“There is no dose–response relationship in psychopharmacotherapy” vs “pharmacotherapy in psychiatry is based on ligand–receptor interaction”: a unifying hypothesis and the need for plasma concentration based clinical trials

Abstract

The main argument against therapeutic drug monitoring (TDM) in psychiatry, which aims at helping the prescribing physician to optimize her/his pharmacotherapy (Baumann et al. 2004), is that “there is no dose–response relationship in psychiatry”. This claim of a lack of any dose–response relationship is voiced, e.g., in a recent meta-analysis (Adli et al. 2005) which suggests that increasing the dose of an antidepressant after an insufficient response to a mid-range dose only led to a clinical improvement in the case of tricyclic antidepressants but not selective serotonin reuptake inhibitors (SSRIs). Consequently, these authors recommend that SSRI dosage may be reduced without loss of efficacy. Internally inconsistent, they also suggest using TDM to monitor the dose—which would be completely unnecessary if no dose response relationship existed. In the same vein, Rasmussen and Brosen (2000) summarize their review of the literature as follows: “In several studies there has not been found a clear relationship between clinical efficacy and plasma concentration, nor any threshold that defines toxic concentrations. The available data do not suggest that any benefit be obtained from routine monitoring of SSRI plasma levels.” In contrast, TDM is based on the implicit assumption that all pharmacotherapies are based on ligand–receptor interaction, i.e., show a range of concentrations in which the therapeutic effect is a linear function of the concentration of the ligand (i.e., the medication). As we will show in the following, both positions are correct and only very little additional information is necessary to reconcile both positions in a unifying hypothesis that, if implemented in future clinical trials, may lead to a considerable saving both in costs and patient risk. An impressive amount of data from randomized controlled clinical trials indeed shows that clinical effect is most often not linearly correlated with dose. This lack of a dose–response relationship can be plausibly explained if one considers that for all psychopharmacologic medications so far investigated, the plasma level—after the extracellular space in the brain the second-best pharmacokinetic compartment to correlate concentration with effect—varies up to 20-fold among subjects/patients at each administered daily dose. Figure 1 gives the respective data (Tanum et al. 2010) for the selective serotonin reuptake inhibitor (SSRI) citalopram, arguably one of the most prescribed antidepressants worldwide; Fig. 2 the data for the partial D2 receptor agonist aripiprazol (Grunder et al. 2008); and Fig. 3 our own (C.H.) data on the SSRI paroxetine. To emphasize again, this is not an effect restricted to the compounds shown in this Commentary. The up to 20-fold interindividual variation of plasma levels at each single dose is a phenomenon that has been observed for all psychopharmacological medications so far. With a 20-fold interindividual variation in plasma levels after a single dose, it is not surprising that data based on dose and not on plasma level yield dose–response relationships that V. Eggart :G. Zernig (*) Experimental Psychiatry Unit, Center of Psychiatry and Psychotherapy, Department of General Psychiatry and Social Psychiatry, Medical University Innsbruck, Innrain 66 a, room 35-G1-010, A-6020, Innsbruck, Austria e-mail: gerald.zernig@me.com