Let’s look at the following example. Say the right upper extremity electrode (white, by the American Heart Association standards) is placed on the right shoulder, and the left lower extremity electrode (red) is placed on the left chest around the third intercostal space (I commonly see this sort of placement). Even if the monitor is displaying lead II, because of the electrode placement, what you will see will more closely reflect lead I.

The same applies to the precordial leads. If the electrode corresponding to the monitor’s V5 is placed on the sternum, you won’t be seeing lead V5. This sounds obvious, yet I see wildly misplaced leads every week.

Even misplaced leads will usually allow for accurate diagnosis of arrhythmias, but the sensitivity for detecting myocardial ischemia may be compromised. For example, loss of R wave progression in the precordial leads can indicate an acute myocardial infarction. However, if those leads are placed incorrectly, you may never get an accurate picture of the R waves and subsequent changes.

ECG filtering

Where the right lower extremity lead (green) goes also matters because this lead is used to help filter out electrical noise. If it is placed close to the right upper extremity lead, or too close to the site of electrocautery, the signal processing can be degraded.1

ECG filtering is a fascinating and complex subject. The monitors we routinely use in the operating room and for telemetry apply different filters than those used when recording a 12-lead ECG. This is the reason that one may see what appears to be ST-segment elevation on the monitor only to have it be absent on the 12-lead ECG. For more reading on ECG filtering, see here, here, and here.

Aside from breast and thoracic surgery, it is usually possible to have optimal lead placement. If the surgeon is operating on the abdomen, place the lower limb leads on the hips. For breast surgery, one can place the leads more distally on the arms even if an axillary dissection is being performed. To the surgeon, if you’re not cutting through the ribs, do you really need me to move the precordial leads?

Modern breast cancer surgery with sentinel lymph node biopsy carries a lower risk of lymphedema of around 3% for a partial mastectomy and up to 8% with a sentinel lymph node biopsy. The incidence of lymphedema increases further if more axillary nodes are excised.2

Share

What should you do?

Ask your patients if they have a history of lymphedema. If they have no history, then it is fine to place IVs and measure blood pressure as you normally would.

Even if a patient has a history of lymphedema in one arm, you can still place IVs in that arm. The majority of the evidence has shown no association between IV placement in the ipsilateral arm and lymphedema. For example, one retrospective study comparing 5,153 IV placements in the ipsilateral arm to 2,743 IV placements in the contralateral arm found no significant difference in the rate of complications.

You can also take single blood pressure measurements on the same side as the lymphedema.

Whether you want to take repeated blood pressure readings on the side of lymphedema, such as would be done under anesthesia, may be a more nuanced topic. While single blood pressure measurements have not been shown to cause or worsen lymphedema, studies have not examined the effects of continuous blood pressure measurements on the incidence of lymphedema. Given that lymph nodes are usually removed proximal to the blood pressure cuff, and given that pressure is often used as a treatment for lymphedema, it seems reasonable to think that cyclic blood pressure measurements would not worsen lymphedema. However, here I cannot offer any evidence-based guidance.3

Let’s put this outdated teaching to rest and assure patients that it is safe to have IVs placed and blood pressure measurements taken on the same side as their prior lymph node excisions even if they have a history of lymphedema.

The American Society of Anesthesiologists (ASA) practice guideline for GLP-1 RAs published in 2024 lists risk factors that may indicate that patients taking these medications have residual gastric contents present.1 The problem with these risk factors is that many patients have residual gastric contents without having any of the listed risk factors present. As we will see, details about the dosing of semaglutide, in particular, offer no insight into the likelihood of having increased residual gastric contents.

To evaluate the incidence of residual gastric contents in patients taking semaglutide, a group of researchers performed a prospective, multicenter, matched study using gastric ultrasound to examine patients prior to surgery. Patients were included if they were taking semaglutide for diabetes or weight management and were undergoing general anesthesia for elective surgery, and had held one dose of the medication. They were matched based on age, diabetes status, and BMI to other patients undergoing elective surgery who were not taking any GLP-1 RAs. Patients with conditions known to delay gastric emptying (e.g., Parkinson disease, hypothyroidism, etc.) were excluded. A total of 88 patients were enrolled (44 in each group) and underwent gastric ultrasound in the preoperative setting.

A full stomach, defined as having ≥ 1.5 mL/kg of gastric volume, was observed in 49% (n = 21) of the patients taking semaglutide and in 18% (n = 8) of the controls (OR 4.29; 95% CI 1.63–11.29, p = 0.003). Solid gastric contents were observed in 85% (18/21) of the semaglutide patients with full stomachs. Notably, these results are in the setting of holding one dose of semaglutide and with a median fasting duration from solids of 14.4 hours in the semaglutide group and 15.1 hours in the controls—far exceeding fasting guidelines.2 This study further supports the notion that holding one dose of GLP-1 RAs does nothing to reduce the risk of residual gastric contents. It also supports the need for longer fasting durations in these patients.

The authors found no significant association between the route of administration (oral vs. subcutaneous), duration of use, dose, or time since the last dose of medication. Even patients who had been taking semaglutide for years still had increased residual gastric contents. I hope this makes clear that the risk factors listed in the ASA practice guideline cannot be used to exclude the risk of having increased residual gastric contents.

One should assume that many, if not most, patients taking GLP-1 RAs have full stomachs. Imaging modalities, with gastric ultrasound being the only feasible option, represent the only way to risk-stratify these patients and exclude the presence of increased residual gastric contents.

For a more in-depth discussion of GLP-1 RAs and their management, see my previous post on the topic.

• ASA 1 is a healthy person.

• ASA 2 has mild systemic disease.

• ASA 3 has severe systemic disease.

• ASA 4 has disease that is a constant threat to life.

• ASA 5 is someone who is not expected to survive without the proposed surgery.

• ASA 6 is a brain-dead patient whose organs are being donated.

In January 2026, the ASA updated their table of examples for the ASA physical status classification system. One change that caught my eye was the classification of obstructive sleep apnea (OSA). Mild and moderate OSA count as an ASA 2 if a patient is using continuous positive airway pressure (CPAP), but count as an ASA 3 if the patient is not using CPAP. Severe OSA falls under ASA 3 regardless of CPAP use. Here are the other new examples for adults:

• ASA 2: CHF NYHA class 1, mild cognitive dysfunction

• ASA 3: CHF NYHA classes 2 and 3, significant cognitive dysfunction,1 poorly controlled diabetes mellitus or hypertension with or without end-organ dysfunction

• ASA 4: CHF NYHA class 4, uncompensated cirrhosis, severe cognitive dysfunction

For pediatrics, having a risk of malignant hyperthermia (e.g. family history) is now an ASA 2.

For pregnant patients, preeclampsia without severe features is now classified as an ASA 3 instead of an ASA 2. For ASA 5, amniotic fluid embolism was added to the examples.

I wish they had added concrete values to descriptions such as “poorly controlled diabetes mellitus” and “uncompensated cirrhosis.” What hemoglobin A1c or MELD score constitutes those descriptions? And where does an overweight but otherwise healthy adult fit? A pediatric patient with an “abnormal BMI” would be at least an ASA 2, but why do the adult examples fail to comment on the BMI range of 25–29.9?

Why ASA 3 lacks clarity

People who fall into categories 1, 2, 4, and 5 are pretty clear. Where I find the system fails to give meaningful stratification between patients is in the ASA 3 category. Take, for example, the following two patients.

Person one is a male with mild OSA who is not using CPAP. Based on the new examples, this patient is an ASA 3 due to noncompliance with CPAP. If he used CPAP, then he would be an ASA 2.

Person two is a female with hypertension, coronary artery disease (CAD) with two stents (placed more than 3 months ago), diabetes with a hemoglobin A1c of 9.8%, daily acid reflux symptoms, an occasional smoking habit, stage 3 chronic kidney disease, hepatic steatosis, and a BMI of 45. This person is also an ASA 3, yet is far sicker than person one.

Having such a broad range of patients in the ASA 3 category decreases its usefulness in understanding patient comorbidities. For this reason, I don’t put a lot of stock into the ASA status unless it is a 1, 4, or 5. But what if there could be more granularity by breaking down an ASA 3 into ASA 3.0 and ASA 3.5?

I would propose that if a patient has certain diseases, such as CAD, a prior stroke, or chronic obstructive pulmonary disease (COPD), they are automatically an ASA 3.5. Additionally, if a patient has five or more severe diseases (I have no evidence-based reason for choosing the number five), that also bumps them up to an ASA 3.5. In this way, a clearer separation between patients could make the ASA 3 classification more useful.

What do you think about this idea? How much attention do you pay to the ASA score?

Share

Cefazolin is a preferred antibiotic for surgical site infection (SSI) prophylaxis. In many patients with penicillin allergies, anesthesiologists and surgeons alike use alternatives such as clindamycin and vancomycin. Some studies have shown a higher incidence of SSI when using alternatives to cefazolin. However evidence on this is mixed and may be seen only in certain subspecialties. A major downside to using clindamycin is that it increases the risk of Clostridioides difficile infection when compared with cephalosporins. For these reasons, cefazolin is more desirable.

The good news is that there is no need to avoid using cefazolin, even in patients with a known history of anaphylaxis to penicillin. The 2022 drug allergy practice parameters update from the American Academy of Allergy, Asthma, and Immunology states: “Patients with a history of urticaria to a penicillin can receive any cephalosporin routinely without prior testing. In contrast, for those rare patients with a history of anaphylaxis to penicillin, a non–cross-reactive cephalosporin (e.g., cefazolin) can be administered routinely without prior testing.” A recent study in joint replacement surgery showed no increased incidence of anaphylaxis when using cefazolin in patients with a known history of anaphylaxis to penicillin.

Spread the word to your colleagues that it is safe to use cefazolin in patients with a penicillin allergy.

To argue that one does or does not need to treat pain while under general anesthesia may be an argument over the definition of pain. If pain is subjective, maybe we should stop referring to objective measurements of heart rate and blood pressure as pain. Instead, let’s refer to them as they are—tachycardia and hypertension—and manage them with cardiovascular-specific agents.1 As you will see below, using a non-analgesic medication instead of an opioid can actually lead to less postoperative pain.

I think the important question to ask is this: does using opioids, non-opioid analgesics, or non-analgesic medications (such as sympatholytics) offer the best postoperative pain control when used to treat elevated sympathetic responses? The opioid-sparing effects of esmolol can help us shed light on this topic.

Esmolol’s opioid-sparing effects

Higher intraoperative opioid doses lead to increased postoperative pain and postoperative opioid consumption. For this reason, minimizing intraoperative opioid use can lead to better patient outcomes.2

Multiple trials and meta-analyses have demonstrated esmolol’s opioid-sparing effects. A meta-analysis from 2018 showed that intraoperative esmolol infusions decreased intraoperative opioid use (standardized mean difference [SMD] -1.60; 95% confidence interval [CI] -2.25 to -0.96; p < 0.001) and reduced postoperative opioid use (SMD -1.21; 95% CI -1.66 to -0.77; p < 0.001). A SMD of ≥ 0.8 is considered a large effect. A more recent meta-analysis from 2025 showed that intraoperative esmolol infusions reduced intraoperative opioid use by 32% (mean difference of -12.89 IV morphine milligram equivalents [MME]; 95% CI -24.74 to -1.05; p < 0.001) and reduced postoperative opioid use by 38.6% (mean difference -3.03 IV MME; 95% CI -4.29 to -1.76; p < 0.001).

Does esmolol have analgesic properties?

The best data we have in humans suggest esmolol does not have analgesic properties. Researchers set out to test this using the cold pressor test in a randomized, placebo-controlled crossover study. Fourteen healthy, conscious volunteers each underwent three separate trials consisting of either a normal saline infusion (the control), an esmolol infusion (10 mcg/kg/min following a 0.7 mg/kg bolus), or a remifentanil infusion (0.2 mcg/kg/min following a normal saline bolus). The primary outcome was maximum perceived pain as measured by the numeric pain rating scale (NRS-max).

The mean NRS-max was similar in the esmolol and placebo groups (8.5, SD 1.4 vs. 8.4, SD 1.3; p = 0.83). When compared to placebo, the remifentanil infusion had a lower mean NRS-max of 5.4 (SD 2.1; p < 0.001). These data suggest that esmolol does not have analgesic properties. Instead, the opioid-sparing effects of esmolol are thought to be due to decreased total intraoperative opioid use, leading to less opioid-induced hyperalgesia and tolerance.

Do we need to treat pain under general anesthesia?

If patients experience better outcomes when managing hemodynamic extremes, commonly referred to as pain, without analgesics (or while minimizing analgesics), perhaps we should be more specific with our terminology. Instead of referring to pain during general anesthesia, we could talk about sympathetic responses such as tachycardia and hypertension. These perturbations can be managed effectively with multiple drug classes. Maybe this simple terminology change will push more people to use esmolol and other sympatholytic agents to minimize opioid use, leading to lower postoperative opioid requirements.

Share

In June 2023, the American Society of Anesthesiologists (ASA) released a consensus-based guideline on the management of patients taking GLP-1 receptor agonists. The concern with these medications is that the delayed gastric emptying leads to residual gastric contents, which may increase the risk of aspiration. This original guideline recommends holding one dose of the GLP-1 RA prior to surgery, but does not recommend changing fasting times.

In anesthesia, one of the first risks to manage in patients is that of aspiration. Fasting guidelines (also called NPO guidelines) are strictly adhered to for this very reason. The NPO guidelines have always had the caveat that patients with delayed gastric emptying may need longer fasting times. The guidelines state, “Anesthesiologists should recognize that these conditions can increase the likelihood of regurgitation and pulmonary aspiration and should modify these guidelines based upon their clinical judgment.” A previous version states, “The guidelines may not apply to, or may need to be modified for, patients with coexisting diseases or conditions that can affect gastric emptying or fluid volume (e.g., pregnancy, obesity, diabetes, hiatal hernia, gastroesophageal reflux disease, ileus or bowel obstruction, emergency care, or enteral tube feeding) and patients in whom airway management might be difficult.”

Requiring patients to abstain from food for longer than 8 hours, even if the optimal duration was unknown, would have been the most cautious approach, even if it was unenjoyable for patients. Instead of taking this approach, the ASA stated, “There is no evidence to suggest the optimal duration of fasting for patients on GLP-1 agonists. Therefore, until we have adequate evidence, we suggest following the current ASA fasting guidelines.” The irony of this is that in the very same guidelines they recommended holding one dose of GLP-1 RAs prior to surgery despite having no evidence that this would minimize residual gastric contents.

What is the harm in holding a dose of GLP-1 receptor agonists?

Every GLP-1 receptor agonist causes delayed gastric emptying. When patients start on these medications, they often experience nausea and vomiting. For this reason, patients are started on the lowest dose and slowly dose-escalated no faster than one increment every four weeks as patients begin to tolerate the medications and the side effects diminish. With each escalating dose, the side effects often temporarily return. When two or more doses of the medication are skipped, a person can lose tolerance, increasing the risk of severe side effects once the GLP-1 receptor agonist is resumed. In this situation, it is often recommended to revert to a lower dose and re-escalate after four weeks. This can pose problems such as needing a new prescription, cost to the patient, worsened glucose control, and weight regain. The logistical problems could lead to even more time off the medications.

In patients receiving GLP-1 RAs for diabetes, the most obvious harm from holding them is poorly controlled blood glucose. Given the risk of complications from uncontrolled diabetes, such as the increased risk of wound infections from hyperglycemia, it would be in a patient’s best interest to continue these medications in the perioperative setting.

Holding GLP-1 RAs for five half-lives can lead to normalization of gastric emptying. For semaglutide dosed once weekly, this amounts to being off the medication for five weeks. Again, the downsides of poorly controlled glucose and weight regain make this a poor approach to management since both hyperglycemia and increasing body mass index are associated with worse perioperative outcomes.

The ASA’s initial recommendation to hold one dose of GLP-1 RAs is all downside with no benefit.

The ASA revised its practice guideline in December 2024, issuing a multi-society guideline. One important change is that they state, “GLP-1 RA therapy may be continued preoperatively in patients without elevated risk of delayed gastric emptying.” The risk factors they state include: escalation phase of dosing, higher dose, weekly dosing (as opposed to daily dosing), gastrointestinal symptoms (nausea, vomiting, abdominal pain, dyspepsia, constipation), and medical conditions that may cause delayed gastric emptying. Who taking a GLP-1 RA is not at increased risk of delayed gastric emptying when that is the very mechanism by which the drugs work? It is important to note that patients who fail to meet all of these risk factors (i.e., patients on stable doses without gastrointestinal symptoms) can still have delayed gastric emptying.

This part of the guideline is silly because there is no way to risk stratify these patients using the stated risk factors. A recent study found no correlation between short- and long-acting of GLP-1 RA, dose, and duration of use with a percentage of patients in all categories experiencing residual gastric contents. Instances of delayed gastric emptying can occur even after patients have been taking GLP-1 RAs for more than three years. The idea that one can accurately identify delayed gastric emptying without performing imaging (such as bedside gastric ultrasound) is not rooted in reality.

The second part of the guideline is more useful, suggesting that patients could undergo 24-hour liquid fasts or rapid sequence induction to minimize risks of aspiration. Ultimately, these actions are left up to shared decision-making.

Shortly after these guidelines were released, the British Association of Anaesthetists released guidelines in January 2025 with similar recommendations. They recommend continuing GLP-1 RAs in the perioperative setting without giving any exceptions. They also note that “upper gastrointestinal symptoms alone should not be used to determine gastric content.” In my opinion, these guidelines offer much better and clearer recommendations.

What does the evidence say?

Perhaps the best single paper on this topic comes from Oprea and colleagues and was published in the British Journal of Anaesthesia in July 2025. Their group, the Society for Perioperative Assessment and Quality Improvement (SPAQI), released their own multidisciplinary consensus guidelines along with a review of the current evidence.

Without making this post needlessly long, I will briefly summarize the evidence.

• Most of the aspirations in the literature occurred during MAC (i.e., sedation) cases.

• One study showed aspiration pneumonia events occurred at a similar rate to a historical cohort (4.6/10,000 vs. 4.8/10,000).

• Multiple studies, usually done during endoscopies (EGDs) with or without colonoscopies, have shown that patients on GLP-1 receptor agonists have residual gastric contents when fasting from solids for 8 hours (70% of patients, n=10) and when fasting from solids for up to 16 hours (13.1% after 10 hr, n=84; 5.4% after 12 hr, n=205; 56% after 13 hr, n=62; 24.2% after 14 hr, n=33; 19% after 16 hr, n=90).

• Patients with obesity and diabetes who were not on GLP-1 receptor agonists also had residual gastric contents at an alarmingly high rate.

  • In obesity and type 2 diabetes mellitus (T2DM): 5.1% had residual gastric contents after no solids for 14 hr, n=371.

  • In obesity: 20% after 8 hr, n=10.

  • In T2DM: 0.49% after 13 hr, n=205.

  • In obesity and T2DM: 5.0% after 16 hr, n=102.

• Only one study showed no residual gastric contents in all patients taking GLP-1 receptor agonists after having the patients fast from solids for 24 hours and from liquids for 12 hours (n=57).

We know that patients taking GLP-1 RAs have an increased incidence of residual gastric contents even when fasting. But what was an even bigger insight for me was the percentage of patients not on GLP-1 RAs who also had increased residual gastric contents. This could be interpreted as alarming and signaling that we should lengthen NPO times, particularly in diabetic patients who may have undiagnosed gastroparesis, or it could potentially be seen as reassuring, indicating that maybe a low volume of gastric contents isn’t as big of a risk factor for aspiration as once thought.1

The authors of this paper recommend that patients be fasted from solid food for 24 hours. This was based on the one study which failed to find residual gastric contents when patients performed a 24 hour fast from solid food and a 12 hour fast from liquids. The authors arbitrarily (and acknowledge that it is arbitrary) recommend fasting from liquids for 4-8 hours (4 hours for no/low-carbohydrate clear liquids and 8 hours for high-carbohydrate clear liquids). The optimal timing for fasting from clear liquids alone is unclear because all the included studies also had patients consuming solids.

Some countries, such as Australia and New Zealand, have adopted guidelines similar to these recommendations, asking patients to fast from solids for 24 hours.

What amount of residual gastric contents increases the risk of aspiration?

Perhaps this is the question we should be asking. Unfortunately, I do not have an answer for you. What remains unclear is whether patients taking GLP-1 RAs have an increased risk of aspiration. The data are conflicting, yet, most studies have not shown an increased incidence of aspiration.

One meta-analysis (April 2025) found no difference in incidence of aspiration (OR 1.04, 95% CI 0.87–1.25) despite the patients taking GLP-1 RAs having a higher incidence of residual gastric contents. However, the authors note that the certainty of evidence was low. Another meta-analysis (July 2025), looking only at endoscopies, found an increased incidence of pulmonary aspiration among patients taking GLP-1 RAs (OR 2.29, 95% CI 1.36-3.87).

I see people arguing that because we have not seen a clear increase in aspiration events, there is no need to treat these patients differently. I would be cautious with that approach. Because aspiration is a rare event, it may be a long time before we can know with high-quality evidence whether GLP-1 RA use increases aspiration risk. The precautionary principle would suggest we do our best to prevent aspiration events in these patients unless we have clear, high-quality evidence that patients taking GLP-1 RAs are not at increased risk of aspiration.

Where does this leave us?

What is clear is that patients should continue taking their GLP-1 receptor agonists in the perioperative period. There is no benefit in holding one dose of these medications and there is potential harm.

If fasting guidelines are meant to minimize residual gastric contents then the clear guidance would be to have these patients undergo prolonged fasting from solid food. Based on the available evidence, fasting from solids for 24 hours would be necessary. If fasting guidelines are instead meant to minimize aspiration events without respect to residual gastric contents, then I think we need to start questioning what amount of residual gastric contents (and what type) are clinically significant.2

Oprea and colleagues published a correspondence this month addressing some of the criticisms of their article. It is a well-written response, and one I highly recommend you read even if you do not read the primary paper(s). They write:

“The current literature supports modifying fasting recommendations for patients taking GLP-1 RAs as the most important mitigating factor. Although rare, aspiration pneumonia can result in hospitalisation, ICU admission, increased healthcare costs, and even death. When faced with such a prospect, a 24-h clear liquid diet does not seem to pose an equal risk or inconvenience.”

The bottom line is this: the most risk-mitigating approach based on the evidence would be to have patients fast for longer (ideally no solids for 24 hours and no liquids for perhaps longer than 2 hours). Would that be unpopular? Absolutely. Would it be safer? That remains less clear.

What are your thoughts on this topic? What are your organizations recommending?

Recently, I was maintaining general anesthesia for a routine case when the pulse oximeter suddenly decreased and the waveform deteriorated. My initial thought was that it might not be a true desaturation because the waveform was poor. Up to this point, the vitals had been stable—like train tracks. I checked the sensor and looked back at the monitor to see the end-tidal CO₂ abruptly dropping. That’s when I knew the pulse ox reading was real and the patient probably didn’t have a pulse.

By chance, a few months prior, I ran a code in the PACU. I walked away from that experience reflecting on how out of practice I was. So I made a short checklist for myself (see the end of the post for the checklist), and I started reviewing it regularly with a visualization exercise where I pictured myself executing each step. This brings us to my first lesson: regular mental rehearsal is paramount to high performance. A cardiac arrest in the operating room is a high-acuity, low-occurrence (HALO) event. If you want to be good at managing these types of situations, practice (ie, simulation) and mental rehearsal are necessary.

When I realized my patient was coding, I resorted to a simple “in case of emergency” checklist I keep in my head. Here is that checklist.

• Call for help

• Oxygen to 100%, max flow

• Cycle blood pressure cuff / verify transducers at correct height

• Change NIBP to cycle every 1 minute

• Notify surgeon

• Consider taking patient off ventilator

• Consider turning off anesthetic

• Consider turning off IV drips

My second lesson is to call for help first and to second-guess yourself second. There was a brief moment when I thought, “Is this data real?” While this is a reasonable thought, the most important thing to do is to get additional help. If it turns out to be a false alarm, no big deal. Conversely, time spent trying to confirm potentially suspect data is time wasted while the patient deteriorates. There is a saying in aviation, “Each subsequent bad decision reduces the remaining number of available good decisions.” Calling for help early preserves your pool of good decisions.

When I was a resident, I experienced multiple codes where the person in charge did not want an arterial line placed, despite having sufficient help to do so. I never understood this. Having an A-line gives real-time feedback about compression effectiveness, allows one to quickly and accurately perform pulse checks (simply look at the A-line during the briefest pause in compressions; no waveform = no pulse; and if systolic BP < 60 mmHg you should resume compressions), and allows easy blood sampling to guide treatment. My third lesson is to ALWAYS place an arterial line (using ultrasound) as soon as possible.

In this particular arrest, having point-of-care cardiac ultrasound was useful in confirming the diagnosis of a pulmonary embolism. I also strongly think that every cardiac arrest should have a central line placed, if sufficient help is available. One person could double-puncture the groin placing femoral central and arterial lines using just a central line kit; this can be a very efficient use of personnel and time.

This brings me to my last, and perhaps most important lesson: teamwork is everything. Anesthesia and surgery are team sports. Emergencies take that level of cooperation and coordination to the next level. My patient had a good outcome because many physicians, nurse anesthetists, nurses, scrub techs, and other healthcare providers were able to help out. Without this help, none of the other lessons I have outlined here would have been effective.

Hopefully, this sort of event rarely happens to you, but if it does, I hope this post and these checklists have helped you better prepare. If you’re looking to improve your code skills, I highly recommend checking out the various related episodes of Scott Weingart’s EMCrit podcast.

Here is my cardiac arrest checklist which I regularly review.

• Take charge

  • Verify patient is actually coding

  • If already on monitor, check for shockable rhythm

  • Ensure compressions adequate

  • Get crash cart

  • Assign rolls - recorder, compressors, monitor, drugs, airway

  • Drugs - epi, antiarrhythmics, etc.

  • Anterior-posterior (AP) pad placement is more effective than anterior lateral (AL) and should be used first even though it takes more time to place AP

  • Dual sequential defibrillation is more effective than single defib

  • Antiarrhythmics should be given right away, if indicated

• Maintain situational awareness

  • Gather patient history

  • Assess for reversible causes

• Further organize help

  • Get ultrasound

  • Intubate

  • A-line - perform ASAP

  • Check glucose

  • Labs - CBC, CMP, LFTs, coags, type & screen, troponin, BNP, ABG, lactate

  • Central line

  • POCUS (exclude tamponade, pneumothorax; RUSH exam for bleeding)

• Additional items

  • Consults (ICU, cardiology)

  • IV infusions (pharmacy)

  • Blood bank

If you have suggestions for improving these checklists, please leave them in the comments below.

However, as I learned more, the controversy over etomidate potentially increasing mortality in critically ill patients made me pause. I came to think that ketamine likely provided all of the hemodynamic benefits without the downside of adrenal suppression (whether this is clinically significant remains unclear). Despite recent studies failing to show an increase in mortality when using etomidate for intubation, my approach of minimizing potential risks led me to largely avoid it in favor of ketamine. Will the results of the newly published RSI trial push me back to using etomidate for intubation in critically ill patients?

Both EMCrit and PulmCrit have great discussions on this trial, and The Bottom Line has a detailed breakdown of the numbers. I would direct readers to their posts for a more in-depth analysis. In short, this was a multi-center trial that randomized 2,365 emergency department and ICU patients to receive either ketamine or etomidate for intubation. The primary outcome was 28-day mortality, with a secondary outcome of cardiovascular collapse—a composite of systolic blood pressure < 65 mmHg between induction and two minutes post-induction, new or increased vasopressor requirement, or cardiac arrest.

A non-significant 1% increase in mortality was seen with etomidate (28.1% with ketamine vs 29.1% with etomidate). I want to be reassured by this evidence, but since the study was powered to detect a larger mortality difference of 5.2%, it remains possible etomidate increases mortality. As Josh Farkas writes, “Searching out tiny mortality differences pushes us to the edge of the scientific method, because these hypotheses aren’t testable. This is almost more of a philosophical question than a scientific one. Ultimately, I don’t know exactly what to make of this.” I don’t think this trial moves the evidentiary needle on this topic.

A statistically significant increased rate of cardiovascular collapse (see above definition) was seen in the ketamine group (22.1%) compared to the etomidate group (17.0%), with an absolute risk difference of 5.1% (95% CI, 1.9 to 8.3). How meaningful is this? Most of the difference in the composite secondary endpoint was due to new or increased vasopressor use in the ketamine group (21.3% vs 15.9%; see image below).1 This difference could be due to differing patient characteristics. In the subgroup analysis, there were more patients in the ketamine group with sepsis (30.6% vs 20.9%) and with APACHE II scores ≥ 20 (31.4% vs 20.7%).

Perhaps a bigger reason there was more hypotension in the ketamine group was the drug dosing. Ketamine was dosed from 1.0-2.0 mg/kg of actual body weight, and etomidate was dosed at 0.2-0.3 mg/kg of actual body weight. The median dose of ketamine was 140 mg, 1.6 mg/kg (IQR 1.4-2.0 mg/kg), and the median dose of etomidate was 20 mg, 0.28 mg/kg (IQR 0.24-0.31 mg/kg). A large percentage of patients in both arms received greater than the maximum recommended doses of their respective induction agents.2 Only 5.8% of patients in the ketamine group received < 1.0 mg/kg. From my perspective, it would be difficult to exclude ketamine (over)dosing as the primary driver of increased post-induction hypotension.

What should we conclude from the RSI trial?

Unfortunately, it was underpowered to detect a mortality difference, so concerns about etomidate’s effect on mortality remain unresolved. This was a well-done trial with, in my mind, one major exception—the dosing of ketamine. Maybe ketamine causes more hypotension or maybe it was due to the higher dosing used. In critically ill patients, I would use ≤ 1 mg/kg of ketamine; this represents < 25% of the patients in the ketamine group. I wanted this trial to give me evidence to use etomidate again, but instead I’m left with more questions than answers.

What are your thoughts? Is this trial going to change your practice?

In 2025, a group of researchers conducted a prospective, multi-center, international study comparing three different fasting times in pediatric patients aged < 16 years. The guidelines at each participating center dictated whether patients were in one of three groups: clear liquids allowed until the patient was taken to the operating room (called the sip-til-send group), clear liquids allowed and encouraged ≥ 1 hour before induction, and clear liquids allowed ≥ 2 hours before induction of anesthesia (the control). The primary outcome was pulmonary aspiration defined as regurgitation or vomiting leading to respiratory symptoms requiring “high-dependency unit or ICU admission,” gastric contents seen in the trachea, or radiologic evidence of aspiration pneumonia.

In total, 306,900 anesthetics were included with 34,028 in the sip-til-send group, 251,021 in the ≥ 1 hour group, and 21,851 in the control group. There were 420 aspiration events categorized as either transient (n = 286, 68%), requiring escalation of care (n = 94, 22%), or requiring intensive care (n = 40, 9.5%).1 Aspiration incidence in the sip-til-send (1.18:10,000) and the ≥ 1 hour (0.96:10,000) groups was non-inferior to the control group (1.83:10,000) with 95% CIs of -1.48 to 3.63 and -0.34 to 3.76, respectively. When looking only at aspirations leading to ICU admission, the sip-til-send and ≥ 1 hour groups were again non-inferior to the control group.

Most of the aspiration events occurred in children fasting ≥ 2 hours (76%) compared to 5.9% in the sip-til-send group and 18% in the ≥ 1 hour group. The most common complications related to aspiration were hypoxemia (33%), laryngospasm (15%), aspiration pneumonia (5.4%), and bronchospasm (4.4%).

The main limitation of this study is that it was not randomized, as groups were allocated based on entire centers. It is possible that centers with more liberal fasting policies differed in other meaningful ways. Actual fasting times were analyzed for each aspiration event, but data were unavailable for 114 of the 420 cases. Lastly, because patients aged ≥ 16 years were excluded, we are unable to know if these results would be similar in adults.

The European Society of Anesthesia and Intensive Care (ESAIC) currently recommends that children be encouraged to drink clear liquids until 1 hour before anesthesia. The results of this study, and the authors’ own conclusion, support liberalizing pediatric fasting guidelines to ≤ 1 hour. If a sip-til-send policy were universally adopted, it could likely lead to fewer cancelled and delayed surgeries and higher patient satisfaction. I hope this new evidence can support meaningful change at your institution.

During fasting or stressful events such as an infection, ketone levels in the blood can increase to meet the body’s energy demands. Normally, insulin acts to decrease ketone production by inhibiting lipolysis, but SGLT2 inhibitors reduce this negative feedback. Simultaneously, the increased glucagon promotes further ketogenesis, leading to an increased risk of ketoacidosis. Because of the glycosuria effect, this ketoacidosis is usually euglycemic, although diabetic ketoacidosis (requiring blood glucose ≥ 200 mg/dL) can also occur.

In 2020, the FDA issued guidance recommending SGLT2i be held for three days prior to surgery to reduce the risk of euglycemic ketoacidosis. Since that time, research has questioned the benefit of this recommendation. As with any decision, the benefits of reducing the harms of an uncommon side effect should be weighed against the downsides of having poorly controlled chronic disease (e.g. blood glucose) in the perioperative setting.

A group of researchers set out to determine if patients with type 2 diabetes who were taking SGLT2i at the time of emergency surgery had a higher incidence of diabetic ketoacidosis than diabetic controls not taking SGLT2i. They performed a retrospective cohort study including 34,671 patients, of whom 2,607 were using SGLT2i. Inclusion criteria included undergoing one of 13 different emergency surgeries. Patients prescribed SGLT2i were only included if a recent prescription would have lasted through the date of surgery. Because of the emergency nature, it was assumed these patients were taking their SGLT2i; however, exact timing remains a limitation of this study. The primary outcome was diabetic ketoacidosis or acidosis (defined by ICD-10 codes) within 14 days postoperatively.1

They found the unadjusted incidence of acidosis to be 4.9% in those taking SGLT2i and 3.5% in the unexposed group. After adjusting for covariates, SGLT2i use was not associated with an increased risk of acidosis (3.8% vs 3.5% in the unexposed group). The major limitations of this study were its retrospective, non-random nature, the exclusion of non-diabetic patients taking SGLT2i, and the lack of access to laboratory data to further inform the primary endpoint.

Subscribe now

How should guidelines be changed?

Multiple medical groups in Europe recommend holding SGLT2i for only 24 hours and/or only holding them for patients undergoing major surgery or who are hospitalized with serious illness. The French Diabetology and Anesthesiology society recommends continuing SGLT2i for ambulatory surgeries and only holding them on the day of surgery for minor and major surgeries.

It may be reasonable to only hold SGLT2i for major surgeries or when patients are expected to continue fasting into the postoperative period. Considering that people routinely fast overnight even when not undergoing anesthesia and many also skip breakfast each day without negative consequences, it seems unreasonable to hold SGLT2i in patients undergoing minor procedures when they will resume eating the same day.

If a patient fails to hold their SGLT2i, measuring serum or urine ketones could help guide decision-making. In the U.S., Northwell Health System created a protocol to test serum β-hydroxybutyrate (BHB) in patients who did not hold SGLT2i. Patients with BHB concentrations < 1.5 mM can proceed with surgery. For patients with BHB concentrations of 1.5-3.0 mM, anion gap is also assessed. At the University of Pennsylvania, a BHB cutoff of 2.0 mM with anion gap > 12 leads to postponement of surgery. While research has yet to determine the optimal BHB level to mitigate risk, these protocols have helped avoid cancellations in low-risk instances.

Until the FDA changes its recommendation, it may be prudent for institutions to create their own guidelines weighing the risks of ketoacidosis against those of cancelled surgeries and poorly controlled disease resulting from held medications.

Let’s start with electrical tape and home electricity. Electrical tape is an excellent insulator, adding such a high resistivity that, in ideal conditions, a single layer can provide substantial protection from household voltages of 120-240 V at 60 Hz.1 If the electrocautery used in surgery operated under these household conditions, applying electrical tape—not medical tape—to jewelry could (theoretically) offer some benefit. However, electrocautery differs significantly from your home’s lightbulbs.

With bipolar electrocautery, the current exits one electrode and returns to an adjacent electrode at the tip of the unit. With monopolar electrocautery, current leaves the single tip electrode and returns through the return electrode—commonly and incorrectly called the “grounding pad.”2 The return electrode pad should offer the lowest resistance path for the current to travel.

The concern with metal jewelry is two-fold. Jewelry placed close to other conductive objects can create an unintended return path for current, resulting in a burn at the site of the jewelry. Even without forming an alternate return path, metal jewelry can concentrate current density beneath it, heating the metal and the adjacent tissue enough to cause a burn. To prevent this, the return electrode should be well-adhered to dry skin, and all metal not removed from the patient’s body should be kept well away from conductive material (e.g. IV pole, bed rail, etc.). While the likelihood of an alternate, lower resistance current path being created with, for example, the IV pole with its plastic or rubber wheels is very low, it is not impossible.

Electrical tape will not provide protection against burns because electrocautery units operate at a frequency in the hundreds of kHz to low MHz range—much higher than a household’s 60 Hz. As frequency increases, more current can flow through the insulation via capacitive coupling. At very high frequencies, displacement current comes into play allowing current to flow across air gaps due to the creation of electric fields. The way to prevent current from moving in this manner is to have sufficiently large air gaps between adjacent conductive surfaces—in other words, keep the IV pole far enough away from the metal ring on a patient’s finger, and even more importantly, have a well-placed return electrode pad.

In contrast to electrical tape, medical tapes have no insulating properties; they will not stop current leakage, and they will not prevent burns. If you are going to tape metal jewelry to prevent it from getting caught on objects or falling out, at least do not fool yourself or the patient into thinking you are providing safety from burns. Patients who cannot remove, or who refuse to remove, metal jewelry should be educated on the risk of burns. If there is a high concern for burns, bipolar electrocautery should be used instead of monopolar.

All general anesthetic agents have the potential to cause delirium.1 This risk is higher with sevoflurane maintenance compared with total intravenous anesthesia (TIVA) using propofol, and it is higher in both pediatrics and older adults. The important question is, does premedicating a patient with midazolam increase the risk of delirium beyond what would occur during general anesthesia alone?

Researchers examined this question by conducting a retrospective cohort study in patients ≥ 70 years old undergoing non-cardiac surgery under general anesthesia. Patients were divided into two groups based on whether they received midazolam premedication or not. The primary outcome was delirium determined by one of several standardized methods performed routinely in the PACU and following admission. Exclusion criteria included patients with dementia, surgeries lasting < 30 minutes, craniotomies, direct postoperative admission to the ICU, and postoperative mechanical ventilation.

A total of 1,973 patients were included with a median age of 75 years. Forty percent of the patients received midazolam. The overall incidence of postoperative delirium was 15.3% (n = 302). Midazolam was not associated with an increased risk of delirium adjusted odds ratio 1.09 (95% CI 0.82-1.45; p = 0.538).2 This persisted in subgroup analysis looking at patients older than 80 years, patients with preoperative cognitive impairment, and those undergoing high-risk surgeries.

The study is notably limited by its retrospective nature with potential for confounding. Those receiving midazolam were more likely to be younger and have lower ASA scores, yet were more likely to undergo high-risk surgeries. These differences could have affected the results even though the authors attempted to control for them. The exclusion criteria of surgeries < 30 minutes leads one to wonder if midazolam could increase delirium risk in short cases where it may not be worn off by the time of emergence. Additionally, stratification by the agents used to maintain anesthesia could have been informative. Perhaps midazolam does not increase the risk of delirium with volatile agents but does with TIVA.

Despite its limitations, this study adds to prior research which has shown similar results. Given this, if I have a patient who may benefit from preoperative anxiolysis, I do not consider age as an absolute contraindication to midazolam use.

A group of researchers set out to determine if a simple PEEP estimating strategy using BMI could give anesthesiologists a more practical means of setting PEEP. They conducted an RCT comparing PEEP of 5 cm cm H₂O (PEEP-5 group) to a PEEP set equal to BMI divided by 3 (PEEP-BMI/3 group) with 30 participants in each arm. Participants were undergoing elective, non-cardiac, non-thoracic surgery under general anesthesia with endotracheal intubation. Laparoscopic surgery and any surgery requiring more than 10° of Trendelenburg or reverse Trendelenburg were excluded. Pregnant patients and those with BMI > 60 kg/m² were also excluded. Ultrasound was used pre- and post-operatively to determine lung aeration and airway pressures were measured.

The PEEP-BMI/3 group had lower driving pressures (PEEP-BMI/3 group median 7.9 cm H₂O compared to PEEP-5 group median 8.9 cm H₂O, p = 0.027) meeting the study’s primary endpoint. The intervention group also had higher compliance (PEEP-BMI/3 group mean 0.95 mL/cm H₂O/kg of predicted body weight compared to PEEP-5 group mean 0.83 mL/cm H₂O/kg, p = 0.02), improved aeration (p ≤ 0.01), and decreased atelectasis (no new atelectasis in PEEP-BMI/3 group compared to 7 new cases in PEEP-5 group, p = 0.011). The median PEEP in the PEEP-BMI/3 group was 9 cm H₂O with a range of 7-13 cm H₂O, which represents a fairly modest increase.

These results may seem promising as a way of estimating PEEP, but the limitations should give us pause. Despite being a positive study, the authors note that the small difference in driving pressures may not be clinically impactful. Furthermore, the mean (SD) BMI across all participants was 27.8 kg/m² (4.8). This represents a population fairly close to normal weight individuals. Whether this BMI PEEP strategy would provide optimal PEEP, or perhaps too much PEEP, in individuals with class 2 and 3 obesity is not clear. Two groups I am particularly interested in, those undergoing laparoscopic surgery and those in steep Trendelenburg positioning, were both excluded. Taken together, these limitations significantly reduce the ability to generalize these results across patient populations.

In a reply to this article, Kozanhan and colleagues pointed out the limitations of using BMI to assess body habits. Body mass index does not distinguish between lean mass and adiposity. Nor does it separate patients who have central versus peripheral obesity—the former likely having a more negative effect on respiratory mechanics. They hypothesized that using waist circumference or waist-to-hip ratio may be more appropriate than using BMI.

While an accurate and accessible way to determine PEEP for all populations remains to be developed, what this paper emphasizes to me is the importance of using higher PEEP values for those with overweight and obesity. Hopefully, in the years ahead, we will have a simple and accurate way to determine optimal PEEP across patients, surgeries, and patient positions.

The article by Lam and colleagues is a systematic review and meta-analysis examining aspiration rates between different fasting conditions. The authors were critical of prior research that focused on surrogate outcomes such as residual gastric contents or gastric pH claiming these endpoints have never been shown to predict aspiration in humans. Their main argument appears to be that more research is needed to recommend fasting. However, it is unclear whether they are arguing for NPO times shorter than currently recommended 2 hours for clears or if they are arguing to encourage more people to follow guidelines allowing clear liquids up to 2 hours before surgery?

Lam and colleagues included 17 studies from January 2016 to December 2023 with 801 patients in the experimental group and 990 in the control group. The meta-analysis was confusing because some studies compared American Society of Anesthesiologists (ASA) fasting guidelines to non-fasted patients and others compared strict NPO after midnight to clear liquids up to 2 hours before surgery (the current ASA recommendations). The analysis did not clearly separate patients who were fasting and patients who were not.

The authors excluded studies prior to 2016 because:

The systematic review that was performed to support the 2017 ASA fasting guideline did not report any studies of witnessed aspiration, we began our search 1 year before the guideline's publication assuming that (1) there were few studies reporting aspiration as an outcome for the time period that the literature was reviewed by the ASA and (2) older literature would not apply to current anesthesia practice.”

The problems I see with this argument are 1) multiple studies exist prior to 2016 looking at aspiration [1] [2] [3] [4], and 2) even though anesthetic medications have evolved, the act of removing a patient’s reflexes thereby increasing the likelihood of aspiration has not changed.

Statistical power

In support of the authors’ hypothesis, the study found no statistically significant difference between aspiration events in the experimental group (n = 4, 0.50%) and the control group (n = 7, 0.71%). The absolute difference between the groups is 0.0021. This is very small. To detect a statistically significant difference with ⍺ = 0.05, a power calculation estimates each group would need about 21,405 patients. This is roughly 24x more patients than were included in this study. Enrollment of this size is precisely why surrogate endpoints are so often used.

Note the wide confidence interval suggesting the lack of power.

A type II error occurs when a study claims there was no difference when in fact there was a difference. It is impossible to say from this research if we would find a difference between the groups, so we cannot say that a type II error was committed. These results are inconclusive.1 However, the authors erroneously conclude, “The lack of evidence demonstrating a relationship between fasting and aspiration suggests that fasting policies should be liberalized…”

Looking deeper at the included studies

Only two of the included 17 studies were identified as having aspiration events. Of these, two aspiration events in the experiential group and four in the control group were from a study looking at aspiration in mechanically ventilated SICU patients who had tube feeds stopped for < 6 or ≥ 6 hours prior to surgery. The paper describes: "aspiration was identified in the progress notes and the discharge summary." It does not detail if aspiration events occurred before, during, or after intubation, during anesthesia, or while not intubated. Nor does it detail whether these patients aspirated around the cuffed ETTs. Generalizing from an intubated ICU population to fasting guidelines for non-intubated patients undergoing non-emergency surgery seems precarious.

The only other study with "aspiration" events showed an event rate of 2% (n = 2) in the experimental group and 3% (n = 3) in the control group. This rate is roughly 100x that reported in prior literature of 0.01% - 0.24%. This should give anyone pause, and indeed the devil is in the details. Sharma et al. compared strict NPO after midnight to patients allowed to have clear liquids until 2 hours before surgery (current ASA guidelines). They write:

To evaluate regurgitation during anesthesia, a piece of turnsole paper was inserted at the end of the pharynx, and in case of color change toward acidic pH, an event of regurgitation was reported as positive.

Although Table 3 lists two aspirations in the experimental group and three in the control group, the discussion indicates these events were simply positive color change on the pH paper rather than pulmonary aspirations. The authors report, “In our study, we found no evidence of aspiration or related events in any patient in both groups.”

In summary, one article showed aspiration events in a different patient population and the other explicitly stated it did not have aspiration events. What are we to make of a meta-analysis which did not clearly sort the fasting groups and which had no [relevant] events in either arm?

What is the argument in favor of liberalizing fasting times?

I am unconvinced that fasting causes acute harm. The citations the authors included do not support their claims that:

Fasting adversely affects patients by its impact on either health-related quality of life, postoperative outcomes, or metabolic complications such as hypoglycemia and insulin resistance.

One reference showed higher patient satisfaction when allowed to consume clear carbohydrate beverages up to 2 hours before surgery. This is in line with current fasting guidelines. Another reference showed surgical outcomes were improved with decreasing percentages of HbA1c—which, while important, is not relevant to acute fasting. The third citation found acute increases in insulin resistance when patients were fasted for 24-to-72-hours, discussing this largely in the context of postoperative fasting.

Exercise acutely causes many changes which could be perceived in isolation as harmful. It causes hypertension, tachycardia, sympathetic stimulation, inflammation, and increased oxidative stress. Yet we know exercise is one of the most potent interventions for health. Can it truly be argued that acute insulin-resistance from an overnight fast is harmful? Furthermore, humans fast for more than eight hours every night without metabolic harm. Intermittent fasting or time-restricted eating is commonplace and has been shown in multiple studies to improve weight and metabolic markers (although the effect is more likely due to overall caloric restriction rather than timing of food consumption).

For patients taking insulin, there is a real concern of hypoglycemia while fasting. Beyond this, is there actually clinically meaningful harm from short-term fasting?

The argument in favor of fasting

Aspiration is a rare event with potentially catastrophic consequences including death. As the saying goes, no one has performed a randomized controlled trial in humans looking at the effectiveness of parachutes. Do we need to perform an RCT to conclude that fasting is the safest option? Lack of evidence does not necessarily mean there is lack of benefit.

However, we now know from the recent research on GLP-1 receptor agonists that many more patients are presenting to surgery with increased residual gastric contents. So far, rates of aspiration have not been shown to be significantly above historical averages. For this reason, I am open to the idea that there may be a higher cutoff of residual gastric volume which is considered low risk for aspiration. Perhaps fasting is not as critical as we have led ourselves to believe. More research should be done. In the meantime, I still think the most risk-mitigating approach is to have people fast prior to surgery.2

As the authors note, “anesthesia-associated aspiration is and always was rare.” Rare events will always be difficult to study. When rare events can have catastrophic outcomes, it will always be prudent to safeguard against them. While the authors may not have committed a type 2 error, they still fooled themselves into thinking they had evidence of no benefit when what they really had were inconclusive results—a significantly less splashy headline. As Richard Feynman once said, “The first principle [of science] is that you must not fool yourself and you are the easiest person to fool.”

A recent paper by Kakalecik and colleagues sought to shed light on the subject by retrospectively looking at patients with tibia fractures who did or did not receive regional anesthesia. Acute compartment syndrome is most commonly seen following long bone fractures. Because of the small fascial compartment surrounding the tibia, tibia fractures have a higher risk of developing compartment syndrome. The study included non-pregnant, adult patients with tibia fractures at a single level-one trauma center from January 2015 until April 2022. Patients with known neurological injury, ipsilateral knee dislocations, and those who underwent prophylactic fasciotomy were excluded. Missed acute compartment syndrome was defined as having a motor deficit at three-months postop.

Subscribe now

Patients had received either a single shot regional block with 15-20 mL of 0.2% ropivacaine (n = 48), and 0.2% ropivacaine peripheral nerve catheter without an initial bolus (n = 538), or no regional anesthesia (n = 176).1 The incidence of missed compartment syndrome was 0.7% (4/610, 95% CI, 0.2%-1.7%) in those receiving regional anesthesia compared to 1.7% (3/181, 95% CI, 0.4%-4.8%) in those not receiving regional with no significant difference (p = 0.19). All of the missed cases in the regional group were in those who had nerve catheters. Not surprisingly, patients who had regional anesthesia received lower morphine milligram equivalents (MME) in the first 24 hours compared to those without regional.

The major limitation of this study is its retrospective nature allowing for possible confounding. When looking at the injury severity score (ISS) of patients, this becomes a concern. Patients who received regional anesthesia had a lower mean ISS (mean difference -7.2, 95% CI, -8.6 to -5.9) and lower incidence of Gustilo-Anderson type IIIA or IIIB injuries. Unfortunately, criteria detailing who could receive regional anesthesia were not included in this paper.

While the lack of significant difference between the groups is encouraging, I do not think there is enough information to say anyone with a tibia fracture can safely receive regional anesthesia. These results suggest that regional anesthesia is safe enough to warrant an RCT looking at single shot blocks and peripheral nerve catheters in patients with tibia or other long bone fractures. Until that time, these patients should be offered regional anesthesia on a case-by-case basis after discussion with the surgeon.

In two randomized controlled trials, suzetrigine was used to treat pain following either abdominoplasty or bunionectomy with participants being divided into three arms: suzetrigine, hydrocodone 5 mg + acetaminophen 325 mg, and placebo. All patients in the bunionectomy group had a continuous popliteal sciatic nerve block catheter with ropivacaine 0.2% which was discontinued on postop day one prior to randomization. All groups were allowed to have ibuprofen to treat breakthrough pain. No other NSAIDs or opioids were permitted.

The primary endpoint was the time-weighted sum of the pain-intensity difference over 48 hours (SPID48) following the first dose of the randomized drug. With this metric, a higher number shows greater pain relief, and a negative SPID would indicate worsening pain. The least squares mean difference compared to placebo was 48.3 (95% CI 33.6 to 63.1 p < 0.0001) following abdominoplasty and 29.3 (95% CI 14.0 to 44.6 p < 0.0002) following bunionectomy. There was no significant difference when comparing suzetrigine to the hydrocodone + acetaminophen groups.

Side effects were similar with suzetrigine and placebo. The most common side effects were nausea, vomiting, constipation, headache, dizziness, and hypotension. The incidence of nausea or vomiting was significantly lower with suzetrigine compared to hydrocodone (n = 91 vs n = 150 with abdominoplasty and n = 39 vs n = 71 with bunionectomy for suzetrigine and hydrocodone, respectively).

While the similar efficacy to hydrocodone makes suzetrigine seem like a promising new medication, some important caveats and unanswered questions remain. Rescue medication (ibuprofen) use was similar between suzetrigine and placebo. However, rescue medication use was not compared to the hydrocodone groups. If that group received significantly less ibuprofen than the suzetrigine groups, it could make suzetrigine appear more effective than it really is. This is the biggest point giving me pause. On the other hand, if everyone in the suzetrigine group had been given acetaminophen, would that group have had superior pain control compared to hydrocodone?

Other major limitations include a patient population in which most participants were female (98% in abdominoplasty and 85% in bunionectomy groups) and white (70 and 71% in the two trials). Patients taking chronic opioids or NSAIDs were excluded limiting the ability to generalize the results to a chronic pain population. Suzetrigine’s use as part of a multimodal approach was similarly not evaluated.

Fortunately, more clinical trials are ongoing. I hope we will see suzetrigine evaluated alongside regional anesthesia to ensure its safety when co-administered with local anesthetics which are non-specific voltage-gated sodium channel blockers. In the meantime, I would approach suzetrigine use with cautious optimism.

A 2024 paper looked at whether scopolamine use was associated with an increased risk of delirium and mortality, among other outcomes, within seven days postoperatively. Patients who underwent inpatient or outpatient surgery (urology, colorectal, orthopedics, neurosurgery, vascular, and thoracic) and received scopolamine within 24 hours of surgery were enrolled. Exclusion criteria included cardiac surgery, patients in the ICU, palliative care patients, and those who had received scopolamine or antipsychotics within the prior three months.

Individuals were grouped into cohorts by age (20-29 years, 30-39 years, … 70+ years) and propensity score matched to yield 201,908 pairs. Across all age cohorts scopolamine use was associated with increased postoperative mortality. The relative risk of mortality increased with increasing age cohort. Among those aged 30-39 years, the relative risk was 2.414 (p < 0.0001). Among the 70+ cohort, the relative risk of 7-day postoperative mortality was 4.126 (p < 0.0001). Only the 20-29 cohort was not statistically significant (RR 1.923, 95% CI 0.984-3.757, p = 0.0514); this cohort also had the least number of individuals and may have been underpowered to detect a difference.

The authors only found a statistically significant increased rate of delirium in those aged 60+ years. However, scopolamine use was associated with new antipsychotic use across every cohort which could suggest that delirium was more prevalent than was recorded in the database from which the data was gathered. When looking at other complications, the risk of readmission was higher in the scopolamine group across all cohorts except those aged 20-29 years, and the risk of urinary retention was increased across all cohorts.

Perhaps the biggest limitation with this paper is that scopolamine dose, route of administration, and duration of use were not evaluated. The inclusion criteria of scopolamine within 24 hours of surgery leads one to reasonably assume it was administered as a patch for prevention of PONV, yet the route and total dose remain a major limitation. In the absence of a randomized controlled trial which may never occur, this evidence was enough for me to stop using scopolamine.

Given these complications, and the many alternative drugs to prevent and treat PONV, the time has come to retire scopolamine in favor of safer antiemetics.

Aprepitant given at a dose of 40 mg is non-inferior to ondansetron in preventing nausea but superior in preventing vomiting with effects lasting 24-48 hours. A higher dose of 125 mg of aprepitant was not more effective than the 40 mg dose.

When looking at the pharmacokinetics, you see that 150 mg of fosaprepitant yields an area under the concentration-time curve (AUC) of 37.4 (± 14.8) mcg•hr/mL at 20 minutes after infusion. Compare this to the AUC of 7.8 mcg•hr/mL for 40 mg of oral aprepitant. Fosaprepitant results in a level ~4.8x higher than aprepitant suggesting a dose closer to ~31 mg IV would be effective at preventing PONV.

Fortunately for us, we don’t need to guess at the dosing. In 2022, Aponvie, an IV version of aprepitant (not a prodrug like fosaprepitant), was approved for prevention of PONV at a dose of 32 mg IV. The AUC at that dose was 7.8 mcg•hr/mL—the same as aprepitant.

All this suggests that when using fosaprepitant for the prevention or treatment of PONV, a reasonable dose would be 32 mg or perhaps 40 mg IV. Doses higher than that are unlikely to confer additional benefit.

Share

Leave a comment

PEEP has many benefits and little downside. It improves pre-oxygenation in obese patients. It decreases atelectasis and post-operative pulmonary complications and it can improve oxygenation. To anyone not routinely using PEEP, I ask, why not?

Why would someone want to avoid PEEP? High levels of PEEP can impede venous return and worsen hypotension. If a patient has significant hypotension, lowering the PEEP to 5 may be reasonable, but I would argue vasopressors, fluid, and blood are more effective treatments depending on the cause. If a patient has a significant air leak (e.g. bronchopleural fistula) or a tension pneumothorax, PEEP could worsen the condition.

A new systematic review and network meta-analysis from March 2025 reaffirms the benefits of PEEP. Their findings showed any PEEP is better than no PEEP for reducing postoperative pulmonary complications including pneumonia. Higher PEEP (defined in the study as > 5 cm H2O) is better than lower PEEP at preventing atelectasis.

The study defined low tidal volume as 4-8 mL/kg of ideal or predicted body weight and high tidal volume as > 8 mL/kg of ideal or predicted body weight. PEEP was chunked into three categories: 0, 1-5, and > 5 cm H2O. Postoperative pulmonary complications was a composite of many different outcomes which were defined differently by various studies included.

• For preventing post-op pulmonary complications and pneumonia:

  • Any PEEP is better than no PEEP (moderate certainty)

  • Low tidal volume is better than high tidal volume (moderate certainty)

• Low tidal volume with PEEP is better at reducing hypoxemia compared to high tidal volume with no PEEP (high certainty)

• PEEP > 5 cm H2O is superior to no PEEP for prevention of atelectasis (moderate certainty)

As with all studies there are limitations. I would have liked to see PEEP of 5 as a standalone group since that is probably the most commonly used PEEP setting. Recruitment maneuvers with PEEP did not show clear benefit over PEEP alone, however, this was probably inadequately parsed out. I would like to see it tested in specific populations (obesity, morbid obesity, laparoscopic surgery, Trendelenberg positioning, and combinations of the previous characteristics). I suspect recruitment maneuvers combined with increasing PEEP would show more benefit in obese and morbidly obese individuals in Trendelenberg than in people with normal BMI regardless of positioning.

My conclusions from this study are: use at least a PEEP of 5 and tidal volumes not greater than 8 mL/kg of ideal body weight for all patients. If you don’t use PEEP or if you use high tidal volumes, please share your evidence for supporting such practices.

As always, if you notice any errors, please reach out to me.