Insufficient oral intake

Several factors may directly lead to diminished food intake and thereby insufficient energy intake, e.g. dysphagia, nausea, xerostomia and changes in taste and smell⁽ Reference Blum, Omlin and Baracos ^{7 )}. Other factors may have an indirect influence on energy intake by affecting appetite and the drive to eat, e.g. pain, fatigue and psychological problems⁽ Reference Bye, Jordhøy and Skjegstad ^{8 )}. Studies have suggested that a high prevalence of taste changes are associated with poor intakes in advanced cancer and that individuals reporting taste changes may ingest as little as 3765·6–4602·4 kJ/d (900–1100 kcal/d)⁽ Reference Hutton, Baracos and Wismer ^{9 )} with other studies reporting energy deficits of up to 230 120 kJ (55 000 calories) during chemoradiotherapy in head and neck cancers⁽ Reference Kubrak, Olson and Jha ^{10 )}. A wide variation in energy intake has been reported in cancer ranging from 16·74 to 221·75 kJ/kg per d (4 to 53 kcal/kg per d)⁽ Reference Bye, Jordhøy and Skjegstad ^{8 )}.

Tumour-related mechanisms

Tumour-related mechanisms include obstruction of the gastrointestinal tract causing dysphagia or odynophagia, as seen in oesophageal and head and neck cancers. Weight loss can be attributed to the physiologic abnormalities associated with the tumour, such as malabsorption, vomiting, diarrhoea, anorexia or the side effects of anti-cancer treatment, including chemotherapy, radiotherapy and surgery⁽ Reference Ryan, Healy and Power ^{11 –} Reference Ryan, Rowley and Healy ^{13 )}. Oral and gastrointestinal symptoms have early effects on changes in weight without regard to nutrition status and treatment modalities. Patients with weight loss have also been shown to have more depression, abdominal fullness, taste changes, vomiting, dry mouth, dysphagia or loss of appetite⁽ Reference Blum, Omlin and Baracos ^{7 )}.

Host response to tumour

A variety of metabolic and endocrine changes, and activation of catabolic pathways also account for weight loss, which is typically greater than would be expected from the prevailing level of oral intake⁽ Reference Blum, Omlin and Baracos ^{7 )}. Pro-inflammatory cytokines (secreted by either immune cells or tumours) play a central role in mediating the metabolic, physiologic and behavioural features of cancer-induced weight loss. They are key signals for lipolysis and proteolysis. Cytokines have three major effects: (1) altering macronutrient metabolism, (2) depressing appetite and (3) initiation of an acute phase protein response. This acute phase protein response is energy-intensive with high requirements for essential amino acids. The need for amino acids drives the loss of muscle. Along with the acute phase protein response, changes in intermediary metabolism also occur, most notably, in protein metabolism. Anorexia results from pro-inflammatory cytokine activity and has both central and peripheral aspects. The central effect is at the level of the hypothalamic nuclei, which control feeding behaviours. Mediators such as corticotropin-releasing hormone, serotonin or leptin may be directly or indirectly involved in producing anorexia.

In addition, inflammatory cytokines directly induce signalling pathways that up-regulate enzymes inducing muscle protein turnover. The pathway most involved in muscle wasting is protein degredation that is mediated by the activation of the ubiquitin-dependent proteasome pathway. Additionally there are abnormalities in protein synthesis and degredation and amino acid metabolism along with increased apoptosis and an impaired capacity for regeneration⁽ Reference Argilés, Busquets, Stemmler and López-Soriano ^{14 )}.

Skeletal muscle loss is accompanied by a profound loss of white adipose tissue due to a variety of factors. Lipid-mobilising factor is produced by cachexia-inducing tumours and is involved in the degradation of adipose tissue, with increased oxidation of the released fatty acids through an induction of uncoupling protein expression⁽ Reference Sanders and Tisdale ^{15 )}. In addition, there is an increase in lipolytic activity through activation of hormone-sensitive lipase; a decreased anti-lipolytic effect of insulin on adipocytes; and an important decrease in lipoprotein lipase activity⁽ Reference Argilés, Busquets, Stemmler and López-Soriano ^{14 )}. Consequently lipid uptake is severely hampered. Also de novo lipogenesis in adipose tissue is also reduced⁽ Reference Argilés, Busquets, Stemmler and López-Soriano ^{14 ,} Reference Argilés, López-Soriano and Almendro ^{16 )}. There are also recent data to suggest that white adipose tissue cells undergo a browning process and brown adipose tissue is associated with increased expression of uncoupling protein 1, which switches the use of mitochondrial electron transport from ATP synthesis to thermogenesis, resulting in increased lipid mobilisation and energy expenditure⁽ Reference Argilés, Busquets, Stemmler and López-Soriano ^{14 )}. Adipose tissue also infiltrates muscle resulting in reduced muscle density⁽ Reference Stephens, Skipworth and Macdonald ^{17 )}.

Energy expenditure and physical activity

The resting energy expenditure (REE) in cancer patients is determined by the type of tumour. REE can be unchanged, increased or decreased in relation to the predicted energy expenditure⁽ Reference Langius, Kruizenga and Uitdehaag ^{18 –} Reference Okamoto, Sasaki and Johtatsu ^{20 )}. In about 25 % of patients with active cancer, REE measured by the gold standard method, indirect calorimetry, is more than 10 % higher, and in another 25 % it is more than 10 % lower than predicted energy expenditure⁽ Reference Arends, Bodoky and Bozzetti ^{21 )}. REE has been shown to be significantly elevated in patients with an acute phase response⁽ Reference Wu, Huang and Xiao ^{22 )}.

Although REE may be elevated in cancer patients, the total energy expenditure is often decreased through reduction in the physical activity level (PAL) and it has been reported that that weight-losing cancer patients reduce the magnitude of their energy deficit by a reduction in PAL. Healthy sedentary adults have a PAL between 1·4 and 1·5⁽ Reference Black, Coward, Cole and Prentice ^{23 )}. In a study of hypermetabolic, cachectic pancreatic cancer patients, it was shown that the measured mean PAL was much lower (mean 1·24) than that recorded in healthy adults of similar age (mean 1·62)⁽ Reference Moses, Slater and Preston ^{24 )}. This level of physical activity is comparable with that observed in spinal cord injury patients living at home⁽ Reference Mollinger, Spurr and Ghatit ^{25 )} or that observed in cerebral palsy (mean PAL 1·23)⁽ Reference Stallings, Zemel, Davies, Cronk and Charney ^{26 )}. Levels of physical activity as low as this may exacerbate muscle wasting⁽ Reference Biolo, Ciocchi and Stulle ^{27 ,} Reference Kortebein, Symons and Ferrando ^{28 )} and it is well understood in any individual, that a lack of physical activity will cause deconditioning and deterioration in muscle mass⁽ Reference Fearon ^{29 )} which in turn impacts the ability to exercise. This vicious cycle leads to progressive decline in physical activity and of muscle mass.

Cancer cachexia

The wasting in cancer is often termed cancer cachexia (CC) and over the last few years several definitions of CC have emerged but they all share two common features: involuntary weight loss of muscle (and fat) and inflammation. Debate continues in the scientific literature surrounding its best definition and classification. Earlier papers described CC as ‘a wasting syndrome involving loss of muscle and fat directly caused by tumour factors, or indirectly caused by an aberrant host response to tumour presence’⁽ Reference MacDonald, Easson, Mazurak, Dunn and Baracos ^{30 )} but more recent definitions describe it as ‘a complex metabolic syndrome associated with underlying illness and characterised by loss of muscle with or without loss of fat mass’⁽ Reference Evans, Morley and Argilés ^{31 )}. In addition to muscle and adipose tissue, other organs are affected by the cachectic process. Abnormalities in heart function, alterations in liver protein synthesis, changes in hypothalamic mediators and activation of brown adipose tissue are also involved in the cachectic syndrome⁽ Reference Argilés, Busquets, Stemmler and López-Soriano ^{14 )}. Therefore tissues and organs such as adipose tissue, brain, liver, gut and heart are all directly involved in the cachectic process.

In 2011, a consensus definition of CC was published by Fearon et al.⁽ Reference Fearon, Strasser and Anker ^{32 )} which proposed new criteria for diagnosing CC. Briefly this definition requires any one of the following criteria to be met: (1) involuntary weight loss >5 % over the last 6 months in the absence of simple starvation; (2) weight loss >2 % with a BMI < 20 kg/m²; (3) weight loss >2 % with an appendicular skeletal muscle index consistent with sarcopenia as measured by dual-energy X-ray absorptiometry (DXA; males < 7·26 kg/m² and females < 5·45 kg/m²).

It is important to recognise that cachexia actually represents a spectrum of conditions. The first stage is ‘precachexia’. Here, cancer cells release substances that initiate inflammatory actions, and the body's immune cells are mounting a response to the presence of cancer cells. Patients may first notice weight loss, sometimes even before the cancer diagnosis has been made. The ‘cachexia syndrome’ is characterised by weight loss in combination with evidence of systemic inflammation and reduced food intake. People with ‘advanced/refractory cachexia’ have depletion of fat reserves, severe muscle wasting and immunocompromise, and they are likely to die primarily as a result of these issues⁽ Reference Fearon, Strasser and Anker ^{32 )}.

Whether or not a patient progresses down the pathway of worsening cachexia depends on the success of treating the primary disease and on addressing cachexia as well. Not all patients will progress down this spectrum to full cachexia. Some will die of their primary disease before they develop advanced cachexia, others will stabilise as a result of treatment of their primary disease. Such heterogeneity makes it difficult to target prophylactic therapy successfully. Prophylaxis would be best initiated in the precachexia phase yet there are few robust predictors to guide such a strategy.

Malnutrition, cancer cachexia and survival

If left untreated, tumour-related weight loss will progressively worsen. Such progression will negatively impact patient outcomes, particularly survival. There is a vast array of published studies showing reduced survival in weight-losing cancer patients and in those with cachexia. Studies have consistently demonstrated that the prognosis for cancer patients with weight loss is worse than that for weight-stable patients⁽ Reference Deans, Tan and Wigmore ^{33 –} Reference Bachmann, Heiligensetzer and Krakowski-Roosen ^{36 )}. Table 1 summarises the available evidence for reduced survival in a variety of cancers which report hazard ratios of death ranging from 1·63 to 4·2⁽ Reference Peng, Hyder and Firoozmand ^{37 ,} Reference Prado, Lieffers and McCargar ^{38 )}. A recent large paper by Martin et al.⁽ Reference Martin, Senesse and Gioulbasanis ^{5 )} on 8160 cancer patients developed a robust grading system which could predict survival and was based on percentage weight loss and BMI. This research showed that weight-stable patients with BMI > 25 kg/m² had the longest survival and percentage weight loss values associated with lower BMI categories were related to shorter survival. This study demonstrated that the loss of the ability to maintain weight (even a weight loss of 2·4 %) is significantly related to reduced survival and those cancer patients with the lowest BMI and highest level of weight loss experienced survival rates consistent with Common Terminology Criteria for Adverse Events grade 4 which is consistent with a life-limiting toxic side effect from chemotherapy⁽ Reference Martin, Senesse and Gioulbasanis ^{5 )}.

Table 1.

Sarcopenia, weight loss and cachexia; predictors of reduced survival in cancer

GI, gastrointestinal; CT, computed tomography; SMI, skeletal muscle index; L3, third lumbar vertebra; DXA, dual-energy X-ray absorptiometry; HU, Hounsfield units; Q4, fourth quartile.

Overweight and obesity in cancer

There is now convincing evidence that a substantial proportion of cancers are attributable to obesity. Indeed the World Cancer Research Fund^{( 39 )} estimates that between 15 and 45 % of cancers are directly related to obesity. Despite the fact that the majority of cancer patients present with involuntary weight loss at the time of diagnosis⁽ Reference Martin, Senesse and Gioulbasanis ^{5 )}, in the era of obesity cancer patients may not look malnourished and many in fact are overweight or obese making the identification of ‘at risk’ patients harder for medical teams caring for them. Recent studies have reported that between 40 and 60 % of cancer patients are overweight or obese, even in the setting of metastatic disease⁽ Reference Martin, Senesse and Gioulbasanis ^{5 ,} Reference Ramos Chaves, Boléo-Tomé and Monteiro-Grillo ^{40 –} Reference Martin, Birdsell and MacDonald ^{42 )} (see Table 2). As the most clinically relevant phenotypic feature of CC is muscle loss, identifying those with muscle loss becomes a huge challenge in overweight and obese patients as the loss of muscle is masked by excessive adiposity.

Table 2.

Recent large studies highlighting the problem of increased BMI in cancer patients

GI, gastrointestinal.

Sarcopenia

Sarcopenia comes from the Greek ‘sarx’ for flesh, ‘penia’ for loss and was first defined by Evans et al. in 1993⁽ Reference Evans and Campbell ^{43 )}, as ‘the age-related loss in skeletal muscle mass, which results in decreased strength and aerobic capacity and thus functional capacity’. Today more commonly it is defined as having an appendicular skeletal muscle mass (kg/height (m)²) less than two standard deviations below the mean of a young reference group⁽ Reference Baumgartner, Koehler and Gllagher ^{44 )}.

The best way to diagnose sarcopenia is by direct measurement of lean mass by either DXA or computed tomography (CT). Although DXA scans produce highly reproducible and accurate results, they lack the ability to discriminate among the lean tissue and adipose tissue sub-compartments. The third lumbar vertebra has been validated as the standard landmark for body composition analysis because in this region, skeletal muscle and adipose tissue correspond to whole body tissue quantities⁽ Reference Mourtzakis, Prado and Lieffer ^{45 ,} Reference Shen, Punyanitya and Wang ^{46 )}. CT image analysis at third lumbar vertebra can distinguish adipose tissue (including visceral, subcutaneous and intramuscular) from skeletal tissue (including psoas, paraspinal muscles, transversus abdominus, external and internal obliques, rectus abdominus)⁽ Reference Mourtzakis, Prado and Lieffer ^{45 )}.

Specific tissues are identified based on their anatomical features and then demarcated and quantified based on pre-established thresholds of Hounsfield units using commercially available imaging analysis software. CT Hounsfield unit thresholds used are −29 to 150 for skeletal muscle⁽ Reference Heymsfield, Smith and Aulet ^{47 )}, −190 to −30 for subcutaneous and intra-muscular adipose tissue⁽ Reference Heymsfield, Smith and Aulet ^{47 )} and −150 to −50 for visceral adipose tissue⁽ Reference Miller, Jones and Yanocski ^{48 )}. Individuals may be compared directly on this basis of total third lumbar vertebra skeletal muscle cross sectional area (cm²), skeletal muscle index (cm²/m²) or approximate whole-body tissue masses estimated from lumbar areas using predetermined regression equations⁽ Reference Mourtzakis, Prado and Lieffer ^{45 )}. These regression equations were first developed for a cancer population using CT scans by Mourtzakis et al.⁽ Reference Mourtzakis, Prado and Lieffer ^{45 )} who converted the cut points developed in a healthy cohort by DXA⁽ Reference Baumgartner, Koehler and Gllagher ^{44 )} from kg/m² to cm²/m² using regression equations. Since then several cut points for sarcopenia have been published⁽ Reference Fearon, Strasser and Anker ^{32 ,} Reference Prado, Lieffers and McCargar ^{38 )} and more recently sex-specific and BMI-specific cut points for sarcopenia were defined by Martin et al. (<43 cm²/m² for men (BMI < 25 cm²/m²), <53 cm²/m² for overweight/obese men and <41 cm²/m² for women)⁽ Reference Martin, Birdsell and MacDonald ^{42 )}.

Despite the presence of increasing prevalence of overweight and obesity in cancer patients, sarcopenia is present in between 20 and 70 % depending on the tumour type and sarcopenia definition used. There are over twenty-five studies in the literature reporting rates of sarcopenia in different cancer populations but studies are difficult to compare due to the variety of cut-off points for sarcopenia that were used. Of these  twenty-five studies, ten used the cut points according to Prado et al.⁽ Reference Prado, Lieffers and McCargar ^{38 )} (38·5 cm²/m² for females and 52·4 cm²/m² for males); five used the Baumgartner cut points⁽ Reference Baumgartner, Koehler and Gllagher ^{44 )} which were converted to cm²/m² using Mourtzakis regression equations⁽ Reference Mourtzakis, Prado and Lieffer ^{45 )} (38·9 cm²/m² for females and 55·4 cm²/m² for males); two used Martin et al.⁽ Reference Martin, Birdsell and MacDonald ^{42 )} (<43 cm²/m² for men (BMI < 25 cm²/m²), <53 cm²/m² for overweight/obese men and <41 cm²/m² for women), and the remainder used either Fearon et al.⁽ Reference Fearon, Strasser and Anker ^{32 )} (<55 cm²/m² for men and <39 cm²/m² for women) or were based on one muscle group⁽ Reference Peng, Hyder and Firoozmand ^{37 )}, quartiles of muscle mass⁽ Reference Prado, Lieffers and McCargar ^{38 )} or optimal stratification analysis⁽ Reference Miyamoto, Baba and Sakamoto ^{49 )}. Thus, cross-comparison of rates of sarcopenia across different studies is hampered. Fig. 1 describes the rate of sarcopenia in fifteen different cancer populations using the most commonly used definitions⁽ Reference Prado, Lieffers and McCargar ^{38 ,} Reference Mourtzakis, Prado and Lieffer ^{45 )}. Figures are compared with a reference healthy population of 378 adults who took part in the New Mexico Elder Health Survey between 1993 and 1999 and had their body composition measured by DXA⁽ Reference Gurney ^{44 )}. As can be seen in Fig. 1 the majority of the studies describing sarcopenia in oncology populations place cancer patients akin to those healthy elderly aged late 70s to early 80s.

Fig. 1.

(Colour online) Summary of studies reporting sarcopenia in a variety of cancers compared with healthy controls. For cancer populations sarcopenia defined by computed tomography scan using Prado et al.⁽ Reference Prado, Lieffers and McCargar ^{38 )} values of 38·5 cm²/m² for females and 52·4 cm²/m² for males; Mourtzakis et al.⁽ Reference Mourtzakis, Prado and Lieffer ^{45 )} values of 38·9 cm²/m² for females and 55·4 cm²/m² for males. Reference ranges for healthy population based on appendicular lean mass as assessed by dual-energy X-ray absorptiometry of <5·45 kg/m² for females and <7·26 kg/m² for males.

As discussed previously, an additional challenge for cancer care providers is the fact that sarcopenia is a phenomenon that may be obscured within the bulk of body weight and body weight change and this is now recognised as a clinically important phenomenon. Fig. 2 depicts several CT scans at lumbar vertebra 3 in cancer patients. It shows the variation (Fig. 2(a)) for four males with identical amounts of skeletal muscle. Fig. 2(b) highlights skeletal muscle variation in three female patients with identical BMI.

Fig. 2.

(Colour online) (a, b) CT scans at lumbar vertebrae 3. Fig. 2(a) shows four male sarcopenic patients with identical skeletal muscle index (SMI) ranging across different BMI categories. Fig. 2(b) shows three female patients with identical BMI but varying SMI. Muscle depicted in red in all images.

Body composition, sarcopenia and dosing of chemotherapy

Body composition has emerged as an important predictor of anti-cancer drug efficacy and toxicity. Historically anti-cancer treatment is dosed using body surface area (BSA) which is calculated using the formula; BSA = 0·007184 × weight^0·425 × height^0·725. This equation was published by du Bois and du Bois in 1916; however, the objective at this time was not to develop a formula to dose anticancer agents⁽ Reference Van Vledder, Levolger and Ayez ^{50 ,} Reference Du Bois and Du Bois ^{51 )}. It was first introduced in oncology in order to derive a safe starting dose, and the use of BSA today as a means to individualise the dose of chemotherapy has been questioned⁽ Reference Martin, Birdsell and MacDonald ^{42 )}. There is growing evidence that this approach is invalid⁽ Reference Gurney ^{52 )} as there can be a 4–10-fold variation in drug clearance with medications like chemotherapy which have a narrow therapeutic index.

Many cytotoxic drugs are largely metabolised and excreted by the liver and kidney and BSA is not a good indicator of this function⁽ Reference Gurney ^{52 )}. Renal function, if based on serum creatinine concentration, may be overestimated in sarcopenic patients which may result in overdosing of renally excreted drugs, such as carboplatin⁽ Reference Takimoto ^{53 )}. There is growing literature to suggest that lean body mass or fat free mass (FFM), which is mainly composed of skeletal muscle and metabolic tissues such as liver and kidney⁽ Reference Trobec, Kerec Kos and Von Haeling ^{54 )}, may be a better basis for normalising drug dosages in cancer patients, especially of hydrophilic drugs⁽ Reference Roubenoff and Kehayias ^{55 ,} Reference Morgan and Bray ^{56 )}. Likewise, increased adipose tissue may increase volume of distribution for highly lipophilic drugs prolonging their elimination half-lives.

The reason for the increased toxicity may be as a result of inappropriate dosing of chemotherapy drugs based on BSA, which has a weak correlation to whole body FFM. Prado et al.⁽ Reference Prado, Lieffers and McCargar ^{38 )} reported that individual variation in FFM could account for up to three times variation in effective volume of distribution for chemotherapy administered per unit BSA in sarcopenic obese patients with solid tumours of the respiratory and gastrointestinal tracts. Thus, administering the same dose of chemotherapy drug in a patient with low FFM compared with a patient with normal FFM would increase the risk of chemotherapy toxicity⁽ Reference Prado, Lieffers and McCargar ^{38 )}.

The discrepancy between BSA and body composition has recently been investigated by Stobäus and co-workers⁽ Reference Stobaus, Kupferling and Lorenz ^{57 )} in a cross-sectional study of 630 cancer patients. They reported a significant discrepancy between BSA and body composition with more than 30 % of patients differing considerably from the established mean of their respective BSA category⁽ Reference Stobaus, Kupferling and Lorenz ^{57 )}. This implies that using the BSA in these patients under and over estimates body composition, which in theory leads to an over and under dose of chemotherapy. The extent of lean tissue loss in cancer patients varies from individual to individual and by basing chemotherapy dose on BSA the patient may experience increased toxicity particularly with hydrophilic drugs.

Impact of sarcopenia on tolerance to cytotoxic chemotherapy

Variability in body composition of cancer patients may be a source of disparities in the metabolism of cytotoxic agents. Identification of those with sarcopenia is important because sarcopenia is associated with elevated toxicity to chemotherapy⁽ Reference Prado, Lieffers and McCargar ^{38 ,} Reference Prado, Lima and Baracos ^{58 –} Reference Prado, Baracos and McCargar ^{62 )}. There is mounting evidence that sarcopenia increases the risk of toxicity to many drugs including Epirubicin⁽ Reference Prado, Lima and Baracos ^{58 )}, Capecitabine⁽ Reference Prado, Baracos and McCarger ^{63 )}, Sunitinub⁽ Reference Cushen, Power and Teo ^{59 )}, Sorafinib⁽ Reference Antoun, Baracos and Birdsell ^{60 ,} Reference Mir, Coriat and Blanchet ^{61 )}, 5-FU and leucovorin⁽ Reference Prado, Baracos and McCargar ^{62 )}. Table 3 summarises the available evidence for sarcopenia and dose-limiting toxicity in a variety of cancers and different chemotherapy drugs. Future clinical trials investigating dose reductions in patients with sarcopenia and dose-escalating studies based on pre-treatment body composition assessment have the potential to alter cancer treatment paradigms⁽ Reference Prado ^{64 )}.

Table 3.

Summary of studies in various cancers showing toxicity in a variety of chemotherapy drugs according to sarcopenic status

NS, non-significant; 5-FU, 5-fluorouracil; FP, fluoropyramidine.

In addition to the argument that pharmacokinetic parameters can explain the higher risk of toxicity in sarcopenic patients it is also important to note that sarcopenic patients are excessively fragile and highly susceptible to acute medical events that exacerbate chemotherapy-related toxicity⁽ Reference Antoun, Borget and Lanoy ^{65 )}. Mechanisms outside the pharmacokinetic hypothesis linking sarcopenia to toxicity include metabolic alterations found in cancer or the role of systemic inflammation. The latter has been shown to decrease liver cytochrome activities and drug clearance and may modify drug exposure⁽ Reference Alexandre, Rey and Girre ^{66 )}. Inflammation can also increase the deposition of fat intramuscularily which then leads to reduced muscle density. Skeletal muscle density is assessed by analysing pixel values on CT images with lower density reflecting fatty muscle infiltration⁽ Reference Goodpaster, Kelley and Thaete ^{67 )}. This adds a qualitative dimension to the measurement of body composition.

In more recent years, the importance of muscle density has come to the forefront with much recent research highlighting its importance in progression-free survival, treatment response and overall survival. Higher skeletal muscle density has been associated with improved survival in renal cell carcinoma⁽ Reference Prado, Lieffers and McCargar ^{38 )}, melanoma⁽ Reference Sabel, Lee and Cai ^{68 )} and in a number of others solid tumours⁽ Reference Martin, Birdsell and MacDonald ^{42 )}. Skeletal muscle density could potentially be a more accurate measure of muscle function and therefore precede the development of sarcopenia. Low skeletal muscle density may be reflective of patients with a lower performance status.

Sarcopenia and survival

It has been reported that sarcopenia is independently prognostic of lower survival in obese patients with solid tumours of the lung and gastrointestinal tract. Prado et al.⁽ Reference Prado, Lieffers and McCargar ^{38 )} reported that patients suffering from sarcopenic obesity had a shorter survival. The median survival was reduced in sarcopenic obese patients to 11·3 (95 % CI 7·4, 15·2) months compared with 21·6 (95 % CI 16·9, 26·3) months in non sarcopenic obese patients (P < 0·0001)⁽ Reference Prado, Lieffers and McCargar ^{38 )}. Sarcopenia has also been identified as a predictor of overall survival after cytoreductive nephrectomy for metastatic renal cell carcinoma, following pancreatic surgery and following curative resection of colorectal cancer⁽ Reference Peng, Hyder and Firoozmand ³⁷ ,Reference Antoun, Baracos and Birdsell ⁶⁰ ,Reference Sharma, Zargar-Shoshtari and Caracciolo ^{69 )}.

Martin et al.⁽ Reference Martin, Birdsell and MacDonald ^{42 )} showed similar results. This is a study with a larger sample size of cancer patients (n 1473). The most striking feature of this analysis is the shorter overall survival in patients with weight loss, sarcopenia and low muscle density 8·4 months (95 % CI 6·5, 10·3) compared with an overall survival of 28·4 months (95 % CI 24·2, 32·6, P < 0·001) in other patients regardless of body weight⁽ Reference Martin, Birdsell and MacDonald ^{42 )}.

Oral nutritional supplements

Unfortunately for those with established cachexia a meta-analyses of thirteen randomised controlled trials have concluded that weight, body composition and functional outcomes fail to improve with oral nutritional supplements alone⁽ Reference Baldwin, Spiro, Ahern and Emery ^{88 )}. The clinical characteristic of cachexia is that it cannot be successfully treated with nutrition alone as there is partial block to the accretion of lean tissue in cancer patients. The use of anti-inflammatory agents (either drug agents or nutritional agents such as EPA) have therefore been advocated in the multi-pronged approach to care⁽ Reference de Aguiar Pastore Silva, Emilia de Souza Fabre and Waitzberg ^{89 ,} Reference Ryan, Reynolds and Healy ^{90 )}.

Physical activity

Efforts should be made to encourage physical exercise (within the capacities of the patient) as a means of preserving and restoring muscle mass and to reduce systemic inflammation⁽ Reference Aapro, Arends and Bozzetti ^{77 )}. The optimum means of achieving this is not yet clear but is the subject of major ongoing studies. Studies targeting cachectic patients have demonstrated that even in advanced disease peripheral muscle has the capacity to respond to exercise training⁽ Reference Maddocks, Murton and Wilcock ^{91 )}. Benefits of exercise include enhancing muscle protein synthesis, attenuating the catabolic effects of cachexia, and modulating levels of inflammation. However, there are challenges and limitations to cachectic patients engaging in exercise and many are not able or willing to undertake programmes.

Multimodal approaches

As multi-factors are responsible for the development of cachexia, it has recently been argued that optimal cachexia intervention should target all components; multimodal therapy for a multidimensional problem⁽ Reference Fearon ^{29 ,} Reference Solheim and Laird ^{92 ,} Reference Balstad, Kaasa and Solheim ^{93 )}. This multimodal approach would incorporate oral nutritional supplements, exercise and anti-inflammatory medications and has been advocated in the prevention and treatment of CC. Underpinning this approach are meta-analyses suggesting that each of the interventions components, while not unequivocally associated with improved outcome, is more likely than not to be beneficial⁽ Reference Aapro, Arends and Bozzetti ^{77 ,} Reference Baldwin, Spiro, Ahern and Emery ^{88 ,} Reference Maddocks, Murton and Wilcock ^{91 ,} Reference Ries, Trottenberg and Elsner ^{94 ,} Reference Solheim, Fearon, Blum and Kaasa ^{95 )}.

The MENAC trial (a multimodal intervention of exercise, nutrition and anti-inflammatory medication plus standard care v. standard care alone) to prevent/attenuate cachexia will test the efficacy of a multi-dimensional approach during chemotherapy. Patients with advanced lung and inoperable pancreatic cancer about to start chemotherapy will be randomised to standard care or standard care plus oral nutritional supplements enriched with EPA, exercise and non-steroidal anti-inflammatory drugs. The primary objective is to determine whether the intervention achieves a difference in skeletal muscle index at 6 weeks.

Adequate protein alone only slows the loss of muscle mass. Exercise (both resistance and aerobic) in combination with adequate energy and protein intake is the key component of the prevention and management of sarcopenia. Preliminary results from the MENAC trials have shown that cancer patients can gain muscle mass by eating sufficient protein and engaging in 2 × 30 min walks per week and 3 × 20 min sessions of stair climbing in their own homes (resistance exercise (KC Fearon et al., unpublished results).

Nutritional targets in cancer patients

We still know very little about the specific nutritional requirements of cancer patients. In general ranges are given; for example total daily energy expenditure in cancer patients is thought to range between 104·6 and 125·52 kJ/kg per d (25 and 30 kcals/kg per d)⁽ Reference Bozzetti ^{96 )}. The optimal nitrogen supply for cancer patients cannot be determined at present and current recommendations are based on expert opinion and range between a minimum of 1 g/kg per d and a target supply of 1·2–2 g/kg per d⁽ Reference Bozzetti ^{96 –} Reference Baracos and Mantavani ^{99 )}. As we have argued, in another article on this subject⁽ Reference Prado, Cushen, Orsso and Ryan ^{100 )} lean body mass drives protein requirements, and current recommended ranges for energy and protein do not take into account body composition and the likely important role that varying degrees of wasting or obesity (or both sarcopenic obesity) have on requirements⁽ Reference Prado, Cushen, Orsso and Ryan ^{100 ,} Reference Prado and Heymsfield ^{101 )}. Therefore targeted energy and protein recommendations should be used, so that protein intake is targeted to protect muscle mass and energy recommendations are calculated according to the overweight or obesity status⁽ Reference Prado, Cushen, Orsso and Ryan ^{100 )}. For those patients who are obese (between 17 and 19 % of cancer patients) there are no specific guidelines as to the energy intake or the target body weight. It is unclear whether cancer patients experiencing obesity related morbidity benefit from maintenance of their heavy body weight or whether some (limited) degree of weight loss could in some way be desirable. However, separate from energy considerations it is reasonable to evaluate protein intake and increase it with the purpose of preventing erosion of lean body mass⁽ Reference Gioulbasanis, Martin and Baracos ^{41 )}.