Effect of kidney transplantation on sleep quality in patients admitted in ICU:a cross-section study in China

Background Although evidences showed that sleep disorder is common in patient with end stage renal disease (ESRD), less is known about their sleep quality after early post-kidney transplantation (kTx) especially in Intensive Care Unite (ICU). Thus, the purpose of this study is to investigate sleep quality of kTx recipients in ICU and explore factors related poor sleep, second, to measure the correlation of subjective sleep quality and sleep architecture assessed by PSG in kTx recipients. Methods This study recruited participants from Participants required to complete the Pittsburgh sleep quality index(PSQI) and demographic questionnaire as baseline data and received one night of Polysomnography (PSG) in the ICU within 96 hours of surgery, during which time sound and light data were monitored. After that Richards Campbell sleep questionnaires (RCSQ) also need completed. 26 participants self-reported sleep quality and sleep efficiency based on RCSQ was at middle level (49.2 ± 25.6mm), and 14/26(53.8%) kTx recipients in ICU were poor sleepers defined by RCSQ <50. PSG showed that most kTx recipients in ICU had shallow sleep with mainly stage 2 sleep time (80.90 ±70.10 min), lower total sleep time (136.50 ±86.41 min), higher awakening frequency after sleep onset (8.87 ±5.92 times) and long awaken time (94.67 ±75.09 min) when a sleep disruption occured. multiple linear regression analysis showed that self-reported noise and pain were the significant factor affecting sleep(P < 0.05).


3
Conclusion Subjective sleep quality based on RCSQ scored better than objective one measured by PSG in kTx recipients, sleep disruption always remained a substantial problem and affected by self-reported noise and pain.

Background
Sleep disorder is one of the most common comobidities in patients with end-stage renal disease (ERDS), and the prevalence rate is gradually increasing. It causes abnormal sleep and poor sleep quality, which could generate some undesirable health consequences, such as fatigue [1] , depression [2] , decreased quality of life [3] , premature mortality [4] as well as major economic consequences [5] .
Kidney transplantation (kTx) is the main treatment for patients with ERDS. It seems that successful transplantation could restores normal metabolism and is likely to improve sleep quality [6] . Most studies suggested that sleep quality improved in kTx recipients (30%-52%) [1,7] than patients with ERDS (49%-78%) [8][9][10] . Our previous research [11] also showed the percentage of "poor sleep" in kTx recipients during follow-up was much lower about 29.2% (defining as global Pittsburgh sleep quality index (PSQI) >7). These results from previous reports all seem to be very good, but most of the subjects were only targeted at recipients with longer transplant times, none studies cared about individuals who stayed in ICU at earlier postoperative times.
Poor sleep could be defined using specific diagnostic criteria subjectively and objectively. In general, subjective questionnaires are the simple and convenient methods for preliminary sleep screen and diagnosis. Commonly used questionnaires include PSQI [12] , Epworth Sleepiness Scale [13] , and the Berlin Questionnaire [14] . 4 However, all the screening survey questionnaires only provided general described or categorized results about sleep symptoms and there is no physiological analysis of sleep. Polysomnography (PSG) is a significant tool for sleep related evaluations and sleep studies. It could provided quantitative information of sleep and its deviation from the normal by analyzing sleep physiological parameters. Although PSG is the gold standard for diagnosing sleep disorders, only a few published articles [15][16][17][18] reported PSG assessment of sleep in kTx recipients for it requires long wear times that limited its use in sleep research. However, Considering that PSG can still provide relatively detailed sleep physiological information and intuitively reflect the patients' sleep situation at night in ICU. Therefore, the purposes of this study were: 1.
to describe the sleep quality characteristics of kTx recipients in ICU based on subjective questionnaires and PSG, and to explore factors related poor sleep.

2.
to measure the correlation of subjective sleep quality and sleep architecture assessed by PSG in kTx recipients.

design, setting and participants
This was a prospective, cross-sectional study using convenient sampling.

variables and measurement
Eligible participants needed to complete the demographic, medical history questionnaire and self-reported sleep quality as baseline data before surgery . Then an objective and subjective sleep assessment were respectively performed at postoperative transplantation ICU. Meanwhile, Pain, sound, and light were also collected during the sleep monitoring.

Demographics, clinical characteristics
Age, gender, height, weight, occupation, status of marriage or fertility, degree of education, monthly income, the type of sleep and waiting time for transplantation were collected by interviewing. Primary disease, the type of dialysis before kTx, times of operation, creatinine (μmol/l), creatinine clearance rate (ml/min), hemoglobin (g/l), immunosuppressors and sleep drugs before and after operation were also retrieved from the participants' hospital information system.

self-reported sleep quality
The self-reported sleep quality of kTx recipients before and after surgery was investigated using the PSQI and Richards Campbell sleep questionnaires (RCSQ).
Self-reported sleep quality of kTx recipients within one month prior to surgery and 96 hours after surgery using the PSQI and RCSQ.
The PSQI is one of the most widely used and most easily self-reported sleep quality questionnaires. It has been validated and used in a variety of clinical populations by differentiating between ''poor-'' and ''good-''quality sleepers. The PSQI includes 19 questions to assess seven domains of sleep: subjective sleep quality, sleep latency, duration, efficiency, disturbances, use of sleep medication, and daytime dysfunction over the past month. Scoring is based on a 0-3 Likert scale, which a score of 3 reflects the negative extreme. Global PSQI scores range from 0 to 21 with a cutoff score of >5 indicating poor sleep in the general population [12] . Liu et al [19] used the PSQI > 7 as the cutoff to differentiate the poor sleep quality in the Chinese version of PSQI , which showed a higher sensitivity (98.3%) and specificity (90.2%).
RCSQ was developed by Richards et al [20] specifically for patients admitted in ICU.
The RCSQ scale consisted of 5 items including sleep depth, difficulty falling asleep, number of awakenings, difficulty in re-sleeping, and overall sleep quality. Each item was scored using visual analogue scale from 0~100mm. Patient marks the distance on the line that represents each item score. The total RCSQ score is the average score of 5 items, with higher scores indicating better sleep. Naik et al [21] defines good sleep in one study as a total RCSQ score of ≥50 (sensitivity 88.24%, specificity 86.67%), patients with total RCSQ score <50 were considered to have poor sleep.The Chinese version of the RCSQ scale has a content validity of 0.840 and a Cronbach's α coefficient of 0.874 [22] .

objectively measured sleep quality
PSG is known as the current gold standard for measuring sleep. Attended one overnight PSG would be performed for each participant. This technique employs numerous collections of surface electrodes, which measures physiologic parameters of sleep including brain dynamics of EEG, eye movements, muscle activities, heart physiology, and respiratory function [23] . Finally, data from the PSG were

Other variables
Pain, sounds and lights are the common factors affecting sleep, these variables would be also collected when PSG was performed. Visual analogue scale (VAS) were used to assess the pain degree of recipients, it's easy and simple to use for post-surgery patients. Sounds were detected by digital sound level meter (SMART AR844), lights by luminometer (TES-1330A).

Analysis of sleep micro-structure
The

Statistical analysis
All data were reviewed by the double check. If the missing values in the questionnaire was >20% or key values were missing in the PSG, the data in this case was invalid. Statistical analyses were performed using SPSS 25.0 (SPSS Inc., Chicago, IL). The continuous data were described as mean and standard deviation, while the categorical data were described as frequencies and percentages. 8 Correlation between RCSQ and PSG was tested using the Spearman correlation test.
Multivariable linear regression analysis was used for exploring the influencing factors of subjective sleep quality.

Sample characteristics
40 participants were enrolled, of those 8 were excluded for invalid RCSQ questionnaires and 6 for incomplete PSG data. Finally, 26 participants were eligible.
The study profile, including reasons for loss to follow-up is summarized in figure 1.
The participants had a mean age of 38.08 ± 9.68 year, 17/26 were male and their waiting time for kTx less than 1 year, 24/26 individuals received hematodialysis (HD), the mean HD time was 13.7 ± 11.6 months, 9/26 had a bad sleep quality (PQSI>7) before the transplant (Table 1).

subjective sleep quality in post-operative ICU
The mean RCSQ score of kTx recipients in ICU was 48.8 ± 20.6 mm, which indicated that self-reported sleep quality and sleep efficiency was on the middle levels.
According to the sleep disruption definition( total RCSQ<50 mm) by Naik, 14/26(53.8%) individuals were poor sleepers when stayed in ICU at earlier postoperative times.

objective sleep quality and their association with RCSQ score
According to the analysis of PSG sleep characteristics, it showed that kTx recipients In the next step, the association between RCSQ score and PSG sleep characteristics was analyzed. Known from the original questionnaire, RCSQ item corresponds to multiple sleep characteristics of PSG in each sleep domain. As presented in the

subjective sleep quality and associated factors
Correlation analysis was performed for each predictor and potential confounder variables, only pain (Spearman r=0.42, P=0.03) and noise (Spearman r=0.67, P=0.00) correlated with RCSQ score.
Furthermore, pain and noise with multivariable linear regression were analyzed. The result showed that higher RCSQ score was significantly associated with higher pain and noise (Table 4).

Discussions
Although there is a growing evidence show the importance of sleep quality on kTx recipients [2,3,6,24] , most of them only focused on patients from outpatient or clinic with some subjective sleep questionnaires. Few studies investigated the sleep quality of kTx recipients who stayed in postoperative ICU, especially rarely used objective sleep assessment tool. This study provides a current characteristic of the quality and quantity of sleep of kTx recipients in ICU based on the subjective and objective methods.
This study found that self-reported sleep quality and sleep efficiency of recipients was at a middle level which were consistent with previous reports [21,25] . Study [26] has shown that as the transplantation time prolongs, sleep quality of kTx recipients will gradually improve, but the symptoms of sleep disorder will be still worse than the general population. Difference from the previous researches, this study described the sleep characteristic and level using RCSQ rather than PSQI in the people who stayed at transplantation ICU. RCSQ described a specific and concurrent sleep pattern of an overnight sleep monitoring in ICU, and the mean total score (49.2 ± 25.6mm) obtained in our study was lower than mean score (51.42 ± 21.48mm) obtained in other study [25] . Both these two study sites were in the surgery ICU, but our participants were younger (38.1±9.7 vs 57.72 ± 14.81 yr), it is a well-known fact that perception of sleep quality decrease with advancing age [27] .  [28] , which manifest that reduced REM sleep, sleep fragmentation, increased arousal, and deep sleep reduction. was evident in ICU. In fact, some studies [29][30] reported that short and fragmented sleep architecture had occurred in patients with chronic dialysis before kTx. For the problem of poor sleep in the ICU,various factors have been proposed, including patient-related factors and environmental factors. There is an agreement in the literature that environmental factors might greatly impact the patients' quality of sleep in ICU [31] .
To further analyze the relationship between subjective sleep and objective sleep, bivariate analysis was used. Due to a single item of RCSQ may correspond to multiple sleep characteristics of PSG, so a single sleep item cannot be used as a proxy for a PSG sleep characteristic. The results showed that all RCSQ items are less correlated with corresponding PSG characteristics with no statistically significant.The study was different from the Richards et al [20] study which demonstrated there was a moderate correlation between RCSQ and PSG, The five item RCSQ score can predict 33% of the variance in the sleep efficiency index.
Several reasons could explained it when there was a subjective-objective mismatch.
First of all , age could affected the perceived sleep. In one study [32] , the relation between subjective sleep quality and PSG variables in older and younger women was investigated, it found that perceived poor sleep was related to short TST, long awakening time, low sleep efficiency and high frequency of awakenings. For those older women, even they had lower TST and sleep efficiency and longer awakening time, their rate of good sleep was still higher than younger women, because older women would adjust objective criteria for good sleep downwards. Second, cognitive and psychological could be another factors including time perception. For poor sleepers, they always tended to underestimate nocturnal sleep and overestimated awake times related to their own sleep problem [33] . In our study, The small sample size, the younger subjects may be factors influencing inconsistency between the subjective and objective sleep. Also Consider that the kTx recipients wore with a PSG monitor for the first time and the "first-night effect" could increase both stage 1 sleep time and REM latency [34] .
Moreover, multivariable linear regression analysis showed that sleep quality was affected by noise and self-report pain. The noise detected by digital sound level meter was middle level (47.34 ± 3.01dB), it equivalent to average noise levels in ICU which can reach 55 to 65 dB over a 24-hour period [35][36] . Noise has been considered to be a key factor of poor sleep in the ICU in some studies. It shows that noise is responsible for 11% to 24% of the total number of arousal [37] . However in our study, the recipients' subjective perception of sound was much lower than the Pain was known as one of most terrible experience in ICU and had a negative effect on the sleep quality. The result also showed that self-report pain was associated with sleep quality though pain detected by VAS at a relative low level. It means that in the state of sedation and analgesia, the patient still has a certain perception of pain, which may be related to the shallower analgesic intensity and kidney transplant recipients may be unable to communicate their level of pain if sedated and anesthetized.

Strengths and Limitations 13
As far as we know, this study was the first study to explore the sleep of kidney transplant recipients in ICU by using both subjective and objective methods in China. It can be seen that the objective outcome of the patient's sleep was not ideal, the subjective sleep score was still at a low level, and the degree of association between subjective and objective sleep was inconsistent, which was similar to the results of other studies. This suggests that the patient's response to subjective sleep is more in line with their sleep experience. In transplanted patients, difficult sleep is more likely to be connected with the existence of some psychological problems than with sociodemographic or biochemical parameters [8] .
Although patient's sleep in ICU is difficult to achieve by changing the physiological indicators that affect sleep, it may be effective to change the external physical environment in ICU and do some the psychosocial interventions.
Several limitations of this study should be considered when interpreting our results.
First of all, admittedly, our study sample was small. When analyzing the influencing factors of sleep, some indicators such as hemoglobin, hormonal dosage were not included in the regression equation. However, these were related factors affecting sleep in other studies. Molnar et al [38] reported that Hb was significantly correlated with AHI and was higher in patients with apnea. A potential explanation for this association is the recurrent reduction of oxygen saturation caused by apnea, which stimulates erythropoietin production. Corticosteroids are well-known to cause sleep disturbance, insomnia, and unpleasant dreams, which insomnia is the most frequently reported sleep-related side effect of steroids and immunosuppressive medications [39] . As this study included these indicators but did not produce a dominant result, this may be related to the sample size of this study is too small. 14

Conclusions
The quantity and quality of kidney transplant recipients sleep in ICU by using both objective and subjective assessment methods were not in consistency, while the subjective sleep quality based on RCSQ score better than objective one measured by PSG. However, sleep disruption always have remained a substantial problem and affected by self-reported noise and pain. Importantly, though the RCSQ is easy and simple to used for patients in ICU, the current study has questioned about the association between PSG and RCSQ score. Therefore, self-reported sleep of patients