Dr. Sandson, it seems that drug-drug interactions (DDI) were in our profession’s consciousness more in the past than they are now. Is this because we have less to worry about than we used to?
Dr. Sandson: I think that there is a false sense of security, partly because with the increasing sophistication of some of the drug interaction computer programs, people have decided that that the computer is going to figure out for them if there are any interactions. But most of these programs are overly sensitive but not really specific, so users end up with alert fatigue and may not pay as much attention as they should to a potential interaction.

What are some of the major interactions among psychiatric drugs that we should be aware of?
Dr. Sandson: Certainly among the antipsychotics, mood stabilizers, and antidepressants, there is plenty to be concerned about.

Are there any particular antipsychotics that we need to worry about?
Dr. Sandson: The most pervasive issue to worry about is smoking. Smoking is a potent inducer of P450 1A2 [sometimes called the CYP1A2 enzyme]. This means smoking has the capability to cut drug levels of almost every typical antipsychotic, as well as the atypicals olanzapine (Zyprexa) and clozapine (Clozaril), roughly in half.

Quite a lot of schizophrenic patients smoke. So how should we deal with this?
Dr. Sandson: I don’t think that this represents any absolute or relative contraindication to using any of these drugs. It is just something that the clinician should be aware of in terms of  appropriate prescribing. A typical scenario might be when a two-pack-a-day smoker who is on clozapine gets admitted—not because of treatment noncompliance, but for some other reason. With the cessation of smoking, enzymes in the body return to their lower baseline quantitative level, which produces a net decrease in metabolic activity vis-à-vis this drug, and then the blood level can rise quite rapidly. This could lead to side effects or something more striking, like a seizure.

Conversely, I assume when patients start smoking, their serum levels decrease.
Dr. Sandson: Yes, and that is more insidious and easy to miss. Let’s say you have someone in the nonsmoking hospital environment, and you have him or her pretty well treated with olanzapine 20 mg a day, for example. He or she makes no bones about resuming a two-pack-a-day habit the nanosecond after discharge. At that point, a predictable problem is going to arise, when a few days after discharge his or her antipsychotic blood level is cut in half. And that is something that we as clinicians really need to be mindful of and anticipate. We need to ask people: Are you going to resume smoking? Are you interested in help quitting?

What should we do about those patients?
Dr. Sandson: If I really believe that the current dosage is appropriate and that a significant cut in the antipsychotic blood level would be deleterious to the patient, I would actually increase the dosage as he or she walks out the door. Just give an extra 10 mg to grow on—he or she might be a little groggier the first two or three days, but it will pass.

Does it matter how much a person smokes?
Dr. Sandson: I believe it is a dose-dependent phenomenon, although we don’t have any firm means of quantifying this.

Are there any other clinically relevant DDIs that come up with antipsychotics?
Dr. Sandson: When you are coadministering potent enzymatic inhibitors, such as fluoxetine (Prozac) or paroxetine (Paxil), you can get new side effects, like EPS and hyperprolactinemia, due to increases in antipsychotic blood levels. Even ethinyl estradiol, often found in birth control pills, can produce increases in clozapine levels. Levels of antipsychotics can be reduced by drugs that induce enzymes that catalyze their metabolism, such as anticonvulsants like phenytoin (Dilantin), carbamazepine (Tegretol), or phenobarbital (Luminal); or antituberculosis drugs like rifampin (Rifadin).

What about drug interactions involving mood stabilizers?
Dr. Sandson: Carbamazepine is one of our most frequently used inducers. Its presence in patients’ regimens can lead to substantially lower blood levels of many drugs. lamotrigine (Lamictal) is a little dicey in terms what will raise or lower blood levels; in particular, valproic acid (Depakote) can double lamotrigine levels. While ethinyl estradiol can increase the effects of many antipsychotics, it acts as an inducer of lamotrigine’s metabolism and thus decreases levels of lamotrigine, as do phenytoin and phenobarbital.

What kind of drug interactions occur with lithium?
Dr. Sandson: Our most common problem with lithium involves caffeine, which has a very significant influence on lithium levels. With an increase of maybe two more cups of coffee than is usual, you can drop lithium levels by about half. If a patient has his or her level drawn before the morning coffee, it may not be reflective of the rest of the day. Likewise, if you are administering the lithium based upon a period of high caffeine consumption and then your patient decides to go caffeine free, this could produce lithium toxicity.

Anything else to worry about specifically with lithium?
Dr. Sandson: Most diuretics can potentially derange lithium levels. The thiazide diuretics will increase it by 25% to 40% on average, whileosmotic diuretics like the xanthenes drop lithium levels. Associated with diuretics are the drugs that affect angiotensin and aldosterone balance—like ACE inhibitors and angiotensin receptor blockers (ARBs)—which also increase lithium levels by about 25% to 40% on average.

Another big one with lithium is nonsteroidal anti-inflammatories, right?
Dr. Sandson: The nonsteroidals, with the exceptions of aspirin and sulindac (Clinoril), tend to increase lithium levels to a variable degree, anywhere from 20% to 200%, and one can’t reliably predict the magnitude of this interaction. Occasional use is no big deal, but if your patient is taking an NSAID on a standing basis, I think you should prospectively decrease the lithium level by a third, and then recheck that lithium level a week after the nonsteroidal has been started.

Are there concerns with psychostimulants like methylphenidate or the stimulant-like drugs such as modafinil (Provigil) and armodafinil (Nuvigil)?
Dr. Sandson: Stimulants are generally more often victims—their levels are altered by other drugs—than perpetrators. Stimulant blood levels can be raised by any of the potent CYP2D6 inhibitors, such as fluoxetine, paroxetine, or bupropion. However, these drugs have high therapeutic indices, so this is often not a compelling clinical concern.

I understand that Provigil and Nuvigil can induce the metabolism of the phosphodiesterase drugs for erectile dysfunction, like Viagra, Levitra, and Cialis.
Dr. Sandson: Theoretically, this could be a meaningful interaction, since the PDE inhibitors are metabolized by the CYP3A4 enzyme. However, I have yet to uncover, despite having looked for it, any case reports that demonstrate 3A4 inducers depriving phosphodiesterase inhibitors of their efficacy. What I can tell you conversely is that there is a literature about various CYP3A4 inhibitors, such as ritonavir (Norvir) and grapefruit juice, greatly increasing phosphodiesterase inhibitor levels, but apparently this is well-tolerated and although it is demonstrable, it does not seem to be something that is clinically significant.

All these interactions are obviously tough to keep track of. In which patients should we be most vigilant of DDIs?
Dr. Sandson: I think we are moving into the very important domain of clinical wisdom at this point. It is not enough to know that a DDI is possible or even actual, but it all needs to be taken in the context of: how pharmacodynamically hardy is my patient? For example, what is the therapeutic index of the victim of the drug interaction? If I have a young, very healthy patient and I have a bunch of drugs, none of which has a particularly low therapeutic index, there might not even be a particular need to evaluate this list with great rigor, because the worst case scenario is not very bad at all. The story can be very different with people who are more medically compromised, or who are taking drugs that can have much more dire effects or for whom the stakes are higher if there is a lack of efficacy. So the lens through which one evaluates the DDI profile for each patient is going to differ based upon the characteristics of each and every patient.

Is there anything we should be aware of in terms of prodrugs?
Dr. Sandson: The analgesic prodrugs like tramadol, codeine, and hydrocodone are particularly important. These drugs rely upon conversion via the enzyme 2D6 into active analgesic products. So any enzymatic inhibitors will thwart that and make these drugs less effective. These enzymatic inhibitors include fluoxetine, paroxetine, and bupropion. On the other hand, it has been found that if someone is a nonresponder to the prodrug blood thinner clopidogrel (Plavix), co-administering something like rifampin or St. John’s Wort can change him or her into a responder by virtue of the CYP3A4 inductive capabilities of these drugs.

Very interesting. Any final words of wisdom on drug interactions?
Dr. Sandson: The best DDI stories are not the ones where someone gets into trouble and after the fact you retrospectively evaluate how and why it happened. The best stories, albeit far less dramatic, are the ones where you anticipate the problem, prevent it, and nothing happens. It doesn’t have as much eye-appeal, but certainly that is a victory for the physician, and most importantly, for the patient.

Thank you, Dr. Sandson.