Well, no way I can summarize in my own words but it's been a while since she sent it to me and I'm sure it's been published by now. I've taken out all identifying information and here is the meat of the info:
"Question 1: An 8 year old foster child is hospitalized for severe, explosive outbursts that have made it difficult for her to maintain placements. Her fuse is so short that there is no time to intervene between whatever her trigger and the subsequent rage. Her attention deficit and oppositional defiant disorder are treated with a stimulant and alpha agonist. Numerous SSRIs, second and now first generation neuroleptics have been directed toward her “irritability.” Would pharmacogenetic testing add to her care? When does it make sense for a psychiatrist to order pharmacogenetic tests?
Answer 1:
The choice of a psychotropic medication is complex and influenced by many things. These include: Age (which effects things like hepatic metabolism), family history and illness characteristics e.g. co-morbidities, severity and duration of illness; Medication compliance (a drug can’t work if you don’t take it). Pharmacologically, drug level, onset of action, cumulative drug exposure and drug interactions influence response. The information one gets from pharmacogenetic testing does not address all of these factors.
Currently available genetic testing provides a significant amount of data but with varying utility. Many of the current genetic testing companies include both established cytochrome p450 enzymes as well as other less well established drug metabolizing enzymes and genetic variants that have yet to be established in terms of medication response, or side effects of treatment response The combination of drug metabolizing enzymes and these other genetic variants in terms of basic biological pathways and mechanisms of action of our psychotropic medications is unknown. The utility of the composite risk genetic profiles is also undefined.
At this time the studies validating utility of genetic testing above and beyond treatment as usual are mostly limited to treatment resistant patients, almost entirely in major depression, and the majority of the samples are adult and primarily Caucasian. No pediatric studies are available and unfortunately a significant number of the studies that have been done were written and conducted by employees of the genetic testing companies.
Some clinicians find pharmacogenetic testing useful after the first medication failure, and some wait until multiple medications have failed or produced serious side effects before thinking about pharmacogenetic testing. Perhaps if you have a treatment resistant patient, like the child described above, or someone with lots of unexpected side effects with a history of other complications with standard medication treatment, or family history of negative responses to medications, drug metabolizing enzymes may be to blame and pharmacogenetic testing may provide benefit. However, the companies test more than just these enzymes and base their guidance on the larger panel with less available evidence. Unfortunately, specific data provided from the companies that do pharmacogenetic does not necessarily address these problems.
Question 2: Do the pharmacogenetic tests help make diagnoses or predict side effects?
Answer 2: At this time any genetic test that states it can predict diagnosis (including ADHD), treatment response or side effects in pediatric populations is unlikely to have sufficient data across any psychiatric diagnosis to be reliable. In addition the vast majority of studies of treatment response have failed to include complexities that we face as child psychiatrists regularly: developmental age, race, family history and context. Even for the “cleanest” psychiatric diagnosis without any evidence of early life trauma there is not enough data to suggest that genetic testing will advance our ability, beyond the use of careful diagnostic clarity, developmental history and family history of both diagnosis and treatment response.
Of the available evidence, it is highest for CYP2D6 and CYP2C19 and tricyclic antidepressants and SSRIs. The Clinical Pharmacogenetics Implementation Consortium (CPIC) provides guidelines for adjusting tricyclic antidepressants and SSRIs based on CYP2D6 and CYP2C19. They provide a strength of recommendation for each genotype and provide references for all recommendations but do not provide guidance on when to test. However, the French National Network of Pharmacogenetics recommends CYP2D6 and CYP2C19 genotyping before initiating an antidepressant treatment, especially in patients with a high risk of toxicity. The FDA recommends testing of HLA alleles prior to use of carbamazepine in genetically at-risk populations, i.e. patients of Asian descent, due to a high risk of serious and sometimes fatal dermatologic reactions. The FDA provides recommendations for maximum doses in poor metabolizers for some psychotropic medications (e.g. citalopram).
Question 3: What do the green, yellow and red markers mean?
Answer 3: The algorithms that are used to bin medications on commercially available tests into bins such as “use with caution” or “use as directed” are proprietary and unable to be validated by external cohorts, and the reports seldom include a strength of recommendation or evidence to back up the binning. Medications in the green “use as directed” bin may not have any genetic associations that influence them, but that doesn’t automatically make them the best treatment for that patient – there may be higher risk of side effects or lower efficacy of those medications compared to some in the yellow or red bins. They also may not be approved for use for the patient’s indication or age. Medications in the yellow or red bins may be safer and more effective than those in the green bin when dosed appropriately based on genotype.
The accuracy of variants tested is regulated, but the interpretations of genetic variants to metabolizer status is not regulated and varies, and the interpretations of variants/metabolizer status to effect on medication varies. Additionally, variants tested for in each panel are different, and since the presence of the “normal” allele, *1, is indicated by the lack of other alleles, the number of alleles tested affects the prediction of the normal allele. One thing to remember when looking at the reports available from pharmacogenetic tests is that just because a drug is a substrate for an enzyme doesn’t mean that genetic variants in that gene influence response/toxicity.
To date it is unclear if greater information is available about exactly how this modeling was done and how these studies account for any racial, sex or developmental differences. Further none of these studies take into account concurrent medication use, dietary factors (Brussel sprouts, grapefruit juice, vitamins, supplements), and other medical conditions that likely directly interact with the end metabolic processes controlling these medications.
In summary:
1. Genetic testing for drug metabolizing enzymes may have specific utility, at this time, for individuals with abnormal side effect profiles and treatment resistant conditions. It is not supported as first line approach in terms of cost effectiveness, particularly in pediatric populations.
2. Careful attention to the specific genotypes that are included in the testing and the variable data in support of their involvement in medication response and the specific diagnostic condition is needed. Inclusion of genetic variants with insufficient or unsupported relevance to drug effectiveness is a current limitation of commercially available genetic testing panels and limits their utility.
3. Data may become useful in the future and we should begin to think about creating data sets that are representative of our clinical populations in terms of race, sex, genographic distribution and life histories that are clearly relevant in terms of prognostic responses, metabolic processes and ideally the strength of our ability to effectively utilize and implement the best quality of personalized medicine.
Pay careful attention to dietary factors that may confound"
