The basic principle of treating bipolar illness is simple: Maximize mood stabilizers; avoid antidepressants. The corollary is: Antipsychotics are not mood stabilizers.
So treatment means taking true mood stabilizers, defined as drugs which have been proven to prevent depressive or manic episodes. There are only four such agents: lithium, valproic acid (divalproex, Depakote), carbamazepine (Tegretol), and lamotrigine (Lamictal). All patients with bipolar illness should be on at least one of these four agents.
Antidepressants should be avoided. In the best-designed and largest studies, they are proven ineffective in acute bipolar depression, meaning for immediate short-term benefit of depressive symptoms in bipolar illness. They are proven ineffective in maintenance prevention of new episodes in bipolar illness, meaning they do not prevent future depressive episodes. Besides their general inefficacy, they are also harmful in some people. According to the best estimates, about one quarter or more of persons who have bipolar illness get immediately manic (or hypomanic) on antidepressants. This is a short-term effect, which may or may not be harmful in itself, but mania (or hypomania) is often followed by depression. This leads to the second main harm with antidepressants in bipolar illness: they cause more and more mood episodes (both depressive and manic/hypomanic) over time. In about one-quarter of persons with bipolar illness, this long-term worsening occurs and leads to a rapid-cycling course (meaning four or more mood episodes per year, as opposed to the natural history of untreated bipolar illness, which leads to about one mood episode per year). In other words, paradoxically, antidepressants not only cause mania, they worsen depression for the long-term, in about one-quarter of persons with bipolar illness
Antipsychotics are in between mood stabilizers and antidepressants. They don’t harm the illness, like antidepressants; but they are not effective by themselves for prevention of mood episodes, like true mood stabilizers. Thus, they can be used in some persons, and can be very helpful, but only as add-on treatments to true mood stabilizers, not in place of true mood stabilizers. There is a common misconception that some antipsychotics are mood stabilizers, based on FDA indications for maintenance treatment, leading to pharmaceutical marketing, and the support of some academic experts. But these claims are based on long-term maintenance studies which have serious scientific problems, rendering them questionably valid. This problem is discussed here.
In sum, give one of the four main mood stabilizers to all persons with bipolar illness (of any type: type I versus type II is not a reason to ignore this rule, as discussed here). Give antipsychotics, only as add-on treatments, to some persons with bipolar illness who do not improve sufficiently with the main mood stabilizers. Avoid antidepressants in the vast majority of persons with bipolar illness, under almost all circumstances.
As with all rules, there are exceptions, but being an exception means being used in a minority of persons, not the majority. The use of antidepressants in bipolar illness in some exceptional cases is discussed here. Unfortunately, instead of being the least frequently used class of medication in bipolar illness, antidepressants are the most frequently used class. That practice, which reflects poor treatment, even though it is the standard of care, should be reversed, based on the more detailed evidence and rationale provided here.