This study used a cross-sectional descriptive design. Patients were recruited from one medical ward (in a university hospital), two heart failure clinics (one university and one county hospital) and five primary health care centres in the southeast of Sweden.

The template is used to format your paper and style the text. All margins, column widths, line spaces, and text fonts are prescribed; please do not alter them. You may note peculiarities. For example, the head margin in this template measures proportionately more than is customary. This measurement and others are deliberate, using specifications that anticipate your paper as one part of the entire journals, and not as an independent document. Please do not revise any of the current designations.

Insomnia was measured with the well-validated Uppsala Sleep Inventory-Chronic Heart Failure (USI-CHF) [2, 17]. The USI-CHF consists of 26 items of which four focuses on insomnia (i.e., difficulty initiating sleep, difficulty maintaining sleep, non-restorative sleep and early morning awakenings). The items about insomnia are to be answered on a five-point likert-type scale; from no problems (1), small problems (2), some problems (3), great problems (4), to very great problems (5). In the statistical processing and analysis of the insomnia variables of the USI-CHF scores of 1 - 3 were considered to represent “no complaints” and scores 4 and 5 “major complaints”.

Excessive daytime sleepiness was measured with the well-validated Epworth Sleepiness Scale (ESS) [18,19]. The 8 items focuses the expectation of “dozing” in a variety of situations. The probability ratings are zero (0), slight (1), moderate (2), or high (3) in the hypothetical situations. The ratings can be summarized to a total score of 24, with a cut-off value of >10 indicating excessive daytime sleepiness.

In the present study depressive symptoms was measured by the Mental Health Scale (MH-5), from the 36-Item Short Form Health Survey [20]. The 5 items focuses the anhedonic aspect of depression (i.e., sadness and loss of pleasure) and the scores can range from 0 (the worst) to 100 (best mental health). With a cut-off point of ≤60 the MH-5 seems valid and reliable to use in the screening for depressive symptoms indicating a high probability of a diagnosis of major depression [21-23]. The MH-5 can be said to assess a depressive symptomatology which might indicate the presence of different types of depressive disorders such as subsyndromal or minor depression, dysthymia or major depression [15]. Due to this the term depressive symptoms is used in this study. Hence, participants with a score ≤60 on the MH-5 scale were therefore labeled as having depressive symptoms and those with a score >60 were labeled as having no depressive symptoms.

HF nurses and primary health care nurses acted as coordinators at each center. Eligible patients with HF were found through a screening process on wards and heart failure clinics at the two hospitals and through the Diagnosed Related Groups register at each of the five primary health care centers. The medical records were scrutinized by an experienced heart failure nurse, together with one of the nurse-coordinators and a physician. All participants gave their informed written consent. The study protocol was approved by the ethics committee at the faculty of health sciences, University of Linköping, Sweden, and is in accordance with the provisions of the Helsinki declaration.

Descriptive statistics were used to describe differences between the HF population with and without depressive symptoms. Categorical variables were analyzed with the Chi-square test. Continuous variables, if normally distributed, were analyzed with the Student’s t-test (2-tailed) and presented as means and standard deviation (SD). The Mann-Whitney non-parametric test were used on skewed continuous variables, and presented as the median (md) and the 25th 75th percentiles. To determine the interrelationship between difficulties initiating sleep, difficulties maintaining sleep, early morning awakenings and non restorative sleep a Spearman’s rank correlation analysis was done. Binary regression modeling was used to analyze the independent relation between depressive symptoms and the insomnia variables. The model was adjusted for covariates that in Table 2 had a univariate p value < 0.15 (i.e. NYHA class). The significance level was set at 0.05. Statistical computing was done with the SPSS 18.0.

Table 2. Characteristics of HF patients (n = 212) and unvariate differences between patients with and without DS.

IHD—Ischaemic Heart Disease; MI—Myocardial Infarction; HT—Hypertension; SBP—Systolic Blood Pressure; DPB—Diastolic Blood Pressure; DIS— Difficulty Initiating Sleep; DMS—Difficulty Maintaining Sleep; EMA—Early Morning Awakenings; NRS—Non Restorative Sleep; EDS—Excessive Daytime Sleepiness.

The mean age of the patients was 76 years, 61% males and 42% lived with a partner (Table 2). A total of 94% were in NYHA classes II-III. Ischaemic heart disease (44%) mostly explained by myocardial infarction (41%) and hypertension (34%) were the most common causes of HF (Table 2). A total of 82% of the patients had their diagnosis of HF based on radiographic evidence and/or echocardiography.

Of the patients 34% (n = 72) suffered from depressive symptoms. Regarding age, partnership, sex and aetiology of HF, there were no significant differences between those with and without depressive symptoms (Table 2). A higher proportion of the non-depressed patients were in NYHA class II and a higher proportion of those with depressive symptoms were in NYHA class III (p = 0.006). Those with depressive symptoms had a lower systolic blood pressure (p = 0.04) compared to those without depressive symptoms (Table 2).

The most common insomnia type was non restorative sleep, which was reported as a major complaint by 44% of the patients. Major complaints of difficulties maintaining sleep, difficulties initiating sleep and early morning awakenings were reported by 23%, 19% and 15% of the patients, respectively. Patients with depressive symptoms had significantly more major complaints of non restorative sleep (p < 0.0001), difficulties maintaining sleep (p = 0.003), difficulties initiating sleep (p = 0.009) and early morning awakenings (p = 0.004) than to those without depressive symptoms (Table 2). Fully one fifth (22%) reported excessive daytime sleepiness. There was no significant difference in excessive daytime sleepiness reported between the patients with and without depressive symptoms (Table 2).

To examine which of the insomnia variables that independently was associated to depressive symptoms, all variables were included in the same regression model. However in this analysis none of the insomnia variables became significant. This can probably be explained because of an interrelationship between the insomnia variables (Table 3). Therefore separate models including the insomnia variables were calculated with adjustments for NYHA classes, with NYHA class II as the referent. In the unadjusted model all insomnia variables had a significant relation to depressive symptoms. Major complaints of non restorative sleep more than five folded the OR to have depressive symptoms. The OR to have depressive symptoms for major complaints of early morning awakenings almost increased by three fold, while the OR for major complaints of difficulties initiating sleep and difficulties maintaining sleep increased more than two fold respectively (Table 4). After adjustment for NYHA class, all insomnia variables were still significantly related to depressive symptoms. In the adjusted model major complaints of non restorative sleep more than five folded the odds to suffer from depressive symptoms. Reports of major complaints of early morning awakenings, difficulties initiating sleep and difficulties maintaining sleep respectively more than doubled the odds for having depressive symptoms. However, NYHA class by itself had a significant association to depressive symptoms only in the models for difficulties initiating sleep and difficulties maintaining sleep (Table 4).

Since many patients suffer from more than one type of insomnia, we studied the relationship between the amount of insomnia subtypes and depressive symptoms (i.e., the higher degree of insomnia the higher proportion of patients with depressive symptoms). In this model insomnia was categorized into three groups; the first group included those who did not report to suffer from a major complaint in any of the insomnia groups (non insomnia), the second group (mild insomnia) included those that had a major complaint in one or two of the insomnia groups; and the third group (severe insomnia) consisted of those who had a major complaint in three or four of the insomnia groups. The percentages of patients with depressive symptoms increased proportionally from 24% in the no insomnia group to 74% in the severe insomnia group (Figure 1). Compared to the non insomnia group, the patients with mild and severe insomnia had more than two fold and seven fold increased risk respectively to suffer from depressive symptoms. In the adjusted modelTable 3. Interrelationship between the insomnia subtypes in patients with HF.

DIS—Difficulty Initiating Sleep; DMS—Difficulty Maintaining Sleep; EMA— Early Morning Awakenings; NRS—Non Restorative Sleep; EDS—Excessive Daytime Sleepiness. ^**p < 0.01.

Table 4. Odds ratios (OR) for the association between the insomnia variables and depressive symptoms (DS) in patients with HF.

DIS—Difficulty Initating Sleep; DMS—Difficulty Maintaining Sleep; EMA— Early Morning Awakenings; NRS—Non Restorative Sleep; EDS—Excessive Daytime Sleepiness.

NYHA class did not have any significant relation to depressive symptoms (Table 5).

The prevalence of depressive symptoms in this study was 34%, which is comparable with the reported median prevalence of 42% found in studies that have used dimensional depression screening tools [15]. Of the patients in the present study 44% complained of non restorative sleep, while 23%, 19% and 15% of the patients complained of difficulties maintaining sleep, difficulties initiating sleep and early morning awakenings, respectively. Patients with depressive symptoms also reported more insomnia. Principe-Rodriguez et al. [7] found that 31% of the HF patients presented symptoms of insomnia, but the different insomnia types were not described. Another study found 37%, 35% and 36% of the patients reporting difficulties initiating sleep, difficulties maintaining sleep and early morning awakenings [1]. This is higher rates than in our study. However, the 44% who

Table 5. Odds ratios (OR) for the association between severity of insomnia and depressive symptoms (DS).

^*The group includes patients not suffering of insomnia (i.e. DIS, DMS, EMA, NRS); ^**The group includes patients suffering from major complaint in one or two of the insomnia variables; ^***The group includes patients suffering from major complaint of three or all of the insomnia variables.

reported non restorative sleep in our study are at the same level as those who reported that they felt tired or fatigued in the morning at least 3 - 4 times/week in Principe-Rodriguez et al. [7], and the 45% who reported insufficient sleep in the study of Redeker and Stein [1]. In total this indicates, that about 15% - 30% of the patients with HF suffer from insomnia and that almost 50% still are sleepy and fatigued when they wake up in the morning. The high prevalence of insomnia found also suggests that a large proportion of these patients also are at risk for suffering from depressive symptoms.

We found a strong association between insomnia and depressive symptoms. The OR to suffer from depressive symptoms for patients complaining of difficulties initiating sleep, difficulties maintaining sleep and early morning awakenings increased by more than two fold. These rates corresponds almost the figures found by LeesmanLeegte et al. [24] who reported the OR between sleep disturbance and depressive symptoms to be three folded. However, in the study of Redeker et al. [1] the OR between difficulties initiating sleep and/or difficulties maintaining sleep and depressive symptoms was five folded [16]. In our study a five folded increased risk was found for the association between non restorative sleep and depressive symptoms. One explanation for this may be explained by that fact that non restorative sleep might include feelings of fatigue and tiredness, which both are symptoms of HF without the co-existence of a sleep disturbance. However, NYHA class, a measure of functional capacity, had no significant relation to depressive symptoms in the separate non restorative sleep model. Moreover the relationships between insomnia and depressive symptoms persisted after that NYHA class was taken into consideration. This corresponds to the study of Redeker et al. [16] who also reported the relationship between depression and insomnia to persist after adjustments for clinical variables such as NYHA class and ejection fraction. In another study patient’s perception of their sleep explained 30% of the variance in mental health compared to the 15% explained by NYHA-class, age, gender and co-morbidity [25]. Hence, insomnia in HF patients may not only be a marker for severity of HF. Furthermore, in the present study we also analysed the association between severity of insomnia and depressive symptoms. Those with severe insomnia had a seven fold increased risk to suffer from depressive symptoms. This has not before been reported in patients with HF highlights the importance of interventions such as cognitive behavioural therapy (CBT) that can be used to target both insomnia and depressive symptoms. In older adults with arthritis and insomnia, CBT improved sleep as well as pain [26]. In patients with HF, CBT has been found to improve depressive symptoms and health-related quality of life [27,28]. However, the numbers of CBT studies performed on patients with HF are few and none have examined of the effects on sleep disturbances and depressive symptoms.

In future studies it is of importance to consider that there are several different factors that can cause difficulties initiating sleep, difficulties maintain sleep and early morning awakenings (i.e. secondary insomnia). Theoretically, insomnia can be described as the combination of sleep interfering and sleep interpreting processes [29]. The sleep interfering processes include different types of arousals (i.e., physiological, emotional and cognitive). A physiological arousal may be explained by an elevated baseline level of the stress-response system, as well as a slow return to the baseline level after being aroused. Emotional and cognitive arousals may be explained by a chronic tendency to worry, an elevated response to stressful events, problems, or increased dysphoric mood. The sleep interpreting processes imply that the insomniaacs’ perceptions of their own sleep are misinterpreted [29]. Sleep interfering processes leading to negative thoughts and anxiety, i.e., emotional and cognitive arousal in patients with HF can be can be associated with problems to perform activities in daily life, thoughts about HF including deterioration, medication and an unknown life situation [8,16]. Dysrhythmia, coughing and dyspnoea or have also been described causing a sense of panic during the night. Nocturia or inability to sleep in a supine position can also cause nightly awakening and be possible sleep interfering processes [8,30].

However, not all patients with a disturbed sleep will develop insomnia and/or depressive symptoms. In an elderly non HF population factors describing a psychological vulnerability (i.e., somatisation, intrusive and irresistible thoughts and worries) were associated to insomnia [31]. In another study of non-HF patients showed that insomniacs and good sleepers did not differ in the frequency of experienced daily stressors. The difference was in how they coped and perceived the stressors in daily life [32]. Poor psychological functioning in elderly people with and without HF has been found to mediate the association between insomnia and health related quality of life [33]. Thus, an HF patient who complains of insomnia should therefore be carefully evaluated for presence of depressive symptoms with common and validated specific instruments for insomnia and depression [15,34,35]. Treatment of insomnia in clinical practice often includes sleep-medications. However, in elderly with or without HF sleep-medication should be used carefully. This is because of the risk for polypharmacy and side-effects, such as sedation, drowsiness or somnolence [36]. Moreover, sleep medications have in elderly, independent of insomnia, been found to cause depression [37]. CBT (e.g., reduction of dysfunctional believes about sleep) can therefore be preferable. CBT studies in HF populations are missing but in other populations face to face CBT interventions have been successfully [38,39]. In a cardiovascular setting the accessibility of face to face CBT is limited due to lack of competence in CBT among the personnel. CBT provided by the internet has shown promising results to reduce insomnia as well as depressive symptoms [40-42].