Sleep Disorders and Cognitive Impairment in Peritoneal Dialysis: A Multicenter Prospective Cohort Study Open Access

Patients with chronic kidney disease (CKD) experience a high burden of sleep disorders caused by disease, treatment-related factors, and psychological factors. The prevalence of sleep disorders ranges from 20 to 70% among patients with CKD and from 44 to 95% among those undergoing dialysis [1-3]. Sleep disorders influence patients with CKD or maintaining dialysis in terms of mortality, quality of life, psychological status, and other aspects [4].

Evidence shows that there are associations between sleep disorders and cognitive impairment among patients with CKD and those on hemodialysis (HD). A cross-sectional study discovered that sleep-disordered breathing is associated with cognitive impairment, especially impaired verbal memory, in advanced CKD patients [5]. The other prospective cohort study indicated that patients undergoing HD with subjective sleep difficulties reported perceptual and memory problems [6]. It is noteworthy that dialysis patients have been found to suffer from cognitive impairment with an incidence of 27–67% [7, 8]. Cognitive impairment has also been shown to be an independent predictor of mortality and technique survival in this population. The use of peritoneal dialysis (PD) especially depends on normal cognitive functioning as it is a home care therapy requiring patients to self-monitor and self-manage their treatment.

Therefore, we aimed to investigate the associations between sleep disorders and general and specific cognitive function through a multicenter longitudinal PD cohort study.

This was a multicenter prospective cohort-study. Five PD centers from 5 provinces participated. Data from each center were collected within a strict quality control framework. Demographic characteristics, comorbidities, biochemistry, along with assessments of cognitive function, depression, and sleep disorders, were collected at baseline. All of the participants were followed prospectively. We repeatedly measured the biochemical data, cognitive function, depression status, and sleep quality of patients who were still on PD between March 2015 and November 2015.

The PD patients were enrolled between March 2013 and November 2013. The inclusion criteria were: age ≥18 years, having been undergoing PD for 3 months or longer and being clinically stable, and ability to complete all measurements and questionnaires as required. Patients were excluded if they had a systemic infection, acute cardiovascular events, active hepatitis or cancer, surgery or trauma in the month prior to the study, or any other study-obstructive condition such as severe eyesight loss, language incompatibility, illiteracy, a mental disturbance (preexisting dementia or confusion and various mental disorders), and upper limb disability. All of the participants received conventional glucose-based, lactate-buffered PD solutions (Ultrabag; Baxter Healthcare, Guangzhou, China).

Demographics and comorbidities, including age, gender, educational level, durations of PD, body mass index (BMI), systolic and diastolic blood pressure, primary kidney disease, presence of diabetes (DM), and history of cardiovascular disease (CVD), were recorded. CVD was recorded if one of the following conditions was present: angina, class III–IV congestive heart failure (New York Heart Association), transient ischemic attack, history of myocardial infarction or cerebrovascular accident, and peripheral arterial disease [9].

The biochemical data included serum sodium, serum albumin, calcium and phosphate, triglycerides, total cholesterol, high-sensitivity C-reactive protein (hs-CRP), and hemoglobin, which were calculated as the mean of measurements taken over the preceding 3 months. Biochemical profiles were investigated using an automatic Hitachi chemistry analyzer.

Assessments of cognitive function, depression status, and sleep disorders were performed in a separate room by trained medical staff in the morning. The sleep quality questionnaire included six parts to evaluate six major categories of sleep disorders. The insomnia part [10] covers nighttime symptoms of insomnia, daytime consequences of disturbed sleep, and frequency. The restless legs syndrome questionnaire includes International Restless Legs Syndrome Study Group (IRLSSG) criteria [11] and the IRLSSG rating scale [12], where a higher score in total identifies a much heavier disease burden. Excessive daytime sleepiness was assessed using the Epworth Sleepiness Scale (ESS) [13], with scores ≥10 indicating the presence of the disease (the higher the value, the larger the tendency to experience excessive sleepiness). The validity and reliability of the Chinese version of the ESS had also been evaluated in PD patients [14].Possible narcolepsy was screened for using the International Classification of Sleep Disorders (ICSD) questionnaire [15]. Sleepwalking and nightmares were assessed by Hatoum’s nighttime disturbance questions [16]. The part on possible rapid eye movement behavior disorders consists of reporting symptoms of movement of the body or limbs or the presence of potential harmful aggressive behaviors related to dreams [17].

The overall cognitive function was assessed using the Modified Mini-Mental State Examination (3MS) [18]. Global cognitive impairment was defined as a score of <80 on the 3MS based on previous observational studies of cognitive function [19]. Because the mean scores on the 3MS vary by education, we used a 3MS cut-off point of <75 for individuals with less than a high school education and a 3MS cut-off point of <80 for individuals with a high school education [20].

Specific cognitive functions were measured as executive function, immediate memory, delayed memory, visuospatial skill, and language ability. Trail-Making Tests A and B [21] were used to test executive function including decision-making and processing speed. Executive dysfunction was defined as a Trail-Making Test A duration of >75 s and a Trail-Making Test B duration of >180 s [22]. Subtests of the Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) were adopted to assess immediate memory (list learning and story memory), delayed memory (list recall, list recognition, story recall, and figure recall), visuospatial skill (figure copying), and language ability (picture naming and semantic fluency) [23]. The reliability and validity of the RBANS have already been proven in Chinese populations [20, 24]. Raw scores were transformed to age-standardized T scores for all subtests of the RBANS. T scores less than 1 SD below the published mean in an education-grouped Chinese population on a subtest indicated impairment [25].

Depression status was assessed using Zung’s Self-Rating Depression Scale [26], which has been validated in general Chinese population and for various medical illnesses.

Continuous data are presented as means ± SD, except for durations of PD, RRF (residual renal function), and high sensitivity C reactive protein (hs-CRP) levels, which are given as medians and interquartile range (IQR) due to a high skewness. Categorical variables are presented as frequencies and proportions. Student’s t test, the Mann-Whitney U test, and the χ² test were used to compare differences in basic characteristics. A paired Student t test and the Wilcoxon signed rank test were employed to compare cognitive function parameters between baseline and the 2-year follow-up.

Using linear regression models, we explored the association between sleep disorders and cognitive function at baseline. Both univariable and multivariable linear regression models were used. Factors reported to be associated with cognitive impairment in previous studies, including demographic data (age, gender, educational level, and BMI), comorbidity status (DM, CVD, and depression status), and laboratory parameters (serum albumin, sodium, and hs-CRP) were all included in the models as independent variables. HR and 95% CI were calculated.

We then used multivariable linear regression models to further analyze risk factors for worsening general and specific cognitive function, respectively, in the longitudinal study. The change of score in each cognitive parameter was used as a dependent variable. The same recognized cofactors mentioned above were included in the models as independent variables. The analysis for each cognitive domain was adjusted for the baseline level of the corresponding cognitive parameter.

The level of significance for all statistical tests was set at 0.05. All of the analyses were performed using SPSS for Windows, version 22.0 (SPSS Inc., Chicago, IL, USA).

Of the 667 patients who were eligible for this study, 493 (73.9%) gave consent to participate. Among the 493 patients, 458 (92.9%) completed baseline cognitive and sleep testing. Demographic and laboratory data for our participants were in accordance with the general characteristics of the PD population in China, with mean values as follows: age, 51.6 years; PD duration, 25.1 months; BMI, 22.9; hemoglobin level, 104.9 g/L; and serum albumin level, 36.2 g/L. Of these patients, 53.1% were male, 14% applied assisted PD, 23.6% had DM, 21.0% had a history of CVD, and 52.4% had an education of high school diploma or higher (Table 1).

A total of 165 patients were excluded for various reasons during the follow-up period (Fig. 1). Therefore, the remaining 293 patients were assessed for cognitive function and sleep disorders repeatedly. There were no significant differences in general or specific cognitive function parameters between the excluded patients and the remaining patients. The DM prevalence and serum hs-CRP levels were significantly higher in the excluded group (p = 0.006 and p = 0.04, respectively). No significant difference was found in other demographic data or biochemical data between the 2 groups (for details, see [27]; Table 1).

In this PD data set, the prevalence of having at least one of the six assessed sleep disorders at baseline was 65.3% (Fig. 2b). As shown in Table 1, compared to patients without sleep disorders, patients with at least one of the six assessed sleep disorders tended to be male, be more reliant on assisted PD, and have diabetes and depression. They also tented to have lower levels of total urea clearance per week (Kt/V).

The scores of each sleep disorder at baseline are shown as means ± SD (Fig. 2a). In this PD data set, among the assessed sleep disorders, the most common one was excessive daytime sleepiness (47.6%), followed by insomnia (22.6%), sleep walking and nightmares (19.1%), restless legs syndrome (7.9%), possible rapid eye movement behavior disorders (7.2%), and possible narcolepsy (4.7%) (Fig. 2b). As shown in Figure 2c, having one sleep disorder accounted for 48% of the sleep disturbance group, having any two combined sleep disorders accounted for 26%, and suffering from any five combined sleep disorders accounted for 2%.

At the second survey, general and specific cognitive function and sleep disorders were repeatedly assessed in 293 patients (Table 2). When comparing scores for each test, we observed a worse performance on the Modified Mini-Mental State Examination (3MS) and excessive daytime sleepiness. However, we observed a better performance at the 2-year follow-up on Trail-Making Tests A and B, the immediate memory test, and visuospatial skills. The scores for delayed memory and language skills, insomnia, restless legs syndrome, possible narcolepsy, sleep walking and nightmares, and possible rapid eye movement behavior disorders were not significantly different.

Associations between sleep disorders and cognitive function at baseline are shown in Table 3. After adjusting for recognized confounders, possible narcolepsy was significantly associated with worse general cognition as assessed based on 3MS scores (p = 0.002).

During the follow-up, the longitudinal relationship between sleep disorders and cognitive dysfunction was also explored. Using a multivariable linear regression model, sleepwalking and nightmares were found to be related to declining delayed memory (p = 0.02 and p = 0.044, respectively). However, restless legs syndrome seemed to be associated with better executive function as assessed by both Trail-Making Test A and Trail-Making Test B. Beyond that, other sleep disorders were neither hazardous nor protective factors for the monitored sleep dysfunction (Table 4).

As far as we know, this is the largest multicenter prospective cohort study to measure the association between sleep disorders and cognitive impairment in patients undergoing PD. In this PD data set, the prevalence of having any of the 6 sleep disorders we assessed was 65.3%. The most common specific disorder was excessive daytime sleepiness (47.6%), followed by insomnia (22.6%). At baseline, possible narcolepsy was associated with general cognitive dysfunction. Moreover, in the longitudinal analysis, sleepwalking and nightmares were risk factors for a decline in delayed memory capacity.

The prevalence of possible narcolepsy in our study was similar to the results of previous studies in HD cohorts (1.4–15.9%) [17, 28, 29]; it was higher than in the general population (0.03–0.06%) [30]. The higher rate of narcolepsy in patients with end stage renal disease (ESRD) might be due to a changed orexin level, an autoimmune disorder, infections, toxins, or psychological stress, among other factors [31]. Whether experiencing narcolepsy during bag change could increase the risk of PD-related complications is unknown in previous literature. According to our data, possible narcolepsy was significantly associated with general cognitive function. The research on the association between possible narcolepsy and cognition is relatively new, and, consequently, data concerning patients with ESRD or CKD are sparse. Past studies conducted in the general population have found that narcoleptics have deficits in attention and memory [30].

The underlying mechanism of this reno-cerebral syndrome remains unclear. One hypothesis is that a hypocretin deficiency could drive the medial prefrontal cortex (a cortical region involved in associative function and attention) dysfunction due to less excitatory action and a consequent deficit of attention and memory [32]. It is suggested that acute sleep deprivation [33]and chronic sleep disorder [34] increase the cerebral amyloid-β level, yet a relatively short-term partial sleep deprivation increases the orexin concentration [35], which is a significant indicator of cognitive impairment. Sleep disorders could be associated with cognitive impairment in patients with CKD. Cognitive decline and sleep disturbances could be manifestations of brain dysfunction in CKD. There are numerous co-contributors to sleep disorders and cognitive impairment in the progression of CKD, such as anemia, uremia, protein energy wasting, neuropsychological aspects [36], secondary hyperparathyroidism [37], unstable hemodynamics [38], systemic inflammation, and a frequent presence of electrolyte disorders [39]. Behind these clinical associations, there are also common pathologic structure changes in the regional brain. Past studies have suggested that an injured hippocampus, small-vessel ischemic brain disease, and deep white matter demyelination could be seen in CKD patients with sleep disorders, which is vital for executive and memory function [39-42]. As narcolepsy is a severely disabling disorder, we encourage PD patients with potential symptoms to be assisted with bag exchange and exit site care in case of infectious and noninfectious complications. More research concerning this issue in a selected group would be urgent.

Sleepwalking and nightmares were found to be risk factors for a decline in delayed memory in our study. A review of the literature found no current evidence of this in a renal-compromised population; however, a correlation between global cognitive function and vivid dreams and nocturnal restlessness has been observed in patients with Parkinson disease [43]. The mechanisms underlying these associations remain to be determined; one hypothesis is that dysfunctions in the cholinergic systems and their projections to subcortical and cortical regions might play a role [44].

Our longitudinal study indicated that restless legs syndrome might be associated with a better executive function, which is against previous data revealing that executive functions are decreased in individuals with restless legs syndrome [45]. The potential mechanisms are not clear. Of note, there was a general improvement in executive function in the whole cohort, which is partly related to the amelioration of uremic symptoms during long-term dialysis treatment, as supported by a past study [46]. Also, we need to collect more information on nutrition, dialysis adequacy, and physical activity among patients with restless legs syndrome in a future study. The exploration of the association between restless legs syndrome and executive function should consider more potential confounders accordingly.

Several strengths of the present study are listed below. First, this study shows the association between sleep disorders and cognitive function in a large PD sample. Second, our cognitive measures included a broad and comprehensive range of tests encompassing a variety of cognitive domains, such as executive, memory, and language functions. Third, recognized confounders for cognitive impairment in general and dialysis populations were controlled in multivariable models.

We are also aware of limitations of this study. First, due to the nature of observational studies, it cannot be determined whether sleep disorders are pathogenic factors or solely risks factors for cognitive impairment. Second, almost 40% of the participants were not included in the second assessment, which could underestimate the trend in decreasing cognitive function as the drop-outs tended to be sicker at baseline. Besides, this study was performed exclusively in a Chinese population of patients who were treated with continuous ambulatory peritoneal dialysis (CAPD), and patients with severe cognitive impairment were excluded, thereby raising the possibility of selection/ascertainment bias. Moreover, polysomnography as a gold standard was not applied in this study. Since it is cumbersome, time consuming, expensive, and not applicable outside of the home environment, a panel of questionnaires consisting of multiple dimensions were measured. Most of the questionnaires have verified its validity and reliability in populations with CKD or under renal replacement therapy [14, 17, 47, 48].

In conclusion, we are the first study to demonstrate that sleep disorders and cognitive impairment have non-negligible associations in a PD population, as certain sleep disorders might cause specific domains of cognitive functioning to deteriorate. Moreover, possible narcolepsy, sleep walking, and nightmares were less commonly studied, although they had high potential risks especially among individuals undergoing PD. Therefore, more attention needs to be focused on this aspect so that more light can be shed on the pathophysiology, precaution, and comprehensive management pertaining to this field.

The authors would like to express their appreciation to the patients, doctors, and nursing staff of the Peritoneal Dialysis Centers of Peking University First Hospital. The authors would also like to express their gratitude to Prof. Yingdong Zheng of the School of Public Health of Peking University for his dedication in the statistical analysis in the study.

The ethics committee of Peking University First Hospital approved this study (approval No. 2013[587]). Patients gave written consent for their information to be stored in the hospital database and for it to be used in research.

The authors have no conflict of interests to declare.

This study was in part supported by New Century Excellent Talents of the Education Department of China and an ISN Research Award from the ISN GO R&P Committee. The funders had no role in the study design, data collection and analysis, the decision to publish, or the preparation of this paper.

Research idea and study design: J. Dong, Y. Zhao, and Y. Zhang. Data acquisition: Y. Zhao, Y. Zhang, Z. Yang, Z. Xiong, J. Liao, L. Hao, G. Liu, Y. Ren, Q. Wang, L. Duan, and Z. Zheng. Statistical analysis: Y. Zhao, Y. Zhang, and J. Dong. Supervision or mentorship: J. Dong. Each author contributed important intellectual content during drafting or revision of this paper and accepts accountability for the overall work by ensuring that questions pertaining to the accuracy or integrity of any portion of this work are appropriately investigated and resolved.

Youlu Zhao and Yuhui Zhang contributed equally to this work.