Pharmacogenomics: personalising medicine

Most people would be shocked to learn that, half the time, the pills and doses prescribed by a doctor probably won’t work for them. But that’s the current reality of medical practice. We work at a population level rather than a personal one. Genomics, though, offers us the ability to deliver more tailor made treatments, where doses and agents are selected according to a person’s make up. So what’s involved in making it happen? Sandosh Padmanabhan is Professor of Cardiovascular Genomics and Therapeutics at the University of Glasgow…

Sandosh - Pharmacogenetics is a field which combines two specialist areas: pharmacology and genomics. The problem with pharmacology or drug treatment is when you go to a doctor, you are prescribed a drug and in almost 40 to 60 per cent of times that drug may not work for you. You don't know that; the doctor doesn't know that. It's difficult to predict.

Chris - Is that because, Sandosh, we are all different: I look different to you because - genetically - I'm different from you, and this means that on the inside I'm a bit different from you as well; and hence what might have been tried and tested on you, hasn't been tried and tested on me and therefore won't work. It's as simple as that...

Sandosh - Absolutely. If you look at your genome, humans are 99.9% identical, but that is this 0.1%, which is different. And this is natural variation. And when you talk about variability in drug response, it's that difference which changes certain proteins in your body that's targeted by the drugs. And if there is a slight variation in that protein or the protein which metabolises the drug, then you see variation in drug response.

Chris - So what you're saying to me is that at the moment what we are effectively doing is I'm taking you by the hand to the shoe shop and I'm not waiting for them to measure your feet. I'm just grabbing a pair of shoes off the shelf and saying you will wear these. That's like giving a drug indiscriminately to somebody. And some people's feet are not gonna fit. And as a result there are gonna be blisters - side effects - of wearing the wrong shoes. And if on the other hand we can look at the genetic code and ask what pair of shoes fits that person best, we we've got a chance to tailor things and minimise the side effects?

Sandosh - Absolutely. And that's a perfect example. And when you go to the shoe shop and if you blindly take different shoes, it may not fit and you'll have to try a lot of shoes to find the right fit, and that's trial and error. But if you have your feet measured and then you go to the right size, then you don't have that much of trial and error to get you on the right dose and the right drug. So it's not just about giving you the right drug, it's about giving the patient the right dose. So we know what drug to give you, but you may be a person where your liver enzyme is unusually overactive because of some genetic variation. So in your case you may actually require a higher dose of the drug, or conversely, you may actually require a lower dose of the drug because the enzyme is so active, it's just removing the drug from the circulation.

Chris - And this is happening a lot in clinical practice. You were saying that 40 to 60% of the time drugs won't work. Is that just down to this effect then that, that when I go to the GP and I'm prescribed something for a condition, there are quite high odds then that that agent may not, or at least the initial dose and the initial choice of agent may not work for me and I may have to be back a few times to get things right?

Sandosh - Correct. And I'd like to quote William Osler who said that if it were not for the individual variability in drug response, medicine would have been a science and not an art <laugh>. And that's what we practice. We practice the art of medicine. We know from clinical trials that this drug will work for this condition, but is that drug the right drug for this patient? And what's the right dose for the patient? That's trial and error.

Most people would be shocked to learn that, half the time, the pills and doses prescribed by a doctor probably won’t work for them. But that’s the current reality of medical practice. We work at a population level rather than a personal one. Genomics, though, offers us the ability to deliver more tailor made treatments, where doses and agents are selected according to a person’s make up. So what’s involved in making it happen? Speaking with Chris Smith, Sandosh Padmanabhan is a Professor of Cardiovascular Genomics and Therapeutics at the University of Glasgow…

Chris - We've been doing this kind of thing in terms of practicing medicine and giving drugs some pills for things for, for a really long time. We've only begun to think like this though much more recently. So how much do we know, how much do we understand; how much, in terms of our pharmacological repertoire, can we actually throw genetics at? So you can use my DNA code to make an informed choice about the drugs that you would like to give me?

Sandosh - The recognition that genetics could play a role, it's not recent. It goes as far back as Pythagoras, who said certain people when they eat fava beans will have hemolysis and pass dark urine. We've known about genetic variations that affect drug response and even dramatic side effects right from the 1950s. So there's a lot of knowledge, there's a lot of science. We have a lot of good examples where genomics can predict adverse drug reaction. And some of those have come into clinical practice. But it's only a handful of genetic variants that are currently used in clinical practice. There is a, a lot of genetic variants out there. So if you look at our drug regulatory agency, so the FDA in the US or the MHRA in the uk, there are nearly 200 drugs where the FDA recommend on the drug label that these drugs have pharmacogenetic implications. It's there on the drug label. But in terms of clinical practice, we have not yet applied pharmacogenetics into our practice.

Chris - Why is that? Is that a practical thing? In the sense that when I went to medical school, we'd really just embarked on the human genome project. It was gonna take us years and cost us about 3 billion. Now you can do it in a day and it costs maybe hundreds to a thousand pounds to do a genome. Is it just that we've been waiting for the technology to catch up or is there another reason why this is not yet very widely implemented?

Sandosh - So this is a complex area and the reasons are multifactorial. Cost is important because there is a genetic test required before prescribing and it's an upfront cost. Secondly, we physicians always follow evidence from clinical trials, which means that for each drug you'll have to do a randomised clinical trial, and it's more expensive. You can't expect pharma companies to do drug trials if it's going to reduce the market for the drugs. Thirdly, in medical school, your genetic education is probably in the first three years or two years after that you don't do a lot of genetics. So now you have all the GPs, specialists, coming out of medical school who have not been trained in genetics. So when you want to implement genomics, you'll face resistance because people will have to train more. On the other side is you have a genetic test with adds cost and most of the drugs we commonly use are off patent. They're very cheap. So that is an argument that, okay, if we just monitor the patients trial and error may be cheaper than genotyping. And then the other thing about data privacy, your genetic information is held in your health records. None of these are huge problems, but there are a multitude of problems and all of them need to be addressed for this to be implemented and people to understand that the benefit of pharmacogenetics is a long-term benefit. Everybody benefits, but it's an investment for the future.