Imaging regional cerebral blood flow

The evidence we have examined indicates that meaningful neural events may entail activity in a distributed network in such a manner that at each of the cortical loci involved there is coherent activation of neurons in an area extending over several millimetres. One corollary of this is that the patterns of activation are accessible to macroscopic imaging techniques that measure local changes in regional cerebral blood flow (rCBF) associated with changes in the metabolic requirements of active neurons. Techniques such as PET and fMRI can be used to provide images of rCBF with a spatial resolution of several millimetres. SPET provides images of rCBF with somewhat less sensitivity and resolution. It is likely that the demand for increased rCBF during neuronal firing will be greatest in the vicinity of the synapses, since synaptic processes consume a substantial amount of energy.
In principle, the activation of either excitatory neurons or inhibitory neurons can cause an increase in rCBF. In practice, it is likely that whenever neurons in a particular cortical location are engaged in meaningful processing, local circuits that include both excitatory and inhibitory neurons will be active, since a combination of excitatory and inhibitory neurotransmission would generally be required to produce meaningful temporal and spatial patterns of neural activity.
None the less, because the major long-range fibres that bring specific information from a distant site to the local circuit are excitatory glutamatergic fibres, and these incoming fibres usually form synapses with many local neurons it is reasonable to assume that an increase in local CBF reflects an increase in excitatory input to that location, while a decrease in local CBF reflects a decrease in excitatory input to that location.
Studies using PET confirm that mental processes are associated with activation at a distributed set of cerebral loci, with relatively extensive activation at each locus. For example, in studies of word generation carried out in collaboration with colleagues at Hammersmith Hospital (Frith et al, 1991a, 1991b), we found that during the internal generation of words beginning with a specified letter, there is activation of extensive regions in the left dorsolateral and medial frontal
cortex, posterior cingulate and thalamus, while there is a diminution of activity in the superior temporal gyrus. The extent of activation in these regions was substantially greater than would be accounted for by smoothing arising from sources such as the limited spatial resolution of the PET camera. In these studies, the pattern of cerebral activity associated with the internal generation of words was determined by subtracting the pattern of activity during the articulation of a list of words provided by the experimenter from the pattern of activity during the articulation of words generated by the subject.

Imaging neuroreceptors

Positron emission tomography (PET) and single photon emission tomography (SPET) can be used to measure the distribution of radioactively labelled tracers in the brain. When the tracer is a ligand for a specific neuroreceptor, the distribution of the labelled tracer provides a quantitative measure of the spatial distribution of accessible receptors. However, not only are there many technical difficulties in preparing a labelled tracer that binds with high specificity to the receptor of interest, and in executing the imaging procedure, but there are intrinsic uncertainties regarding the behaviour of ligands in vivo that complicate the interpretation of the observed binding of the administered ligand. In particular, the administered ligand must compete with any endogenous ligand that is present. A decrease in the amount of tracer that binds to receptors in a particular condition may reflect either a decrease in receptor number or an increase in the concentration of endogenous ligand occupying the receptors, thus making fewer available.
The situation is made even more complex by the fact that competition between endogenous ligand and administered ligand does not appear to obey the laws of equilibrium binding that are observed in vitro. If the administered radio-ligand is present at trace concentrations that occupy only a minute proportion of the available receptors, the number of available receptors should be determined by the total number of receptors and the concentration of endogenous ligand. The relative affinities of the endogenous ligand and the administered ligand for the receptor should have a negligible effect on the observed density of available receptors. The number of available receptors should be the total density of receptors minus the proportion occupied by the endogenous ligand, irrespective of the presence of trace amounts of the radio-ligand. However, that is not what is observed. The binding of administered ligands that bind very strongly to the
receptor appears to be unaffected by the presence of endogenous ligand, whereas ligands that bind weakly are sensitive to the presence of endogenous ligand.
For example, the dopamine D2 ligand N-methylspiperone, which binds very tightly to the D2 receptor, is apparently insensitive to the presence of endogenous ligand, whereas raclopride, which binds relatively weakly, is affected by endogenous ligand, although less strongly than would be predicted. This has led to apparently contradictory findings regarding D2 receptor density in the basal ganglia in schizophrenia, measured by different investigators using different tracers. Wong et al (1986), using N-methylspiperone, found an apparent increase in D2 receptor
density in previously unmedicated schizophrenic patients relative to healthy subjects, while Farde et al (1988), who employed raclopride, found no generalised increase, although they did find an increase in the left putamen relative to the right.

link fatigue to other markers of disease activity.

Several studies have attempted to link fatigue to other markers of disease activity. These include general measures of nutritional status such as weight and albumin, and tumour-specific markers of disease. In general these studies show disappointingly few clear associations:Mendoza et al. (1999) showed a relationship between albumin levels and fatigue for both solid and haematological malignancy, but Monga et al. (1999) found no association between fatigue and haematocrit or body weight in patients with carcinoma of the prostate. Monga’s study also did not find any association between prostate-specific antigen and fatigue in this group. Similarly Stone et al. (1999) found no association between subjective fatigue and malnutrition.
Should we be surprised by the somewhat unspectacular associations between disease activity and fatigue? In a review of the literature Wessely et al. (1998) assessed the association between many different physical diseases and fatigue and in general found a consistent lack of association between disease severity and fatigue. This held for renal failure (Brunier and Graydon 1993), heart failure (Wilson et al. 1995), Parkinson’s disease (Friedman and Friedman 1993), and rheumatoid arthritis (Belza et al. 1993; Belza 1995).Whilst all these diseases are in themselves powerful risk factors for the development of disabling fatigue, there is little association between disease severity and severity of fatigue, or indeed other important symptoms.
There are probably a number of reasons for this failure to find an association between disease severity and symptom severity. Firstly, it may be an artefact of study design. Many studies selected homogeneous populations of patients among whom the disease status may not vary sufficiently to demonstrate the importance of severity as a risk factor for the symptom. Secondly, most studies are cross-sectional, and a true relationship between disease severity and fatigue may get lost in the ‘noise’ of interindividual differences. If longitudinal studies were used to follow patients across the course of their illnesses these might demonstrate rather more convincing associations. Thirdly, it may simply be that any association is overwhelmed by the importance of psychological and behavioural factors—the presence of depression, interindividual differences in terms of self-efficacy, the effects of behavioural change and deconditioning, and so on.

Sociodemographic risk factors for fatigue in cancer

The pattern of increased risk of fatigue among women which was so clearly demonstrated in population-based studies of healthy individuals is not so obvious in cancer-related fatigue. Of nine studies identified which assessed the relationship, five found no difference between men and women (Hickok et al. 1996; Smets et al. 1998a; Glaus 1998; Donnelly et al. 1995; Stone et al. 1999), whilst the remaining four (Vogelzang et al. 1997; Smets et al. 1998b; Loge et al. 1999; Akechi et al. 1999) showed that women had more fatigue than men. The key difference between papers which appear to show a difference and those which do not is the disease stage or time from treatment—those that found a difference in rates between men and women tended to have taken samples with earlier disease or ‘survivors’ in whom the disease had remitted. It may be that with less aggressive or remitted disease, the pattern of fatigue becomes closer to that of the general population.
The pattern of fatigue in cancer according to age is confusing. There might be an expectation that younger cancer patients suffer less fatigue, but Hickok et al. (1996) found a U-shaped distribution of fatigue in cancer patients according to age, with fatigue symptoms at their highest in relatively young adults (aged less than 50) and the over 70s. Glaus (1998) found a slightly more complex biphasic relationship, with younger patients having more affective symptoms of fatigue whilst the older group had more physical symptoms of fatigue. Two other studies reported a relationship between age and fatigue—Loge et al. (1999) found that older patients were more fatigued and Vogelzang et al. (1997) found younger patients were more fatigued. The remaining studies showed no relationship (Andrykowski et al. 1998; Smets et al. 1998a; Stone et al 1999; Cimprich 1999; Hann et al. 1999).
One other remaining sociodemographic characteristic was widely studied, namely educational status. Here there appears to be a consistent association, with most studies reporting greater fatigue in the less well educated (Irvine et al. 1994; Mast 1998; Loge et al. 1999) although, as ever, there are exceptions, with one study (Akechi et al. 1999) describing the opposite relationship.

Prevalence: how common is fatigue in cancer?

The simple answer to this question is ‘very’. However, there are a number of important obstacles to determining the prevalence of fatigue in cancer patients. In part it is difficult to gain a reliable picture of the prevalence of fatigue in cancer because most studies which look at fatigue in cancer assess the symptom in special groups. It is likely that the prevalence of fatigue in cancer patients varies significantly accordingto site and type of cancer, stage of disease, presence of medical co-morbidity (anaemia,infection, metabolic disturbance), and psychiatric disorders (especially depression andanxiety).
it makes the results difficult to interpret except to those familiar with the questionnaires involved.
Some studies have assessed the prevalence of fatigue against other common symptoms:
Ng and von Gunten (1998) found that fatigue and weakness were the two commonest symptoms in hospice patients, with a prevalence of over 80%, whereas painwas only present in approximately half. Despite the almost universal presence of fatigue in this population, it was relatively infrequently rated as the ‘main symptom’— only being reported by 6%, as opposed to 11% for pain (the most common ‘main symptom’ was respiratory problems, present in 22%). Other studies which have assessed multiple symptoms also point to fatigue being the most common symptom of cancer irrespective of diagnosis or stage (Kurtz et al. 1994; Savage et al. 1997; Newell et al. 1998), except where the symptom was ascertained from case notes (Savage et al. 1997), implying that medical staff tend not to question patients about it, or ignore the symptom if it is spontaneously reported.
Relatively few studies have attempted to compare patients with cancer and other control populations: Mendoza et al. (1999) and Stone et al. (1999) (see Table 1.2) are exceptions and took healthy controls as comparison groups. Both studies found very significant differences in prevalence rates of severe fatigue between these groups. There can be little doubt that cancer is strongly associated with fatigue when healthy control populations are used. In contrast Andrykowski et al. (1998) compared levels of fatigue in 88 women with breast cancer and a similar number of women who had attended a breast clinic for benign breast disease.Whilst the women with breast cancer had more fatigue than those with benign disease the difference was slight and not statistically significant on several of the fatigue scales used. This may say more about their control group of symptomatic women consulting for a breast problem than about the cancer patients.
Two studies (Mendoza et al. 1999; Stone et al. 1999) looked at the distribution of fatigue among patients with cancer. In the general population the distribution is strongly skewed, with a long tail going to the right of the distribution. These studies indicate that the entire distribution is shifted to the right in cancer patients. Hence the increased prevalence of fatigue in cancer patients is not due to a relatively few sufferers who are very fatigued, but to a general shift in the population’s experience of the symptom.

Uncertainties regarding assessment, treatment,and future care

The complexity of fatigue in cancer and the wide range of different mechanisms means that accurate assessment, with an understanding of the likely cause(s) and associated factors, is critical to considering treatment options. There has been a lack of information on this in the past, often leading doctors and nurses to feel hopeless in the face of fatigue.
But some possible methods of assessing fatigue and monitoring progress are now available. Measuring the severity and consequences of fatigue is complex; fatigue is difficult to measure, because of the different ways in which it is interpreted and understood. Yet this is vital if treatments and their outcomes are to be monitored over time. Treatment options are varied, depending on the likely cause. The evidence base for treatments is only now developing; studies are often lacking and are difficult to conduct.
However, correction of simple causes, such as sleep interruption or anaemia, if identified, is a common first step. There are also general non-pharmacological measures such as adapting activities of daily living and occupational therapy to help match clinical function and symptom status with the expectation of patients and families. There is a wide range of pharmacological and non-pharmacological interventions available. Corticosteroids have been proposed as a treatment, and are often used. It is argued that they may decrease fatigue, either by inhibiting the release of tumour-induced substances or by a central euphoria effect. Amphetamines have been favoured in some settings, in particular in North American countries. Many other pharmacological treatments are emerging for the treatment of fatigue and its related symptoms.
Because of the lack of concentration of research and clinical effort in cancer-related fatigue some aspects the findings are preliminary and in some areas they contradict. Here we have sought to present the best knowledge and to debate opposing evidence, in the hope that it will provide doctors, nurses, and all those involved in caring for patients and families with mechanisms for improving care, as well as identifying important future lines of enquiry through research. The appraisals may begin to break the past, sometimes nihilistic, attitude to the assessment, discussion, and treatment of fatigue and encourage future trials and investigation, so that ultimately the effects of this most common and neglected symptom may be reduced.

Fatigue: a complex symptom with wide-reaching effects

Fatigue is associated with increased distress due to some other symptoms, including
pain. It often clusters with cachexia and anorexia, and is difficult to distinguish
between them . It has profound effects on everyday functioning and,
perhaps consequently, service use. It reduces quality of life and increases suffering.
Fatigue has been associated with hospital admission and increased stress to caregivers
(Hinton 1994; Robinson and Posner 1992). The needs of lay caregivers in this context
are often overlooked . A deeper comprehension of these factors is
important in assessing patients, planning care, and in designing and testing future
treatments for fatigue.
But do the effects of fatigue stretch even wider, reaching well beyond traditional
symptom boundaries? Having energy and vitality is an important part of self-image.
Fatigue is often seen as a sign of impending deterioration. So if doctors and nurses are
to discuss fatigue with patients, grasping their interpretation and understanding of
fatigue is essential . Fatigue can have a profound meaning for patients
living with fatigue, and for their carers or family, which also need to be considered and
assessed.