- Email: [email protected]
Methodological issues in studies on rapid onset of depression
S. 17 Pindolol augmentation
-1
Methodological of depression
-from
issues in studies on rapid onset
M.T. Isaac. Uniuersit?, Hospital Lewisham,
London, UK
Any serious attempt to ascertain the upon set to of action of antidepressants are immediately confronted with munerous methodological issues. Among the forefront of these issues is the importance of controlling for the effect of placebo. It has been recognised for many years that placebo response in trials of antidepressants may run at levels of 30 or even 40%. One way around the problem may be to incorporate a run in phase where only placebo is given to the subject. Someone who showed, say, more than 20% improvement in the primary outcome variable would either be eliminated from the study or identified as a placebo responder. As this run in period could be made before randomisation, the intention-to-treat analysis need not be affected. However, there are counter-arguments. One may argue that placebo run in constitutes an unethical deception of the trial subject. Also, one may be denying treatment to a patient simply for perceived methodological correctness. Moreover, it is not necessarily warranted to assume that all placebo response occurs during the first week of treatment. It may take place later, or for longer (although not much longer: placebo response is seldom, if ever, prolonged). My own view is placebo run in periods in depression are seldom absolutely necessary, and that tight control of the study variable is usually sufficient, in a study population of adequate size, to allow the separation of placebo response. It may be that this is simpler to do where one is testing the potential acceleration effect of an augmenting agent on a standard antidepressant. A good example is the use of pindolol (or placebo) as an augmentor of antidepressants. Here, the presence or absence of the augmentor is the only experimental variable, and every subject is treated with a standard antidepressant. Also, we are looking at rate of change of response, rather than absolute response, in the population, and it may that this helps to circumvent problems with placebo response. Of course, it is a sine qua non that observations (study visits) are frequent enough to pick up early changes in the outcome variables. Our own practice in this setting has been to have visits at days 0 (baseline), 4,7, 10 (optional), 14,21,28 and 42 in a typical six week study. To illustrate such methods in more detail, 1 describe preliminary results of a continuing study of the augmentation of the antidepressant milnacipran with pindolol. Decisions on dosage of augmentor are clearly fimdamental, and 1 shall discuss these briefly, too. Perhaps one of the most important methodological innovations, only recently coming into more regular use, goes beyond the standard sixweek trial period to examine longer-term outcome. Following patients from our earlier pindololiparoxetine study for one year (Tome & Isaac 1998, International Clinical Psychopharmacology, in press), we developed a method for evaluating the value of increased cost of pharmacological augmentation that, taken for 6 weeks, accelerated the action of an antidepressant. This tested the hypothesis that, if onset of action is taken into account, any added direct costs of the augmenting agent are offset by longer term cost effectiveness. Patients received the SSRI antidepressant paroxetine and pindolol (or placebo). Atter 6 weeks, patients were offered SSRI alone on an open label basis for up to 6 months. The results of this study have been reported elsewhere. At that point they were discharged to their General Practitioner (GP) or local psychiatric services and subsequently assessed by us at one year. We used techniques of decision analysis, cost-effectiveness and cost-benefit and included a sensitivity analysis. The direct costs over one year of SSRI and augmenting agent, if the acceleration effect is taken into account, reflected greater cost-effectiveness than antidepressant alone. The costeffectiveness analysis was still positive in both cases. We concluded that the direct costs of treatment are higher than those previously calculated for antidepressants; but that the rate of onset must be taken into account.
References [I]
Tome,
Isaac
et al.,
1997,
Int. Clin.
Psychophannacol.
12: Xl-89
the bench to the clinical arena
[s.17.03)
Pindolol augmentation: in the brain
s91
Single or multiple targets
F. Artigas. Department of Neurochemistry, Institute de lnuestigaciones BiomPdicas de Barcelona, CSIC, 08034 Barcelonu, Spain In recent years, the use of the intracerebral microdialysis technique has revealed some unpredicted actions of antidepressant drugs. Thus, single treatment clinically-relevant doses of SSRls, non-selective 5-HT uptake inhibitors or MAOls enhanced the extracellular 5-HT concentration in the raphe nuclei of the midbrain (that contain serotonergic cell bodies) but not in forebrain areas rich in nerve terminals. The limited effect of single doses of these agents in brain regions rich in serotoninergic terminals is due to the fact that they behave as indirect agonists of the somatodendritic ~-HTIA autoreceptors in the midbrain raphe nuclei. The activation of such receptors reduces serotonergic cell firing and terminal 5-HT release, thus offsetting the S-HT increments produced in forebrain by reuptake or MAO inhibition. The combined administration of uptake or MAO inhibitors and 5HTIA receptor antagonists enhances the action of the former, as assessed by neurochemical techniques. Recent data also supports the enhancement of the behavioural actions of SSRIs by the selective ~-HT~A antagonist WAY 100635. At neurochemical level, WAY 100635 potentiate the SSRIinduced increments of extracellular 5-HT preferentially in frontal cortex or dorsal striatum, two forebrain areas innervated mainly by axons of serotonergic neurones of the dorsal raphe nucleus (DRN). In contrast, the potentiation elicited in dorsal hippocampus - innervated mainly by the median raphe nucleus - is much lower. This regional selectivity is consistent with the greater inhibition of 5-HT release produced by SSRls and selective ~-HT~A agonists in DRN-innervated areas. Pindolol potentiates the effects of SSRls and SNRls in striatum and hypothalamus but not in hippocampus. We observed a paradoxical reduction of the effect of paroxetine by the concurrent administration of the mixed @-adrenoceptor/S-HTIA antagonists (-)alprenolol, (-)pindolol and (-)tertatolol. Since (-)pindolol enhances the action of SSRls in striatum and (-)tertatolol elevates extracellular 5-HT by itself in this area, it is likely that the lower effects of paroxetine in frontal cortex when associated with (-)pindolol or (-)tertatolol derive from the involvement of cortical &adrenoceptors. An inhibitory action of these agents at the level of cell body seems unlikely since the same neuronal groups at the DRN innervate the striatum and frontal cortex. Several uncontrolled and controlled clinical trials have reported an early onset of the antidepressant action of SSRls and other 5-HT-acting drugs when used in combination with pindolol (2.5 mg tid). In some of them, pindolol addition also enhanced the efficacy of the treatment. At the time of writing, 11out of 14 reports have shown a beneficial action of pindolol. Our data indicate that pindolol addition is effective to accelerate and enhance the effects of fluoxetine in untreated patients (N = 1 IO) and in some treatment-resistant patients. However, in a recently finished study, pindolol was not more effective than placebo in SRI-resistant patients on the short term (N = 80). The reasons for the discrepancres between studies are still unclear and may possibly be related to the patients’ characteristics and different design of the trials (e.g., exclusion of comorbid or chronic patients, number of previous episodes, degree of chronic&y, etc). A distinct neurobiological alteration may also be considered as a potential source of differences (i.e., some patients may not respond to manipulations of the serotonergic system and may require interventions through other transmitter systems). There has been some concern as to whether the dose of pindolol used in clinical studies is sufficient to interact with 5-HTl.4 receptors. We have employed receptor autoradiography to examine the affinity of (-)pindolol for S-HTIA receptors in the brain of four different species, rat, guinea pig, monkey and human. (-)Pindolol displaced the binding of the selective agonist [3H]R-OH-DPAT and antagonist [3H]WAY 100635 to pre- and postsynaptic receptors with a fairly good affinity. The K, values found for pre- and postsynaptic receptors are comparable in all four species. In human brain, the Ki for the displacement of the [“HISOH-DPAT binding to hippocampus (CA]. CA3 and dentate gyrus) and DRN was between 6.5 and 13.5 nM. These affinities are below the plasma
-1
Methodological of depression
-from
issues in studies on rapid onset
M.T. Isaac. Uniuersit?, Hospital Lewisham,
London, UK
Any serious attempt to ascertain the upon set to of action of antidepressants are immediately confronted with munerous methodological issues. Among the forefront of these issues is the importance of controlling for the effect of placebo. It has been recognised for many years that placebo response in trials of antidepressants may run at levels of 30 or even 40%. One way around the problem may be to incorporate a run in phase where only placebo is given to the subject. Someone who showed, say, more than 20% improvement in the primary outcome variable would either be eliminated from the study or identified as a placebo responder. As this run in period could be made before randomisation, the intention-to-treat analysis need not be affected. However, there are counter-arguments. One may argue that placebo run in constitutes an unethical deception of the trial subject. Also, one may be denying treatment to a patient simply for perceived methodological correctness. Moreover, it is not necessarily warranted to assume that all placebo response occurs during the first week of treatment. It may take place later, or for longer (although not much longer: placebo response is seldom, if ever, prolonged). My own view is placebo run in periods in depression are seldom absolutely necessary, and that tight control of the study variable is usually sufficient, in a study population of adequate size, to allow the separation of placebo response. It may be that this is simpler to do where one is testing the potential acceleration effect of an augmenting agent on a standard antidepressant. A good example is the use of pindolol (or placebo) as an augmentor of antidepressants. Here, the presence or absence of the augmentor is the only experimental variable, and every subject is treated with a standard antidepressant. Also, we are looking at rate of change of response, rather than absolute response, in the population, and it may that this helps to circumvent problems with placebo response. Of course, it is a sine qua non that observations (study visits) are frequent enough to pick up early changes in the outcome variables. Our own practice in this setting has been to have visits at days 0 (baseline), 4,7, 10 (optional), 14,21,28 and 42 in a typical six week study. To illustrate such methods in more detail, 1 describe preliminary results of a continuing study of the augmentation of the antidepressant milnacipran with pindolol. Decisions on dosage of augmentor are clearly fimdamental, and 1 shall discuss these briefly, too. Perhaps one of the most important methodological innovations, only recently coming into more regular use, goes beyond the standard sixweek trial period to examine longer-term outcome. Following patients from our earlier pindololiparoxetine study for one year (Tome & Isaac 1998, International Clinical Psychopharmacology, in press), we developed a method for evaluating the value of increased cost of pharmacological augmentation that, taken for 6 weeks, accelerated the action of an antidepressant. This tested the hypothesis that, if onset of action is taken into account, any added direct costs of the augmenting agent are offset by longer term cost effectiveness. Patients received the SSRI antidepressant paroxetine and pindolol (or placebo). Atter 6 weeks, patients were offered SSRI alone on an open label basis for up to 6 months. The results of this study have been reported elsewhere. At that point they were discharged to their General Practitioner (GP) or local psychiatric services and subsequently assessed by us at one year. We used techniques of decision analysis, cost-effectiveness and cost-benefit and included a sensitivity analysis. The direct costs over one year of SSRI and augmenting agent, if the acceleration effect is taken into account, reflected greater cost-effectiveness than antidepressant alone. The costeffectiveness analysis was still positive in both cases. We concluded that the direct costs of treatment are higher than those previously calculated for antidepressants; but that the rate of onset must be taken into account.
References [I]
Tome,
Isaac
et al.,
1997,
Int. Clin.
Psychophannacol.
12: Xl-89
the bench to the clinical arena
[s.17.03)
Pindolol augmentation: in the brain
s91
Single or multiple targets
F. Artigas. Department of Neurochemistry, Institute de lnuestigaciones BiomPdicas de Barcelona, CSIC, 08034 Barcelonu, Spain In recent years, the use of the intracerebral microdialysis technique has revealed some unpredicted actions of antidepressant drugs. Thus, single treatment clinically-relevant doses of SSRls, non-selective 5-HT uptake inhibitors or MAOls enhanced the extracellular 5-HT concentration in the raphe nuclei of the midbrain (that contain serotonergic cell bodies) but not in forebrain areas rich in nerve terminals. The limited effect of single doses of these agents in brain regions rich in serotoninergic terminals is due to the fact that they behave as indirect agonists of the somatodendritic ~-HTIA autoreceptors in the midbrain raphe nuclei. The activation of such receptors reduces serotonergic cell firing and terminal 5-HT release, thus offsetting the S-HT increments produced in forebrain by reuptake or MAO inhibition. The combined administration of uptake or MAO inhibitors and 5HTIA receptor antagonists enhances the action of the former, as assessed by neurochemical techniques. Recent data also supports the enhancement of the behavioural actions of SSRIs by the selective ~-HT~A antagonist WAY 100635. At neurochemical level, WAY 100635 potentiate the SSRIinduced increments of extracellular 5-HT preferentially in frontal cortex or dorsal striatum, two forebrain areas innervated mainly by axons of serotonergic neurones of the dorsal raphe nucleus (DRN). In contrast, the potentiation elicited in dorsal hippocampus - innervated mainly by the median raphe nucleus - is much lower. This regional selectivity is consistent with the greater inhibition of 5-HT release produced by SSRls and selective ~-HT~A agonists in DRN-innervated areas. Pindolol potentiates the effects of SSRls and SNRls in striatum and hypothalamus but not in hippocampus. We observed a paradoxical reduction of the effect of paroxetine by the concurrent administration of the mixed @-adrenoceptor/S-HTIA antagonists (-)alprenolol, (-)pindolol and (-)tertatolol. Since (-)pindolol enhances the action of SSRls in striatum and (-)tertatolol elevates extracellular 5-HT by itself in this area, it is likely that the lower effects of paroxetine in frontal cortex when associated with (-)pindolol or (-)tertatolol derive from the involvement of cortical &adrenoceptors. An inhibitory action of these agents at the level of cell body seems unlikely since the same neuronal groups at the DRN innervate the striatum and frontal cortex. Several uncontrolled and controlled clinical trials have reported an early onset of the antidepressant action of SSRls and other 5-HT-acting drugs when used in combination with pindolol (2.5 mg tid). In some of them, pindolol addition also enhanced the efficacy of the treatment. At the time of writing, 11out of 14 reports have shown a beneficial action of pindolol. Our data indicate that pindolol addition is effective to accelerate and enhance the effects of fluoxetine in untreated patients (N = 1 IO) and in some treatment-resistant patients. However, in a recently finished study, pindolol was not more effective than placebo in SRI-resistant patients on the short term (N = 80). The reasons for the discrepancres between studies are still unclear and may possibly be related to the patients’ characteristics and different design of the trials (e.g., exclusion of comorbid or chronic patients, number of previous episodes, degree of chronic&y, etc). A distinct neurobiological alteration may also be considered as a potential source of differences (i.e., some patients may not respond to manipulations of the serotonergic system and may require interventions through other transmitter systems). There has been some concern as to whether the dose of pindolol used in clinical studies is sufficient to interact with 5-HTl.4 receptors. We have employed receptor autoradiography to examine the affinity of (-)pindolol for S-HTIA receptors in the brain of four different species, rat, guinea pig, monkey and human. (-)Pindolol displaced the binding of the selective agonist [3H]R-OH-DPAT and antagonist [3H]WAY 100635 to pre- and postsynaptic receptors with a fairly good affinity. The K, values found for pre- and postsynaptic receptors are comparable in all four species. In human brain, the Ki for the displacement of the [“HISOH-DPAT binding to hippocampus (CA]. CA3 and dentate gyrus) and DRN was between 6.5 and 13.5 nM. These affinities are below the plasma