Dr Rubin reviews his presentation from the American College of Gastroenterology Scientific Meeting on therapeutic drug monitoring (TDM), including the appropriate application of proactive TDM in the treatment of inflammatory bowel disease.

David T Rubin, MD, FACG, is the Joseph Kirsner Professor Medicine, chief of gastroenterology, hepatology, and nutrition, and director of the Digestive Diseases Center at the University of Chicago.

TRANSCRIPT:

Hello, I'm Dr. David Rubin from the University of Chicago, and I'm at the American College of Gastroenterology annual meeting in Las Vegas, Nevada.

During the postgraduate course, I presented an update on therapeutic drug monitoring in inflammatory bowel disease. This is an area of immense challenge and some controversy related to how we can apply obtaining serum concentrations of some of our biological therapies in an attempt to improve response, preserve response, and reduce loss of response over time.

In my presentation, I provided a number of new updates. One was starting with our old friend, the thiopurine-based therapies. I shared that there is a second genetic polymorphism that predicts early leukopenia.

The polymorphism that most people are quite aware of is TPMT. That's been around for many years, but more recently, has been described NUDT15, which predicts early leukopenia and has a prevalence in those with Asian ancestry of between 10 and 13%. It is, however, found in some people who are from other areas of the world as well.

This new polymorphism is available to be tested, but the practical approach recognizing that this may complicate use of the therapy is to start with a low dose of the thiopurine, if you're going to use it, and check a CBC after a week or two, and then slowly increase your dose while monitoring them.

The other update I shared was from a study in the United Kingdom called the PANTS study. The PANTS study was a prospective study to determine some of the risk factors for primary and secondary nonresponse of infliximab and adalimumab.

In addition to reporting out those risk factors, the authors and investigators also identified a unique haplotype, HLA-DQ1*05, which was associated with an increased risk of immunogenicity—developing antidrug antibodies. This particular haplotype increased the risk of having antidrug antibodies almost 2-fold over those who did not have the haplotype.

While this is of great interest, I also commented in my lecture that it's unclear if that's sufficient to change practice yet. This will be available as a commercial test that you can order for your patients.

At the current time, with the available data on anti-TNF therapy, where we've recommended that most patients, when they start therapy, would use combination therapy with an immunomodulator to avoid immunogenicity, it's unclear whether knowledge that somebody doesn't have that haplotype would take away that strong evidence that we've used to make those recommendations.

My general recommendation to the audience was that they should stay tuned about this particular test and that we'll be using it to understand future approaches to management.

I then spent some time reviewing 2 additional approaches to therapeutic drug monitoring in patients with IBD. In the first approach, it was the traditional approach that we've learned and talked about, which is reactive therapeutic drug monitoring. This is when patients are on a therapy and lose response, or have an inadequate response while they've started therapy.

In this scenario, I outlined 3 specific questions that you should be asking. The first one is, is the patient actually inflamed when this is happening? In other words, are there symptoms due to active inflammation as opposed to other things, like scar tissue, or potentially, an infection, or a bacterial overgrowth?

The second question is, are they infected? Remembering that patients with inflammatory bowel disease have a higher risk for certain enteric pathogens and infections, specifically Clostridioides difficile, but there's the possibility of other things, including CMV and general gut pathogens.

The third question you must ask when people are not responding to therapies is, where is the drug? If they're truly inflamed, they're not infected, then you say, is the patient taking their therapy as prescribed? If they are, is the drug getting cleared too rapidly, or is it perhaps being metabolized too rapidly? Is something happening to the therapy?

That's where reactive therapeutic drug monitoring might apply. The majority of data exist for anti-TNF therapies and, therefore, we have some algorithms that tell us what to do.

If you have a patient who's lost response, and, for example, they're on infliximab, and you measure a trough concentration of the drug, and you find out that there's no drug present, and the patient has developed antidrug antibodies, this is a sign that the reason they're not responding is because the drug is being cleared by those neutralizing antibodies and you should move on to another therapy.

If the drug was working, you can stay within the anti-TNF class, but a major lesson we've learned is that you should then make sure that when you use a second anti-TNF, you do combination therapy so that you can prevent antibodies from occurring again.

In the second scenario for reactive therapeutic drug monitoring, when you measure the anti-TNF drug level, if you find out that there's lots of drug present -- I went through in my lecture the different levels you might consider —in general, let's say that the drug level is more than 10 or even 20, and the patient's inflamed and not infected, then, you might say, "The drug is present, the patient is inflamed, this mechanism must not be working." You would switch to a different mechanism.

The third scenario is where it's mixed, where you get a drug level that might be low, there might be some antibodies forming, and you can't tell, "Are these antibodies the reason the drug level's low, or is it the disease, or a combination of factors?"

In this scenario, sometimes we need to follow with a second drug level to see where things are trending. Sometimes, we would consider then adding an immunomodulator.

I made the point that if the drug level is zero, adding an immunomodulator won't help, but if the drug level is detectable with some antibodies, adding an immunomodulator has been shown to suppress those antibodies and let the drug level go back up.

The other major area I discussed therapeutic drug monitoring was one that's more controversial and has some confusing evidence. That has to do with doing proactive therapeutic drug monitoring. In other words, in a stable patient or one who you're starting therapy on, measuring a drug level to know what's going to happen to them in the future.

There are some strong retrospective data from Boston in the Beth Israel Deaconess Hospital that demonstrate keeping the infliximab level in a range between 5 and 10 at trough led to increased persistence of that therapy and reduced hospitalization and surgery, compared to waiting until the patient was sick, and then trying to adjust the drug and figuring out what to do next.

That's very compelling. However, there have been several prospective attempts to do this with adjustments in drug level or dosing changes that have failed to demonstrate the overall benefit of this approach.

The way I tried to explain it to the audience is, I said that I think it's reasonable to use this in your patients who are high-risk for loss of response or who have a lot to lose if you get it wrong, meaning that they're going to face surgery or we've run out of drug options, so you want to make sure this one is working. In the usual patient who's very stable on therapy, I didn't advocate for us using regular drug levels all through their maintenance strategies.

The other place where this has some good evidence is when you're doing loading. Predominantly, this is with infliximab, although there are some emerging data with ustekinumab. During loading of the drug, if you check a level of infliximab, for example, at week 6 before the third loading dose, or even at week 8, it can be correlated to the subsequent ongoing response at weeks 30 and at 1 year.

Knowing what the drug level is early is assumed to then help us know if we should do proactive dose adjustment to keep them on therapy out to a year. I say assumed because as much as that makes sense, there hasn't been a study yet that showed that that early level, followed by a proactive dose adjustment, leads to that subsequent outcome. There's some areas here that need further clarification.

I ended with one additional pearl for the audience related on some emerging evidence, which is how to restart a therapy after a drug holiday. The dogma had been that if an anti-TNF was discontinued, it would therefore mean you couldn't go back to it because of a high risk of antidrug antibodies. We've now learned a few things.

Number 1 is that vedolizumab and ustekinumab both have such low immunogenicity that these drugs appear to be safe to be used as monotherapy and also may be restarted without much concern about developing antidrug antibodies. The majority of evidence for the previous concerns was with anti-TNF.

I shared an algorithm that Maria Abreu and I developed and have shared our overall experience from the University of Miami and the University of Chicago, showing that when you restart an anti-TNF after a drug holiday of more than 6 months, after you give 1 loading dose—so an infusion or the injection of an injectable anti-TNF—7 to 10 days later, you check a drug level. You wait for that result before you give them their second loading dose. If there's drug present and no antibody, you can continue, and it's likely they'll tolerate the therapy and do well.

If there are antidrug antibodies and no detectable drug, then you've induced what's called an anamnestic response, which is the immune system waking up to this exposure of antigen and you should not continue and it's time to move on to another therapy.

This restarting of therapy after a drug holiday is the last point I made about how to use therapeutic drug monitoring.

I know I've spoken about quite a bit. If you're interested, the lectures that were presented at the postgraduate course will also be available through the ACG online Universe, so stay tuned for that as well. Thank you very much.