Table 3 Proposed criteria for classifying lithium response

After all the study participants have completed 2 years of prospective follow-up, clinical response status will be determined by a panel of five experts (who are blinded to biomarkers findings). Using a “best estimate” method (Fennig et al. 1994), the panel will review all available clinical data and provide a consensus classification for each participant according to one of three categories: (i) good responders (GR), (ii) partial responders (PR), (iii) non-responders (NR). If no clinical follow-up data are available, the participant will be categorized as unclassifiable (UC)

Overview of procedure:

(i) The panel will first assess clinical outcomes using internationally agreed consensus definitions of clinical recovery, remission, and partial and full relapse (Tohen et al. 2009) and review illness activity for the 2 years prior to and 2 years after initiation of lithium treatment. Illness activity is a composite measure (e.g. incorporating number of syndromal episodes of bipolar disorder (BD), number of days ill or well during a specified time, etc.). The percentage change in illness activity will be calculated by estimating the difference in the level of illness activity in the 2 years after lithium initiation compared with the 2 years prior to lithium initiation

(ii) Next, the panel will use the evaluation of the clinical outcome to classify the participant in the relevant responder category

For example, for GR:

Classification of response: The expert panel will be asked to classify individuals according to the following steps:

1. All study participants will be classified into the most appropriate responder category irrespective of their exposure to lithium

2. The experts will consider responder status in the context of level of exposure to lithium (e.g. dose, duration of treatment, adherence and serum levels, adjunctive treatments, etc.)

3. Further classifications will examine different conceptualizations of response (e.g. time to remission or relapse, etc.) and other outcomes or putative changes of interest (e.g. sequence of change of symptoms, etc.)

The individual meets criteria for sustained remission, namely they experience euthymia (defined as a score on the Quick Inventory of Depressive Symptoms (qIDS) < 6 and the Bech Rafaelson Mania Scale (BRMS) < 7) for a minimum period ≥ 8 consecutive weeks during the 2-year follow-up period without evidence of relapse into a syndromal episode of BD at any time after achieving sustained euthymia

AND No additional mood stabilizer medication is prescribed after initiation of lithium treatment