Novel Program Effective for Anxiety Disorders in Primary Care
Anxiety disorders are commonly seen in primary care, whether they present on their own, or comorbid with other illnesses. There’s good evidence for both cognitive-behavioral therapy (CBT) and pharmacotherapy in anxiety disorders, but many patients don’t receive such treatment.
A team led by Michelle Craske of UCLA recently evaluated an evidence-based treatment program for anxiety in primary care patients. More than 1,000 patients from primary care practices throughout the U.S. were enrolled, all of whom suffered from one or more of the following: generalized anxiety, panic, social phobia, and PTSD. Patients were randomly assigned to either the active intervention, “CALM” (Coordinated Anxiety Learning and Management), or treatment as usual.
Patients who were randomized to CALM treatment had the option of choosing CBT, medication treatment, or a combination of both. (The majority chose the combination, followed by CBT only, while less than 10% opted for meds only.) The CBT entailed six to eight sessions of computer-aided therapy provided by “anxiety clinical specialists,” who were primarily either social workers or registered nurses with no prior expertise in CBT. They were given a three day intensive CBT training specifically for this research study. The medication treatment was based on an algorithm directing PCPs to start with either SSRIs or SNRIs, and to push the dose as high as the patient could tolerate before switching meds.
The control group received “usual care,” which was whatever treatment their primary care physicians were able to provide, including referral to a mental health specialist.
Did CALM make these patients calmer? Mostly. For all patients except those with primary PTSD, the CALM group reported better outcomes than the controls on standard rating scales such as the Panic Disorder Severity Scale–Self-Report. The effects were moderate with numbers-needed-to-treat (NNT) of five to seven. Most of the benefits were maintained 18 months later. Unfortunately, the authors did not report outcomes separately for CBT vs combination vs meds only, presumably because there weren’t enough subjects to make such comparisons statistically meaningful (Craske MG et al, Arch Gen Psychiatry 2011;68(4):378–388).
TCPR’s Take: We already knew that CBT and drugs work in anxiety. This study shows that these treatments can be implemented effectively in primary care settings. More novel is the suggestion that only three days of training is needed for nurses and social workers to provide effective computer-aided CBT. The computer program used can be viewed on YouTube at www.youtube.com/user/CHAMMPTraining.
Serotonin Transporter Genotype: Involved In Depression After All?
Demonstrating the value of genetic testing in psychiatry is a tricky proposition. The latest case in point is the debate over whether determining variations in the gene coding for the serotonin transporter pump (5-HTTLPR) is useful for predicting future risk for depression.
A widely cited 2003 study showed that the homozygous SS genotype of the 5-HTTLPR gene predisposes people to developing depression if they are subject to stressful life events (Caspi A et al, Science 2003;301(5631):386–389). As a result of the excitement generated by this finding, dozens of studies tried to replicate the result, but two large meta-analyses of these studies, both published in 2009 quashed the enthusiasm, both concluding that there was no “gene × environment” interaction with the S allele (Munafò MR et al, Biol Psychiatry 2009;65(3):211–219; Risch N et al, JAMA 2009;301(23):2462–2471).
Now, these wet blankets are being called into question. A brand new study analyzed all 54 studies they could find (total n=40,749), while the previous meta-analyses only included five and 14 studies, respectively. The new study (Karg K et al, Arch Gen Psychiatry 2011;68(5):444–454) included childhood maltreatment and other specific medical stressors (such as hip fracture, interferon-alfa treatment, and heart disease) rather than limiting the studies to those that measure recall of general “stressful life events.”
This larger study found that the presence of the S allele seems to make patients more sensitive to childhood maltreatment and to specific medical conditions (with p=0.00007 and p=0.0004, respectively). There was less evidence for an association with stressful life events, although this was significant, too (p=0.03). Although odds ratios were not reported, and the 54 studies varied widely in how they measured stressors or depression scores, the findings were generally consistent across all published papers. The authors even calculated that an additional 729 negative unpublished studies would be necessary to negate their results, demonstrating no obvious publication bias.
The authors conclude that the 5-HTTLPR genotype does, after all, moderate the relationship between stress and depression. Readers may also recall a similar reversal involving 5-HTTLPR genotype and SSRI response, in which a larger study showed no relationship between antidepressant response and genotype, after isolated smaller studies had shown a correlation (TCPR, November 2010).
TCPR’s Take: The S allele of the 5-HTTLPR gene may interact with specific early life stressors in the development of depression. But given the complexity of genetic studies and the history of reversal of findings after efforts to replicate, we don’t recommend that you refer your patients for genetic analysis quite yet.
Combined Antidepressants No More Effective than Monotherapy
The well known Star*D trial yielded disappointingly low remission and response rates in depressed patients who were put on citalopram (Celexa) for eight weeks. Recently, some clinicians have advocated starting depressed patients with a combination of antidepressants, in the hopes that targeting multiple neurotransmitters will boost efficacy. While this makes sense theoretically, the practice has thus far been tested only in a couple of small short-term trials, with positive results. A recent NIMH-funded study represents the first large study of combination treatment vs monotherapy as an initial treatment for depressed patients.
A total of 665 patients with chronic or recurrent depression were randomly assigned to one of three treatments: escitalopram (Lexapro) 10 mg to 20 mg + placebo; Lexapro + bupropion SR (Wellbutrin) 300 mg to 400 mg; or venlafaxine XR (Effexor) 150 mg to 300 mg + mirtazapine (Remeron) 15 mg to 45 mg (this last combination has achieved anecdotal renown as “California Rocket Fuel”).
Dosing was increased if patients did not meet remission criteria as the study progressed. Patients were aware of the first medication (Lexapro or Effexor) but were blind to the second medication in their pair, though study physicians were aware of the second medication. Empirically supported psychotherapies for depression or other antidepressants were forbidden during the study. Outcomes were assessed at both 12 weeks and seven months.
There was no difference in response (52% for all groups at 12 weeks, 57% to 59% at seven months) or remission rates (38% to 39% at 12 weeks, 42% to 47% at seven months) among treatments. Further, all groups had roughly equal change on depression rating scales and measures of quality of life, anxiety, and work/social adjustment. Dropout rates also did not differ between treatments. Effexor-Remeron and Lexapro-Wellbutrin had higher side effect frequency and intensity compared to Lexapro-placebo at both 12 weeks and seven months, with the highest intensity, frequency and burden of side effects occurring in the Effexor-Remeron group (Rush AJ et al, Am J Psychiatry online ahead of print).
TCPR’s Take: A couple of much smaller studies showed promising results for combination therapy but the current large trial severely dampens any enthusiasm for starting patients on combined treatment. We recommend following the standard practice of starting with a single antidepressant and tinkering with combinations only if there is no significant response.