Sleep Disturbances after Posterior Scoliosis Surgery with an Intraoperative Wake-up Test Using Remifentanil

The study was approved by the institutional review board at the University of Muenster, Muenster, Germany (reference number 2008–095-f-S).

The quality of sleep was assessed using the Pittsburgh Sleep Quality Index (PSQI) questionnaire. The PSQI records seven components of sleep quality, retrospectively over a period of 4 weeks: the frequency of sleep-disturbing events, self-assessment of sleep quality, the usual sleeping times, sleep latency and sleep duration, use of sleeping medication, and tiredness during the day. Nineteen self-rated questions, each one scored from 0 to 3, are combined to form seven corresponding component scores, each one ranging from 0 to 3. The global score is generated by summing up all seven component scores and ranges from 0 to 21, with higher values corresponding to reduced sleep quality. There is an empirical cutoff value of 5 that makes it possible to classify patients into “good” and “poor” sleepers. In the first description of the PSQI, healthy individuals who were not awaiting surgery had a PSQI score of 2.67 (± 1.7).12,13The following statistical data are available in the literature on the reliability and validity of the PSQI questionnaire: The stability of the PSQI has been tested in three studies and assessed as satisfactory, with values between 0.82 and 0.89 for the overall PSQI score. The internal consistency and homogeneity of the overall score are 0.77. The sensitivity of the overall PSQI score for various samples of sleep-disturbed patients was always higher than 80% (80–100%), and the specificity was similarly high (83–87%).12,14–16Various studies used the PSQI score in preoperative or postoperative patients before.17–19The PSQI questionnaire is included in the current article as  appendix 1, and the scoring instructions are included as  appendix 2.

After receiving detailed information and providing written informed consent to participate in the study, the patients completed a PSQI questionnaire preoperatively, as well as 3 months, 6 months, and 1 yr postoperatively. The encoded questionnaires were analyzed by an independent investigator unaware of the study protocol. In addition, data were gathered on patients’ age, weight, and sex, as well as on the duration of surgery and anesthesia, the vertebral levels operated on, the amount of blood loss, the opioid used, and its dosage. In addition, the wake-up test times were documented. The patients were also explicitly asked on the first day after the operation whether they had any memory of the wake-up test.

The patients were divided into three groups according to the type of surgery performed and the anesthesiologic standards at our hospital, i.e. , posterior instrumentation with an intraoperative wake-up test, using remifentanil (PR group, n = 32), and anterior instrumentation without an intraoperative wake-up test, using sufentanil (anterior group, n = 40). After analysis of the initial results, another group with posterior instrumentation, an intraoperative wake-up test, and sufentanil (PS group, n = 19) was also investigated.

Anesthesia was performed in all groups in accordance with the following protocol: On the morning of surgery, patients received 0.3 mg/kg midazolam (Hoffmann-La Roche AG, Grenzach-Wyhlen, Germany) orally for premedication. Anesthesia was induced with 5 mg/kg thiopental (Inresa Arzneim Ltd., Freiburg, Germany), 0.3 μg/kg sufentanil (Janssen-Cilag Ltd., Neuss, Germany) in the anterior and PS groups or 0.5 μg · kg⁻¹· min⁻¹remifentanil (GlaxoSmith Kline Ltd., Munich, Germany) in the PR group, and 0.15 mg/kg cisatracurium (GlaxoSmithKline Ltd.). Anesthesia was maintained using the volatile anesthetic sevoflurane (Abbott Ltd., Wiesbaden, Germany) with an end-tidal concentration of 1.5–3.0 vol% and—depending on the group assignment—intermittent sufentanil administration in the anterior and PS groups or a continuous intravenous infusion of remifentanil in the PR group (0.2–0.5 μg · kg⁻¹· min⁻¹). The dosage of each opioid was guided by the patient’s individual requirements. The procedure was performed during controlled hypotension (mean arterial pressure 50–65 mmHg) in all groups.20If necessary, the α-inhibitor urapidil was used to guarantee this range of mean arterial pressure.

The wake-up test was performed in accordance with the following protocol: At the surgeon’s request, anesthesia was reduced. In the PR group, the remifentanil infusion was stopped. Sevoflurane was decreased by stopping the supply and increasing the oxygen flow to 4 l/min in the PR and PS group until spontaneous breathing occurred and the patient followed the instruction to move the extremities. Immediately after the surgical team had checked intact voluntary movement of all extremities, anesthesia was deepened with an intravenous administration of propofol (Fresenius Kabi Ltd., Bad Homburg, Germany; 1–2 mg/kg) and sufentanil (0.3 μg/kg) or remifentanil (1 μg/kg). In addition, 2–3 mg midazolam was administered intravenously. The wake-up test times were recorded.

After the wake-up test, anesthesia was continued as described above. Postoperatively, the patients were moved to the recovery room and then discharged to the peripheral wards after having met the criteria for discharge, i.e. , consciousness, hemodynamic stability, sufficient breathing, as well as unimpaired swallow and cough reflexes. After posterior surgery, patients were mechanically ventilated in the recovery room for 2 more hours. All patients were treated according to the following pain management protocol: 20 mg/kg paracetamol was administered orally four times per day. In addition, 0.2 mg/kg piritramide was administered intravenously every 6 h from day 1 until 3 after surgery. The dosage was reduced to 0.1 mg/kg on the fourth day. Piritramide was stopped at the fifth postoperative day.

Statistical analysis was performed using SPSS for Windows, version 13.0.1 (November 2004; SPSS, Inc., Chicago, IL). Categorical variables are expressed as frequency and percentage, whereas continuous variables and PSQI scores are represented as median and interquartile range [25th percentile; 75th percentile]. Before statistical testing, each continuous variable was analyzed exploratively for its normal distribution (using the Kolmogorov–Smirnov test). The Mann–Whitney test was applied for comparison of nonparametric variables between two study groups. The nonparametric patients’ baseline characteristics and PSQI scores were assessed using the Kruskal–Wallis test. The Friedman signed rank test was used to compare the nonparametric time-dependent variables. The chi-square test was performed for comparison of categorical variables.

Pairwise multiple comparisons after the Friedman and Kruskal–Wallis tests were performed using the Miller “simultaneous statistical inference” procedure.21,22The rank average differences for any pair were compared with the critical constant. The null hypothesis was rejected if the rank average difference of any pair exceeded the critical constant. This meant that the difference between the compared samples was significant. Differences were considered as statistically significant at P < 0.05.

There were no differences between groups with regard to age, weight, sex, or wake-up test times (table 1). The duration of surgery and anesthesia was significantly shorter in the anterior group as compared with the posterior groups. The opioid dosage, the number of vertebral levels operated on, and the amount of blood loss were also significantly lower in the anterior group. The patients included in the current study did not report any symptoms or memories of the wake-up test. All of them felt comfortable with the postoperative pain management.

Preoperatively, there were no differences between groups with regard to PSQI values (anterior: 3 [3; 5]; PS: 4 [2; 6]; PR: 4 [3; 5]; P = 0.828); median values were below the cutoff value of 5. The patients in the PR group had a significantly diminished quality of sleep up to 6 months postoperatively in comparison with the situation before the surgical procedure (P < 0.05). The maximum PSQI value was observed at the first postoperative questionnaire 3 months postoperatively (7 [6.25; 9]). After 1 yr, the PSQI value decreased below the cutoff value of 5 again and thus into the range of the preoperative values (4 [3; 6.75]). The detailed PSQI values for each patient within the PR group are provided in figure 1.

Among the patients in the anterior group, there were no significant differences in sleep quality in comparison with the baseline values during the entire 1-yr follow-up period. After 3 months, there was a trend toward a slight increase in PSQI values, although they did not increase above the cutoff value of 5 (4 [3; 5]). However, values returned below baseline values after 1 yr (3 [2; 5]). After 3 and 6 months, the PSQI values in the PR group were significantly higher as compared with those in the anterior group (P < 0.05).

No significant differences in PSQI values were observed in the PS group during the 1-yr follow-up period compared with preoperative values (4 [2; 6]; P = 0.772). The median PSQI score was always below the cutoff value of 5. In comparison with the anterior group, no significant differences in PSQI values were seen at any time point. When the results for the PS group were compared with those for the PR group after 3 and 6 months, PSQI values in the PS group were found to be significantly below those in the PR group (P < 0.05). The PSQI scores in the various groups at the individual questionnaire time points are shown in figure 2. Detailed information about the results within the different components of the PSQI questionnaire is presented in  appendix 3.

The current study demonstrates that patients who have undergone posterior instrumentation spinal fusion with an intraoperative wake-up test using remifentanil slept significantly more poorly up to 6 months postoperatively than did patients with an anterior instrumentation without a wake-up test. Notably, postoperative sleep disturbances after the wake-up test were reduced to values noticed without any wake-up test when sufentanil instead of remifentanil was used for intraoperative analgesia. These data strongly suggest that not only the wake-up test itself, but also the opioid choice may contribute to postoperative sleep disturbances in patients undergoing scoliosis surgery.

Comparing the results of the anterior and the PR group, the wake-up test may be a cause for the impaired postoperative sleep quality after posterior scoliosis surgery. Therefore, the renunciation of the wake-up test might represent a way of prevention. There are alternative procedures for neurologic monitoring, such as somatosensory evoked potentials and motor evoked potentials.23However, the use of these methods is associated with much higher technical effort and costs than the wake-up test.

In addition, the group differences associated with the two different surgical approaches have to be considered. During anterior surgery, the convex side of scoliosis is compressed. In contrast, traction is applied to the dura mater and spinal cord when the vertebral column is distracted during posterior corrective spinal fusion.5This meningeal stress as well as the more severe surgical trauma in the PR group, reflected by an increased blood loss and a higher number of vertebral levels operated on as compared with the anterior surgical procedure (table 1), may also in part be responsible for the postoperative sleep disturbances. However, these theses are only speculative, because no data are available regarding these issues. Because no differences in the PSQI values occurred between the posterior group, which received sufentanil (PS group) instead of remifentanil, and the anterior group at any time point, the wake-up test is probably not the only cause for the diminished postoperative sleep quality in the PR group.

The improved postoperative sleep quality in patients of the PS group compared with the PR group suggests that the choice of opioid may contribute to the patients’ postoperative sleep quality, because the two groups differed only in the analgesic agent used. Direct inhibiting effects of central acetylcholine release by opioids can be a cause of sleep disturbances. In this context, it has been reported that opioids, including remifentanil,24may contribute to a reduction in deep sleep and rapid eye movement sleep.25,26One potential mechanism for this is inhibition of acetylcholine release in the medial pontine reticular formation, contributing to arousal state disruption and impaired memory function.27However, this effect has been described for morphine and fentanyl, but not for remifentanil.28 

Inadequate analgesia during the wake-up test might be another explanation for the poorer sleep observed in the PR group. It is known that despite low-dose remifentanil administration and unaltered hemodynamic parameters, increases in patients’ serum epinephrine and norepinephrine concentrations occur during the wake-up test in comparison with the time of the skin incision.29In addition, remifentanil has a very short, context-sensitive half-life of 3–4 min.30As a consequence, the opioid is no longer effective shortly after infusion stop for the wake-up test. By contrast, the effective concentration of sufentanil only declines by half after 33.9 min,10suggesting that even after discontinuation of the volatile anesthetic, adequate analgesia was present during the wake-up test in the anterior and PS groups.

In patients of the PR group, pain, which is not remembered explicitly because of the subsequent administration of propofol, midazolam, and remifentanil, might have promoted implicit memory. Andrade and Deeprose6have distinguished between explicit, retrievable, or conscious memory and implicit, nonretrievable, or unconscious memory. An absence of postoperative memory of the wake-up test thus does not exclude intraoperative consciousness.6Implicit memories can trigger sleep disturbances and even posttraumatic stress disorder.31It has been described that despite a lack of hemodynamic changes, some patients recalled colors that had been mentioned during the wake-up test after scoliosis surgery, without having any explicit memories of the wake-up test or the operation.32 

Kress et al.  33recently showed that continuously sedated patients in the intensive care unit had fewer posttraumatic stress disorder symptoms when they received daily “wake-ups.” At first glance, this seems to contradict the current results. However, these patients had factual memories of their stay. In contrast, our patients did not have any factual or explicit memories. In this context, the results of Jones et al.  34seem to be closer to the patients included in the current study. The authors reported that intensive care patients with delusional memories of their intensive care unit stay were more likely to have posttraumatic stress disorder–related symptoms than were those with factual memories. Because no tests for implicit memory were performed in this study, its influence on patients’ postoperative sleep remains uncertain.

The recommendation for remifentanil in anesthetic procedures involving intraoperative wake-up tests is justified by the significant shortening of the arousal time. However, in the studies concerned, either total intravenous anesthesia was used5,9or total intravenous management was compared with volatile anesthesia.8In the current study, no difference in wake-up test times between administration of combined sufentanil and sevoflurane and the combination of sevoflurane and remifentanil occurred.

A major limitation of this study is the lack of assessment of the level of sedation and analgesia during the wake-up test. However, it is a matter of controversy whether an adequate method for monitoring the depth of anesthesia exists currently.35,36The most commonly used method, the Bispectral Index, is also limited by its dependence on external factors such as the anesthetic agents used, the patient’s age, the position of the electrodes, and so forth.36A randomized, controlled, prospective study revealed no lower incidence of awareness with Bispectral Index monitoring (BIS®; Aspect Medical Systems, Newton, MA). The authors did not recommend this method as a part of standard practice.37In fact, there is not only interindividual but also intraindividual variability in sedation at the same Bispectral Index value.38Even at a Bispectral Index value of 40, explicit recall is described.39Monitoring the quality of analgesia also is extremely difficult. The questionnaires were disclosed for practical and ethical reasons, because the patients were intubated and had a large surgical wound at the moment of the wake-up test. Measurement of opioid blood concentrations during the wake-up test would provide only poor evidence, because there is wide interindividual variability in the response to opioids.40 

Further limitations include the absence of randomization and the later investigation of the PS group. Therefore, evident as well as unknown confounders might have influenced the results. Another point of criticism is that no investigation of an anterior group using remifentanil was performed in the current study. A posterior group without a wake-up test was not investigated, because the alternative technique of evoked potentials is not established at our institution. In addition, it might not be sufficient to disclose explicit recall by asking the patient on the first postoperative day, because explicit recall might occur after several days.41However, none of the patients reported having explicit recall afterward. The PSQI scores in this study are lower than those described in other trials (12.0 [10; 16],4212.5 (± 3.8)43). One difference is that this investigation describes long-term effects of one special event, whereas most of the other studies have been concerned with short-term effects or chronic diseases. Furthermore, the patients included in this study were younger and had no primary insomnia or diseases causing sleep disturbances. The administration of midazolam after the wake-up test must be critically questioned, too. More recent research results suggest a dissociated amnesia process for midazolam, i.e. , suppression of conscious memory only.44Information that is recorded via  routes that bypass consciousness can cause negative psychological consequences.45 

The findings of the current study suggest that the use of remifentanil in posterior scoliosis surgery with an intraoperative wake-up test may result in a higher incidence of postoperative sleep disturbances than the use of sufentanil. However, little is known about the complexity of the interplay between the choice of anesthetic, surgery, intraoperative awareness, and postoperative sleep disturbances. In view of these issues and the current study design, it is only possible to generate hypotheses on the basis of the current data. Further research is needed to determine the interaction between intraoperative wake-up tests and postoperative sleep disturbances in more detail and to clarify whether the choice of opioid affects postoperative sleep quality.