Sleeping Patterns in Patients Undergoing Total Knee Arthroplasty: A Prospective Study Using a Wearable Device

Sleep disturbance is a common, poorly understood complaint following total knee arthroplasty (TKA). We characterized sleep patterns in a cohort of TKA patients.

We prospectively monitored sleep quality in 93 patients who underwent primary, uncomplicated TKA for primary osteoarthritis. Outcomes included the Pittsburgh Sleep Quality Index (PSQI) recorded two weeks preoperatively to six weeks postoperatively and six objective sleep quality variables measured with a wearable device. Subjective and objective outcomes were compared to the 2-week preoperative baseline using mixed-effects linear regression models. The potential association of PSQI with clinically relevant variables, including having a history of sleep disorder, Visual Analog Scale (VAS) pain scores, and opioid consumption, was assessed.

The highest mean PSQI scores, indicating worse perceived sleep, were observed between the second and fourth postoperative weeks (10.6 and 10.4, respectively). The worst sleep quality detected by the wearable device was during postoperative week one (mean change compared to two weeks preoperatively in sleep score = -7.7), with a progressive trend to normalizing by week six (mean change sleep score= -4.0). When patients who did and did not have sleep disturbances were compared, we observed neither any significant differences in subjective and objective sleep outcomes nor in opioid consumption patterns. The linear mixed-effects model demonstrated a significant relationship between PSQI and VAS over the study period, estimating a 0.33-point increase in PSQI (indicating worse perceived sleep quality) for every 1-point increase in VAS pain score (mixed-effect linear regression model estimate: 0.33; 95% confidence interval: 0.17 to 0.48; P < 0.001).

We delineated sleeping patterns in TKA patients utilizing patient- and wearable-reported data. Patients perceived the worst sleep quality during the second and fourth postoperative weeks, whereas objective data indicated the worst sleep in the first week. Patients who have a history of sleep disorders did neither report nor objectively experience worse sleep patterns perioperatively.