Study design and aim
The present study is a prospective cohort study investigating follow-up data on newly diagnosed patients with BD. The overall aim of the present study is to elucidate the clinical consequences of the modification to the current DSM-5 criterion A when focusing on newly diagnosed bipolar patients.
Setting of the study
The present study investigated data from the ongoing longitudinal Bipolar Illness Onset study (BIO), which aims to identify biomarkers for BD (8). Details on methods and procedures in the BIO-study are fully described elsewhere (8). The BIO-study protocol was approved by the Committee on Health Research Ethics of the Capital region of Denmark (protocol No. H-7-2014-007) and the Danish Data Protection Agency, Capital Region of Copenhagen (RHP-2015-023). All participants provided written informed consent. The BIO-study complied with the Declaration of Helsinki principles (Seoul, October 2008).
Participants
We included 373 patients with newly diagnosed BD. Recruitment of participants took place at the outpatient Copenhagen Affective Disorder Clinic from June 2015 to November 2019. The Clinic provides treatment for patients with newly diagnosed BD and receives patients from the Capital region of Denmark, an area covering 1,6 million and all the psychiatric centers in the Region (9). All patients referred to the clinic as newly diagnosed bipolar patients were routinely asked for inclusion in the BIO-study.
At baseline, the initial diagnostic assessment of participants was conducted by an experienced specialist in psychiatry using the Structures Clinical Interview for DSM-IV-TR Axis I Disorders (10), categorizing patients into bipolar disorder I (BD I) or bipolar disorder II (BD II). The diagnosis of BD was confirmed in a semi-structured research-based interview by MDs or MScs in Psychology PhD students using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) providing an ICD-10 diagnosis (3). Inclusion criteria were an ICD-10 diagnosis of BD or a single manic episode and age 15–70 years.
No attempts to balance the prevalence of bipolar subtypes were taken. Patients with BD not otherwise specified and patients with cyclothymia were not offered treatment in the Clinic and thus not included in the BIO-study. Exclusion criteria were organic BD secondary to brain injury.
Patients with BD received treatment as usual in the clinic without interference from study investigators.
Besides the clinical assessment at inclusion (baseline), patients were assessed during remitted depressive, manic or mixed phases and and once yearly for up to five years of follow up.
Procedures
At each assessment the current affective state of each participant was evaluated according to ICD-10 (manic, hypomanic, depressive, mixed episode, remission). The severity of manic and/or depressive symptoms were evaluated using the 11-item clinician administered Young Mania Rating Scale (YMRS) (11) and the 17-item Hamilton Depression Rating Scale (HAMD-17) (12), respectively, covering the precceding three days. Both rating scales have been extensively validated for assessing manic and depressive symptoms, respectively (11, 13). Patients were required to have a completed YMRS score at their first visit (baseline) to be included in the present study, as this was the basis of inclusion criteria.
Definitions
We used the same methodology and definitions as in our prior publication on patients with chronic BD (6). The individual ratings on item 1 (Mood), item 2 (Energy/Activity) and item 5 (Irritability) on the 11-item YMRS allowed an evaluation of the presence of fulfilled criterion A of hypomania/mania according to DSM definition (see table 1).
A score equal or above 2 on item 1 (Mood) was considered reflective of increased/elevated mood. A score equal or above 2 on item 2 (Energy/Activity) was considered reflective of increased motor activity or energy. A score equal or above 4 on item 5 (Irritability) was considered reflective of increased irritability (see table 1).
Thus, study definition of fulfilling DSM-IV criterion A for a hypomanic/manic visit was defined as item 1 (Mood) 2 and/or item 5 (Irritability) 4, presumably adequate to meet the DSM-IV definition of criterion A. Study definition of fulfilled DSM-5 criterion A for a hypomanic/manic visit was defined as item 1 (Mood) 2 and/or item 5 (Irritability) 4 AND item 2 2 (Energy/Activity), presumably adequate to meet the DSM-5 definition of criterion A (see table 1).
Depressive symptoms were defined as an HAMD-17 total score 15.
Statistical analyses
A total of 849 visits by 373 participants were analyzed. Data analyses were performed based on participants (373 patients) using logistic regression models and based on visits (849 visits) using linear mixed-effects models.
Descriptive data were presented as median and interquartile range (IQR) or count and percentage (%). A power of 80% and a significance level of alfa=.05 were chosen for all statistical analyses. Results from analyses were presented as odds ratios with 95% confidence limits.
All statistical analyses were performed using software programs (SAS Enterprise Guide version 7.15, SAS institute, Cary, North Carolina).
Missing data
One subject had missing data on the variables “age of onset” and “age at first hypomanic/manic episode” and additional one subject had missing data on “delay in diagnosis”. These missing data were handled by single imputation with the median value of the cohort. Twelve subjects had missing data on family history of BD, 4 had missing data on no. of previous hospitalizations / admissions, and 4 had missing data on education years. These missing data were likewise handled by imputation before inclusion in the logistic regression model. A maximum of 7 missing values on the individual YMRS items were present throughout all visits (849) and were handled in the mixed model analysis.
Logistic regression
A modified YMRS score for each visit was calculated based on the total YMRS score, minus item 1, 2, and 5, as these items were used to define the outcome, i.e., DSM-IV and DSM-5 criterion A. The mean modified YMRS score was found for each participant as a mean of the modified YMRS scores across their visits.
Logistic regression models were performed in order to investigate if any of independent variables could predict DSM-5 Criterion A fulfillment during follow-up. The outcome was defined as fulfilling DSM-5 criterion A at least once during follow-up in study (yes/no). Univariate and multivariable analyses were performed, and data were represented unadjusted and fully adjusted for
bipolar subtype, sex, no. of visits contributed, age first hypomanic/manic episode (years), no. of visits with depressed symptoms, mean modified YMRS total score, family history of BD, delay in diagnosis, education years. BD subtype, family history of BD, and sex were treated as binary variables, whereas the remaining variables were treated as continuous variables in the univariate and multivariable analysis. Variables included in the multivariable analysis were included based on clinical relevance and/or significance in univariate analysis (p<0.2).
Mixed models
Based on visits, we examined the association between individual items on the YMRS. Visits data (849 visits) represent repeated measures data and were analyzed accordingly, using generalized linear mixed-effects models, specifying each participant as random effect. The data represented zero-inflated data. Thus, the outcome was dichotomized, and the mixed models were specified as binomial.
The first analysis investigated the individual YMRS items (predictors) and their association with experiencing increased mood (outcome variable). The outcome variable mood (item 1) was defined as elevated (>0) and not elevated (=0). Univariate as well as multivariable linear mixed-effects models were conducted, and results presented unadjusted and fully adjusted for the other YMRS items. The analysis was repeated with irritability (item 5) as the outcome.
The analyses were specified with mood and irritability, respectively, as outcome variable, the individual YMRS items as predictors, and a random intercept for each participant to account for multiple observations made on the same subject.