It has come to our attention that pharmacogenetic tests for drugs used for the treatment of ADHD and analgesics have been removed by other pharmacogenetic platforms. We believe that these actions are appropriate given the low levels of evidence for many of these gene-drug associations. Our software for precision prescribing, TreatGx, covers drugs that also belong to these categories; however, we would like to assure our users that we have selected only those with strong evidence and functional significance.
In the TreatGx ADHD treatment algorithm, we use pharmacogenetics and other factors to personalize prescribing options. For ADHD, TreatGx gives weight-based and age-based dosing for some drugs, in addition to adjusting drug dose based on renal and hepatic impairment.
TreatGx translates gene-drug interactions into actionable results backed up with the highest levels of evidence for clinical utility. We follow a rigorous review process of peer-reviewed articles, systematic reviews, randomized controlled trials, and professional guidelines such as those published by the Clinical Pharmacogenetics Implementation Consortium (CPIC) and the Dutch Pharmacogenetics Working Group (DPWG). Physicians, pharmacists, epidemiologists, and geneticists are involved in the curation process.
Although there is much interest in the pharmacogenetics of ADHD stimulants, we have not included these drugs in our software due to their low level of evidence. A vast amount of pharmacogenetic studies have been done on methylphenidate and dextroamphetamine, two of the most commonly used ADHD stimulants; however, the evidence is not yet conclusive. The DPWG evaluated therapeutic dose recommendations for methylphenidate based on CYP2D6 genotypes and found that there is not a gene-drug interaction. Amphetamines are metabolized through CYP2D6 but there has been no correlation with clinical response based on CYP2D6 genotype. Other gene-drug associations with stimulants such as dextroamphetamine have minimal evidence supporting them and therefore, it is not possible to give clinical recommendations based on genotype for these drugs.
Non-stimulant drugs such as atomoxetine (CPIC level A, PharmGKB level 1A) and desipramine (CPIC level B, PharmGKB level 1A) show high level of evidence for drug-gene interactions and are included in TreatGx. The recommendations provided in TreatGx are based on the guidelines published by CPIC.
Analgesics comprise a wide range of medications and not all of them have strong pharmacogenetic evidence. We have focused our attention on those medications that do. For instance, it is well-known that CYP2D6 interacts with the metabolism of codeine (CPIC level A, PharmGKB level 1A), and that this translates into a clinical outcome: CYP2D6 poor metabolizers have a higher risk of having insufficient pain relief. Therefore, codeine, and other drugs known to be metabolized by CYP2D6 such as oxycodone (CPIC level A, PharmGKB level 2A) and tramadol (CPIC level A, PharmGKB level 1B), are part of the TreatGx medication decision support system.
Another group of analgesic drugs covered by TreatGx are those metabolized by CYP2C9. Nonsteroidal anti-inflammatory drugs (NSAIDs) such as celecoxib (CPIC level B, PharmGKB level 2A), diclofenac (CPIC level C, PharmGKB level 2A) and flurbiprofen (CPIC level B/C, PharmGKB level 3) have a moderate level of evidence that suggest that CYP2C9 poor metabolizers should be administered these drugs with caution. Response to other NSAIDs, such as naproxen and ibuprofen, have not been shown to be significantly influenced by CYP enzyme variation. In the absence of pharmacogenetic information, other patient-specific factors are used by TreatGx to offer safe and effective medication options with optimized dosing based on age, weight, kidney function, and additional relevant information.
GenXys has always followed the principle of only including pharmacogenetic markers that have high levels of evidence for a gene-drug response association. Using genes with lower levels of evidence to adjust drug therapy risks offering patients inaccurate information and potentially may cause the patient to have less-than-optimal medication therapy.
Pharmacogenetics is a valuable tool to help personalize and optimize drug therapy, but we risk eroding the faith in this approach if we suggest implementing a change based on a gene-drug association that is later found to be incorrect. As the number of available pharmacogenetic tests increases, transparency with recommendations is the best approach. Offering levels of evidence and references for each pharmacogenetic-based clinical recommendation allows clinicians to ensure these interventions will be evidence-based and rational.