PEER Online, the successor to ‘referenced EEG’ (or ‘rEEG’), has been studied for over 20 years and the evidence supporting the effectiveness of using a physician based outcomes database to guide treatment is significant.

PEER Online is similar to a standard QEEG in that it uses QEEG output variables, but differs from a standard QEEG in that it references the QEEG to a normative and then symptomatic database. By comparing a given patient’s QEEG to a database of QEEGs of subjects who have tried and responded to a specific medication, PEER Online can provide useful information regarding the response of neurophysiologically similar patients to a wide number of medications. PEER Online may thus have the advantage of providing physicians with useful information as to medication outcomes before a medication regime is started. As important as the positive findings (e.g. Sensitive finding) from PEER Online may be, physician users have also reported that negative findings (in which neurophysiologically similar patients reported resistant outcomes for certain medications) can be extremely useful in reducing trial and error pharmacotherapy. It has also been used to help select the medication that best matches the QEEG brainwave pattern, regardless of “symptom clusters,” currently used for diagnostic nomenclature.

The initial publication of results using the information from an EEG-based database to guide treatment was the Suffin and Emory article in which they examined attentional and affective disorders and their association with pharmacotherapeutic outcomes (Suffin and Emory, 1995). This was a retrospective analysis of treatment outcomes in sequential patients with Attention Deficit Disorder and Affective Disorders. One hundred medication-free patients meeting DSM-III-R criteria for attentional and affective disorders underwent pretreatment QEEG, where the data was submitted to a normative database. Following the EEG, attention-disordered patients were first treated with a stimulant, secondarily with an antidepressant, and tertiarily with an anticonvulsant. Affectively disordered patients were treated initially with antidepressant, and secondarily augmented with anticonvulsant or lithium. Tertiary treatment was a stimulant. Patients were assessed up to 6 months. A Clinical Global Improvement score was assigned. Similar Neurometric subgroups were identified within both the attentional and the affectively disordered patients. Without regard to DSM-III-R diagnosis, there were robust correlations between Neurometric subgroup membership, responsivity to selected pharmacologic agent class(es), and clinical outcome resulting in an 87% response to antidepressants. Another subgroup was 100% responsive to stimulants and a third was 85% responsive to anticonvulsants/lithium. A fourth subgroup was 80% responsive to anticonvulsants. Patients with similar Neurometric features responded to the same class(es) of psychopharmacologic agent(s) despite their DSM-III-R classification.

Other smaller, preliminary studies have suggested a potential role in using this information for medication selection for depression, to name just one psychiatric disorder. A prospective, randomized, controlled 25-week study at a Veterans Affairs hospital consisted of two groups of treatment-resistant depressed patients (n=6 control, n=7 experimental) (Suffin, et al. 2007). The trial used these quantitative EEG features to guide prescribing of psychotropic medications, while the control group received treatment as usual. The results indicated that six of seven subjects augmented with QEEG data received ratings of moderate to marked improvement on both the Ham-D and Beck’s Depression Inventory. Only a single subject in the control group did this well. When unblinded, that same subject had been treated successfully with the medication that was consistent with the EEG patterns in the report. Pilot studies using these QEEG variables in eating disorders (Greenblatt, et al. 2011) and substance abuse (Schiller, 2008; Shaffer, et.al. 2005,) demonstrated similar promising results.

Another pilot study (DeBattista, et al. 2008) was conducted to compare this same methodology with the Texas Medication Algorithm Project (TMAP) algorithm for patients with treatment-resistant depression. This 10-week study (n=18) found the data derived from the QEEG variables resulted in statistically greater change from baseline scores for the Quick Inventory of Depressive Symptomatology-Self Report-16 (QIDS-SR16) and the Quality of Life Enjoyment and Satisfaction Questionnaire-Short Form (Q-LES-Q-SF) than TMAP-guided therapy. It also found that five subjects in the TMAP group received a successful TMAP therapy that was identical to what would have been prescribed with the information obtained for the variables from the referenced-EEG database (DeBattista, et al. 2008).

Based on these earlier studies which suggested that using the information from shared outcomes based on QEEG patterns could assist clinicians in being cautious yet efficacious in their choice of medications, especially for treatment-refractory patients, a larger trial was run. In a multicenter, randomized trial, DeBattista found the referenced-EEG database treatment group (experimental) was compared with an optimized treatment based on the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study guidelines (control) initially funded by the National Institutes of Health (NIH) in patients with treatment-refractory major depressive disorder (DeBattista, et al. 2011). The experimental group’s selection led to statistically better outcomes compared with the control group. The improvement in the group guided by the QEEG data over the control group was significant as early as two weeks and the improvement continued throughout the 12-week study and demonstrated significantly greater improvement on both primary endpoints (QIDS-SR16 and Q-LES-Q-SF, i.e. the same as in the STAR*D study) of P <0.0002 compared with control, as well as statistical superiority in nine of 12 secondary endpoints. Because there are innumerable combinations of therapies available from the databases’ report, the DeBattista study was not intended to evaluate any one medication against another, but to examine it as a tool to improve outcomes in this difficult population.

In aretrospective chart review in the treatment of depression in Eating Disorders, Greenblatt reported on 22 patients with a 2-year previous history to using the referenced-EEG database to guide treatment and followed them for 2-5 years. Patients demonstrated significant decrease in depressive symptoms (Ham-D), severity of illness (CGI-S), and overall clinical global improvement (CGI-I). The group also had substantially fewer inpatients, residential and partial hospitalization days within the 2-year follow up period compared to the two years prior to the use of the report. (Greenblatt, et al. 2011)

Without objective physiologically based clinical evidence to support clinical decisions, treating physicians can only respond to their patients in the best way possible, trial and error of approved and un-approved treatments.

Psychiatry is plagued by the lack of a clinically useful biomarker assessment system to guide treatment. Instead, pharmacotherapy is typified by an inductive trial and error process resulting in significant additional morbidity, mortality, and costs due to failed medication trials. Despite extensive research resulting in more than 130 FDA-approved medications used in treating psychiatric disorders, evidence-based guidance is largely lacking for selecting amongst these options to treat refractory patients.

Lack of effectiveness

In 2007, the Agency for Healthcare Research and Quality (AHRQ), an agency under the Department of Health and Human Services (DHHS), released a report on a systematic review they conducted on “Comparative Effectiveness of Second-Generation Antidepressants in the Pharmacologic Treatment of Adult Depression”. In this extensive review of 293 published articles, AHRQ noted almost 40% of patients did not respond to these treatment medications and that over 50% of the patients failed to achieve remission (AHRQ; Effective Health Care #7, Executive Summary, 2007). Although not an impressive response rate, the authors were encouraged by a new study that that was not included in their review due to its late publication; the STAR*D study.

STAR*D (Rush, 2006) is the largest study to date evaluating pharmacotherapy for depression and was designed to provide guidance in selecting the best ‘next-step’ treatment for the many patients who failed to get adequate relief from their initial SSRI treatment. By enrolling over 4,000 patients first treated with escitalopram, STAR*D ensured a sufficient number of treatment failures for its step 2, 3, and 4 comparisons. Discouragingly, there were no significant differences found in the five next-step comparisons of eleven pharmacologically distinct treatments even though the N per each specific treatment ranged from 51 to 286 patients and was therefore more than sufficient to identify differences if any existed. STAR*D does document some of the negative consequences from failed medication trials (see table below). These include progressive step-by-step:

— Decreasing remission and response rates

— Increasing relapse rates; and

— Increasing dropout rates despite STAR*D’s exemplary free, acute and continuing care

In recent re-analyses of the STAR*D data, several studies have question the effectiveness of the treatments (Pigott et al., 2010; Pigott, 2011) and certainly has brought into question the tolerability of the treatments given the relapse and dropout rates noted in the table above. Pigott notes that ‘In stark contrast to STAR*D’s report of positive findings supporting antidepressants’ effectiveness, only 108 of its 4,041 patients (2.7%) had an acute-care remission, and during the 12 months of continuing care, these patients neither relapsed nor dropped out.’

In response to the STAR*D reports showing limited efficacy for a variety of antidepressant strategies, the NIMH funded the recently published CO-MED study (Rush et al., 2011) testing whether starting patients with several antidepressants at the same time would be associated with increased efficacy. Six hundred and sixty-five (665) patients with MDD were randomized to a 12-week acute treatment and participants who experienced substantial benefit in the acute phase were enrolled in an additional 16-week continuation treatment. There was no significant difference between the response or remission rates observed in the three arms in either the acute phase or the continuation phase (12-28 weeks).

Clearly, even in the most optimistic evaluation of available treatments for patients with any treatment resistant disorder, an effective rate of 40% indicates that most medications are simply not sufficiently effective to alleviate the patient’s symptoms and most certainly are not providing a remission of disease. This is particularly clear when it is reported that in antidepressant trials there is a placebo response rate of between 35% – 45% (Fava, et al., 2003). Therefore, it can be convincingly argued that these treatments for antidepressants are no more effective than no medication at all.

Recent research raises doubts about the degree to which psychopharmacological treatment has kept pace with our advances in understanding the brain and psychiatric disorders. There is a plethora of new psychiatric medications, but there is growing recognition that these new generation medications may not hold significant advantages over older medications despite their higher costs. It has been observed (DePaulo, 2006) that, when viewed together, the three major studies (STAR*D, CATIE (Stroup, 2003), and STEP-BD (Goldberg, 2009)) question whether or not modern pharmacological provide increased benefits for the additional costs. While none of these studies (STAR*D, CATIE, and STEP-BD) have focused primarily on comparing older and newer treatments, such contrasts do not suggest any dramatic advantages for the newer medications.

Low tolerability

When treating patients with mental disorders, particularly patients who are treatment resistant, it’s common for patients to stop taking their medications for various reasons, whether it is a lack of effectiveness or side effects of the medication. This lack of adherence to the treatment can collectively be considered tolerability to the treatment, and the evidence for low tolerability of these treatment medications is significant.

With the STAR*D study, over 60% of the patients dropped out of the study prematurely, even though they were provided the free care throughout the study, and were managed at a level far higher than could be expected in the standard care of private practice. Whereas it is difficult to know if they dropped from the study due to the ineffectiveness of the treatment or due to side effects of the medication, there is clear evidence that the side effect profile of these treatments is significant. In the AHRQ systematic review, it was noted that 61% of the patients in the efficacy trials they reviewed had at least 1 adverse event. Further, the same report notes that between 16% and 6% of severe adverse events was reported for the second-generation antidepressants.

In the CO-MED study, although there were no significant differences in efficacy, there was a significant difference in the side effect burden for the patients. In the two polypharmacy arms the maximum side effect burden was between 10% and 15% in both the acute and continuation phases, however in the single treatment arm the side effect burden was only 4% in the acute phase and 5% in the continuation phase. This large, well designed, adequately powered study suggests that for a significant number of patients with MDD, adding medication only adds to their side effects without an increase in efficacy, thus increasing to the low rate of tolerability.

Summary

In 2009, Dr. Thomas Insel made similar observations to those of DePaulo, noting that second-generation medications have consistently demonstrated no significant advantage compared with first-generation medication in multiple comparative effectiveness studies funded by the NIMH (Insel, 2009). He also felt that current medication regimes help too few people improve from the perspective of side effects, and a recent prospective study has now even called into question the widely held belief that second-generation antipsychotics produce a lower incidence of tardive dyskinesia (Woods, 2010).

It is increasingly clear that the reason there are so many treatment resistant patients (patients who have failed on two or more medication trials) is due to the lack of effectiveness of the medications and/or the high side effect burden of the treatments. The evidence points to a poor understanding of the underlying physiological causes of the symptoms of these disorders and therefore the ineffective diagnosis as the root cause of the ineffectiveness of the current medication treatments. Without objective physiologically based clinical evidence to support clinical decisions, treating physicians can only respond to their patients in the best way possible, trial and error of approved and un-approved treatments.

EEG recordings with quantitative analysis yield more information than can be appreciated by simple visual inspection.

Since the 1970’s EEG recordings have been digitized and subjected to quantitative analysis (QEEG), which yields more information than can be appreciated by simple visual inspection. QEEG extends EEG technology beyond qualitative identification of abnormality and allows for comparison of an individual patient’s EEG pattern with large public databases of age matched, asymptomatic (control) group EEG values (Duffy et al. 1979; John et al., 1977).  Medication-induced changes in EEG and QEEG data have been reported for a broad range of antidepressants, benzodiazepines, stimulants, antipsychotics, lithium salts, and anticonvulsants (Herrmann et al., 1979, Itil et al., 1973, 1979; Saletu et al., 1987; Small et al., 1989; Struve, 1987). These drug changes are specific in regard to effects on distinct components of the EEG pattern and are dose-dependent, reversible upon medication withdrawal, and measurable across psychiatric syndromes and in asymptomatic volunteers.

In those studies obtaining baseline, medication-free EEGs, investigators demonstrated unique QEEG features that could be used to aid in treatment guidance. For example, patients with major depressive illness with excess alpha wave magnitudes were retrospectively reported responsive to antidepressants that reduce alpha magnitude (Ohashi, 1994). Similarly, patients with obsessive-compulsive disorder with excess alpha wave activity responded best to antidepressants (Prichep et al., 1993).

In contrast, a subgroup of patients with obsessive-compulsive disorder with elevated theta wave activity did not respond favorably to antidepressants that are known to increase theta wave magnitudes and would therefore be predicted to exacerbate the symptoms of these patients (Prichep et al., 1993). Finally, patients with attention deficit hyperactivity disorder showed a general decrease in EEG patterns and were predicted to respond favorably to methylphenidate, an agent that accelerates EEG frequencies (Satterfield et al., 1973).Subsequent work in depression led to a large multicenter trial (n = 375) (Leuchter et al., 2009) that examined “cordance” – a frontal QEEG index’s ability to predict response to the SSRI, escitalopram, after one week of treatment. Subjects were subsequently randomized to remain on treatment or crossed over to alternative treatment. Seventy-five (75) subjects remained on escitalopram and 52% responded to therapy after 49 days. In these subjects, the QEEG pattern indicated response with a 74% accuracy, demonstrating that changes in asymmetry composite EEG index one week into pharmacological treatment can be predictive of positive response to escitalopram at the end of treatment.

These studies found that there were significant EEG heterogeneities within neuropsychiatric disorders. The existence of these subgroups suggests that different patients within the same neuropsychiatric disorder would differentially respond to medications and indicates that EEG patterns could predict the most effective pharmacotherapy for a specific patient.

EEG collection and Z scores

The first step of the assessment is to collect 21-channel, awake, eyes-closed, digital electroencephalographic (EEG) recordings on subjects who have either washed out their medications for five half-lives or who are currently medication free. The results are reviewed in raw form by an electroencephalographer to ensure that there are no abnormalities that would affect the data from going through the referenced-EEG database. The EEG is then screened to remove any “artifacts” that may exist in the EEG record. These artifacts include such things as muscle twitches, eye blinks, and periods of drowsiness.

Neurometric analysis involves computation of a series of measures that mathematically describe the EEG. These measures are then compared with a database of “normal” EEGs. There are approximately 1,200 measures derived from the EEG component wavelengths and amplitudes. These measures fall into four main categories, i.e., power, coherence, symmetry, phase and frequency. Power is the sum of the amplitudes of the wavelengths in each band, computed on an absolute and relative basis. Relative power indicates the percentage of total power in each band. Coherence measures the synchronization of electrical activity between two channels. In mathematical terms, coherence is the phase shift between similar wavelengths at the two channels. Symmetry measure the ratio of power between a symmetrical pair of electrodes and, lastly, frequency measure the average frequency of the EEG component wavelengths with each band.

Most Neurometric features are highly non-Gaussian in their characteristics. For this reason, the neurometrics are log-transformed to make the distribution more normal (Gaussian) in nature. Many quantitative EEG features also vary consistently with age. To account for the difference between the age of the patient and the age of the subjects in the normative database, these quantitative EEG features are age-regressed using a linear regression equation to yield a “standard-age” quantitative EEG feature. The comparison of the actual values of the Neurometric variables with norms is expressed as a Z score which is defined as:

Z = observed value - normative mean
standard deviation

Development of pattern variables

Neurometrics analysis outputs approximately 2,400 variables (known as univariables) that describe the EEG. To make this data utilizable, reference EEG transforms this data into a smaller set of multivariables (or pattern variables). These pattern variables preserve the information contained in the set of quantitative EEG univariables while retaining some degree of physical interpretation. As such, the data are not simply “mined” to come up with combinations of variables that are indicative of one state or another; instead they are combined according to anatomical location. In some cases, factor analysis is employed to give a greater weight to those univariables that preserve the largest amount of total information of all the univariables in an anatomic group. In other cases, the univariables in an anatomic group are combined in a nonlinear fashion to increase the separation of observed clusters within the data. At present there are 74 pattern variables.

Correlation of pattern variables with patient outcomes

The referenced – EEG variables for historical subjects with known positive and negative clinical outcomes to various psychotropic medications are examined in order to develop a model that will allow the prospective determination of likely patient medication responsivity to these medications. The variables are examined by stratifying the distribution according to the individual medication responsivities represented. Before utilizing this apparent relationship, the appropriateness of the pattern variables are checked. Tests of skewness and kurtosis are conducted for each of the pattern variables to ensure that the original variable distribution is Gaussian. Having ensured a Gaussian distribution, mathematics can be applied that provide a comparison of other subjects with similar patterns demonstrating whether the pattern variable value for the current test in question belongs to the distribution represented by a particular medication or belongs to the distribution defined by some other group (the rest of the population). This procedure is done for all medications represented in the database and for all of the pattern variables that serve as indicators for those medications. The weightings then are averaged to calculate a “score” for each medication.

Calibration against patient records

The final step is to calibrate this score against actual patient records to determine what level of score translates into a specified likelihood response to the medication. For purposes of communication, for the PEER Online service, three levels of responsiveness were created. The first is “sensitive” or “S”. This level suggests that the indicated medication, or group of medications, produced a positive outcome to treatment in 80% or more of cases. “Intermediate” or “I”, the second level, indicates that the responsivity was in less than 80% of cases but more than 35% of cases. The third level, “resistive” or “R” indicates that a response to the medication is seen in fewer than 35% of cases. In other terms, if we formulate Ho (the null hypothesis) in such a way that Ho is true if the patient is not actually responsive to the medication, then the model is calibrated to all for a type I error rate of no more that 20% in the region indicated as “S” and a type II error rate of no more that 35% in the region designated as “R”.

To calibrate the report generator model against these standards, the outcomes database is queried for all subjects’ responses that were not used in the construction of the actual model. This dataset is known as the validation sample. This sample is then divided into two subsets, the first of which is known as the tuning sample and the second is the final validation sample. To complete the model development, the scoring model is run against the tuning sample and the resulting distribution of scores is compared against the known responses. Thresholds for scores are then empirically set to implement the standards of S, I, and R described above, and which are common in such medical reports as, for example, antibiotic sensitivity results. Final validation of the model is made by running the processes, complete with the thresholds that were set, against the final validation sample. In order to preserve the fully prospective nature of this validation, no adjustment of the model parameters, including the thresholds, is made after this process. If the results of this “run” meet the specifications for the previous clinical correlations, the model is then ready to be used for a new patient.

The PEER methodology does not take into account the diagnosis of the patient when offering objective data on any specific medication. Response research has shown, and industry experience corroborates, that diagnosis is often an unreliable predictor of the treatment most likely to be successful for the individual patient. This is one of the fundamental improvements that shared quantitative EEG features correlated to long-term clinical outcomes brings to the practice of psychiatry.

This process can provide objective neurophysiologic data to assist in avoiding the unnecessary risk that comes with the practice of trial and error psychopharmacology, which is also seen through the efficiency of treating a patient, thus reducing suffering and medical costs. The report is unique to each patient’s quantitative EEG features.

Despite the prevalence of mental illness, the treatment of mental illness, and more specifically the most prevalent behavioral disorders of depression, anxiety, bipolar disorder, and has been problematic.

Mental illness is a profound burden on the world’s health and productivity. In fact, it ranks second only to cardiovascular disease in established market economies (Murray & Lopez, 1996).  Despite the magnitude of the problem, the treatment of mental illness, and more specifically the most prevalent behavioral disorders of depression, anxiety, bipolar disorder, and ADHD, has been problematic. The reasons are multifaceted; first, there are no objective decision support tools to support treatment. Unlike other areas of medicine, psychiatry has lacked biomarkers that can help guide pharmacotherapy with similar reliability as bacterial assays that guide antibiotic treatment. Second, the treatment guidelines most widely used for behavioral disorders like depression are based upon little or no clinical evidence. (Insel, 2013).  And finally, diagnoses for behavioral disorders are based upon symptoms and not the underlying physiological etiology. Psychopharmacotherapy is inductive and assumes that certain behavioral symptoms respond to a specific medication class. This selection process is highly subjective.

Without the objective physiologically tests available to other medical specialties or clinical decision support tools such as evidence based treatment guidelines to follow, treating physicians can only rely upon personal experience or antidotal information. In essence, physicians must resort to a trial and error process where they try a treatment and wait and see if it is effective. As noted by Thomas Insel, MD, Director, National Institute of Mental Health in the NIMH Director’s Blog of August 30, 2010, “Clinicians must often resort to trial and error before finding a treatment regimen that works, often subjecting patients to weeks of ineffective treatments or adverse side effects in the process.”

As the term implies, with trial and error it is not uncommon for the first, second, or even third treatment medication to not be sufficiently effective to provide significant improvement. Even with multiple medications administered sequentially, many do not respond adequately to pharmacotherapy (Warden et al., 2007).  The STAR*D study (Rush et al., 2006), which included more than 4,000 subjects with depression, observed a only a 50% response to the primary medication, and for those patients who failed the primary treatment medication, only an additional response of around 25% was observed when switched either to two other classes of antidepressants or cognitive behavior therapy.

Therefore a significant percentage of patients with non-psychotic behavioral disorders are considered ‘Treatment Resistant’ (TR). Treatment resistant are those patients that fail to achieve a significant improvement on the initial and/or subsequent treatment regimens. The prevalence of TR patients varies within the different disorders, but published data suggests that for patients with depression up to 50% did not respond to initial treatment. For patients with Bipolar Disorders, 42% of patients failed to achieve recovery (NIHM-funded STEP-BD Best Practice Treatment Pathway (Perlis, 2006)), 10-40% of Anxiety patients do not respond to treatment and many more have residual symptoms (Bystritsky et al., 2006), and for patients with AD/HD, between 30-40% of patients do not respond to repeated trails of medications (Doyle, 2006).

Since the current standard of care for psychiatric patients is trial and error, and because of this process many patients do not find significant improvement to their mental disorders, treating physicians must resort to other methods to find relieve for their patients suffering. More and more, physicians are using multiple medications simultaneously, a practice known as ‘Polypharmacy’. In a study undertaken between 1996 and 2005, psychiatrists significantly increased their use of polypharmacy such that outpatient visits resulting in two or more prescribed psychotropic drugs increased from 42.6% in 1996 to 59.8% in 2005 and psychiatric visits resulting in three or more such drugs being prescribed doubled, increasing from 16.9% to 33.2% (Mojtabai et al., 2010).Additionally, physicians are frequently using off label medications to treat treatment resistant patients. Off-label prescribing of psychopharmacology is common and perhaps necessary. Off label uses of psychotropic medications can be part of a good psychopharmacology practice. This is not an unusual practice but he standard of care for psychiatry because of the evidence compared to the unmet needs of patients. The difficulty is that there is little evidence supporting many of the treatments for particular disorders and these medications are powerful psychotropic medications that have a significant side effect profile.

As a result of the trial and error treatment process, patients are prescribed different medications for weeks at a time. The problem, aside from the patient not receiving relief from their disease while trying to find the ‘right’ treatment, is that the medications being prescribed, both on and off label medications, are powerful drugs that have significant side effect profiles, including the potential increase of suicidality. The side effect burden of these medications can be significant, both medically and economically.

Without objective clinical decision support tools that are based upon evidence, trial and error will continue to be the standard of care for psychiatric medicine. As Thomas Insel, MD, PhD noted in his keynote address to the American Psychiatric Association at their annual meeting in 2005, “We need to develop biomarkers, including brain imaging, to develop the validity of these disorders. We need to develop treatments that go after the core pathology, understood by imaging”.

The current model for treating patients is based upon symptom-based medical practice and pathophysiologic measurement. Unfortunately, psychiatry is unique among medical specialties in its lack of such measurements. This has hindered both the advancement of psychiatry in general and clinical practice in particular.

The current medical model for treating patients is based upon symptom-based medical practice and pathophysiologic measurement. Psychiatry is unfortunately unique among medical specialties in its lack of such measurements. This has hindered both the advancement of psychiatry in general and clinical practice in particular. Thus, in the current therapeutic model for psychiatry, and in particular of any treatment resistant disorder, psychiatrists much choose treatment medications based upon patient symptoms, response to previous treatment, and behaviors. Significant limitations result because these illness features do not have a simple relationship with medication response. Without a defined pathologic abnormality to treat, or a physiologic marker to guide treatment, psychiatrists have been forced into the position of choosing between large numbers of psychotropic medications without ample evidence to support their choices.

This is particularly true for Major Depressive Disorder (MDD), and specifically for Treatment Resistant Depression (TRD). Whereas the same holds true for bipolar disorders and other psychotic disorders, MDD is very representative of the issues.

Reimbursement and quality control for the treatment of depression by many payers in the US is guided by:

The American Psychiatric Association (APA) treatment guidelines (APA, 2008)

The Texas Medication Algorithm Project (TMAP) treatment algorithm (Crismon, 1999)

Results from the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) Study (Rush, 2006)

These guidelines and reports attempt to provide insight regarding the treatment of depression from the first line therapy through more complex scenarios involving TRD.

The APA guidelines document is a practical guide to the management of major depressive disorder for adults over the age of 18 and represents a synthesis of current scientific knowledge and rational clinical practice. This guideline strives to be as free as possible of bias toward a theoretical posture, and it aims to represent a practical approach to treatment.

The TMAP is a treatment algorithm that constitutes the most extensive and comprehensive development and implementation to date of medication algorithms for persons with serious mental illness. An algorithm is a rule or set of rules that is applied to solving a problem. Medication algorithms are a subset of practice guidelines. They are distinguished by an exclusive focus on medications and by a more step-by-step approach to clinical decisions. Current projects address the treatment of Schizophrenia, bipolar disorder, and major depression. TMAP was initiated by the Texas Department of Mental Health and Mental Retardation in collaboration with a consortium of Texas academic medical centers. The development of the TMAP algorithms incorporated expert panels, literature review, and consensus conferences. The Texas Implementation of Medication Algorithms (TIMA) is the practical, clinician-targeted implementation of TMAP.

STAR*D was set up to evaluate clinical strategies to improve outcomes for patients with TRD, determine the best next-step treatments for depressed patients who do not respond satisfactorily to earlier treatment attempts, and compare relative efficacy and patients’ acceptance of different treatment strategies to relieve depression. Focusing on the common clinical question of what to do next when patients fail to respond to a standard trial of treatment with antidepressant medication, STAR*D aimed a defining which subsequent treatment strategies, in what order or sequence, and in what combination(s) are both acceptable to patients and provide the best clinical results with the least side effects.

Operationally, those who did not respond to a line of treatment were then assigned to either a) augmenting the first antidepressant with other medications or psychotherapy, b) changing to a different antidepressant or psychotherapy, c) adding psychotherapy or discontinuing the first antidepressant medication while switching to psychotherapy, d) switching to another antidepressant, e) augmenting the first antidepressant with other medications, or f) augmenting first antidepressant with other medications or switching to another antidepressant. So after failing the first line of therapy, patients were randomized to the next-step (Level 2) treatment strategies based upon the above options. After indicating which options are or are not acceptable to them, patients will be randomly assigned to a treatment option within those strategies that are deemed acceptable to and medically safe for them. Patients who did not have a satisfactory therapeutic response to their Level 2 treatment were presented with additional treatment options as a third step (Level 3). Again, they were assigned randomly to one option, which included medication switching and medication augmentation, at Level 3 of the treatment protocol. Similarly, Level 4 treatment options were provided for patients who did not respond satisfactorily to the Level 3 of the treatment protocol. STAR*D reported Level 3 and Level 4 results in three trials; two trials for Level 3 and one trial for Level 4(Fava, 2006; McGrath, 2006; Nierenberg, 2006; Rush 2006). Level 3 and Level 4 randomized controlled trials in STAR*D referred to trials in patients who had previously failed two and three antidepressant medication trials, respectively.

Whereas all of these guidelines are valuable, all are problematic. While the APA guidelines provide evidence support for their first line treatment recommendations, no evidence is provided for the TRD treatment recommendations and did not reference any studies specifically evaluating TRD. TMAP was primarily designed using expert panel consensus, which is considered a lower level of evidence (Level 4) on the scale of evidence used by the Centre for Evidence Based Medicine (CEBM). The remission rates in STAR*D using Hamilton-D scores in all treatment arms in the three trials ranged between 6.9% – 24.7%. These rates are relatively low as response rates in excess of 20% are often observed in placebo control treatment arms (Maes, 1996). This implies that the results for STAR*D trials for Level 3 and Level 4 can be considered negative trials.

Other independent analyses of STAR*D have highlighted the disappointing outcomes from the study, particularly its low level of sustained improvement during the follow-up (Fava, 2007; Ghaemi, 2008; Pigott, 2010; Pigott, 2011). Unfortunately, the overly favorable interpretations of the results of STAR*D (Gaynes, 2008) have fostered further trial and error, frequently with no solid clinical or scientific rationale.

References

American Psychiatric Association (APA); Practice Guideline + Resources for: Treatment of Patients with Major Depressive Disorder. Second Edition, July 2008

Crismon, M., et al. The Texas Medication Algorithm Project: Report of the Texas Consensus Conference Panel on Medication Treatment of Major Depressive Disorder. Journal of Clinical Psychiatry, 1999

Depression Guideline Panel Clinical Practice Guideline 5: Depression in Primary Care, vol. 2: Treatment of Major Depression. DHHS, Public Health Service, Agency for Health Care Policy and Research, 1993

Fava, M., et al. A comparison of mirtazapine and nortriptyline following two consecutive failed medication treatment for depressed outpatients: a STAR*D report. American Journal of Psychiatry, 2006

Gaynes, B., et al. The STAR*D study: Treating depression in the real world. Cleveland Clinic Journal of Medicine, 2008

Ghaemi, S., et al. Why antidepressants are not antidepressants: STEP-BD, STAR*D, and the return of neurotic depression. Bipolar Disorders, 2008

McGrath, P., et al. Tranylcypromine versus venlafaxine plus mirtazapine following two failed medication treatments for depression: A STAR*D Report. American Journal of Psychiatry, 2006

Nierenberg, A., et al. A comparison of lithium and T3 augmentation following two failed medication treatments for depression: A STAR*D Report. American Journal of Psychiatry, 2006

Perlis, R., et al. Predictors of recurrence in bipolar disorder: primary outcomes from the systematic treatment enhancement program for bipolar disorders (STEP-BD). American Journal of Psychiatry, 2006

Rush, A., et al. STAR*D Investigators’ Group. Sequenced treatment alternatives to relieve depression (STAR*D): rationale and design. Controlled Clinical Trials, 2004

Rush, A., Trivedi, M., Wisniewski, S., et al. Acute and Longer-Term Outcomes in Depressed Outpatients Requiring One or Several Treatment Steps: A Star*D Report. American Journal of Psychiatry 2006; 163:1905-1917

Rush, A., et al. Combining medications to enhance depression outcomes (CO-MED): acute and long-term outcomes of a single-blind randomized study. American Journal of Psychiatry, 2011

Stroup, T., et al. Schizophrenia Bulletin, 2003

Trivedi, M., et al. TMAP procedural manual: depression module, physician algorithm implementation manual. Physician’s Manual. Dallas: University of Texas Southwestern Medical School, 1998

Warden, D., et al. The STAR*D project results: a comprehensive review of findings. Current Psychiatry Reports, 2007

Although the practice of polypharmacy is growing and being institutionalized by the National Committee for Quality Assurance (NCQA) and virtually mandated by the Affordable Care Act of 2010, there’s a growing number of reports and meta-analyses that the practice of polypharmacy is lacking solid evidence of effectiveness, and worse, an increasing side effect burden that would off-set any benefit.

Although different authors have defined polypharmacy with different meaning, most often, the definition of polypharmacy is made with regard to the specific number of medication prescribed, and most commonly, it is the use of two or more medications to treat the same condition.

There has been a rapid increase in the use of polypharmacy in psychiatry. The reasons may be multifactorial, such as an increasing number of available medications targeting new and different symptoms and receptors, or even the pressure on psychiatrists to focus on medication treatment. Regardless of the reasons, the trend is clear, with psychiatrists now frequently seeing patients presenting on multiple psychiatric medications (Hoffman, 2011).

Polypharmacy of difficult psychiatric patients is rapidly becoming the norm. In a study undertaken between 1996 and 2005, psychiatrists significantly increased their use of polypharmacy such that outpatients visits resulting in two or more prescribed psychotropic drugs increased from 42.6% in 1996 to 59.8% in 2005 and psychiatric visits resulting in three or more such drugs being prescribed almost doubled, increasing from 17% to 33% (Mojtabai, 2010).

The rationale for polypharmacy or augmentation strategies to enhance retention or to increase remission rates is supported by findings from empirical research. As noted by Yury et al., 2009, most patients with unipolar depression do not remit with initial antidepressant monotherapy; second, no monotherapy medication is robustly different from others in achieving remission; third, the lack of response with antidepressant monotherapy leads to high dropout rates among depressed patients; and fourth, the emergence of adverse side effects (e.g., agitation, insomnia) or persistence of some initial baseline symptoms (e.g., anxiety, insomnia) may lead to premature discontinuation from monotherapy.

Yury went on to note that there are significant limitations to the American Psychiatric Association (APA) guidelines to the use of augmentation to treat depression. “First, researchers have noted a lack of data to inform the sequence in which augmentation strategies should be implemented or for identifying the types of patients for whom specific strategies might be most helpful (Fava et al., 2003;). Further, it appears that the majority of medications suggested as augmenting agents have no studies, or few examining them as augmenting agents (Thase, 2001). Thus, it is an irony of current psychiatric practice that the most common augmentation strategies for the treatment of depression may be those with the least evidence of efficacy (Thase, 2004).” As emphasized in Pigott et al., 2010, the APA’s continuation phase guideline is profoundly misguided because there is no apparent benefit for most patients from continued antidepressant drug treatment, yet this practice unnecessarily exposes such patients to significant risks.

The STAR*D study was a large trail that included some augmentations. In the trial, patients who failed to respond to the step-1 medication and agreed to be randomized to augmentation strategies received either 12 weeks of a step-2 medication added to their step-1 medication. Such augmentation resulted in a 39% remission rate, but with a subsequent relapse of up to 67%. Subsequent augmentation had even lower remission and higher relapse rates. As noted by Yury et al., “one might consider the results to be disappointing and reflect and unfavorable risk-benefit ratio given that about 4% of the patients experienced a serious adverse event from the first augmentation and that about 13% had to discontinue the study because of intolerance of side effects” (Yury, 2009). Even more concerning is that the STAR*D study found that 8.6% of step-1 patients reported an increase in suicide ideation during the acute phase treatment (Perlis, 2007) and in another report (Nierenberg, 2010) 71.3 % who had a remission during the acute-care treatment reported increased weight gain and 71.7% reported residual symptoms of sleep disturbance despite having received remission (Pigott, 2011).

Although the practice of polypharmacy is growing and being institutionalized by the National Committee for Quality Assurance (NCQA) and virtually mandated by the Affordable Care Act of 2010, there is a growing number of reports and meta-analyses that the practice of polypharmacy is lacking solid evidence of effectiveness, and worse, an increase side effect burden that would off-set any benefit. Both Hoffman and Pigott have recently noted that patients either wash out or off medication may achieve similar benefit to those of polypharmacy (Hoffman, 2011; Pigott, 2011).

References

Centorrino, F., et al. Multiple versus single antipsychotic agents for hospitalized psychiatric patients: Case-controlled study of risks versus benefits. American Journal Psychiatry, 2004

DePaulo, J., et al. Bipolar disorder treatment: an evidence-based reality check. American Journal of Psychiatry, 2006

Fava, G., et al. Can long-term treatment with antidepressant drugs worsen the course of depression? The Journal of Clinical Psychiatry, 2003

Frye, M., et al. The increasing use of polypharmacotherapy for refractory mood disorders. Journal of Clinical Psychiatry, 2000

Gaynes, B., et al. The STAR*D study: Treating depression in the real world. Cleveland Clinic Journal of Medicine, 2008

Goldberg, J., et al. Depressive illness burden associated with complex polypharmacy in patients with bipolar disorder: findings from the STEP-BD. Journal of Clinical Psychiatry, 2009

Gunderson, J. Clinical Practice. Borderline personality disorder. New England Journal of Medicine, 2010

Helwick, C., et al. Polypharmacy is common in psychiatry, but is more better? Medscape Medical News

Hoffman, D., Schiller, M., Greenblatt, J., Iosifescu, D. Polypharmacy or medication washout: an old tool revisited. Neuropsychiatric Disease and Treatment 2011:7 639-648

Kingsbury, S., et al. Psychopharmacology: rational and irrational polypharmacy. Psychiatry Service, 2001

Mojtabai, R., et al. National patterns in antidepressant treatment by psychiatrists and general medical providers: results from the national comorbidity survey replication. Journal of Clinical Psychiatry, 2008

Mojtabai, R., et al. National trends in psychotropic medication polypharmacy in office-based psychiatry. Archives of General Psychiatry, 2010

Nierenberg, A., et al. Residual symptoms after remission of major depressive disorder with citalopram and risk of relapse: A STAR*D report. Psychological Medicine, 2010

Pigott, H., et al. STAR*D: A tale and trial of bias. Ethical Human Psychol. Psychiatry, 2011

Pigott, H., et al. Efficacy and effectiveness of antidepressants: current status of research. Psychother. Psychosom, 2010

Rush, A., et al. STAR*D Investigators’ Group. Sequenced treatment alternatives to relieve depression (STAR*D): rationale and design. Controlled Clinical Trials, 2004

Rush, A., Trivedi, M., Wisniewski, S., et al. Acute and Longer-Term Outcomes in Depressed Outpatients Requiring One or Several Treatment Steps: A Star*D Report. American Journal of Psychiatry 2006; 163:1905-1917

Rush, A., et al. Combining medications to enhance depression outcomes (CO-MED): acute and long-term outcomes of a single-blind randomized study. American Journal of Psychiatry, 2011

Stroup, T., et al. The NIMH Clinical Antipsychotic Trials on Intervention Effectiveness (CATIE) project: schizophrenia trial design and protocol development. Schizophrenia Bulletin, 2003

Thase, M., et al., Remission rates during treatment with venlafaxine or selective serotonin reuptake inhibitors. British Journal of Psychiatry, 2001

Thase, M. Therapeutic alternatives for difficult-to-treat depression: A narrative review of the state of the evidence. CNS Spectrums, 2004

Yury, C., et al. Meta-analysis of antidepressant augmentation: piling on in the absence of evidence. Ethical Human Psychol. Psychiatry, 2009