If SSRIs have previously been administered, without much evident success, the task at hand is to carefully examine the patient's course during the SSRI trials. We have observed several scenarios in which a patient's anxiety disorder has not responded to SSRI therapy.
In the first situation the SSRI prescribed for an anxiety disorder induces mania, hypomania, rapid cycling, or mixed states. In most instances, a history of a bipolar mood disorder, if not noted previously, can be elicited on more thorough inquiry. According to DSM, this form of bipolarity does not definitively indicate bipolar disorder, but, in our experience, predicts a positive response to treatments used for bipolar depression, including atypical antipsychotics and anticonvulsants. Young patients with or at high risk for bipolar disorder may be particularly vulnerable to SSRI antidepressant-induced mania and thus should be closely monitored if given SSRIs. In their preliminary study, Baumer and colleagues15 did not find that serotonin transporter polymorphism significantly influenced vulnerability to antidepressant-induced mania.
In another scenario, the patient is intolerant of the SSRI and a "paradoxic" reaction develops that may include the "jitteriness syndrome." The paradoxic reaction consists of marked sensitivity to the SSRI with development of increased anxiety, jitteriness, or a feeling of "jumping out of one's skin," worsening depression, and irritability. The paradoxic reaction may be responsible for the emergence of suicidal ideation in the context of SSRI administration.5,16
The occurrence of a paradoxic reaction to an SSRI does not bode well for future SSRI treatments, and we include it as a possible side effect when communicating information before we begin treatment with an SSRI. This form of treatment resistance is best managed by starting with an atypical antipsychotic or an anticonvulsant, such as gabapentin(Drug information on gabapentin) or topiramate (although neither is particularly effective in bipolar disorder), or lamotrigine(Drug information on lamotrigine). An SSRI can be added later with reduced likelihood of a paradoxic reaction.
In a third variation, the patient experiences tachyphylaxis or loss of response to the SSRI. Studies by Quitkin and colleagues17 indicated that placebo response was often accompanied by an acute, remitting-relapsing pattern, whereas true drug response was associated with slow and persistent response. Response to SSRIs that mimics the placebo response described by Quitkin and colleagues may follow a similar course of treatment. As occurs with placebo, many patients who experience tachyphylaxis are found to have experienced an initial period of excessively positive mood within days following the initiation of an SSRI. Although bupropion18 has been added to SSRIs to regain treatment response, it is our contention that tachyphylaxis indicates a fundamental inability of serotonin mechanisms to maintain treatment response and requires intervention with agents that work on glutamate (lamotrigine), g-aminobutyric acid (GABA; eg, valproate(Drug information on valproate)), serotonin (5-HT)1a receptors and 5-HT2 receptors (ziprasidone and aripiprazole(Drug information on aripiprazole)), histamine (quetiapine, olanzapine(Drug information on olanzapine), clozapine(Drug information on clozapine)), and dopamine(Drug information on dopamine) D2 receptors (atypical antipsychotics as a group).
In yet another situation, the patient experiences very little, if any, response to the SSRI, indicating that attempting to boost synaptic serotonin has failed as a strategy. This may be explained by a biologic model of SSRI treatment resistance (Figure 1) based on the work of Commons and colleagues.19 In this model, stress activates neurons in the prefrontal cortex, leading to an increased release of the excitatory amino acid, glutamate, with potential for cytotoxicity.
The hyperglutamatergic state, arising from activation of prefrontal cortical efferent glutamatergic-releasing neurons, has at least 2 effects: first, to stimulate local increase of serotonin in the dorsal raphe. This increase of serotonin activates local 5-HT1a autoreceptors, with the consequence of shutting down raphe neuronal firing, leading to a serotonin deficit state with reduced inhibition of fear-related structures, such as the amygdala.20 Second, glutamatergic stimulation of the amygdala leads to an increased release of the stress neuropeptide, corticotropin releasing factor (CRF), which induces anxiety states when administered to experimental animals. The combined effects of these regional neurochemical changes may contribute to the clinical features observed in human anxiety.
In this model, when an SSRI is administered, since raphe neurons are shut down by glutamate, insufficient synaptic serotonin is available for the therapeutic effect achieved by serotonin reuptake inhibition. In contrast, behavioral effects of stress models in animals are reversible by tianeptine(Drug information on tianeptine),21 an antidepressant, not available in the United States, that enhances rather than blocks serotonin reuptake.
Treatment resistance and neurotrophic compromiseComorbid conditions other than mood disorders may complicate the potential treatment response in patients with symptoms of anxiety. These include comorbid alcohol(Drug information on alcohol) and drug use disorders, somatic conditions such as altered thyroid function, and psychotic disorders such as schizophrenia. At the core of treatment resistance in anxiety disorders may be a process called neurogenesis,22-25 or formation of new neurons in the hippocampus and other areas of the brain. It has been observed that a host of factors known to exacerbate anxiety disorders and promote treatment resistance also suppress neurogenesis. The precise role of these neurons continues to be debated, with one limitation being that their proliferation is confined to a small area within the hippocampus, the dentate gyrus. However, a number of neurocircuits, inclusive of the prefrontal cortex, anterior cingulate cortex, amygdala, hypothalamus, locus caeruleus, nucleus accumbens, and hippocampus itself, intersect at the dentate gyrus (Figure 2).
Perhaps what is more salient is that the process of neurogenesis is accompanied by maturation of germinal precursor cells that differentiate into new blood vessels (angiogenesis), new glial cells (gliogenesis) that are a critical support for optimal neuronal function, and new oligodendrocytes (oligodendrogenesis), cells that preserve myelin integrity in the white matter of the brain, facilitating communication between neurons.26-28 Neuroscience is now providing a framework for the clinician, with the possibility that protecting neurotrophism in the brain will become a treatment goal. Some factors that impede neurogenesis and may lead to treatment resistance include hypothyroidism, insulin resistance, sleep deprivation, and inflammation. Factors that stimulate neurogenesis include antidepressants, 5-HT1a agonism, estrogen, valproate, lithium(Drug information on lithium), typical (but not atypical) antipsychotics, and GABA, as well as learning and environmental enrichment.24
Treatment considerationsWe recommend the following targeted approach to treatment-resistant anxiety disorders, assuming that the patient is currently taking an SSRI with or without benzodiazepines. First, eliminate any reversible causes, such as stress, substance abuse, sleep deprivation, or hypothyroidism, that may lead to treatment resistance by mediating excitotoxic influences on the brain. Next, block not only serotonin but also norepinephrine(Drug information on norepinephrine). We have reported positive effects by combining SSRIs3 and tricyclic antidepressants29 when only partial response was observed with the SSRI alone. The next step is to block glutamate. At present, we have available 2 medications with clear antiglutamatergic properties, lamotrigine and riluzole(Drug information on riluzole).30 Other anticonvulsants,31 such as valproate, gabapentin, pregabalin(Drug information on pregabalin), and topiramate(Drug information on topiramate), may be considered at this point but are generally ineffective for comorbid depression. The next strategy is to target the dopamine systems using atypical antipsychotics,32 and to simultaneously block 5-HT2 receptors.
The atypical agents have a host of other effects that may contribute to their anxiolytic properties, including partial 5-HT1a agonism, antihistaminergic effects, anticholinergic effects, and serotonin and norepinephrine reuptake inhibition. Should the use of these agents not be effective, backups include lith-ium, monoamine oxidase inhibitors, buspirone(Drug information on buspirone), and electroconvulsive therapy. Other modalities that are being explored include transcranial magnetic stimulation, vagal nerve stimulation, and deep brain stimulation.33
ConclusionIt may be that the psychopharmacology of the treatment-resistant anxiety disorder with high comorbidity may occur, figuratively, in layers, with certain symptoms responding earlier than others, necessitating a step-wise strategy to target successive receptor systems. In Figure 3, we schematically integrate a number of central concepts that aim to enlighten the clinician's understanding of the anxious patient's condition, and how ineffective and effective treatment may impact the brain.
With respect to treatment-resistant anxiety disorders, the psychiatrist's role is 3-fold: first, to diagnose bipolar comorbidity; second, to protect neurons from excitotoxicity; and third, to promote neurotrophism.
Dr Coplan is professor of psychiatry and director of the division of neuropsychopharmacology at SUNY Downstate Medical Center in Brooklyn, NY, and director of the nonhuman primate facility. He reports that he is on the Speaker's Bureau for GlaxoSmithKline, BMS, Pfizer Inc, and AstraZeneca; and he has received an investigator initiated grant from Pfizer Inc and GlaxoSmithKline.Dr Reddy is attending psychiatrist in the division of neuropsychopharmacology at SUNY Downstate Medical Center in Brooklyn, NY. She reports that she has no conflicts of interest concerning the subject matter of this article.