At the doctor’s office, after the first contact:

Doctor: “I understand from our consultation that you are suffering since some months from a major depressive episode and that you may benefit from a specific antidepressant treatment. I have to add that however the treatment will take a few weeks to exert its effect and there is a chance of about 50 % that it will not benefit you because of either lack of efficacy or poor tolerability.”

Patient: “I see Doctor, do you have a suggestion for a prescription which could fit my situation?”

Doctor: “Yes, indeed, based on your disease profile and the information you provided me, I have in mind a few options, but now we have a further help coming from the knowledge of your genetic status. In fact we now can know in advance which of the possible options are better indicated for your case.”

Patient: “What should I do then?”

Doctor: “If you agree, I will collect with this buccal swab a specimen and send to the lab. We can meet again next week and I will prescribe you a medication”.

One week later.

Patient: “Hello Doctor, I am here again as we agreed.”

Doctor: “Welcome, indeed I received the output from the lab. Your genetic data indicate as best options some compounds, and, paired with the clinical information I already collected, it seems that this is the most appropriate treatment for you. In fact it could lead to some unwanted side effects such as… and there is no guaranty of efficacy, but I would rate the expected global benefit above the other treatment options we have. See you next week and tell me how it is going.”

There is no better way to clarify which are the clinical implications of pharmacogenetics than a narrative. Indeed what is reported above is not what could happen in the future, but what is already happening in many locations around the world. At present thousands of subjects have been treated according to pharmacogenetic indications, and, from what is reported in published papers, they received a relevant benefit in terms of efficacy and tolerability. The optimal drug is chosen at the beginning of the treatment, thus reducing the trial-and-error procedure routinely in use and reducing the time to improvement, with evident benefits for the individual suffering and the societal costs.

The aim of this final chapter is to discuss aspects related to what is described above, the clinical application of pharmacogenetics.

In the previous chapters, you were updated on the most recent scientific information in the field; as you may have noticed, there are good and bad news. Good news are the many DNA variants which have been found to be influencing efficacy and tolerability of many psychotropic drugs, at a variable degree of confidence. The bad news is exactly the same.

There are in fact too many variants affecting these traits; it is not therefore a straightforward prediction such as in other fields of medicine: “you have this variant and this drug is good/bad for you.” Dozens of variants have shown to have a variable degree of influence on individual drug reactions but not a single one with a large and universal effect. Each one of the variants received at least one sample where the effect was not detectable, but on an aggregate level, each of them seems to have a little influence. Little influence means that a single variant does not dramatically change the individual reaction to treatment; thus they should be combined in a proper way in order to obtain a clinically useful tool, and the degree of confidence is also variable, with none achieving a high and unequivocal impact. Further genetic source of variants has not been much considered in literature. Apart from SNPs, there are many other sources of variation; without much detail, we should consider methylation, copy number variation, and de novo mutations, not to mention all possible modulating factors in the transcription process. Appropriate study design is also needed to progress our knowledge [2].

Therefore at present we are not sure which and how many of the variants listed in the previous chapters should be included in the practical screening tool; also we are not sure as well how they should be combined for a clear clinical indication. The most simple hypothesis is to make a summary, i.e., if a subject has 5HTTLPR long variant has an increased possibility to benefit from SSRIs, if the same subject has also the GNB3 rs5443 (C825T) T variant, the possibility to benefit from SSRIs is increased even more. But we are not sure yet that this is the correct strategy; possibly there is a ceiling effect and the combination of variants should not be linear but following a more specific model. At present this problem is faced by offering a warning or suggestion on groups of drugs based on the global genetic profile instead of offering indication for a specific compound..

In fact, there are only few cases in which the indication is clear and straightforward; first of all is the case of carbamazepine, unfortunately a drug not much used in psychiatry. In this case the FDA stated about 8 years ago in the drug label that: “Patients with ancestry in genetically at-risk populations [Asians] should be screened for the presence of hla-b*1502 prior to initiating treatment with tegretol. Patients testing positive for the allele should not be treated with tegretol unless the benefit clearly outweighs the risk” (http://​www.​accessdata.​fda.​gov/​). This is a clear example how pharmacogenetics may help to reduce the risk of severe unwanted effects in the clinical practice. However apart from this very special case, the genetic indication for carbamazepine does not offer further help for efficacy or in other populations.