Peripheral nerve injury is a rare1,2 but serious complication of regional anesthesia. Needle penetration into a nerve is an infrequent occurrence while performing a peripheral nerve block (PNB). It does not necessarily result in fascicular damage,3,4 and the outcome is usually benign. Nevertheless, this complication remains a significant legal issue associated with regional anesthesia, with devastating consequences for the patient. Several methods are currently available to prevent nerve injury while performing PNB, including testing for paresthesia, using a peripheral nerve stimulator and/or ultrasound guidance, and monitoring the injection pressure while administering the local anesthetic. Unfortunately, no nerve localization or monitoring technique has been shown to be clearly superior for reducing the frequency of clinical injury,5 and nerve injury may occur even when these techniques are applied.6-10

Interestingly, a study using a porcine model showed that electrical impedance (EI) measured by nerve stimulators was significantly lower extraneurally than intraneurally when using ultrasound guidance.11 Electrical impedance describes tissue opposition to alternating current and depends mainly on tissue composition, primarily due to the variation in water and lipid content.12 When the tip of the insulated needle is outside the nerve, current flow spreads in all three dimensions in the tissue surrounding the nerve; therefore, EI is low. Conversely, the needle tip is isolated when it penetrates the nerve, and the low-impedance path for the electrical current is no longer available; therefore, EI is high. During PNB, nerve stimulators are designed to deliver specific calibrated electrical signals, i.e., they produce a constant current (I) regardless of variations in resistance (R) via insulated needles. Recently, nerve stimulators available in clinical practice (Stimuplex HNS 12, B. Braun Medical, Bethlehem, PA, USA) have displayed EI values in real time. Electrical impedance for pulsatile stimulation is calculated according to Ohm’s law by dividing the maximum voltage developed at the end of the current pulse (V) by the known applied current intensity (I).13

We hypothesized that intraneural needle tip placement (the epineurium is then crossed) would be associated with an increase in EI and that the EI variations could help in the diagnosis of accidental nerve puncture.

Methods

We carried out a prospective observational study in our institution (Limoges University Hospital Center, France) during May to November 2010 to assess the feasibility of using EI during nerve localization. This study was approved by the CHU de Limoges Ethics Committee (number 47-2010-05) on May 5, 2010. Since this was an observational study, the Ethics Committee also approved the authors’ proposal to obtain patient consent verbally. Patients were given an information sheet about their participation in the study.

We compared variations in EI between patients with no suspicion of intraneural puncture vs patients with suspected nerve puncture. Since we did not know the precise incidence of nerve puncture during PNB, we were unable to calculate a priori the number of patients needed to show a difference of EI between the two groups.

All patients requiring PNB could be included in our study. In addition, different types of nerve blocks were performed depending on the surgical procedure (interscalene, supraclavicular, infraclavicular, axillary, and nerve blocks in the forearm, femoral, popliteal, and tibial blocks at the ankle). Ultrasound guidance (S-Nerve, Sonosite Inc., Bothell, WA, USA) with a HFL 38x 13-6 MHz linear array probe was used for nerve or plexus localization, and only an in-plane needle approach was used. Nerve stimulation (Stimuplex HNS 12, B. Braun Medical, Bethlehem, PA, USA) was also used for all PNBs. To standardize EI measurements, the nerve stimulator was set at 1 Hz, 0.5 mA, and 0.1 msec, and if muscular twitches were obtained, the intensity of the minimal stimulating current was recorded. Only short bevelled needles were used, and length was chosen depending on block site and the patient’s morphology. The type of local anesthetic solution and the agents used for sedation were chosen by the anesthesiologist in charge of the patient. For patients with blocks involving more than one injection, we made measurements for the first injection only.

Demographic data and information about the PNB performed (type of block and local anesthetic solution used) were collected. Our primary outcome was relative EI variation. For EI measurements, we made a first measure at a distance of 0.5-1 cm from the nerve; this measurement was then used as a reference value outside the nerve. A second measurement was taken either immediately before local anesthetic injection if no nerve puncture was suspected, or immediately before repositioning the needle if nerve puncture occurred. Finally, the relative variation between the two measurements was calculated using the following formula: EI at injection – EI at distance from the nerve / EI at distance from the nerve. The criteria for nerve puncture were: pain or paresthesia in the sensory territory of the targeted nerve; motor responses with a minimal stimulating current lower than 0.4 mA; needle tip seen inside the nerve using ultrasound guidance; and/or nerve swelling after injection of local anesthetics. If at least one of the above criteria occurred, the patient was included in the “suspected nerve puncture group”. All other patients were included in the “no nerve puncture group”.

Statistical analysis

The statistical analyses were performed using R software version 2.10.1 (R foundation for statistical computing, www.r-project.org). The continuous covariates were expressed in terms of median and [quartiles], and the categorical variables were expressed in terms of number (%). A Wilcoxon paired test was used to compare EI at distance vs during injection, and a generalized linear model was used to investigate the influence of demographic covariates on relative EI. Results were expressed in terms of β (standard deviation [SD]). A Mann-Whitney test was performed to compare the median relative EI variation in the “no nerve puncture group” vs the “suspected nerve puncture group”. The P values are two sided, and 95% confidence intervals were calculated when relevant. A receiver operating characteristic curve (ROC) was constructed using Medcalc® (Mariakerke, Belgium) version 11.3.6 software to determine a threshold EI value for nerve puncture diagnosis with satisfactory sensitivity, specificity, and positive and negative predictive values.

Results

One hundred thirty-three patients participated in the study and 135 PNBs were performed. One patient received a rescue block and one patient underwent two surgeries. Five patients had two nerves blocked in the forearm with one injection for each nerve; therefore, two pairs of EI values were measured in these patients, and as a result, 140 pairs of EI values were collected.

Fifty-three (40%) women and 80 (60%) men were included in the study; only eight patients (6%) were diabetics. Demographic data are shown in Table 1. One hundred PNBs were performed on the upper limb, and 40 were performed on the lower limb (Table 1). The median relative EI variation for all patients was −8% [−28; −4%].

Table 1 Patient (n = 133) demographic data and data regarding the blocks (n = 140) performed
Full size table

For all patients, EI reference values (measured 0.5-1 cm from the nerve) were extremely heterogeneous. Median EI was 14.0 kΩ and ranged from 5.2-81.3 kΩ. Furthermore, extraneural EI references were significantly higher in the popliteal fossa than in the other anatomical locations (Mann-Whitney test, P < 0.05). Median EI was less on injection (12.2 kΩ [9.0; 15.5 kΩ] vs at a distance from the nerve (14.0 kΩ [Q1; Q3: 12.0; 18.0]; P = 0.0009), Wilcoxon rank test (for dependent measures).

Generalized linear analysis showed that neither sex [male vs female, β = −0.06 (SD 0.05); P = 0.27] nor age [per year increase, β = 0.0004 (0.001); P = 0.77], nor body mass index (BMI) [per unit increase, β = −0.006 (0.006); P = 0.36], nor diabetes (no vs yes, β = 0.17 (0.10); P = 0.10] affected EI relative variation measurements.

We observed at least one suspicious sign of intraneural puncture during each of 21 nerve blocks (Table 2). Six patients experienced pain or paresthesia in the sensitive distribution of the blocked nerve during the procedure, and in 15 cases, motor responses were observed with stimulating current < 0.4 mA. Two patients had both pain and paresthesia with a current < 0.4 mA. No nerve swelling was diagnosed after injection of 1 mL of the local anesthetic solution, although we twice observed accidental intraneural needle tip penetration. No nerve puncture was suspected in the remaining 119 blocks.

Table 2 Characteristics of nerve blocks with suspicion of nerve puncture
Full size table

The EI variations were significantly different between the two groups. In the suspected puncture group, there was median increase of +6.6 % [−20.0; 36.0%], whereas the median EI variation was −10.0 % [−28.0; 0.0%] in the no puncture group (P = 0.02) (Fig. 1). Absolute values for EI were 15.5 kΩ [12.0; 18.0%] in the suspected puncture group and 12.0 kΩ [8.9; 15.1%] in the no puncture group (P = 0.013).

Fig. 1
figure 1

Relative electrical impedance variation (%) depending on whether a nerve puncture criterion is present (right) or not (left). The line in the rectangle represents the median; the rectangles show the quartiles, and the whiskers represent the minimum and maximum values

Full size image

A ROC curve analysis was performed (Fig. 2). The area under the curve was 0.67 (95% confidence interval 0.58 to 0.75; P = 0.03). With a +4.3% threshold, the sensitivity, specificity, positive predictive values, and negative predictive values were 57%, 82%, 36%, and 92%, respectively.

Fig. 2
figure 2

Receiver operating characteristics curve (plot of sensitivity values against 1 – specificity values) for several relative electrical impedance variations

Full size image

Discussion

Our study suggests that EI variations could be useful during the performance of ultrasound-guided nerve block. An increase in EI might be indicative of intraneural location, and extraneural positioning of the needle is usually associated with a decrease in EI.

Electrical impedance was first used by Thomas et al. in 1987 to identify the epidural space.14 More recently, two other studies evaluated EI variations measured by nerve stimulators. In one study, accidental vascular puncture was diagnosed,15 and in the other study, nerve puncture was detected during PNB in pigs.11 In our study, we assessed the use of EI in regional anesthesia in humans. Ultrasound evidence of intraneural puncture requires observer expertise,16 and other technologies are needed to help improve PNB safety.17,18

Our results confirm those of Tsui et al. 11 in an animal model. They found that the variation in EI displayed by the nerve stimulator is greater when the needle tip crosses the epineurium. We observed that the extraneural EI measured in our study (14.0 kΩ [12; 18 kΩ]) was lower than that reported by Chin et al. 15 [23.5 (8.3) kΩ]. One explanation for the difference could be that Chin et al. collected each measurement during supraclavicular nerve blocks, whereas we measured EI at many different body sites. The wide heterogeneity of EI values in our work (5.2-81.3 kΩ) might be explained by the various blocks performed. Indeed, a study that evaluated tissue EI along the median nerve revealed significant differences when EI was measured at the elbow vs the axilla,19 and this may be explained by differences in the distribution of muscle or fat. Electrical properties depend on the type of tissue through which the electric current travels.12 Moreover, muscles have a non-homogeneous electrical behaviour with a conductivity varying with muscle fibre orientation.20 This can explain why we observed highly variable EI values from one anatomical site to another in a same person or from one patient to another. Therefore, it is more relevant to detect relative rather than absolute values to compensate for intra-individual and interindividual sources of EI variations. We also observed that extraneural EI was higher when measured in the popliteal fossa than in other sites, probably because of the large amount of fat tissue surrounding the nerve. It is possible that EI has a better diagnostic performance when used in anatomical sites where extraneural tissue EI is low. Nevertheless, our small sample size did not allow us to verify this hypothesis. In order to establish optimal standardization of the extraneural EI measurement, it may be best to measure it in the subcutaneous fat that is present at the site of the nerve blockade.

The measurement of EI may be subject to errors because of electrical interference, a less than ideal electrical pathway through tissues, and differences between set and delivered current.21 The sources of error that can affect variations in EI measurements are not known at present.

We intended to determine a variation cut-off that would detect intraneural puncture reliably despite the small sample size in the “suspected nerve puncture” group. We focused on negative predictive value because, in our view, certainty that the needle is positioned outside the nerve is more useful than confirmation of a damaged nerve. We propose that a +4.3% cut-off is associated with a very good negative predictive value. We are aware that this is more of a theoretical value and that it is difficult to reach this value precisely in routinely practice. Nevertheless, one of our main objectives was to determine a first cut-off value for the diagnosis of intraneural injection.

Clearly, the most important limitation of our work lies in the choice of nerve puncture criteria. Since no gold standard is currently available, we relied on the criteria usually applied to suspect accidental nerve puncture during PNB. Each criterion is arguable4,16,22; but for ethical reasons, we could not perform voluntary nerve punctures, which could have reduced the variations in EI between the two groups. With our nerve puncture criteria, it is impossible to determine how often the diagnosis of nerve puncture was correct.

It was also not possible to distinguish between intra- and extrafascicular punctures. Nerve puncture does not invariably result in nerve damage, and it appears that intrafascicular injection is more deleterious than intraneural extrafascicular punctures.4 To date, no technique can convincingly make the difference. Therefore, as mentioned by Tsui et al.,11 further work should focus on differences between intra- and extrafascicular EI.

In conclusion, the present study suggests that EI measurements in regional anesthesia may be used to detect the difference between an extraneural or intraneural location and that nerve stimulators commonly used during PNB may be able to detect those differences. Thus, EI measurement may enhance PNB safety, although this needs to be confirmed by further work. We propose a relative EI cut-off of 4.3%, which is associated with a very good negative predictive value. Nevertheless, the optimal EI cut-off may depend on other factors, especially the type of block performed. We also propose that future work should focus on a single anatomical site.