Patient selection and data extraction
The patients were enrolled between September 2019 and January 2020. The study was approved by the Regional Ethical Review Committee at the Beijing Chao-Yang Hospital, Capital Medical University approved this retrospective study (2019-ke-273).
Patients were included according to the following inclusion criteria: (1) 18 years or older;(2) American Society of Anesthesiologists (ASA) physical status Ⅰ-Ⅲ;(3) undergoing elective non-cardiac surgery with general anesthesia at Beijing Chao-Yang Hospital. Exclusion criteria were as follows: (1) history of dementia or severe cognitive impairment;(2) history of the psychiatric disease;(3) history of organic sleep disorder ;(4) taking antipsychotic medication; (4) history of drug abuse.
Data collection
We collected the patient's general information and preoperative clinical data including gender, age, body mass index (BMI), smoking history, alcohol consumption, years of education, preoperative comorbidities (coronary artery disease, hypertension, diabetes), preoperative Visual Analog Scale (VAS) score. We also reviewed and analyzed patients’ assessment of depression, anxiety, and insomnia using the Insomnia Severity Index (ISI), Patient Health Questionnaire-9 (PHQ-9), Generalized Anxiety Disorder-7 (GAD-7) respectively one day before surgery.
The surgery and anesthesia data including ASA physical status classification, type of surgery, duration of surgery, anesthesia types, and postoperative analgesia were also collected. We included postoperative data including VAS scores in the postoperative anesthesia recovery room (PACU), 24, 36, 48, 60, 72, and 84 hours after surgery respectively. Besides, we recorded postoperative nausea and vomiting (PONV), rescue analgesia, and length of hospital stay (LOS). The assessments of ISI, PHQ-9 and GAD-7 were extracted from the paper-based questionnaire. Other data were extracted from electronic medical records.
Exposure of interest
For this study, we identified the mental disease according to the scores of ISI, PHQ-9 and GAD-7. The ISI used a cut-off value of 8 points, with a score ≥ 8 points being insomnia and less than 8 points being normal[18]. And 5 points was used as the cut-off value for both the PHQ-9 and GAD-7, with a higher score indicating moderate to severe depression or anxiety[19,20]. Patients were separated into non-mental diseases group and mental diseases group (with anxiety, depression or insomnia) based on the assessments.
Outcomes
We compared the effect of preoperative mental disease on postoperative pain. And propensity score matching (PSM) analysis was performed to reduce selection bias. The primary outcome was postoperative pain according to VAS 24h after surgery. Secondary outcomes included VAS at other time points and incidence of PONV, rescue analgesia, and LOS.
Statistical analysis
Statistical analysis was performed with IBM SPSS Statistics version 25.0 (IBM Corp., Armonk, NY, USA). Continuous variables are expressed as a median and interquartile range for non-normal distribution variables and mean values with standard deviation for normal distribution variable. Categorical variables were expressed as counts and percentages. We compared the categorical variables with the chi-square test or Fisher's exact test. Levene's test was used to test for equality of variances of continuous variables. Independent t-tests were performed for continuous variables with symmetric distributions. And the Mann - Whitney U test was used to compare continuous data with the asymmetric distribution.
We performed PSM analysis using multivariable logistic regression model by R (R Foundation for Statistical Computing, version 3.5.1) to build matched groups of patients to compare postoperative acute pain between patients with or without any of mental diseases. Variables were identified in the present study. These variables used for matching included ASA physician status, surgery site, preoperative pain. The matched group of patients from non-mental disease group were in Group A and the matched patients from mental disease group were in Group B. To investigate the effect of each mental diseases (anxiety, depression and insomnia), PSM were used which variables were also included ASA physician status, surgery site, preoperative pain. When we analyzed 1 mental disease, the other 2 mental diseases were also included in variables in order to minimize their effects. Matching analysis was performed based on each patient's estimated propensity score using a nearest neighbor 1:1 matching with a caliper range of 0.2. We tested the patients demographic and perioperative characteristics of patients before and after matching to ensure well-matching. Two-sided P values of less than 0.05 were regarded to be of statistical significance.