this review, the relationship between pre-existing sarcopenia and worse DFS was less clear. Although the data are suggestive of an association, it did not reach statistical significance. This is likely a result of the high level of heterogeneity between the studies reporting this outcome. One clear contributor to this heterogeneity was the difference in cut-off values used to define sarcopenia between studies. As such, patients classed as sarcopenic in one study may not have been classed as such in another. Takeda et al. 33 highlighted this by showing that, had they used the most common cut-off values of 52.4 cm 2 /m 2 for males and 38.5 cm 2 /m 2 in females, 21 70.1% of patients in their study would have been classified as sarcopenic. Instead, Miyamoto et al. and Peng et al. defined low skeletal muscle mass using sex-specific cut-off values below the lowest quartile, something that has been applied in other studies looking at the relationship between outcomes and low skeletal muscle mass. 42,45 The high level of heterogeneity in studies reporting survival-based outcomes may also be due to individual variation in biological responses to neoadjuvant treatment. Previous studies have shown a significant association between the systemic inflammatory response caused by the presence of a tumour and the degree of muscle mass loss in patients with CRC, 46,47 with some suggesting that the metabolic and inflammatory response to a cancerous tumour may, in part, be reversed by successful neoadjuvant treatment. 41 In support of this suggestion, Heus et al. 48 analysed the effect of chemoradiation on the preoperative body composition of patients with LARC and reported an increase in skeletal muscle area following chemoradiotherapy. In addition, De Nardi et al. 35 reported that the majority (62.5%) of patients who did not respond to nCRT suffered muscle mass losses >2%, whereas 68.2% of responders (defined as tumour regression of 50% or more) showed no muscle loss. Finally, if neoadjuvant treatment can mitigate the atrophic effects of a cancerous tumour burden, 49 this may explain why, of 30 patients with pre-existing sarcopenia in the study by De Nardi et al., 35 six gained muscle mass during neoadjuvant chemotherapy.
In addition to emerging evidence highlighting the importance of muscle mass for survival outcomes after cancer treatment, it is perhaps unsurprising that muscle mass also has a role to play in determining how well patients will tolerate neoadjuvant treatment. For example, Cespedes Feliciano et al. 50 found that lower muscle mass was associated with higher odds of early discontinuation, treatment delays, and dose reductions independent of age, sex, and cancer stage. These impacts on neoadjuvant treatment are likely related to the pharmacokinetics and metabolism of many chemotherapy agents, as diminished muscle mass will reduce drug metabolism and clearance and as such lead to greater toxicity. 50 Clearly, both the effect on treatment protocols and the physiological outcomes (i.e., toxicity) have potential to negatively impact DFS and ultimately OS.
To date, few studies have looked at the impact of change in body composition during nCRT on those with resectable CRC. Although the focus of this review was to assess the impact of pre-existing sarcopenia on survival outcomes in patients undergoing nCRT before resectional surgery, it should be acknowledged that recent evidence suggests that the degree of muscle mass loss during nCRT is likely a more powerful prognostic factor than low skeletal muscle mass at any specific time point. 32 42,[51][52][53] all showed an association between the degree of skeletal muscle loss during nCRT and worse survival outcomes. Although all studies included in this review were retrospective and as such the conclusions drawn are limited, these findings concur with other studies in oesophageal and non-resectable colon cancer and reinforce the suggestion that if losses of muscle mass could be prevented, or even reversed, during neoadjuvant treatment, prognostic outcomes for cancer patients could be enhanced.
Despite the mounting evidence relating poor clinical outcomes and low skeletal muscle mass in cancer, there is still no universally accepted nutritional therapy for maintaining or enhancing muscle mass in these patients. 52,54 Although some evidence is emerging to suggest that protein intake does need to be higher (<1.4 g/kg/body weight) in patients with cancer and sarcopenia, 55 further high-quality research is needed to address the practical challenges of implementing this into the current clinical settings in a way that is acceptable for patients. Similarly, although there is growing recognition of the importance of muscle mass in the field of oncological exercise prehabilitation, the time constraints of cancer treatments and the side effects associated with neoadjuvant therapy present challenges in creating a universally accepted regimen that produces meaningful results. As such, the main focus of preoperative optimization remains improving cardiorespiratory fitness. [56][57][58] This review has several methodological limitations. Firstly, due to no universally agreed values of skeletal mass index to define sarcopenia, multiple definitions were included in this study. Secondly, because all the studies were retrospective, there is a higher risk of bias due to confounding factors and absent data. Thirdly, all studies were single-centred and had limited numbers of patients, which prevented multivariable analysis being performed. Additionally, it must be acknowledged that the focus of this review was low muscle mass and not muscle function. This is primarily because all studies included used muscle mass as the sole measure of sarcopenia. With muscle mass being a parameter that can be measured retrospectively, on imaging modalities that are universally used in clinical practice, reporting this parameter is easier than for measures of muscle function. However, with a consensus across the literature that both muscle mass and function impact overall health and that changes in muscle function are key to the diagnosis of sarcopenia, conclusions drawn from this review may be limited.
In conclusion, this review evidences a significant association between pre-existing low muscle mass and poorer OS in patients with LARC who undergo nCRT. Large-scale, prospective studies considering both skeletal muscle mass and muscle function in individuals with CRC are required to gain a greater understanding of the true impact of sarcopenia on clinical outcomes in this sizeable patient group. In addition, further research is required to establish exercise, nutritional, and/or multimodal therapy to prevent or mitigate both muscle mass and function losses in these patients.