Fatigue is experienced by patients inmany phases of cancer— at presentation, with chemotherapy, with radiation therapy, after surgery, in survivors, and in advanced cancer. The universality of fatigue as a symptom and its effects on the quality of life in patients with advanced cancer [Lundh Hagelin C, Seiger A, Furst CJ, Quality of life in terminal care—with special reference to age, gender, and marital status, Support Care Cancer 14:320–8, 2006] makes it important, but fatigue has been a very difficult symptom to define and study. One approach to understanding fatigue has been to characterize the epidemiology of fatigue and other symptoms. Many studies in different groups of cancer patients have shown statistically significant correlations between fatigue and other symptoms. This approach has been advanced by the current paper by Yennurajalingam et al. [Yennurajalingam S, Palmer JL, Zhang T, Poulter V, Bruera E, Association between fatigue and other cancer-related symptoms in patients with advanced cancer, Supp Care Cancer, 2008]. They analyzed cancer patients who had enrolled for clinical trials for fatigue, who were rigorously evaluated for other possible causes of fatigue, and assessed for fatigue and symptom covariates obtained with widely accepted instruments, the Edmonton Symptom Assessment Scale and the FACIT Fatigue subscale. Such a relatively uniform and carefully selected group should make it easier to detect associations. On the other hand, findings from a research population may not always be generalizable to the overall population. Yennurajalingam et al. confirmed many of the associations between symptoms by using stepwise linear regression models in an adequately powered population, with the FACIT F fatigue subscale score as the dependent variable. Anxiety, drowsiness, anorexia, and well being were correlated with fatigue, but interestingly, depression, dyspnea, and insomnia were not present in the final model. One finding from this approach was that the R for their model was 0.21, which the authors suggest accounts for a small amount of the variance. This small R2 may not necessarily indicate low levels of association of the underlying variables with fatigue, but could be an artifact. Sources of variation include measurement error both in fatigue and predictor variables in addition to variables that have not been included for analysis. Patient-reported outcomes, like other variables, also have error terms. These sources of measurement error may include different perceptions of fatigue by patients, and within a single patient, that person’s perception can vary over time. Variations from measurement error can sometimes contribute considerably to the variance, and this should be noted when interpreting the R term. We still do not know whether the relationship between fatigue and the other symptoms is causal or if these symptoms are part of a more general response. Fatiguerelated symptom clusters will be helpful in clinical assessment and management of patients and allow advances in the understanding of other symptoms to generate new hypotheses about fatigue. To gain further insights into fatigue, we will eventually need to consider other study designs. If one thinks of fatigue as a common form of illness behavior in response to various kinds of stress, then a context-related definition of fatigue patients may be Support Care Cancer (2008) 16:1095–1096 DOI 10.1007/s00520-008-0483-4
Read more