This little book tells you about bright light therapy for depression and sleep, how it can be used, and why we need it. CHAPTER 5 BRIGHT LIGHT FOR DEPRESSION: THE SCIENTIFIC EVIDENCE 5.A.     Is bright light only good in winter? 5.B.     A perspective on antidepressant              medication

Table of Contents Chapter 1 We Wilt In The Dark Chapter 2 Depression & The Droopy Dims Chapter 3 Treating Depression With Light Chapter 4 Using Antidepressant Drugs & Counseling Chapter 5 Bright Light For Depression: The Scientific Evidence Chapter 6 Advanced & Delayed Sleep Phase: Bright Light Treatment Chapter 7 Side Effects of Bright Light Chapter 8 Good Sleep Habits Chapter 9 About This Web Book: Printing & Copying Chapter 10 About Dr. Kripke

Bright light treatment for winter depression has been supported by dozens of studies demonstrating beneficial results (though there have been a few studies with unsuccessful results).   The Clinical Practice Guidelines issued by the U.S. Department of Health and Human Services recognize bright light as a generally accepted treatment for winter depression. 

Until recently, there were rather few studies of bright light treatment of nonseasonal depression, however, there are now at least 15 controlled studies showing that bright light reduces symptoms in nonseasonal depression.  Scientists rate how depressed patients are both before and after treatment using descriptive methods called depression rating scales.  For example, a doctor who talked to the patient give scores for how much the patient seemed sad, guilty, without appetite, suicidal, and so forth to add up a total depression score.  One of the most widely used scoring methods is the Hamilton Depression Rating Scale or HDRS.  Because depressed people usually recover spontaneously, given enough time, the depression ratings of patients who are given no active treatment usually drop over time.  Thus, even when depressed people who volunteer for research studies are given an inactive placebo pill or placebo (inactive) light, their depression ratings decrease after 8 to 16 weeks.  In a clinical trial, volunteers are randomly given either the active treatment (such as bright light or an antidepressant drug) or the inactive placebo for contrast.  To calculate how much of the average patient’s recovery was due to the active treatment, I considered the baseline depression rating (e.g., the HDRS score) as the 100% reference, and then the depression ratings at the end of the trial were computed for active treatment and for control treatment as a percentage of this baseline.  Then, the net benefit of the active treatment was computed as the percentage reduction of depression ratings with active treatment minus the percentage reduction of depression with the control treatment.  

Our first studies of bright light tested only one single hour of bright light to treat hospitalized patients with nonseasonal depressions.   The patients were awakened to receive the light treatment from 2 hours to 1 hour before the patient's usual planned time of awakening, so that these patients experienced 1 or 2 hours of wake therapy at the end of the night.  We now call getting patients up early “wake therapy,” because to call this helpful treatment “sleep deprivation” gives the wrong impression.  Most patients were drug free, but some were taking antidepressants during light treatment.  As compared to a control hour of light placebo, which presumably produced the same sleep curtailment, the bright light reduced mood ratings about 12%.  As will be explained, a 12% net gain as compared to placebo is similar to benefits achieved by antidepressant medications after weeks of treatment, so it was remarkable that such substantial benefit could be obtained with one single hour of bright light.

In an extension of our initial studies, 25 drug-free patients were treated with bright light each day for one week, compared to 26 patients treated with a dim-light placebo.    Depression ratings were 18% lower after bright light than after placebo, a benefit which was statistically significant.  These and other data suggested that one-week treatment produced more benefit than only one hour of bright light.  More recently, another one-week study of unmedicated inpatients observed a somewhat larger net advantage of 24.2%, which was likewise statistically significant.    Even a study reporting no statistically significant benefit achieved a 12.2% net advantage in HDRS ratings of the bright-light treated group  , so that the failure to achieve statistical significance could have been partly due to an insufficient numbers of subjects.  These studies of drug-free depressives were consistent in demonstrating advantages of bright light treatment.

Two European studies were important because they examined effects of bright light (as compared to a dim-light placebo) in patients who were also receiving antidepressant medications.  In both of these studies, the net relative advantage of bright light over dim light was 27%.   Since the medication-only groups also did well in these combination studies, the additional improvement gained by the light-treated patients was especially impressive.  A third European study also demonstrated substantial benefits of bright light compared with placebo among patients simultaneously receiving antidepressants.    Not only do these studies show that bright light improves the antidepressant response of patients who are receiving antidepressant medications, but they leave the impression that the benefit of bright light may be greater when patients are also receiving antidepressants.

A new light triple treatment for nonseasonal depression has recently been tested.    Ironically, this new light treatment was developed by Dr. Neumeister working under Dr. Kasper at the University of Vienna--the same place where Sigmund Freud trained so many years ago.  The Vienna psychiatrists treated patients in the hospital with serious nonseasonal depressions who were being treated with antidepressant drugs but had not yet responded.  At the start, the doctors in Vienna awakened these patients in the middle of the first night and kept them awake for the rest of that night, while starting bright light treatment and also continuing with antidepressant drug treatment.  About 70% of these patients felt dramatically better on the day after they had been awakened early, and they continued to feel better.  About 35% of their depressive symptoms were relieved immediately.  It had been previously known in Europe that such awakenings often relieved depression on the day of the early wakening, but the patients usually had relapsed almost completely the next day.  Because of this relapse, few doctors in America thought that wake therapy was really very useful.  If adding bright light can prevent the relapse, we have a new way to relieve the symptoms of severe depression in one day.  There is nothing like it.  This excellent response to bright light combined with wake therapy and antidepressants has now been reproduced by other studies at European hospitals.   Our group has also reproduced this effect in a small study of outpatients, who awakened themselves at 2 AM in their own homes.    Unfortunately, the controlled studies of the triple therapy have not yet extended beyond 2 weeks of bright light. 

Two studies have shown that bright light may be useful for depressed elderly in nursing homes.  

In addition to these studies, there has been one controlled study showing that 10,000 lux produced more benefit than 2,500 lux, when treatment was only 30 min. per day.    This study in effect confirms that bright light is an active treatment.  Also, two studies have found shorter hospitalization times needed when depressed patients were in bright rooms than in rooms with darker windows.    A well-designed study showed that bright light was useful for premenstrual depression.  

There have been only three studies with good scientific methodology which failed to confirm significant benefits of light for nonseasonal depression.  Two of these studies indeed showed bright light benefit by every measure, but the statistical evidence was insufficient,  possibly because these studies needed more patients.  The other was an unlucky study, where by chance, the patients who received bright light had a poorer prognosis at the outset than the patients with whom they were compared. 

Unfortunately, there has been little controlled study of bright light effects beyond one or a few weeks, partly because the one-week results are so persuasive.  Longer-term studies are needed.  Anecdotal clinical experience suggests that the light benefit is maintained with long-term treatment or may augment.

In summary, there is now extensive evidence that bright light treatment reduces symptoms for both nonseasonal and seasonal depression.  For nonseasonal depression, the triple combination of bright light, ½ night’s wake therapy, and medication produces approximately a 35% reduction in symptoms in 1 week.  There is a need for longer-term studies.

5.A.  Is bright light only good in winter?

The earliest controlled report of bright light treatment for seasonal affective disorder (SAD) described remarkable 52% net benefits within one week, however, the bright-light-treated patients were given high expectations which were not matched in the placebo-treated group. Patients with high expectations often report improvement through the power of positive thinking, but benefits of positive thinking need to be separated scientifically from the effects of bright light.  A review of later studies showed diminishing net benefits compared to the initial report, even though several of the additional studies reviewed may also have failed to adequately equalize placebo expectations and even to balance placebo assignments.  Since clinical trials of SAD have often induced positive expectations with newspaper recruitment, since the volunteers cannot be literally blind to treatment, and since SAD (by definition) tends to remit spontaneously, the problems of biased expectations and placebo responses have been a continuing problem.

Recently, two clinical trials of SAD patients devoted extremely careful attention to controlling placebo effects in assessing light-treatment benefits.  One of these trials showed only a 6.4% net benefit of bright light after 5 weeks of treatment.   After 10-14 days, the second trial showed a 35% net benefit of morning light treatment as compared to a placebo dosage of negative ions, with a 31% net benefit of evening light.   Thus, the net benefit for 2-5 weeks of bright light treatment of SAD appears to be in the range of 6% to 35%, when expectations and randomization are carefully controlled.  These well-controlled results with winter depression do not appear superior to results with nonseasonal depression. 

A recent study indicated that bright light treatment of SAD works best when morning bright light is given so early that it requires waking up early (i.e., a bit of wake therapy is combined), however, it is not clear how long awakening so early can be maintained.

Even if we were sure that SAD patients responded better to light than other depressed patients, it would often hard to tell what type of depression a person has.  Follow-up of SAD patients by Dr. Rosenthal’s research group showed that the majority of patients first defined as SAD eventually displayed summer symptoms, if their illnesses continued to recur, and many needed antidepressant drugs.  Over follow-up, such patients may be perceived as nonseasonal according to accepted criteria.  Many SAD patients eventually require both bright light treatment and antidepressant medications outside the winter months.  Because of controversial criteria and complex clinical course, it may often be impossible to define whether a depressed patient does or does not have SAD, so restricting light therapy to SAD would not be practical. 

Even when seasonal pattern can be distinguished, there is no assurance that light treatment will work better than in nonseasonal patients, nor does seasonality exclude the usefulness of antidepressant medication in addition to bright light.

In conclusion, in my opinion, bright light is useful for people who are depressed, whether or not we think that they have SAD or winter depression.  It is likely that for both nonseasonal and SAD patients, a combination of bright light, antidepressant drugs, and wake therapy is advisable.

5.B.  A perspective on antidepressant medication

To have perspective on the results of bright light treatment, it is useful to understand the benefits which result from antidepressant drugs.  Hundreds or perhaps thousands of controlled trials of antidepressant medications have been reported, with a great preponderance of evidence that medicated patients improve somewhat faster than comparison patients given placebo.  Placebo means an inactive, dummy treatment.  Nevertheless, many physicians suppose that the benefits of antidepressant drugs are greater than such trials actually demonstrate. 

Because antidepressant medications may require 6 to 16 weeks to achieve substantial benefit, patients given placebo (dummy pills) for the same interval often display spontaneous remission.  They get better by themselves, perhaps helped by hope and by encouragement from the researchers.  The alleviation of symptoms attained during antidepressant drug treatment in most studies is due more to this spontaneous remission than to medication benefits.  The benefits of antidepressant drugs are only clearly understood when the percentage remission of symptoms achieved with placebo is subtracted from the remission accompanying medication.

Only recently have overall and unbiased assessments of antidepressant drug effects  become available.  A compendium of antidepressant drug trials in thousands of patients reported to the U.S. Food and Drug Administration provided a general summary of antidepressant drug effects.   The advantage of this report over previous compendia was the inclusion of rather unsuccessful studies which the pharmaceutical manufacturers were obliged to report, though they may not have wished to see such results published.  In these studies, the 8-week net benefit of antidepressant drugs on the Hamilton Depression Rating Scale was only 8-12% better than the result with placebo. Another analysis of some of the same data found that the placebo groups had improved 58% percent at the last measurement and the drug-treated groups had improved 69% percent, a drug-related benefit of only 11%.   An analysis of fluoxetine results reached similar estimates, even though some of the authors were employees of the makers of Prozac.   Incidentally, this analysis found only 7 clinical trials comparing fluoxetine to placebo, whereas we have at least 15 comparing bright light to placebo in nonseasonal depression.  A comprehensive meta-analysis using response criteria obtained rather similar results and also demonstrated that there has been a bias to publish more successful results and to leave less successful results unpublished. While these meta-analyses certainly demonstrate that antidepressant drugs have significant benefits for nonseasonal depression, the size of the antidepressant net benefit after 8 weeks or more (approximately 8-19%) certainly does not appear superior to the 1-week benefit of bright light treatment (approximately 12-35%). 

It is not wise to emphasize comparison of antidepressant drug benefits with bright light benefits, because there have been no direct randomizing comparisons of the two treatment approaches.  Although the benefits of bright light might be greater and are certainly more rapid than benefits of medication, much more is known about the long-term benefits of antidepressant drug treatment, which has been much more extensively studied.  Moreover, there really is no reason to inquire whether one treatment of depression is better than the other.  The important point is that bright light and antidepressant drugs are best used in combination, probably also combining at least a half-night of wake therapy.

Continued in Chapter 6

Brighten Your Life, in all its formats, including this eBook, Copyright ©1997-2002 by Daniel F. Kripke, M.D., all rights reserved.