The relation between insomnia and depression in the subacute phase after stroke

ABSTRACT

Prevalence rates for both depression and insomnia the first year after stroke are around 30%, significantly impacting the prospects of recovery, rehabilitation, and quality of life. Furthermore, the risk of insomnia and depression becoming chronic is high in the subacute phase post-stroke. This cross-sectional observational study investigated whether insomnia and depression are related in the subacute phase post-stroke, using validated instruments. Sixty-six outpatient stroke survivors participated. Depression was measured using the depression subscale of the Hospital Anxiety and Depression Scale (HADS-D) and insomnia severity with the Insomnia Severity Index (ISI). A multiple linear regression analysis was used to examine the association between the dependent variable post-stroke depression and the independent variables insomnia and pre-stroke depression treatment. Results showed that insomnia (β = 0.48, t = 4.40, p < 0.001) and pre-stroke depression treatment (β = 0.24, t = 2.28, p = 0.026) were both significant predictors of depression. Participants with more insomnia complaints and participants with pre-stroke depression treatment had more depression symptoms post-stroke. Therefore, it is important to be alert in the subacute phase post-stroke of both, insomnia and depression complaints.

KEYWORDS:

• Stroke
• Depression
• Insomnia
• Rehabilitation

Introduction

Stroke is a worldwide leading cause of mortality and disability (Cai et al., 2019). As the number of stroke survivors is increasing (Feigin et al., 2019), more patients need to cope with disability. This will lead to substantial economic costs for post-stroke care and to a challenge for the healthcare system to provide adequate rehabilitation and long-term care (Rajsic et al., 2019). Due to the distinct cerebrovascular nature and etiologies of strokes, studying a stroke-specific population enables targeted investigation into clinical symptoms, which may not directly translate to larger acquired brain injury cohorts (Harris et al., 2022).

Post-stroke depression is often associated with a reduction in rehabilitation treatment efficacy (Gillen et al., 1999), poor functional outcomes (Hama et al., 2007), lower levels of participation or quality of life (Ayerbe, Ayis, Wolfe, et al., 2013), recurrence of stroke and increased mortality (Cai et al., 2019; Jørgensen et al., 2016; Thomson, 2014; Yuan et al., 2012). Reported prevalence rates of depression in the first year post-stroke range from 29% to 33% (Ayerbe, Ayis, Crichton, et al., 2013; Hackett et al., 2005). Actual rates are expected to be even higher, as under-reporting and under-recognition of post-stroke depression are presumed due to neurological disability (Hackett et al., 2005). Most depressive symptoms start within three months after stroke and are characterized by a short duration (Ayerbe, Ayis, Wolfe, et al., 2013). Post-stroke depression displays a high recurrence rate in the general stroke population and only 15% to 57% recover from depression in the first year after stroke (Ayerbe, Ayis, Crichton, et al., 2013). Moreover, 12% of the patients who have not recovered from the initial depression within the first six months post-stroke develop a chronic, persistent depression (Berg et al., 2003; Robinson et al., 1987; Schepers et al., 2009). Over time, research has identified several risk factors associated with post-stroke depression, such as genetic influences, and the extent of functional impairment (Robinson & Jorge, 2016). While social support is frequently found and implicated as a risk factor for post-stroke depression it is not as well-established (Hackett et al., 2005; Morris et al., 1991; Robinson & Jorge, 2016). One of the most common risk factors of post-stroke depression is pre-stroke depression with one out of six patients with a lifetime history of pre-stroke depression developing depression again following a stroke (Ayerbe, Ayis, Wolfe, et al., 2013; Kutlubaev & Hackett, 2014; Taylor-Rowan et al., 2019). Therefore, considering pre-stroke depression is important when assessing the likelihood of post-stroke depression.

A second, common and severe problem in stroke survivors is sleep disturbances, which are also associated with poor recovery (Duss et al., 2017), an increased risk of a recurrent stroke (Gottlieb et al., 2019; Wallace et al., 2012), and with depression (Davis et al., 2019; Fleming et al., 2021; Nakase et al., 2016; Suh et al., 2016; Zhao et al., 2021). A frequently occurring sleep disturbance following a stroke is insomnia (Hermann & Bassetti, 2016; Palomäki et al., 2003; Pérez-Carbonell & Bashir, 2020). Insomnia is characterized by dissatisfaction with sleep duration or quality, and difficulties initiating or maintaining sleep, along with substantial distress and impairments in daytime functioning (Morin et al., 2015). Prevalence rates of insomnia in the first year post-stroke are about 30% to 38% (Glozier et al., 2017; Leppävuori et al., 2002). In addition, 16% of people with a stroke reported chronic insomnia over the first six months (Glozier et al., 2017). This resembles the prevalence rates of post-stroke depression. Insomnia and depression most often appear in the subacute phase i.e., three to six months post-stroke, and are both at risk to become chronic during this phase.

Post-stroke insomnia and depression appear to be associated (Baylan et al., 2020). Sleep problems specified as insomnia and hypersomnia are one of the criteria used to classify a depressive disorder based on the Diagnostic and Statistical Manual of Mental Disorders-V (DSM-V; (American Psychiatric Association & Association, 2013)), and “disturbed sleep” is listed in the ICD-10 as an “associated symptom” of depression (WHO, 2020). Thus, although the concepts are associated, insomnia does not necessarily have to be present in order to classify a depressive disorder, and not all patients with insomnia meet the classification of a depressive disorder. Although studies in other populations consistently find associations between insomnia and depression (Nutt et al., 2008), their association following stroke requires further investigation.

To our knowledge, the studies of Leppävuori et al. (2002) and Glozier et al. (2017) are among the few studies that investigated insomnia and depression in the subacute phase post-stroke. Assessment in the subacute phase post-stroke is crucial for reliability, and informing early interventions to prevent long-term symptoms, as the acute phase’s rapid changes and medical interventions may obscure accurate evaluation (Grefkes & Fink, 2020). Three months post-stroke 51% of the insomniac patients suffered from minor depression and 34% were classified with major depression following the DSM-IV classification (Leppävuori et al., 2002). Furthermore, post-stroke insomnia increases the risk of developing depression within the first year post-stroke by six times (Glozier et al., 2017). Additionally, patients with stroke and chronic insomnia were more likely to be depressed and anxious (Glozier et al., 2017).

According to (Baylan et al., 2020), post-stroke insomnia research is still in its infancy and the quality of post-stroke insomnia studies is variable. The studies that have been conducted thus far differ strongly in the way in which post-stroke insomnia and depression are operationalized. Leppävuori et al. (2002) used the DSM-IV as a classifying tool for (primary) insomnia, in addition to diverse depression rating scales that did not control for somatic overlap in stroke patients. Glozier et al. (2017) used a validated depression rating for the stroke population, however, participants were only eligible between 18 and 65 years of age, while most stroke patients are older than 65 years. In addition, they did not use a standard research assessment of insomnia. In insomnia research in the general population, it is standard to use validated instruments measuring insomnia (Buysse et al., 2006; Ibáñez et al., 2018; Morin et al., 2011; Riemann et al., 2017; Wong et al., 2017). It is important to use validated instruments in the stroke population as well to obtain consistent results. Additionally, to prevent circular reasoning when measuring depression in stroke patients, it is important to use an instrument that measures depressive symptoms that show limited overlap with somatic consequences of stroke. Variable operationalization of studies might limit conclusions or generalizations and thus more studies with validated diagnostic classification tools without overlap are needed.

Therefore, in this study, the relationship between insomnia and depression was examined, in the subacute phase (i.e., three to six months post-stroke) using validated diagnostic classification tools. We hypothesized that the severity of insomnia is positively associated with the severity of depressive symptoms. Furthermore, we expected that a history of pre-stroke depression is positively associated with post-stroke depression as well. In the current study, age and sex served as control variables. Age and sex show an identified but weak association with post-stroke depression (Essau et al., 2010; Hackett et al., 2005), yet considering their significant impact on insomnia (Sivertsen et al., 2009) and major depressive disorder (Salk et al., 2017), they are included as control measures.

Methods and materials

Participants

Patients who suffered an ischemic stroke and visited the outpatient neurological unit or rehabilitation program between June 2020 and October 2021 and who approved to be contacted, were eligible for participation in this study. Inclusion criteria were: (1) diagnosis of an ischemic stroke, based on medical records and confirmed by a neurologist; (2) three to six months post-stroke; (3) 18 years or older. Exclusion criteria were: (1) hospitalization; (2) severe aphasia; (3) diagnosis of other brain diseases, such as objectified dementia; epilepsy; brain tumour; or history of traumatic brain injury; (4) serious psychiatric disorders such as schizophrenia or bipolar disorder; (5) insufficient knowledge of the language based on clinical judgment; (6) use of sleep medication.

Procedures

The study protocol received approval from the Medical Ethics Committee of Zuyderland Medical Center (METC-Z) with the registration number Z2020058 on June 2, 2020, and was subsequently registered in The Netherlands Trial Register under the registration number NL8475. Patient recruitment occurred either during a six-week post-stroke follow-up visit at the neurological unit by a nurse practitioner or during an outpatient rehabilitation program by a medical psychologist, which took place between 6 weeks and 3 months post-stroke. The nurse practitioner or medical psychologist introduced the patient to the study, provided the information letter and informed consent form, and asked the patient for permission to be contacted by a member of the research team. Patients who gave their permission to be contacted were approached by phone, during which additional information was provided, and upon agreement to participate, the patients were asked to return the signed informed consent form via email or post. Next, demographic variables, including age, sex, level of education (low, or high), and civil status (single, living together/married, widow(er), divorced) were collected via a phone call. Verhage’s education scoring system, consisting of seven levels was used to quantify education (Verhage, 1964). This was translated as having a low (Verhage 1–5) and a high (Verhage 6–7) level of education. During the phone call alcohol use (number of alcoholic beverages used per week), nicotine use (number of cigarettes used per day), caffeine use (number of caffeine consumptions per day), and lifetime drug use (yes or no) were registered, as were time since stroke, current and past treatment of depression at the time of inclusion, and diagnosis of a sleep disorder. Subsequently, an appointment was scheduled three to six months post-stroke, during which participants completed questionnaires about sleep quality (Pittsburgh Sleep Quality Index (PSQI), insomnia (Insomnia Severity Index (ISI), and depressive symptoms (Hospital Anxiety Depression Scale-Depression (HADS-D)). This single measurement point encompassed the comprehensive collection of all relevant data for the study. Questionnaires were filled out either at home or at the clinic, utilizing paper and pencil or using an online platform; both location and the method of presentation of questionnaires were based on the participants’ preferences.

Measures

Hospital anxiety and depression scale: Depression was assessed with the depression subscale of the HADS (Zigmond & Snaith, 1983). The HADS depression subscale consists of 7 items. Each item is scored on a four-point Likert scale and the score ranges from 0 to 21. A score ≥ 8 is an indicator of depression in people with acquired brain injury including stroke (Whelan-Goodinson et al., 2009). The HADS is a reliable measure and has been validated in the acquired brain injury population (Whelan-Goodinson et al., 2009).

Pre-stroke depression: A lifetime history of pre-stroke depression treatment was used as a possible predictor for post-stroke depression. Patients were asked if they had a history of treatment for pre-stroke depression (yes/ no). We specifically asked about depression treatment because asking about a history of depression can be subjective, as individuals may have varying interpretations of what constitutes depression and when they consider themselves to be experiencing it. This ensured a more standardized and clinically relevant assessment.

Insomnia severity index: Post-stroke insomnia severity was measured with the Insomnia Severity Index (ISI). This seven-item questionnaire assesses the nature, severity, and impact of insomnia (Morin et al., 2011). Each item is scored on a five-point Likert scale, yielding a total score range from 0 to 28. The scores can be categorized as absence of insomnia (0–7); sub-threshold insomnia (8–14); moderate insomnia (15–21) and severe insomnia (22–28). An ISI-score of ≥15 is indicative of insomnia. The ISI corresponds with the DSM-V criteria and International Classification of Sleep Disorders third edition (ICSD-3) criteria for insomnia (Wong et al., 2017) and has valid and reliable psychometric properties, also in individuals with stroke (Gagnon et al., 2013; Morin et al., 2011; Niu et al., 2023). The ISI is a recommended diagnostic self-report instrument to detect insomnia (Buysse et al., 2006; Riemann et al., 2017; Wong et al., 2017).

Pittsburgh sleep quality index: The PSQI is a widely used subjective sleep questionnaire comprising 19 items assessing seven scales related to sleep problems. These scales include overall sleep quality, sleep onset latency, sleep duration, sleep efficacy, sleep disturbances, use of sleep medication, and daytime dysfunction. A global score is derived by summing up item scores on these scales, ranging from 0 to 21. A global score of five or higher indicates poor sleep quality. The questionnaire demonstrates high internal consistency with a Cronbach’s alpha of 0.83 (Buysse et al., 1989; Mollayeva et al., 2016). The Dutch version of the PSQI has been utilized to assess sleep quality in patients with acquired brain injuries (Hofman et al., 2007). In the current study, the global score, poor sleep quality (score of >5), and quantitative data from the first four questions are described: bedtime (in hours), minutes to sleep, wake-up time (in hours), and total sleep time (in hours).

Statistical analysis

Descriptive statistics were used to describe the sample. The data were examined for normality, curvilinearity, homoscedasticity, and multicollinearity, no assumptions were violated. Independent samples T-tests were used to compare depression scores (HADS-D, dependent variable) between males and females, and between pre-stroke depression treatment (yes/ no). Pearson correlation coefficients were used to assess the correlations between the depression (HADS-D) and the independent continuous variables age, and insomnia (ISI).

Subsequently, a multiple linear regression analysis was conducted to investigate the relationship between the dependent variable depression (HADS-D), and the independent variables insomnia (ISI) and pre-stroke depression treatment. Age and sex were included as control variables. Statistical significance was reached at p < 0.5 (two-tailed). Data were analysed using IBM SPSS software version 27.0.

A sample size calculation was performed using G*Power version 3.1.9.6 to estimate the number of participants required for the study focusing on the analysis that needed the most power. Based on a linear multiple regression analysis (fixed model, R² deviation from zero) including two predictors (insomnia and pre-stroke depression treatment) and two control variables (age and sex), an alpha of 0.05, a medium effect size (beta = 0.2) and power of 0.8, 65 participants needed to be included in the study.

Results

Participant characteristics

Ninety-one stroke patients were notified of the study. After checking the inclusion and exclusion criteria, three patients were excluded. One patient died before inclusion. Sixteen patients did not fill in the questionnaires. After inclusion, four patients dropped out due to non-responding and one patient had to be excluded due to missing data. Leaving the data of 66 participants to be included in this study.

Participant characteristics are shown in Table 1. The sample included 26 females and 40 males with a mean age of 69.3 years. The mean time between the stroke and assessment was 102.3 days. Most participants filled in the questionnaires at home (67%), and most of the participants used the paper and pencil version (83%). In terms of psychiatric properties, fifteen participants (23%) had had a pre-stroke depression treatment, and three participants (5%) were treated for depression during the study. Ten participants (15%) had at least mild symptoms of depression (HADS-D > 8) and ten participants (15%) were classified with insomnia (ISI ≥ 15). Of the total sample, 9% had complaints of insomnia only, another 9% had complaints of depression only and 6% of the participants had both complaints of insomnia and depression. Four participants (6%) were diagnosed with a sleep disorder; three with obstructive sleep apnea syndrome and one with insomnia. Thirty-one participants (48%) reported a poor quality of sleep (PSQI > 5).

Table 1. Participant characteristics (N = 66).

	N (%) or range	Mean (SD)
Age (years)	50–92	69.33 (10.99)
Sex (female)	26 (38.8%)
Civil status
Single	2 (3%)
Married/Living together	53 (79.1%)
Widowed	4 (6%)
Divorced	5 (7.5%)
Education
Low	49 (74.2)
High	17 (25.8)
Time since stroke and assessment (days)	90–180	120.30 (25.27)
Pre-stroke depression treatment (yes)	15 (22.7%)
Current depression treatment (yes)	3 (4.5%)
Diagnosed sleep disorder (yes)	4 (6%)
Caffeine (p.d.)	0–12	3.3 (2.4)
Alcohol (p.w.)	0–24	1.7 (3.8)
Nicotine (yes)	10 (14.9%)
Drugs (yes)	0
Depression (HADS-D)	0–19	7.95 (3.90)
Clinically significant (≥8)	10 (15.2%)
Mild (8–9)	5 (7.6%)
Moderate (10–14)	5 (7.6%)
Moderately severe (15–19)	1 (1.5%)
Severe (≥20)	0 (0%)
Insomnia (ISI)	0–20	6.32 (5.68)
Clinically significant (≥15)	10 (15.2%)
Subthreshold (8–14)	13 (19.7%)
Moderate (15–21)	10 (15.2%)
Severe (≥22)	0 (0%)
Sleep quality (PSQI)	0–15	6.66 (3.97)
Clinically significant (>5)	31 (48.4%)
Bedtime (hours)	20:00–01:00	23:00 (1.05)
Minutes to sleep	0–150	28.52 (32.64)
Wake-up time (hours)	5:00–11:00	7:45 (1.04)
Total sleep time (hours)	3–10	6.92 (1.50)

Notes. N (%): number of participants; %: percentage of participants. HADS-D: Hospital Anxiety and Depression Scale – Depression; PSQI: Pittsburgh Sleep Quality Index; ISI: Insomnia Severity; p.d.: per day; p.w.: per week.

Relation between depression and possible predictors

Independent sample t-tests showed that participants who had received pre-stroke depression treatment scored significantly higher on the HADS-D than participants who had not received pre-stroke depression treatment (t [64] = 2.34, p = 0.023). There was no difference in depression score (HADS-D) between males and females (t [64] = 0.35, p = 0.71; Table 2).

Table 2. Differences in depression (HADS-D) scores between male and female and between pre-stroke depression treatment and no pre-stroke depression treatment.

	N	HADS-D score
		Mean (SD)	t	df	p
Pre-stroke depression treatment
No	51	3.08 (3.33)	2.34	64	0.023*
Yes	15	5.67 (5.02)
Sex
Male	40	3.53 (4.08)	0.36	64	0.71
Female	26	3.88 (3.66)

Notes. HADS-D = Hospital Anxiety and Depression Scale – Depression; SD = standard deviation; *p < 0.05.

Pearson correlations showed a significant positive correlation between the HADS-D and insomnia (ISI; r = 0.49; p = 0.001). The HADS-D score and age were not significantly correlated (r = −0.07; p = 0.27).

Relation between depression, insomnia and pre-stroke depression treatment

Multiple linear regression analysis was used to investigate the relationship between depression (HADS-D) and insomnia and pre-stroke depression treatment (Table 3). The model explained 30.3% of the variation of the HADS-D score (F [4] = 6.61, p < 0.001). Insomnia (β = 0.48, t = 4.40, p < 0.001) and pre-stroke depression treatment (β = 0.24, t = 2.28, p = 0.026) were both significant predictors of depression (HADS-D score). A higher score on the ISI questionnaire was associated with a higher depression score on the HADS-D subscale. Pre-stroke depression treatment was associated with a higher depression score on the HADS-D subscale as well.

Table 3. Results of multiple linear regression analysis examining the relationship between the dependent variable depression (HADS-D), and independent variables insomnia and pre-stroke depression treatment controlled for age and sex.

Variables	Model R^2= 0.303 Adjusted R^2 = 0.257 F = 6.614 p < 0.001**
	B	β [95% CI]	t	p
Age	0.02 [−0.06 to 0.10]	0.07 [−0.16 to 0.29]	0.58	0.56
Sex	0.09 [−1.63 to 1.81]	0.01 [−0.21 to 0.23]	0.10	0.92
Insomnia (ISI)	0.33 [0.18–0.48]	0.48 [0.26–0.70]	4.40	<0.001
Pre-stroke depression treatment	2.30 [ 0.28–4.31]	0.25 [0.03–0.47]	2.28	0.026

Notes. HADS-D: Hospital Anxiety and Depression Scale – Depression; ISI: Insomnia Severity Index; B: unstandardized regression coefficient; β: standardized regression coefficient.

Discussion

The main objective of this study was to determine the relationship between depression and, insomnia and pre-stroke depression treatment in the subacute phase (i.e., three to six months) post-stroke using validated instruments for depression and insomnia. In line with the hypotheses, results showed that insomnia severity and pre-stroke depression treatment are both positively related to symptoms of depression post-stroke. Participants with more insomnia complaints had more depressive symptoms post-stroke. Additionally, participants with pre-stroke depression treatment had more depressive symptoms post-stroke.

Our findings are in accordance with other studies in the subacute phase post-stroke, showing an association between insomnia and depression and in which a history of pre-stroke depression appeared to be a risk factor for developing depression post-stroke (Ayerbe, Ayis, Wolfe, et al., 2013; Glozier et al., 2017; Kutlubaev & Hackett, 2014; Leppävuori et al., 2002; Taylor-Rowan et al., 2019). However, in our study the number of participants scoring above the cut-off for both depression and insomnia was low (6%). Nevertheless, the number of participants scoring above the cut-off on one of the questionnaires was also relatively low, approximately 15% for both measures, which might explain why there was little overlap. Our results are also in line with studies in the general population in which increased frequency of insomnia is related to increased levels of depression (Taylor-Rowan et al., 2019). Considering the increased risk of developing a chronic, persistent depression (Berg et al., 2003; Robinson et al., 1987; Schepers et al., 2009) or chronic insomnia within the first six months after stroke (Glozier et al., 2017) and the impact on rehabilitation (Gillen et al., 1999), this study underlines the importance of addressing both depression and insomnia in the subacute phase post-stroke.

Our insomnia rates (15%) based on the ISI score (ISI > 15) resemble the insomnia prevalence rates (16%) over the first six months post-stroke found by Glozier et al. (2017). However, our rates are lower compared to the insomnia (37.6%) and depression (34.4%) rates found by Leppävuori et al. (2002). The difference might be explained by the operationalization and the classification of insomnia. Leppävuori et al. (2002) defined insomnia when there was a presence of insomnia complaints, which is a broader classification, while we used a stringent cut-off to define the presence of an insomnia disorder. This is a valid cut-off to diagnose insomnia in stroke patients (Niu et al., 2023). This shows the importance of strict and uniform measurement to increase the construct validity and greater clarity regarding the classification (disorder vs symptoms).

Our depression rates (15%) based on the HADS score (HADS ≥ 8) were lower compared to other studies in the subacute phase, which reported depression rates between 20% and 30% using similar measures (Liu et al., 2023; Stokman-Meiland et al., 2022). Our lower ratings may be because our recruitment occurred in a standard follow-up visit tailored to all stroke patients. In contrast, other studies often include patients at later stages, such as during rehabilitation, where milder cases that do not require rehabilitation are typically absent. Nonetheless, it is noteworthy that our results are based on a stroke sample that was relatively mild in depression ratings compared to other studies.

Strengths and limitations

One of the strengths of this study is the use of validated questionnaires. The ISI is a standardized questionnaire following the DSM-V criteria and ICSD-3 for insomnia and has been shown to have reliable psychometric properties in stroke patients (Niu et al., 2023; Wong et al., 2017). The HADS is validated for stroke patients (Sagen et al., 2009) and explicitly measures depression (Skilbeck et al., 2011; Zigmond & Snaith, 1983). By using the HADS we prevented circular reasoning because the HADS measures depressive symptoms that show limited overlap with somatic consequences of stroke, nor with post-stroke sleep disturbances. However, we acknowledge the possibility that post-stroke physical symptoms may have influenced the depression score, thereby introducing potential confounding factors. Furthermore, somatic symptoms can be more present in depressed stroke patients compared to non-depressed stroke patients, which is something to consider in future research (de Man-van Ginkel et al., 2015). Another strength of this study is the relatively stable time frame of three to six months post-stroke, not including the impact of the acute phase.

This study has several limitations. First, a cross-sectional study design was used, which does not imply causality. However, a cross-sectional study can show prevailing characteristics in a population at a specific point in time. Secondly, inclusion took place during the COVID-19 pandemic, therefore, patients might have been more hesitant to participate. However, this is speculative since we do not know how many patients declined to be contacted about the study, and for what reason. Furthermore, it is relevant to note that while we excluded participants using sleep medication, we maintained those using antidepressant medication at the time of testing. Although secondary analyses showed that the exclusion of participants using antidepressant medication did not change our results, we recognize the importance of future research being more concise, considering the potential impact of both sleep and antidepressant medication usage, or adopting a consistent approach to their inclusion or exclusion. In addition, by specifically asking about depression treatment we might have missed patients with a depression diagnosis who did not receive treatment. Although we might have missed patients with pre-stroke depression, and thus assigned them to the wrong group, pre-stroke depression treatment still predicted current depression scores in our sample, suggesting that the number of misclassified patients could not have been substantial. Lastly, it should be noted that the relationship between depression and insomnia is intertwined, complex, and multifactorial. Other factors could have influenced the relationship between insomnia and depression. This includes for example fatigue or anxiety, but also multiple lifestyle factors such as nicotine use or alcohol consumption and other known risk factors for post-stroke depression such as genetic factors, stroke severity, the severity of the functional impairment and social support. Furthermore, the COVID-19 pandemic itself might have influenced sleep quality and depression negatively. However, our sample size was too small to investigate the influence of other moderators

Clinical implications

This study, combined with the results of previous studies leads to several clinical implications. First, since 15% of the participants of our study met the clinical diagnosis for insomnia and several other studies have shown that insomnia can develop in the subacute phase after stroke (Glozier et al., 2017; Leppävuori et al., 2002) follow-up of insomnia symptoms during this phase is important. The ISI seems to be a valid instrument for this population (Niu et al., 2023). Furthermore, the subacute phase, compared to the acute phase, shows more stability concerning neurological status, psychological reactions, and environmental conditions. Nonetheless, initially “good sleepers”, patients with non-insomniac symptoms directly after stroke, still have a 23% rate of developing insomnia within six months post-stroke (Glozier et al., 2017). Furthermore, in our study, nearly half (48%) of the participants experienced poor sleep quality (PSQI > 5). In the general population, insomnia has been shown to have a causal role in the development of depression (Baglioni et al., 2011; Li et al., 2016; Taylor et al., 2005). In addition, insomnia, or even a subjective experience of poor sleep, is a major risk factor in the general population for relapse of depression, even after successful depression treatment (Dombrovski et al., 2008; Jindal & Thase, 2004; Perlis et al., 1997; Reynolds 3rd et al., 1997; Riemann & Volderholzer, 2003). This underscores the importance regular follow-ups and early interventions targeting acute sleep problems post-stroke to prevent the development of chronic sleep problems (e.g., insomnia) and depression. As far as treatment is concerned, cognitive behavioural therapy for insomnia (CBT-I) lowered depression symptomatology in the general population (Ballesio et al., 2018; Van der Zweerde et al., 2019). Moreover, CBT-I has also proven to be an effective treatment in the stroke population for both, insomnia, and depression (Herron et al., 2018; Nguyen et al., 2019).

Future research could examine the trajectory of insomnia and depression following stroke through the implementation of multiple measurements over time. Such an approach would yield valuable insights into whether one symptom precedes the other and guide the identification of appropriate early interventions. In addition, risk factors for post-stroke depression such as pre-stroke depression, genetic factors, stroke severity, the severity of functional impairment, and social support could be included, and the possible influence of moderating factors, such as lifestyle and personality (Hackett et al., 2005; Morris et al., 1991; Robinson & Jorge, 2016).

Conclusion

In conclusion, there is a positive relationship between insomnia and depression at three to six months post-stroke. Participants with more post-stroke insomnia complaints had more depressive symptoms and participants with pre-stroke depression treatment had more depression symptoms post-stroke. In research and the clinic, it is therefore important to be alert in the subacute phase that insomnia and depression complaints often co-exist.

Acknowledgements

The authors would like to thank R. Nieling, Nurse Practitioner and Dr. S.Z. Stapert, medical psychologist for their contribution to this study.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.