Question And Answer Session

January 3, 2006 Presentation:
The Role Of Pharmacogenomics In Bipolar And Other Psychiatric Disorders

Welcome to AACC’s Expert Access Live Online program

The Role Of Pharmacogenomics In Bipolar And Other Psychiatric Disorders

This month's expert is Neil Sandson, MD. View the presentation and direct your questions to our online expert. AACC would like to thank Bayer HealthCare Diagnostics for making this program possible.

You refer to azithromycin and pravastatin as "safer" somatic agents. What makes them safer?
Washington, DC

Neil Sandson: Azithromycin is "safer" than its fellow macrolides (erythromycin and clarithromycin) in that it is not an inhibitor of either P450 3A4 or P-glycoprotein, and therefore less apt to increase the levels of any 3A4 or P-gP substrates. The other macrolides are potent inhibitors of both. Ooops. I forgot to address the pravastatin part. Pravastatin relies on phase II metabolism and renal excretion, as it is thus less susceptible to DDIs that would increase it's levels and thus predispose to the rhabdomyolysis that can occur with statins. Atorvastatin, simvastatin, and lovastatin are all 3A4 substrates and fluvastatin is a 2C9 substrate. Inhibitors of these enzymes can predispose to the rhabdo problems when these statins are used.

One issue that's alluded to in your excellent presentation, but not expressed, is that clinical PGx testing has the potential for helping a medical facility save money by avoiding adverse drug interactions and streamlining therapy selection. Can you point us to some evidence-based literature that details these economic outcomes, or perhaps offer some of your own experiences in this realm?
Leeds, UK

Neil Sandson: Unfortunately, PGx is still too expensive to make true screening practical. Thus, we usually use it to make sense of intolerance or other adverse outcomes after the fact and avoid them in the future, rather than predict and avoid them in the first place. I don't have any great stories of how avoiding ADRs through PGx, but here's some evidence and case vignettes that might suggest some utility for PGx testing. 27 drugs are frequently cited in ADR studies. 59% (16/27) are metabolized by at least one enzyme with a variant allele causing poor metabolizer (PM) status. 69% of these drugs (11/16) are metabolized by P450 2D6. In contrast, only 7-22% of randomly selected drugs are metabolized by enzymes with this genetic variability. 100 consecutive admissions to state hospital in Kentucky genotyped for 2D6 side effects and total treatment costs tracked PMs (n=12) had more side effects (n.s.) UMs/PMs on 2D6 drugs cost $4000-6000 more to treat longer hospital stay for PMs (24d) than for EMs (17d) Need larger N to detect sig difs (1500-2000) Example 1 Over a 10 month period, a 9 year-old boy who was receiving fluoxetine for ADHD and OCD experienced progressive metabolic toxicity marked by bouts of gastrointestinal distress, low-grade fever, incoordination, and disorientation. Generalized seizures were observed, and the patient lapsed into status epilepticus followed by cardiac arrest and subsequently expired. On autopsy, fluoxetine and norfluoxetine levels were vastly higher than expected. Post-mortem dx: death due to fluoxetine toxicity. Initial suspicion was directed toward the adoptive parents who controlled access to medications. Further testing revealed a 2D6 genotype consistent with PM status, and the investigation of the adoptive parents was terminated. Example 2 A patient was given small doses of codeine for a cough due to bilateral pneumonia, soon after which the patient developed life-threatening opioid intoxication. Genotyping revealed 3 or more functional alleles for 2D6 (UM phenotype). Codeine transforms to morphine via 2D6 Codeine + 2D6 UM phenotype can yield increases in morphine levels up to 800%! Example 2 (cont.) This is an especially important example, as there were more prescriptions filled for hydrocodone in 2004 than for any other drug. Hydrocodone  hydromorphone via 2D6. The basic message is that is you have reason to suspect a polymorphism, perform testing, and know what substrates are metabolized by given enzymes (see below), then one can get some utility out of this testing. 2D6 – Notable substrates Anti-arrhythmics Aripiprazole ß-blockers Codeine  morphine Hydrocodone  hydromorphone Oxycodone (NOT a prodrug) Phenothiazines Risperidone Simvastatin SSRIs Tamoxifen TCAs Tramadol  M1 Venlafaxine 2C19 – Notable substrates Diazepam Omeprazole (and other proton pump inhibitors) Phenytoin Proguanil (prodrug) Sertraline Tertiary amine TCAs Another point is that at this time, I think we will get a lot more utility out of more frequent use of TDM than PGx testing.

If you'll allow me to chime in again, I would also be interested in any studies or personal experiences relating PGx testing to CLINICAL outcomes.
Leeds, UK

Neil Sandson: Please see previous response.

How pervasive is the a state dependent relationship of drug to benefit? In other words does the activity of a drug depend on the state on the state of a patient--same dose of valium has a greater effect if the anxiety is greater
gaithersburg, MD

Neil Sandson: This is a complicated question that I think combines 2 ideas that might be separate. "Benefit" and "activity" of a drug are related, but the former issue is much more complicated. One example that comes to mind is analgesia. It's much easier to treat pain "early" than when it has gained momentum, but I'm not at all sure that functionally relates to true drug "activity" per se. Certainly we have a healthy lore that acutely manic folks can tolerate high lithium levels (over 1.0) that then generate side-effects after the person has become euthymic. This question strays into the domain of pharmacodynamics rather than kinetics, and that is a harder realm than no one understands all that well.

Please comment also on the pharmacogenomics of marijuana and other 'self-medications' for bipolar disorder.

Neil Sandson: THC is a substrate of 2C9, but I'm not sure of its metabolism there (is it a prodrug for example?). I don't know of specific genetic vulnerabilities that link MJ use with bipolarity, although I wouldn't be surprised to learn that such exists. Bipolar patients are often not picky about finding substances to produce "state change", and I suspect there's nothing specific to MJ in this respect in this population.

My son has ADHD and Asperger's syndrome. He also has a tic disorder. He surrently is taking Stratera for inattentiveness. What are the other drug options for this that will not exacerbate his tics?

Neil Sandson: I'm not a child psychiatrist, but I believe that clonidine will prove helpful for both tics and ADHD, although the effect size is not tremendous. Actually, one can often get away with giving stimulants to kids with both ADHD and tic disorders and NOT exacerbate the tics, although it's always a risk. Additionally, one can give a stimulant and if tics emerge, then an antipsychotic agent (atypical preferably) can be added.

Hi There: I have a question to ask you regarding dosing for kids with attention deficit hyperactivity disorders. Although your presenation focussed more on phychiatric disorders, perhaps you may be able to provide some insight into this area. What do you think is the value of testing for polymorphisms in the DRD4 seretonin receptors for kids undergoing treatment with methylphenidate for ADHD. Several recent reports have provided contradictory evidences regarding the presence of the 7R repeat and methylphenidate dosage. Given that Ritalin dosage is quite extensive in the U.S. do you think it is prime time to develop a pharmacogenetic test for this treatment modality.

Neil Sandson: This isn't really my area, but here goes. When it comes to testing for polymorphisms that affect kinetic handling of drugs and that therefore exert an influence through higher or lower blood levels than expected, then I think that TDM is a lot more practical at this point. I do think there is an emerging and potentially valuable role for testing for polymorphisms that address pharmacodynamic differences, from tolerability at a given blood level to efficacy, but all of the psychiatric genetic testing is both in its infancy and only available if one is involved in a research protocol or BOTH rich and well-connected. In the tx of ADHD, I think we are a long way from gaining a definitive advantage through means beyond careful and judicious dosing by a skilled and attentive clinician, and responding to what happens as treatment progresses.

Thank you for an eye-opening presentation, Dr. Sandson. This is a little bit outside the scope of your materials, but should I be using CYP genotyping to titrate dosage of psychoactive therapeutics?
Cleveland, OH

Neil Sandson: If this info is available to you for whatever reason (prior testing, etc.), then it can be useful to know if a patient is a poor, extensive, or ultrarapid metabolizer for a given enzyme(s). However, I don't think we are yet at the point where it is advisable to perform widespread screening to obtain this info. It's too expensive and the info is rarely truly critical for appropriate care. Dosing should generally occur with the assumption that your patient is in the unlucky small percentage of poor metabolizers for the pertinent enzyme(s), and proceed from there. First do no harm!

My father (30 yrs bipolar) is currently taking Lithium, Seroquel, Trimodal (for pain) and Zantex (ulcer). Does this combination work okay? He was taking Hadol but his new dotcor took him off. What do you think of Hadol?
Vacaville, Ca

Neil Sandson: Haldol is not an intrinsically "bad" drug, but all other things being equal, seroquel is probably a better choice if it is felt that an antipsychotic should be part of his regimen. Given that he was taking tramadol, switching the haldol was a good idea, as the reduced haloperidol metabolite is a potent inhibitor of 2D6, and it would therefore be expected to interfere with tramadol's transformation to M1, with an accompanying loss of analgesic efficacy. The current regimen actually looks quite good.

My husband is bipolar (and so was one of his sister's and brother). Can my children become bipolar?
St. Marys, Ks

Neil Sandson: The short answer is yes. The exact likelihoods are difficult to impossible to determine, given the multifactorial nature of inheriting and developing the clinical illness. This is exactly the sort of question that a genetic counselor would be helpful with. I don't know where you are, but in my area I would contact Holly Peay at Greater Balt Med Center (443-849-2000 and ask the operator to connect you to her office). If you're not near Baltimore, then Holly should still be able to refer you along.

Does the litium level need to change during the 2 times/yr the bipolar cycle escalates?
Vacaville, Ca

Neil Sandson: I'm not sure I understand the question. If you mean that a particular patient has a tendency to get manic twice each year, then yes, it could be helpful to increase the Li level in response to the episode, or if the periodicity is quite predictable over time, then even conceivably increase the dose and level in anticipation of this to help prevent it. We are not talking about profound changes, however. An increase from say 0.8 to 1.0 or 1.1 would be about as far as I would advise.

Since any pharmaceutical compound may have dozens of genes that are required to produce a desired therapeutic result, it would seem that any single pharmacogenetic test might not offer a complete story of a patient’s response phenotype. Are SNP or multi-gene “panels” in the offing, and would this type of technology be practical for clinical use?
Sacramento, CA

Neil Sandson: Somewhere in our pretty distant future I'm sure something like what you are describing will exist. However, that is vary far away at this point. There are some tests, like for the promotor region of the serotonin transporter and variants in CRH receptors that seem to test for variants that seem to correlate with illness likelihood, severity, and responsiveness to specific (or any) agents. First we have to figure out what we are even really treating most of the time before we can make phenomenologic-genetic correlations that will yield pervasively useful and reliable testing.

When using Depakote and Lamictal for bipolar disease, does Seroquel have an effect? If so, what would one expect?
Boulder, CO

Neil Sandson: Generally speaking, quetiapine should help. It can be calming and help with sleep, it has an FDA antimanic indication, and folks from Astrazeneca are making cautious/coy statements about prospects for effectiveness in the depressive pole, although none of that has yet been published in a peer reviewed journal. It's a reasonable combo.

On slide 5: Inhibition of hydrocodone resulted in a loss of analgesia. Doesn't hydrocodone itself have analgesic activity? Typically, metabolism to hydromorphone is slow; shouldn't the analgesic effect of hydrocodone persist?
Willow Grove, PA

Neil Sandson: My understanding is that hydrocodone is only a fraction as potent a mu agonist as hydromorphone (dilaudid). As for the rate of the transformation, stats like hydrocodone half-life will only tell part of the story. I think generating even a little hydromorphone goes a much longer way analgesically than persisting higher levels of hydrocodone.

Does your facility test for all of the genes (or gene products) referenced in your presentation? Is this testing readily available from refernce labs?
Houston, TX

Neil Sandson: Heavens no. At this point, one can readily test for 2D6 and 2C19 via the amplichip. Even if your lab doesn't have the chip, they can readily send it to Quest or some other outfir that does, generally for $250 per enzyme. There are a number of commercial outfits that test for the range of P450 enzymes and some other enzymes, but stuff like P-glycoprotein and testing for various receptor polymorphisms is usually done in research labs. One notable exception is the Mayo clinic, which is very far ahead of the pack in terms of extensive PGx testing. If you want to see where the rest of the country will be in 20+ years, chack out what they are doing now (I assume they have a website).

In your opinion, what is the role (and or how critical) is the role of the clinical pharmacist in working with physicians to help patients avoid drug interactions and or suggest best possible drug/ dose to use?
Houston, TX

Neil Sandson: Pharmacists have the potential to be enormously helpful to clinicians, but they have to want to be helped. That takes time, interest, awareness, and initiative. Systems like the VA, where pharmacists are an integrated part of the treatment team derive great benefits from this, but it's the rare place that appropriately utilizes their pharmacists in the absence of such structured involvement. The range of ways that they can help is basically anything you would reasonably suspect.

Thank you for an excellent collection of interesting DDI cases. Given the complexity of pharmacogenomics related DDI, please comment on the merit of inter-disciplinary approach to bring pharmacogenomics to a clinical level for enhanced patient drug safety.
Milwaukee, WI

Neil Sandson: If we were more oriented toward patient safety (and we aren't that oriented toward it, despite the awareness-raising efforts of the Institute of Medicine), then such an approach would be critical. For instance, I can evaluate the pharmacokinetic likelihood of blood level changes given my eclectic approach to this topic, despite being a mere shrink. However, that still falls short of being able to evaluate the subsequent significance of these potential changes, let alone performing the risk-benefit analysis that would enable me to determine the advisability of accomodating a regimen that contains potential DDIs versus changing one or more of the agents in the regimen. However, this again takes time, interest, effort, and other such things that are often not deemed "time-effecient".

There is now an FDA approved test that can show a patients metabolizer status (Roche's AmpliChip CYP450 Test) for 2D6 and 2C19. How do you envision psychiatrists using the result of the test (i.e patient comes back as a poor metabolizer) to provide better therapy for patients? Is not having exact dosing recommendations going to be a hurdle for psychs? Extensive knowlegde like yours of drug drug interactions is not pervasive in every physician out there.
San Francisco, CA

Neil Sandson: I think you've hit on a key limitation of current PGx testing, whether for P450s or even the emerging tests for receptor polymorphisms, and that is that they suggest how to avoid problematic medication choices, but they arent that helpful in directing us toward superior choices. In short, they help tell us what not to do, but not what to do. I really believe that computer programs will have to be the answer to making this info available and helpful to most clinicians, but there is an intrinsic tension between completeness and efficiancy that makes it hard to find "one size fits all" programs. However, we are getting smarter at program design and I am hopeful that good programs of the future, integrated into electronic medical record and order entry systems, will go a long way toward predicting and preventing many DDIs.

You mention that lithium levels are decreased by caffeine; do you recommend restricted caffeine intake? Also, how about the affect of OTC supplements, such as omega-3; what effect, if any, do they have on lithium or would you recommend restricted use/

Neil Sandson: Omega 3s are fine. I recommend them. The evidence base is suggestive of benefit, and they certainly can't hurt. As for Li and caffeine, it's not so much that patients should caffeine restrict as they should be consistent in the amount of caffeine they should ingest while on Li. Abrupt increases in caffeine can decrease Li levels by 50% and the reverse is equally true. If "caffeine consistency" is too elusive for a given patient, then restricting or eliminating it from the diet is probably prudent.

Please disregard this question if it is a duplicate - my "send" function may have malfunctioned on the earlier question. Our company, NeuroMark Genomics, has a test available for the Serotonin Transporter Gene, the 5HTT-LPR. The short form of this polymorphism has been shown to be associated with slowed and reduced response to SSRI's, antidepressant-induced mania and higher susceptibility to stress leading to depression. There is significant published research regarding these indications. Interestingly, the reduced effect is compensated for by adding propranolol according to three or more studies. Do you think this test may be useful for psychiatrists or general practitioners treating unipolar and bipolar depression? It is not an expensive test, turnaround time is 24 hours. Kim Bechthold, CEO
Boulder, CO

Neil Sandson: I'm really happy to hear from you. I would love to know how to access this test, as I had believed it to be beyond the availability of most clinicians and patients. My e-mail is Please send me info! Also, I had not heard about the propranolol wrinkle, but I knew the other features that you mentioned in your question. I really do think there might be utility for a test like this. If a patient has failed one or 2 SSRIs, then it would be good to know that they had a ss genotype before wasting more time and effort on further SSRI trials (or really any antidepressants that rely extensively on 5-HT reuptake blockade, like venlafaxine). One study suggested that ss genotype patients do respond to mirtazapine, which makes sense since its mechanism of action does not involve 5-HT reuptake blockade. I'm less clear on its utility for bipolar depression, especially as I try to avoid antidepressants in treating bipolar depression, although not inflexibly so.

What is the interaction between Prozac and diazepam? Due to similar metababolic pathways, does the concurrent use increase the diazepam in-vivo half-life? Thanks.
Medford, OR

Neil Sandson: It is certainly true that fluoxetine increases diazepam levels and half-life through inhibition of P450 2C19 > 3A4. While the metabolism of flu is heterogenous, I'm not certain to what extent inhibition at these 2 enzymes is competitive and/or allosteric.

My son has a mild form of ADD. He has tried several different meds (Concerta and one other). He says he does not like the way they make him feel e.g. loss of appetite and a "hyped up feeling". What are your thoughts regarding the meds?
Walnut Creek CA

Neil Sandson: Well, clonidine or guanfacine (note to prior questioner, guanfacine, another centrally acting alpha-2 agonist, is another possibility besides clonidine for comorbid ADHD and tics that won't worsen tics) will not make him feel hyper, although it may prove sedating. Various antidepressants can also help in ADD, and there is a variety of levels of probable activation to choose from amongst these. It sounds like he doesn't like stimulants. As an isolated fact, that would lead me to wonder about the diagnosis, but of course the diagnosis could be completely accurate and he still could respond to stimulants as you've described. I'm not a child psychiatrist, so I'm a little out of me element here.

Page Access: