TCR: Dr. Sandson, you've just published the definitive casebook on drug interactions (1), and you obviously have thought long and hard about the issue. What do you think psychiatrists really need to know about drug-drug interactions?
Dr. Sandson: It's hard to imagine practicing safely without at least a basic familiarity with recurring, common P-450 inhibitors, inducers, and substrates. I'm not talking about the extensive roster of several hundred agents, but in psychiatry we use about 20 drugs over and over again, and there might be another 10 or 15 drugs that would commonly run in company with the psychotropics we use. We should focus on these agents, and especially on those that have low therapeutic indices.

TCR: Can you define what you mean by "low therapeutic index"?
Dr. Sandson: A drug where the ratio between a lethal dose and an effective dose is relatively low, like Lithium. These are the medications where it pays to know your interactions backwards and forwards. For the general psychiatrist, it's not worth using up brain space on drug interactions that aren't likely to be of clinical significance. For example, I rarely discuss Effexor (venlafaxine) or Remeron (mirtazepine) when I give talks on drug interactions, because they are unlikely to be involved in clinically significant situations.

TCR: Isn't it also true that certain drug interactions that we are often warned about, like the protein binding interaction between Zoloft and Coumadin, turn out to be rarely of clinical significance in our patients?
Dr. Sandson: That touches on an important point, namely, variability. If you combine two different agents across the population, it may be that in only 10 to 15% of all those individuals is it going to produce a clinically meaningful interaction. Faced with that kind of statistic, some clinicians are going to say, "That's an unacceptably high incidence rate and I need to orient my practice such that I avoid or minimize that interaction." And then you have another group of clinicians who have used this combination a number of times, and, by chance, have never seen this interaction, and who therefore minimize it and say, "This isn't anything I need to be overly concerned about." So there's going to be a range of responses. I personally would be in the more conservative group of psychiatrists who would check a Prothrombin Time (PT) before and after adding an SSRI (other than Luvox, Celexa, or Lexapro, which have low protein binding) to Coumadin.

The Concept of “Variability”
“If you combine two different agents across the population, only 10 to 15% of all those individuals are going to show a clinically meaningful drug interaction.”

TCR: Are there any other important protein binding interactions that might be underappreciated in psychiatry?
Dr. Sandson: Yes, there's an important interaction between aspirin and Depakote. When aspirin is given at antipyretic doses you can increase the free fraction of Depakote up to four-fold. I've seen patients who have been chronically confused on this regimen, and when you check their free valproate level, it's 40, which is four times the upper limit of the standard free valproate level.

TCR: When we normally order a valproate level, we're given the total level?
Dr. Sandson: Yes. And in this particular patient, the total level, which represents the bound plus the unbound portion, was in the 70s, which was "therapeutic."

TCR: Do you have a list in mind of drugs in psychiatry that we should be most aware of in terms of plasma protein binding interactions?
Dr. Sandson: The ones to be most aware of are Dilantin, Digoxin, aspirin, Coumadin, Depakote, Prozac, Paxil, and Zoloft. They are all highly protein bound.

TCR: So what do you recommend we do in those cases of patients who are on Coumadin or Digoxin, in whom we are about to prescribe an SSRI?
Dr. Sandson: I would say that unless you have some history-based reason to go with Prozac, Paxil or Zoloft, that you go with one of the other SSRIs.

TCR: What about the other new generation antidepressants. Are they highly protein bound?
Dr. Sandson: For the most part, they are not. Nor are the tricyclics or the MAOIs.

TCR: Well this is a beautiful list, very manageable. But moving to the more complicated issue of P450 interactions, this is a big, huge, confusing topic. Help us!
Dr. Sandson: It is a confusing topic. Many clinicians are appropriately relying on auxiliary brains, like epocrates (epocrates.com), micromedex (micromedex.com), physicians on-line, and other programs that do drug-drug interactions. I'd also recommend two excellent websites. The first is www.drug-interactions.com. It's maintained by David A. Flockhart, MD, PhD, at the Indiana University School of Medicine. If you click on each medication in the table, it will hyperlink you to a literature reference. My colleague, Dr. Jessica Oesterheld, runs another wonderful website, www.mhc.com/Cytochromes, which provides some of the most current and detailed information on P450-based interactions, as well as information on phase II metabolism and the P-glycoprotein transporter.

TCR: I have to say that I've been finding your new book extremely helpful clinically. I look in the index and look for vignettes related to the drug I'm prescribing, and it clues me into what symptoms to look for. Plus, it's much more interesting to read a story than a chart! But moving on to the dreaded P450 system, let's start with 2D6.
Dr. Sandson: The 2D6 inhibitors that we need to watch out for are Prozac, Paxil, high dose Zoloft, and Wellbutrin.

TCR: Wellbutrin? That's interesting, because people tend to think that Wellbutrin is completely clean.
Dr. Sandson: Exactly, and it came off the "good" list in the year 2000, which is one of the few times we have a drug company to thank for being honest and forthright-the good folks at Glaxo cued us into that aspect of this drug. The most pertinent 2D6 substrates are beta-blockers, tricyclics, and a couple of analgesics, like hydrocodone (Vicodin and others) and Ultram (tramadol).

TCR: So that's the cast of characters. What's the plot?
Dr. Sandson: The plot is that when you add an inhibitor to a substrate, you will increase the blood level of the substrate significantly. There are numerous examples of clinicians adding Prozac to beta-blockers, causing people to become hypotensive, or someone being put on Paxil when they're on tramadol and not achieving adequate analgesia, because tramadol has to be converted to M1 for adequate pain control.

TCR: What should we do when we add an SSRI to a beta-blocker?
Dr. Sandson: You either very carefully monitor pulse and blood pressure, or you instruct them to have their primary care doctor monitor this. You can also advise people that if they feel acutely dizzy to call you or their internist. Another option is for you to communicate with the internist before starting the SSRI and see if it makes sense to prospectively decrease the beta-blocker dose. Of course, one may also select a non-2D6 inhibiting SSRI, such as Celexa (citalopram) or Lexapro (escitalopram).

TCR: What about 3A4?
Dr. Sandson: It's difficult to give a concise list of substrates, because there are so many of them. Some of the more important ones are calcium channel blockers (eg, diltiazem), tricyclics, HIV protease inhibitors, methadone, Ambien, Sonata, Risperdal, and Seroquel. But it's more important to know the inhibitors and inducers of 3A4. Important inhibitors are Prozac, Luvox, Serzone, all macrolide antibiotics with exception of zithromax, and certain antifungals like ketoconazole. Grapefruit juice is a potent inhibitor and this has led to its exclusion from many hospital diets.

TCR: Do you ask your patients if they drink grapefruit juice?
Dr. Sandson: I usually do, especially if they are on a 3A4 substrate. 3A4 inducers include many anticonvulsants, such as Phenytoin, Tegretol, to a limited degree Trileptal (oxcarbazepine), phenobarbital, rifampin, St. John's Wort, and Topamax (topiramate). Next in importance is 1A2, with substrates including Clozapine, Zyprexa, and caffeine. 1A2 inducers include Prilosec (omeprazole) and rifampin, but cigarette smoking is the biggest issue here. Luvox is a huge inhibitor of 1A2 and is the key one we need to be aware of.

TCR: Luvox seems to come up over and over again in drug interactions.
Dr. Sandson: It's true: Luvox is a mischievous medication, and I have to have an awfully compelling reason to use it!

TCR: What about the “2C's”?
Dr. Sandson: Dilantin is both substrate and inducer of 2C9 and 2C19; Depakote and Luvox are 2C9 inhibitors, and important substrates include many oral hypoglycemics (glipizide, glyburide, glucotrol) and Coumadin.

TCR: Any parting words of wisdom?
Dr. Sandson: I would say that SSRIs interactions are the things psychiatrists should really know cold, in terms of what you should commit to memory. It's a very good investment of brain space. Psychiatrists should train themselves to become familiar with some core elements and then have a source that they trust and turn to in rapid manner to look stuff up, even with the patient in the room.

^{1. Sandson NB: Drug Interactions Casebook: The Cytochrome P450 System and Beyond. Washington, DC: American Psychiatric Publishing, Inc., 2003.}