How does scientist test to find out drug interaction?

[kenny1999]

There are possible up to hundreds of thousand drugs used in medicine, then there are just too many combinations between two drugs. Do they test each new drug with hundreds of thousands of other approved drugs before introducing the new drug to the market?

How about three or more drugs? I mean it's not uncommon that a patient may be taking more than one drug because they may have a lot of diseases. How do they know if the new drug will not have problem when it is used together with possibly two or more drugs? For example, drug A + drug B + drug C or even with drug D or more. There are just countless combinations.

[studiot]

4 hours ago, kenny1999 said:

There are possible up to hundreds of thousand drugs used in medicine, then there are just too many combinations between two drugs. Do they test each new drug with hundreds of thousands of other approved drugs before introducing the new drug to the market?

How about three or more drugs? I mean it's not uncommon that a patient may be taking more than one drug because they may have a lot of diseases. How do they know if the new drug will not have problem when it is used together with possibly two or more drugs? For example, drug A + drug B + drug C or even with drug D or more. There are just countless combinations.

 

Hundreds of thousands is a bit of an overestimate.

Quote

Oxford Science

The British National Formulary (BNF) is a reference book containing the standard list of medicines used in UK prescribing. It gives information on the indications, dosages and side effects for over 70,000 medicines.

 

But that is still a large number so you are right it is not practcable to pretest all possible interactions.

But many potential interactions are predictable by biochemical means and some will never occur simply because there is also redundancy in the drugs list.

In other words if there are say 5 antibiotics all 5 would not be normally prescribed at once for the same condition.

Some interactions are beneficial eg omeprazole with drugs like dichlorofenac to prevent the latter attacking the stomach lining.

In some cases different routes into the body can be used (oral, intramuscular, intravenous, patches, inhalation, suppositories etc) can be used to avoid a potential unwanted interaaction.

The body itself also plays a part, as only some people experience any given interaction.

As a result of this latter, there is a comprehensive (in the UK at any rate) reporting system for any new information that arises.

 

Sadly despite all this, something serious occasionaly goes unanticipated.

 

Doubtless others will have additional views on the subject.

[exchemist]

5 hours ago, kenny1999 said:

There are possible up to hundreds of thousand drugs used in medicine, then there are just too many combinations between two drugs. Do they test each new drug with hundreds of thousands of other approved drugs before introducing the new drug to the market?

How about three or more drugs? I mean it's not uncommon that a patient may be taking more than one drug because they may have a lot of diseases. How do they know if the new drug will not have problem when it is used together with possibly two or more drugs? For example, drug A + drug B + drug C or even with drug D or more. There are just countless combinations.

From a brief web search on the subject of drug-drug interactions, it looks as if this subject is a real issue in modern medicine  and not tremendously well controlled. For instance I found this: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3500158/.

While it looks as if quite a lot can be predicted by knowing the modes of action of a given pair of drug on the body, clearly this is not really sufficient. I tried to find out how the regulations for drug approval handle this issue but the relevant pages from the FDA seem not to be readily available, while I could not find anything on the UK's MHRA, apart from the "Yellow Card Scheme", which is for reporting adverse events seen in clinical practice, i.e. after drug approval has already been granted.

I'd actually be very interested to see comments from one of our forum experts in this sort of thing. Perhaps @CharonY may know something about this. 

 

 

Edited July 30, 2022 by exchemist

[CharonY]

On 7/30/2022 at 2:52 AM, exchemist said:

While it looks as if quite a lot can be predicted by knowing the modes of action of a given pair of drug on the body, clearly this is not really sufficient. I tried to find out how the regulations for drug approval handle this issue but the relevant pages from the FDA seem not to be readily available, while I could not find anything on the UK's MHRA, apart from the "Yellow Card Scheme", which is for reporting adverse events seen in clinical practice, i.e. after drug approval has already been granted.

There are quite a number of ways to predict and/or focus on specific drug-drug-interactions, but it is correct that generally a complete knowledge of all potential interactions is not required or even achievable.

Very roughly speaking, a focus is on transporters, which determine the transport of substrate to and from tissues, as well as enzymes, that together determine when and where and how much of a given drug distributes through your body. If a drug therefore alters enzyme or transporter abundance, they can the pharmacokinetics of other drugs.

One way to look at those is to use known index perpetrators (known substances that are either inhibitors or inducers for certain transporters and pathways) and see if a given drug has changed characteristics.

Another way to priorize interaction studies if certain drugs are likely going to be prescribed in tandem. In order to predict issues there are also so-called physiologically-based pharmacokinetic models, which the gold standard theoretical framework to evaluate potential drug-drug interactions, though in recent times machine-learning approaches are on the rise (no idea how they are evaluated by regulators, though).

And as you noted, sometimes interactions are only seen after they are already use in practice. In short, there is only so much we know and/or can predict when it comes to drug-drug interactions (and there is also the related area of drug-food interactions) and at least to my knowledge (which is at best peripheral) there is no perfect pipeline to assess those.

[exchemist]

10 hours ago, CharonY said:

There are quite a number of ways to predict and/or focus on specific drug-drug-interactions, but it is correct that generally a complete knowledge of all potential interactions is not required or even achievable.

Very roughly speaking, a focus is on transporters, which determine the transport of substrate to and from tissues, as well as enzymes, that together determine when and where and how much of a given drug distributes through your body. If a drug therefore alters enzyme or transporter abundance, they can the pharmacokinetics of other drugs.

One way to look at those is to use known index perpetrators (known substances that are either inhibitors or inducers for certain transporters and pathways) and see if a given drug has changed characteristics.

Another way to priorize interaction studies if certain drugs are likely going to be prescribed in tandem. In order to predict issues there are also so-called physiologically-based pharmacokinetic models, which the gold standard theoretical framework to evaluate potential drug-drug interactions, though in recent times machine-learning approaches are on the rise (no idea how they are evaluated by regulators, though).

And as you noted, sometimes interactions are only seen after they are already use in practice. In short, there is only so much we know and/or can predict when it comes to drug-drug interactions (and there is also the related area of drug-food interactions) and at least to my knowledge (which is at best peripheral) there is no perfect pipeline to assess those.

Thanks, very helpful.

So, if I try to summarise, as it is impossible to evaluate all drug-drug combinations, knowledge of the mode of transport and mode of action of a drug is used to predict what interactions with other drugs might be expected, and priority is given to checking these combinations. And then, after introduction of the drug, there is a catch-up process, to flag any further interactions discovered in clinical practice.

What seems still unclear is what requirements there are, if any, to check out potential interactions as part of the regulatory approval process.