A Website of Neuraxiom LLC

Neuraxiom LLC

A Washington State Company


Contact Me

The Neuraxiom Privacy Policy


The Curious Events Surrounding

A Case of Acute Respiratory Failure following Interscalene Block

November 1, 2008

I  write about this case because the scenario that precedes it is quite common and most of the time events never align to arrive at the rather dramatic point that was reached in this case.  But the potential exists in every interscalene brachial plexus block done on a patient with COPD.  I suspect you can go along through an entire career without encountering this, or see it tomorrow, but the opportunity is not rare.

The patient came to the operating room for acromioplasty and rotator cuff repair of the left shoulder.  He was in his 50’s, his height and weight were average and had no known allergies.  He had a 40 year smoking history of 1-2 packs per day and was recognized as having COPD but had not been on home oxygen therapy and only recently been started on inhaled steroids and bronchodilators.  Preoperative chest x-ray showed hyperinflation.    The assigned anesthesiologist reviewed his charts, interviewed him and formed a plan consisting of a general anesthetic supplemented for postoperative analgesia with an interscalene brachial plexus nerve block.

The Block

The patient agreed and while in the pre-operative unit, monitors were applied and the patient was sedated with 2 mg of the midazolam and 50 mcg of fentanyl. The left neck was prepped with chlorhexidine and sterile gel was applied.  A shallow (8-10 mHz) ultrasound probe was placed against the neck and the brachial plexus was identified, first in the supraclavicular space, then followed cephalad into the interscalene space.

A lidocaine skin wheal was place and 22 gauge short bevel needle was inserted into the interscalene space in an out-of-plane approach and the block was performed using a total of 30 mls bupivacaine 0.375% with epinephrine 1:400,000.  During the block there were repeated aspirations of the syringe to ensure that no vascular structures were entered and the needle was repositioned several times in order to optimally place the local anesthetic boluses.  Sedative effect was light and the patient was speaking and answering questions throughout the procedure.  He reported nothing unexpected and only a light amount of discomfort.  The block was uneventful,  The needle was removed and effects of the block were seen immediately in the inability of the patient to use his shoulder muscles to raise his left arm. The patient was counseled for what to expect from the block over the next 24 hours and was observed for approximately 5 to 6 minutes if there were any acute effects from the injected local anesthetic.

An Unexpected Result

Approximately 10 minutes later patient was wheeled to the operating room and the gurney was placed next to the operating table and the patient was in the process of moving across to the table when he suddenly sat up, clutched his chest, and said "I can't breath".  He then collapsed back onto the bed pulling out his intravenous catheter in the process. 

The anesthesiologist immediately began ventilating with a bag and mask while the circulating nurse pushed the "Code Blue" button for help. Personnel assembled, the code cart arrived. a second anesthesiologist restarted the IV.  As soon as the IV was running, intralipid solution 20% was begun and eventually ran in to the amount of 150 mL (even though the EKG showed a regular rhythm, normal intervals and a rated of around 105 bpm).  Simultaneously a muscle relaxant was given and the patient's trachea was intubated and ventilated with 100% oxygen.  Non-invasive blood pressure revealed hypotension with the lowest pressure seen was 65 systolic prior to the IV restart and intubation. 

The patient had become markedly cyanotic after his initial collapse with the lowest pulse oximetry showing saturations in the 50's.  His saturations rose immediately with intubation and ventilations.  Auscultations revealed no wheezing or rales.  The scheduled surgery was abandoned and the patient was transferred to the postoperative care unit and placed on a ventilator assist-control . An arterial line was started in left radial artery and arterial blood gases done. The differential diagnosis of the time of the patient's collapse was pulmonary embolism, myocardial infarction, pneumothorax, and hemidiaphragmatic paresis. 

A chest x-ray was taken in PACU to rule out pneumothorax, and a 12 lead EKG was done to rule out MI, both were negative for an explanation for the collapse. Troponins, and CPK were negative for indication of MI. The arterial blood gases after his resuscitation revealed a pH of 7.30, a pCO2 of 65, pO2 of 443, a base excess of -2.9 and a bicarbonate of  31 on an FiO2 of 0.8.

One possible explanation was the patient having had a pre-existing, unrecognized paresis or paralysis of the contralateral hemidiaphragm.  If the interscalene block paralyzed the hemidiaphragm on the operative side, acute hypoventilation would ensue.  After arrival in the postoperative care unit, when the patient was more aware and able to follow instructions, an ultrasound survey of the diaphragm was performed using a low frequency (2-4 mHz) curved array probe in M-Mode ultrasound.  We found that the contralateral diaphragm was not paralyzed and the hemidiaphragm on the operative side was paralyzed as expected presumably from the block.  Images from the ultrasound  surveys of the hemidiaphragms are shown below.



The patient was placed in the intensive care unit overnight while the block dissipated and he was extubated without incident the next morning. The sensory effects of the interscalene brachial plexus block lasted until approximately 3 am the next day, for a block of 20 hours in duration.    The patient was discharged soon after extubation with no other problems noted.

So How did We Get Here?

A probable explanation for this scenario, considering the results of the tests, is that the patient had fairly significant COPD and was a carbon dioxide retainer prior to surgery.  Prior to the block the patient was sedated with IV midazolam and fentanyl which probably caused more hypoventilation in turn increasing the patient’s carbon dioxide and decreasing the patient’s already blunted sensitivity to the rising CO2 level.   Finally as a side effect of the interscalene block, the local solution (bupivacaine) spread around the front of the anterior scalene muscle and came into contact with the phrenic  nerve, creating a conduction block and a paresis or paralysis of the hemidiaphragm on the side of the block.  This lack of movement in the hemidiaphragm further reduces the tidal volume and causes another increase in the amount of CO2 remaining in the venous blood.   Eventually the spiralling levels of carbon dioxide and the accompanying acidosis reaches a level that causes widespread vasodilatation and syncope.

The pulling out of the intravenous line was just a happy coincidence.

What can we do about this?

It has long been recognized that an increased risk exists for respiratory complications in patients with pre-existing pulmonary disease following interscalene block.  This procedure is commonly performed because the risk benefit balance is in favor of the block over the use of narcotics in the patient with obstructive pulmonary disease.  So what are our options in terms of the care of the patient with obstructive pulmonary disease undergoing shoulder arms surgery needing an interscalene block?  It doesn't seem prudent to deny patients with COPD the benefits of regional block but there must be a way to minimize or manage the inherent risk in the block. 

A provisional plan may be to perform incentive spirometry prior to interscalene block in patients whose history indicates they are at risk for problems related to loss of function of a hemidiaphrgam.  After the block the patient can use the spirometer again to compare their ability to move air.  This may give the clinician a "heads-up" in time to provide enough ventilatory support to avoid a more dramatic course. 

In addition, it is probably a good idea to modify the usual interscalene technique in favor of a low-volume technique (as discussed on the Neuraxiom pages "Low Volume Interscalene" and the "PSIB Project").  Some time in the past it became dogma that hemidiaphragmatic paralysis inevitably follows interscalene block just as gravy follows pork-chops, but early last year the "Phrenic Sparing Interscalene Block" (PSIB) project conducted here on the Neuraxiom site showed that a simple low volume technique could avoid this complication in the majority of cases.  This axiomatic finding was once again "discovered" in a paper in a major professional journal .  (During the work on this project a friend, Dr. Andre Ceccoli in France carried out a wonderfully simple series of surveys mapping the variation in position of the phrenic nerve in relation to the brachial plexus usng ultrasond and a nerve stimulator.  The results of this series are quite worth looking at and can be found at

In the low volume modification we use between 5 and 10 mls and control the injected local so that it stays within the brachial plexus sheath as much as is possible.  By using a lower volume and avoiding overflow of the local solution across the anterior side of the anterior scalene we can, in the majority of cases, avoid local anesthetic contact with the phrenic nerve, and thereby avoid the hemidiaphragmatic paralysis.  There will be cases where contact is unavoidable because in a small percentage of anatomical variations the phrenic actually travels with the brachial plexus between the scalenes, but according to our small sampling in the PSIB project the hemidiaphragm will remain functioning in about 75% of blocks done with 10 mls or less (without taking the time to map the position of the phrenic using a nerve stimulator).

Why not use the low volume technique all of the time then?  Because we sacrifice duration the the low volume.  The low volume technique yields a solid block with good coverage that lasts between 8 and 10 hours usually.  It's possible that this block may be lengthened a bit by using some local cooling at the site of the block but probably not more than a couple hours at best. 

Still these steps may be enough to avoid or mitigate the loss of additional respiratory function caused by sedation and the occasional loss of a hemidiaphragm.


To some extent this particular event probably unfolded as it did as a fault of just bad timing, and to that degree the clinical situation is probably much more common than is recognized.  If the patient had reached the operating room a few minutes earlier he may have been asleep and mechanically ventilated. If this had happened before the patient reached the point hypoventilation that pushed his carbon dioxide level to the critical level, the respiratory acidosis would never have occurred.  At the end of the case, the weakened respiratory effort would have been recognized and the patient would have remained on a ventilator until he recovered enough to be extubated (a much more common scenario).



The Strange Case

of the

Delayed Onset of Motor Block in the Hand following Interscalene Block

February 22, 2008

Jack Vander Beek

I’d like to present an interesting case which gives hints to the actions of local anesthetics in regional blocks.  It involves a young man about 19 years old who came to surgery for repair of a rotator cuff tear and who received an interscalene block with bupivacaine.

This young man was very calm and even enthusiastic about the block and the interscalene block procedure under ultrasound guidance went well.  Bupivacaine 0.375% with epinephrine, 25 mls, was used.  The local anesthetic volume was divided as injected;  between the brachial plexus sheath and the anterior scalene, within the sheath, and between the sheath and the middle scalene.   The onset of the block was fairly rapid, and within a few minutes he could not raise his arm.  He could however still move his fingers and even produce a forceful grip although he reported some numbness in his thumb and forefinger.  He was reassured that it was common to be able to still be able to move his hand and fingers.

A few minutes later the patient was rolled back into the operating room where he was given a general anesthetic and the surgical procedure was performed without incident and took a little over an hour.  He required very little general anesthetic during the procedure and after transfer to recovery room he awakened very quickly and reported no pain at all in his surgical shoulder and arm and he could still move his hand and fingers. He spent approximately an hour in the recovery room and was transferred back to the outpatient unit to complete preparations for his discharge to home. 

After approximately 2 more hours he noted he could no longer move his hand and fingers and had no grip.   His concerned mother notified the nurse who, in turn, called the anesthesiologist who was still working on scheduled cases in his room.  The anesthesiologist reassured the unit nurse that this happened sometimes and that it would “wake up” when the block wore off.  The patient was not upset or concerned by this extension of the block, he was in fact happy that his shoulder did not hurt and was anxious to be discharged.

The patient was discharged and the block lasted through the night and completely dissipated the following day with no sequela. 

Okay, big deal, right?  Maybe not big deal, but interesting deal.  Most conventional wisdom holds that a block, regional or other, caused by local anesthetic is “set” after a given amount of time.  For bupivacaine, 40-50 minutes is usually given as the time in which the molecules have found their musical chairs and are comfortably seated for the duration.  AND YET, here is a case where the block progressed for at least 4 hours after placement.

Like a Sponge

The direction of onset is not surprising.  Proximal to distal progression makes sense since nerves that will distribute to more proximal end organs will exit the brachial plexus nerve bundle earlier and therefore need to be arranged more superficially to make exit easier.  Since they are more superficial the local anesthetic solution will come in contact with those early departure axons more quickly because of their exposure.  Conversely the nerves in the bundle that will be the last to depart the bundle will be those located most centrally in the nerve bundle.  This is the most efficient evolutionary arrangement available. 

When looking at nerve groups (bundles and fascicles) exiting the cord for distribution down an extremity, we would expect to see the following characteristics.

  • The closer to the spinal cord, the larger the overall bundle.
  • The closer to the cord more axon and glial tissue and the less connective tissue (accounts for more hollow appearance on ultrasound)
  • As distance from the cord increases, more axon is leaving the bundle to connect to end organs, less nervous tissue (axon & glial) is present, more organizing and supporting connective tissue is present.
  • As distance increases overall diameters will begin to shrink.

Therefore; the closer to the cord the farther it is, in terms of diffusion, from the outside of the nerve group to the inside.  So if you are surrounding a bundle, for instance the trunks of the brachial plexus, with local anesthetic, the diffusion of the local from the outside to the inside will result in a stabilization of the axonal membranes in the most proximal distributions to the most distal, in that order.  The most distal (most internal) axons will only become blocked when and if the anesthetic molecules reach them.  In my opinion the efficiency of the diffusion of the local through the substance of the radius of the bundle will depend upon;

  • The beginning concentration of the local anesthetic solution being used. (density of the molecules available, therefore the gradient of the solute at the leading edge of the solution injected)
  • Total volume and distribution of volume of the local anesthetic injected for the block.  (higher volumes can delay redistribution away from the site by vascular absorption and normal circulation of the interstitial fluid)
  • The pH of the nervous tissue and interstices of the area being injected for the block (slight variations in the area toward the more acidic will cause more the local anesthetic molecules to become ionized and thereby unavailable for crossing the lipid membrane and participating in the block.  This is usually seen clinically as a delayed onset of a block.  Tissue acidosis can be caused by conditions such as trauma, nearby infection, or peripheral micro-vascular disease.)
  • The physical size (radius) of the target nerve. 
  • The distribution of the local solution around the circumference of the nerve bundle.  Local solution which is placed only on one side of the circumference of a nerve will take longer to completely diffuse through a nerve than solution placed in more than one place or even complete around the nerve.  (How much this really effects block efficiency in real life is probably negligible.)
  • The number and character of barriers that the local molecules must cross on their passage to the center of the nerve.  As discussed above, while more distal nerve bundles are smaller in diameter they will contain more connective tissue packaging that resists diffusion.
  • Block efficiency is greatly influenced by the proximity of the local solution injection to the target nerve bundle but this only sets up the conditions described above for the passive phase of diffusion into the nerve.  

In an empty sink basin, put a drop of black ink on a wet sponge, the ink molecules will gradually diffuse to the center of the sponge.  Whether the ink color will be noticeable at the center will depend on how big the drop, how dark the ink is, and how thick the sponge is.  If you place the sponge in a slow moving stream of water and apply the ink you have a situation closer to the regional nerve block.

The brachial plexus model provides the ideal example for tailoring the placement of the local to the site of the desired effect.

It is really is true that wrist & hand pain is best addressed from a brachial plexus block at  the axillary approach (or even below), mid-arm & elbow pain to mid-upper arm is best handled at the supraclavicular level of the brachial plexus, and pain of the shoulder & clavicle to the sternal border can be covered from the interscalene approach.

Sure, a lot of times the hand is numb after an interscalene, but sometimes it’s not and it’s always numb after the axillary level block.  So why tempt fate if you don’t have to.  Using these 3 approaches to the brachial plexus any of the situations can be handled with confidence and when the unusual need arises, such as a cardiac patient with no neck, and bad bad COPD comes to you for the total shoulder arthroplasty, you will know that a high supraclavicular approach with less than 20 mls stands a good chance of covering the pain while missing the phrenic (possibly) and that an axillary approach is probably not going to get you there.


So, Maybe it’s the Placebo Effect! (WoooOOOoooOOO ← scary other-worldly sound)

So applying the scenario described above to the problem child of blocks; the Sciatic. 

The sciatic is a large caliber nerve and yet many times we make the mistake of expecting it to block like a small nerve bundle.  I don’t know about your situation but at our house there is an ongoing “discussion”, especially with a few orthopedic surgeons as to whether a sciatic block is necessary or desirable for pain control after total knee arthroplasties.  The logic goes that most people don’t have posterior pain after the operation and you only need to observe patients in recovery to understand this.  Okay, well I guess some patients do well with a femoral only but a lot require quite a bit of narcotic for something after femoral-only block. 

The femoral-only viewpoint was bolstered by observations that patients with failed sciatic blocks (quite common at our operating room in the earlier days) were completely comfortable.   The qualification for inclusion to the failed-sciatic club was that the patient could move and feel their feet.  We were expecting every sub-gluteal sciatic to diffuse to the center of this large nerve and thereby block the foot.  The group lament went like this; “Well, it doesn’t look like your sciatic block worked, but luckily, the patient didn’t need it anyway.  Look how comfortable he is!”

Using the model of the above 19 year old guy’s brachial plexus, all of these comfortable post-op total knee patients who could move their feet, actually had very good sciatic blocks even though they could move their feet.  It is, in fact, remarkable that so many people with sciatic blocks had blocks that progressed all of the way through to their feet.  It is certainly not necessary to block every fiber in the sciatic in order to have a comfortable post-op total knee patient and in fact a numb foot is not appreciated by physical therapy the following morning when they are trying to ambulate the patient.


The Conclusions

The block progression isn’t over ‘til it’s over.  While some, maybe most, blocks are “set” after 30-40 minutes, it’s a continuum— some require less time, some require more. 

The block may be therapeutically effective without being complete.   The aim should be comfort.  Remember that the majority of the time we want a pain block, we’ll take full sensory, but we need not demand motor.


free counters