Science for Health
Over the last decade, huge steps have been made in deciphering the human genetic code, with the promise that we will finally understand what “makes” a human. Predictably perhaps, the completion of the sequencing of the human genome has raised as many questions as it answered. The number of genes we carry is of the same order of magnitude as in a mouse; far fewer than had been anticipated. Furthermore, the genetic differences between individuals seem too few to account for the wide variations seen between individuals. Nevertheless, the search through our genetic code for genetic mutations that may cause disease continues. What issues will this new knowledge raise for future doctors? The way that recent medical technology has affected the world of clinical medicine gives us some clues.
Scientists have known for a long time that the gene is not the only determinant of an organism’s exact physical form, also referred to as its phenotype. Most research into embryological development is carried out in inbred mouse strains. These strains are the result of many generations of brother-sister matings, to reach a point where all members of the mouse colony carry more or less exactly the same copies of each gene. Thus genetic variation in such strains has been removed. However, subtle individual variation between siblings is still found, for example in the structure of the heart. As this cannot be due to genetic differences, scientists have described so-called epigenetic mechanisms, which are not directly due to the genetic code, to account for this variation. For example, the outflow tract of the heart goes on to form the aorta, the major blood vessel leading from the cardiac ventricles to the body. In its development the patterns of blood-flow are a crucial factor in determining how this most important of blood vessels forms. In a chick embryo, insertion of a tiny glass bead into the outflow tract dramatically changes the formation of the aorta, without having directly altered gene expression in the heart. Scientists are finding many similar mechanisms that are important in developmental biology, which can generate complexity of phenotype from the seemingly limited palette of the genome.
The simplistic idea that a gene encodes a protein which then goes on to make an organ is obviously not the whole story. What about the idea that a single defective gene goes on to cause a single disease? There are certainly many examples of a small change in the sequence of an individual gene leading to a unique disease. For example, cystic fibrosis is caused by single point mutations in a huge gene encoding a protein that crosses cellular membranes and acts as a pump redistributing ions. Thus a diagnosis of cystic fibrosis can be made by sequencing the gene. If you’re unlucky enough to have inherited the gene, then you’ve got the disease. End of story. Or is it? If individual patients carrying the mutant gene are examined, then no two are exactly alike, presumably as a result of all the other genes they inherited from their parents, which act to mitigate, or worsen, the effects of the mutant cystic fibrosis gene. Then there is the age-old debate about the relative contributions of nature versus nurture, the relative importance of the effects of the genes we inherit versus the effects of the environment we live in - how much of our environment (tobacco and alcohol consumption, environmental pollution, educational opportunities) ends up contributing to our adult phenotype?
he discordance between our genotype, the identity of the genes we carry, and our phenotype has presented some real practical difficulties for doctors. For example, in the field of cardiology, there has been impressive progress in discovering the genes that lead to inherited diseases of the heart muscle, termed cardiomyopathy. (Cardiomyopathy is merely a translation of “heart muscle disease” into Latin. Many medical terms are merely Latinised descriptions designed to sound impressive to the uninitiated). When one member of a family is diagnosed with this devastating disease, doctors will look for a disease-causing gene in both the patient and also other members of the family in order to tell carriers of the “bad gene” that they are at risk of future disease. However, this information is sometimes not as helpful as it may seem. Seemingly well people who have been told they carry a “disease causing” gene ask, quite understandably, how long it will take to develop the condition, and what the chances are that they will develop the disease. Both are very difficult questions to answer honestly. When cardiomyopathy genes were first discovered fifteen to twenty years ago, examination of some families revealed something surprising. The same disease-causing gene had different effects on different family members. Some carriers of the disease gene would require heart transplantation, while other close relatives with the same gene would remain well. It is known that some carriers of “disease-causing” genes live normal lives. The effect of the other genes that individuals in these families inherited is perhaps modifying the effect of the disease gene, in ways that we cannot understand at the moment.
We currently know around half the genes that cause cardiomyopathy. So another practical issue arises in the clinic. If you as the doctor don’t know what gene caused the cardiomyopathy that is affecting several members of a family, is there any useful information that can be given to other members of the family to let them know if they are at risk of developing the condition? The development of the echocardiograph, or cardiac ultrasound, has helped to some degree. Family members are screened with echocardiography to see if their hearts “look different” from normal hearts. If these people are told that they have abnormal looking hearts, a presumptive diagnosis is often made, with terrible emotional consequences, not to mention problems with obtaining life insurance. What do you as a doctor tell patients whose hearts look “a bit different”? This is where real difficulty arises. One of the universal truths is that every aspect of human form, function performance and behaviour exists on a spectrum. At some point someone has to decide (sometimes quite arbitrarily) what constitutes “normal” and “abnormal”, in just about every aspect of medicine.
One debate in medicine about what is normal concerns “non-compaction” of the left ventricle in cardiology. When the current generation of echocardiographs was first used, a novel type of cardiomyopathy was found, in which the wall of the left ventricle (the chamber of the heart pumping blood to the body from the lungs) appeared spongy. It appeared as if the normally thin spongy layer that forms a rim inside the left ventricle had enlarged. The term “left ventricular non-compaction” was coined, and a new cardiomyopathy joined the list. Debate then started in academic journals about how best to measure the degree of non-compaction: should the length that the spongy layer protrudes into the left ventricular cavity be the measurement of choice, or should one take account of the volume of the layer in three dimensions? To confuse matters even further, echocardiographers started noticing something unexpected. A proportion of people without cardiomyopathy who were having echocardiographs for different reasons turned out to have hearts that were almost identical in echocardiographic appearance to hearts with ventricular non-compaction, yet they did not suffer the loss of pumping function that is the hallmark of a cardiomyopathy. This led researchers to start studies in which all-comers to hospital were examined with echocardiography, to try and get an idea of the normal range of left ventricular spongy layer size. While these studies are currently underway, it is already pretty clear that some people without cardiomyopathies are walking around with “clearly abnormal” non-compacted layers of their left ventricles. Are these people at risk of developing the condition? If they are, why hasn’t medicine picked up this problem previously? Is non-compaction a separate cardiomyopathy at all? There are more questions than answers here.
This is far from an isolated example. Whenever a new medical test finds its way from the research laboratory into the clinic, there is always a period of assessing the range of results of the test in the normal population. A useful test will have very little overlap between results in diseased and non-diseased populations. Less useful tests will have more overlap, leading to greater confusion. More problems occur when the data from the test isn’t simply a number, but a picture. The results of diagnostic imaging methods such as echocardiography or MRI are notoriously difficult to interpret, and every time a new imaging machine arrives in hospital, doctors have to learn what is “normal” and “abnormal” in the new images, and the overlap between images of disease and the extremes of the normal range of human anatomy.
Screening programs that cover the general population will therefore often turn up unexpected and inexplicable results. A very good example of this is screening athletes for cardiac abnormalities to prevent sudden cardiac death (SCD). As the news media are always keen to note, there is a low but appreciable rate of sudden death in young athletes from undiagnosed cardiac disease. Problems in the electrical conduction system, heart muscle, heart valves and holes-in-the-heart can all cause catastrophic electrical failure under extreme cardiac stress in otherwise well people, leading to sudden death. Some conditions that can lead to SCD leave diagnostic clues in the overall appearance of the body, most do not. One common example of a condition with features outside the heart is Marfan syndrome, whose sufferers have long limbs, and can be extremely tall, in addition to the cardiac malformation that predisposes them to SCD. Ironically it is these features that make some Marfan sufferers more likely to play competitive sport, in sports such as basketball, where great height is an advantage. In a case like Marfan syndrome, it is relatively easy to identify the individuals from their overall appearance, and then examine the heart with echocardiography to look for cardiac problems. What can be done to prevent SCD for athletes with no outward sign of cardiac disease? Historically, two approaches have been taken. Firstly, careful autopsies should be performed on the unfortunate individuals who have suffered SCD, to make a post mortem diagnosis of the cause; secondly, echocardiography should be performed on everyone who plays sport competitively.
Post mortem examination of hearts from young sportsman SCD victims has shown that the causes of death are cardiomyopathy, valve disease and “unknown” in more or less equal proportions. Surprisingly, for such a young and healthy group, a significant number of these deaths are caused by coronary artery atherosclerosis, the thickening of the lining of the arteries supplying blood to the muscle of the heart, and normally associated with smoking, obesity and old-age. A very interesting question is how representative this athletic group are of the general population. Are athletes a self-selecting group with abnormally efficient hearts that have a peculiar tendency to malfunction under stress, or are they like the rest of us? The evidence is that they probably do represent a proportion of the general population, as the rates of SCD are broadly similar in non-athletes, but these equally tragic events are not as high profile in non-athletes. In that case, the finding that some members of this young, healthy group already have a dangerous build-up of atherosclerosis has clear implications for the rest of the population. The “epidemic” of childhood obesity is a well-rehearsed media scare story, based on hard facts. There are growing numbers of obese children, who have detectable atherosclerosis and abnormal arterial function even before their tenth birthday. Some studies have shown that obese ten year olds have arteries that behave more like those of a forty year old smoker. Given that a proportion of even the super-healthy suffer from advanced atherosclerosis, doctors are worried that the young obese are going to keep the NHS very busy when they grow up.
Echocardiographic screening of athletes is carried out in several countries. Experience has shown that if everyone who performs sport at a certain level is screened with echocardiography, a proportion of “abnormal” hearts can be identified, and their owners told not to play sport. If the abnormality is pronounced enough, an implantable defibrillator can be implanted, to shock the heart back into a normal cardiac rhythm should a cardiac arrest occur. So why not simply perform echocardiographs on every athlete in the UK? For that matter, why not try and identify future SCD patients in the general population? Inevitably, the answer is cost. Analysis of the cost-effectiveness of screening programmes (the rather brutal process of calculating the expenditure needed to save a year of life) has led to the conclusion that screening every high-school athlete with the best available test, echocardiography, is not cost-effective compared with other health-care expenditure such as cancer chemotherapy or maternity services. A charity in the UK is raising funds to perform echocardiography screening in the absence of government support. Some countries, notably Germany and Italy, have elected to invest in screening of young athletes for cardiac disease anyway, and have noticed a decrease in SCD in this population as a result.
One region of Italy found that by identifying and disqualifying two percent of athletes from competition, the rate of SCD in that region decreased by 89%. However, the rate of SCD prior to this screening was nowhere near two percent of those screened. Had the screening program been overzealous in disqualifying those with abnormally appearing echocardiographs? It turns out that around 1% of healthy newborn children have hearts that appear abnormal to echocardiography, and these “abnormal” appearances persist into adulthood. One of the really interesting questions in cardiology is whether these people represent a group that will suffer cardiac problems in the future, or whether they are merely extreme variations of normal. Only time will tell, and the difficulty encountered in discussing this with these unfortunate people (who are not patients, because they don’t have an actual disease) is here to stay for the foreseeable future. To complicate matters even further, distinguishing “normal” from “abnormal” is even more difficult in athletes, as the athletic heart has thicker muscle due to training, and can be indistinguishable in appearance from certain cardiomyopathies as a result. These diagnostic decisions matter greatly, especially as some athletes can have extremely lucrative careers put at jeopardy by the interpretation of their scans. Some medically disqualified athletes have mounted legal challenges to their diagnoses, adding another complication.
The margins of “normality” in medicine are a fraught environment in many ways. Many pressures are brought to bear on the cut-off points that determine normality versus disease. A very good example of this is the area of psychiatry. What determines whether the behaviour of an individual is considered to be an extreme of normal, or disease? Open any psychiatry textbook from the 1950s, and you will find homosexuality listed as a mental disorder. In 1973 the American Psychiatric Association removed homosexuality from its diagnostic manual, no longer regarding it as a disorder. Clearly society has moved on and the borderline of normal behaviour has been shifted. It has been frequently argued that behavioural continua of this type are in some way “soft”, or not objective, and thus liable to periodic shifts in accepted “normal values”. We have already seen, however, that even “harder” measurements such as echocardiography are just as open to debate.
Yet another pressure on the borderlines delineating disease and health has been commercial. Over the last decade, there has been relentless pressure from the pharmaceutical industry for more of us to consume more of their products. One way of doing this is to “disease-monger”, or, to put it another way, invent new diseases that correspond to marketed drugs. The condition “erectile dysfunction” came into being shortly after Viagra was marketed, to persuade patients and governments that impotence in older men was in fact a disease, and should therefore have drugs to treat it paid for by the healthcare system. There are many other examples of this medicalisation, such as growth hormone for the abnormally short, new drugs for “restless legs” while you sleep, and prescription of antidepressants for “social anxiety disorder”. The seemingly trivial matter of “what is normal” in many contexts is anything but a simple question.
What hope is there that scientific advances will help clarify some of these issues? If current trends continue, the cost of sequencing an individual’s genome will reduce to a manageable amount, in the range of one to two thousand dollars. An attempt to find all the genetic changes for common disease is already underway in the so-called “genome wide association screens”. Most of the common diseases such as atherosclerosis, high blood pressure and senile dementia are not caused by single, but multiple gene changes. Presumably we will eventually also be able to identify the important genes that modify the severity of effect of “disease genes”. The hope is that by uncovering the most important genetic changes associated with common diseases, we may be able to predict what diseases we are at risk of in the future, and modify our lifestyles, or take medication accordingly. Experience of the effects of advances in science and technology of the type described above leads to the thought that any advances will be paid for with another set of unanswerable questions. It seems that the margins of medical “normality” will remain blurred for a good few years to come.
This essay was published in the Mill Hill Essays 2008
© MRC National Institute for Medical Research
The Ridgeway, Mill Hill, London NW7 1AA