ORIGINAL ARTICLE

Incidence of Local Anesthetic Systemic Toxicity and

Postoperative Neurologic Symptoms Associated With 12,668
Ultrasound-Guided Nerve Blocks
An Analysis From a Prospective Clinical Registry
Brian Daniel Sites, MD, Andreas H. Taenzer, MS, MD, Michael D. Herrick, MD, Constance Gilloon, MD,
John Antonakakis, MD, Janeen Richins, MD, and Michael L. Beach, MD, PhD
per 1000).3 This incidence does not seem to be dramatically
Background and Objectives: There are varying reports on the in- different from historical frequencies reported using conventional
cidence of major morbidity associated with peripheral regional anes- nerve localization techniques. For instance, in 43,946 patients
thesia. Our objective was to contribute to the knowledge regarding the identified by organizational self-reporting, Auroy et al4 reported
incidence of local anesthetic systemic toxicity and postoperative neu- a 0.14-per-1000 rate of PONS lasting longer than 6 months.
rologic symptoms in the setting of ultrasound-guided peripheral regional With respect to LAST, there are currently only 2 large
anesthesia. clinical registries in the peer-reviewed literature that contain
Methods: During an 8-year period, 12,668 patients undergoing pe- patients receiving UGRA.5,6 In a mix of both ultrasound-guided
ripheral regional anesthesia were evaluated. Using a clinical registry, and nerve stimulatorYguided blocks (n = 7156), Barrington et al5
incidence rates of postoperative neurologic symptoms, local anesthetic reported an overall LAST incidence of 0.98 per 1000 (95% CI,
toxicity, pneumothorax, and vascular trauma were calculated. Univariate 0.42Y1.9). Rates of LAST events were not statistically different
analysis was used to identify risk factors for postoperative neurologic between the nerve stimulation and the ultrasound subgroups.
symptoms. We defined postoperative neurologic symptoms as any sen- With respect to the overall incidence of LAST, the findings of
sory or motor dysfunction present for more than 5 days and anatomically Barrington et al are consistent with those of Auroy et al who
consistent with the possibility of contribution from the nerve block. documented a rate of 0.8 per 1000 associated with nonultrasound
Results: The incidence (per 1000 blocks) of adverse events across techniques.4 The second clinical registry belongs to Orebaugh
all peripheral regional anesthetics was 1.8 (95% confidence interval et al,6 who reported a reduction in LAST associated with the
[CI], 1.1Y2.7) for postoperative neurologic symptoms lasting longer use of ultrasound. They reported 0 LAST events in 2146 blocks
than 5 days, 0.9 (95% CI, 0.5Y1.7) for postoperative neurologic when ultrasound and nerve stimulation were used. This stands
symptoms lasting longer than 6 months, 0.08 (95% CI, 0.0Y0.3) for in contrast to 5 LAST events in 3290 blocks when landmark-
seizure, 0 (95% CI, 0Y0.3) for pneumothorax, 0.6 (95% CI, 0.2Y1.2) for based nerve stimulation techniques were used alone.
unintended venous puncture, 1.2 (95% CI, 0.7Y2.0) for unintended Our fundamental objective was to contribute to the knowl-
arterial puncture, and 2.0 (95% CI, 1.2Y3.0) for patients having unin- edge regarding the incidence of LAST and PONS in the setting
tended paresthesia during block placement. There were no cardiac arrests. of a dedicated ultrasound-guided peripheral regional anesthesia
Conclusions: In the setting of a surgical procedure, ultrasound-guided practice. To this end, we analyzed our prospectively maintained
regional anesthesia is associated with the risk of long-term postoperative clinical registry consisting of 12,668 ultrasound-guided pe-
neurologic symptoms. Local anesthetic systemic toxicity, however, is ripheral regional blocks. In addition to the incidence rates of
extremely uncommon. major morbidity, we report the incidence of unintentional vas-
(Reg Anesth Pain Med 2012;37: 478Y482) cular puncture, pneumothorax, and unintentional paresthesia.

METHODS
L ocal anesthetic systemic toxicity (LAST) and postoperative
neurologic symptoms (PONSs) are considered the 2 major
morbid events associated with peripheral regional anesthesia.1,2
Data were analyzed for patients receiving a peripheral
regional anesthetic at Dartmouth Hitchcock Medical Center
between July 2003 and February 2011. After approval by
Although rare, these events continue to be reported.3 Dartmouth College’s Committee for the Protection of Human
Large prospective data collections examining these events Subjects, we queried our prospective clinical registry. All pe-
in adult patients are sparse. With respect to PONS, the frequency ripheral regional anesthesia data were prospectively gathered
during ultrasound-guided regional anesthesia (UGRA) has been and entered into an electronic database by members of the re-
reported as 0.4 per 1000 (95% confidence interval [CI], 0.08Y1.1 gional anesthesia service. These members included a group of
specialized staff physicians as well as rotating fellows and
From the Department of Anesthesiology, Dartmouth-Hitchcock Medical
residents. Data entry was composed of peripheral nerve blocks
Center, Lebanon, NH. done in the block area, operating room, and recovery room. The
Accepted for publication April 26, 2012. database platform was based on Microsoft Access (Microsoft
Address correspondence to: Brian Daniel Sites, MD, Dartmouth-Hitchcock Corporation, Redmond, Washington) and consisted of pre-
Medical Center, Department of Anesthesia and Pain Management, One
Medical Center Dr, Lebanon, NH (e-mail: brian.sites@hitchcock.org).
defined fields mandating direct entry or selection from drop-
The authors declare no conflict of interest. down menus. Our database tracked patient demographics, block
Funding provided by the Department of Anesthesiology, Dartmouth-Hitchcock characteristics, operator characteristics, and morbidity events.
Medical Center, Lebanon, New Hampshire. During the study period, there were between 5 and 6 full-
Copyright * 2012 by American Society of Regional Anesthesia and Pain
Medicine
time anesthesiologists responsible for the regional anesthesia
ISSN: 1098-7339 service. Our regional anesthesia practice is an academic model
DOI: 10.1097/AAP.0b013e31825cb3d6 in which fellows and residents are under the supervision of staff

478 Regional Anesthesia and Pain Medicine & Volume 37, Number 5, September-October 2012

Copyright © 2012 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
Regional Anesthesia and Pain Medicine & Volume 37, Number 5, September-October 2012 Morbidity of UGRA

anesthesiologists. A staff anesthesiologist has discretionary au- a total knee arthroplasty and femoral nerve block would not
thority regarding when to take over a regional anesthetic pro- have been included as a PONS event. Our clinical registry did
cedure from a trainee. In addition, advanced fellows were allowed not track, nor did it try to determine, relative contributions
to supervise regional anesthetic procedures by residents. Medical of various possible etiologies for PONS. We further categorized
students and certified nurse anesthetists do not participate in PONS as long term or possibly permanent if it was present
regional anesthetic procedures at our institution. For the dura- for more than 6 months (PONS-L). Local anesthetic systemic
tion of the study period, standing sedation orders were in place toxicity was defined as any event in which the patient experi-
for patients having regional anesthesia. These orders consisted enced unconsciousness, arrhythmias, seizure, or cardiac arrest
of 1 to 4 mg of intravenous midazolam and 25 to 100 Kg of associated with the injection of local anesthetic. Unintentional
intravenous fentanyl, which were titrated for patient comfort vascular puncture was considered to have occurred when blood
and anxiolysis. All of the nerve blocks were performed using was aspirated or flowed in the extension tubing connected to
ultrasound guidance, plus or minus the use of nerve stimulation. the block needle. Arterial blood was presumed to be bright red
For a given nerve block, our practice uses 1 or a combination in comparison to darker venous blood. All immediate compli-
of the following local anesthetics: 1% lidocaine, 1.5% lidocaine, cations, such as LAST, clinically symptomatic pneumothorax,
2% lidocaine, 2% chloroprocaine, 3% chloroprocaine, 0.2% unintentional vascular trauma, and unintentional paresthesia were
ropivacaine, 0.5% ropivacaine, 0.25% bupivacaine, and 0.5% entered into the database in real time.
bupivacaine. None, 1, or more of the following additives were
added to the local anesthetic: clonidine, dexamethasone, epi- Statistical Analysis
nephrine, and nalbuphine. All analyses were done using STATA (StataCorp LP, Col-
For inpatients, postoperative evaluation consisted of a daily lege Station, Texas). Continuous variables are reported as mean
visit by a physician member of the regional anesthesia team. (SD). Morbidity rates are reported as mean per 1000 blocks, with
Sensory and motor examinations were conducted daily on these a corresponding 95% CI. Categorical exposures (sex, block type,
patients. For ambulatory patients, a telephone call evaluation was local anesthetic type, additives, paresthesia, neuropathy, chronic
conducted within 3 days of discharge. This telephone call was pain) that were potential risk factors for PONS were analyzed
conducted by a nurse member of the regional anesthesia team using a Fisher exact test. Continuous exposures (age and needle
and was targeted to confirm sensory and motor block resolution. attempts) that were potential risk factors for PONS were analyzed
In addition, we tracked patients about whom we were made aware using a 2-sample t test. Odds ratios were calculated as appro-
of possible PONS by patient self-reporting or surgeon notifi- priate. Two-tailed P e 0.05 were considered statistically signif-
cation. If PONS was suspected, the patient was followed up icant without adjustments for multiple comparisons.
by the surgeons, and recommendations were made for further
medical consultation and diagnostic testing. Beyond the afore-
mentioned process, we did not have a systematic examination RESULTS
process to identify nonYpatient-reported and nonYsurgeon-reported We analyzed data of a total of 12,668 patients in our da-
possible PONS. Table 1 summarizes our postoperative follow- tabase. Men constituted 49.9% of the population with missing
up process and provides an estimate of the rate of success of sex data on 197 patients (1.55%). The mean age was 55 (17)
this process. years. The mean weight was 88 (23.5) kg. There were 184 blocks
Postoperative neurologic symptom was defined by 2 criteria. performed on patients 16 years old or younger. Additional
First, there had to be a patient-reported or evaluator-identified characteristic of our cohort can be found in Table 2.
sensory or motor dysfunction present at a minimum of 5 days after The most commonly performed blocks were single-injection
surgery. Second, the neurologic dysfunction had to have an an- femoral (34.6%), single-injection interscalene (15.8%), single-
atomic basis to support the possibility of a block contribution. injection supraclavicular (11.9%), and single-injection lateral
As an example, a common peroneal nerveYrelated foot drop after popliteal (7.7%) blocks, making up almost 70% of all blocks

TABLE 1. Postoperative Patient Evaluation

History Physical Audit Results, %

Event Taken Examination Potential Action Taken for Positive Finding Success (n)*
Nurse telephone call for Yes No Regional team investigation and database entry 93 (76)
ambulatory patients†
In-patient anesthesia Yes Yes Entry into database 90 (112)
visit‡
Ambulatory postoperative Yes Yes Communication by surgical team member to regional team 98 (195)
surgical clinic visit§ member and entry into database
*Our audit consisted of 200 randomly selected patients. There were 82 ambulatory patients and 118 inpatients. We performed electronic chart reviews
to indicate whether the patient had a documented follow-up interaction. Audit results were based on the year 2006 and later because this time interval was
when the nursing postoperative assessment was available electronically. Before this, the nursing follow-up was on a paper record.
†Questions posed to patients on telephone call: (1) When did the block wear off? (2) Do you have any residual numbness, tingling, or unusual
sensation in the area of the block? (3) Do you have any weakness? 4() Please rate your satisfaction with your regional anesthesia experience.
‡The regional anesthesia team conducted daily clinical rounds and documented a history and physical examination on patients who underwent
a peripheral regional anesthetic. The questions asked of inpatients were designed to match that of the nursing ambulatory follow-up telephone call. All
potential complications were recorded and entered into the database.
§All patients were scheduled for ambulatory postoperative surgical follow-up. At this visit, patients received postoperative surgical histories and
physical exams. We worked closely with our surgical colleagues to receive electronic communication regarding potential cases of postoperative
neurologic symptoms.

* 2012 American Society of Regional Anesthesia and Pain Medicine 479

Copyright © 2012 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
Sites et al Regional Anesthesia and Pain Medicine & Volume 37, Number 5, September-October 2012

(95% CI, 0.0Y0.4) for seizure, 0 (95% CI, 0Y0.3) for pneumo-
TABLE 2. Characteristics of Patients thorax, 0 (95% CI, 0Y0.3) for wrong block location, 0.6 (95% CI,
0.2Y1.2) for unintended venous puncture, 1.2 (95% CI, 0.7Y2.0)
Characteristic* n % for unintended arterial puncture, 2.0 (95% CI, 1.2Y3.0) for
Hypertension 3832 30 patients having unintended paresthesia during block placement,
CAD 1391 11 and 0 for cardiac arrest (95% CI, 0Y0.3).
Diabetes 1398 11 Using univariate analyses, we found no correlations be-
Chronic Pain 831 7 tween the risk for PONS and paresthesia at time of block
Neuropathy 788 6 placement (P = 0.9), chronic pain (P = 0.2), preexisting neu-
ropathy (P = 0.5), sex (P = 0.2), age (P = 0.8), type of local
COPD 691 5
anesthetic used (P = 0.2), use of block additives, such as clo-
Depression 502 4 nidine (P = 0.8), or the number of block attempts (P = 0.99). In
Renal disease 429 3 addition, a subanalysis of PONS-L revealed that none of these
Ambulatory status† 82 41 factors were associated with an increased risk for injury. There
*Information entered into registry based on medical record and patient was a trend toward an increased incidence of PONS-L in the
interview. continuous nerve blocks (2.4/1000) compared with single-
†Based on a random audit sample of 200 patients. Ambulatory status injection nerve blocks (0.7/1000; odds ratio [OR], 3.5; 95% CI,
was not tracked within the clinical registry. 0.6Y14.4; P = 0.08).
CAD indicates coronary artery disease; COPD, chronic obstructive We conducted an exploratory analysis to determine if there
lung disease were specific blocks that were associated with a higher incidence
of PONS. Interscalene blocks (continuous and single injection
combined) were associated with an elevated risk of long-term
performed. Continuous catheters accounted for 9.8% (femoral PONSs (3.1/1000; OR, 6.6 [95% CI, 1.8Y26.3; P = 0.002]
7.5%, interscalene 1.8%) of all blocks. 68% of blocks were compared with not having an interscalene block 0.5/1000).
performed by a resident, 25% by a fellow, and 7% by a staff Two patients had persistent paresthesia in the first to third
anesthesiologist. fingers after interscalene blocks and shoulder arthroscopies. One
Of the 12,668 blocks, a single local anesthetic was used patient had persistent numbness in the fifth finger and forearm
in 10,251, 2 different local anesthetics were used in 2343, and after a shoulder arthroscopy with an interscalene block. There
information is missing for 74. Of the blocks that were performed was 1 report of ear and jaw numbness that persisted after an
with a single local anesthetic, bupivacaine was used in 7564 interscalene block for a total shoulder arthroplasty. Also, after
(73.8%), ropivacaine was used in 2433 (23.7%), lidocaine was an interscalene block for a total shoulder arthroplasty, 1 patient
used in 238 (2.3%), and chloroprocaine was used in 16 (0.2%). with preexisting vitamin B12 deficiency neuropathy reported de-
All blocks were placed with ultrasound imaging. In 4972 blocks creased sensation in the lateral antebrachial cutaneous dermatome
(39.25%), a nerve stimulator was used, in addition to the ultra- and the inability to extend the thumb. After an interscalene block
sound. A local anesthetic additive was used in 4949 (39.1%) of for a shoulder arthroscopy, 1 patient developed biceps weakness
the nerve blocks. The mean total volume of local anesthetic and an electromyogram (EMG) finding of a root-level brachial
across all blocks was 25.2 (9.5) mL. Of all blocks, 3187 (25%) plexus injury. As previously reported, 1 patient, after an inter-
blocks were placed as primary surgical anesthetics, with the re- scalene block and total shoulder arthroplasty, sustained a per-
maining 9455 (75%) placed primarily for postoperative analgesia. manent brachial plexus injury with incapacitating motor and
Table 3 summarizes the incidence of PONS associated with sensory loss.7 After a popliteal sciatic block for resection of a
our regional anesthesia practice. The incidence (per 1000 blocks) Haglund deformity, 1 patient had decreased plantar flexion and
of adverse events across all peripheral regional anesthetics was sensory loss in the tibial and common peroneal nerve distribu-
1.8 (95% CI, 1.1Y2.7) for PONSs lasting longer than 5 days, 0.9 tions. After a popliteal sciatic nerve block for an ankle ligament
(95% CI, 0.5Y1.7) for PONSs lasting longer than 6 months, 0.08 repair, 1 patient sustained a foot drop associated with an EMG

TABLE 3. Postoperative Neurological Symptoms by Block Type

Block Type n (%) PONS PONS-L

Femoral 4377 (34.6) 4 (0.9) [0.2Y2.3] 1 (0.2) [0Y1.2]
Interscalene 2003 (15.8) 7 (3.5) [1.4Y7.3] 5 (2.5) [0.8Y5.8]
Supraclavicular 1508 (11.9) 3 (2.0) [0.4Y5.8] 0 (0) [0Y2.4]
Popliteal (lateral) 977 (7.7) 4 (4.0) [1.1Y10.4] 3 (3.1) [0.6Y8.9]
Femoral (continuous) 952 (7.5) 1 (1.0) [1.1Y10.4] 1 (1.0) [1.1Y10.4]
Interscalene (continuous) 230 (1.8) 3 (12.0) [2.7Y37.6] 2 (8.7) [1.0Y31.1]
Axillary 42 (0.3) 1 (23) [0.6Y125.7] 0 (0) [0Y84]
Other 2579 (20.4) 0 0
Total 12,668 (100) 23 (1.8) [1.1Y2.7] 12 (0.9) [0.5Y1.7]
Values for PONS and PONS-L are expressed as n (n/1000) [95% CI].
Other indicates remaining blocks (n) consisting of ankle (1147), wrist (536), popliteal-prone (376), sciatic-nonpopliteal (147), obturator (101),
paravertebral (62), infraclavicular (60), transversus abdominus plane (53), infraclavicular-continuous (32), popliteal-continuous (17), inter-
sternocleidomastoid (14), cervical plexus (13), supraclavicular-continuous (13), lateral femoral cutaneous (6), ilioinguinal (1), and suprascapular (1);
PONS, any duration postoperative neurologic symptoms; PONS-L, postoperative neurologic symptoms lasting longer than 6 months.

480 * 2012 American Society of Regional Anesthesia and Pain Medicine

Copyright © 2012 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
Regional Anesthesia and Pain Medicine & Volume 37, Number 5, September-October 2012 Morbidity of UGRA

TABLE 4. Characteristics Associated With the Twelve Long-Lasting Neurological Symptoms

# Block Procedure Solution Additives NS Comorbidities EMG

1 Femoral ACL repair 0.5% B 30 mL C, 50 Kg Yes HTN Yes
2 Popliteal-lateral Ankle debridement & ligament repair 0.5% B 30 mL No No None Yes
3 Popliteal-lateral Ankle ligament repair 0.5% B 30 mL No No None Yes
4 Interscalene Shoulder arthroscopy 0.5% B 30 mL C, 50 Kg Yes None Yes
E, 150 Kg
5 Interscalene Shoulder arthroscopy 0.5% R 30 mL No No None Yes
6 Interscalene-continuous Total shoulder arthroplasty 0.5% R 30 mL No No Vitamin B12 deficiency Yes
7 Popliteal lateral Resection of Haglund deformity 0.5% R 30 mL No No HTN No
8 Interscalene Total shoulder arthroplasty 0.5% B 30 mL C, 50 Kg No DM, COPD, HTN No
9 Interscalene-continuous Total shoulder arthroplasty 0.5% B 30 mL No No Fibromyalgia Yes
10 Femoral-continuous Total knee arthroplasty 0.2% R 30 mL No No None Yes
11 Interscalene Shoulder arthroscopy 0.5% B 30 mL C, 50 Kg No HTN Yes
12 Interscalene Total shoulder arthroplasty 0.5% B 30 mL C, 50 Kg Yes MS Yes
E, 150 Kg
Solution: B indicates bupivacaine; R, ropivacaine.
Additive: C indicates clonidine; E, epinephrine.
Comorbidities: COPD indicates chronic obstructive pulmonary disease; DM, diabetes mellitus; HTN, hypertension; MS, multiple sclerosis.
EMG indicates electromyogram; NS, nerve stimulation.

confirming pathology of the common peroneal nerve. This in- of the American Society of Regional Anesthesia (ASRA) with
jury was still present after 18 months. One patient had persistent respect to the injection sequence. That is, the ASRA Joint Com-
decreased temperature sensation in the dorsum of the foot after mittee on UGRA recommends that practitioners cease the injec-
a popliteal sciatic block for ankle debridement and ligament tion and reposition the needle if the local anesthetic is not
repair. After femoral blocks (1 for a total knee arthroplasty and sonographically visualized.8
1 for an ACL repair), 2 patients sustained EMG confirmed With respect to PONS, Brull et al9 recently reviewed 16
femoral neuropathies associated with significant quadriceps cohort and case-control trials conducted between 1995 and 2005
weakness. Table 4 summarizes additional key details of patients to examine the incidence of PONS associated with peripheral
with PONS-L. regional anesthesia. These trials identified 399 events in 22,414
nerve blocks, resulting in an injury rate of 18 per 1000. However,
only 1 reported nerve injury that lasted longer than 12 months
DISCUSSION (0.04/1000). We have identified a long-term PONS rate of 0.9
Our clinical registry provides additional insight into the per 1000 based on a 6-month definition. Our rate represents a
contemporary risks associated with UGRA. The 2 main findings potential 22-fold increase in long-term PONS in comparison
of this analysis are (1) the incidence of LAST was very low, at to the meta-analysis of Brull et al. Obvious difficulties arise when
0.08 per 1000, and (2) the overall incidence of PONS lasting trying to compare rates and events between institutions, such as
greater than 6 months was 0.9 per 1000. durational definitions (ie, 6 versus 12 months), neurologic injury
Our finding of 1 LASTevent in 12,668 is consistent with the definitions, completeness of the postoperative evaluation, and
low incidence reported by both Barrington et al5 and Orebaugh sensitivity of the reporting infrastructure. One concern we have
et al.6 Barrington et al reported 3 major LAST events of 8189 regarding the external validity of the data of Brull et al is that,
blocks, whereas Orebaugh et al recorded 0 in 2000 ultrasound- even if no blocks were performed, we would expect to see a higher
guided blocks. It is interesting to note that most LAST events in risk of nerve injury associated with 44,859 operative procedures.
the data of Barrington et al reflect the performance of axillary From a comprehensive clinical registry at the Mayo Clinic,
brachial plexus blocks. In our cohort of patients, we rarely Jacob et al10 reported 47 unresolved nerve injuries in 12,998
performed an axillary plexus block because of our preference for total hip replacements, an incidence rate of 3.6 per 1000 with
supraclavicular block. a 95% CI of 2.6 to 4.8 per 1000. Thus, if we assumed con-
Our 1 case of LAST was in the context of a continuous servatively that the baseline rate across all surgery was 50% less
femoral block, and this brings to light some of the distinct than the lower CI bounds of the data from Jacob et al, we would
limitations of ultrasound technology. In this patient, the out-of- expect to see approximately 29 long-term injuries in the 22,414
plane needle insertion technique, in which the local anesthetic patients summarized by Brull et al. Thus, we call into question
was injected through the block needle before catheter advance- the data collection strategies and definitions used by the indi-
ment, was used. The local anesthetic was not visualized, although vidual authors cited in the meta-analysis of Brull et al. Variances
the needle tip was apparently seen. The ultrasound beam gen- in the rigor of data collection, completeness of follow-up,
erated a short-axis view of the shaft of the needle in which the assignment of causality, and definitions of nerve injuries are
tip was actually intravascular. This limitation of ultrasound illus- likely responsible for differences in risk reporting, especially in
trates the fact that a 2D ultrasound beam provides little volume the setting of rare occurrences.
information. When imaging the block needle in short-axis, it can We would like to emphasize that our incidence of PONS
be challenging to confirm needle tip location. The events linked represents a perioperative phenomenon in which the exact eti-
to this case of LAST also seem to validate the recommendations ology of the nerve injury is, as usual, unclear. We subscribe to the

* 2012 American Society of Regional Anesthesia and Pain Medicine 481

Copyright © 2012 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.
Sites et al Regional Anesthesia and Pain Medicine & Volume 37, Number 5, September-October 2012

theory of Hebl,11 in that perioperative nerve injury is multifac- ondary to failed follow-up. This failed follow-up could have
torial and likely occurs as a result of multiple results, such as arisen from several scenarios, such as inaccurate anesthesia as-
preexisting lesions, surgical trauma, block related trauma, local sessment, patient lost to follow-up, or poor surgical communi-
anesthetic effects, and position-related trauma. Thus, beyond cation. The end result is that our findings should be considered
methodical differences, our apparent higher rate of PONS an underestimation of the true long-term PONS rate.
compared with historical norms may reflect differences in our In conclusion, in our academic UGRA practice, we report 1
institution’s surgical approaches, nerve block approaches, an- case of LAST in 12,688 patients. The incidence of long-term
esthetic management, underlying patient characteristics, as well PONS was found to be 0.9 per 1000, which is higher than that
as our definition of what constitutes a reportable adverse re- deduced from historical controls and textbook references. Given
gional anesthetic event. From the detailed chart review and the low absolute rate of events, the ability to identify independent
interviews of the 12 patients who sustained long-term neuro- predictors of LAST and PONS will depend on the collaboration
logic symptoms, it is not possible to rule out involvement of of multiple centers. Such collaboration would ideally exist
the peripheral nerve block. With respect to the nerve block through the use of a shared clinical registry in which data ele-
approach, it would be too simplistic to summarize our tech- ments and outcome measures could be standardized. Given the
nique with 1 description. As is true at most institutions, we have popularity of peripheral regional anesthesia, our shared efforts
a varying and evolving practice with multiple providers. The would likely translate into the ability to truly risk adjust and,
common approach, however, is the use of ultrasound guidance perhaps, make a safe practice even safer.
with or without nerve stimulation. When PONS met our working
definition, we deliberately did not attempt to distinguish between REFERENCES
a primary surgical-, positional-, or nerve blockYrelated etiology. 1. Neal JM, Bernards CM, Hadzic A, et al. ASRA Practice Advisory
This approach was influenced by the fact that it is often im- on Neurologic Complications in Regional Anesthesia and Pain
possible to determine the exact etiology or to rule out multiple Medicine. Reg Anesth Pain Med. 2008;31:404Y415.
causes. Further, it is our conviction that from a patient’s per- 2. Di Gregorio G, Neal JM, Rosenquist RW, Weinberg GL. Clinical
spective, this is a largely irrelevant distinction. Our fundamental presentation of local anesthetic systemic toxicity: a review of published
objective should be to provide patients with a conservative es- cases, 1979 to 2009. Reg Anesth Pain Med. 2010;35:181Y187.
timate of the risk of regional anesthesia in the context of a sur-
3. Neal JM. Ultrasound-guided regional anesthesia and patient safety: an
gical intervention. Hence, we are reporting all injuries where the
evidence-based analysis. Reg Anesth Pain Med. 2010;35:S59YS67.
nerve block could have conceivably contributed to the adverse
event versus reporting only those that are certainly attributable to 4. Auroy Y, Benhamou D, Bargues L, et al. Major complications of regional
the block. This inclusive approach should be more helpful in the anesthesia in France: the SOS Regional Anesthesia Hotline Service.
process of shared decision making compared with an approach Anesthesiology. 2002;97:1274Y1280.
that only counts PONSs that are, with absolute certainty, due to 5. Barrington MJ, Watts SA, Gledhill SR, et al. Preliminary results of the
the nerve block. Australasian Regional Anaesthesia Collaboration: a prospective audit
Our results also corroborate the relative safety of the of more than 7000 peripheral nerve and plexus blocks for neurologic
ultrasound-guided supraclavicular nerve block in pneumotho- and other complications. Reg Anesth Pain Med. 2009;34:534Y541.
rax risk, with 0 cases in 1508 blocks. Conventional landmark 6. Orebaugh SL, Williams BA, Vallejo M, Kentor ML. Adverse outcomes
techniques have been cited with a risk rate as high as 6%, making associated with stimulator-based peripheral nerve blocks with versus
this block unpopular in the preultrasound era.12 From a retro- without ultrasound visualization. Reg Anesth Pain Med.
spective view, Perlas et al13 reported 0 cases of pneumothorax in 2009;34:251Y255.
510 patients. The ability to easily image the first rib and pleura 7. Koff MD, Cohen JA, McIntyre JJ, Carr CF, Sites BD. Severe brachial
likely contributes to the safety of this block. It should be noted plexopathy after an ultrasound-guided single-injection nerve block for
that our results of 0 cases in 1508 is still consistent with a total shoulder arthroplasty in a patient with multiple sclerosis.
pneumothorax rate as high 2.4 per 1000, which represents the Anesthesiology. 2008;108:325Y328.
upper bound of the 95% CI. 8. Sites BD, Chan VW, Neal JM, et al. The American Society of Regional
Our data represent a retrospective cohort study that is Anesthesia and Pain Medicine and the European Society of Regional
subject to the distinct limitations related to the lack of ran- Anaesthesia and Pain Therapy joint committee recommendations for
domization and possible confounding. The P value of 0.002 education and training in ultrasound-guided regional anesthesia.
associated with an increased PONS risk from an interscalene Reg Anesth Pain Med. 2010;35:S74YS80.
block should be viewed with caution because of the multiple 9. Brull R, McCartney CJ, Chan VW, El-Beheiry H. Neurological
blocks that were examined and the lack of adjustment for mul- complications after regional anesthesia: contemporary estimates of risk.
tiple comparisons. In addition, we suspect that the apparent re- Anesth Analg. 2007;104:965Y974.
lationship between interscalene blocks and PONS is confounded
10. Jacob AK, Mantilla CB, Suiggum HP, Schroeder DR, Pagnano MW,
by multiple variables, such as the surgical procedure. Given the
Hebl JR. Perioperative nerve injury after total hip arthroplasty:
small number of events, multivariate adjustment was not pos-
regional anesthesia risk during a 20-year cohort study.
sible. Our technical performance of UGRA likely changed
Anesthesiology. 2011;115:1172Y1178.
during the 8-year study period, as did the anesthesia and sur-
gical providers. How these institutional changes affected 11. Hebl JR. Ultrasound-guided regional anesthesia and the prevention of
morbidity and quality is unmeasured. Furthermore, we did not neurologic injury: fact or fiction? Anesthesiology. 2008;108:186Y188.
track several variables that could have affected our morbidity 12. Brown DL, Cahill DR, Bridenbaugh LD. Supraclavicular nerve block:
rates, such as evolving ultrasound technologies, nuances of anatomic analysis of a method to prevent pneumothorax.
supervision styles, and specifics of novice behaviors. Finally, Anesth Analg. 1993;76:530Y534.
our incidence rate of long-term PONS should be viewed as a 13. Perlas A, Lobo G, Lo N, Brull R, Chan VW, Karkhanis R.
best-case scenario. That is, our process of identifying patients Ultrasound-guided supraclavicular block: outcome
who sustained neurologic injury likely missed some cases sec- of 510 consecutive cases. Reg Anesth Pain Med. 2009;34:171Y176.

482 * 2012 American Society of Regional Anesthesia and Pain Medicine

Copyright © 2012 American Society of Regional Anesthesia and Pain Medicine. Unauthorized reproduction of this article is prohibited.