Postoperative REM sleep inhibition: Association with pain or cortisol?

Objectives: While severe sleep disturbance is common after surgery, the etiology of this sleep disturbance is poorly understood.1 Pain is most often cited by patients as the reason for their poor postoperative sleep, but this relationship has not been rigorously examined.2 The surgical stress response might contribute to postoperative sleep disturbance since administration of exogenous cortisol to healthy volunteers suppresses REM sleep.3 This ongoing study is will test the hypothesis that postoperative REM sleep inhibition is associated with pain scores and urinary free cortisol levels. Methods: Following IRB approval and informed consent, 7 patients undergoing surgery for benign gynecological conditions through a low abdominal incision were studied. Patients received general anesthesia supplemented with epidural fentanyl or bupivacaine. Postoperative analgesia was provided by epidural infusion of either fentanyl or bupivacaine. Sleep was monitored using standard polysomnography on the preoperative night and the first three postoperative nights from 10 p.m. until 6 a.m. Pain scores (VAS) at rest and with coughing were obtained immediately before and after the sleep period. Overnight pain score was estimated by the average of the evening and morning scores. The urinary free cortisol levels were measured using RIA and expressed as ug/hr. Sleep was scored using the Rechtschaffen and Kales criteria in 30 second epochs to determine the percentage of the recording time spent in REM sleep. All values are expressed as mean + SD. Results: The % REM sleep decreased from 8.2 % ±6.0 on the preoperative night to 0.1 % ±0.15 on the first postoperative night (p<0.01 Tukey/Kramer multiple comparison procedure). The % REM sleep was not different from baseline on postoperative nights 2 and 3. Pain at rest was not significantly different on any of the postoperative nights (one-way ANOVA for repeated measures p = 0.31). Pain with coughing was significantly worse on postoperative night 2 than on postoperative night 3 (Tukey/Kramer multiple comparison procedure, p < 0.01), but pain on postoperative night 1 was not different from pain on postoperative nights 2 or 3. Urinary free cortisol levels were 3 times higher on postoperative night 1 than on postoperative nights 2 or 3 (one-way ANOVA for repeated measures with post hoc Tukey/Kramer multiple comparison procedure, p < 0.05). Discussion: These data confirm previous observations of profound suppression of REM sleep in postoperative patients. It is interesting that no association was found between REM sleep disturbance and either pain at rest or with coughing, as this association is commonly drawn by patients. REM sleep suppression on the first postoperative night was associated with a three-fold elevation of urinary free cortisol, suggesting either a possible direct inhibition of REM sleep by the surgical stress response or an indirect inhibition of REM sleep by disruption of the circadian rhythm.