TCPR: What are we getting wrong in the way we treat bipolar disorder?

Dr. Gitlin: There’s been a marked decline in the use of lithium over the past 20–30 years. There are reasons not to use it, and certainly the need for lab monitoring is not trivial, but lithium is arguably the gold-standard drug, so we ought to rethink this trend. Lithium is not first line for everybody, but something is wrong when so many patients with bipolar disorder have tried multiple anticonvulsants and antipsychotics without ever having a chance on lithium. The other problem is antidepressants. There’s a disparity between what bipolar experts think and what clinicians do here. Most guidelines say, “Don’t use antidepressants first line,” but in practice they are used much more liberally.

TCPR: What is your view on antidepressants in bipolar?

Dr. Gitlin: My personal view is that antidepressants have been given a much worse rap than they deserve. There are now sufficient showing as monotherapy in patients with bipolar II disorders—and in bipolar I disorder if the patient is on a mood stabilizer (Cheniaux E et al, Exp Opin Drug Safety 2019;8(10):893–913). But they should not be used in bipolar I as monotherapy, and they are not first line in bipolar II depression. We have other options like lamotrigine, lithium, lumateperone, lurasidone, and quetiapine for bipolar depression.

TCPR: What are the risks with antidepressants?

Dr. Gitlin: The risks are mood switching to mania, mixed states, or hypomania, and—over the long term—rapid cycling. The consequences are greater in patients with bipolar I, who are much more likely to switch all the way into the more destructive state of mania. In bipolar II, we see more switches into hypomania, and the overall switch rate is about half for bipolar II versus bipolar I. In bipolar I disorder, half the switches are into mania and half are into hypomania, whereas in bipolar II, 90% of the switches are into hypomania—which is, by definition, a much less destructive state.

TCPR: Do you have a preferred antidepressant in bipolar disorder?

Dr. Gitlin: This varies a lot by patient. In the research, bupropion and SSRIs have the lowest switch rates, and then venlafaxine and—for the highest switch rates—the tricyclics. Studies of MAOIs are few, but their risk appears somewhere between that of the SSRIs and the tricyclics. But there are many antidepressants where we don’t have reliable information: mirtazapine, trazodone, selegiline (an MAOI Emsam patch), and the other SNRIs besides venlafaxine (Barbuti M et al, Eur Neuropsychopharmacol 2023;73:1–15).

TCPR: You mentioned that lithium use is on the decline. That trend has accelerated with the approval of antipsychotics in bipolar disorder.

Dr. Gitlin: Yes, the first atypical to gain approval in acute mania was olanzapine in 2000, but antipsychotics were used for mania long before that. In the 1960s, it was lithium versus chlorpromazine (Thorazine), and when I read those papers, I realized you could have substituted the word olanzapine for chlorpromazine and everything they found would still be correct. Things have not changed much.

TCPR: What are the meaningful differences between lithium and antipsychotics in bipolar disorder?

Dr. Gitlin: Antipsychotics work faster and are better at controlling behavioral agitation. Lithium is slower, but it has more of a normalizing feeling to patients (Bergamelli E et al, CNS Drugs 2021;35(11):1275–1287). Lithium is not sedating the way chlorpromazine and olanzapine are, so patients don’t feel like they “have a blanket over their head” when on lithium. But it often takes more than a few weeks to recover from mania on lithium, and it’s not as good at controlling the agitation or insomnia.

TCPR: That may explain why antipsychotics are preferred on inpatient units. They lead to faster improvement.

Dr. Gitlin: Yes, and some of that is due to their sedating effects. For the nursing team—or the family if the patient is at home—calming the agitation is an important goal. Sometimes patients want that too, but not over the long term, and sedation is not actually necessary for recovery from mania. Less sedating antipsychotics like risperidone and aripiprazole work just as well as sedating ones like olanzapine. Another strategy is to use a benzo along with a nonsedating antipsychotic or mood stabilizer, and then taper the benzo off as the mania improves.

TCPR: Is there a preferred benzo in bipolar disorder?

Dr. Gitlin: There are no clinical data suggesting one is clearly better than another, but in practice, I prefer clonazepam for pragmatic reasons: It only has to be dosed once or twice a day.

TCPR: So antipsychotics work faster in mania, but how do you know what will work best over the long term?

Dr. Gitlin: We know the most about who does well on lithium long term. About 20% of people with bipolar disorder have a lightbulb response to lithium, and the more they fit the textbook case of bipolar disorder, the more likely they are to respond to it. That means they have discrete episodes of classic-looking mania and depression that are separated by clear periods of normal “euthymic” mood. They lack mixed features and rapid cycling. Another predictor is that the depressions tend to come on after the manias in a mania-depression-well interval (MDI) pattern, as opposed to the opposite: depression-mania-well interval (DMI) (Hui TP et al, Acta Psychiatr Scand 2019;140(2):94–115).

TCPR: What about comorbidities?

Dr. Gitlin: Some say that comorbidities like anxiety or substance use disorders predict a poorer response to lithium, but I think they just predict a poorer response to almost anything. There is evidence, though, that valproic acid (Depakote) has a favorable response for those with comorbid alcohol use disorders (Le Fauve CE, Evid Based Ment Health 2005;8(3):79).

TCPR: What works better for patients with mixed features?

Dr. Gitlin: The antipsychotics and anticonvulsants, of which valproate is the anticonvulsant with the most research in mixed features. In rapid cycling, though, it is not as clear. We used to think lithium was less effective in rapid cycling, but Joe Calabrese did a large, 20-month, randomized comparison of lithium and valproate and found that both had similar outcomes and neither worked particularly well (Calabrese JR et al, Am J Psychiatry 2005;162(11):2152–2161). It looks like rapid cycling is a bad patch of the illness that nothing works well in, but the good news is that this is temporary. Most cases resolve within the first few years, so rapid cycling doesn’t define the individual.

TCPR: Do you know of any other markers for antipsychotic or anticonvulsant response?

Dr. Gitlin: No, we don’t have any good data. We tend to use lamotrigine when depressions predominate in bipolar disorder, as they often do in bipolar II.

TCPR: What have we learned about lithium in recent years?

Dr. Gitlin: It looks like we’ve been targeting blood levels that were higher than necessary. For maintenance treatment, the target is now 0.6–0.8 mEq/L, but it used to be thought of as 0.8–1.0 mEq/L (Nolen WA et al, Bipolar Disord 2019;21(5):394–409). Lower levels mean higher adherence and fewer side effects. Another exciting finding is that lithium is neuroprotective, which is ironic considering many patients complain of cognitive dulling on it. That side effect usually improves by lowering the dose, but over the long term, lithium may prevent dementia (see The Carlat Psychiatry Report, September 2022). Lithium also protects the ends of DNA—the telomeres—which are involved in cellular death and aging.

TCPR: Any new risks with lithium?

Dr. Gitlin: This is not a new side effect, but there is greater awareness of hyperparathyroidism on lithium, which can lead to osteoporosis and renal stones if not corrected. Most patients are asymptomatic, or may complain of fatigue or cognitive fuzziness, so you have to detect it through labs. At a minimum, that means monitoring calcium on routine lithium labs, and if it is elevated, I’d check the parathyroid hormone (PTH) and refer to endocrinology. Another update is not a risk, but there’s been some controversy around lithium’s ability to prevent suicide. A lot of studies suggest lithium prevents suicide, and not just because it’s a good mood stabilizer—we see this effect in unipolar mood disorders as well, and it is partly independent of lithium’s mood benefits (Smith KA and Cipriani A, Bipolar Disord 2017;19(7):575–586).

TCPR: What called that into question?

Dr. Gitlin: A VA trial randomized patients with mostly unipolar mood disorders to lithium or placebo, but it was halted for futility after researchers found no difference in the first 519 patients they enrolled, even though one in four of them had suicide-related events (Katz IR et al, JAMA Psychiatry 2022;79(1):24–32). But there were problems with this study—for example, only half of the sample had lithium levels above 0.5 mEq/L.

TCPR: Are there any medical complications that would make you stop lithium?

Dr. Gitlin: That always has to be balanced with the psychiatric risks. Even with renal disease, like when the eGFR (estimated glomerular filtration rate) goes below 60 mL/min—which is the general cutoff for chronic kidney disease—we may have to keep the patient on lithium. Sure, you’d want to consider other options and get the lithium level as low as possible, but if the alternative is frequent mania, hospitalizations, or suicide, they may need to stay on it (Salgado ME et al, Int J Bipolar Disord 2014;2(1):12).

TCPR: Is long-term prevention with a mood stabilizer always necessary in bipolar disorder?

Dr. Gitlin: For bipolar I, the answer is generally yes. For bipolar II, the risk/benefit ratio is different because—although recurrence of mood episodes is high—patients are not going to have the kinds of manias that can destroy their lives. Some practice guidelines outside the US say you can consider taking patients with bipolar I off a mood stabilizer after just one manic episode as long as it wasn’t severe—like floridly psychotic or requiring hospitalization. But the relapse rates are very high, so I recommend long-term prevention even after one manic episode. The problem is that the first episode usually occurs in the teenage or early adult years when people have an inflated feeling of being in control of their lives. They think that the first episode was due to a life event, and they often won’t agree to long-term treatment until the second or third episode.

TCPR: Are psychiatrists overdiagnosing or underdiagnosing bipolar disorder?

Dr. Gitlin: For bipolar I patients, we do a reasonably good job. Nobody is going to miss florid mania, although we may not pick up on it if it’s buried in the patient’s history and we don’t get old records or speak to the family. Bipolar II is easier to miss because hypomania can be a very subtle state and can be positive—and not feel like psychopathology. That’s where a close friend or relative’s observations are critical. If I’m wondering whether we’re dealing with hypomania or another form of affective lability like borderline personality disorder, I will ask to speak with someone who knows the patient well. First, I reassure the patient that I’m not going to reveal any information and invite them to be there while we talk. I find outside observers are critical for two diagnoses: bipolar II and ADHD. Yes, it’s critical, but what really confirms the diagnosis is long-term observation.

TCPR: How long do you need before you’re fairly confident in the diagnosis?

Dr. Gitlin: It depends. If the patient is a good historian, it may be clear in the first interview. On the other hand, I’ve treated patients with “unipolar depression” for two to three years before realizing they had hypomania. One diagnosis to pay close attention to is depression with psychotic features. That often evolves into bipolar disorder, schizoaffective disorder, or schizophrenia (Wood AJ et al, Front Psychiatry 2021;12:734272).

TCPR: You sound excited about your work. How do you keep from getting burned out?

Dr. Gitlin: Variety. I see patients, I teach, I do some administrative stuff, I write, and so I’m never doing one thing all the time. Also, I get to see people recover. If I were doing the same thing all day and only saw patients with chronic, treatment-resistant disorders, things might be different.

TCPR: Thank you for your time, Dr. Gitlin.