To investigate perioperative sleep quality of total knee arthroplasty (TKA) patients managed with enhanced recovery protocol and analyze its risk factors to provide scientific base for clinical intervention in future.

A general survey, Pittsburgh sleep quality index (PSQI), Epworth sleepiness scale(ESS), visual analogue scale(VAS) score, the center for epidemiologic studies depression scale(CESD) and Zung Self-rating anxiety scale(ZSAS) were employed to measure sleep quality in 118 TKA patients who fulfilled the inclusion and exclusion criteria at preoperative two weeks, the day of surgery and postoperative one, two, four, eight, and 12 weeks. The influencing factors were analyzed by multivariate logistic regression. Inclusion criteria: older than 18 years; unilateral elective primary TKA. Exclusion criteria: the patients were undergoing revision surgery or bilateral TKA; regularly using sleep-related medication or having mental diseases.

The prevalence of sleep disturbance was as high as 53.4% preoperatively, which peaked 87.3% on the day of operation, then began to decline, but re-peaked two weeks after operation, and then rapidly decreased. PSQI and ESS score significantly increased postoperatively from baseline to two weeks before decreasing rapidly to a lower level than preoperative baseline at four weeks postoperatively(t=7.136, 5.269, both P<0.001) and a positive correlation was found between ESS and PSQI scores(r=0.380, P<0.001). VAS score decreased from baseline over all time points in the postoperative period and a positive correlation was found between VAS and PSQI score(r=0.198, P=0.032). Multivariate logistic regression analysis demonstrated that pain, anxiety and depression were three risk factors associated with postoperative sleep disturbance[odds ratio(OR)=1.384, P=0.007; OR=1.260, P=0.038; OR=1.203, P=0.049].

PSD existed in more than half of the patients undergoing TKA perioperatively which may worsen afterwards and would last longer than two weeks. Pain, anxiety and depression are the three independent risk factors.