Comparing Youth With Bipolar Disorder to Non-Bipolar Youth Referred for Bipolar Disorder

Abstract

Objective

Bipolar disorder (BD) diagnoses require episodes of hypomania and mania as well as depressive episodes. Given the overlap of BD symptoms with symptoms of other psychiatric conditions among youth, misdiagnosis is common. This topic was examined in a large sample of youth clinically referred for BD.

Method

Participants were 394 clinically referred youths ages 13 to 20 years, including 255 with confirmed BD and 139 for whom BD was not confirmed (non-BD). Participants and their parent/guardian completed a semistructured diagnostic interview and dimensional scales. Demographic and clinical variables were compared between BD and non-BD groups. Following correction for multiple comparisons, significant variables associated with BD diagnosis (p < .05) in univariate analyses were evaluated in multivariable analyses.

Results

Compared with the BD group (n = 255), the non-BD group (n = 139) had significantly lower current mania symptom severity, family history of hypomania/mania, current lithium treatment, and lifetime bulimia nervosa, whereas most severe past global functioning was higher and current oppositional defiant disorder was more common in the non-BD group compared with the BD group. Use of second-generation antipsychotics was high in both groups. Common reasons for not diagnosing BD in the non-BD group included not meeting duration criteria for a hypomanic/manic episode and manic-like symptoms being better explained by other psychiatric disorders.

Conclusion

Youth with and without BD did not differ in the vast majority of clinical variables examined. Frequent use of second-generation antipsychotics in non-BD youth may relate to characterization of overlapping comorbidity symptoms as manic symptoms. Both groups have complex presentations, necessitating psychosocial and pharmacological treatments.

Diversity & Inclusion Statement

We worked to ensure sex and gender balance in the recruitment of human participants. We worked to ensure race, ethnic, and/or other types of diversity in the recruitment of human participants. One or more of the authors of this paper self-identifies as a member of one or more historically underrepresented racial and/or ethnic groups in science. We actively worked to promote sex and gender balance in our author group. We actively worked to promote inclusion of historically underrepresented racial and/or ethnic groups in science in our author group. While citing references scientifically relevant for this work, we also actively worked to promote sex and gender balance in our reference list. While citing references scientifically relevant for this work, we also actively worked to promote inclusion of historically underrepresented racial and/or ethnic groups in science in our reference list. The author list of this paper includes contributors from the location and/or community where the research was conducted who participated in the data collection, design, analysis, and/or interpretation of the work.

Plain language summary

A total of 394 youth, aged 13-20 years, were recruited from a subspecialty adolescent bipolar disorder clinic at a teaching hospital in Toronto, Canada, over a 12-year period. Participants were clinically referred for assessment and/or treatment of bipolar disorder. A bipolar spectrum disorder diagnosis was confirmed for 255 of the 394 youth. The authors examined a broad range of demographic, clinical, and familial characteristics, and over 90% did not yield significant between-group differences. The most common reasons for not confirming a bipolar disorder diagnosis were insufficient and/or fleeting manic symptoms and manic-like symptoms being better explained by other psychiatric disorders. Overall, both groups had complex presentations, emphasizing careful assessment and the need for psychosocial and pharmacological treatments in both groups.

Bipolar disorder (BD) is a chronic mood disorder that is associated with significant burden of mood symptoms and is accompanied by functional impairment, multiple comorbid psychiatric and physical comorbidities,¹ and high risk of suicide.² Onset of BD in childhood or adolescence is reported by 32% to 65% of adults with BD.³ Such cases of youth-onset BD, relative to adult-onset BD, are known to be more severe as reflected by increased mood symptom burden, high rates of comorbidities, and increased suicidality.4, 5, 6, 7, 8

Although early diagnosis and treatment of BD are imperative, misdiagnosis is common. In adults with BD, unipolar depression is the most frequent misdiagnosis, followed by schizophrenia.⁹ Other conditions that have overlapping symptoms/presentations include anxiety disorders, primary psychotic disorders, alcohol or substance use disorders, and personality disorders. Differential diagnosis in youth can be particularly challenging, given the number of non-BD psychiatric conditions that have overlapping symptoms with BD.¹⁰ In addition to episodes of hypomania and mania that are the most distinguishing characteristics of BD, youth with BD often exhibit mood lability, irritability, impulsivity, and aggressive behavior.^11,12 Similar symptoms are observed in youth with attention-deficit/hyperactivity disorder (ADHD), oppositional defiant disorder (ODD), and severe mood dysregulation/disruptive mood dysregulation disorder.^10,13

In the past 2 decades, evidence has been reported of significantly increased outpatient and inpatient diagnoses of BD in the United States, with convergent findings also reported in Denmark, Germany, and Portugal.14, 15, 16, 17 Leibenluft et al. have conducted a series of clinical and neurobiological studies comparing youth with BD (type I and II) to youth with chronic irritability who do not experience discrete episodes of hypomania/mania.^13,18,19 Youth with chronic irritability were selected as a comparison group for youth with rigorously diagnosed BD due to findings suggesting that the rapid increase in the use of the BD diagnosis in young people was driven in part by the application of this diagnosis to youth with severe chronic irritability.^13,19 However, less is known regarding youth who are thought to have BD for reasons other than chronic irritability.

In adults, there is a low rate of confirmation of BD diagnosis by psychiatrists following primary care referral for assessment of BD, a pattern that has not improved with time.²⁰ When evaluating youth who present with symptoms that overlap between BD and other psychiatric conditions (eg, irritability, hyperactivity, difficulty concentrating), it is particularly important to determine whether they occur, or worsen, at the same time as the change in mood.^21,22 Establishing the presence of episodes of hypomania/mania is fundamental to diagnosing all bipolar spectrum disorders.²²

Although scores on certain self-report and particularly parent-report questionnaires (eg, Child Behavior Checklist, Mood Disorders Questionnaire parent version, General Behavior Inventory) differ on average between youth with BD compared with youth with other conditions such as major depressive disorder (MDD) and ODD, these questionnaires cannot be used as a substitute for a comprehensive diagnostic assessment.²³ Ultimately, there is no short-cut to confirming an accurate diagnosis of BD other than the application of rigorous diagnostic assessment that confirms the presence of a sufficient number of symptoms, their duration, and their functional impact and that ensures episodicity.²²

The present study is based on a large, single-site sample of youth referred to a specialized clinic and research institute for assessment and treatment of BD from 2009 to 2020. The goals of the study are twofold. First, we sought to compare youth with vs without BD in terms of demographic, clinical, and familial characteristics. Although the specific comparison group in the current study (youth referred for BD and found not to have BD) differs from prior studies, we generated hypotheses based on prior studies that included MDD and severe mood dysregulation comparison groups. We hypothesized that compared to youth with BD, youth without BD would have lower rates of family history of BD²⁴ and lower rates of anxiety disorders,²⁵ disruptive behavioral disorders (ODD, conduct disorder [CD]),²⁶ and suicidal thoughts and behaviors. Second, we sought to examine the reasons that youth referred for BD were deemed to not have BD based on structured diagnostic interview.

Method

Participants

Participants were 394 clinically referred youths ages 13 to 20 years, including 255 with confirmed BD and 139 for whom BD was not confirmed (non-BD). Participants were recruited from a subspecialty adolescent BD clinic at a tertiary academic health sciences center over a 12-year period from 2009 to 2020. Approximately 50% of referrals were from other child and adolescent psychiatrists, and the remainder were from pediatricians and primary care physicians. All referral information was reviewed by the senior author (B.I.G.) to confirm that sufficient information was available to warrant a subspecialized BD assessment. Other referrals were deferred to the general outpatient youth psychiatry assessment pool. Informed consent was obtained from participants and at least 1 parent/guardian in the study, which was approved by the institutional research ethics board.

Measures

All assessments were conducted by researchers with at minimum a bachelor’s or master’s degree in a health science field who were trained and supervised by the senior author (B.I.G.). Youth were interviewed first, followed by interviews with at least 1 parent/guardian. The Schedule for Affective Disorders and Schizophrenia for School-Age Children–Present and Lifetime Version (K-SADS-PL) https://www.pediatricbipolar.pitt.edu/clinical-services/clinical-tools), a semistructured interview, was used to determine current and lifetime diagnoses of BD and other psychiatric disorders.²⁷ Diagnoses, derived from interviews with youth and their parents in combination with available medical records, were made according to DSM-IV criteria for data collected from 2009 to June 2018 and DSM-5 criteria for data collected from July 2018 to 2020. Diagnosis of BD not otherwise specified (BD-NOS) was based on operationalized criteria from the Course and Outcome of Bipolar Illness in Youth (COBY) study for duration of symptoms (minimum 4 hours/day) and number of hypomanic days (minimum 4 in lifetime),²⁸ but retaining the symptom count requirements of 3 when elation was the primary symptom and 4 when irritability was the primary symptom.

Diagnostic decisions were made based on all available information, with clinical judgment applied when conflicting information was provided during the evaluation process. If there were differing symptoms reported by the youth vs the parent, both parties were brought together to reconcile these differences. Diagnoses were then confirmed during a consensus meeting with licensed child and adolescent psychiatrists following completion of the K-SADS-PL interview. BD onset was defined as the age at which the youth first experienced mania or hypomania or when the operationalized criteria for BD-NOS were first met. The K-SADS-PL was used to obtain information regarding demographic data and clinical characteristics (eg, lifetime medication history and psychiatric hospitalization). A Safety Assessment Form was also administered by the interviewer to obtain information about lifetime suicide attempts, nonsuicidal self-injury, suicidal ideation, and legal history. In addition to the K-SADS-PL, the extended K-SADS Depression Rating Scale²⁹ and extended K-SADS Mania Rating Scale were used. Mood symptoms, including the most severe past episode and current episode (worst week in the past month), were also assessed using the Depression Rating Scale and Mania Rating Scale.

The Family History Screen³⁰ was used to gather information about psychiatric status and history of first- and second-degree relatives from the youth and parent(s)/guardian(s). Socioeconomic status was determined based on the 4-factor Hollingshead Scale.³¹ The Children’s Global Assessment Scale (CGAS)³² was used to assess the youth’s global functioning, including current, most severe past, and past year highest level episode. The Children’s Affective Lability Scale (CALS)³³ was used to measure affective regulation and is based on self-report and parent/guardian-report. The Life Problems Inventory (LPI)³⁴ self-report was used to identify dimensions related to impulsiveness, emotional dysregulation, identity confusion, and interpersonal difficulties.

Study data were initially collected in paper format and subsequently using REDCap electronic data capture tools.³⁵ REDCap is a secure, web-based software platform designed to support data capture for research studies, providing an intuitive interface for validated data capture, audit trails for tracking data manipulation and export procedures, automated export procedures for seamless data downloads to common statistical packages, and procedures for data integration and interoperability with external sources.

Statistical Analysis

Bivariate analyses were undertaken using t tests or Pearson χ² tests for continuous and categorical variables, respectively. A false discovery rate approach was used to correct for multiple comparisons. Variables with significance at p ≤ .05 threshold after correcting for multiple comparisons were then entered in a forward stepwise logistic regression analysis controlling for age. Lifetime (most severe past) mania was not included in this regression as it was highly correlated with BD diagnosis. SPSS default values of inclusion (p < .05) and exclusion (p > .1) were used for the regression. All statistical analyses were conducted using IBM SPSS Version 25 (IBM Corp., Armonk, NY).

Results

Demographic Data

Demographic and clinical variables across groups are presented in Table 1. Of 394 youths referred for BD/suspected BD, 64.7% (n = 255) were confirmed as having BD, and 35.3% (n = 139) were not. Mean age was significantly higher in the BD group vs the non-BD group (t = 3.28, p = .001). There were no significant between-group differences in self-reported sex, socioeconomic status, race, or living status.

Table 1.

Demographic and Clinical Variables for Youth With Suspected Bipolar Disorder (BD)

	Groups				Statistics
	BD (n = 255)		Non-BD (n = 139)		Test statistica	p	pFDR	Effect sizeb
	Mean	(SD)	Mean	(SD)	t
Demographic characteristics
Age, y	16.9	(1.5)	16.3	(1.5)	3.28	.001	.008	0.4
	n	(%)	n	(%)	χ2
Sex, female	179	(70.2)	99	(67.9)	0.05	.83	.95	0.01
	Mean	(SD)	Mean	(SD)	U
SES status, rankc	4.1	(1.0)	3.9	(1.1)	16,316.0	.21	.50	0.19
	n	(%)	n	(%)	χ2
Race, White	199	(78.0)	108	(77.7)	0.006	.94	1.00	0.004
Living status	141	(55.3)	73	(52.5)	0.25	.62	.83	0.03
Clinical characteristics
Bipolar subtype
BD-I	71	(27.8)	—		—	—	—	—
BD-II	85	(33.3)	—		—	—	—	—
BD-NOS	99	(38.8)	—		—	—	—	—
	Mean	(SD)	Mean	(SD)
Age of BD onset, y	14.8	(2.6)	—		—	—	—	—
Age of MSP episode	15.3	(2.3)	—		—	—	—	—
	n	(%)	n	(%)	χ2
Current comorbid diagnosis
MDD	—		82	(59.0)	—	—	—	—
ADHD	77	(30.2)	50	(36.0)	1.36	.24	.52	0.06
ODD	69	(27.1)	61	(43.9)	11.5	.001	.008	0.17
CD	14	(5.5)	17	(12.2)	5.64	.02	.08	0.12
Generalized anxiety disorder	151	(59.2)	90	(64.7)	1.16	.28	.57	0.05
Social anxiety disorder	92	(36.1)	64	(46.0)	3.76	.05	.17	0.10
Agoraphobia	25	(9.8)	17	(12.2)	0.56	.46	.69	0.04
Panic disorder	58	(22.7)	36	(25.9)	0.49	.48	.71	0.04
PTSD	20	(7.8)	11	(7.9)	0.001	.98	1.00	0.001
OCD	36	(14.1)	13	(9.4)	1.88	.17	.41	0.07
Anorexia nervosa	5	(2.0)	2	(1.4)	0.14	.71	.87	0.02
Bulimia nervosa	8	(3.1)	1	(0.7)	2.36	.13	.34	0.08
Eating disorder NOS	16	(6.3)	7	(5.0)	0.25	.62	.83	0.03
Lifetime comorbid diagnosis
MDD	208	(81.6)	108	(77.7)	0.85	.36	.65	0.05
ADHD	110	(43.1)	63	(45.3)	0.18	.68	.85	0.02
ODD	76	(29.8)	63	(45.3)	9.49	.002	.01	0.16
CD	16	(6.3)	20	(14.4)	7.13	.008	.04	0.14
Any anxiety disorderd	206	(80.8)	114	(82.0)	0.09	.77	.90	0.02
	Mean	(SD)	Mean	(SD)	χ2
Number of anxiety disorders	1.9	(1.5)	2.0	(1.4)	6.79	.34	.65	0.13
	n	(%)	n	(%)	χ2
Generalized anxiety disorder	160	(62.7)	91	(65.5)	0.29	.59	.82	0.03
Social anxiety disorder	97	(38.0)	67	(48.2)	3.82	.05	.17	0.10
Agoraphobia	26	(10.2)	18	(12.9)	0.69	.41	.68	0.04
Panic disorder	66	(25.9)	39	(28.1)	0.22	.64	.83	0.02
PTSD	23	(9.0)	12	(8.6)	0.02	.90	.98	0.006
OCD	46	(18.0)	16	(11.5)	2.89	.09	.26	0.09
Substance use disordere	89	(34.9)	43	(30.9)	0.64	.43	.68	0.04
Anorexia nervosa	6	(2.4)	1	(0.7)	1.38	.24	.52	0.06
Bulimia nervosa	20	(7.8)	1	(0.7)	9.05	.003	.02	0.15
Eating disorder NOS	32	(12.5)	19	(13.7)	0.1	.75	.89	0.02
Other lifetime characteristics
Suicide attempt	59	(23.1)	22	(15.8)	2.08	.15	.38	0.08
Nonsuicidal self-injury	140	(54.9)	73	(52.5)	0.04	.85	.95	0.01
Suicidal ideation	160	(62.7)	74	(53.2)	1.42	.23	.52	0.06
Smoking	115	(45.1)	65	(46.8)	0.1	.75	.89	0.02
Sexual abuse	19	(7.5)	10	(7.2)	0.02	.89	.98	0.007
Physical abuse	21	(8.2)	15	(10.8)	0.63	.43	.68	0.04
Legal history	92	(36.1)	49	(35.3)	0.04	.84	.95	0.01
	Mean	(SD)	Mean	(SD)	U
Mood symptom severity and functioning
Current—past month
Mania	17.7	(13.1)	7.7	(8.4)	9,870.5	<.001	.008	—
Depression	20.0	(12.5)	19.6	(10.4)	17,172.0	.69	.85	0.04
Global functioning	53.1	(11.4)	56.5	(11.0)	14,084.5	.001	.008	−0.3
					t
Lifetime—MSP episode
Mania	29.7	(9.3)	12.0	(9.2)	18.11	<.001	.008	1.91
					U
Depression	30.8	(10.8)	28.0	(9.4)	14,392.5	.003	.02	0.28
Global functioning	41.9	(8.9)	46.0	(9.8)	13,852.0	.001	.008	−0.44
Global functioning—highest past year	62.1	(12.2)	62.6	(12.6)	16,475.0	.38	.66	−0.04
Dimensional traits
Child Affective Lability Scale
Adolescent	31.2	(17.0)	32.2	(18.3)	15,409.5	.67	.85	−0.06
Parent	23.7	(16.1)	25.4	(16.6)	15,061.5	.41	.68	−0.10
Life Problems Inventory
Impulsivity	32.5	(14.8)	33.3	(14.6)	15,865.5	.55	.79	−0.05
Emotional dysregulation	38.0	(16.4)	40.6	(16.2)	14,686.0	.08	.24	−0.16
Identity confusion	43.7	(17.1)	46.6	(17.3)	14,868.5	.12	.33	−0.16
Interpersonal problems	37.1	(16.0)	39.6	(16.6)	14,907.0	.13	.34	−0.15
Family psychiatric historyf
	n	(%)	n	(%)	χ2
Mania/hypomania	114	(44.7)	29	(20.9)	22.12	<.001	.008	0.24
Major depressive episode	200	(78.4)	103	(74.1)	0.95	.33	.64	0.05
Suicidal ideation	110	(43.1)	52	(37.4)	1.22	.27	.57	0.06
Suicide attempt	75	(29.4)	35	(25.2)	0.84	.36	.65	0.0
Any anxiety disorderd	156	(61.2)	85	(61.2)	<0.001	1.0	1.00	<0.001
ADHD	88	(34.5)	44	(31.7)	0.33	.57	.81	0.03
CD	34	(13.3)	21	(15.1)	0.24	.63	.83	0.02
Substance use disordere	130	(51.0)	78	(56.1)	0.95	.33	.64	0.05
Psychosis	57	(22.4)	18	(12.9)	5.16	.02	.08	0.11
Schizophrenia	21	(8.2)	8	(5.8)	0.81	.37	.66	0.05
Current treatmentg
Any psychosocial treatment	212	(83.1)	127	(91.3)	4.55	.033	.12	0.11
SGA	73	(28.6)	23	(16.5)	7.12	.008	.04	0.13
Lithium	24	(9.4)	2	(1.4)	9.28	.002	.01	0.15
Valproate	1	(0.4)	0	(0.0)	0.55	.46	.69	0.04
Lamotrigine	9	(3.5)	3	(2.2)	0.57	.45	.69	0.04
SSRI antidepressants	27	(10.6)	27	(19.4)	5.94	.02	.08	0.12
Non-SSRI antidepressants	9	(3.5)	5	(3.6)	0.001	.97	1.00	0.002
Stimulants	16	(6.3)	16	(11.5)	3.31	.07	.22	0.09
Lifetime treatment historyg
SGA	142	(55.7)	54	(38.8)	10.20	.001	.008	0.16
Lithium	49	(19.2)	9	(6.5)	11.63	.001	.008	0.17
Valproate	17	(6.7)	2	(1.4)	5.36	.02	.08	0.12
Lamotrigine	15	(5.9)	8	(5.8)	0.003	.96	1.00	0.003
SSRI antidepressants	108	(42.4)	73	(52.5)	3.74	.05	.17	0.10
Non-SSRI antidepressants	35	(13.7)	19	(13.7)	<0.001	.99	1.00	0.001
Stimulants	51	(20.0)	39	(28.1)	3.31	.07	.22	0.09
Psychiatric hospitalization	123	(48.2)	40	(28.8)	14.04	<.001	.008	0.19

Note: Boldface indicates significant results. ADHD = attention-deficit/hyperactivity disorder; CD = conduct disorder; FDR = false discovery rate; MDD = major depressive disorder; MSP= most severe past; NOS = not otherwise specified; OCD = obsessive-compulsive disorder; ODD = oppositional defiant disorder; PTSD = posttraumatic stress disorder; SES = socioeconomic status; SGA = second-generation antipsychotic; SSRI = selective serotonin reuptake inhibitor.

^aHomogeneity of variance not met, Welch test used.

^bPartial η² or Cramer V.

^cHigher values for SES (Hollingshead-Redlich scale) indicate higher educational level and occupation.

^dAny anxiety disorder includes generalized anxiety disorder, social phobia, separation anxiety disorder, agoraphobia, panic disorder, and/or anxiety disorder NOS.

^eSubstance use disorder includes alcohol abuse/dependence, recreational drug abuse/dependence, and prescription medication abuse/dependence.

^fFirst- and/or second-degree relatives for BD, only second-degree relatives for non-BD.

^gNumbers do not sum to the sample size because participants could be taking more than 1 class of medication.

Diagnoses

In the BD group, 27.8% (n = 71) had BD type I (BD-I), 33.3% (n = 85) had BD type II (BD-II), and 38.8% (n = 99) had operationalized BD-NOS. The mean (SD) age of BD onset was 14.8 (2.6) years. There was no significant difference in the lifetime prevalence of major depressive episodes between the BD group (81.6%) and the non-BD group (77.7%). The non-BD group had a significantly higher prevalence of current and lifetime ODD (χ² = 11.5, p = .001; χ² = 9.5, p = .002) and CD (χ² = 5.6, p = .02; χ² = 7.1, p = .008) and a significantly lower prevalence of lifetime bulimia nervosa (χ² = 9.1, p = .003). Current CD was no longer significant after correction for multiple comparisons. There were no significant between-group differences for other comorbidities.

Due to the assessment of autism spectrum disorder (ASD) being initiated several years into recruitment of the current sample, this information was missing for 131 participants. As such, ASD is not included in the primary analytic model. Within the subset of youth for whom ASD was assessed, rates of ASD were similar among youth with BD (4/118, 3.4%) and non-BD youth (6/145, 4.1%) youth (χ² = 0.1, p = 0.75).

Clinical Characteristics

There were high rates of lifetime suicide attempts, nonsuicidal self-injury, and suicidal ideation in both groups, but there were no between-group differences. The BD group had higher current and most severe past mania symptom severity (U = 9,870.5, p < .001; t = 18.1, p < .001) and most severe past depressive symptom severity (U = 14,392.5, p = .003) as well as lower most severe past global functioning (U = 13,852.0, p = .001) and past month global functioning (U = 14,084.5, p = .001). There were no significant between-group differences in current depression severity.

Family Psychiatric History

Family history of hypomania/mania and psychosis was significantly higher in the BD group (χ² = 22.1, p < .001, χ² = 5.2, p = .02). Family history of psychosis was no longer significant after correction for multiple comparisons. There were no significant between-group differences in family psychiatric history of a major depressive episode, anxiety disorders, ADHD, CD, substance use disorder, schizophrenia, or suicidal ideation or attempts.

Treatment History

The BD group had greater current use of second-generation antipsychotics (SGAs) (χ² = 7.1, p = .008) and lithium (χ² = 9.3, p = .002) and lower current any psychosocial treatment (χ² = 4.6, p = .03) and use of selective serotonin reuptake inhibitor (SSRI) antidepressants (χ² = 5.9, p = .02). For lifetime treatment, SGAs (χ² = 10.2, p = .001), lithium (χ² = 11.6, p = .001), and valproate (χ² = 5.4, p = .02) were significantly more common in the BD group. The findings for current psychosocial treatment, SSRIs, and lifetime valproate treatment were not significant after correction for multiple comparisons. The lifetime use of antidepressants, stimulants, or lamotrigine did not differ between the groups. Finally, lifetime psychiatric hospitalization was significantly more common in the BD group (χ² = 14.0, p < .001).

Dimensional Self-Report and Parent-Report Measures

There were no significant group differences in youth- and parent-reported affective lability. There were also no between-group differences for the adolescent-reported LPI subscales.

Multivariable Logistic Regression

The multivariable regression model (including all variables in Table 1 with false discovery rate–corrected p ≤ .05), controlling for age, accounted for 42.9% of the variance (Table 2). The non-BD group had significantly lower current mania symptom severity (odds ratio [OR] = 0.91, 95% CI 0.89-0.93, χ² = 47.2, p < .001), family history of hypomania/mania (OR = 0.29, 95% CI 0.16-0.50, χ² = 18.9, p < .001), current lithium treatment (OR = 0.15, 95% CI 0.03-0.69, χ² = 5.9, p = .02), and lifetime bulimia nervosa (OR = 0.08, 95% CI 0.01-0.73, χ² = 5.1, p = .02). The non-BD group had significantly higher (ie, better) most severe past global functioning (OR = 1.04, 95% CI 1.01-1.07, χ² = 5.2, p = .02) and current ODD (OR = 2.85, 95% CI 1.62-5.00, χ² = 13.3, p < .001). Current global functioning, lifetime lithium treatment, lifetime SGA treatment, ODD, CD, and most severe past depressive episode severity were not significant.

Table 2.

Multivariable Analyses Examining Demographic and Clinical Variables of Non–Bipolar Disorder (BD) Diagnosis in Youth With Suspected BD

Variables	OR	Wald χ2	p	95% CI of OR
				Lower	Upper
Current diagnosis of ODD	2.849	13.261	<.001	1.622	5.004
Global functioning (most severe episode)	1.038	5.174	.023	1.005	1.072
Current mania rating score	0.913	47.203	<.001	0.890	0.937
Psychiatric hospitalization	0.586	2.973	.085	0.319	1.076
Family history of hypomania/mania	0.286	18.881	<.001	0.163	0.503
Current lithium treatment	0.146	5.887	.015	0.031	0.691
Lifetime diagnosis of bulimia nervosa	0.084	5.096	.024	0.010	0.722

Note: Variables at p ≤ .05 threshold after correcting for multiple comparisons were entered in a forward stepwise logistic regression analysis controlling for age except for lifetime mania rating score. ODD = oppositional defiant disorder; OR=odds ratio.

Reasons for Not Diagnosing BD

The reasons for not diagnosing BD in the non-BD group, which are not mutually exclusive, are presented in Table 3. Half of the non-BD group had manic-like symptoms that were better attributed to other psychiatric disorders (n = 71, 51.1%). One-fourth had sufficient number of manic symptoms, but had never reached the minimum threshold of 4 hours in a single day (n = 35, 25.2%). Another common reason was that the description of manic symptoms was deemed not reliable enough to be confident of BD diagnosis (n = 32, 23%). Although case-specific aspects that undermined reliability were not recorded, this determination was based on factors such as insufficient autobiographical detail (ie, description of episodes and/or their consequences), inconsistency of narrative (eg, provided different answers to the same questions when posed by research interviewer vs clinician), and plausibility (eg, parent described youth as acutely manic who had no difficulties throughout a school week, youth endorsed pronounced manic symptoms not noticeable to others). Less common reasons included having less than 4 cumulative lifetime days of hypomanic symptoms (n = 12, 8.6%) and having manic symptoms deemed secondary to psychotropic medication (n = 15, 10.8%) or substance use (n = 5, 3.6%).

Table 3.

Reasons Bipolar Disorder (BD) Diagnosis Was Not Given for Non-BD Group

Reasons	Number of respondents (n = 139)
	n	(%)
Symptoms cannot be differentiated from and/or are better explained by other psychiatric disorders (ie, ADHD, ODD, CD, MDD, ASD)	71	(51.1)
Diagnostic criteria for mania are insufficient	47	(33.8)
Never experienced manic symptoms for >4 h in 24-h period	35	(25.2)
Only 1-3 cumulative lifetime days meeting manic symptom criteria	12	(8.6)
Information obtained from the source is not reliable or sufficient to confirm diagnosis of BD	32	(23.0)
Manic symptoms occurred in association with psychotropic medication (stimulants, antidepressants)	15	(10.8)
Manic symptoms occurred only in the context of substance use	5	(3.6)

Note: Multiple responses were permitted. The 112 participants had one reason, 23 participants had two reasons, and 4 participants had three reasons. ADHD = attention-deficit/hyperactivity disorder; ASD = autism spectrum disorder; CD = conduct disorder; MDD = major depressive disorder; ODD = oppositional defiant disorder.

Discussion

This study focused on youth referred for assessment and/or treatment of BD, comparing participants with vs without confirmed BD. Among a broad array of demographic, clinical, and familial characteristics examined, greater than 90% did not yield significant between-group differences. Compared with the BD group, the non-BD group had significantly lower current mania symptom severity, family history of hypomania/mania, current lithium treatment, and lifetime bulimia nervosa, whereas most severe past global functioning and current ODD were significantly higher in the non-BD group compared with the BD group. The most common reasons for not diagnosing BD in the non-BD group were not meeting duration criteria for a hypomanic/manic episode, and manic-like symptoms being better explained by other psychiatric disorders. The current study adds to the existing literature by focusing on a unique comparison group of youth referred for, but found to not have, BD, and by including a description of reasons for not diagnosing BD.

Among all youth referred for BD, diagnosis of BD was confirmed for 64.6%. In previous studies, 27% to 73% of referred adults with suspected BD were diagnosed with BD.^20,36 In a prior youth study, the concordance rate of BD diagnoses by community providers and by clinicians at a mood disorders clinic was only 26%.³⁷ We speculate that the higher rate of BD diagnosis confirmation in the current study may be due to the use of prescreening referral materials before assessment. Although precise information on this topic is not available, we estimate that approximately 50% of referrals were declined due to lack of sufficient evidence from the referral information to suggest that a BD-focused assessment was justified. Based on this estimate, the rate of confirmed BD diagnoses among all patients referred for BD would be approximately 1 in 3, which aligns with the prior finding of 26%.³⁷

In terms of clinical characteristics, non-BD youth displayed a higher prevalence of ODD. There is prior evidence that comorbid disruptive behavior disorders are common among youth with BD and are predictors for subsequent BD.^38,39 In addition, there are multiple overlapping symptoms between BD and disruptive behavior disorders.^1,40 Taken together, the increased prevalence of ODD, but not other comorbidities, in the non-BD group suggests that present findings align with prior speculations that symptoms of disruptive behavior disorders are a common driver of false-positive diagnoses of BD²¹ and in turn excessive use of mood-stabilizing medication.⁴¹

The findings of lower manic and depressive symptom severity, greater functional impairment, and less family history of BD in the non-BD group are intuitive and align with prior findings. In accordance with our results, previous studies reported that youth with BD-I had more severe depressive symptoms and functional impairment than youth with MDD.^42,43 Furthermore, family history of BD is a leading risk factor for BD.⁴⁴ It is also notable that more than 80% of youth in the BD group had previously been diagnosed with MDD, and depressive symptoms more often preceded manic symptoms in these young people. In addition, greater than 40% of youth with K-SADS-PL–confirmed BD had been prescribed SSRI antidepressants.

This study found that both groups were comparable in terms of comorbid anxiety disorders and mood lability. While anxiety disorders^26,45 and mood lability⁴⁶ are precursors of BD, these precursors lack specificity and are associated with multiple subsequent psychiatric disorders.⁴⁷ There were also high rates of suicidal thoughts and behaviors, legal problems, and medication use, including SGAs, in both groups. Altogether, youth referred for BD, including those found not to have BD, have substantial burden of psychiatric symptoms and risk-related behaviors and require pharmacological and psychosocial treatment.

Although not the focus of this study, it is important to note that a previous publication based on a largely overlapping sample reported that youth with BD-NOS are generally similar to youth with BD-I and BD-II, including similar rates of suicidal ideation, suicide attempts, comorbidities other than anxiety, and family history of BD.⁴⁸ Most subtype-related differences were based on expected characteristics of BD-I (eg, manic symptom severity, psychiatric hospitalization). In addition, BD-II and BD-NOS had more depression, anxiety, and emotional dysregulation.⁴⁸

The common reasons that participants in the non-BD group were not diagnosed with BD reinforces the importance of abiding by previously described clinical principles.²² For example, overlapping symptoms of comorbid conditions should be counted toward hypomania/mania only if they meaningfully increase from baseline concurrent with other criterion A manic symptoms (ie, elation, increased energy) and/or specific symptoms of mania (eg, reduced need for sleep, grandiosity). Additionally, it is important to not diagnose BD when manic symptoms are fleeting and/or nonrecurrent. The COBY-operationalized criteria for BD-NOS (“other specified bipolar and related disorder” in DSM-5) provides a good benchmark for setting a minimum threshold for BD diagnoses, given evidence of clinical, familial, and genetic similarity with BD-I and BD-II.48, 49, 50 Finally, lack of a reliable history of manic symptoms was a common reason for not diagnosing BD. Efforts aimed at establishing a clear autobiographical epoch, reconciling discrepant reports by parents/guardians vs youth, and ensuring that described manic symptoms are noticeable to others and impact functioning can help to optimize diagnostic accuracy. While no single strategy is superior in all cases, there is evidence that parent report is of particular value when assessing BD in youth.²³

The results of this study should be interpreted in the context of the following limitations. First, the study focused on youth who were referred to a tertiary subspecialty clinic and was based on a sample with middle-to-high socioeconomic status and with limited ethnoracial diversity. As such, findings from this sample may not generalize to settings with different sampling frames. Similar studies in other settings and with diverse samples are warranted. Second, it is possible that participants in the non-BD group may subsequently develop BD. Third, this study was cross-sectional and retrospective. Future longitudinal studies are warranted to evaluate the temporality of comorbidities, treatments, and stressors in relation to the onset of BD. Fourth, due to the dense findings and data in the current study, this study did not examine potential differences in ODD symptomatology across the BD subtypes and non-BD group. However, this is an important topic, especially in the context of diagnosis of BD. Therefore, future studies including fine-grained detail regarding ODD dimensions are warranted. Fifth, the high rate of referrals from other child and adolescent psychiatrists and the prescreening of referrals likely biased the study toward increased similarity between the BD and non-BD groups. Present findings likely do not extend to the broader population of all youth who are referred to general youth outpatient settings with a query of BD. Nonetheless, in our view, it is precisely this almost BD group that is most likely to be potentially misdiagnosed with, and treated for, BD.

In summary, youth clinically referred for BD for whom BD diagnosis was not confirmed were similar to youth with confirmed BD on the vast majority of variables examined. The current study provides insights regarding factors that may contribute to false-positive diagnoses of BD and potentially unnecessary use of specific medications (eg, SGAs) and/or unnecessary avoidance of other medications (eg, SSRIs, stimulants). Both groups had a substantial burden of comorbidities, risk-related behaviors, and functional impairment despite high rates of pharmacological and psychosocial treatment. As such, continued research is warranted to identify, validate, and implement interventions that are needed to improve symptoms, safety, and functioning among youth in both groups.

CRediT authorship contribution statement

Jangho Park: Writing – original draft. Alysha A. Sultan: Writing – original draft, Writing – review & editing, Formal analysis. Aaron Silverman: Writing – review & editing. Eric A. Youngstrom: Writing – review & editing. Vanessa Rajamani: Writing – review & editing, Data curation. Mikaela K. Dimick: Writing – review & editing. Benjamin I. Goldstein: Writing – review & editing, Supervision, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization.