Rajnish Mago, MD

Is your patient concerned about daylight savings time? This handout offers tips for getting a good night's sleep.

Train yourself to be awake and sleepy at regular times

• Stick to approximately the same times to wake up and to go to bed. Don’t vary this schedule much on weekends

• Wake up at a fixed time every day-regardless of how long you slept during the night. This is hard to do but will be really helpful by making you sleep better and better over the next few days

• Rest as much as you need to during the day, but do not lie down or nap. Otherwise, your sleep cycle will be worse

• Close the blinds; wear earplugs if necessary

• Sleep in a cool but not cold bedroom. A temperature of about 65°F is ideal for most people

• If you cannot sleep, get up and go into another room

• Only get into bed when you are sleepy

• Stay away from caffeine after about 3 pm

• Establish a bedtime “ritual” that you follow every evening. This may include a warm bath or shower, some light reading, or listening to soothing music

• Avoid alcohol. It may help you to fall asleep, but you will probably have a restless night

," by Rajnish Mago, MD, on which this Tipsheet was based.

This article was originally posted on February 12, 2012 and has since been updated.

Articles

Concerned about daylight savings time? This patient handout offers tips for getting a good night's sleep. 

Sleep Hygiene: Tips on Getting a Restful Night's Sleep

Lamotrigine is an important medication option for the treatment of bipolar disorders. However, many clinicians also use it in patients with a (unipolar) depressive disorder who have not responded adequately to conventional antidepressants. That several retrospective chart reviews of such use have been published indirectly indicates that such “off-label” use is not infrequent.

However, is such use scientifically justified?

Case reports, case series (including “chart reviews”), and uncontrolled, open-label clinical trials are useful and important for preliminary assessment of the efficacy of a treatment, especially to guide further research, but are 

true evidence that a treatment works and should 

 usually lead to change in regular clinical practice.

Studies without randomization, a control group, and double blinding are subject to a series of biases that frequently lead to positive results, a high proportion of which are subsequently disproven or at least not verified (Mago, Lane, and Mahajan, 2009, unpublished data). The history of medicine is littered with treatments that were thought to work based on considerable clinical experience or uncontrolled studies but were later shown to be ineffective.

In assessing the evidence for efficacy of a treatment, much time will be saved and many errors prevented if strong consideration is given to results of double-blind, randomized, controlled trials (RCTs), rather than to other types of research, when deciding whether to change clinical practice.

This is not to say that RCTs do not have limitations or may not be biased in any way, but only that conclusions that can be drawn from studies without a control group, randomization, and blinding are definitely much more limited.

A systematic search of MEDLINE found 4 published RCTs of the use of lamotrigine for MDD (

), all of which used lamotrigine as an add-on to an antidepressant. However, 2 of these studies included patients with bipolar disorder, making them potentially misleading.

Despite some similarities and overlap between MDD and bipolar disorders, there are important differences in their response to many available treatments. Therefore, efficacy studies should ideally not combine unipolar MDD and bipolar disorder patients. In such cases, look to see whether data on one of the subgroups can be extracted independent of the other, without which the results for the unipolar and bipolar patients analyzed together are very hard to interpret.

None of these RCTs found the addition of lamotrigine to the anti­depressant to be superior to the addition of placebo on the 

 out-come measure. Since statistical significance is based on probability, doing many statistical tests on the same data is like tossing a coin until one gets a desired outcome. If you get one chance to toss a coin, there is a 50% chance of getting a “heads.” However, this is no longer true if you have 2 or more tries. The cutoff of 

 < .05 means that there is less than a 5% chance that these results could have been obtained by chance alone if there was, in fact, no difference between the two groups. However, this cutoff assumes that only one statistical test was done. The more statistical tests, the more likely it is that one of these tests will be positive just by chance.

The more statistical tests are done, the higher the bar should be set. The combined 

 value for all the tests should be less than .05. For example, if 5 outcome measures are statistically tested, the cutoff 

 value should be .05 divided by 5, or .01 (called a Bonferroni correction). An alternative way to avoid this problem of multiple testing that is commonly used is to define a “primary” outcome measure in advance (as was done in these studies), use the cutoff of .05 for the primary outcome measure, and consider any other statistical testing as requiring subsequent confirmation. In that case, the primary outcome measures for these studies should be our focus.

Now that clinical trials are registered online before they are begun, it is easier to verify that the primary outcome measure was, in fact, specified in advance. As noted above, none of these RCTs found a statistically significant difference between lamotrigine and placebo on the primary outcome measure.

The largest and most methodologically sound study was that by Barbee and colleagues.

 One reason for this was that they not only considered patient report of failure to respond to an antidepressant but also first prospectively treated these patients with paroxetine alone. Only patients whose Hamilton Depression Rating Scale score was 15 or greater after 8 weeks of taking paroxetine alone were randomized to addition of lamotrigine or placebo. The wisdom of this approach is supported by the fact that nearly half of the patients in whom paroxetine monotherapy was started improved enough to not make it to the randomized phase.

 In studies of treatment-resistant depression, it is becoming standard to not consider a patient treatment-resistant simply on the basis of previous history without blinded, prospective treatment with an antidepressant.

The study by Barbee and colleagues did not find that the addition of lamotrigine was superior to the addition of placebo on the Montgomery Asberg Depression Rating Scale, the Hamilton Depression Rating Scale, the Clinical Global Impression–Severity Scale, or the Clinical Global Impression–Improvement Scale.

 If a study does not find a statistically significant difference, an important question to ask is if the study was adequately powered (ie, if it had a sufficient number of patients to detect a meaningful difference if one had existed).

Before the start of their study, Barbee and colleagues had calculated that 90 subjects who were in the double-blind phase would provide enough statistical power to detect an expected difference between the groups. Therefore, the study was not underpowered.

On the basis of currently available evidence, clinicians are urged to not prescribe lamotrigine for MDD. If a detailed and thorough clinical evaluation finds no specific evidence of bipolar disorder (mania or hypo­mania) in a particular patient, it is not appropriate to prescribe lamotrigine for that patient, since there is no evidence that that particular patient has bipolar disorder.

Barbee JG, Jamhour NJ. Lamotrigine as an augmentation agent in treatment-resistant depression. 

 Rocha FL, Hara C. Lamotrigine augmentation in unipolar depression. 

 Gutierrez RL, McKercher RM, Galea J, Jamison KL. Lamotrigine augmentation strategy for patients with treatment-resistant depression. 

 Barbee JG, Thompson TR, Jamhour NJ, et al. A double-blind placebo-controlled trial of lamotrigine as an antidepressant augmentation agent in treatment-refractory unipolar depression.

Normann C, Hummel B, SchÃ¤rer LO, et al. Lamotrigine as adjunct to paroxetine in acute depression: a placebo-controlled, double-blind study. 

Barbosa L, Berk M, Vorster M. A double-blind, randomized, placebo-controlled trial of augmentation with lamotrigine or placebo in patients concomitantly treated with fluoxetine for resistant major depressive episodes. 

 Santos MA, Rocha FL, Hara C. Efficacy and safety of antidepressant augmentation with lamotrigine in patients with treatment-resistant depression: a randomized, placebo-controlled, double-blind study. 

Lamotrigine is an important option in bipolar disorders. However, many clinicians also use it in patients with a (unipolar) depressive disorder who have not responded adequately to conventional antidepressants.

Lamotrigine for Major Depressive Disorder Is Inappropriate

A landmark paper published in 2006 reported that the use of SSRIs after the 20th week of pregnancy is associated with an elevated risk of a rare but serious abnormality-persistent pulmonary hypertension of the newborn (PPHN).

 PPHN involves persistence after birth of the normal fetal high pulmonary vascular resistance. This leads to right-to-left shunting of blood, low pulmonary blood flow, hypoxemia, and severe respiratory failure.

 At least 6 subsequent studies have addressed this association between use of SSRIs and PPHN, with apparently contradictory results, which has led to uncertainty about the association.

The FDA reviewed the literature on the subject and issued a notification in December 2011, stating, “given the conflicting results from different studies, it is premature to reach any conclusion about a possible link between SSRI use in pregnancy and PPHN.”

 The agency advised health care professionals “not to alter their current clinical practice of treating depression during pregnancy.” Health care professionals were also advised to “weigh the small potential risk of PPHN that may be associated with SSRI use in pregnancy against the substantial risks associated with under-treatment or no treatment of depression during pregnancy.”

These studies offer an excellent opportunity to understand which study designs are best-suited to evaluate rare outcomes. If an outcome occurs rarely (eg, PPHN), an extremely large number of patients with and without the risk factor (eg, treatment of the pregnant woman with an antidepressant) would have to be followed to be able to evaluate whether patients who did or did not have the risk factor were equally at risk for the outcome.

Early in an investigation of a potential association, it is not viable to study very large numbers of subjects. Instead, a case-control study is often used. Case-control studies are well suited to initial investigation of rare outcomes and outcomes that take a long time to develop (eg, lung cancer with smoking). The researchers start with a sufficient number of subjects who have the outcome of interest, identify an appropriate control group without the outcome, and then assess the proportion of subjects with or without the outcome who were exposed to the risk factor being studied.

Because case-control studies can have certain problems, such as recall bias (subjects may be more likely to remember exposure to the risk factor than the controls), their findings usually need to be confirmed by large cohort studies. In cohort studies, subjects with or without the risk factor are identified and followed up over time to determine whether they develop the outcome of interest.

When studying potential risk factors, confounders need to be taken into account. The confounder is a third factor that is associated with both the putative risk factor and the outcome and, therefore, results in an apparent association between the putative risk factor and the outcome.

Factors other than the putative risk factor being evaluated that are associated with an outcome (ie, potential confounders) can be controlled for in several ways. First, inclusion criteria can limit the subjects to a more homogeneous subgroup (eg, the studies cited in this article usually only included infants born after 33 or 34 weeks of gestation). Second, in case-control studies, the controls can be matched to cases on specific factors. To illustrate, Chambers and colleagues

 matched the controls to the cases on the hospital of birth and the approximate date of birth. Third, the potential confounders can be controlled for in the statistical analysis.

 shows the various published studies of the potential association of PPHN with maternal use of an antidepressant during the pregnancy. The studies variably controlled statistically for some potential confounders. For example, while one study controlled for maternal diabetes, race, BMI, and date of birth, another controlled for high maternal age, first parity, maternal BMI, maternal smoking, and year of birth.

 Unfortunately, none of the studies took all possible confounding factors into account. Apparent risk factors for PPHN include male sex, black or Asian race, high pre-pregnancy BMI, high infant weight, cesarean section, maternal diabetes or asthma, and maternal use of aspirin and NSAIDs.

 It is therefore possible that some or all of the apparent association between antidepressant use during pregnancy and PPHN is a result of these confounding factors (or other, unknown confounding factors) that were not adequately controlled for.

 
Studies of potential association of persistent pulmonary hypertension of the newborn with maternal use of an antidepressant
 

In its notification of December 2011, the FDA referred to various studies, all of which had very few patients with PPHN.

 The FDA data summary did not include 2 more recently published large cohort studies that 

 find an association between the use of SSRIs and PPHN.

 All 4 of the studies with a substantial number of infants with PPHN found a statistically significant association between maternal use of antidepressants and PPHN in the infant.

 The results of the 3 studies with only a few infants with PPHN showed no statistically significant association.

 This illustrates the main point of this article: 

for rare outcomes, either a very large number of subjects must be studied in a cohort study or a case-control study that starts with an adequate number of cases must be used.

Finally, a few brief comments about issues that are not the focus of this column but are important. It is essential to keep in mind that the risk of PPHN in infants of mothers treated with serotonergic antidepressants during pregnancy, even if real, is quite small: the risk is about 0.05% to 0.1%.

 It appears that the risk is greater with antidepressant use during late pregnancy but may be elevated with use during early pregnancy as well.

No significant differences were found between the various SSRIs regarding risk of PPHN.

 While one study suggests that non-SSRI antidepressants may not be associated with increased risk of PPHN, the power of the study to detect such effects was limited and this finding has not been verified.

 Tricyclic antidepressants have been associated with higher rates of congenital malformations and a variety of neonatal complications.

 Thus, these agents are not recommended as alternatives to SSRIs in pregnant women.

Untreated mental disorders (eg, MDD) may also have negative consequences for the pregnancy, although this has not been well studied. From data on groups of patients, it is hard to predict for a particular patient what the risk of not treating with an antidepressant will be. Some patients may not exhibit behaviors that can mediate the effect of depression or other mental disorders on pregnancy outcomes; it is not clear whether the depression will lead to poor outcomes for the fetuses or neonates in these cases.

Do not assume that antidepressants are effective for all patients with MDD or that an effective treatment is being withheld; only a minority of patients are likely to have a true antidepressant response. Lastly, withholding treatment is not the only alternative to using a serotonergic antidepressant in a pregnant woman. Bupropion 

 be safer, and psychotherapy may be a viable alternative treatment for many patients.

 Chambers CD, Hernandez-Diaz S, Van Marter LJ, et al. Selective serotonin-reuptake inhibitors and risk of persistent pulmonary hypertension of the newborn. 

 Occhiogrosso M, Omran SS, Altemus M. Persistent pulmonary hypertension of the newborn and selective serotonin reuptake inhibitors: lessons from clinical and translational studies. 

Källén B, Olausson PO. Maternal use of selective serotonin re-uptake inhibitors and persistent pulmonary hypertension of the newborn. 

 Andrade SE, McPhillips H, Loren D, et al. Antidepressant medication use and risk of persistent pulmonary hypertension of the newborn. 

 Wichman CL, Morre KM, Lang TR, et al. Congenital heart disease associated with selective serotonin reuptake inhibitor use during pregnancy. 

Reis M, Källén B. Delivery outcome after maternal use of antidepressant drugs in pregnancy: an update using Swedish data.

 Kieler H, Artama M, Engeland A, et al. Selective serotonin reuptake inhibitors during pregnancy and risk of persistent pulmonary hypertension in the newborn: population based cohort study from the five Nordic countries. 

 Wilson KL, Zelig CM, Harvey JP, et al. Persistent pulmonary hypertension of the newborn is associated with mode of delivery and not with maternal use of selective serotonin reuptake inhibitors. 

 US Food and Drug Administration. FDA Drug safety communication: selective serotonin reuptake inhibitor (SSRI) antidepressant use during pregnancy and reports of a rare heart and lung condition in newborn babies. http://www.fda.gov/Drugs/DrugSafety/ucm283375.htm#data. Accessed June 14, 2012.

Hernández-Díaz S, Van Marter LJ, Werler MM, et al. Risk factors for persistent pulmonary hypertension of the newborn. 

 Delaney C, Cornfield DN. Risk factors for persistent pulmonary hypertension of the newborn. 

It appears that the risk is greater with antidepressant use during late pregnancy but may be elevated with use during early pregnancy as well.

Antidepressants and Persistent Pulmonary Hypertension of the Newborn

This is the third in a series of occasional columns that aim to help clinicians interpret research related to a clinical question. Perhaps more important, the tips and discussion of the issues aim to improve clinicians’ understanding of research methodology and critical appraisal of the literature in general. Your questions, comments, and suggestions are eagerly solicited at

While SSRIs and SNRIs are valuable in the treatment of major depressive disorder (MDD), partial response or nonresponse occurs in many patients.

 In a survey of clinicians, the most frequently chosen agent for addition to an SSRI after inadequate response was bupropion (30%).

 In that survey, bupropion was chosen as the “augmenting” agent by more experienced clinicians; by US rather than Canadian clinicians; and by clinicians from community, individual practice, or group settings rather than from academic settings.

Many treatment strategies that are widely practiced are not based on scientific evidence. Make it a habit to question treatments that are routinely used. Some of these treatments may in fact turn out to be effective, but some others will be shown to be myths. The history of medicine is full of treatments that were widely used for years before being shown to be myths.

Is the strategy of adding bupropion to an SSRI or SNRI based on reliable scientific studies or mainly on the hypothesis that combining antidepressants with different mechanisms of action may result in greater efficacy than using either one of them alone?

 have suggested that augmentation of SSRIs or SNRIs with bupropion may be efficacious.

When trying to determine whether a treatment is efficacious or not, look for only double-blind, randomized, controlled trials (RCTs). Case reports and uncontrolled studies can only suggest that a treatment 

 work, but such treatments are frequently disproved. Busy clinicians can save a lot of reading time and avoid confusion by focusing 

Before looking for evidence, we should define our questions. Is ad­dition of bupropion to an SSRI or SNRI helpful in patients (1) with minimal or no response to the SSRI or SNRI? (2) with response to the SSRI or SNRI (greater than 50% improvement) but failure to achieve remission? or (3) who are just starting treatment with the SSRI or SNRI, ie, from the outset? (The possibility that bupropion may treat some of the adverse effects of serotonergic antidepressants is beyond the purview of this column.) Importantly, we must distinguish from 

 from an SSRI to bupropion (ie, is any perceived benefit from the addition of bupropion to an SSRI better than simply switching to bupropion?).

A systematic search of the MEDLINE and Scopus databases was done (last on December 8, 2011). Many clinicians will be surprised to know that no RCT of the addition of bupropion to an SSRI or SNRI has ever been conducted. This includes a comparison to simply continuing the SSRI or SNRI, switching to bupropion, or other possible comparison strategies. There are, however, other types of studies that shed some tentative light on the issue and a discussion of their limitations may be of general educational value.

The STAR*D study included the addition of bupropion in patients with MDD who did not achieve remission after 12 weeks of treatment with citalopram.

 This strategy was compared with augmentation of ci­talopram with buspirone. However, in this large sample (N = 565), no statistically significant difference was found in remission (the primary outcome measure) between the two groups.

If no statistically significant difference is found between two treatments, we 

 conclude that they work equally well or equally poorly. Without a placebo control group, it is not possible to say whether this particular study in this particular set of patients had the ability to detect a meaningful effect or difference. This concept is called “assay sensitivity” and is analogous to a control window in a home pregnancy test that indicates that the test is working correctly. In clinical trials, if one or more of the treatments is superior to placebo, this indicates that the “assay” is working.

However, in the STAR*D study, citalopram plus bupropion 

 found to be better than citalopram plus buspirone on several secondary outcome measures.

 Why do researchers differentiate between primary and secondary outcome measures? Because hypothesis testing is based on probability, and the more statistical tests of comparison that are done, the greater the likelihood that one test will be “significant” by chance alone. Secondary outcome measures should either be corrected for multiple comparisons (a simple way for readers to do this is to divide 0.05 by the number of comparisons and use that 

 value as the cutoff for statistical significance) or be considered tentative and in need of confirmation.

In the STAR*D study, the only difference that met this lower cutoff was number of dropouts because of intolerance (

 < .0009) (ie, patients receiving citalopram+buspirone dropped out because of adverse effects more often than those receiving citalproam+bupropion). Notably, the proportion of patients who had a response (50% or greater reduction) on self-rated depression did not differ between the two groups. Thus, we cannot conclude with confidence from this study that the combination of an SSRI and bupropion is efficacious on the basis of either response or remission as the outcome.

A recent, small clinical trial from India randomized 60 patients with MDD who had not responded adequately to a 4-week trial of an SSRI to the addition of either bupropion or placebo.

 The study found that add-on bupropion was significantly superior to add-on placebo in reducing the severity of depression. However, a few serious problems with the study design prevent us from drawing confident conclusions.

-blind rather than a double-blind manner, which is not a minor problem. Remember that investigators are likely to have conscious and unconscious biases in favor of the treatment they are evaluating, perhaps even more so than patients. This bias not only can affect investigators’ ratings but also may be inadvertently communicated to the patients. Therefore, double-blind is not an optional feature but rather is crucial to the conduct of scientifically valid research.

Second, it would be more appropriate to set the minimum duration of the trial of the SSRI at 6 (or even 8) weeks before patients are considered to have an inadequate response to the antidepressant.

Third, the SSRI trial was received before patients entered this study. Thus, their adherence and the trajectory of improvement could not have been reliably determined. It is now standard to conduct a 

trial of the antidepressant; all patients also receive a placebo to determine whether patients have an adequate response to the SSRI.

Fourth, during the 4 weeks of the trial, the dose of the SSRI was increased if needed. This added a potential source of bias that is hard to evaluate.

Recently, the Combining Medications to Enhance Depression Outcomes (CO-MED) study looked at our third issue above.

 This randomized single-blind trial assessed whether the combination of escitalopram and bupropion 

in patients with MDD who were not treatment-resistant was superior to escitalopram and placebo. There was no difference in remission rates after 12 weeks of treatment.

How would the single-blind nature of administration of bupropion or matching placebo in this study affect our interpretation of the result?

For any design feature, ask yourself if it 

 the findings of the study. In this case, if the investigator knew which patients were getting a placebo and which were getting bupropion and an unconscious bias operated, wouldn’t the investigator be less likely to rate the patients who were receiving placebo as improved? If so, lack of investigator blinding would make it more likely that in combination with escitalopram, bupropion would be found more efficacious than placebo. Since this was not found, we can conclude that the single-blind actually strengthens the conclusion that bupropion is 

 more efficacious than placebo when added to escitalopram from the outset.

If, in the future, bupropion augmentation is more convincingly shown to be superior to augmentation with a placebo, the question remains about the mechanism of this augmentation, which, in turn, leads to some important clinical questions:

1. Is the addition of bupropion more efficacious than increasing the dose of the SSRI or SNRI? Bupropion is a moderately potent inhibitor of cytochrome P-450 2D6. Therefore, it may increase the levels of some antidepressants, including duloxetine.

 However, another study did not find plasma levels of paroxetine to be elevated after addition of bupropion.

 Conversely, fluoxetine, fluvoxamine, sertraline, and paroxetine may, at least in theory, increase bupropion levels by inhibiting cytochrome P-450 2B6 isoenzyme.

2. Is the addition of bupropion more efficacious than switching to bupropion? That is, is the addition of bupropion effective only because bupropion alone is efficacious in these patients-or can we truly refer to the combination as 

 Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. 

 Mischoulon D, Nierenberg AA, Kizilbash L, et al. Strategies for managing depression refractory to selective serotonin reuptake inhibitor treatment: a survey of clinicians. 

 Lam RW, Hossie H, Solomons K, Yatham LN. CitalÂ­opram and bupropion-SR: combining versus switching in patients with treatment-resistant depression. 

 Trivedi MH, Fava M, Wisniewski SR, et al; STAR*D Study Team. Medication augmentation after the failure of SSRIs for depression. 

 Gulrez G, Badyal DK, Deswal RS, Sharma A. Bupropion as an augmenting agent in patients of depression with partial response. 

. 2011 Sep 3; [Epub ahead of print]. doi:10.1111/j.1742-7843.2011.00788.x.


 Rush AJ, Trivedi MH, Stewart JW, et al. Combining medications to enhance depression outcomes (CO-MED): acute and long-term outcomes of a single-blind randomized study. 

 Sandson NB, Armstrong SC, Cozza KL. An overview of psychotropic drug-drug interactions. 

 Kennedy SH, McCann SM, Masellis M, et al. Combining bupropion SR with venlafaxine, paroxetine, or fluoxetine: a preliminary report on pharmacokinetic, therapeutic, and sexual dysfunction effects. 

While SSRIs and SNRIs are valuable in the treatment of major depression, partial response or nonresponse occurs in many patients. Research has found that bupropion was the most frequently chosen agent for addition to an SSRI after inadequate response.

Bupropion After Nonresponse or Partial Response to an SSRI or SNRI?

For the first installment of this series, click 

This is the second installment of a new series in which clinically relevant research is briefly discussed and, perhaps more important, a few tips on how to read and interpret research studies are presented. Your feedback, suggestions, and questions are eagerly solicited at 

More efficacious treatments are sorely needed for the treatment of the major depressive episodes of bipolar I and bipolar II disorders (“bipolar depression”). For example, lamotrigine has been found to be more effective than placebo for delaying occur-rence of both depressive and, to a lesser extent, manic episodes during 18 months in patients with bipolar I disorder (who were recently or currently in a manic, mixed, or depressive episode and whose condition was first stabilized with the use of lamotrigine).

 Its efficacy in the acute phase of bipolar depression is, however, much less clear.

The published literature on the treatment of acute bipolar depression with lamotrigine was somewhat misleading until recently. One clinical trial found lamotrigine to be more efficacious than placebo on at least some of the outcome measures.

The American Psychiatric Association’s practice guideline for bipolar disorder advocates using lamotrigine as one of the first-line options for bipolar depression.

 However, 4 other randomized, placebo-controlled clinical trials (similar in size to the published one) were 

 published (although the results of 1 of these 4 were mentioned as part of a review article). These 4 trials did not find lamotrigine to be more efficacious than placebo on most of the outcome measures. This was a prominent example of publication bias in which studies that fail to find a treatment to be effective are much less likely to be published, thus significantly biasing the published literature in favor of the proposed treatment.

TIP: In assessing whether a particular treatment is efficacious for a particular disorder, or reading a review or meta-analysis that attempts to do this, it is essential that 

 randomized, controlled clini-cal trials on the topic be taken into consideration. Unfortunately, this is commonly not done, so readers must be vigilant for this.

Finally, in 2008 the results of all 5 randomized, placebo-controlled studies of the efficacy of lamotrigine monotherapy for acute bipolar depression were published in a single article.

These studies randomized about 200 patients each and lasted for 7 to 10 weeks. Three of the studies included only patients with bipolar I disorder, 1 included only those with bipolar II disorder, and 1 included patients with both. The final dosage of lamotrigine was 200 mg/d in 3 studies, 1 study compared 50 mg/d and 200 mg/d with placebo, and 1 study used a flexible dosage between 100 and 400 mg/d.

In each study, 1 of the 2 clinician-rated depression severity scales that were completed in all 5 studies (

 [HDRS] and the Montgomery-Asberg Depression Rating Scale) was identified as the “primary” outcome measure. Other outcome measures (for example, the Clinical Global Impression–Severity scale, and the Clinical Global Impression–Improvement scale) were considered “secondary” outcome measures.

TIP: Why have multiple outcome measures? In clinical trials, it is standard practice to measure the same outcome by more than 1 method and to evaluate a variety of outcomes. This is because different outcome measures may capture a different aspect of change and to a different extent. For example, rating scales for severity of depression that cover a specified list of symptoms in a specified way may reach different conclusions from the Clinical Global Impression–Improvement scale in which clinicians assess patients and simply rate the patient as being “minimally improved,” “much improved,” “very much improved,” and so on.

TIP:Why would 1 of the measures be identified as the “primary” outcome measure? Statistical testing is based on probability. The more times you flip a coin, the more chances that there will be at least 1 tail. Similarly, if multiple statistical tests are done to compare several different outcome measures, this increases the likelihood that at least 1 of the tests will be “statistically significant” just by chance. Therefore, researchers should identify in advance that 1 of the comparisons is the “primary” outcome measure. 

The difference in improvement in severity of depression between patients treated with lamotrigine or placebo, on the basis of the depression rating scale that was the predefined primary outcome measure for that study, was not statistically significant in any of the 5 studies (ie, was not such that it was unlikely to have been due to chance). In 1 study, lamotrigine was statistically significantly better than placebo on improvement as assessed by important secondary outcome measures.

 In the other 4 studies, however, lamotrigine was not statistically significantly better than placebo even on the secondary outcome measures.

TIP: When a treatment is found not to be statistically significantly more efficacious than placebo, this could be because either (1) the patients receiving the drug were not improving much, or (2) the patients receiving the placebo were improving almost as much as those receiving the drug. This distinction should be obvious but is commonly ignored.

Looking at the mean improvement on the various outcome measures in these 5 studies, it appears that the mean improvement with lamotrigine in all 5 studies was broadly similar; however, in 4 of the studies, patients who received placebo also improved nearly as much, thus largely obliterating the drug-placebo difference.

TIP: When drug-placebo differences are small, or when different studies on the same question don’t agree-both of which were true in this case-one approach may be to combine the data from all patients in all the studies. This is called a “pooled analysis.”

A pooled analysis of individual data of the 1072 patients who had participated in these 5 trials was in fact performed and published in 2009.

 On the basis of a 50% reduction in the HDRS score, 44% of patients who received lamotrigine were “responders” compared with 35% who received placebo, and this difference was statistically significant (ie, unlikely to be due to chance).

TIP: In assessing how well a treatment works, having large numbers of patients in a study or pooling data from various studies may allow us to conclude-even for small differences between the treatment groups-that these differences are probably real. We must keep in mind, however, that such statistically significant differences may nevertheless be of too small a magnitude to be of clinical significance.

In this pooled analysis, there was only a 9% difference in the proportion of patients who were responders (had a 50% or more reduction in the severity of depression) in the lamotrigine and the placebo groups.

TIP: In all instances, it is very important that we assess the efficacy of a treatment not only for the patient group as a whole but also for clinically relevant 

 on the basis of our knowledge of the disorder. In many cases, a subgroup can be identified that is much more likely to respond to a particular treatment. Clinicians may then think differently about the proposed treatment for each of the subgroups. 

In a planned subgroup analysis for patients with more 

 depression (HDRS score ≥ 23), lamotrigine was statistically significantly better than placebo. In this subgroup, 46% of the patients in the lamotrigine group were responders compared with 30% in the placebo group, a 16% difference. On the other hand, a higher proportion of patients with less severe depression (ie, HDRS ≤ 24) responded to placebo (45%) and there was no statistically significant difference between the proportion of responders who received lamotrigine (48%) and the proportion of those who received placebo.

TIP: Since post hoc subgroup analyses can be a way of trying to salvage a study that failed to find a statistically and/or clinically significant effect, it is important to see whether a subgroup analysis was planned (as in this study) or was part of a “fishing expedition” after the study. If it was not planned, we should (1) think about whether the subgroup makes sense on the basis of our knowledge of the disorder; and (2) consider the findings of the subgroup analyses to be tentative or “hypothesis generating,” with verification in future studies being essential.

Putting the results into clinical context

A key factor in considering the potential use of lamotrigine to treat acute bipolar depression is that lamotrigine takes several weeks to titrate. On the one hand, in 7- to 10-week studies this reduces the amount of time that patients receive the final dose of lamotrigine being tested. On the other- hand, a longer duration increases the chances that patients in the placebo group may improve as well, because of spontaneous improvement, or the supportive experience of being in a clinical trial, or both.

A 9% drug-placebo difference overall is just below what is commonly considered to be clinically meaningful. (A 10% to 20% drug-placebo difference is commonly considered clinically meaningful, depending on the seriousness of the disorder, cost and adverse effects of the treatment, and so on.) However, the drug-placebo difference was more substantial in patients who had more severe depression.

While this article focuses mainly on lamotrigine, medications recommended as first-line treatment options for acute bipolar depression also include lithium, quetiapine, olanzapine, and the SSRIs.

 Of these, the olanzapine/fluoxetine combination (OFC) and quetiapine have received FDA approval for use in bipolar depression. Critical review of the data that support the efficacy of these and other treatments is beyond the purview of this article. It seems that some of these alternative treatments may be effective. For example, the proportion of responders in the initial 2 quetia-pine trials was 61% versus 39% for those who received placebo, a difference we would consider clinically meaningful.

 However, there is a paucity of what the clinician sorely needs-unbiased studies that compared the efficacy of these treatments in the same study (“head-to-head”). I am unaware of any study that compares the efficacy of lamotrigine with that of quetiapine for acute bipolar depression.

TIP: Since study samples and study design can vary considerably, it is highly inadvisable to do a simple comparison of the efficacy results for 2 medications if they were from 2 different studies. One medication should be considered more efficacious than the other 

 if the 2 were directly compared in the same study and the study design was not substantially biased in favor of 1 of the medications.

Lamotrigine and OFC were directly compared in the same clinical trial after 7 and 25 weeks of treatment.9,10 OFC was found to be statistically significantly more efficacious than lamotrigine on several, but not all, outcome measures. At the end of 6 months, the proportion of patients who were responders was 64.4% for OFC and 55% for lamotrigine. This difference, while

 significant, is again just below the threshold usually accepted as being 

 significant. On the other hand, OFC also had more treatment-emergent adverse effects and a greater incidence of clinically significant weight gain (defined by convention as 7% or more of the patient’s baseline weight; 34% with OFC vs 2% with lamotrigine) and of new-onset hypercholesterolemia (16% with OFC vs 4% with lamotrigine).

When all is said and done, should we use lamotrigine as one of the treatment options for acute bipolar depression? First, not all potential treatments for acute bipolar depression are as effective as quetiapine and OFC seem to be. Second, in addition to data on efficacy of these medications, the considerable differences in the incidence of adverse effects that may be medically serious is critical to the choice of the medication for a particular patient. Finally, if a particular medication has already been considered as the one that a particular patient will probably take for maintenance treatment, while this is not absolute, this is an important factor in support of its use for the acute depression as well. Thus, lamotrigine is appropriately considered by various practice guidelines as 

 of the first-line options for acute bipolar depression.

 Goodwin GM, Bowden CL, Calabrese JR, et al. A pooled analysis of 2 placebo-controlled 18-month trials of lamotrigine and lithium maintenance in bipolar I disorder. 

 Calabrese JR, Bowden CL, Sachs GS, et al. A double-blind placebo-controlled study of lamotrigine monotherapy in outpatients with bipolar I depression. 

 American Psychiatric Association. Practice guideline for the treatment of patients with bipolar disorder (revision). 

 Nassir Ghaemi S, Shirzadi AA, Filkowski M. Publication bias and the pharmaceutical industry: the case of lamotrigine in bipolar disorder. 

http://www.medscape.com/viewarticle/579046

 Calabrese JR, Huffman RF, White RL, et al. Lamotrigine in the acute treatment of bipolar depression: results of five double-blind, placebo-controlled clinical trials. 

 Geddes JR, Calabrese JR, Goodwin GM. Lamotrigine for treatment of bipolar depression: independent meta-analysis and meta-regression of individual patient data from five randomised trials. 

 Yatham LN, Kennedy SH, Schaffer A, et al. Canadian Network for Mood and Anxiety Treatments (CANMAT) and International Society for Bipolar Disorders (ISBD) collaborative update of CANMAT guidelines for the management of patients with bipolar disorder: update 2009. 

 Weisler RH, Calabrese JR, Thase ME, et al. Efficacy of quetiapine monotherapy for the treatment of depressive episodes in bipolar I disorder: a post hoc analysis of combined results from 2 double-blind, randomized, placebo-controlled studies. 

 Brown EB, McElroy SL, Keck PE Jr, et al. A 7-week, randomized, double-blind trial of olanzapine/fluoxetine combination versus lamotrigine in the treatment of bipolar I depression. 

This is the second installment of a new series in which clinically relevant research is briefly discussed and, perhaps more important, a few tips on how to read and interpret research studies are presented. Your feedback, suggestions, and questions are eagerly solicited at rajnish.mago@jefferson.edu.