TCPR: What is nightmare disorder?
Dr. Krakow: This is a sleep disorder in DSM-5 characterized by repeated, distressing nightmares. Some patients awaken from the dreams and others do not, but either way they usually remember the dreams, sometimes vividly. These nightmares often involve themes of threat, fear, and other distressing emotions. An important element is that the nightmares cause impairment during the day—either from the dream content or the sleep disruption that goes along with it.

TCPR: Does it include sleep terrors?
Dr. Krakow: No. In sleep terrors (which used to be called “night terrors”) the patient doesn’t fully awaken, and they don’t remember the dream material very well if at all.

TCPR: How is nightmare disorder different from PTSD?
Dr. Krakow: The way the DSM lays it out, ­nightmare disorder can’t be diagnosed if the dreams occur during PTSD and reenact themes from a traumatic event. In practice, though, posttraumatic nightmares and nightmare disorder respond to the same therapy—imagery rehearsal therapy (IRT)—and in both disorders the patient’s distress improves after eradication of nightmares. That’s true in other psychiatric disorders as well where nightmare disorder can be diagnosed, like depression, anxiety, bipolar disorder, schizophrenia, and borderline personality disorder (Gieselmann A et al, J Sleep Res 2019;28(4):e12820). This is important to recognize, because treatment of nightmares decreases these other psychiatric symptoms, too.

TCPR: Interesting. Can you explain this further?
Dr. Krakow: In the past, it was thought that you had to treat the underlying disorder for the nightmares to get better, but two psychiatrists, my mentors Robert Kellner and Joseph Neidhardt, showed the reverse. Depression and anxiety symptoms got better when nightmares improved with a therapy they developed in 1987 called imagery rehearsal therapy (IRT). When we expanded our research program in the early 1990s, Dr. Kellner encouraged me to try it in PTSD. We did a study on 168 women who were survivors of sexual assault. It was a randomized controlled trial where they received three sessions of group treatment of IRT (7 total hours) or wait-list control. IRT improved nightmares and decreased PTSD symptoms as well. The effect sizes were in the large range for nightmares, sleep quality, and PTSD symptoms (Krakow B et al, JAMA 2001;286(5):537–545).

TCPR: Was the work more challenging in PTSD?
Dr. Krakow: We had to do more education and cognitive restructuring, as many of these patients had come to believe their nightmares were caused exclusively by their PTSD and would not improve unless the PTSD was treated. Indeed, they would debate this point with us in sessions and often declare, “My therapists said I needed to treat the trauma before my nightmares could resolve.” The irony was not lost on them—seeking treatment for nightmares after years of psychotherapy, such as EMDR, exposure therapy, and medication, and most of all “talk therapy.”

TCPR: Psychiatrists are pretty good at recognizing nightmares in PTSD, but what should we look for in other disorders?
Dr. Krakow: Nightmares are very common in patients with depression, anxiety, and insomnia if you ask about them. They are more common in people who’ve been through lots of stress and have difficulty coping with stress. People don’t bring them up because they don’t know there is treatment for bad or disturbing dreams, or they presume that drugs or psychoanalysis are the only options, even though high-level evidence for these traditional approaches is sparse.

TCPR: Nightmares are also part of our normal experience, so how do we know when they need treatment?
Dr. Krakow: The answer is frequency and distress. Here’s a pearl: If a patient has a bad dream at least once a week, the chance that it is bothering them is extraordinarily high. Maybe it’s making them afraid to go to sleep and they have insomnia, or their sleep is fitful and they feel crappy in the morning. Or the nightmare comes back to them during the day and makes them feel sad or fearful. Ask the patient, “Do you think about your nightmare content at all during the day?” If they say “Yes,” they would probably benefit from an intervention.

TCPR: What else do patients with nightmare disorder complain of?
Dr. Krakow: They have higher levels of anxiety, depression, somatization, and hostility (Kellner R et al, Am J Psychiatry 1992; 149(5):659–663). Their impairment in those respects is about equal to what you’d see for your average patient with anxiety or depression who comes to an outpatient psychiatric practice, so it is significant. But to reiterate, it is extremely rare for a patient to enter therapy requesting treatment for nightmares.

TCPR: Is this a vicious cycle where a negative mood sets them up for nightmares, those nightmares worsen their mood further, and it spirals from there?
Dr. Krakow: I think that cycle is part of it. I think almost all nightmare patients have something in their lives that is very difficult to cope with, to emotionally process, even if it is not a DSM trauma. But I think the larger part, from the vantage point of sleep medicine, is that disturbing dreams often seem to take on a life of their own to become an independent disorder that fuels or worsens insomnia.

TCPR: Is the idea that the nightmare was originally functional as a way to process the stressful event?
Dr. Krakow: Yes, it’s likely that the dream started out that way. We don’t know all the functions of dreams, but it seems dreams and REM sleep help people learn, remember, and process emotional events. In fact, we often wonder if REM sleep—where dreaming occurs most memorably—is a kind of psychotherapy while you’re sleeping. So the question is, why do some trauma survivors suffer nightmares chronically? Most people have nightmares in the first month after a traumatic event, but for 80%–90% of them those nightmares go away 2–3 months later.

TCPR: IRT is based on a behavioral model of nightmares. Tell us about that.
Dr. Krakow: The idea is that nightmares become habitual, like any learned process. This is how we explain it to patients: “Maybe these nightmares keep happening because a habit forms in your brain where, when you are keyed into something during the daytime, your body says, ‘Well, that was an anxiety-producing event, so I’m just going to have another nightmare tonight.’” And that is actually how we introduce IRT—as an idea that nightmares are a learned behavior. This concept can be a breath of fresh air for some or a deal-breaker for others. Unfortunately, conventional wisdom about nightmares has led most sufferers to imagine there cannot possibly be a learned-behavior model of nightmares. As such, the more entrenched they are in this view, the more education and restricting we need to work through.

TCPR: What happens in the therapy?
Dr. Krakow: The main technique is that patients take a nightmare and change it in some way—in any way they want. Then they’ll rehearse the new dream in their imagination during the daytime.

TCPR: Can you walk us through the steps?
Dr. Krakow: We start with the education piece on nightmares as a learned behavior that originally served a purpose. Next we want the patient to become comfortable with imagery. It is much like guided imagery. We’ll have them imagine a scene and engage their full senses in it: “Picture in your mind’s eye how to drive from your house to your favorite restaurant,” or “Picture going from the meadow down to the beach and listening to waves.” This part may take longer in PTSD, because these patients are prone to flashbacks and intrusive memories when they allow their imagination to run far afield.

TCPR: What do you do when distressing thoughts intrude?
Dr. Krakow: You check for that. You want patients to be in control, so you have them stop and get grounded. Specifically, they stop the session, open their eyes, and take some relaxed breaths. Then they go back to the imagery when they are ready. When they become more accomplished with imagery, we want to encourage them to acknowledge a distressing image and then choose to move away from it and on to a new and pleasant image. Some patients learn this more advanced strategy very quickly, while others might need months and lots of coaching or counseling to achieve this aim.

TCPR: So first they practice imagery. What’s next?
Dr. Krakow: That usually takes 2 sessions, so by the third session they are comfortable with imagery and have practiced it on their own. Then we say, “Okay, I want you to select a dream and write it down. Next I want you to change that dream ‘any way you wish,’ following Dr. Neidhardt’s original instruction. And then I want you to rehearse the new dream you created in your mind’s eye.”

TCPR: Are they supposed to start with a repetitive nightmare?
Dr. Krakow: No. It can be any nightmare. Theoretically, they don’t even have to choose a ­nightmare. We think the imagination process is where the action is. What matters is that they are engaging their imagination actively, instead of it taking hold of them and terrorizing them every night.

TCPR: Are you desensitizing them to the nightmare?
Dr. Krakow: No. This is not desensitization. This is new learning. In fact, I only have them write down the nightmare once, when they are first learning the therapy. I would never ask anyone to write a bad dream twice, because when they write down a nightmare it could overstimulate them. We don’t want them to rehearse the nightmare.

TCPR: You ask patients to “change the nightmare.” Do they ever have trouble coming up with new material?
Dr. Krakow: They may say, “Well, change what?” The best response we’ve found is very open: “Change the nightmare any way you wish.” We want the ideas to come from the patient, not us. Once we adopted this instruction, we saw wild changes. One woman presented with a repetitive nightmare where she was chased to the edge of a cliff. What was the new ending she created? “I jumped off the cliff.” Her nightmares cleared up fairly soon thereafter.

TCPR: So it’s more important for them to take control and choose the ending than to create a happy ending?
Dr. Krakow: Exactly. It’s empowering, but remember it’s not about directing them to choose an ending. Changing it any way you wish means literally that. When most people hear that instruction, a lightbulb goes off and they run with it. There are a handful of patients who resist and say, “What do you mean, ‘change it’? I can’t change my dream. That’s what happened to me.” I say, “Well, then let’s not pick that traumatic replay dream.” And we may end up using a neutral dream or a fantasy if their nightmares trigger too much anxiety. In the end, we want them to have a “new dream,” and that is what they rehearse outside of session. And this new dream need not be static. Its imagery may evolve as they practice it. That’s good—we want to see change, because part of the patient’s problem is that their dream life and their imagination have become stuck.

TCPR: Is it important to rehearse the dream just before bed?
Dr. Krakow: Not at all.

TCPR: Really?
Dr. Krakow: We don’t tell patients when to rehearse the dream or for how long. If we do, they’ll likely run into obstacles like “I didn’t have time to do it.” In the trials, the average patient rehearsed for less than 5 minutes every other day; many only rehearsed 1 minute every other day. We want to get them back in control of their imagination.

TCPR: What are some problems patients run into when they try this therapy?
Dr. Krakow: Some patients have trouble imagining anything. They say, “My screen is black.” It may be they have so much anxiety they are just suppressing everything. I tell them, “What you need to do is just let your imagery session go longer. Just practice longer.” But these patients might need more guided imagery sessions under the direction of a therapist.

TCPR: How widespread is IRT use?
Dr. Krakow: IRT has been declared the number one nonpharmacologic treatment for chronic nightmares for about a decade, and 99% of all research studies on IRT show decreases in nightmares and distress. Its use became fairly widespread in the US military after the war in Iraq, when soldiers were returning home complaining of PTSD, sleep disorders, and nightmares. We conducted 10 trainings at leading bases in the US in 2013 and 2014. In fact, I’ve got one coming up at Fort Campbell in August. But well before these trainings, VA medical centers frequently contacted me for trainings and use of our audio workbook, Turning Nightmares Into Dreams. European countries are very active in using IRT, in part because there are so many dream science research centers in Germany, the Netherlands, Switzerland, and England.

TCPR: Are there patients who are not appropriate for this therapy?
Dr. Krakow: I’d want to know that the patient can cope with the imagery work, so I may not do it with someone who is severely depressed or suicidal. I’d be careful in patients with unstable PTSD who are having flashbacks, dissociation, or panic attacks on a regular basis. Those cases are better managed by a therapist with expertise in PTSD. IRT has been most widely promulgated in sleep clinics in the US by clinicians with little training in psychotherapy, although mental health professionals have applied it successfully to more severe cases.

TCPR: Can you summarize the research benefits of IRT?
Dr. Krakow: IRT decreased nightmares in virtually all populations studied. After about 2 weeks, patients have fewer nightmares or less intense nightmares (usually both). Most of the benefits kick in between 2 weeks and 2 months. Their sleep also improves, and they have less depression, anxiety, somatization, and hostility (Krakow B et al, Behav Res Ther 1995;33(7):837–843). And in PTSD, there is decreased PTSD symptom severity; by comparison, the improvement in PTSD appears to be about the same magnitude of sertraline’s impact on PTSD.

TCPR: Do patients continue the rehearsal technique on their own after they recover?
Dr. Krakow: Many do, particularly if the nightmares flare up again. But they also start using their ­daytime imagery to solve problems, which is exactly what most of us do in everyday life. For example, when you misplace something, you try to find it in your mind’s eye: “Where was I when I lost it?” That’s a good use of imagery, and patients have reported an increase in this activity after using IRT.

TCPR: Is it fair to say that you’re teaching nightmare sufferers how to daydream?
Dr. Krakow: Absolutely. One of our patients came back a month after IRT and said, “It was amazing. I was in this conflict with my boss and I pictured (‘daydreamed’) having a better conversation with him. And then when I went and saw my boss, I had a better conversation with him.” And she was ecstatic.

TCPR: Anything else we should know?
Dr. Krakow: Yes. Since starting this work, we were stunned to learn that nightmares are also a sign of sleep apnea in a very high proportion of cases, whether they occur during PTSD or nightmare disorder. The rate of sleep apnea is very high in PTSD, up to 80%. These patients don’t look like they have sleep apnea—they are often young and thin—but you don’t want to miss this diagnosis.

TCPR: Thank you, Dr. Krakow.

Editor’s note: To learn more about imagery rehearsal therapy, visit www.barrykrakowmd.com.