Risk factors of sleep disturbances in elderly patients after thoracic surgery: a preliminary clinical trial

Background To identify the risk factors of sleep disturbances in elderly patients after thoracic surgery. Methods We enrolled 200 patients, all aged > 65, who underwent surgery and had American Society of Anesthesiology physical status II–III. We recorded general information, surgical diagnosis, type of operation, surgical duration, bleeding, nerve block, and dexmedetomidine dose given by controlled intravenous analgesia (PCIA). We used the Pittsburgh Sleep Quality Index (PSQI) at the end of PCIA to evaluate subjective sleep quality; we also recorded postoperative pain, nausea, and vomiting. We divided patients into a non-sleep disturbances group and a sleep disturbances group with PSQI 5 as the cutoff.


Background
Insomnia is a subjective experience in which sleep time and sleep quality are not satis ed even though there are appropriate sleep opportunities and sleep environment, affecting social function during the day. Studies showed that, in the previous month, 5.4% of Chinese people experienced varying degrees of sleep disturbances, 25% met diagnostic criteria for insomnia, and 50% over the age of 65 suffered from insomnia [1,2]. Insomnia affects daytime mental state, including fatigue, depression, or irritability, and reduces life quality. It can also lead to cognitive impairment, reduce work e ciency, alertness, and ≥ judgment, especially for postoperative sleep disturbances. Postoperative sleep disturbances include insomnia, hypersomnia, narcolepsy, changed sleep structure and increased frequency waking [3,4].
Postoperative sleep disturbances caused by general anesthesia might increase postoperative complications such as postoperative fatigue, severe anxiety and depression, delirium, and even severe stroke [5][6][7].
There are few studies of postoperative sleep disturbances in elderly patients. The aging of China's population and their particular medical environment will increase postoperative sleep disturbances in elderly patients. We aimed to identify the risk factors of sleep disturbance in elderly patients after thoracic surgery to provide reference for the development of effective preventive measures.

Characteristics of participants
We enrolled patients classi ed as American Society of Anesthesiologists status or , were aged 65 years and older and were scheduled for elective thoracic surgery. We excluded patients with communication disturbances, inability to cooperate (e.g., language comprehension disturbances, mental illness), unstable angina pectoris or myocardial infarction within 3 months, New York Heart Association grade ≥3, severe cardiovascular and cerebrovascular diseases, those who underwent an unplanned second operation in the post-anesthesia care unit (PACU), those transferred to intensive care unit (ICU) after surgery, and those transferred from the ward to ICU after surgery.
We recruited patients in the PACU admission after August 2020. A nurse anesthetist and a PACU nurse scrutinized the medical records. During the reviewer training, the study protocol, de nitions, and the case record form were explained. Initially, both reviewers practiced completing the case record forms for 20 cases together. Then, each reviewer independently examined half of the remaining 180 medical records to prepare a form for each case. Each reviewer subsequently checked every 20 th form completed by the other reviewer, and a third reviewer (an anesthesiologist) randomly checked 60 of the total 200 case record forms. Finally, 200 forms were converted to electronic data for input to PASW SPSS Statistics for Windows ver. 26.0 (SPSS Inc., Chicago, IL, USA). We recorded the following: age; sex; smoking history; drinking history; chronic insomnia; hypertension; diabetes; American Society of Anesthesia classi cation; body mass index (BMI); surgical diagnosis; type of operation; surgical duration; bleeding; nerve block; whether dexmedetomidine was included in PCIA; postoperative pain, nausea and vomiting; and the patient's subjective sleep quality evaluated by Pittsburgh Sleep Quality Index (PSQI) after 48 hours.

Preoperative preparations and anesthesia protocol
No patients received pretreatment before admission to the operating room. All patients were monitored using electrocardiography, pulse oximetry, invasive blood pressure recordings, and bispectral index. The anesthesiologist determined whether to perform PCIA with dexmedetomidine or thoracic paravertebral block after surgery, depending on the patient's condition. Subsequently, we administered a comprehensive evaluation of satisfaction with analgesia from patients and their family members at the end of PCIA (very satis ed; satis ed; general; dissatisfaction). We grouped very satis ed and satis ed as satisfaction. We considered this group as having no pain after surgery; If patients reported general or dissatisfaction, we classi ed them as dissatis ed, and de ned this group as having postoperative pain. We used the Pittsburgh Sleep Quality Index (PSQI) questionnaire to evaluate their subjective sleep quality, including sleep quality, sleep time, sleep time, sleep e ciency, daytime dysfunction, sleep disturbances, and hypnotic drug use. The scale consists of nine questions, with 18 items, and each index is scored 0-3. The total score of each index is 21 points [8]. When the PSQI score is greater than or equal to 5, the patient is de ned as having sleep disturbances. We divided patients into a non-sleep disturbances group and a sleep disturbances group.

Sample size calculation
The sample size was calculated with "Power and Sample Size.com," an online power and sample size calculator. The primary objective of this study identi ed the risk factors of sleep disturbances in elderly patients after thoracic surgery. The calculation of the sample size was based on the pilot study nding that the incidence of sleep disturbances in elderly patients after thoracic surgery was approximately 45%.
It was thus assumed that the incidence of sleep disturbances in elderly patients after thoracic surgery was 60%. To detect a clinically signi cant difference with a power of 80%, an alpha error of 0.05 (twosided), a total of 87 patients were required. To compensate for data that may be found to be missing during the retrospective review, the sample size was adjusted upward to 200 cases.

Statistical analysis
The data analysis was performed using the independent sample Student t-test, the Pearson's chi-square test, and the Mann-Whitney U-test, and it employed the program PASW SPSS Statistics for Windows ver. 26.0 (SPSS Inc.). All descriptive data were presented as case (%), mean ± standard deviation, and median (interquartile range). Univariate analyses of each factor were performed to calculate the p-value, OR, and 95% con dence interval (CI). Multiple logistic regression analysis was performed by choosing the factors with a p-value of <0.10; p<0.05 was considered to be signi cant.

Results
In total, 200 patients met the inclusion criteria, and 16 patients declined to participate. We enrolled a total of 174 patients, ve of whom developed delirium. The nal analyses included 169 patients (Fig. 1).
The incidence of sleep disturbances in elderly patients after thoracic surgery was 45%. We performed univariable and multivariable analyses by selecting factors whose p-values were < 0.10: age, chronic insomnia, hypertension, diabetes, BMI, surgical diagnosis, type of operation, surgical duration, bleeding, nerve block, whether dexmedetomidine was included in PCIA, and postoperative pain (Table 1). We performed multiple logistic regression analysis by selecting factors whose p-values were < 0.10: chronic insomnia, diabetes, BMI, surgical diagnosis, type of operation, surgical duration, bleeding, nerve block, dose of dexmedetomidine in PCIA, and postoperative pain. Only nerve block (p < 0.001; OR, 0.085; 95% CI, 0.046-0.155) and dexmedetomidine (p = 0.003; OR, 0.772; 95% CI, 0.609-0.979) were marginally signi cant preventive factors for sleep disturbances in elderly patients after thoracic surgery. Chronic insomnia, BMI, diabetes, surgical diagnosis, type of operation, surgical duration, bleeding, and postoperative pain were independent risk factors of postoperative sleep disturbances in elderly patients (Table 2).

Discussion
Postoperative sleep disturbances are manifestations of postoperative brain dysfunction, especially for elderly patients, and they are among the important risk factors for delirium, which had gradually attracted attention. Studies con rmed that extensive surgical trauma often leads to postoperative sleep disturbances, increasing the possibility of postoperative cognitive impairment [9]. The expansion of surgical incisions and the complexity of surgical types, the amount of bleeding increased, and the operation time prolonged all increase the degree of severity of surgical trauma of elderly patients; these factors signi cantly increase the incidence of sleep disturbances in elderly patients. The operation time was associated with the duration of postoperative sleep disturbances, and this may have been due to the extensive surgical trauma and the severity of the patient's condition [10].
In the present study, we found the perioperative application of thoracic paravertebral block, dexmedetomidine, and reducing postoperative pain reduced the incidence of postoperative sleep disturbances in elderly patients; postoperative pain is the most harmful factor associated with postoperative sleep disturbances [11]; pain can prolong sleep latency and reduce total sleep time, while sleep disturbances increase pain sensitivity and reduce pain threshold, including pain experienced the following day. The degree of pain is also predicted by postoperative sleep quality [12]. General anesthesia disrupts sleep/wake-up cycles and other circadian rhythms [13], leading to postoperative fatigue, severe anxiety and depression, delirium, and even severe stroke [5,7]. Volatile general anesthetics, including sevo urane, iso urane, and halothane, may cause short-term sleep disturbances and fragmentation [14]. Dexmedetomidine can induce bionic sleep [15], that is, natural sleep, and it does not produce dependence, tolerance, or addiction [16]. For elderly patients with minimal trauma, the application of total intravenous anesthesia combined with dexmedetomidine may reduce the incidence of postoperative sleep disturbances.
Sleep disturbances are associated with increased age [17]. In the present study, we showed that people with preoperative insomnia were more likely to have postoperative sleep disturbances. The reason may be that melatonin levels in the elderly are signi cantly lower than those of young people [18]. The physiological reserve of elderly patients decreases, and the prevalence of in rmity and comorbidities increases, all of which may affect the brain and central nervous system, leading to postoperative sleep disturbances [19]. Diabetes mellitus is a risk factor for long-term use of opioids and persistent pain after surgery; patients with diabetes are in chronic systemic proin ammatory states, characterized by high concentration of the in ammatory cytokine IL-6; persistent immune responses related to host defense prolong peripheral sensitization leading to pain hypersensitivity [20], and this further aggravates postoperative sleep disturbances.
We found that many factors in uenced postoperative sleep disturbances, including surgical anestheticand patient-related factors. Other studies showed that postoperative sleep disturbances might be affected by psychological and postoperative environmental factors such as ward lighting and sound; elderly patients are a special population. We should comprehensively consider the factors associated with their postoperative sleep disturbances.
Some limitations of this study should be noted. First, postoperative sleep disorders are affected by many factors, including surgery, anesthesia and patient factors, as well as environmental factors which was not noticed by this study. Second, our study did not continue to track the recovery of patients with postoperative sleep disorders, which is the main direction of our future research.
Chronic insomnia, BMI, diabetes, surgical diagnosis, type of operation, surgical duration, bleeding, and postoperative pain are independent risk factors of postoperative sleep disturbances in elderly patients. Nerve block and the dose of dexmedetomidine in PCIA are protective.

Consent for publication
All copyright owners agree to submit the paper to BMC Anesthesiology for publication. The author agrees that the copyright and related property rights of the paper shall be transferred to BMC Anesthesiology from the date of o cial publication of the paper.