Our polypharmacy state: understanding drug-drug interactions (DDIs) in the drug development process

Updated 2020. The FDA has now approved the DDI regulations. Below is the article as originally published.

In 2018 the FDA and several other regulatory bodies are expected to enact and enforce more laws related to the requirements for nonclinical drug-drug interaction (DDI) studies prior to an IND. A key component of developing a drug for market-use is to test for how it might interact with other drugs or substances in the body.

Here’s an overview of what a DDI is and how it effects the drug development process.

What is a drug-drug interaction (DDI)?

To help us understand the importance of DDI studies, we need to meet Tony, a made-up patient. One day Tony walks into his physician’s office and, after a thorough consultation with his doctor, is prescribed fluvoxamine, a common selective serotonin reuptake inhibitor (SSRI) used globally to treat obsessive-compulsive disorder and other major depression and anxiety disorders. Tony also struggles with gastric reflux and uses over-the-counter omeprazole to treat heartburn. And, like many of us, Tony loves real coffee, not that infamous decaf variety.

Tony knows that his morning work meetings are usually better with his caffeine boost and his after lunch meetings usually go much better with pretreatment of omeprazole for his favorite food. But this week Tony feels extra jittery and irritated in the morning and has had a consistent afternoon headache despite following his normal routine.

So what gives? It has to be the side effect of that new drug the doctor prescribed, right? It’s true. Tony’s discomfort is related to the new medication, but it is not directly caused by fluvoxamine itself. It is actually fluvoxamine’s interaction with his caffeine and omeprazole intake that changed his pharmacokinetics, resulting in much higher drug levels than Tony’s body was used to.

The overall effect of this interaction may be dependent on the patients CYP2C19 genotype (Yasui-Furukori, Takahata et al. 2004). Figure 1 displays how Tony’s body may have metabolized the drug differently over time, based on how omeprazole interacted with his coffee and the fluvoxamine for someone with Tony’s genotype.

A phone call to his doctor and a new prescription in hand, Tony’s life is back to normal, but what if Tony’s saga could have been avoided?

Is it possible to know how a drug compound might interact with any and every substance—or a myriad of substances—prior to being released to market?

A multiplicity of polypharmacy patients

The specific DDI described in Tony’s story is well-documented, but unfortunately attracts little attention, partly because of the increasing number of variables that need to be tested.

As our population ages, approximately 33 percent of patients are considered polypharmacy (Bjerrum, Gonzalez Lopez-Valcarcel et al. 2008). And the older you get, the higher the percentage is that you will be taking at least five different medications (Charlesworth, Smit et al. 2015):

65-69 ages: 25% of population have at least five different medications
70-79 ages: 46% of population have at least five different medications
Furthermore, it is not uncommon for some patients to be prescribed more than 20 different medications at a time.

Imagine how difficult it is for both the patient and the physician. The physician has to manage the healthcare for a patient who takes more than 20 different medications (provided, of course, that the patient has reliably disclosed all their medications, both prescribed and over-the-counter). And the patient has to keep track of multiple pages of instructions in print that is nearly unreadable: “Take this medication in the morning; this one with food; and this one without food.” While the efficient and ubiquitous daily pill boxes can help patients, the sheer complexity of this dosing regimen causes great confusion and opens the door to significant DDI risks.

Polypharmacy patients are at an increased risk for major drug-drug interactions

It may not be a surprise given our discussion that many patients have experienced a problem with how their drugs interact with each other and other substances (like coffee) in the body.

In fact, the percentage of patients experiencing both minor, moderate and a major drug-drug interaction increases based on the number of medicines prescribed (Bjerrum, Gonzalez Lopez-Valcarcel et al. 2008).

In some of these situations, the problem is linked to a serious injury or death. Unfortunately, when there is a problem with a medication, the physician may mistakenly treat the symptoms of the DDI with a new medication, compounding the problem even further. The fact is that in many of these cases, the problem was preventable.

Thankfully, Tony, in our example, contacted his physician right away after experiencing symptoms. But what if he hadn’t?

How could this DDI be prevented? Was the DDI due to a lack of clear information in the package insert? Did the patient disclose all over-the-counter medications? This is clearly a complex problem, so where should we start?

Earlier DDI studies and predictive PK/PD modeling might help

Let’s start with something within our control as scientists: the data being generated to support the development of drugs.

In order to improve the predictability of drug-drug interactions and to better inform physicians, patients, and regulatory agencies of potential problems with a drug under specific circumstances, you should conduct DDI investigations earlier in the development of new drugs, preferably prior to the IND submission.

The FDA seems to agree with this model. Last fall, the FDA proposed new in vitro drug transporter regulations requiring that all DDI studies be completed prior to an IND being submitted (instead of waiting until later in the drug development process). These regulations are widely expected to go into effect by the end of 2018.

Another tool available to help determine a compound’s potential interaction with other substances is the use of predictive PK/PD modeling and simulation of preclinical tests prior to clinical or FIH studies. Population modeling can help predict how a drug might interact within a specific population or variation of a population.

The polypharmacy trend won’t be disappearing any time soon; however, it’s our responsibility drug developers to understand how a potential new substance will interact with other drugs.


Bjerrum, L., et al. (2008). “Risk factors for potential drug interactions in general practice.” Eur J Gen Pract 14(1): 23-29.

Charlesworth, C. J., et al. (2015). “Polypharmacy Among Adults Aged 65 Years and Older in the United States: 1988-2010.” J Gerontol A Biol Sci Med Sci 70 (8): 989-995.

Yasui-Furukori, N., et al. (2004). “Different inhibitory effect of fluvoxamine on omeprazole metabolism between CYP2C19 genotypes.” Br J Clin Pharmacol 57(4): 487-494.

You may also like...

Popular Articles...