Can Physical Activity, Sleep Parameters, and Sleep–Wake Patterns Predict Outcome of Combined Chronotherapy in Mood Disorder During Routine Clinical Practice? An Exploratory Study

Alterations in circadian rhythms, clinically reflected in diurnal mood variations and sleep difficulties, are common in mood disorders [1]. These disruptions are particularly pronounced in Major Depressive Disorder (MDD) and Bipolar Disorder (BP), where they are considered a core pathophysiological feature of the disorder [2]. A well-established, intensive therapy that specifically targets resetting the circadian rhythm is combined chronotherapy (CCT). The combination of repeated sleep deprivation (SD) and light therapy (LT) is a type of CCT used to treat severe depressive symptoms with the intention of improving mood and stabilizing sleep and sleep–wake patterns [1]. Interest in CCT started with the discovery of the rapid positive effect on mood after a single night of SD, which diminishes following a night of recovery sleep [3]. To prolong the beneficial effects on mood, SD has been combined with other forms of CCT such as LT [4].

Currently, CCT is used in the clinic to treat patients suffering from a depressive episode in both bipolar and unipolar depression [5,6]. Despite the known potential, its use in the clinic is limited [7]. This is likely partially due to the limited information available regarding which patients will benefit from CCT, a gap that if addressed, could inform personalized treatment decisions for individual patients. Some evidence for markers that possibly predict a positive response to SD alone and combined with LT is reported for a patient’s self-reported chronotype (evening-type) and positive diurnal mood variation (mood best in the evening) [8]. Furthermore, a higher level of observed physical activity on the day before SD predicts a positive response to SD [9]. Identifying ways to measure chronotype, diurnal mood variation, and physical activity in a clinical setting may help predict CCT response more precisely in individual patients. Actigraphy provides an unobtrusive way to measure physical activity, sleep parameters, and sleep–wake patterns [10]. Time series data assessed with actigraphy are used to calculate non-parametric markers of sleep–wake pattern stability [11]. As CCT is used to stabilize patient’s sleep and sleep–wake patterns, it is expected that these markers change after CCT. Supporting this hypothesis, in their systematic review, Panchal and colleagues (2022) showed that therapies specifically targeted at the circadian rhythm (e.g., light therapy, cognitive behavioral therapy for insomnia, blue blocking glasses) stabilized sleep and sleep–wake patterns as indexed with actigraphy-derived markers in patients diagnosed with a Bipolar Disorder [12].

Previous research established the efficacy and feasibility of CCT [4,8,12]. A study that included actigraphy assessments at baseline before CCT for patients with a unipolar depression assessed sleep characteristics and did not analyze whether these could predict a positive response to CCT [13]. These studies have not yet incorporated actigraphy to assess potential predictive markers of individual patients’ responses to CCT. Moreover, like most medical studies, these studies used a group-based (nomothetic) approach to contribute to knowledge about generalized patterns across the sample. However, treatment in daily clinical practice demands insights into the psychopathological characteristics of a specific individual patient. To enrich personalized medicine, research approaches that take into account between-patient heterogeneity, as well as within-patient longitudinal changes during treatment and symptom fluctuations, are urgently needed [14,15]. A well-established research approach to do so is the individual-based, single-case design approach, which has wide applicability in health-related research and practice [16,17]. In this approach, the value of a study lies not primarily in the number of subjects, but in the number of repeated observations collected per subject.

In this exploratory clinical study, actigraphy and questionnaires were used to monitor a small number of patients treated with CCT for their depressive symptoms in regular clinical practice. The study had the following aims: (1) examine whether actigraphy can be used to obtain markers on physical activity, sleep, and sleep–wake patterns during CCT; (2) compare actigraphy-derived markers before and after CCT; (3) explore whether actigraphy could be used to search for predictors of response to CCT in future studies.

In this exploratory study, we demonstrated that actigraphy is an easy-to-use tool for obtaining markers of physical activity, sleep parameters, and sleep–wake patterns in patients undergoing CCT in regular clinical practice. The results indicated that baseline actigraphy data may provide useful information systematically related to CCT response, as patients who responded positively to CCT had a higher baseline FI and spent more time in bed, while exhibiting lower SE. We tentatively conclude that it may be valuable to investigate these markers as predictors of response in individual patients in a larger study. These findings could contribute to the development of personalized treatment strategies for depression.

A notable advantage of using actigraphy for assessing physical activity, sleep and sleep–wake patterns is the amount of objective information that can be passively collected during routine care. This adds to routine clinical data collection with self-report questionnaires on mood and sleep measures, as often this subjective information does not strongly correlate with objective measurements [25]. Another advantage, although not formally assessed, is that patients reported no side effects or made other negative comments related to wearing the actigraphs. Actigraphy is increasingly utilized in research to assess physical activity, sleep parameters, and sleep–wake patterns in mood disorders, underscoring the importance of exploring its potential for other applications, such as identifying subgroups for clinical staging [26] or, as suggested in the current paper, predicting individual treatment response.

In the group of patients receiving CCT, we found an overall decrease in depressive symptoms, FI, and IV after CCT treatment. The 33% response rate in this sample was lower compared to earlier reports of 40–60% [5]. Yet our sample was small, and since data were collected in routine care, not all data were complete. The percentage of response was similar in the study of Sikkens et al. (2019) performed in the same hospital, who reported a response rate of 34% to CCT including 26 depressed patients [6]. Moreover, there is a relatively high number of patients with treatment-resistant depression that are treated at our tertiary academic psychiatric center. The patients of the current study had all previously received at least two types of antidepressant medication. To compare, in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial, the remission rate at levels 3 and 4 (third and fourth switch of antidepressant medication because of nonresponse) was 14% and 13% [27]. A previous study by Benedetti et al. also reported a lower response rate (50% reduction in depressive symptoms on the last day of combined CCT) for medication-resistant patients (44%) versus non-resistant patients (70%) [28]. Two out of three of our responders in the final week of CCT were no longer responders in the week after CCT. This reflects the difficulty in maintaining CCT’s effect after the initial response, which has been reported before. A significant decrease in depressive symptoms one week, but not two weeks after treatment with SD was reported in a meta-analysis on the efficacy of SD, combined with standard care in the treatment of depression [29]. Adding other chronotherapeutic approaches may be helpful to maintain the effect for a longer period, such as sleep phase advance or social rhythm therapy [5,6].

Our study had multiple limitations. First, the study had a small sample size with a large age range and there was no control group. Because of that, the data collected could not be tested by statistical analyses and results could not be compared between men and women, making the results tentative and potentially limiting their replicability in larger samples. Additionally, the sample consisted of mixed diagnoses (MDD, BP), resulting in a less homogeneous sample, which may further limit the interpretability and generalizability of the results. To ensure transparency, we combined the visual inspection of the data in the results section with complete patient-level results provided in the Supplementary Materials (Figure S1). Nevertheless, replication in a larger sample is necessary, together with the use of robust research designs that capture both between-patient heterogeneity and within-patient longitudinal changes during treatment, as well as symptom fluctuations [30]. Second, in addition to the small sample size, data were not complete for every patient, which could have introduced a potential bias in the interpretation of the results. For example, because of the missing data, response to CCT was calculated using either the timepoint at the end of the CCT or the subsequent week. The percentage of responders may have been different if the response was calculated using the same timepoints for every patient. The main reason for the missing data is that the study was performed during regular care and not as part of a clinical trial. However, this real-world setting provides valuable insights into the practical feasibility of incorporating innovative technologies like actigraphy into routine clinical practice. Moreover, an MCID for the IDS-SR has not been established yet. As explained in the Method section, based on the MCID scores of other depressive symptom scales, a decrease of ≥30% in IDS-SR score was considered a significant decrease in depressive symptoms in this study [20]. Using another percentage might have shifted the categorization of the patients into the responder/nonresponder groups and the subsequent visual trends described in the results because of the small sample size. Third, most patients were using medication during the CCT period (e.g., benzodiazepines, antidepressants, mood stabilizers; see Table 1), which may have influenced physical activity, sleep, and sleep–wake patterns [31]. Because this study was conducted in a routine clinical setting, these effects could not be controlled for, but they should be considered when interpreting the findings. A notable strength of our study is that, to our knowledge, we were the first to explore physical activity, sleep parameters, and sleep–wake patterns on an individual patient level with actigraphy in CCT in a routine clinical setting. Systematic information assessments at the individual level can enrich and guide personalized treatment strategies, individualized predictions, and shared decision-making [32]. Nevertheless, both group-based and single-case design approaches have their limitations and strengths [14]. Ideally, they should complement each other to bridge the research and clinical practice gap, facilitating personalized medicine.

From this exploratory study, we tentatively conclude that actigraphy may be a useful add-on tool to CCT to assess physical activity, sleep, and sleep–wake patterns during regular clinical practice. In line with previous research, CCT was found to stabilize depressed mood, sleep, and sleep–wake patterns. Furthermore, patients with more disrupted sleep–wake patterns prior to treatment may benefit most from CCT. In the future, actigraphy may hold the potential to predict individual patient responses to CCT. This could support the development of more personalized treatment strategies for depression.