Functional somatic symptoms and syndromes are a major health issue. They are common, costly, persistent and may be disabling. Most of the current literature pertains to specific syndromes defined by medical subspecialties. Indeed, each medical subspecialty seems to have at least one somatic syndrome. These include: irritable bowel syndrome (gastroenterology); chronic pelvic pain (gynaecology); fibromyalgia (rheumatology); non-cardiac chest pain (cardiology); tension headache (neurology); hyperventilation syndrome (respiratory medicine) and chronic fatigue syndrome (infectious disease). In 1999, Wessely and colleagues concluded on the basis of a literature review that there was substantial overlap between these conditions and challenged the acceptance of distinct syndromes as defined in the medical literature (Reference Wessely, Nimnuan and SharpeWessely et al, 1999). They proposed the concept of a general functional somatic syndrome. But is there any empirical evidence for such a general syndrome? Is it even a useful concept? Five years on, Professor Simon Wessely, King's College London, revisits this debate. He is opposed by Dr Peter White from St Bartholomew's Hospital and Queen Mary School of Medicine and Dentistry, London.
When Mike Sharpe, Tok Nimnuan and I proposed in the Lancet (Reference Wessely, Nimnuan and SharpeWessely et al, 1999) that the classification of the world of unexplained syndromes was a mess, we had little idea of how controversial it would prove to be. This debate is, therefore, a welcome opportunity to make it clear what we did, and did not, say.
Our starting point was that every medical specialty has its own ‘unexplained’ syndrome, by which we mean a diagnostic label used in that clinic to describe patients with symptoms, disability, but no clear-cut biomedical mechanism to explain their distress. Gastroenterologists see people with irritable bowel syndrome, rheumatologists see fibromyalgia, infectious disease specialists frequently diagnose post-viral fatigue syndrome, and so on and so forth.
However, as time passed, more and more reports described the overlaps between two or more of these syndromes. If one studied a population with the label of fibromyalgia, many also fulfilled criteria for chronic fatigue syndrome. Patients presenting with irritable bowel syndrome have high rates of tension headaches. Women with chronic pelvic pain also had marked myalgic symptoms. And again, so on and so forth. Indeed, the literature on these syndromes showed numerous similarities. The epidemiology seemed similar. Putative mechanisms overlapped (abnormalities of serotonin function, for example, have been found in many of these syndromes). Outcomes seemed similar, and not encouraging. All had links to depression and anxiety. And there was considerable overlap in successful treatment strategies, especially those involving some variety of active rehabilitation, such as cognitive-behavioural therapy (CBT).
So why did our article provoke so much reaction? (Although, to be fair, this was rarely among professionals, most of whom had no problem in accepting our thesis, particularly if they belonged to that disappearing breed, the general physician.) There were two reasons. First, many sufferers did not accept the thesis, and continue to have strong emotional attachments to their label. A person who believed powerfully that they were a victim of multiple chemical sensitivity, for example, did not always take kindly to the view that they also had links to atypical chest pain or irritable bowel syndrome, even if they had the symptoms of both. Explanations advanced for these syndromes clearly differ - a patient attending a gynaecology clinic with pre-menstrual syndrome has a very different idea of why they are sick than a person attending a clinic for chronic fatigue syndrome. Nevertheless, our argument that the symptoms overlap to a very great extent is not disproved by this, even if the explanations advanced by patient or professional differ.
Second, some felt that what we were saying was that all these syndromes are psychiatric. Those offended by this idea were often those who also equated the word psychiatric with imaginary or non-existent, a regrettable view sadly made more regrettable because it is also held by some professionals. But this has never been our argument. What we said is that all of these syndromes still fall under the title of ‘unexplained’ since no consensual scientific explanation has been advanced for any of them that meets with universal acceptance. Unexplained means what it says on the tin, and is not a code for psychiatric, still less for ‘all in the mind’.
Five years later, Sharpe and I stand by our thesis. We are not saying that all these syndromes are the same. We do believe that in time differences will emerge that will enable us to divide up the unexplained cake better than at present. We believe that better understanding and classification will result from an improved understanding of mechanisms. Conversely, we do not expect that improved understanding will come from further statistical manipulations of symptoms and their occurrence. The symptom-based classifications that we have now are more a reflection of professional specialisation and access to care, and do not cleave nature at the joints.
Wessely and colleagues (Reference Wessely, Nimnuan and Sharpe1999) have suggested that patients with diverse medically unexplained symptoms ‘may have... a general functional somatic syndrome’. Telling patients that they have a psychosomatic disorder is usually the first step in a deteriorating doctor-patient relationship. To most lay people, psychosomatic means malingering or ‘all in the mind’. Lumping functional somatic syndromes together as a general functional somatic syndrome conceptually supports mind/body dualism, feeding this misapprehension. A general functional somatic syndrome can be consistent only with psychogenesis, since it is difficult to conceive of a pathophysiological mechanism that would be common to all functional somatic syndromes. The alternative of deconstructing or splitting functional somatic syndromes into their constituent parts leads both to a more sophisticated understanding of these illnesses and to better treatments, using the biopsychosocial approach.
Wessely and colleagues suggest that case definitions of functional somatic syndromes overlap. However, there is little overlap in the core symptoms of the two most common syndromes: irritable bowel syndrome and fibromyalgia. The apparent overlap is also confounded by both co-morbid mood disorders and selection bias. Primary care and community studies find lower rates of overlap of functional somatic syndromes than do secondary care studies (Reference Jason, Taylor and KennedyJason et al, 2001; Reference Whitehead, Palsson and JonesWhitehead et al, 2002).
Most doctors are either ‘splitters’ or ‘lumpers’ as they classify ill-health. Historically, more progress has been made through splitting illnesses rather than lumping them together. Take the example of dropsy (generalised oedema), which was thought to be a single disease, until it was divided into heart, kidney and liver causes. Psychiatric taxonomies are similarly misleading and dualistic. A somatoform disorder can only be so classified in the absence of an adequate physical explanation (World Health Organization, 1992). Furthermore, a somatoform pain disorder can only ‘... occur in association with emotional conflict or psychosocial problems that are sufficient to allow the conclusion that they are the main causal influences’ (World Health Organization, 1992). How can the clinician be sure that the psychosocial problem actually caused the illness? Somatisation is generally more useful when regarded as a process that is essentially independent of diagnosis, and which can therefore be applied to a patient with any medical condition.
The concept of a general functional somatic syndrome does not lead to better understanding of aetiology. For instance, there is a five-fold risk of chronic fatigue syndrome in patients suffering from infectious mononucleosis (Reference White, Thomas and AmessWhite et al, 1998), whereas there is no evidence that fibromyalgia is caused by infections (Reference Rea, Russo and KatonRea et al, 1999). Lumping fibromyalgia and chronic fatigue syndrome together as a general functional somatic syndrome would have reduced the chance of finding this effect (because of dilution). Moreover, the risk factor of childhood sexual abuse varies six-fold across different functional somatic syndromes (Reference Romans, Belaise and MartinRomans et al, 2002). It is only by separating general functional somatic syndrome into its separate disorders that we will advance understanding of causation. We have started to understand the pathophysiology of fibromyalgia as central nervous system supersensitivity due to brain neuroplasticity (Reference Gracely, Petzke and WolfGracely et al, 2002). We have gone further by starting to deconstruct individual functional somatic syndromes, such as chronic fatigue syndrome, into aetiologically different disorders.
The concept of a general functional somatic syndrome does not lead to better treatments. Antidepressant efficacy ranges widely in different functional somatic syndromes, and may be more accurately predicted by the presence of comorbid mood disorders. A recent systematic review showed that ‘... psychosocial treatments have not yet been shown to have a lasting and clinically meaningful influence on the physical complaints of polysymptomatic somatisers’ (Reference Allen, Escobar and LehrerAllen et al, 2002). A recent large trial of treatment of Gulf War syndrome found no significant differences between CBT and control treatments (Reference Donta, Clauw and EngelDonta et al, 2003). An accompanying editorial by Hotopf (Reference Hotopf2003) correctly attributed this lack of efficacy of CBT to not using an illness-specific model for CBT. In contrast, CBT is effective when specifically designed to help improve the physical functioning of patients with chronic fatigue (Reference Whiting, Bagnall and SowdenWhiting et al, 2001).
Finally, the concept of a general functional somatic syndrome does not predict prognosis, which varies by specific functional somatic syndrome. Fibromyalgia runs a persistent and chronic course, whereas irritable bowel syndrome runs an intermittent course with recovery being more common. The concept of a general functional somatic syndrome, therefore, reduces the accuracy of prognosis.
I conclude that the concept of a general functional somatic syndrome is unhelpful in understanding illness, aetiology, treatment and outcome, thus failing four of Kendell's tests of clinical validity (Reference KendellKendell, 1989). Illnesses with unexplained physical symptoms are best considered in an integrated way, paying equal attention to body, mind and social context.