YOU have a headache and you buy an over-the-counter 鈥渆xtra strength鈥� painkiller. Among other ingredients, it may contain codeine, a well-known drug for pain relief, but one without analgesic properties itself. Codeine is a pro-drug: only after the body transforms codeine into morphine is there any pain relief. But nearly 1 out of every 10 people of European ancestry do not produce the right enzyme to turn codeine into morphine. For them, codeine has no effect. You might like to know if you are one those people.
Other examples of inappropriate treatments are less benign. Take the case of children with acute lymphoblastic leukaemia, treated with 6-MP, a drug that interferes with the accurate synthesis of DNA. The problem with this drug is that if it accumulates in the body it can prove toxic and even fatal, so the body must convert it into an inactive form. However, about 1 in 300 people do not correctly make the necessary detoxifying enzyme, so for them sustained treatment could prove deadly. Genetic testing before 6-MP treatment is now routine in the US but not elsewhere.
These are extreme examples of genetic differences that influence how people respond, but they point to a large general problem. For all the benefits medicines bring some patients, their overall performance is very mixed. In some people, they may not work at all, or may be harmful. For example, in the UK, up to 10,000 people may die because of adverse drug reactions (100,000 per year in the US).
Advertisement
This is not surprising given how blindly medicines are used. For many diseases and conditions, including hypertension and epilepsy, there is a wide range of available drugs. In many cases, doctors try different medicines randomly until they find a drug or drug combination that works. There is little doubt that future doctors will view this as the pharmaceutical equivalent of using leeches for everything.
Not only is the choice of drug a matter of trial and error, but often so is the dose. For many common drugs, including certain antidepressants, antipsychotics, anti-epileptics and opioids such as morphine, identifying the right dose can take a long time, often causing unnecessary suffering.
The reasons why different patients need different doses vary. Some of it is genetic, and some of it we can predict right now. For example, people who metabolise fluoxetine (Prozac) fast need a dose up to four times as large as those who metabolise it slowly. But rarely does anyone check how fast an individual metabolises Prozac before prescribing the standard dose. Doctors tend to choose the drug and dose irrespective of genotype, despite the fact that we know of at least 50 genetic differences between individuals that may affect drug response. Virtually none of this information is being used in British clinics. Most doctors and patients don鈥檛 realise these genetic differences exist, and wouldn鈥檛 know how to adjust for them if they did. Indeed, for the majority of known variants, there has not yet been enough basic research to determine their clinical implications.
But even more important than what we know is what we don鈥檛 know. Pharmacogenetics has been around for about 50 years. The human genome project has made it possible to hunt systematically for relevant genetic differences, so you would be forgiven for assuming that such searches are now being carried out for most major prescription medicines.
鈥淩arely does anyone check how fast an individual metabolises Prozac before prescribing the standard dose鈥�
Unfortunately, nothing could be further from the truth. Some companies remain ambivalent: after all, where is the commercial advantage in costly pharmacogenetic studies on medicines that have already been approved? Academia is more interested, but researchers lack the funding and integrated research programmes needed to do the work correctly.
But pharmacogenetics is worth doing for its benefit not only to patients, but to society. A recent study in the UK found that adverse reactions to medicines account for 1 in 16 of all hospital admissions and cost 拢466 million a year. There is also growing support for the idea that this research may lead to better treatments much sooner than disease genetics because while some gene variants have been found to be risk factors for disease, it is unclear how this relates to treatment or prevention.
The bottom line is that there is an urgent need for basic research into the genetics of drug response, and to equip healthcare providers with the tools they need to use the information. This is not easy: it will require more funding and accurate patient response data. Applying the research to clinical practice requires large prospective studies and training for doctors, and overall we need more cooperation between industry and academia.
But the promise is great: more effective use of existing medicines, quicker, cheaper and better development of new medicines, fewer side effects for patients, and treatment tailored to individuals. It is past time to get serious about pharmacogenetics.
