In the present study, individuals were required to have a DSM-IV-TR diagnosis of BDI or BDII in a current depressive episode, according to the structured clinical interview for DSM-IV axis I disorders (SCID-I) (15). Patients began the lithium treatment on a dose of 450 mg/day and subsequent dosage adjustments were permitted in a flexible manner to a dose of ≤900 mg/day, based on the clinical efficacy (response/remission rate) of individual patients and the level of lithium in their plasma. Measurements for the plasma lithium levels were obtained on days 7 and 14, and at the endpoint. Short-term use of benzodiazepines/hypnotics for a maximum of five consecutive days (once per day) was permitted during the follow-up period. Subjects with BDII had not taken any psychopharmacological treatment for at least six weeks prior to the study while subjects with BDI may have been treated with a maximum of one mood stabilizer or one antipsychotic agent. All patients had not been previously treated with lithium. Lithium was obtained from Eurofarma (São Paulo, Brazil). The current study was a single center study conducted at the Institute of Psychiatry of the University of São Paulo (São Paulo, Brazil), clinical trial number NCT01919892.

Male and female outpatients, aged between 18–45 years were eligible to participate in the study. All subjects were recruited and followed up at the Mood Disorders Program, Laboratory of Neuroscience LIM27, Institute of Psychiatry of the University of São Paulo (São Paulo, Brazil) between August 2010 and June 2012. Clinical assessments included: the score for the 21-item Hamilton Rating Scale for Depression (HAM-D) as the primary outcome, and the Young Mania Rating Scale (YMRS) and the Clinical Global Impression (CGI) scale scores as secondary outcomes. A HAM-D score of ≥18 was also required for inclusion. All adverse effects were recorded using the Udvalg for Kliniske Undersogelse (UKU) side-effects rating scale. Early improvement was defined as ≥20% improvement from the baseline HAM-D score following one week of lithium treatment. The response rate was defined as a reduction of ≥50% in the HAM-D score at endpoint and the remission rate was characterized as a HAM-D score of <8 at endpoint. There was no run-in period. At the onset of the study, all subjects had <3 mood episodes in their lifetime and an illness duration of ≤5 years. Exclusion criteria included: rapid cycling in the past 12 months, previous head trauma, a current Axis I psychiatric disorder other than BD, subjects submitted to electroconvulsive therapy, significantly abnormal laboratory test results and any chronic medical condition. The diagnosis was determined by experienced psychiatrists who held an inter-rater reliability score of >0.9 for HAM-D and YMRS. The first assessment was carried out at baseline [visit (V)1] and subsequent visits for clinical assessment, dose adjustment and monitoring of the levels of lithium were performed at the end of weeks 1 (V2), 2 (V3), 4 (V4) and 6 (V5). The present study was approved by the local institutional ethics committee of the Clinics Hospital (São Paulo, Brazil) and all patients provided written consent prior to participation in the study.

The primary outcome was the degree of change in depressive symptoms (HAM-D scores) analyzed using mixed-effects random regression. The mixed-effects regression was used to analyze changes in HAM-D assessments over the 6 weeks of follow-up and their association with time and clinical variables that were potentially associated with response to lithium treatment including: age, gender, duration of illness, history of psychosis, BD subtype and plasma lithium levels at week 6. Inter-group comparisons of continuous variables with a Gaussian distribution were performed using the Student’s t-test. The analyses of continuous variables with a non-parametric distribution were conducted using the Mann-Whitney U Test (the Kolmogorov-Smirnov test was used to verify the normality of the data within each study group). The chi-square test was used for the comparison of categorical factors. The Pearson coefficient was used to evaluate the correlation between the relative change in HAM-D scores over time and other demographic and clinical variables. Statistical significance was set at P<0.05 (two-tailed) and data are presented as means ± standard deviations. All statistical analyses were carried out using SPSS software, version 16.0 (SPSS, Inc., Chicago, IL, USA).

A total of 42 patients were accessed for eligibility to participate in the current study following a telephone screening at the Mood Disorders Program, LIM27, University of São Paulo. Of these, 31 patients were included in the present study. Two drop-outs were observed: one patient left due to personal problems unrelated to the study and the other patient was excluded due to noncompliance to the treatment (undetectable blood lithium levels). Thus, the final sample at endpoint comprised of 29 acutely depressed bipolar patients.

Table I summarizes the clinical and demographic data for the subgroups of patients with BDI and BDII. The final study group comprised 21 females (72.4%). The patients had a mean age of 28.4±5.4 years and a diagnosis of BDI (n=11; 37.9%) or BDII (n=18; 62.1%). At baseline, the patients had a mean duration of illness of 36.1±19.6 months and presented mean HAM-D and YMRS scores of 22.5±3.5 and 6.1±5.6, respectively. Out of the 29 patients that completed the current study, 21 were naïve to lithium treatment (72.4%) and 25 (86.2%) had been medication-free for at least six weeks prior to participating in the study. Three of the patients had YMRS scores >8. Furthermore, only four (13.8%) patients had a previous history of psychotic symptoms.

The 29 patients who completed the study were followed-up with lithium monotherapy (n=25) or lithium therapy associated with another mood stabilizer or antipsychotics (n=4). The mean plasma level of lithium at V5 was 0.49±0.20 mEq/l.

A significant reduction in HAM-D scores was observed over the six weeks of follow-up in the whole group of BD patients (P<0.001). Following six weeks of lithium treatment, the response and remission rates in the total study population were 86.2% and 62.0%, respectively. The mean HAM-D and YMRS scores following six weeks of lithium treatment were 7.3±5.9 and 3.8±8.6, respectively. When comparing BDI and BDII, similar remission rates were observed at endpoint; 63.6 vs. 61.1%, respectively (Fig. 1). The CGI scores, which were 4.21±0.55 pretreatment, were significantly decreased at endpoint (2.14±0.99; P<0.001) with no significant difference identified between the patients with BI and those with BII (P=0.86). None of the patients experienced a switch to (hypo)mania during the follow-up period.

At week 6, 15 patients had lower (Li<0.5 mEq/l) and 14 presented higher (Li≥0.5 mEq/l) plasma lithium levels. The plasma lithium levels did not differentially impact the number of patients who responded (n=13 responders with lower levels vs. n=12 responders with higher levels; P=0.94). Furthermore, the use of hypnotics (n=7), if necessary, was not observed to influence the clinical outcome (HAM-D scores; P=0.37). No other significant association was observed in the mixed-effects regression. No significant differences were observed in patients in remission (n=18) vs. those not in remission (n=11) at endpoint (a HAM-D score <8) with regard to oral dosage, blood lithium levels, gender and bipolar disorder subtype (data not shown). No association was observed between HAM-D scores and lithium dose (P=0.99) or blood levels (P=0.87) at endpoint.

With regard to the potential predictors of response rate, only a previous history of psychosis was observed to be significantly associated with a faster reduction in HAM-D scores in response to lithium treatment (F=3.68, P=0.008). None of the patients experienced a switch to (hypo)mania during the follow-up period. Baseline YMRS scores did not influence the response rate (data not shown). Common adverse effects observed were polydipsia/polyuria (62.1%), cognitive complaints (41.4%), nausea (31.0%), increased oniric activity (31.0%) and sedation (31.0%). In subjects with higher lithium levels compared with those with lower levels, an increased prevalence of nausea (50.0 vs. 13.3%; P=0.03), restlessness (42.9 vs. 6.7%; P=0.02) and headaches (42.9 vs. 6.7%; P=0.02) was observed. No patient drop-outs due to side-effects occurred during the follow-up period.