Despite the evidence that bipolar misdiagnosis has not decreased in the past decade, some clinicians and researchers express concern about possible overdiagnosis of bipolar disorder. This concern is especially voiced in relation to discussion of broadening the definition of the bipolar spectrum.
Differential diagnosis of depression
Primary: Et ology not definitive
Secondary: Clear psychosocial stress, clear medical illness
FIGURE 1-3. How the polarity-based approach to bipolar disorder in DSM-IV-TR can lead to misdiagnosis.
*A diagnosis of exclusion.
Source. From Goodwin FK, Ghaemi SN: "An Introduction and History of Affective Disorders," in Oxford Textbook of Psychiatry, Vol 1. Edited by Gelder MG, Lopez-Ibor JJ Jr, Andreasen NC. Oxford, England, Oxford University Press, 2000, pp 677-680.
In contrast to the extensive literature described above that supports the notion of continued underdiagnosis of bipolar disorder, there is a much more limited body of literature providing empirical evidence of overdiagnosis of bipolar disorder. A MEDLINE search of "overdiagno-sis" and "misdiagnosis" for bipolar disorder identifies only three reports of possible overdiagnoses of bipolar disorder, two of which are case reports in letters to the editor. In the only published study (Krasa and Tolbert 1994), of 53 adolescents admitted with a clinical diagnosis of bipolar disorder, 72% met DSM-III-R criteria, suggesting that 28% were mistakenly diagnosed with the condition. As noted below, however, it is controversial whether DSM criteria for bipolar disorder, developed through study of adults, are equally applicable in pediatric populations. In the cases reported as possible overdiagnoses, one is not easily convinced that those patients indeed do not have bipolar disorder. For instance, it appears that the main reason why the author of one letter to the editor thought that the patient did not have bipolar disorder is due to lack of response to divalproex and a good response to antide-pressants (Hutto 2001). Of course, medication response is quite nonspecific and, though informative, is not useful as the sole source of data for diagnostic validity.
As noted throughout this chapter, multiple sources of data are needed to establish diagnostic validity. Other studies, albeit with a broader approach to bipolar diagnosis than in DSM-IV-TR (i.e., excluding the duration requirements), report much higher rates of underdiag-
nosis of bipolar disorder in children. In one study of 82 children (mean age 10.6 years) diagnosed with DSM-IV-TR criteria but excluding the duration requirements, 90% had not been previously diagnosed with bipolar disorder (Faedda et al. 2004a).
In searching for possible unpublished data regarding the diagnosis of bipolar disorder, we identified a poster presented at a bipolar conference in which 70 adult inpatients with clinical diagnoses of bipolar disorder were rediagnosed using a structured clinical interview for DSM-IV-TR (SCID) in preparation for a research study (Berns et al. 2003). Only 78% were diagnosed, on the basis of DSM-IV-TR criteria, with any form of bipolar disorder. Interestingly, the researchers also observed that some patients hospitalized with nonbipolar diagnoses appeared likely to have bipolar disorder. In seven cases where such underdiagno-sis was suspected, five indeed met bipolar disorder criteria with a SCID interview. Hence the authors found evidence of both overdiagnosis and underdiagnosis. The latter study partly confirms the earlier study in adolescence, suggesting that some patients are mistakenly diagnosed with bipolar disorder. If these estimates are correct, we might conclude that about 25% of hospitalized patients diagnosed with bipolar disorder may not have the condition, according to current DSM-IV-TR criteria. On the other hand, about 40% of patients with bipolar disorder, when DSM-IV-TR criteria are used, are initially misdiagnosed with other conditions. Thus underdiagnosis still seems to be more of a problem than overdiagnosis.
Of course, besides the issue of under- or overdiagnosis using our current criteria, there is the larger issue of whether current DSM-IV-TR criteria for bipolar disorder are valid. This is the problem of the bipolar spectrum whose solution will require better clinical and epidemiologi-cal research, along with better nosologically oriented genetic and neu-robiological research.
The question of diagnosis of bipolar depression in children is both quite complex and quite simple. To start with the simple aspect of the problem, bipolar disorder, as noted above, has an earlier age at onset than unipolar depression. As previously described, data from a study of children with mean age of 12.3 experiencing depression suggests that about one-half develop bipolar disorder by age 21 (Geller et al. 2001). Thus, the question of bipolar disorder in children is simply the question of depression: if children are depressed, the likelihood of bipolar disorder is high, perhaps about 50%. If a depressed child also has a family history of bipolar disorder, then the likelihood of bipolar disorder is even higher.
Therefore, in depressed children, clinicians should have a high index of suspicion that the illness is bipolar disorder, even if manic or hy-pomanic episodes have not been demonstrated yet.
The complex part of the problem is whether DSM-IV-TR criteria for mania, developed in adults, are applicable to children. At one level, it would seem logical that adult criteria do not apply in the same way to children. In order to establish the childhood criteria for bipolar disorder, research in children is required, with follow-up into adulthood. Such research has begun, and the developing evidence seems to fall into two categories: In one line of findings, it has been suggested that bipolar disorder in children is best established with the presence of elation, grandiosity, decreased need for sleep, racing thoughts, and hypersexu-ality (Geller et al. 2002). While most clinicians would agree with that definition, the more controversial question is whether other children might have manic presentations without elation and grandiosity, but rather primarily with irritability of mood and aggressive behavior (Wozniak et al. 1995). Clinical experience, dating back to Kraepelin (Trede et al. 2005), and recent research with adults (Cassidy and Carroll 2001) suggest that irritability and aggressive behavior are indeed important components of manic episodes in many patients; furthermore, it appears that euphoric mania represents a minority of manic episodes in adults, with mixed and irritable moods predominating (Cassidy and Carroll 2001). In fact, children may be especially sensitive to developing serotonin reuptake inhibitor (SRI)-induced mania, perhaps more so than adults (Baldessarini et al. 2005). Hence, it would not be surprising if manic episodes in children might also present primarily with irritability rather than euphoria. However, unless a few other manic criteria are present, the DSM-IV-TR definition would not be currently met, leading to the current controversy. Follow-up of such children into adulthood will provide the best evidence of whether such presentations in children indeed represent a childhood variation on bipolar disorder.
Attention-Deficit/Hyperactivity Disorder in Children and Adults
In children, a major diagnostic issue is differentiating attention-deficit/ hyperactivity disorder (ADHD) and bipolar disorder, which is a challenge largely because of the overlapping symptoms of distractibility and hyperactivity. In one study, about 90% of children who met DSM-III-R criteria for bipolar disorder also met criteria for ADHD. In contrast, only about 25% of those who presented with ADHD met criteria for bipolar disorder (Wozniak et al. 1995). Thus, it appears that many children are diagnosable with ADHD alone and seem to respond well to stimulants, while some children who appear to have ADHD may in fact have bipolar disorder. This observation is supported by research indicating that about 9.5% of adults with bipolar disorder can be retrospectively diagnosable with ADHD in childhood (Nierenberg et al. 2005). The use of stimulants, while not diagnostic, can be informative in the setting of other possible manic symptoms and a family history of bipolar disorder. In such children, stimulants are frequently ineffective, are limited in efficacy, or can trigger manic episodes associated with marked irritability and aggression. In one study of 82 children (mean age 10.6 years), 18% appeared to develop mania with amphetamines, though the rate of mania with other antidepressants was higher (44%) (Faedda et al. 2004b). One study found that children with bipolar disorder who were treated with stimulants had an earlier age at onset to their bipolar illness than children with bipolar disorder who had never received stimulants (DelBello et al. 2001). Although this outcome might have been influenced by other factors, one possibility is that stimulants speed the onset of bipolar illness, essentially worsening the course of the illness. Neurodevelopmental risks of stimulant use in long-term treatment of young animals have also been reported (Bolanos et al. 2003) and need to be weighed against the limited benefits in children with risk factors for bipolar disorder (Volkow and Insel 2003).
Therefore, we would generally recommend caution in the diagnosis of ADHD in children and especially in the use of stimulants for treatment of this disorder. In the first nationwide, community-based survey of ADHD diagnoses in the United States (Centers for Disease Control and Prevention 2005), it was found that 7.8% of children ages 4-17 were diagnosed with ADHD in 2003. ADHD was diagnosed in males much more often than in females (11.0% vs. 4.4%), and diagnosed in children ages 9-17 twice as often as in children ages 4-8 (9.7% vs. 4.1%). Children were also more likely to receive the diagnosis if they were white, insured, and below the federal poverty level. About half of those who received the diagnosis of ADHD were given stimulants (4.3%); again, even more so males than females (6.2% vs. 2.4%). The evidence for over-diagnosis of ADHD presented in other studies is somewhat supported by the finding of notable regional differences in diagnoses. The diagnosis of ADHD was more likely in states with the highest poverty levels (highest diagnosis rates of 10%-11% were in Alabama, Louisiana, and West Virginia) compared to wealthier states (lowest diagnosis rates of 5%-5.5% were in Colorado and California). Since there is obviously no biological reason why living in Alabama should predispose one to ADHD more than living in Colorado, social and economic factors lead ing to diagnosis and treatment appear relevant.
Despite the limitations in diagnosing and treating ADHD in children, the evidence for the validity of the diagnosis and the utility of treatment, in those appropriately diagnosed, is much more extensive than in adults. Yet the diagnosis of adult ADHD has become quite popular, coinciding with the marketing of the first medication for adult ADHD, Strattera (atomoxitene), by the pharmaceutical industry in 2002. Although there is some research on the potential validity of adult ADHD, that research is much more limited than childhood ADHD and awaits replication by independent research groups.
In the past, it was generally considered that ADHD did not persist into adulthood. Despite the recent surge of interest in adult ADHD, a recent review supporting the notion of adult ADHD only cited two studies (Wilens et al. 2004). One was a 1985 study of 63 children followed for 15 years that reported that a minority of the children continued to have symptoms into young adulthood (Weiss et al. 1985). Another study involved a 6-year follow-up of adolescents and reported improvement in hyperactivity but some persistence of attentional deficits (Achenbach et al. 1998). Compared with the century-old extensive literature on bipolar disorder (Trede et al. 2005), these data on the course of illness that support the validity of the concept of adult ADHD appear far from convincing, suggesting even more caution in diagnosing and treating adult ADHD.
It is worth noting that adult patients often seeking amphetamine stimulants for ADHD may be engaging in substance abuse. In addition to obtaining an adequate history of past substance use, clinicians may need to discuss with all patients the fact that improvement in attention in response to amphetamine stimulant treatment does not validate a diagnosis of ADHD. As noted earlier, treatment response is one of only four diagnostic validators, and it is the weakest and most nonspecific of them. This is especially the case when one is faced with a drug of abuse. The reason that amphetamines are abused is because they improve attention in normal people, irrespective of any diagnosis (Miller et al. 1989). Thus, such improvement is not diagnostically relevant.
Another topic of heated controversy is whether bipolar disorder is over-diagnosed at the expense of personality disorders, such as borderline personality disorder (Akiskal 2004; Birnbaum 2004; Magill 2004; Paris 2004; Smith et al. 2004). While this argument has been made based on case reports (Bolton and Gunderson 1996), no empirical evidence has been published to that effect. On the other hand, there is extensive re search that borderline personality disorder is overdiagnosed in patients with abnormal and/or unstable moods, with frequent resolution of "borderline" traits when the mood episode resolves with antidepressants or mood stabilizers (Frankenburg and Zanarini 2002; Preston et al. 2004). Thus, the empirical evidence again seems to suggest a need to abstain from diagnosing borderline personality disorder in individuals with major depressive and manic or hypomanic episodes. Instead, it makes sense clinically to treat their mood episodes. If those episodes resolve and the individual continues to have borderline personality disorder traits in the euthymic state, then a true comorbidity may exist.
On the other hand, there certainly are individuals without hy-pomanic or manic episodes but who have long-standing and unchanging borderline personality traits. Those persons are best diagnosed and treated primarily with psychotherapies for borderline personality disorder. Mood lability in that setting, not meeting hypomania or mania syn-dromal definitions, is often a feature of the borderline personality condition. Although some data suggest that mood lability in general may be a predictor of bipolar disorder (Benazzi and Akiskal 2005a), it seems reasonable to emphasize the classic approach of requiring other symptoms to meet syndromal definitions of mood episodes (whether in the narrow DSM-IV-TR model or the broader bipolar spectrum disorder model).
Our perspective is that we need to recognize overlap of the two syndromes where such overlap exists, while also working hard diagnosti-cally to identify those who predominantly have bipolar disorder in some cases or predominantly borderline personality disorder in other cases. This approach is supported by some of the limited, empirical evidence on the borderline/bipolar issue. In patients with bipolar disorder type II, Benazzi found that there was extensive overlap of borderline traits of affective instability, but not with borderline criteria for impul-sivity (Benazzi 2005). Similar results were found in a previous study by another group (Henry et al. 2001a). Thus, mood lability is not a useful diagnostic feature to distinguish the two conditions, but borderlinetype impulsive behaviors (such as cutting) are much less common in bipolar disorder. When clinicians simply pay attention to standard definitions of mania and hypomania, most patients with bipolar disorder, even type II, can reasonably be distinguished from borderline personality disorder (Benazzi 2000).
Again the key is to get beyond simply symptom overlap, such as mood lability, and use the other diagnostic validators, especially course and family history, as described throughout this chapter (Paris 2004).
The association of potential, increased risk of suicide in some persons treated with antidepressants raises the question of whether those depressed patients might have instead experienced mixed states. If a broader definition of mixed states, as described earlier, is valid, then many individuals today diagnosed with major depressive episodes instead might be experiencing depressive mixed states. Antidepressants can cause mixed episodes (Dilsaver and Swann 1995) or worsen them, and the limited data available indicate that antidepressants do not improve depressive symptoms in mixed episodes (Prien et al. 1984). Mixed episodes are known to be associated with a high degree of sui-cidality, perhaps even more so than anergic pure depression (Dilsaver et al. 1994).
Thus, it is not unlikely that many of the depressed children who entered clinical trials with SRI antidepressants actually had bipolar disorder, given the data that about half will likely develop manic or hypomanic episodes (Geller et al. 2001). They would then likely be predisposed to antidepressant-induced mixed episodes, with concomitant suicidality (Berk and Dodd 2005). As discussed previously, similar risks would likely hold for young adults in their 20s (Goldberg et al. 2001).
A related issue is the fact that many clinicians appear to misdiag-nose mixed episodes as pure depressive episodes, and then provide treatment with antidepressants. Data supporting the use of antidepres-sants for depressive mixed episodes are very limited (Brown et al. 1994; Rihmer et al. 1998), but this has not seemed to deter clinicians from this practice. On the other hand, anticonvulsants and antipsychotic agents are likely much more effective (Bowden et al. 1994; Sachs et al. 2002) and need to be more seriously considered for depressive mixed states to the extent that they can be shown to be distinct from pure major depression. In sum, it could be that the problem of antidepressant-induced suicid-ality may really be the tip of an iceberg of inappropriately treated bipolar depression and depressive mixed states.
Was this article helpful?
Bipolar is a condition that wreaks havoc on those that it affects. If you suffer from Bipolar, chances are that your family suffers right with you. No matter if you are that family member trying to learn to cope or you are the person that has been diagnosed, there is hope out there.