Subarachnoid Sufentanil for Early Postoperative Pain Management in Orthopedic Patients: A Placebo-controlled, Double-blind Study Using Spinal Microcatheters

After obtaining approval from the local ethics committee (Ärztekammer Hamburg) and written informed consent, 90 patients (table 1) undergoing elective lower-limb surgery, e.g. , hip and knee replacement, revision hip or knee arthroplasty, or major tumor resection with femoral or knee reconstruction, were included in the study. Exclusion criteria were coagulation disorders, acute neurologic diseases, and a history of acute or chronic consumption of any opioids or α²-adrenergic agonists (e.g. , clonidine).

Patients (n = 80) were orally premedicated with 7.5 mg midazolam (Hofmann-La Roche, Grenzach-Wyhlen, Germany) 1 h before arriving at the anesthesia department, where they were monitored by electrocardiogram, noninvasive blood pressure measurement, and pulse oximetry. A large peripheral intravenous cannula was placed, and an infusion of 500 ml lactated Ringer’s solution was started. CSA was performed with the patient in the sitting position (performed at L2–L3 in 5%, L3–L4 in 65%, or L4–L5 in 30% of patients). A 22-gauge Sprotte needle (Pajunk, Geisingen, Germany) was inserted into the subarachnoid space using the midline approach. When free flow of cerebrospinal fluid (CSF) had been obtained, a 28-gauge nylon catheter (Cospan; Kendall, Neustadt/Donau, Germany) was inserted 3 cm into the subarachnoid space. After the needle and the catheter stylet had been withdrawn, a microfilter was attached to the proximal end. After aspiration of CSF through the catheter, all patients received repetitive injections of 1 ml plain 0.5% bupivacaine to achieve an analgesic level (pin prick) of at least T12 before the operation. When the sensory block had been established, a central venous and urinary catheter were inserted, and the patients were transferred to the operating room.

During surgery, the analgesic level was assessed every 15 min and maintained constant (target T10, with a range between T8 and T12) by subarachnoid injections of 1 ml plain 0.5% bupivacaine every 90 min. Patients received oxygen via  face mask to maintain oxygen saturation greater than 95% and intravenous titration of midazolam at 1-mg bolus doses for sedation. In case of pain or discomfort during the operation, small doses of 0.025–0.05 mg fentanyl (Janssen-Cilag, Neuss, Germany) were injected intravenously. Blood loss less than 1 l was replaced by hydroxyethyl starch infusion and, in case of hemoglobin concentration less than 8 g/dl, by transfusion of autologous or allogeneic erythrocytes.

After completion of surgery, patients were transferred to the postanesthesia care unit. The protocol started when patients could move their legs and rated postoperative pain greater than 50 on a visual analog scale (VAS) ranging from 0 to 100 mm (0 = no pain; 100 = unbearable pain). According to a computer-generated randomization list, patients received a subarachnoid injection of 10 μg sufentanil (Janssen-Cilag), 5 mg plain bupivacaine in a dose of 2 ml of 0.25% bupivacaine (Astra, Wedel, Germany), a mixture of 5 μg sufentanil and 2.5 mg plain bupivacaine (mixture), or saline (placebo), respectively. All drugs were prepared as a 2-ml bolus dose in an unlabeled syringe by an anesthesiologist who was the only person who had access to the randomization list but was not otherwise involved in the study. The subarachnoid dose was injected through the catheter within 1 min by a research fellow who was blinded to the content of the syringes. After the bolus injection, the catheter was cleared by subsequent injection of 2 ml saline.

Data recorded immediately before the first postoperative subarachnoid injection were defined as baseline (time 0). The quality (VAS 0–100) and dermatomal level of analgesia (pin prick and cold swabs), the degree of motor block (Bromage scale: 0 = none, full flexion of both legs against gravity; 1 = partial, patient is able to move feet and knees but is not able to elevate his legs against gravity; 2 = almost complete, patient is able to move feet but not knees; 3 = complete, patient is unable to move feet or knees), hemodynamics (heart rate, mean arterial pressure), and respiratory parameters (respiration rate, oxygen saturation by pulse oximetry) were determined at 5, 15, 30, and 45 min after injection by an anesthesiologist who was also blinded to the applied study drugs.

At each measurement interval and again 120 min after the injection, 5 ml of central venous blood and 1 ml of CSF were taken for measurements of the sufentanil concentration in the sufentanil and mixture groups. The venous blood was centrifuged over 10 min at 4,000 revolutions per minute, and the supernatant serum was removed. All samples were immediately frozen at −20°C until the analyses were performed. The sufentanil concentration in plasma and CSF was measured by a special radioimmunoassay (Janssen-Cilag, Beerse, Belgium). The method of the radioimmunoassay has been described by Michiels et al . 15The lowest detectable sufentanil concentration was 0.02 ng/ml.

In case of recurrent pain (VAS > 25), patients received a subarachnoid injection of 1 ml 0.25% plain bupivacaine on the condition that at least 45 min after the first subarachnoid drug injection had elapsed. Duration of analgesia was defined as the interval between the first (study drug) and the second (0.25% bupivacaine) subarachnoid drug injection.

The incidence of side effects such as nausea, vomiting, pruritus, and dizziness was registered by the same anesthesiologist who assessed the level and quality of analgesia, using a standardized questionnaire. This investigator also registered respiratory depression if the respiratory rate of the patient decreased below 8 breaths/min before the interview, and shivering if the patient showed any sign of muscular trembling. If the oxygen saturation measured by pulse oximetry decreased below 95%, patients were treated with oxygen insufflation (4 l/min) via  a face mask.

After the 80 patients completed the study, a second part of the study was added to evaluate the efficacy of repeated subarachnoid sufentanil injections for postoperative pain relief. Ten patients (table 1) undergoing major lower-limb surgery (hip or knee replacement) underwent the same protocol with microcatheter CSA as previously described. After admission to the intensive care unit, these patients received the first subarachnoid dose of 10 μg sufentanil when postoperative pain was assessed as greater than 50 on the VAS (0–100 mm). The quality of pain relief and motor function of the lower extremity were assessed by VAS and by the Bromage scale (see above) immediately before injection (0 min) and 5, 15, 30, 45, and 60 min after the injection of sufentanil into the microcatheter.

In case of recurrent postoperative pain greater than 25 on the VAS, patients received a maximum of three additional doses of 10 μg sufentanil each. When the maximal dose of sufentanil (total dose, 40 μg) was reached, patients with pain were treated by injection of 1 ml 0.25% plain bupivacaine. The incidence of side effects such as nausea, vomiting, pruritus, dizziness, respiratory depression (respiratory rate < 8 breaths/min), and shivering was registered at the respective times of measurement by an anesthesiologist (in accordance with part 1). The microcatheters were withdrawn on postoperative day 1.

All 90 patients were visited by an anesthesiologist on the morning of postoperative day 2 and before they left the hospital and were asked about postdural puncture headache or any neurologic sequelae.

Normally distributed data were described as mean ± SD. Median (range) was used if nonnormality can be assumed or if data are censored. Categoric data (table 1: sex, American Society of Anesthesiologists class) are described as frequency distribution. Differences within groups were tested by analysis of variance for repeated measurements and post hoc  paired Student t  test. Differences between groups were tested by analysis of variance and the Scheffé F test in case of parametric and normally distributed data. If data were not normally distributed, the Kruskal-Wallis test was applied. Nonparametric data were tested by the chi-square test.

Data of part two of the study were tested by analysis of variance for repeated measurements and post hoc  comparison by paired Student t  test (after each sufentanil injection) and by the Kruskal-Wallis test comparing the four repetitive injections. Nonparametric data were tested by the chi-square test with Fisher exact test for small samples. All differences were considered significant at P  less than 0.05.

Morphometric data, time of surgery, and intraoperative drug administrations were comparable among groups (table 1).

Although heart rate and oxygen saturation measured by pulse oximetry did not change in any group, the mean arterial pressure decreased below baseline in all groups with exception of the sufentanil group (table 2). In the bupivacaine and mixture groups, the respiration rate was lower after drug injection when compared with baseline, whereas the highest respiration rates were measured in the placebo group until the rescue medication was given.

Figure 1shows the VAS mean values that were significantly decreased in comparison with baseline (P < 0.01) after the first 5 min after drug injection in all groups except the placebo group. In groups 1–3, the VAS values decreased further until they reached their nadir 45 min after injection (P < 0.01 vs.  baseline). In the placebo group, the VAS mean values did not decrease below baseline before 45 min after saline injection. Five patients in the placebo group, whose data are not included in figures 1–3, received 1 ml 0.25% bupivacaine within 45 min after saline injection because of unbearable pain.

The mean duration of analgesia (pain on VAS > 25) was 128 ± 61 min in the sufentanil group, 146 ± 74 min in the bupivacaine group, 167 ± 78 min in the mixture group, and 47 ± 62 min in the placebo group (P < 0.01 placebo vs.  other groups). No patient required parenteral analgesics within the first 4 h postoperatively.

Figures 2 and 3show the level of analgesia and the degree of motor block during 45 min after the first subarachnoid injection. Patients with bupivacaine experienced the most cephalad dermatomal levels (median, T6; one patient with T1 at 45 min). In all treatment groups, the median analgesic level had its highest extension 30 min after injection. While patients in the sufentanil group did not show any detectable signs of motor block and patients with the mixture only had a short-term restriction of their motor function after 15 min (Bromage 1), patients in the bupivacaine group experienced more profound motor block (Bromage 2) 15–45 min after the injection (P < 0.05 vs.  other groups).

Figure 4shows the changes of the sufentanil concentrations in plasma and CSF in groups 1 and 3. While plasma concentrations of sufentanil were below or slightly above the level of detection after subarachnoid injection of 10 μg sufentanil (the maximal mean value after 15 min was 0.027 ng/ml), the mean levels were below the threshold of 0.02 ng/ml after injection of 5 μg sufentanil. The CSF concentrations of sufentanil peaked 5 min after injection of either 10 μg (mean, 183 ± 167 ng/ml; maximum, 623 ng/ml) or 5 μg (mean, 120 ± 102 ng/ml; maximum, 328 ng/ml) sufentanil and declined continuously over the following 60 min.

Patients in the placebo group showed a higher incidence of early postoperative nausea compared with patients in other groups (table 3). Patients in the sufentanil group had the highest incidence of pruritus, which was only moderate and did not require any treatment. Respiratory depression (respiration rate < 8 breaths/min) occurred in one patient in the sufentanil group 15 min after injection and lasted for 30 min, and occurred in two patients in the mixture group also 15 min after injection and lasted 10 min. All patients had oxygen insufflation and did not experience oxygen desaturation less than 95%. The respiratory depression was easily treated by asking the patients to breathe deeply.

The demographics of the 10 patients in part 2 were comparable to the preceding patients (table 1). Each sufentanil injection provided immediate and profound pain relief, as shown in figure 5. While the quality of analgesia was not different among the four different injections at any time (P > 0.2), the duration of analgesia was significantly longer after the first injection of sufentanil (198 ± 70 min) when compared with the following injections (second, 115 ± 72 min; third, 114 ± 55 min; fourth, 102 ± 52 min;P < 0.001, first vs.  following injections). Although all patients required the second and third sufentanil injection, only 6 of 10 patients required the fourth sufentanil injection. In four patients, the observation time expired without the need of the last subarachnoid injection of sufentanil or any other parenteral analgesia. Four of the six patients with all the four sufentanil injections did not need further subarachnoid analgesic drugs, whereas two of these patients required 0.25% bupivacaine 60 min after the last sufentanil dose. No patient received additional parenteral analgesics in the intensive care unit.

The mean arterial pressure did not change over time, and no patient showed any signs of motor block (Bromage scale 0) at any time. Table 4gives an overview on the incidence of side effects after the respective sufentanil injections. Moderate pruritus was the most frequent side effect (90% of patients), and nausea and vomiting were observed in 20% of the patients. Respiratory depression (< 8 breaths/min) occurred in three patients within 30 min after the first sufentanil injection. No respiratory depression was registered after the subsequent sufentanil injections. No patient experienced oxygen desaturation less than 95% or apnea after sufentanil injection, although four patients received oxygen via  face mask throughout the time. Medical treatment was not required for any of the registered side effects.

None of the 90 patients suffered from postdural puncture headache or any transient neurologic symptoms, numbness or motor weakness of the lower extremities, or cauda equina syndrome on postoperative day 2 or on the day of discharge.

The present study shows that 10 μg sufentanil, 5 mg bupivacaine, or a combination of 5 μg sufentanil and 2.5 mg bupivacaine injected through spinal microcatheters provide immediate and adequate postoperative pain relief for 2–3 h in patients after major orthopedic lower-limb surgery.

No significant difference between the treatment groups was observed with regard to the quality of pain relief assessed by VAS, the time of onset, and the duration of pain relief. One interesting finding is that insufficient analgesia was associated with a higher incidence of nausea in the placebo group and not with opioid application, as one might have expected. The rapid onset of pain relief in all treatment groups can be explained by the subarachnoid injection of highly lipid soluble drugs such as sufentanil and bupivacaine, which have a high affinity to neural tissues, e.g. , spinal nerves and cord. The fast onset of pain relief by only small doses of local anesthetics or opioids is one of the advantages of subarachnoid over epidural or systemic application. This has already been demonstrated for obstetric analgesia, where subarachnoid techniques—combined spinal–epidural anesthesia more than CSA—have been established for years for this reason. 16–18 

Camann et al.  19demonstrated that the duration of analgesia was significantly longer after subarachnoid application of 10 μg sufentanil when compared with an epidural or intravenous injection of the same dose in parturients. These results in combination with our data demonstrate that sufentanil acts directly on spinal opioid receptors. Campbell et al.  20showed that the combination of 2.5 mg bupivacaine and 10 μg sufentanil provided more extended pain relief than did the subarachnoid injection of 10 μg sufentanil or 2.5 mg bupivacaine alone. Although different doses to our study have been applied and labor pain is not really comparable with postoperative pain after lower-extremity surgery, the obtained quality and duration of pain relief were similar in our study and that of Campbell et al .

It is known from animal experiments and clinical studies with spinal and epidural analgesia that the combination of local anesthetics with spinal opioids can enhance or prolong analgesic effects because of a reduction of the “wind-up” phenomenon, which represents a potentiation of afferent C-fiber–transmitted nociception caused by repetitive pain stimuli. 21–23Although sufentanil did not show significant advantages over bupivacaine in terms of analgesic effectiveness in the present study, sufentanil analgesia was not associated with significant motor block or decreases of mean arterial pressure. Cardiovascular stability was also confirmed after subarachnoid sufentanil by Houweling et al. , 24who compared different modes of sufentanil application—intravenous, epidural, and intrathecal—in patients with aortic aneurysm repair. However, although significant, the decreases in mean arterial pressure in the other groups were not clinically relevant in this study, and they should be avoided in elderly or cardiac patients.

Although the increased incidence of pruritus after subarachnoid sufentanil is tiresome but not harmful, 25and was self-limiting in our patients, respiratory depression after subarachnoid opioid application remains potentially dangerous. There have been reports on high sensory spread and respiratory depression after subarachnoid application of 10 μg sufentanil, 14,26,27although sufentanil, as a highly lipid-soluble opioid, may be associated with a lower risk than morphine; in contrast to morphine, 28there is no report on late apnea after sufentanil. The risk of respiratory arrest after intrathecal sufentanil was calculated with 0.02% by Ferouz et al.  for obstetric patients. 29However, 6 of the 50 patients who received subarachnoid sufentanil in our study showed symptoms of mild respiratory depression. Therefore, surveillance and monitoring of the patients after subarachnoid sufentanil application for at least 30 min is mandatory in this setting. In a pharmacokinetic study by Hansdottir et al. , 30the concentration peak in CSF was reached 39 ± 10 min after subarachnoid injection of 15 μg sufentanil, with a mean residence time of 55 min in the CSF and 6.8 h in the plasma. As a consequence, repetitive intrathecal sufentanil injections may enhance the risk of plasma accumulation because of the rapid CSF clearance and prolonged intravascular half-life of sufentanil. Thus, the plasma rather than the CSF concentrations of sufentanil may cause side effects after repetitive intrathecal injection. 31However, in this study, three additional injections of 10 μg sufentanil were not associated with an increased risk of respiratory depression, because all episodes of respiratory depression occurred after the first rather the following injections. In contrast to the higher sufentanil plasma concentrations after epidural application, 32the plasma levels in this study were significantly lower and may not have contributed to systemic side effects.

The question arises, if a lower dose of sufentanil, e.g. , 5 μg alone, would have been able to provide comparable pain relief while reducing the incidence of respiratory depression. However, two patients in our study also experienced respiratory depression after injection of 5 μg sufentanil plus bupivacaine.

The main advantage of subarachnoid sufentanil was the lack of motor block in this study. Earlier studies could also demonstrate that postoperative pain relief can be performed using bupivacaine infusions 33or small doses of low concentrated plain bupivacaine as repetitive bolus injections 11or via  patient-controlled analgesia 34over a period of 17–24 h. However, significant motor dysfunction or inadvertent extension of the block associated with hypotension occurred in these studies using higher doses, whereas smaller bupivacaine doses required higher parenteral rescue medication. Motor block is unwanted because patients should be able to mobilize or undergo active physiotherapy postoperatively. A higher incidence of thromboembolic events may also result if orthopedic patients are unable to move their lower extremities. The inability to monitor motor function in CSA may be one of the reasons why this technique is not frequently used for postoperative pain relief thus far. Another reason is the fear of CSF infection, although studies show that there is no increased risk for infection during the first postoperative days when spinal catheters are used for pain management. 11,35In our opinion, the minimal risk associated with the use of 28-gauge spinal catheters and isobaric local anesthetics or opioids is much more than compensated by advantages over large-bore catheters, e.g. , the lower incidence of postdural puncture headache. 11,34 

In addition to case reports with meperidine and fentanyl for pain relief using CSA, 36,37three studies dealt with subarachnoid morphine and fentanyl application for postoperative pain management after lower-extremity procedures. 9,10,38Although Niemi et al.  10did not recommend continuous morphine infusions via  28-gauge spinal catheters because of a high incidence of technical problems and side effects, they stated in a later study that intrathecal morphine as bolus dose or infusion provided better analgesia than fentanyl infusions. 38Reuben et al.  9found that 40 μg fentanyl injected through spinal catheters at the end of lower-extremity revascularization procedures provided excellent pain relief within 10 min lasting nearly 5 h. The investigators did not see respiratory depression or oxygen desaturation, and the cardiovascular stability was good. Thus, fentanyl appears to have advantages over morphine and even sufentanil, which is more expensive. However, only 10 patients were examined in each group with adequate analgesia, and the number of patients might have been too small in this trial to observe rare events such as respiratory depression with this opioid. The longer-lasting analgesic effect of fentanyl in comparison to sufentanil in our study can possibly be explained by a lower intensity of pain after vascular surgery than after major orthopedic operations.

However, all clinical studies have not allowed conclusions to be drawn on long-term application of fentanyl or sufentanil with regard to toxicity until now. Although repetitive subarachnoid sufentanil applications in large doses caused severe inflammatory changes such as meningitis and myelitis in sheep, 39more clinically relevant doses of 5–50 μg/day injected over a period of 15 or 28 days did not reveal evidence of neurotoxicity in a dog model. 40In conclusion, we believe that postoperative pain relief with a maximum of 40 μg sufentanil in fractions of 10 μg administered through 28-gauge spinal catheters represents an effective technique for early postoperative pain management in patients after major lower-limb surgery. Although repetitive sufentanil injections appear not to increase the risk of early respiratory depression, monitored surveillance of these patients remains mandatory.

The authors thank Monika Weber, Technical Assistant-in-Chief, Department of Anesthesiology; Frank Kugler and Karoline Fuchs, Research Fellows, Department of Anesthesiology; and the staff of the Postanesthesia Care Unit, Orthopedic Department and the Intensive Care Unit (all from the University Hospital Hamburg-Ependorf, Hamburg, Germany).