Integrating Measurement Into the Management of Bipolar Disorder: An Expert Interview With Gary S. Sachs, MD

August 10, 2006

Editor's Note:

How can patient self-reports on a standard form, aggregate data about routine tests, and other measurements yield better patient outcomes? To find out about the value of integrating measurement into the management of bipolar disorder, Elizabeth Saenger, PhD, Editorial Director of Medscape Psychiatry & Mental Health, interviewed Gary S. Sachs, MD, Associate Professor of Psychiatry, Harvard University; Director, Partners Bipolar Treatment Center, Massachusetts General Hospital, Boston, Massachusetts.

Medscape: What are some of the interesting findings now on bipolar disorder?

Gary S. Sachs, MD: There are a lot of interesting things happening in bipolar disorder research in general, some of which are being presented here in Toronto. One thing researchers are doing that I'm really very excited about -- and a very important theme for our group -- is integrating measurement into management. We have a couple of findings in this area that are really neat.

One set of findings[1] has to do with outcomes when guidelines for routine monitoring of blood indices for patients with bipolar disorder were followed. We've all become increasingly aware of the problem of metabolic syndrome, and of the increased risk for diabetes associated with specific treatment for the general adult population. We also know we're supposed to measure drug levels and the electrolytes of patients being treated with antipsychotic medications. However, recommendations for routine testing were firmed up this year in my facility. We really encouraged patients to come in for testing and made an effort in staff meetings to encourage clinicians to order annual monitoring to establish a routine process for testing at the first visit in any calendar year. The monitoring included recommended routine tests, obtained without any additional clinical indication. We wanted to see what the impact of routine monitoring would be.

When results were analyzed, the findings were quite striking in that abnormal electrolytes were seldom found. A few patients had some abnormal findings; all but 1 (who was bulimic) turned out to be taking diuretic medication. Otherwise, there were no significant findings in anyone else tested.

Additionally, no abnormalities were found through monitoring carbamazepine or lithium levels. However, valproate levels (again, tested without clinical indication) were often determined to be out of range, so it was helpful to monitor these.

We spent over $30,000 monitoring thyroid-stimulating hormone (TSH) levels. Not a single patient was out of range for TSH, so no helpful clinical findings were obtained from this testing.

However, many patients had lipid abnormalities; about 28%. Of interest, the prevalence of abnormal findings was the same among patients currently taking atypical antipsychotic medications as among those not currently taking atypicals.

We're not more interested in saving the lives of people who might be at risk for metabolic syndrome if they're taking or not taking any given particular medication. We're interested in the overall health of all of our patients who are being treated with medication. Lipid screening turned out to be cost-effective, and new, clinically significant findings were obtained.

We also found that patients with elevated lipid levels had big waist circumferences. Our guidelines for routine monitoring will probably be revised next year, limiting some testing to patients determined ahead of time to have risk factors for particular problems. But we will continue to routinely monitor blood lipids and other factors related to metabolic syndrome, such as weight, waist circumference, and blood pressure, in all of our patients.

Medscape: Are there any other things that you would suggest clinicians do on the basis of research findings you're presenting at APA?

Dr. Sachs: Yesterday, we presented data[2] validating a form routinely used in our waiting room. We have developed a standard form that is given out to patients at each follow-up visit. They fill out the form, which we know saves a tremendous amount of clinical time. But we wanted to see what actual data generated through use of the form would look like.

Toward this end, we looked at the way symptoms were reported -- crunched the numbers -- and found that the symptom self-reports elicited via the form were a pretty good indication of what was going on clinically with patients. There was a very high rate of agreement between these self-reports and the clinicians' assessments. This finding makes me think we might be able to start to use a self-report measure not just to measure outcomes when we are doing research, but also to track patients' response to treatment in the clinical context.

The integration of measurement into management allows tracking treatment, in an iterative way, and lets us know whether what we're doing is working or not. We don't want to keep prescribing treatments that aren't working. We also want a very good record of what patients respond to or what they have problems with, whether it's a treatment nonresponse or a medication side effect.

In addition to the study assessing the self-report form, Michael Ostacker[3] from our group presented data on smoking among outpatients with bipolar disorder. He found that smoking and a history of smoking were independent risk factors for having made a prior suicide attempt. One interesting way to think about those findings is that smoking -- or people who become smokers -- could, in a way, be viewed as a biological marker.

Today, we are presenting a poster session of outcome data for medications commonly used in our bipolar clinic.[4] We examined 3 different metrics -- the same ones that the Clinical Antipsychotic Trials in Intervention Effectiveness (CATIE) evaluated: median duration of use, the number of days patients actually stayed on the medication, and the percentage who got to euthymia within the first 3 months. We also looked at what percentage had a euthymic period of 8 months or more within the first year of treatment.

When data from the first year were analyzed, the most interesting findings were that the only statistically significant differences were with the use of lamotrigine. It looked better on all of the metrics than the other medications commonly used in our clinic. I should add that the medications assessed are not used as monotherapy -- evaluation was from whatever point a medication was added to the regimen the patient was already taking. However, the findings from use of lamotrigine were very encouraging in terms of both tolerability and clinical effectiveness, compared with other medications.

Another finding was that there really was no indication of any advantage for olanzapine use compared with any of the other antipsychotic medications -- no significant differences between these treatments were demonstrated. Similarly, no significant differences were seen on any of the outcome measures between standard antidepressants.

Medscape: So, do you think that patients with bipolar disorder possibly differ from patients with schizophrenia in terms of sensitivity to atypical antipsychotics?

Dr. Sachs: It is probable that patients with bipolar disorder are considerably more sensitive to things such as sedation, weight gain, and perhaps other medication-related side effects. But, in many instances, our bipolar patients are having full recoveries. Data show that within about the first 6 months, about 80% were getting to euthymia because of our use of an iterative approach. With this approach, if one medication doesn't work, we try the next one. With use of lamotrigine alone, 81% of our patients got to euthymia within 3 months (vs 60% for lithium; 67% for valproate; 62% for atypical antipsychotics; 68% for antidepressants; and 65% for anxiolytics). Of course, we didn't stop if patients didn't respond to lamotrigine -- they got the next medication and the next, and so on, until a therapeutic response was obtained.

The really good news is that we have a pretty good batting average in the bipolar clinic with respect to treatment outcomes; overall response rates are between 45% and 60%. We now have a lot of tools in our toolbox, all of which work for a substantial percentage of patients. This further encourages us to use the iterative approach -- where we select reasonable treatments, teach patients how to use these, and measure the results. Then, based on findings, we repeat the process as indicated.

Medscape: Is there anything else that you would like to add in terms of the research your group has done and how this might affect clinical practice?

Dr. Sachs: As I've said, the unifying theme for our group is integrating measurement into management. We've been very pleased that other clinicians have been enthusiastic about doing this as well. I can see this approach spreading out and being adopted by more and more clinicians all over. We'll be able to speak the same language whether we're doing patient consults for each other, discussing what's happening with a patient during a case conference, or simply knowing when we refer patients how they're likely to be managed. Having a common language because we use common measures consistently would be good for everybody.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.