Introduction
Cancer-related fatigue (CRF) can occur at all stages of treatment and increases in advanced disease. It affects the majority of patients with cancer. As a symptom it can limit the intensity and duration of anti-cancer treatment, especially in older people. In more advanced illness it is one of the most troubling symptoms and can have a profound effect on functional status. In addition, as the population ages there is an increasing number of patients living with the long-term effects of treatment, which include chronic post-treatment fatigue.
There are a number of specific objectives which the review will cover:
• Defining and measuring clinically significant fatigue in the older cancer population.
• Quantifying the prevalence of CRF in the older person.
• Outlining the postulated mechanisms of CRF with particular focus on the role of inflammatory cytokines.
• Clinical guidance on managing CRF in practice.
• Future directions for research into CRF in the older person.
The search strategy used for this review is detailed in other previously published papersReference Minton, Richardson, Sharpe, Hotopf and Stone1, Reference Minton and Stone2 and can be obtained from the corresponding author. Briefly we used a list of terms for fatigue and neoplasms combined to search Medline, Embase and Cinahl from inception to week 1, November 2010.
Assessment
One of the major difficulties when examining fatigue in the older population is being clear about what the baseline level is. We know that levels of fatigue increase with age in a healthy population.Reference Loge, Ekeberg and Kaasa3, Reference Schwarz and Hinz4 The data taken from a number of cross-sectional studies demonstrate statistically significant differences in the value of fatigue scores between the youngest and oldest age groups. The general trend is that these scores increase with age, peaking in the over-70 age group. There is no specific further data breakdown in this age group.
This age-related phenomenon is likely to be in part due to the natural ageing process with loss of muscle mass and the increasing frequency of co-morbidities.Reference David, Pelosi and McDonald5, Reference Avlund, Rantanen and Schroll6 While fatigue can occur in other medical conditions frequently seen in older people, this review will focus specifically on cancer. However, most of the research has not been specifically carried out in older people. The mean age of research participants in the vast majority of studies is between 55 and 65, and so the findings will need to be extrapolated for higher age groups.
There is an increasingly clear recognition in oncology treatment that a comprehensive assessment of an older person will influence treatments that are offered.Reference Extermann and Hurria7 This will be more than simply assessing someone's performance status. Of necessity, clinical trial data underpinning treatment decisions will often be taken from younger age groups.
The same principle applies to symptom research in cancer. There is evidence highlighting increased frequency of problems associated with cancer treatment specifically in older people.Reference Balducci8, Reference Given, Given, Azzouz, Kozachik and Stommel9 The authors suggest that there is an increase in the complication rate from treatment with increasing age. They also suggest that fatigue is not associated with age per se and is highly correlated with stage of tumour and type of treatment.
These are large datasets with measures on over 1000 patients with a range of tumour types. This is an important finding that distinguishes between increased fatigue related to ageing and the specific nature of CRF.
Clinically significant fatigue
As a result of a baseline level of fatigue and the wide variety of conditions it can occur in, it is important to define severe or clinically significant fatigue in cancer. The pervasiveness of fatigue in so many medical conditions as well as in the general healthy population can lead to confusion over terminology.
There is no agreed consensus as to the definition or measurement of CRF. An illustration of this point is the number of ways CRF can be defined. It is accepted as a separate entity from tiredness experienced by the general population. It is characterized by being pervasive and not relieved by rest.Reference Cella, Lai, Chang, Peterman and Slavin10 CRF can occur at the time of diagnosis,Reference Berger and Walker11 or during surgery, Reference Cimprich12 chemotherapyReference Roscoe, Morrow, Hickok, Bushunow, Matteson, Rakita and Andrews13 or radiotherapy.Reference Geinitz, Zimmermann, Stoll, Thamm, Kaffenberger, Ansorg, Keller, Busch, van Beuningen and Molls14 CRF is also a recognized longer term phenomenon in disease-free patients.Reference Minton and Stone15 However, it is most prevalent with advancing disease.Reference Stone, Hardy, Broadley, Tookman, Kurowska and A'Hern16 It is also associated with increased levels of anxiety and depression.Reference Stone, Richards, A'Hern and Hardy17 CRF in advanced disease rarely occurs in isolation and may be part of a symptom clusterReference Donovan and Jacobsen18 although it is often rated as the most severe symptom.Reference Cleeland, Mendoza, Wang, Chou, Harle, Morrissey and Engstrom19 This is important to note as clinicians often under-rate this symptom and feel under-confident in its management.Reference Stone, Ream, Richardson, Thomas, Andrews, Campbell, Dawson, Edwards, Goldie, Hammick, Kearney, Lean, Rapley, Smith, Teague and Young20
Fatigue in cancer can be defined and measured in a number of ways.Reference Glaus21 The national comprehensive cancer network22 in the USA definition is as follows: ‘Cancer related fatigue (CRF) is a distressing, persistent, subjective sense of physical, emotional and/or cognitive tiredness or exhaustion related to cancer or cancer treatment that is not proportional to recent activity and interferes with usual functioning’. While this definition would suggest that fatigue is secondary to psychological or physical stress, this is not an absolute.Reference Storey, Waters, Hibberd, Rush, Cargill, Wall, Fallon, Strong, Walker and Sharpe23
There is a useful working definition of fatigue based on the European Organisation for Research and Treatment of Cancer (EORTC) approach. This definition was chosen in order to closely mirror the wording of the EORTC Quality of Life Questionnaire (QLQ C30),Reference Aaronson, Ahmedzai and Bergman24 which contains a three-item fatigue sub-scale addressing the symptoms of tiredness, weakness and lack of energy. This working definition of fatigue has the advantage of being simple to understand and easy to remember. The EORTC QLQ C30 has also been reliably translated into more than 50 different languages, facilitating its use in cross-cultural understanding of fatigue. This defines CRF as ‘a subjective feeling of tiredness, weakness or lack of energy’.Reference Radbruch, Strasser, Elsner, Gonçalves, Løge, Kaasa, Nauck and Stone25 However, while this is a pragmatic definition for clinicians and non-specialists alike, it does not capture all of the aspects of CRF. This lack of consensus amongst researchers has led to the development of a myriad of tools to assess CRF and has meant there is no universally agreed definition. This has meant that, in a clinical setting, definitions are often vague.Reference Ream and Richardson26
Unfortunately the same term can be used to describe both an objective physical or mental decrement in performance and a subjective mental state. Both types of fatigue are usually found to a greater or lesser extent in the same individual. The duration of these phenomena will also affect how they are interpreted by the person. However, there is no direct correspondence between the two phenomena. It is quite possible to feel extremely subjectively fatigued but to perform relatively normally on objective tests of physical or mental functioning. It is also possible to have poor physical functioning and not to complain of fatigue at all.
The most common method to identify clinically significant fatigue in clinical practice is using a simple visual analogue scale with cut-off score(s). This is recommended by the NCCN22 in the USA. The question asked is: ‘On a 0–10 scale what is your average level of tiredness over the last week, where 0 is no fatigue and 10 is fatigue as bad as you can imagine?’ A score of 4 or more is used to define clinically significant fatigue. This is a pragmatic approach but will not pick up on the complexities of the symptom.
Our group has used a cut-off score based on population normal distribution.Reference Stone, Richards, A'Hern and Hardy17 We have previously defined severe fatigue as that which is above the 95th percentile on a normal control population distribution of scores. The mean age of the control group was 68 so these data and approach are relevant to the specific assessment of elderly cancer patients. We demonstrated significant differences between the prevalence of fatigue in the control group and all tumour types and stages. The prevalence of severe fatigue increased in advanced disease in comparison with newly diagnosed patients on receiving treatment. This is an important finding as it is much more likely that a diagnosis of cancer will be made at a more advanced stage in an older patient.
Measurement
As a result of the absence of an agreed approach many tools for measuring CRF have been developed. However, most are not widely used outside of a research setting. As we have previously mentioned, it is not sufficient to ask someone about fatigue without some quantification. Assessment tools need to be population specific and cover the relevant aspects of CRF to be sensitive enough to provide a clinically meaningful measure. These scales have usually been validated originally in cancer patients. Other investigators have used scales that were originally validated in non-cancer populations and have then independently validated their use in patients with cancer. These scales can differ widely in the number of items that they contain, the dimensions of CRF that they cover (e.g. physical, affective and cognitive) and their psychometric properties.
We have previously conducted a systematic review of the scales used to measure CRF.Reference Minton and Stone2 The full details of search methodology can be found in the original paper but we used a comprehensive range of search terms to interrogate a number of electronic databases. An ad hoc quality score was created prior to searching to ensure that included scales reached a minimum level of psychometric properties based on the ideal characteristics of measurement scales described in detail by Norman and Streiner.Reference Streiner and Norman27 The paper describing the scale must have contained information on at least three of the following: internal consistency; test–retest reliability; known group validity (discriminant validity); responsiveness to change or convergent validity (against other scales).
The scales were divided into unidimensional and multidimensional for ease of presentation. This division is possibly artificial as CRF has been shown in a recent study to be a unidimensional construct, even after measuring all relevant dimensions of fatigue.Reference Lai, Crane and Cella28
Overall fourteen scales were identified that met the minimum criteria.
The most widely used of these scales are the functional assessment of cancer therapy–fatigue subscale (FACT F)Reference Cella29 and the EORTC QLQ C30 fatigue subscale, which have data from over 10,000 patients between them and have been widely used in intervention studies to treat CRF.Reference Minton, Richardson, Sharpe, Hotopf and Stone1 The FACT F has the advantage of having a validated clinically significant change score – it also covers the social impact of CRF but takes longer to administer than the 3-item EORTC QLQ C30 fatigue subscale. The former should be used in a research setting, the latter could be used to monitor clinical effects. As yet there is no clinical value in distinguishing between patients with predominantly ‘physical’, ‘cognitive’ or ‘motivational’ fatigue. Until the clinical importance of these proposed ‘dimensions’ have been clarified there is little incentive to use these lengthier scales outside of a research study.
The lack of agreement about how to measure and define CRF has led to the development of a diagnostic interview to identify cases of cancer-related fatigue syndrome – CRFS.Reference Cella, Peterman, Passik, Jacobsen and Breitbart30 This approach is based on that taken in chronic fatigue syndrome (CFS).Reference Fukuda, Straus, Hickie, Sharpe, Dobbins and Komaroff31 A structured interview is used to determine if the relevant criteria are met. This is a 10-item interview and takes approximately 10–15 minutes to complete.
In order to be diagnosed as a ‘case’ of CRFS the following criteria must be met:
• For at least 2 weeks in the last month patients must have ‘significant fatigue, lack of energy, or an increased need to rest’ (criterion A).
• To fulfil these diagnostic criteria the subject must have experienced at least six out of ten fatigue-related symptoms on most days or every day for 2 weeks in the previous month (criterion B).
• There must be evidence that the symptoms are a consequence of cancer or cancer therapy (criterion C).
• Finally, the fatigue should not be ‘primarily a consequence of co-morbid psychiatric disorders such as major depression’ (criterion D).
This last criterion requires the co-administration of a structured psychiatric interview based on the criteria listed in the diagnostic and statistical manual (DSM) of mental disorders,32 to rigorously exclude any psychological co-morbidity contributing to fatigue. This type of procedure is a semi-structured interview based on a defined schedule that requires training to administer and interpret correctly. It allows one to rule out or diagnose any psychiatric condition in the DSM. The average interview is between 30 and 45 minutes. This is the most rigorous method for assessing criterion D and allows one to tightly define two groups – i.e. those that do and do not meet the criteria for CRFS. It is, however, a time-consuming method and unsuitable for use outside of a research setting. This has limited its use to a small number of studies. The prevalence of the fatigue syndrome has varied from 10% in women with breast cancer prior to adjuvant chemotherapyReference Andrykowski, Schmidt, Salsman, Beacham and Jacobsen33 to 56% in a group of heterogeneous cancer in-patients.Reference Fernades, Stone, Andrews, Morgan and Sharma34 The mean age of patients was around 60 in both studies, although both included a significant number above the age of 70. These figures are therefore equally likely to apply to an older population.
Cella and colleagues originally suggestedReference Cella, Peterman, Passik, Jacobsen and Breitbart30 that fatigue attributable to cancer diagnosis and treatment tends to be characterized by less prominent cognitive symptoms, onset coincident with cancer diagnosis and treatment initiation, greater synchronization of symptoms with receipt of adjuvant therapy, and little diurnal fluctuation in symptoms. In practice, this can be very difficult to objectively interpret.
One advantage of using this rigorous method is that questionnaires can be co-administered to be subsequently used as a screening tool for CRF. Van Belle and colleaguesReference Van Belle, Paridaens, Evers, Kerger, Bron, Foubert, Ponnet, Vander Steichel, Heremans and Rosillon35 used the FACT F on a mixed cancer population. The psychometric data suggest this technique could be more widely used. Screening using a cut-off score of 34 out of a possible 52 on the FACT F provided 75% specificity and 91% sensitivity for meeting criteria for CRFS.
Our own group used the same technique in breast cancer patients only.Reference Alexander, Minton and Stone36 Screening with FACT F, again using a cut-off score of 35/52, gave a sensitivity of 80%, specificity of 71%, for meeting the CRFS criteria. The similarity in findings across tumour types suggests this could be more widely used for identifying clinically significant CRF. It is a more robust method than simply using a visual analogue scale (VAS), especially in older people where the baseline fatigue will be higher. This is important in defining a clear clinical categorization in order to investigate the underlying mechanisms of CRF.
CRF and associated symptoms
Up to this point we have only focused on the clinical aspects of CRF in isolation. However, there is significant debate as to whether CRF is merely secondary to other conditions. This is certainly relevant as CRF does not occur alone but is associated with other clinical problems.
Prue and colleaguesReference Prue, Rankin, Allen, Gracey and Cramp37 reported that most studies of CRF have found little evidence to relate fatigue severity to basic demographic, disease or treatment-related factors. There is some evidence that fatigue is more common in patients receiving chemotherapy and with more advanced stages of cancer. There is no causal association seen with depression or sleep disturbance. However, the authors cautioned that methodological weaknesses in the literature prevented general conclusions being drawn.
We have suggested throughout the literature review that fatigue and depression are often associated but that no causal relationship was found between the two. Depression, like fatigue, can be both a symptom (of low mood and the associated feelings) and also be described as a distinct syndrome – that of a major depressive illness – based on distinct criteria. The relationship between fatigue and depression is complex. This point is illustrated by data from a large scale study examining CRF. Storey and colleaguesReference Storey, Waters, Hibberd, Rush, Cargill, Wall, Fallon, Strong, Walker and Sharpe23 examined nearly 3000 mixed cancer patients using the EORTC QLQ 30 fatigue subscale and the hospital anxiety and depression scale (HADS).Reference Zigmond and Snaith38 Clinically significant psychological distress was determined using a cut-off score of >15 on the HADS total. CRF was determined using a score >4/10 on a VAS. The prevalence of fatigue using this method was 32%, with just under half of the patients having significant psychological distress. The authors concluded that ‘the association between CRF and emotional distress is strong but they are not equivalent conditions’.
There is similar debate (to depression) as to whether CRF is secondary to sleep disturbance. Insomnia affects between 30 and 50% of newly diagnosed cancer patients and persists for several years after treatment.Reference Savard and Morin39 A similar large scale study to the measurement of CRF and depression was conducted in sleep disturbances.Reference Palesh, Roscoe, Mustian, Roth, Savard, Ancoli-Israel, Heckler, Purnell, Janelsins and Morrow40 The authors examined 923 mixed cancer patients for insomnia. In this study 43% met the criteria for insomnia syndrome. A comparison between this group and the ‘good sleepers’ demonstrated significant differences in fatigue and depression scores. This certainly suggests a relationship between these symptoms but does not provide a causal link.
These are important co-variables to quantify in any examination or assessment of CRF. Although many studies have reported on the correlates of CRF and have speculated on the cause of this symptom, there have been very few studies that have tested specific hypotheses about the underlying mechanism.
Postulated mechanisms of CRF
Despite all the clinical descriptors of this group, the underlying biological mechanisms are poorly defined and understood. This stems from the lack of agreement in defining CRF and its associations. The potential mechanisms causing CRF can be classified as central, peripheral and immunological, although there is a large overlap between them.
Peripheral mechanisms
One of the most common associations examined has been the contribution of exercise tolerance to subjective fatigue. Physiology studies have tended to focus on alterations in maximal oxygen capacity (VO2max). This is defined by the Fick equation: cardiac output multiplied by the difference between oxygenation between arterial and venous blood and usually expressed in terms of oxygen consumed per minute. There are then two important factors which determine its level. The first is the central capacity of the organism to deliver oxygen to skeletal muscle – which is based on circulating volume and oxygenation of blood. The second is the capacity of skeletal muscle to extract and use oxygen.
The first factor has led to much research examining the role of anaemia in CRF.Reference Bohlius, Wilson, Seidenfeld, Piper, Schwarzer, Sandercock, Trelle, Weingart, Bayliss, Brunskill, Djulbegovic, Benett, Langensiepen, Hyde and Engert41 However, it is unclear how the impact of anaemia affects fatigue.Reference Jones, Schenkel, Just and Fallowfield42 The authors of this meta-analysis showed that only 20% of the variance of fatigue could be explained by a haemoglobin concentration of less than 10 g/dl. Many studies have reported a significant (although modest) association between fatigue severity and haematosis.Reference Prue, Rankin, Allen, Gracey and Cramp37
There has been little research focusing on the role of skeletal muscle architecture and CRF. Instead, the focus has been on muscle wasting per se. There is work in related fields such as CFS.Reference De Lorenzo, Xiao, Mukherjee, Harcup, Suleiman, Kadziola and Kakkar43 The authors suggested that deconditioning leads to muscle atrophy which may be reversible with exercise training. However, muscle changes only appear to occur in a minority of patients (approximately 10%).Reference Lane, Barrett, Woodrow, Moss, Fletcher and Archard44 However, there are parallels with CRF and the potential for deconditioning that can occur during treatment. This is also a feature that occurs as part of the natural ageing process.Reference Evans and Lambert45 All of these factors may impair exercise capacity, one of the reasons why so much research has been done in this area. This is especially important when it comes to potential treatments for CRF (of which exercise is one of the main modalities) which will be discussed later.
Central mechanisms
The relationship between the hypothalamic–pituitary axis (HPA) and possible causes of fatigue has been extensively studied in CFS.Reference Cleare46 The HPA is felt to be responsible for the regulation of circadian rhythm and has interactions with the immune system and brain neurotransmitter activity.Reference Sanders and Korf47 The extensive field of research in CFS provides contradictory data on the type and level of HPA disruption.Reference Cleare, Blair, Chambers and Wessely48 This is reflected in the small number of studies conducted investigating this relationship in cancer.
The work carried out in cancer has only been done in women successfully treated for breast cancer, although with an increasing number of cancer survivors this represents a significant number of patients. The applicability of this work to patients with advanced disease remains unclear.
In one studyReference Bower, Ganz, Dickerson, Petersen, Aziz and Fahey49 the authors examined 24-hour salivary cortisol variation. They found some flattening of the HPA response but no marked abnormality.
In another studyReference Bower, Ganz and Aziz50 the authors examined the same group using a social stress test. This involved public speaking and mental arithmetic and had been previously associated with a rise in cortisol in healthy volunteers. They demonstrated a flattened cortisol response to this defined stimulus, suggesting suppression of the HPA axis. However, a more recent study conducted by our groupReference Alexander, Minton, Andrews and Stone51 failed to demonstrate any difference in 24-hour urinary cortisol output in relation to fatigue.
Immunological mechanisms
The majority of ‘mechanistic’ studies have examined immunological disturbances in relation to fatigue. There is no agreement on which cytokines should be measured, which method is chosen for analysis and which measure of fatigue is used. A narrative review of available studies appears to show a consistent association with interleukin-6 and possibly C reactive protein (CRP). In addition, there is an alteration in white count levels, especially lymphocytes. This does suggest that a non-specific immune dysfunction (mediated by B cells) may contribute to CRF. However, with the wide experimental variation, this is far from certain.
The clear advantage of this avenue of research stems from the fact that many of these analytical tests for cytokine levels are either commercially available or can be conducted by the local pathology department. The necessity of making an a priori choice of a small number of tests is the major weakness of this approach. The choice of cytokines to include on the panel is likely to be determined by researcher preference, local resources and available funding rather than being hypothesis driven. There is no limit to how many of these analyses could be conducted, but all such approaches have a significant cost.
There is therefore much debate in the literature as to which cytokines are genuinely associated with CRF.
A recent systematic review and meta-analysisReference Schubert, Hong, Natarajan, Mills and Dimsdale52 found a consistent positive relationship between fatigue and a number of cytokines, but failed to confirm many of the findings of single studies. The authors studied all tumour types at all stages including post-treatment and metastatic disease. They included eighteen studies with over 1000 participants. They cited the wide range of outcome measures and experimental variation in different studies as factors that may explain the lack of consistent findings. Nonetheless, the authors concluded that overall there was evidence of an association between inflammatory markers and fatigue in cancer patients, with the best evidence being for a relationship between fatigue and interleukin-6, interleukin-1a and neopterin (secreted by activated macrophages). All of these cytokines are associated with a prolonged inflammatory response and would fit with our current understanding of the pathophysiology of CRF. However, more work is clearly needed before firm conclusions can be drawn about the relevance of any particular cytokine/inflammatory marker.
One of the major limitations of this type of approach, in general, is that it has focused on one postulated mechanism. It is more than likely there is no single mechanism for post-treatment CRFSReference Andrews, Morrow, Hickok, Armes, Krishnasamy and Higginson53 and that research in this area should examine a number of postulated mechanisms, rather than narrowly focusing on just one. Many published studies have focused on only a handful of cytokines or proteins for analysis. Identification of new proteins and quantification of specific known proteins could be used to generate new hypotheses about CRF mechanisms.Reference Minton and Stone54
Treatment options
In spite of the large number of descriptive studies, potential treatments for the problems illustrated are limited. There have been, however, three recent Cochrane reviews that summarize the evidence for the different treatment modalities. These reviews are systematic and rigorous in their conduct. An exhaustive set of electronic search terms and hand searching of journals is part of the typical process. If it is statistically possible then the data extracted from the included studies are combined quantitatively in a meta-analysis. This is done using the Cochrane software.55 This process ensures that the conclusions are based on objective measures, and is superior to a narrative review.
Our group conducted a review of the drug treatment for CRFReference Minton, Richardson, Sharpe, Hotopf and Stone56 and there have been two complementary reviews examining the role of exerciseReference Cramp and Daniel57 and psychological interventions.Reference Goedendorp, Gielissen, Verhagen and Bleijenberg58 We will now briefly summarize the main findings of each review.
Drug therapy
We analysed all randomized controlled trials (RCTs) of the pharmacological treatment of CRF. Eligible studies included a multi-item fatigue scale as one of the outcome measures. Our results highlighted a lack of well-designed RCTs and robust outcome criteria. Thirty-five trials met our full inclusion criteria. Six different classes of drugs were identified: haematopoietic growth factors, psychostimulants, bisphosphonates, anti-TNFα antibodies, antidepressants and progestational steroids. There were data on over 7000 participants in total. When appropriate, meta-analyses were undertaken by drug class. These provided evidence for the use of three agents: methylphenidate (a psychostimulant), erythropoietin and darbopoetin (both haemopoetic growth factors).
Five studies have been undertaken using methylphenidate. The meta-analysis did show a statistically significant effect size and no increase in adverse events. However, the total number of patients included in these studies was small (approximately 400) and larger RCTs will be needed before methylphenidate could be recommended as a treatment for CRF. While these drugs may be useful for patients undergoing treatment, they cannot be recommended for cancer survivors.Reference Minton, Richardson and Sharpe59
The evidence for the effectiveness of erythropoietin and darbopoetin was much more robust and has been supported by other systematic reviews.Reference Bohlius, Wilson, Seidenfeld, Piper, Schwarzer, Sandercock, Trelle, Weingart, Bayliss, Brunskill, Djulbegovic, Benett, Langensiepen, Hyde and Engert41, Reference Jones, Schenkel, Just and Fallowfield42 Both agents appear to improve CRF in anaemic patients (haemoglobin <12 g/dl). Most studies have been conducted in patients undergoing chemotherapy. Cancer-related anaemia (from whatever cause) does not explain CRF entirely. Most benefit is seen with correction of anaemia between 8 and 10 g/dl. Only 20% of any potential improvement in CRF is explained at higher levels.Reference Jones, Schenkel, Just and Fallowfield42 The potential side-effects of these drugs (most notably thrombo-embolic events) may also limit their routine use. A further reviewReference Bohlius, Schmidlin, Brillant, Schwarzer, Trelle, Seidenfeld, Zwahlen, Clarke, Weingart, Kluge, Piper, Rades, Steensma, Djulbegovic, Fey, Ray-Coquard, Machtay, Moebus, Thomas, Untch, Schumacher, Egger and Engert60 has demonstrated a worse outcome and poorer survival for patients being treated with these drugs. Therefore, these drugs cannot be recommended for treating CRF and have no indication in the older person especially.
One isolated result did support the hypothesis that an inflammatory process may be associated with CRF. One trialReference Monk, Phillips, Waite, Kuhn, Schaaf, Otterson, Guttridge, Rhoades, Shah, Criswell, Caligiuri and Villalona-Calero61 examined the effect of etanercept (a tumour necrosis factor blocker) on 12 patients with mixed tumour types undergoing docetaxel-based chemotherapy. This was an open label study with six patients randomized to receive an additional etanercept infusion. The authors of the trial reported a statistically significant treatment effect on CRF.
The review also demonstrated some important negative results. There was no improvement seen in CRF with progestational steroids or with anti-depressants (paroxetine and sertraline). This suggests that, although fatigue is often associated with depression or cachexia, improvements of appetite or mood do not necessarily result in improvements in CRF.
Exercise
The authors of this Cochrane review examined the role of exercise for improving quality of life, physical functioning and fatigue in patients with cancer.Reference Cramp and Daniel57 The trials in this area have been a mixture of resistance and cardiovascular training and trials differ extensively in terms of frequency, intensity and duration. Twenty-eight studies were identified for inclusion (over 2000 participants), with the majority carried out in breast cancer (16 studies). A meta-analysis of all fatigue data was conducted. At the end of the intervention period exercise was statistically more effective than the control intervention. This overall conclusion does mask some of the problems with the quality of the studies and outcome measures used.
The studies varied widely in terms of exercise regimen and control procedures. There was potential for contamination between groups in the vast majority of studies. The measurement of CRF was often a secondary measure behind an improvement in exercise tolerance. It is unclear from this review how these two outcomes are related. The majority of studies demonstrated an improvement in physical functioning but only a subset showed an improvement in CRF. This is confirmed by a complementary Cochrane review examining the role of exercise in breast cancer patients receiving adjuvant therapy.Reference Markes, Brockow and Resch62 The authors conducted a meta-analysis and demonstrated an improvement in cardiorespiratory fitness but no associated improvement in fatigue symptoms. Overall, the effect size of exercise was small, and was less than that of the methylphenidate studies.
While this type of intervention might seem the most appropriate for the older person, it is unclear what the optimum ‘dose’, type or frequency of exercise should be. There are potentially fewer safety concerns with the use of an exercise intervention than with pharmacotherapy but care still needs to be taken.
Psychosocial interventions
Broadly speaking, psychosocial interventions include psychological, educational and support group studies.Reference Goedendorp, Gielissen, Verhagen and Bleijenberg58 There is a wide variation in study size, design and quality. By their very nature, these studies have complex designs and often used mixed methods. These studies were a heterogeneous category that included cognitive behavioural therapy, supportive and expressive therapies, and psycho-education. Examples of studies ranged between simple interventions such as telephone support to more complex individualized cognitive behavioural therapy sessions. The interventions were carried out both on individuals and in group settings. A large number of studies were conducted on breast cancer patients.
This review is more difficult to interpret than the other two reviews. The authors included all studies examining psychosocial interventions where fatigue was an outcome measure. The result is that, while 27 studies were included, only five studies specifically examined the effect on CRF. These five interventions specific for fatigue were short, consisting of three individual sessions, varying between 10 and 60 minutes, provided by oncology nurses. In general, during these interventions, patients were educated about fatigue, were taught in self-care or coping techniques, and activity management was part of the intervention. As a result, it was not possible to combine these studies in any analysis because of the wide variety of content. This may be an effective treatment approach but is clearly resource intensive and needs to be directed at the most appropriate group, i.e. those who are most fatigued.
The authors of the Cochrane review do confirm the specific nature of CRF and the need for more focused treatments. Out of the remaining 22 studies, only three were effective in reducing fatigue, but these interventions were not specifically designed for CRF. These interventions were aimed at psychological distress, mood and physical symptoms, and varied widely in content.
It is difficult to make treatment recommendations based on the current evidence. There are some annually updated guidelines from the national comprehensive cancer network in the USA.22 However, these are also heavily based on expert recommendation and so must be used with caution. It is increasingly clear that more targeted treatments are likely to be more useful. It would also be helpful to have an objective outcome measure to monitor the effects of more complex interventions. A better understating of the underlying mechanisms of CRF will be vital to achieve these aims.
Conclusions for practice and future research
In clinical practice there is a need to recognize and assess the impact of this distressing symptom in cancer patients. This can occur at all stages of treatment and is a separate phenomenon from ageing. A simple VAS could be used to monitor the effects of any intervention. There is evidence for the use of methylphenidate but caution is required in the older person and those with cardiovascular disease. The risks of polypharmacy may outweigh the potential benefit in this population and it may be that exercise in conjunction with patient self-management and education is the treatment of choice in this group.
There are two major directions for research in this area. The first is to define the clinical phenotype and to specifically examine a range of tumour groups and stages across a full range of age groups. This should be linked to studies examining the underlying mechanisms of CRF. A pragmatic method would be to add mechanistic studies to the quality of life arms of oncology trials.
The other area is to focus on large scale randomized controlled trials examining the role of methylphenidate in all age groups as this seems the most promising avenue on current evidence. This could be combined into a factorial trial with an exercise arm. The current level of evidence makes the need for these types of trials paramount in cancer patients of all ages.
Conflicts of interest
The authors declare no conflicts of interest.
