Keywords

Introduction

Neurologic injuries following peripheral nerve blocks (PNB) are rare, ranging between 2.4 and 4 per 10,000 blocks, but they can be debilitating and, at times, devastating [1–6]. From a health perspective, a rare event can be defined as any event that occurs infrequently (≥1/10,000 to <1/1000) [7]. Rare events do not occur in a predictable pattern; thus, trying to deduce event rates may prove to be erroneous. Predicting neurologic complications following PNBs is subject to the same issues affecting other rare events, such as multiplicity of sources, difficulties in data collection, and variation in statistical analysis. The incidence of the event may be impacted further by any change in the target population or the definition of the problem. Unsurprisingly, no studies to date have investigated neurologic complications following regional anesthesia from a rare event perspective, likely due to the complex interactions of known and unknown factors that influence these complications. Although the use of ultrasound (US) has been shown to reduce the incidence of vascular puncture, LA systemic toxicity [7], and block success [8] we have yet to demonstrate improvements with the introduction of US.

Neurologic injury following PNB is complex and includes needle trauma, pressure injury [9], damage to the vasa nervosum resulting in hematoma formation or ischemia, and finally LA [10] or adjuvant-related toxicity. Other important factors also include patient characteristics [11, 12], type of surgery [13], and the anatomical location of injections. Given the complexity of possible interactions among various factors in regional anesthesia, the complication may be best explained using the same epidemiological principles of disease causation (Fig. 5.1).

Fig. 5.1
figure 1

Epidemiologic triangle demonstrating relationship between causative agents, host factors, and environmental influences on neurological injury

Full size image

Epidemiological principles have been used to determine and study the interrelationship of various factors on the suspected cause of diseases so that control measures can be identified and implemented to prevent and minimize the disease. Typically, the event (complication) is said to occur when there is interaction among an injurious origin (causative agents), a susceptible host (host factors), and suitable circumstances (environmental influences) known popularly as the epidemiological triad [14, 15]. Using this model, risk factors can then be broadly classified to the host (anatomical and biological factors), the injurious agent (mechanical, pressure, and neurotoxic insults), and the environment (guidance techniques, supervision, safety culture). The neurological injury may represent the final result from the interaction between these risk factors. Elimination or minimization of one of the triangle’s component may potentially, in theory, interrupt the interaction and prevent the event from occurring.

In fact, any discussion of epidemiology would be incomplete without mentioning John Snow, a pioneer anesthesiologist, who is also known as the “father of epidemiology” due to his well-known first epidemiological studies conducted in the 1850s [16]. In his studies, Snow used logic and common sense to study the interaction of factors causing disease and to develop preventative measures in ending the cholera outbreak. This work classically illustrates the effective use of epidemiological principles used even today to investigate and control disease and outbreaks. In this chapter, we, therefore, have performed a systematic review to evaluate the pertinent clinical and pathophysiological aspects of neurological complications following PNBs from the perspective of the epidemiological triangle.”

Search Strategy and Selection of Studies

A systematic review of the medical literature (MEDLINE, OVID, and EMBASE) was performed during Nov–Dec 2015 using the search strategy described later. The MEDLINE search used a combination of the following medical subject headings: nerve injury, neurologic injury, peripheral nerve injury, neurologic deficit, paresthesia, neurologic sequelae, pathology, ultrastructure, anatomy, transient neurologic deficit, transient neurologic symptoms, paralysis, nerve block, peripheral nerve block, local anesthetic, local anesthesia, conduction anesthesia, and regional anesthesia. Subsequent searches combined the keywords intraneural injection, epineurium, subepineurial injections, perineurium, intrafascicular injection, extrafascicular injection, injection pressure, ultrasound, neurostimulation, and needles. EMBASE and OVID database searches were performed for the period 1975–2015. We started from the year 1975 since the very first investigations, looking into the factors important to the causation of nerve injury following regional anesthesia in a systematic way, began in 1977 [17].

Both human and animal studies were included in the review. Additional database searches included Cochrane, LILACS, DARE, IndMed, ERIC, NHS, HTA via Centre for Reviews and Dissemination (CRD; York University), which did not produce any additional unique results. The bibliographies of publications included for analysis were also reviewed manually for additional material that may have been missed by the database searches.

Literature Selection

The full text of all articles obtained from the searches was retrieved for critical appraisal. References of all articles were examined to ensure that no original research studies were missed. We included closed claimed analyses, meta-analyses, systematic reviews, randomized controlled trials (RCTs), controlled studies without randomization, observational studies, retrospective studies, comparative studies, and case series for this review. For the purposes of this review, RCTs were defined as such only when they included human subjects; randomized studies of animal subjects were not classified as RCTs. We did not include correspondences, pediatric population, or conference abstracts with incomplete data sets in this review.

Evidence Evaluation

Relevant full-text articles were separated based on literature type (database reviews, human and animal studies) and were subsequently reviewed independently in duplicate. Data were classified based on the epidemiologic triangle : (1) host factors (anatomic, surgical, and patient-specific elements), (2) damage-causing agents (needle, local anesthetic, adjuvants, pressure injury), and (3) environmental factors (methods to detect intraneural injection, safe practices, future technologies). Additionally, data relating to nerve injury and the incidence of unintentional intraneural injection were evaluated separately.

Data were extracted and entered into a database (Microsoft Excel, Microsoft Corp., Redmond, WA). Level of Evidence (Table 5.1) and Grades of Recommendation (Table 5.2) developed by the Oxford Centre for Evidence-Based Medicine were assigned to each study.

Table 5.1 Oxford Centre for Evidence-Based Medicine levels of evidence and grades of recommendation (adapted from www.cebm.net)
Full size table
Table 5.2 Assessment of animal studies investigating neurological injury
Full size table

Furthermore, RCTs included in the current review were assigned Jadad scores (0–5) [18] while case reports were graded by Pierson scale [19] to assess scientific quality. Nonrandomized studies were not assessed for quality. Animal and cadaveric tissue studies were given a lower grade (Level of evidence 5; Grade D) irrespective of the study design.

Selected studies: A total of 3328 abstracts were retrieved from the MEDLINE, OVID, and EMBASE databases. After elimination of 62 duplicates, 3266 articles were screened for eligibility, 206 of which were selected for full-text review . Seven additional articles identified from a manual search of references from relevant articles were included. Seventy nine studies were excluded based on the criteria earlier, leaving 134 full-text articles for review (Fig. 5.1).

A total of 43 animal [9, 17, 20–59] studies (Table 5.2), 60 human [1–6, 60–113], and 8 cadaver/laboratory studies [114–121] (Tables 5.3 and 5.4) 21 case reports/case series (Table 5.5) [122–143] were included for this review. The statement of evaluated outcomes has been summarized in Table 5.6.

Table 5.3 Assessment of human studies investigating neurological injury
Full size table
Table 5.4 Incidence of neurological complications following peripheral nerve blocks
Full size table
Table 5.5 Case reports of neurological dysfunction following peripheral nerve blocks
Full size table
Table 5.6 Summary of evaluated outcomes with statements of evidence and grades of recommendation
Full size table

Among animal studies, eight studies evaluated the impact of needle design on nerve trauma, while seven studies reported on the injection pressure, 21 studies evaluated neurotoxicity of LA/adjuvants, and seven studies evaluated guidance methods such as neurostimulation/US. Of the human studies, six studies evaluated the incidence of unintentional intraneural injections while four studies evaluated the impact of deliberate intraneural injections. A total of 38 studies reported on neurologic complications in relation to PNB, while the remaining 9 reported on methods to detect or avoid intraneural injection.

Incidence of Neurologic Complications Following PNB

Transient neurologic dysfunction following PNBs are more common than long-term dysfunction and usually resolve with time (LOE 1b; Grade A). Long-term postoperative neurologic dysfunction is rare following peripheral nerve blocks (LOE 1b; Grade A). Procedure-induced paresthesia may increase the risk of postoperative neurologic dysfunction (LOE 1b; Grade A). The safety of performing PNB under general anesthesia and its impact on neurologic outcomes is currently unknown (LOE 2b; Grade C).

Retrospective reviews tend to under-report the incidence of neurologic complications due to selection, information, and recall bias, whereas the medico-legal databases may overestimate the incidence due to over-reporting and a lack of denominator for the incidents (Table 5.4). Early attempts to determine the incidence of neurologic sequelae following regional anesthesia came from ASA closed claims analyses [69, 88]. The first closed claims analysis included spinal anesthesia, ophthalmic blocks, and chronic pain blocks, while the latter looked specifically for neurologic complications following PNB. Closed claims analyses of PNBs have shown a trend toward a rise in nerve injury claims over the years (31–51 %), but only a few are thought to be related to the PNB itself [88, 89]. This ambiguity necessitated several prospective studies of block-related neurologic sequelae.

Prospective studies estimate the incidence of long-term neurologic injury following peripheral nerve blocks to be in the range of 2.4–4 per 10,000 blocks [2, 65–68, 144]. Transient neurologic deficits lasting up to 2 weeks occur more frequently following PNB, with an incidence varying between 8.2 and 15 % [3, 145]. Transient neurologic symptoms are known to resolve by 6 months to 1 year [3, 66]. Neither ultrasound nor nerve stimulation guidance affected the incidence of short- or long-term neurologic dysfunction following PNB in one retrospective review [5], although a recent update of the same database showed a lower incidence of short-term neurologic dysfunction with the use of ultrasound guidance [4]. A retrospective database review of ultrasound-guided blocks showed an incidence of long-term neurologic dysfunction of 0.9/1000 [6], which is about 22 times higher than those reported by others [1–3, 67]. Various definitions of long-term neurologic dysfunction (e.g., >6 vs. >12 months) may have accounted for the difference in incidence between these studies.

Procedure-induced paresthesia may increase the likelihood of transient neurologic symptoms following PNB as reported in three prospective cohort studies [3, 102, 144]. Certain peripheral nerve blocks have a predilection for neurologic complications than others. In a retrospective review of 12,668 patients undergoing ultrasound-guided nerve blocks, Sites et al. [6] reported short-term neurologic dysfunction being highest with axillary nerve block (2.3 %), followed by interscalene catheter (1.2 %), popliteal sciatic block (0.4 %), single-injection interscalene block (0.35 %), supraclavicular block (0.2 %), and femoral nerve block (0.1 %). Long-term dysfunction was again common with interscalene catheters (0.87 %), popliteal sciatic block (0.31 %), and single-injection interscalene block (0.25 %). In contrast, supraclavicular, axillary, and femoral nerve blocks rarely caused long-term problems. In an internet-based survey of 36 centers (27,031 patients), Ecoffey et al. [74] reported an overall incidence of postoperative neurologic symptoms of around 0.37 per 10,000 following ultrasound-guided axillary brachial plexus block, most of which were thought to be unrelated to the block. Although the reported incidence indicates a decrease in block-related neurologic symptoms compared to other studies [6], whether or not the observed results are due to ultrasound guidance cannot be extrapolated.

Neurologic complications must increase following prolonged exposure to nerves according to lab studies but there has been conflicting evidence regarding this. While some studies have noted a higher than normal incidence of neurologic complications with the use of catheter in psoas compartment blocks, popliteal sciatic nerve blocks [77, 96], other studies note a very low complication rate [68, 71, 72, 95, 105, 113]. This may be related to the method of data collection and the definition of neuropathy. Future prospective data collection methods are needed to address this issue.

Although there are articles reporting low incidence of neurologic complications following PNB performed under general anesthesia [64, 110], there is limited information on whether blocks performed under general anesthesia increase the risk of postoperative neurologic dysfunction. A retrospective review by Bogdanov et al. [64] did not report neurologic complications following interscalene blocks performed under general anesthesia but two patients in the study by Watts et al. [110] reported long-term neurologic dysfunction. The details of whether these blocks were performed under sedation or general anesthesia are not known from the study. To date, there is no known pathological reason why general anesthesia would directly increase the patient’s susceptibly (host factor) in neurologic injury when receiving regional anesthesia. However, one would expect that general anesthesia would compromise the patient’s (environmental influences) ability to communicate and provide feedback of either early symptoms of LAST or paresthesia from needle–nerve contact. In a recent report, threshold currents that are needed to generate a motor response were higher in an anesthetized patient than those in awake patients. This observation may suggest that there is a possibility of potential error which can be made when using nerve stimulation to locate the nerve when a patient is under general anesthesia [146].

Nevertheless, the current ASRA advisory panel suggested that a conscious patient is preferred while performing PNBs unless in selected patient populations (e.g., dementia and developmental delay) where the risk-to-benefit ratio of performing regional anesthesia under general anesthesia may improve [147].

Lessons from Case Reports

Case reports identify the patient and performance characteristics, neurologic presentation, and subsequent outcomes. A total of 21 case reports/series reported on the occurrence of neurologic complication in 24 patients following PNB (Table 5.5). The majority was middle aged (Median age 50.5 years) and consisted of 12 males and 12 females. Only four of the 24 cases had some signs of intraneural injections while the rest of the cases did not mention the possibility. It is not only those with some form of subclinical or overt neuropathy (n = 5/24 patients) who are susceptible, but quite often it is an otherwise healthy patient who suffers this unfortunate complications. The presence of risk factors may be a bad prognostic sign since only two of these 5 patients had recovery of some nerve function after a prolonged period of time. The most common presentation was persistent weakness (16 cases) followed by pain and paresthesia (three cases) and a combination of both in the remaining. Only 4 patients had catheters placed while the rest had single shot blocks. A total of 12 patients did not have recovery of nerve function back to normal while the rest of the patients had recovery ranging anywhere from 1 week to 2 years. Five blocks were performed under US guidance while 11 cases utilized neurostimulation, 1 case used the combined US+NS technique, 1 case did not document the guidance method used, and 5 cases used the landmark/paresthesia technique.

Benumof [148] reported a case of spinal cord injury following an interscalene block performed under general anesthesia. This case report is an invaluable reminder of the risks associated with RA but is not strictly speaking PN injuries.

Analyzing Neurologic Injury from the Perspective of Disease Causation

Given their complexity, neurologic complications can best be evaluated by the same epidemiological principles of event causation (Fig. 5.1). The epidemiological triangle is a common injury model used to describe the relationship between an agent, a host, and the environment [14, 15]. A neuronal injury is more likely to occur when there is interaction between a susceptible host (inadequately protected nerve), an injurious agent (local anesthetic, needle, or injection pressure), and a hazardous working environment (poor supervision/guidance for locating needle, unsafe practices, unintended exposure). Elimination of one of the triangle’s components should, in theory, prevent the occurrence of the event. Hence, the safest approach appears to be identification of potential risk factors and prevention of their interaction.

Epidemiological Triangle

Host/Biological Factors

The history of neurologic complications is as old as the field of regional anesthesia itself. Early performers of regional anesthesia acknowledged both the possibility of neurologic complications following PNB [149, 150] and the lack of complications following deliberate needle–nerve contact [151]. Various anatomical, surgical, and patient factors may affect the incidence of postoperative nerve injury and include the type of surgery, associated comorbidities, the presence of preexisting neuropathy, and whether temporary or permanent injury is being considered.

Anatomy and Physiology

Not all nerves or nerve blocks are the same since intraneural fascicular topography shows wide variability (LOE 2b; Grade B). The connective tissue content of a peripheral nerve varies depending on the number of fascicles at a given site (LOE 2b; Grade B). Neural connective tissue and number of fascicles increase from proximal part of the nerve distally (LOE 2b; Grade B).

A total of three studies looked into the neural anatomy with relevance to PNB [115, 116, 120]. In most cases, a peripheral nerve is a mixed entity consisting of both sensory and motor components and has both myelinated and unmyelinated axons. Connective tissue covering the axons is present in different layers, providing support and nutrition to the nerves and acting as a protective barrier to the axon (Fig. 5.2). The three protective covers are the epineurium which covers the nerve overall and separates the fascicles, perineurium which lines the fascicles, and the endoneurium which lies inside the fascicles and surrounds the axons. The epineurium—the outer covering of the nerve—encases the fascicular bundles within a connective tissue network known as interfascicular epineurium. The adipose tissue in the interfascicular epineurium acts as a cushion for the fascicles and causes them to slide under or over a slowly advancing needle, protecting the fascicles from needle trauma. The fascicular bundle is in turn encased by multiple layers of cells, known as the perineurium, which act as a functional barrier for the axons and protects against physical and chemical insults. The perineurium bathes the axons in an interstitial fluid which is similar to CSF in composition and is continuous with the neuraxis [152, 153]. Inside the fascicle, myelinated or unmyelinated axons are supported by a network of connective tissue known as endoneurium which also contains the nonfenestrated capillaries that provide nutrition to these tissues. The endoneurium serves a vital role in nerve regeneration by aligning the regrowing axons toward its target. The perineurium maintains an intrafascicular pressure which is reflected in the intracellular pressure of the axons [154, 155]; thus, injection deep to the perineurium generally requires greater injection pressure compared to injection within the epineurium.

Fig. 5.2
figure 2

Electron micrograph of a peripheral nerve stained with osmic acid. (a) The entire nerve is encased in a connective tissue layer, the epineurium (Epi), and the nerve fibers are arranged in fascicles. (b) Each fascicle is surrounded by a cellular layer, the perineurium (red arrow). Blood vessels (BV) can be seen collapsed in the interfascicular epineurium. (c) Axons (Ax) within the fascicle are in an endoneurial network, interspersed with nonfenestrated blood vessels (BV). (Reproduced with permission from the Department of Anatomy and Cell Biology, Schulich School of Medicine and Dentistry, Western University, London, ON, Canada. http://slides.uwo.ca/spinal_cord.html)

Full size image

Nerve composition varies among different nerve types and also within a given nerve. Sunderland [152] noted that, in the upper limb, the fascicular topography of the radial, median, and ulnar nerves varied every 0.25–0.5 mm segment, and the branching pattern was not constant for a given nerve at a given site. While the sizes of individual fascicles are inversely related to their number at a given location along the nerve [152], the connective tissue content and cross-sectional area of a nerve are directly proportional [120]. This suggests that the amount of injury following intraneural injection depends not only on the characteristics of the insult but also on how protected a nerve is at the site of injection. Nerves are thought to be oligofascicular at the level of nerve roots and polyfascicular in areas prone to physical stress, such as the joints. Hence it is common to see hypoechoic (mono/oligofascicular) nerves at the level of roots (interscalene block) whereas they are hyperechoic (multifascicular) near a joint (popliteal nerve block). Moayeri et al. noted a proximal oligofascicular pattern progressing to a polyfascicular pattern in the brachial plexus [115] and sciatic nerve [116]; Sunderland and Ray [120] noted a wide variation in the fascicular pattern of the sciatic and forearm nerves with no consistent pattern in any part of the nerve. Whether neurologic complications are related to the fascicular morphology is currently unknown [97, 99] since proximal blocks (ISB, subgluteal sciatic nerve block) are known to have similar complications as distal blocks (popliteal sciatic, axillary brachial plexus block). Although the connective tissue content increases with age due to endothelial proliferation as a reaction to decreased vascularity of the nerves [156]. This may influence block onset and recovery, but its implications for neurologic injury are currently unknown. Since we did not anticipate any differences between cadaver and live tissue in terms of nerve composition, cadaver studies provided good evidence to support the earlier statements even in the absence of studies of live human tissue.

Surgical Factors

Certain types of surgery are associated with a higher risk of postoperative nerve injury (LOE 2b; Grade B). Peripheral nerve blocks do not increase the risk of postoperative neurologic dysfunction. (LOE 2b; Grade D).

Some surgeries are more prone to nerve injuries than others, especially those involving excessive neural stretch [157], trauma [158], inflammation [80], or ischemia [127] including a prolonged tourniquet time [82, 159]. In a retrospective review of 380,680 anesthetics during a 10-year period, Welch et al. [112] found a 0.3 % incidence of iatrogenic injuries. There was a significant association of iatrogenic injuries with certain types of surgeries, general anesthesia, and epidural anesthesia but a similar association was not found with peripheral nerve blocks. The lack of association between regional anesthetic nerve blocks and iatrogenic injuries is also confirmed by other studies in shoulder [65, 66, 144], knee [82], and hip surgeries [81]. Shoulder surgeries have a predilection for iatrogenic nerve injuries [13, 160] and the incidence can be as high as 8.2 % following anterior stabilization, around 1–4 % following shoulder arthroplasty or 1–2 % following rotator cuff repairs [161]. While Borgeat et al. [66] and Candido et al. [144] noted different incidences of persistent neurologic sequelae unrelated to surgery 1 month after ISB (7.9 % vs. 3.3 %), most of these complications were unrelated to ISB. Further, a retrospective review of 1569 patients undergoing total shoulder arthroplasty by Sviggum et al. also noted no such relationship between interscalene block and nerve injury [104]. While some studies indicate that the likelihood of complete recovery from peripheral nerve injury is lower when the patient had a PNB [82], other studies have not shown a similar association [82].

Neuropathy

Preexisting neuropathy is thought to increase the risk of postoperative neurologic dysfunction following PNB (LOE 5; Grade D). Neuropathic nerves are more prone to the prolonged effects of local anesthetics (LOE 5; Grade D).

Currently, there is no high-quality evidence regarding cause and effect of neurologic sequelae following nerve blocks but most anesthesiologists have a tendency to avoid PNB in patients with neuropathy. Although a retrospective cohort study [79] did not demonstrate worsening of neurologic outcomes following PNB in patients with preexisting neuropathy, a number of case reports [125, 128, 129, 132, 140, 143] indicate that either subclinical or overt preexisting neuropathy may make them susceptible to long-term nerve damage. Hence, the expert opinion regarding regional anesthesia in patients with neurologic disease tends to err toward caution [11, 162]. The degree of neural dysfunction in a chronically compromised nerve may be clinical or subclinical, and any secondary insults such as hypoxia or ischemia, local anesthetic neurotoxicity, or direct mechanical trauma following nerve blockade is thought to exacerbate it [162]. Importantly, the secondary insult need not be at the site of the neural compromise itself, a phenomenon known as “double-crush syndrome ” [163]. In fact, a double-crush injury in the form of two distinct low-grade insults has been shown to be more damaging to the nerve compared to an insult at a single site [164]. Thus, when suspecting underlying chronic neuropathy such as in patients with peripheral vascular disease, mechanical compression, metabolic derangements (diabetes mellitus) or postchemotherapy (cisplatin neurotoxicity), the decision to perform a PNB should be made on a case-by-case basis after thorough physical examination and discussion with the patient and the surgical team [162, 165]. It is generally thought that any evolving lesions or active inflammation of the nerves is a contraindication for PNB [162].

Two animal models of diabetic neuropathy have been tested for local anesthetic neurotoxicity [29, 35]. In the study by Kroin et al., local anesthetics produced a longer mean duration of sensory nerve block in diabetic rats versus nondiabetic rats [35]. Doses of lidocaine (with or without adjuvants) or ropivacaine that did not cause noteworthy nerve fiber damage in nondiabetic rats also failed to produce major pathology in nerves of rats with streptozotocin-induced diabetic neuropathy. The study by Kalichman [29] not only showed a lower conduction velocity in diabetic nerves, but also it had neuronal edema subsequent to extraneurally placed LA in a concentration-dependent fashion. This study along with others indicating that local anesthetic neurotoxicity is directly proportional to the dose and duration of local anesthetic exposure [59, 166], higher LA concentrations should be strongly discouraged for neuropathic patients and deliberate intraneural injections should be avoided based on conventional wisdom.

Causative Agent Factors

The insulting injury to a nerve can be as a result of direct needle trauma, pressure injury, or local anesthetic neurotoxicity. A majority of these factors have been evaluated in animal studies since human studies are not feasible due to obvious ethical concerns and hence most of the evidence is extrapolated to humans. It is difficult to judge as to which factor is the most damaging since most of the evidence originated from different animal models and more than one injurious agent may be evaluated in these studies.

Mechanical Agents

Needle Trauma

Nerve trunks usually slide under an advancing short-bevel needle compared to long-bevel needles (LOE 5; Grade D). Long-bevel needles cause more functional or histological damage compared to short-bevel, pencil-tip, or Tuohy needles but the superiority among the latter three needle types is currently unknown (LOE 5; Grade D). Needle gauge may in itself influence the degree of damage irrespective of needle type (LOE 5; Grade D). When short-bevel needles do penetrate the perineurium, the resultant nerve damage is greater than that of long-bevel needles (LOE 5; Grade D). The amount of damage is greater when the needle bevel is perpendicular to nerve fibers than when it is parallel (LOE 5; Grade D).

Eight animal studies and one cadaveric study evaluated the impact of needle design on nerve injury. The degree of nerve damage from needle trauma depends on the bevel type, the angle of needle insertion, and the needle size (gauge). Long-bevel (14° angle) needles penetrate fascicular bundles through the perineurium, while these fascicles slide under or away from short-bevel (45° angle) needles [17]. Animal [38] and human cadaver [119] studies demonstrate that injection with a long-bevel needle has a greater chance of being intrafascicular and resulting in nerve injury. One animal study showed that even in the absence of direct neural trauma, the presence of perineural hematoma might in itself result in inflammation and structural injury to the nearby nerves [48] and this has been implicated as a possible cause of injury in a case report [127]. Using cadaveric tissue , Sala-Blanch et al. [119] showed that, although fascicular contact is fairly common with intraneural injections, injury to these fascicles rarely occurs. Of the 134 fascicles contacted by the needle, only four were damaged, all from long-bevel needles. In animal studies, needles with a tapered end, such as Whitacre and Sprotte needles , are comparable to each other [37] and to Tuohy needles with respect to neural damage [37, 45, 46]. While two studies show superiority of tapered-tip needles over short-bevel needles in terms of neural damage caused [27, 37], and its effect on nerve conduction [27] another study reported similar neural perforations with tapered-tip and short-bevel needles of the same gauge [46].

The amount of nerve damage following intraneural needle placement is also higher when the bevel is inserted transversely to the nerve fiber compared to insertion along the long axis of the nerve [17, 27, 37]. Regardless of the type, needle gauge is directly proportional to the extent of nerve damage, as demonstrated by the stark difference in the extent of fascicular damage from 22G needles (3 %) and 17/18G needles (40 %) [45]. In general, short-bevel needles are preferred for PNB since they have difficulty penetrating perineurium; however, when short-bevel needles do penetrate the perineurium, the amount of mechanical trauma far exceeds that done by a long-bevel needle [42].

It is important to point out that basic science research using animals or cadaver tissue as a study model, such as the ones described earlier, were considered to be level 5 evidence and given a grade D recommendation irrespective of study design. This is because these studies arguably do not provide direct research evidence in live human subjects, although ethical issues and other difficulties obviously preclude doing these studies in live subjects. Nevertheless, the available evidence is quite convincing despite having a lower grade.

Pressure Injury

Perineural injections require the least injection pressure followed by extrafascicular injection, while intrafascicular injections generate high injection pressure (LOE 5; Grade D). While high injection pressures result in functional and histological nerve damage, intraneural injection with low injection pressures may not necessarily result in nerve damage (LOE 2b; Grade C).

The axons inside the fascicles are under pressure created by the perineurium and hence any injection into the perineurium will probably require higher injection pressure subsequently resulting in pressure injury. The evidence for pressure injury is purely based on animal models [9, 17, 21, 34, 53, 54, 167] and the human evidence is limited to studies looking at pressure monitoring during PNB [75]. In animal studies , low injection pressures (<25.1–27.9 kPa) are noted for injection performed around the nerve without penetration of the outer epineurium, while injection pressures increase slightly (69.8–86.5 kPa) upon entering the epineurium [53, 54]. Selander et al. [167], in a study of intraneural injection at different locations within the rabbit sciatic nerve, showed that a relatively low injection pressure (25–60 mmHg [3.3–7.9 kPa]) was required for subepineurial (extrafascicular) injections and resulted in limited spread of injectate, whereas intrafascicular injections required higher pressures (300–750 mmHg [39.9–99.7 kPa]) and resulted in rapid spread of injectate over long distances within the fascicle. To study the clinical consequence of such injections, Hadzic et al. [9] performed intraneural injections with 4 mL lidocaine in the canine sciatic nerve. Low-pressure (<4 psi) injections (3/7) were extrafascicular while high pressure injection (25–45 psi) (4/7) were intrafascicular in location which was similar to that noted by Selander et al. [167]. In a similar study design, Kapur et al. [34] showed that all intrafascicular injections resulted in clinical deficits in the form of paresis or disability while none of the extrafascicular injections resulted in any neural dysfunction. A study of ultrasound-guided deliberate intraneural injections in piglets with injection pressures <20 psi (~138 kPa) also showed that none of the injected nerves had a breach in the perineurium. Although the nerves showed signs of inflammation for up to 2 days postinjection and changes in nerve architecture under ultrasound for up to 4 days, none of the animals developed any functional deficits [21]. A similar evidence from a human study also showed that a low injection pressure during deliberate intraneural popliteal sciatic nerve block does not necessarily lead to early postoperative neurologic dysfunction [97] but further studies on injection pressure in clinical practice are needed. The pressure measurements following subepineurial injections are similar between those obtained by Vuckovic et al. [53, 54] and Hadzic et al. [9] but are higher than those reported by Selander et al. [167]. This could be related to differences in animal models, syringe, and injectate volumes used in the two studies. Although injection pressures <15 psi is recommended safe in clinical practice, this needs to be further validated.

The generation of high injection pressures during intrafascicular injection can be explained by the high intrafascicular pressure created by the perineurium and may also lead to pressure injury. The low injection pressures needed for perineural injection compared to subepineurial and intrafascicular injections show the potential utility of continuous monitoring of injection pressures during PNB . There is a need for further evidence regarding the short- and long-term safety of low-pressure intraneural injections.

Similar to studies related to needle design (see earlier), it would be difficult and unethical to perform studies in live humans to evaluate injury from high pressure injection. Thus, the published evidence is limited to basic science research using animals and cadaver tissue as study models. However, as with studies of needle design, the available evidence is fairly persuasive despite being assigned a lower grade.

Chemical Agents

Neurotoxicity

All local anesthetics are neurotoxic in increasing concentrations and individual local anesthetics differ in their neurotoxic potential (LOE 5; Grade D). Both extra- and intrafascicular injection of local anesthetic can result in histological damage, but is far greater following intrafascicular injection leading to functional injury as well (LOE 5; Grade D). Both epinephrine and local anesthetics decrease neural blood flow, and their combination has synergistic effects (LOE 5; Grade D).

A total of 21 studies evaluated the neurotoxicity of LA in different animal models. Broadly, the studies looked at comparative neurotoxicity of different LA solutions with or without adjuvants [25, 26, 44, 55, 58, 59], the impact of topical application of LA [22, 23, 29–33, 39, 40, 50, 57], or their intraneural injection [25, 26, 28, 35, 36, 44]. Intraneurally injected LA may often result in histological changes without any functional neuropathy [28, 35, 36]. While there is a general agreement over the increased amount of nerve damage following intrafascicular injection of LA as compared to topical application [44], whether or not LA solutions are more toxic than saline intrafascicularly is currently debated. While Farber et al. [25] in a study of Lewis rats noted intrafascicular injection of LA was more damaging than saline [25], a study by Selander et al. [44] on rabbits showed both saline and 0.5 % bupivacaine to cause equal amount of axonal damage. Although the amount of damage was greater with increasing concentrations of LA indicating that the pressure injury is far more damaging than LA neurotoxicity . The damage following intrafascicular injections is a result of a breach in the blood–nerve barrier and the loss of internal hypertonic milieu [25] compounded by pressure injury, interstitial edema, and direct neurotoxicity, resulting in clinical nerve damage.

At therapeutic doses , all local anesthetic agents exhibit neurotoxic potential [168] and, although debatable, some drugs may be more neurotoxic than others. The direct neurotoxicity of local anesthetics is thought to be related to prolonged increases in cytosolic Ca2+ leading to depletion of adenosine triphosphate, mitochondrial injury, membrane dysfunction, and, ultimately, cell death [169, 170]. Transient neurologic symptoms following spinal anesthesia are thought to represent a mild consequence of local anesthetic neurotoxicity [171], and transient neurologic symptoms following PNB may represent a similar event, with small-diameter axons (pain and temperature) being more affected than large-diameter axons (motor and proprioception) [172].

The neurotoxic effect of local anesthetics is time and concentration dependent in an animal study and in vitro models of cell cultures [59] but whether this holds true in human subjects is not known. While long-acting LA [85] and continuous catheters [6, 68, 72] have been employed safely with a low incidence of long-term nerve damage, some catheter studies [3, 77, 95, 96] and case reports [122, 125, 128, 140] do point toward a fairly high incidence of nerve dysfunction. While Capdevilla et al. [68] in a study of continuous catheters noted a low incidence of long-term neuropathy, bupivacaine infusion was one of the risk factors for the same along with ICU stay and age <40 years. Further prospective studies are needed to know whether prolonged exposure of nerves to different concentrations of LA is safe or neurotoxic.

The local anesthetic neurotoxic potential of individual agents differs depending on the animal model and study methodology but in general, most local anesthetics have vasoconstrictive properties and that includes the common agents such as lidocaine [39], levobupivacaine, and ropivacaine [23], hence making them both directly neurotoxic and have neuronal ischemic effects. Although bupivacaine has a vasodilatory effect on intraneural blood flow [22] and is thought to be less neurotoxic following intraneural injection according to one study [26], another study found it to be more neurotoxic than lidocaine or ropivacaine when injected into the fascicle [25]. Given that local anesthetic neurotoxicity is well documented, deliberate intraneural injection of local anesthetic is still strongly discouraged, despite the fact that most of the evidence comes from animal studies.

Adjuvants

Local anesthetics are more neurotoxic than adjuvants and, while some adjuvants may have neurotoxic potential, others may be neuroprotective (LOE 5; Grade D).

The neurotoxic potential of local anesthetics far exceeds that of any adjuvants used in regional anesthesia [57, 58], and effects on nerve tissue depend on the individual agent. While adjuvants, including opioids, clonidine, dexamethasone, and neostigmine, do not influence the neurotoxic potential of local anesthetics in vitro, drugs such as ketamine and midazolam may themselves be neurotoxic at higher doses [173]. On the other hand, dexmedetomidine was shown to be neuroprotective in rats following intraneural sciatic nerve injection [50]. It was postulated that dexmedetomidine decreased the neurotoxic potential of bupivacaine by decreasing mast cell degranulation at the site of injury. Nevertheless, the current evidence is limited to studies in animal models.

Intraneural Injections

Unintentional intraneural injections occur more often than previously expected (LOE 2b; Grade B), but they may not necessarily result in neurologic dysfunction (LOE 2b; Grade B). Intraneural injections have a rapid block onset (LOE 2b, Grade B).

Six trials studied the incidence of unintentional intraneural injection [73, 78, 91, 94, 98, 99]. Three were performed with the aid of nerve stimulation alone , one was done with ultrasound guidance alone, and two used dual guidance. The results showed that unintentional intraneural injection occurs frequently in both upper and lower limb blocks, with the incidence varying from ~17 % to as high as 66 % [73, 78, 91, 94, 98, 99]. Intraneural injections were also shown to hasten block onset [78, 94, 99], improve block success [108], and have also been shown to prolong block duration in animal models [34]. The incidence of needle nerve contact could possibly be higher with an out-of-plane (OOP) approach (64 % for femoral nerve block) [98] but whether or not this results in an increased incidence of intraneural injections is currently unknown. OOP approaches although have not been shown to increase the incidence of neurologic complications [3].

Irrespective of unintentional or targeted intraneural injections using either low current neurostimulation or US guidance, none of the trials reported long-term postoperative neurologic dysfunction related to PNB [62, 63, 78, 94, 97–100, 108]. However, the follow-up period in some of these studies was not long enough to allow symptoms to develop, and none of the studies were sufficiently powered to assess the incidence of neurologic dysfunction or nerve injury. Hence, it cannot be recommended as safe practice to perform deliberate intraneural injections until data from larger studies are available.

Five studies investigated deliberate intraneural injection [62, 97, 100, 108]. In each one, ultrasound was used to identify intraneural injection, and one study used nerve stimulation in addition to ultrasound [97]. A 10 % incidence of transient neurologic deficit was observed in one of the studies [63], and another study evaluating the deliberate intraneural injections performed under ultrasound versus neurostimulation showed an increased success rate with US but resulted in a higher incidence of paresthesia [101]. None of the studies revealed any increase in neurologic complications during follow-up (1–4 weeks after the procedure). A cadaveric study of interscalene blocks reported a 50 % incidence of subepineural injection when the needle tip was placed adjacent to the brachial plexus trunks [117]. While the results of these studies do not imply that intraneural injection is a safe procedure, they do show that it is a fairly common occurrence and does not always lead to neurologic complications.

The take-home message is not to think that deliberate intraneural injections are safe to perform but to think that it is fairly common in clinical practice to note intraneural injections and it does not necessarily result in neurologic complications. The occurrence of neurologic complications may increase following intrafascicular (subperineural) injections but currently most of the evidence for this is based on animal studies and case reports.

Environmental Influences

The time-honored statement that “an ounce of prevention is worth a pound of cure” is essential when considering the ways to minimize adverse outcomes following intraneural injection. To help reduce or prevent the possibility of intraneural injection, an effective method of detecting and monitoring the presence and extent of intraneural injection is critical, as is the skill and willingness to use it in regional anesthesia practice.

Nerve Stimulation

When used at low currents, nerve stimulation has low sensitivity but high specificity for detecting proximity of the needle tip to the target nerve (LOE 2b; Grade B). Nerve stimulation cannot differentiate between intraneural needle placement and needle–nerve contact (LOE 5; Grade D). Higher stimulating currents are required in diabetic patients for detecting intra- and extraneural needle placement (LOE 2b; Grade C).

For electrical nerve stimulation , the minimal stimulating current intensity is proportional to the square root of the distance between the needle tip and the nerve, provided there is a constant magnitude of charge between the two points. In animal studies, a low stimulating current requirement (<0.2 mA) was originally shown to correlate with histological evidence of nerve injury in 50 % of the study animals, while current intensity >0.5 mA implied extraneural placement [52]. A similar study in humans employing noninsulated needles showed that the median (Range) stimulating current noted when a deliberate paresthesia is obtained was 0.17 (0.03–3.3 mA) [70]. This led to the popular practice of eliciting motor response at stimulating currents between 0.2 and 0.5 mA and deliberately withdrawing the needle when stimulation is obtained at currents <0.2 mA. A number of studies later showed the inaccuracies of neurostimulation both at low and high current stimulation. Even the studies which established the notion that an MSC of <0.2 mA was specific but not sensitive indicator of intraneural needle placement possibly had extraneural needle placements as evidenced by an extraneural injection in 50 % of injections in the animal study [52] and the wide range of MSC noted with the human study [70]. Animal studies have shown that higher stimulating currents are sometimes needed to elicit a motor response following intraneural needle placement [20, 24, 174]. The same phenomenon was observed in 16.7 % of patients receiving deliberate low-pressure intraneural injections during popliteal sciatic nerve block [97]. On the contrary, low stimulation currents have been employed for performing sciatic nerve block [83] and infraclavicular block [84] without evidence of nerve damage.

Recently, Weismann et al. [56] showed that a low stimulating current may indicate either needle–nerve contact or intraneural placement. Hence, a low stimulating current, if present, may only indicate that the needle tip is too close to or within the nerve, rather than differentiating between the two. The noncorrelation of needle tip location and nerve stimulation is due to a variety of factors influencing motor response following stimulation. The stimulating current is influenced by pulse width, interaction of the needle tip with the fascicles, and the degree to which a depolarization or hyperpolarization occurs as a result of the stimulating current [175–177]. Since the minimal stimulating current for each nerve is different [178], a single value cannot be extrapolated for all nerves.

Evidence regarding whether or not diabetic individuals require a higher stimulation threshold is evolving. In animal models of hyperglycemia , when a low stimulation threshold was used to guide the needle, all injections were intraneural, while none of the low current stimulation injections in normoglycemic animals had the same pattern of injectate dispersion [43]. A significant number of diabetic patients undergoing supraclavicular brachial plexus block required a higher stimulation threshold when the needle was placed perineurally (57 % required currents >1.0 mA vs. 9 % nondiabetic) or intraneurally (29 % required currents of 0.5–1.0 mA vs. 2 % nondiabetic) [63]. It has been reported and is worth pointing out that it also has been that the threshold currents used for motor response from nerve stimulation under general anesthesia might be higher than those in awake patients [146]. Thus, their result also suggested that using nerve stimulation as a technique to warn for intraneural placement in patients under general anesthesia may require different parameters compared with patients who are not under general anesthesia.

Injection Pressure Monitoring

High injection pressures are often reached unknowingly by experienced and nonexperienced practitioners (LOE 2b; Grade B). Syringe feel is inaccurate for differentiating tissues, and higher pressures are generated unknowingly (LOE 5; Grade D). Injection pressure can be kept within safe limits reliably by using compressed air injection technique (CAIT) or pressure measurement devices (LOE 2b; Grade C). Opening pressure can detect needle nerve contact reliably in interscalene block (LOE 2b; Grade C).

While intrafascicular injections require higher injection pressures, a low injection pressure has a good negative predictive value for neurologic dysfunction [21, 97]. Two important pressures to monitor when performing a PNB are the opening pressure (OP) and injection pressure (IP). The OP is the pressure in the needle–tubing–syringe assembly before the injectate begins to flow through the needle. A high OP (>20 psi) has been shown to correlate with nerve damage [75]. Once flow has begun, IP at the needle tip depends on various factors, including needle size, length of tubing, and syringe volume. Avoiding high IP is as important as OP in preventing further damage from injectate flow into the perineurium. Simple “syringe feel ” is inaccurate in determining what tissues the performer is injecting into, irrespective of operator experience as shown in an animal model where only 12 of 40 anesthesiologists (30 %) correctly identified intraneural injection using “syringe feel” [107]. Anesthesiologists also vary widely in their perception of injection pressure and the speed of injection. In a study of 30 anesthesiologists performing simulated injections in a lab model, a 20-fold variability in baseline injection pressure and speed of injection was noted. When resistance was increased gradually in a blinded fashion during injection, 70 % of anesthesiologists exceeded the recommended injection pressure of 20 psi [109, 114].

The inaccuracy of “syringe feel” and a wide variability in baseline perception of the performer has led to the use of objective methods and devices to monitor injection pressure during PNB performance. These include the compressed air injection technique (CAIT) [109, 121] and B.Braun’s BSmart™ injection pressure monitor. When using CAIT , a set volume of air is drawn into the syringe containing the injectate, and the air is compressed to a certain percentage of its initial volume when injecting. In vitro evaluation of this technique has been shown to ensure injection pressures substantially below the threshold considered significant for nerve injury, irrespective of the needle or syringe type when the air compression was ≤50 % of the original volume. Currently, no animal or clinical studies have evaluated the technique, so its impact on clinical outcomes is unknown. Recently, the use of the BSmart™ device in patients (n = 16) undergoing ultrasound-guided interscalene brachial plexus block consistently (97 %) revealed an opening pressure of ≥15 psi at the time of needle–nerve contact [75]. Nevertheless, the specificity of using injection pressure monitoring to avoid intraneural needle placement is still suspect. High injection pressures can be caused by contact with fascia, tendon, or bones. Moreover, needle tip pressure may be dependent on the needle–syringe combination [179].

Ultrasound

Ultrasound guidance can detect intraneural injection and is dependent on operator experience (LOE 2; Grade B). Use of ultrasonography does not prevent intraneural injection (LOE 2; Grade B). Long-term neurologic complications following PNB have not declined as a result (LOE 2b; Grade B).

Ultrasound can be a useful tool for avoiding and detecting intraneural needle placement and injection but is not foolproof in preventing intraneural injection. Currently available ultrasound technology cannot differentiate between the different layers of the nerve and therefore cannot distinguish between inter- and intrafascicular injection. Possible ultrasonographic indicators of intraneural injections include visualization of the needle tip within the nerve, increase in the nerve cross-sectional area by at least 15 %, spread of local anesthetic within the epineurium upon proximal-to-distal scanning, and real-time visualization of fascicle separation on injection. It is important to note that, if any of these signs is observed on ultrasound, intraneural injection has already occurred.

When performing PNB, the needle tip is often not visualized on ultrasound, and needle advancement without proper needle tip visualization is a common error that persists even after adequate experience. Surrogate markers , such as increase in cross-sectional surface area or local anesthetic solution found between the fascicles, are therefore used to monitor for intraneural injection. The occurrence of unintentional intraneural injections during ultrasound-guided PNB has been noted frequently in cadaveric studies [117] and the clinical setting [63, 78, 91, 98] and is most likely due to dependence on the practitioner’s expertise in detecting intraneural needle placement or injection. In a study of assessment of intraneural injection by novices and experts, the sensitivity of detecting a low volume (0.5 mL) intraneural injection was 65 % in novices and 84 % in experts, but the specificity of assessment was 98 % irrespective of the level of expertise [86]. Although Bigeliesen et al. [63] showed that intraneural needle tip placement was detected reliably in only 69 % of cases, surrogate markers of intraneural injection (e.g., increase in cross-sectional area of nerve) can detect intraneural injections reliably (94 %) [93, 100]. Ruiz et al. [98] evaluated whether an in-plane (IP) approach to femoral nerve block was better than an out-of-plane (OOP) approach for avoiding needle–nerve contact and intraneural injection. Although they noted a higher incidence of intraneural injections with an OOP approach (64 % vs. 9 % IP), their definition of intraneural injection was the presence of local anesthetic below the nerve, rather than visualization of intraneural needle tip or injectate placement on ultrasound. This, combined with the lack of evidence from other types of PNBs, suggests that further study is needed to conclude with certainty that OOP approaches increase the chances of needle–nerve contact and intraneural injection.

Orebaugh et al. [4, 5] investigated whether the use of ultrasound has led to a decrease in neurologic complications. In both retrospective reviews, no differences in long-term neurologic complications were found between blocks performed under nerve stimulation or ultrasound guidance. Electromyography detected nerve injury following nerve stimulation-guided block in 3/3290 cases, but no long-term neurologic injuries were detected following ultrasound-guided blocks (0/2146). An update in 2012 showed the incidence of nerve injury lasting 6–12 months was significantly higher with nerve stimulation alone (4/5436) compared to ultrasound guidance (1/9069), but no significant difference in the incidence of long-term injuries (>1 year) was observed between the two groups (3/5436 nerve stimulation vs. 0/9069 ultrasound). This has also been supported by a prospective study by Liu et al. [92]. Although the underlying reason(s) for not seeing a reduction in complications despite the increasing use of ultrasound in regional anesthesia practice is unclear, it may explained in part by the old adage, “A tool is only as good as the person using it,” which is highly applicable when it comes to using imaging technologies such as ultrasound.

Monitoring neurologic outcomes following regional anesthesia.

To monitor and manage patients effectively with possible peripheral nerve injury following regional anesthesia, it is important to have a basic understanding about classification and the pathophysiology of neurologic injuries.

Pathophysiology

The overall clinical course of pathophysiology of peripheral nerve injury usually takes 2–4 weeks to manifest and progress [180, 181] for most nerves. However, there is a primary histological change involving physical fragmentation of both axons and myelin, a process that begins within hours of injury (Wallerian degeneration) occurring at the axon distal to the site of injury [181]. For the portion of the nerve proximal to the injury, it also undergoes a retrograde degeneration. Eventually, the axons in the endoneurial network undergo chromatolysis and are replaced by Schwann cells. The process of recovery begins after 4–6 weeks, and the integrity of endoneurial network is crucial at this recovery phase and correlates with clinical recovery (see the section on practical aspects below). If the endoneurium is intact, the regenerating axons grow into them and are subsequently myelinized by the Schwann cells. If there is a disruption of endoneurial network, the regenerating axons grow aimlessly in all directions, resulting in a neuroma. The classification of nerve injury and its subsequent course is described in Table 5.7. For practical purposes, Sunderland’s classification is used to classify and predict outcomes.

Table 5.7 Classification of nerve injury
Full size table

As presented in Table 5.7, nerve injury is not necessarily synonymous with clinical complications and at times may not lead to any detectable clinical symptoms or signs. In other words, the injury may lead to subclinical complications with no overt clinical manifestations. Individuals who present with neurologic symptoms and sequelae may therefore only represent the tip of the iceberg (Fig. 5.3). Thus, it is important to consider and interpret carefully the evidence regarding the incidence of clinical neurologic complications.

Fig. 5.3
figure 3

Schematic diagram of relationship of injury and clinical symptoms

Full size image

Practical Points in Mechanism of Nerve Injury

A neuronal injury is more likely to arise when a negative interaction between a susceptible host (inadequately protected nerve), an injurious agent (local anesthetic, needle, or injection pressure), and a hazardous working environment (poor supervision/guidance for locating needle, unsafe practices, unintended exposure) occurs. Risk stratification by minimizing one of the triangle’s components should, in theory, preclude the manifestation of the event. Hence it is vital to choose a technique tailored to each patient’s existing physiology (nonmodifiable risks) as delineated earlier. The clinician should attempt to minimize all modifiable risks such as needle trauma, pressure injury, and LA neurotoxicity using appropriate monitoring techniques and safe practices. A clear understanding of the procedure by the patient and good communication between the clinician and the patient is vital to detect iatrogenic injury either during performance of the block or in the recovery period. Hence we recommend the following practice points which may help in early identification of neurologic outcomes :

  • Preoperative assessment and documentation of neurologic function (Identify at-risk patient)

  • Clear communication with the patient regarding the block procedures and postoperative recovery of sensory and motor function

  • Minimal sedation during the performance of PNB to permit patient–clinician communication.

  • Use of all available monitoring technique during the performance of PNB. We routinely use US + NS guidance (0.2 mA) for needle placement and employ CAIT for injection pressure monitoring.

  • Close monitoring and adequate follow-up in the event of procedural paresthesia/signs of intraneural injection to ensure recovery of neurologic function

  • Use dilute LA solutions in high risk patients (i.e., preexisting neuropathy and presence of surgical risk for compartment syndrome).

  • Early neurology referral in those patients with red flags for iatrogenic nerve injury.

Classifying and managing patients with neurologic injury can be challenging given that a widely accepted algorithm is lacking for monitoring neurologic recovery following PNB. We present a simplified algorithmic approach for follow-up of peripheral nerve blocks (Fig. 5.4). Most common symptoms following neurologic injury are sensory changes such as persistent numbness, pain, or persistent paresthesia/dysesthesia in the distribution of the nerve block. The presence of motor weakness out of proportion to that from PNB or after the discontinuation of the block should prompt early referral after ruling out mechanical causes such as tight surgical dressing/tourniquet injury. Evolving sensory/motor lesions also mandate early referral since neurologic deficits arising within the first 24 postoperative hours likely represent acute injury. The routine practice in the majority of institutions includes a follow-up visit or phone call on POD-1 to ensure the resolution of block following discontinuation but, many of the sensory-motor disturbances arise several days to a couple of weeks following PNB and such cases need to be referred to neurology for evaluation if it does not resolve within 4–6 postoperative weeks. Neurologists commonly perform nerve conduction studies, evoked potentials, and electromyography which identifies the site of lesion and the timing of injury thereby helping in the diagnosis and prognosis of injury. These tests are invasive procedures and are not without limitations. Nerve conduction studies are useful in evaluating large sensory-motor nerve fibers while unmyelinated fibers may be missed. EMG requires several weeks of denervation before changes can be detected. Hence cases wherein an evolving/nonresolving lesion is suspected or motor weakness is present are referred to neurology and the majority of cases with mild sensory disturbances are managed conservatively with follow-up.

Fig. 5.4
figure 4

Pathway to classify and manage neurological injury following peripheral nerve blocks

Full size image

Conclusion

In summary, long term neurologic complications following regional anesthesia are rare and are usually a result of an interplay between the host (patient) factors, causative agents (mechanical and chemical), and environment (regional anesthesia tools and methods). Many of the factors responsible for the neurologic complications are nonmodifiable and hence screening for at-risk patients is necessary. Unintentional intraneural injections are thought to occur frequently during PNB and intraneural injections may not necessarily result in neurologic complications as long as they are extrafascicular. Most of the evidence for neurologic injury following PNB such as needle design, pressure monitoring, and local anesthetic neurotoxicity arises from animal models and their findings are being extrapolated to clinical practice. Evidence from animal experiments indicates that intrafascicular injections used with higher injection pressures are more likely to result in nerve injury. While technological improvements in regional anesthesia practice continue to improve our ability to detect and prevent nerve damage, preparation, vigilance, and careful observation remain a regional anesthesiologist’s most important tools in ensuring patient safety.