Postoperative pain after major knee surgery is a major concern. It is severe in 60% of patients and moderate in another 30% (1,2). Pain has a major impact on patient satisfaction and postoperative well-being. In addition, pain impairs early intensive physical therapy and rehabilitation—probably the most influential factor for good postoperative knee rehabilitation (2,3).
Continuous peripheral nerve blocks offer the potential benefits of extended postoperative analgesia, few side effects, improved patient satisfaction, and accelerated functional recovery after major knee surgery (4). Continuous femoral nerve block is often used to provide postoperative analgesia in this setting (1,3).
When performing a continuous peripheral nerve block, efforts are made to place the catheter close to the nerve to achieve effective perioperative analgesia. Traditionally, catheter placement is performed through a stimulating needle, followed by injection of the local anesthetic and then blind advancement of the peripheral catheter beyond the needle tip. Secondary analgesic block failure rate (failure of a catheter to produce postoperative analgesia after having provided sufficient intraoperative analgesia with the bolus administration) with this technique ranges from 10% (5,6) up to 40% (7). This may be explained by the fact that the catheter can curl away from the needle during uncontrolled advancement (7). Correct catheter placement is confirmed by testing for a clinical effect of satisfactory analgesia or by sensory modality testing within the desired sensory distribution after injection of the local anesthetic. However, in case of insufficient block, the catheter cannot be further redirected.
It has been argued that a more reliable method to verify correct catheter position at the time of insertion is needed to prevent secondary block failure.
Stimulating catheters were introduced in 1999 (8) and provide the possibility to verify the position the catheter takes during advancement through the cannula. An observational study in 130 patients demonstrated that stimulating catheters appear to be beneficial (9). In this trial, catheters were positioned at different locations using the stimulation catheter technique. The authors found that the ability to elicit a motor response during catheter advancement correlated with successful clinical anesthesia, tested over 48 h, in 124 of 130 cases, and they concluded that the success rate was increased by using electro-stimulation over an in situ catheter. However, the authors also suggested that controlled investigations were necessary to compare this technique with conventional methods, and that femoral block may be an appropriate model to assess the utility of stimulating catheters.
We hypothesized that catheter placement using a stimulating catheter improves onset of sensory and motor blockade after injecting the local anesthetic solution via the catheter and further reduces the rate of secondary catheter failure and improves the quality of postoperative analgesia. Therefore, we compared catheter placement time, onset of sensory and motor block, and efficacy of analgesia using stimulating versus blind catheter advancement in 81 patients undergoing major knee surgery.
This prospective, randomized and observer-blinded trial was approved by our local ethics committee and written informed consent was obtained from each patient. Exclusion criteria were infection near the insertion site, coagulation disorders, preexisting neurological disorders, known allergies to local anesthetics, former operation of the vessels near the insertion site, ASA classification IV or V, age younger than 18 yr, pregnancy, and lactation period. Eighty-one patients undergoing major knee surgery under general anesthesia (total knee replacement or anterior cruciate ligament repair) received a continuous femoral nerve block which was performed by one of four anesthesiologists (AMM, LHJE, CK, HW) with considerable experience in regional anesthesia (having performed >250 peripheral nerve blocks). Patients received oral premedication with clorazepate 20 mg and rofecoxib 50 mg 1 h before the procedure. For all patients, a stimulating catheter, the “Arrow StimuCath continuous nerve block set” with a 17-gauge Tuohy needle of 9-cm length and a 19-gauge catheter (Arrow, Germany; Fig. 1) was used. The technique to identify the femoral nerve with the stimulating needle was standardized and performed according to institutional standards. After aseptic preparation of the puncture site and intracutaneous local anesthesia, the femoral nerve was identified just lateral to the femoral artery below the inguinal ligament in conscious and cooperative patients. The nerve stimulator current was initially set at 1 mA, with 2 Hz and 0.3 ms (Stimuplex HNS 11; Braun, Germany). After initial appropriate motor response was noted with the needle (cephalad patellar movement by eliciting quadriceps femoris muscle twitches), the current was progressively reduced to ≤0.3 mA. Patient randomization was performed using sealed envelopes immediately before performing the block. In the conventional catheter group (CC group), the catheter was blindly advanced 5 cm through the needle without further stimulation of the catheter. With the catheter in place, the active lead of the nerve stimulator was clipped to the stylet of the catheter (Fig. 1). The catheter was stimulated to verify the required current to evoke an appropriate motor response, but no correction of catheter placement was performed. In the stimulating catheter group (SC group), the catheter was slowly advanced 5 cm beyond the needle tip under continuous electric stimulation using an initial current of 0.5 mA that was subsequently adapted according to the motor response achieved. Using the stimulating catheter in this real-time manner, we could immediately see if the desired motor response decreased during catheter advancement, and then either the catheter or the needle was manipulated until muscle twitches of the quadriceps femoris muscle reappeared. The time interval from the first penetration of the skin with the stimulating cannula until its final removal was defined as catheter placement time. A maximal time of 20 min was allowed for the whole procedure. In case of failure, the catheter was left in place and the patient was included in an intention-to-treat analysis. The ease of catheter placement was rated by the anesthesiologists on a four-point Likert scale.
The time until the catheter was in place, and the minimal current that could provoke a correct motor response over the catheter was recorded. All catheters were sutured to the skin to avoid catheter dislodgement.
After a negative aspiration test for blood, 20 mL of prilocaine 2% (400 mg) was injected to initiate the block in both groups. Over a period of 30 min, sensory block was evaluated repeatedly using an ether ball and pinprick perception at 6 defined areas of the skin (anterior, medial, and lateral aspect of the thigh, 10 cm below the groin and above the patella). Sensory block was defined as a loss of cold and prick sensation. Motor blockade was evaluated by stretching the knee, abducting and adducting the leg, and was defined as onset of weakness, according to Bromage scale ≥grade II (10). Sensory and motor testing was performed every 2 min by a blinded investigator who was not present during the placement of the catheter (SW and TK).
To test the hypothesis that multiple manipulations in the SC group could lead to unpleasant sensations, patients were asked to rate the procedure of catheter placement on a visual analog scale (VAS) scoring (ranging from VAS = 0 for not bad at all and tolerable, to VAS = 10 for very painful and beyond endurance).
After the 30-min testing period, a continuous infusion of ropivacaine 0.2% with 6 mL/h was started. In all patients, general anesthesia was performed. After induction with propofol 1–3 mg/kg, 0.5 mg/kg rocuronium bromide, and 5 μg/kg fentanyl as the sole intraoperative analgesic, anesthesia was maintained with desflurane or propofol (Table 1).
Postoperative care was standardized during the first 48 h. The ropivacaine infusion with 6 mL/h was maintained during surgery and for at least 48 h. No bolus application was allowed during this period. On the ward, all patients received a daily dose of rofecoxib 50 mg per os. A patient-controlled analgesia (PCA) with piritramide (an opioid with comparable potency to morphine) was initiated with a bolus of at least 2 mg and a lockout interval of 10 min, and the doses of the opioids administered via the pump were recorded during the study period of 48 h. All patients had identical physical therapy regimens. From the day after surgery until discharge, active and assisted knee flexion and extension exercises were done twice daily. In addition, a continuous motorized motion machine was applied 3 times daily for 30-min duration, with the range of motion set at the level well tolerated by the patient. Maximal bending and stretching was assessed daily by 2 physical therapists (WN and UW) blinded for the study group, for 5 days.
The patients were visited at least twice a day to assess pain scores at rest and maximal bending using VAS scoring (ranging from VAS = 0 for no pain to VAS = 100 for maximal pain) and to evaluate the correct position of the catheter by neurological examination.
The femoral nerve catheters were left in place even after the study period of 48 h, as long as deemed necessary or until local infectious signs occurred. All data were collected by a blinded observer (SW and TK). The results of the study are reported according to the requirements of the CONSORT statement (www.consort.org).
A prospective power analysis revealed that 80 patients provided a 94% chance (power) to detect reduction of mean piritramide consumption by one-third (e.g., from 60 to 40 mg) during the first 48 h postoperatively with a type I error of 0.5, when the standard deviation was not >50% of the means and correction for non-normal distribution of the values was assumed. Sample size calculation was performed using NCSS Trial (Number Cruncher Statistical System [NCSS] Statistical Software, Kaysville, UT).
Continuous data that were repeatedly recorded (VAS pain ratings, bending and stretching of the knee) were analyzed using a two-factorial analysis of variance. Analyses of single measurements were performed using Student’s t-test (if distributed normally) or the Mann-Whitney U-test (in case normal distribution was rejected using the Kolmogorov-Smirnov test). Repeated measurements of dichotomous data (e.g., loss of sensory or motor block) were subjected to a Kaplan-Meier survival statistics using the nonparametric Breslow-Gehan-Wilcoxon test including censored values. The χ2 test (with correction for continuity) or Fisher’s exact test, if appropriate, was used for any other nominal data. Statistical significance was assumed if P < 0.05. All statistical calculations were performed using StatView 4.5 for Windows.
During the study period, 141 patients presented for elective major knee surgery. Of these, 12 patients elected to undergo surgery under spinal anesthesia and 10 patients refused a peripheral nerve block. Of the remaining 119 patients eligible for the study, 15 patients were not included because of organizational reasons (e.g., none of the 4 experienced anesthesiologists was available) and 8 patients refused to participate in the study. Informed consent was obtained from 96 patients. Of these, 11 patients were not randomized because surgery was postponed or a diagnostic arthroscopy was performed. Thus, 85 patients were randomized (CC group: n = 44; SC group: n = 41). This inhomogeneity occurred because no efforts were made to balance the number of patients between the groups, e.g., using a fixed 2 × 40 allocation between the groups in order to ensure the allocation concealment.
There were four subsequent drop-outs that were not included in the final analysis. In two patients of the SC group, surgery was modified intraoperatively into a diagnostic arthroscopy, and a further two patients were excluded because of incomplete recordings (one from each group). Finally, 81 patients were included in the analysis, 43 patients in the CC group, and 38 in the SC group. Postoperatively, catheters were left in place 4 days (4/5; median, 25th/75th percentile). No catheter required removal because of secondary block failure or local infection. No signs of nerve irritation or neurological complication were observed. Demographic data are presented in Table 1. No statistical differences were noted between the groups.
Preoperative Assessment of the Patients
Catheter placement time, need for redirection, current required to elicit a motor response, and catheter duration are presented in Table 2. The current required during catheter stimulation to provoke a motor response was significantly lower in the SC group. There were no other differences between groups.
Patients judged the whole procedure as quite tolerable in both groups. The four anesthesiologists who performed the blocks for this study rated the placement difficulty to be the same in the CC and the SC group (Table 2).
The onset time of sensory block with cold and pinprick testing after catheter placement and injection of the bolus dose did not differ in any of the six tested areas. Furthermore, there were also no differences with respect to the onset of motor block in the three tested movements. For better visualization, three representative areas for sensory block and motor block for bending, abduction, and adduction are shown as Kaplan-Meier curves in Figure 2.
There were no statistically significant differences with respect to postoperative opioid consumption over PCA during the first 48 h after surgery and the VAS scores for pain at rest and at movement during the 5 observational days after surgery (Table 3, Figs. 3 and 4).
An explorative analysis was performed to reveal the possible connection between the current needed to obtain adequate motor response and the clinical success of the catheters defined as onset time of sensory and motor block and the clinical variables evaluated postoperatively (opioid consumption, VAS pain scores, and joint range of motion). For this analysis, the Spearman rank correlation was performed with patients of both groups analyzed together. There was no correlation between the accuracy of catheter placement (determined by the current needed to achieve adequate motor response) and the following variables: onset time of sensory block (ρ = −0.08–0.27), motor block (ρ = 0.11–16), piritramide consumption within 48 h (ρ = 0.17), VAS pain scores at rest (ρ = −0.01) and at movement (ρ = −0.02) on day 2, maximal bending (ρ = −0.11) and stretching (ρ = 0.11) of the knee on day 2, as well as on the other days.
In a similar explorative analysis, the 52 catheters of both groups where motor response of the patella could be evoked with a current of ≤0.5 mA (“positioning successful”) were compared with those 11 catheters in which no response could be obtained with a current up to 5 mA (positioning “not successful”). Onset time for sensory block after “unsuccessful block” was 9 min (6/14; median and 25th/75th percentile) and in the group with “successful block” 6 min (2/12); P = 0.23. The data for the onset times of motor block were 13 min (4/24) and 9 min (3/18), respectively; P = 0.62. Functional outcome with bending and stretching of the knee during physical therapy during the five postoperative days was similar between groups. For example, bending on day 2 in the group in which no motor response could be elicited with 5 mA was 68° (59/80), and in the group with successful catheter positioning it was 70° (60/85). The values for stretching of the knee were 10° (4/10), respectively 5° (5/8).
After major knee surgery, analgesia provided by continuous femoral nerve block is effective, has few side effects, allows for more intense early rehabilitation and accelerated functional recuperation, and shortens hospital and rehabilitation center stay (3). Poorly managed pain may inhibit the early ability to mobilize the knee joint. This in turn may result in capsular contracture which may impair functional outcome (1).
When performing a continuous peripheral nerve block, efforts are made to place the catheter close to the nerve to achieve effective perioperative analgesia with dilute local anesthetic solutions used for postoperative analgesia (7). Previous studies have attempted to obtain correct catheter tip location by means of radiography after injection of contrast media. Results from two studies are completely contradictory: in one study, successful sensory and motor block and the quality of postoperative analgesia depended on the position of the catheter tip near the lumbar plexus (11); in another trial, no evidence supported a relationship between the radiographical image of ideal perineural catheter’s position and its clinical effectiveness (12).
An alternative for assisting with correct catheter placement is ultrasonographic guidance. In one investigation, the onset of sensory blockade was significantly shorter and the quality of sensory block significantly better compared with the nerve stimulator needle-assisted application of local anesthetic (13).
The ability to elicit motor response with a stimulating catheter implies that the catheter is adjacent to a nerve and within the perineural space. In 1999, Boezaart et al. (8) described this new technique of stimulating catheters. An observational study of 130 patients evaluated the ability of catheter stimulation to correctly place catheters (9). The stimulating catheter was advanced 3–5 cm over the needle tip without continuous real-time stimulation. Only 37% of the catheters achieved the desired perineural position at first attempt, and thus multiple attempts had to be performed to correctly place the catheter. The results in our trial were comparable. About half of the catheters (42%) were placed correctly at first attempt in both the CC and SC groups. Using the stimulating catheter technique, it was possible to redirect the catheter in all but two patients to achieve cephalad patella movement (successful placement) with a current ≤0.5 mA.
This is the first randomized controlled study to answer the questions concerning whether the improved accuracy achieved by stimulated catheter positioning speeds the onset of sensory and motor block and improves the quality of postoperative analgesia or enhances functional recovery compared with blind catheter advancement.
Using femoral nerve block in our trial as a model for peripheral nerve blockade with a considerable incidence of secondary block failure [10% (5,6) up to 40% (7)], we found no difference between the CC group and the SC group in onset time and failure rate. The quality and expansion of primary blockade was the same in both groups (Fig. 2). Opioid consumption via the PCA was the same in the CC and the SC group, regardless of the type of surgery performed (Table 3). The mean consumption of piritramide, an opioid with comparable analgesic potency to morphine, seemed to be high with a cumulative use after 24 hours of 34 mg (CC group) and 35 mg (SC group), respectively, and 18 mg versus 9 mg on the second day, from 24 to 48 hours observational period. This could be explained by the fact that we chose a slow standardized infusion rate postoperatively via the femoral catheter, with 6 mL/h ropivacaine 0.2% with no additional bolus administration allowed, to make the 2 groups comparable with respect to opioid consumption. As a consequence of the unlimited access to opioids, there were no significant or clinically relevant differences in postoperative pain scores (Fig. 3). However, during exercise, half of the patients experienced considerable pain (VAS ≥ 50 mm on a 10-mm VAS) on the first postoperative day. Over time, pain scores decreased significantly (0.01 > P > 0.0001) in both groups. During the observational period of 5 days, there was no significant difference in maximal bending and stretching of the knee between the 2 study groups (Fig. 4).
There was no relationship between the current that had to be applied via the stimulating catheter to evoke a motor response and any of the variables determined to judge the success of the catheter, for example, in the comparison of the 52 catheters of both groups where motor response of the patella could be evoked with a current of ≤0.5 mA (positioning successful) and the 11 catheters where no motor response could be obtained with a current as high as 5 mA (positioning unsuccessful). Thus, we can only speculate that these catheters, inserted after successful nerve stimulation via the cannula, were still close enough to the nerve and might just have been separated by a small layer of connective tissue or fat that caused insulation of the catheter tip.
However, our results are only valid for the setting and technique described. For example, the catheters were not inserted further than 5 cm over the stimulating needle tip, to minimize the possibility of curling away from the nerve. This, together with the fact that the success rate of catheter positioning was already very high in the CC group, might be the main reason for the unexpected results of the stimulating catheters.
One might criticize our decision not to stimulate the catheters 24 and/or 48 hours postoperatively to evaluate whether the catheter had maintained its initial position. This was not done because of ethical considerations: the local anesthetic infusion would have had to be stopped for several hours before a reliable measurement could have been performed.
In conclusion, catheter positioning with the stimulating catheter did not improve onset time of sensory and motor block, quality of postoperative analgesia, or functional recovery in patients undergoing major knee surgery under continuous femoral block. Although stimulating catheters are a useful didactive tool, additional investigation is required before we can objectively recommend the use of stimulating catheters for specified indications, unless their cost becomes similar to that of nonstimulating catheters.