Thoughts about our recent discussion of the article "No association between preprocedural fasting and witnessed pulmonary aspiration: A systematic review and meta-analysis."
I’m honored that so many have read our research and commented upon it.
Dr. Nielson-we struggled with the term liberal versus conservative fasting. Admittedly, some of my co-authors didn’t like that terminology, but I insisted on it. I wasn’t sure of a better way to classify comparisons between one group of fasting regiments with others that were quite heterogeneous.
Although the measurements were all different, they represented one approach to fasting, and then another that involved a lesser degree of fasting.
But that wasn’t the major point of the work. The bottom line is that almost every study of fasting used the intermediate outcomes of gastric volume or pH. Neither of these outcomes have been proven to correlate to human disease. Every study cited all, but the most recent ASA guidelines for fasting used either gastric volume or gastric pH as the study end point. Even the most recent ASA guideline, which was only a partial update, only used gastric volume and pH for meta-analytic purposes They did not use other outcomes such as witnessed aspiration in the calculations.
These end points were conjured up from animal studies in the 1970s. What I found surprising in reviewing the literature quite carefully was that almost no one has questioned the use of these intermediate outcomes.
None of the ASA publications discussed the serious limitations the of using untested surrogate outcomes. Nor did they downgrade the evidence scoring sufficiently to account for the fact that untested intermediate outcome was used in the studies. Based on my review, all the studies cited in all of the ASA guidelines should have been rated as very low quality because they relied on intermediate outcomes.
From my perspective – that of someone who managed the EBM portfolio at JAMA for nine years – this is a striking observation. We talked about these kinds of things at the journal all the time. Yet, they are hardly addressed anywhere in the extent literature for fasting before surgery.
From my perspective, there is no evidence to support any particular fasting guideline. That is because of the reliance on these untested intermediate outcomes.
Another observation I found interesting was that no one had done a meta-analysis of witnessed aspiration. Meta-analyses are at their best when they are aggregating studies reporting rare outcomes to help provide sufficient numbers of events to facilitate meaningful conclusions.
When we did that, we found that aspiration is extremely rare. We reviewed the original article that posited a relationship between a full stomach and aspiration and a need for fasting. Aspiration was very rare before they were fasting policies and it remains very rare today despite excessively long fasting times.
Being a surgeon, and a former chief of surgery, I believe that case cancellations because of “inadequate” fasting are not acceptable. I have had to tell many patients that their surgery was delayed or canceled because they ate too soon before their scheduled surgery. I’ve also had to deal with the severe operating room, disruptions and loss of revenue because of these cancellations.
More recently, I was a patient myself and had to sit in the hallway all by myself, waiting for my surgery for about an hour or two because I hadn’t passed that magical eight hour fasting time that was required in the facility where I had surgery.
The attention this article has gotten represents how many people are exposed to the problems caused by fasting policies. I’m hoping our article stimulates serious reconsideration of these fasting policies and a more rational approach to research into how to optimize preoperative management of patients.
In GI there are various protocols for drinking colytely, from finishing the last portion by midnight for a morning procedure, to finishing the last half two hours before. That would make easy accruing a large number of patients with which to make a RCT comparison for conscious sedation.
Great post. And great teaching point in service of reviewing evidence.
As always, the “methods” section is the key. In this case, what exactly did “liberal” mean? Some clear fluids is one thing; steak and egg brekky is likely quite another.
And there’s always a disconnect btw what the physician says (or intends) vs what the patient hears (or wants to hear).
Also, cancelling a procedure means a wasted spot which means another patient on the waitlist is having to wait that much longer.
I’m not a surgeon but I do procedures that require conscious sedation.
And I’m a believer in KISS principle.
My practical take-away is this: I will instruct pts the same way. I will still say NPO from midnight except meds with sips in AM. I will still advise 6 hrs NPO for procedures later in the day. But I will cancel procedures less often by being less strict with those who fail to adhere to the letter of those instructions, who have had clear fluids closer to procedure time than advised.
Speaking of nuance, perhaps one needs to consider what type of elective surgery, how long the wait to re-schedule, the consequences of delaying the surgery, the medical status of the patient, and the facility within which the procedure is being done. There are a lot of moving parts.
I remove wisdom teeth in my office under IV anesthesia. We are rural and the nearest "hospital" is 10 miles up the road. The next nearest "hospital" is an hour away. If we need a true place of higher level of care, that is a 2 and a half hour drive. Considering that I have almost zero external support and that 99.99% of the time what I do can hardly be considered life saving, I have a very, very low threshold for canceling cases where the patient is not NPO for at least 6 hours. But I also have the luxury of saying to them that we can proceed with local anesthesia only or we can re-schedule in just a week or two when you follow directions.
Considering another scenario, a patient might be scheduled for tumor removal in a large center with oodles of support. Perhaps that patient had their latte 3 hours before their surgery. Perhaps postponement might make a huge difference in their outcome. That might be a circumstance where the anesthesiologist would recommend proceeding. There has to be nuance in the decision to proceed or not.
Interesting. As a former lactation consultant, I take an interest because we lobbied hard and mostly successfully for counting breastmilk as a clear liquid for purposes of fasting guidelines. Given the comment below and the article, clearly there is high variability not just in what patients' stomachs actually contain just prior to surgery, but in whether the anesthesiologist even knows what's in there, and also in surgeons'/ anesthesiologists' tolerance for reported or discovered intake. Those intersecting major areas of uncertainty would all have to be sternly controlled for in a large RTC. I'd rather read (1) a comparison of aspiration events in emergency surgeries where fasting could not take place (subdivided by time of day or some such rough proxy for likelihood of a recent full meal), versus aspiration events in planned surgeries where fasting orders were given; and (2) a stringent analysis of all aspiration events over a period of years.
Here’s what the medical literature shows regarding (1) comparisons of aspiration rates in emergency surgeries (where fasting may not occur) versus elective surgeries with fasting orders, and (2) large‑scale analyses of aspiration events over time.
⸻
1. Emergency surgery without fasting vs elective surgery with fasting
• A 2020 retrospective review at a single center in acute care surgery (appendectomy/laparotomy in 2016–17) compared unfasted patients (<8 h NPO) to those fasted (>8 h). There were zero aspiration events in either group—aspiration risk was statistically equivalent (relative risk ≈ 1.0) .
• In procedural/emergency sedation settings (emergency department sedation in adults and children), systematic reviews found only one aspiration case out of over 22,000 sedations, and non‑fasting status was not linked to increased risk .
• In children, a German multicenter prospective observational study (12,093 cases) found regurgitation in 0.26%, suspected aspiration in 0.08%, confirmed aspiration in 0.03%. Emergent status increased risk (odds ratio ≈ 2.8), but overall confirmed pulmonary aspiration remained rare and resolved quickly .
Summary: High‑quality observational data suggest that non‑fasted patients undergoing emergency or urgent surgery do not experience a statistically higher risk of aspiration than fasted patients in elective surgery—particularly when modern airway protections like rapid sequence induction are used.
⸻
2. Rigorous analyses of aspiration events over time
• A new meta‑analysis published June 2025 by UCLA examined 17 studies (801 in experimental arms, 990 controls) on fasting duration and clinical aspiration. Aspiration was rare—0.50% vs 0.71%—and no significant difference was found between fasting regimens and controls. Trial sequential analysis suggested further studies are unlikely to change that conclusion .
• Older reviews estimate aspiration incidence in elective general anesthesia at roughly 1.4–6 per 100,000 cases (~0.0014–0.006%) .
• Large ED procedural sedation series (JAMA Pediatrics, >2,600 pediatric cases) reported no clinically apparent aspiration events and found no association between fasting times and adverse events like vomiting or other sedation complications .
⸻
🔍 Answer to your two questions
(1) Comparison of aspiration events: Emergency (non‑fasted) vs elective (fasted)
• While we lack large randomized controlled trials directly assigning fasting vs non‑fasting in emergency surgery, several retrospective and observational studies show no increased aspiration in non‑fasted emergency cases when proper anesthesia protocols (rapid sequence induction) are used.
• Emergency pediatric cases do show a modest increase in regurgitation or suspected aspiration, but confirmed aspiration remains rare and usually benign  .
• Time‑of‑day stratification (e.g. likelihood of having eaten recently) is rarely detailed in published datasets; up‑to‑date emergency vs elective comparisons typically focus on emergent status rather than specific timing or meal presence.
(2) Period‑wide analyses of aspiration events
• Meta‑analysis covering multiple studies up to 2023 found aspiration is very rare (~0.5–0.7%) in clinical settings and not impacted by different fasting regimens, based on 17 RCTs/observational arms .
• Broader epidemiological data across decades consistently puts aspiration rates in elective GA around 1–6 per 100,000 cases (~0.001–0.006%)  .
⸻
🚑 Additional context: Why emergency cases may not show elevated risk
• In emergency surgery, anesthesiologists typically use rapid sequence induction (RSI) with a cuffed endotracheal tube—this protects the airway even when the stomach is full  .
• The risk of delaying a life‑saving surgery far outweighs the low risk of aspiration. Safety is optimized by airway techniques, vigilant suctioning, and early intervention if aspiration occurs —not by waiting for gastric emptying  .
⸻
✅ Limitations
• No large RCTs directly randomizing fasting vs non‑fasting in emergency surgery, for obvious ethical reasons.
• Time‑of‑day or recent full‑meal versus fasted subgroups are rarely captured in uniform data.
• Most existing evidence comes from observational retrospective or prospective cohort studies, not experimental trials—but these are well‑controlled and consistent across settings.
⸻
###📊 Summary Table
Study Type Setting Fasting Status Aspiration Incidence Key Findings
Retrospective cohort (n=100) Acute-care appendectomy/laparotomy < 8 h vs > 8 h 0 events in either group No increased risk in non‑fasted emergencies 
Pediatric prospective observational (n≈12k) Mixed planned/emergency Varied fasting 0.03% confirmed aspiration Emergent status increased minor events, not aspiration  
Emergency procedural sedation reviews ED adults & children Mostly non‑fasted ~0.005% (1 of ~22k) Aspiration extremely rare despite NPO violation  
Systematic review/meta‑analysis (17 studies) Various elective/emergency settings Reduced fasting 0.5% vs 0.71% No difference between fasting regimens 
⸻
🧠 Bottom line
• No randomized trials directly compare aspiration rates in emergencies stratified by fasting, but multiple robust observational studies show no significant difference in aspiration risk for emergency non‑fasted patients versus elective fasted patients, when proper airway protection is applied.
• Longitudinal and meta‑analytic data confirm fatal or clinically meaningful aspiration is very rare, and strict fasting beyond guideline recommendations does not appear to reduce aspiration rates.
⸻
If you’d like targeted data — for example, aspiration rates by time‑of‑day or meal timing in emergency settings — these likely aren’t captured in the published literature, but ongoing or future large registries using real‑time gastric ultrasound or detailed fasting logs might enable that. Let me know if you’re interested in designing such a study or exploring institutional data.
As an Emergency Physician for almost 30 years, I have heard both sides of this discussion for decades. We don’t have the liberty of patient fasting and intubate ad nauseum in the most dire of circumstances. I have seen the anesthesiologist cancel surgery generously in the silliest of circumstances for a glass of water. A full meal and such I agree with you only for your control and comfort. But, I must say after hundreds of thousands patient encounters, intubating through the brutane era to the rapid sequence with video. After performing my own real time case study analysis for 30 years, I have never had an aspiration. This includes blood, food, gsw, and any foreign object you can conjure up. I find their position amusing at best. My 2 cents.
A proper study of the risks would thus logically require assessing both the risk from aspiration, AND the risk from canceling or postponing an urgent major surgery.
I’m honored that so many have read our research and commented upon it.
Dr. Nielson-we struggled with the term liberal versus conservative fasting. Admittedly, some of my co-authors didn’t like that terminology, but I insisted on it. I wasn’t sure of a better way to classify comparisons between one group of fasting regiments with others that were quite heterogeneous.
Although the measurements were all different, they represented one approach to fasting, and then another that involved a lesser degree of fasting.
But that wasn’t the major point of the work. The bottom line is that almost every study of fasting used the intermediate outcomes of gastric volume or pH. Neither of these outcomes have been proven to correlate to human disease. Every study cited all, but the most recent ASA guidelines for fasting used either gastric volume or gastric pH as the study end point. Even the most recent ASA guideline, which was only a partial update, only used gastric volume and pH for meta-analytic purposes They did not use other outcomes such as witnessed aspiration in the calculations.
These end points were conjured up from animal studies in the 1970s. What I found surprising in reviewing the literature quite carefully was that almost no one has questioned the use of these intermediate outcomes.
None of the ASA publications discussed the serious limitations the of using untested surrogate outcomes. Nor did they downgrade the evidence scoring sufficiently to account for the fact that untested intermediate outcome was used in the studies. Based on my review, all the studies cited in all of the ASA guidelines should have been rated as very low quality because they relied on intermediate outcomes.
From my perspective – that of someone who managed the EBM portfolio at JAMA for nine years – this is a striking observation. We talked about these kinds of things at the journal all the time. Yet, they are hardly addressed anywhere in the extent literature for fasting before surgery.
From my perspective, there is no evidence to support any particular fasting guideline. That is because of the reliance on these untested intermediate outcomes.
Another observation I found interesting was that no one had done a meta-analysis of witnessed aspiration. Meta-analyses are at their best when they are aggregating studies reporting rare outcomes to help provide sufficient numbers of events to facilitate meaningful conclusions.
When we did that, we found that aspiration is extremely rare. We reviewed the original article that posited a relationship between a full stomach and aspiration and a need for fasting. Aspiration was very rare before they were fasting policies and it remains very rare today despite excessively long fasting times.
Being a surgeon, and a former chief of surgery, I believe that case cancellations because of “inadequate” fasting are not acceptable. I have had to tell many patients that their surgery was delayed or canceled because they ate too soon before their scheduled surgery. I’ve also had to deal with the severe operating room, disruptions and loss of revenue because of these cancellations.
More recently, I was a patient myself and had to sit in the hallway all by myself, waiting for my surgery for about an hour or two because I hadn’t passed that magical eight hour fasting time that was required in the facility where I had surgery.
The attention this article has gotten represents how many people are exposed to the problems caused by fasting policies. I’m hoping our article stimulates serious reconsideration of these fasting policies and a more rational approach to research into how to optimize preoperative management of patients.
In GI there are various protocols for drinking colytely, from finishing the last portion by midnight for a morning procedure, to finishing the last half two hours before. That would make easy accruing a large number of patients with which to make a RCT comparison for conscious sedation.
Great post. And great teaching point in service of reviewing evidence.
As always, the “methods” section is the key. In this case, what exactly did “liberal” mean? Some clear fluids is one thing; steak and egg brekky is likely quite another.
And there’s always a disconnect btw what the physician says (or intends) vs what the patient hears (or wants to hear).
Also, cancelling a procedure means a wasted spot which means another patient on the waitlist is having to wait that much longer.
I’m not a surgeon but I do procedures that require conscious sedation.
And I’m a believer in KISS principle.
My practical take-away is this: I will instruct pts the same way. I will still say NPO from midnight except meds with sips in AM. I will still advise 6 hrs NPO for procedures later in the day. But I will cancel procedures less often by being less strict with those who fail to adhere to the letter of those instructions, who have had clear fluids closer to procedure time than advised.
Speaking of nuance, perhaps one needs to consider what type of elective surgery, how long the wait to re-schedule, the consequences of delaying the surgery, the medical status of the patient, and the facility within which the procedure is being done. There are a lot of moving parts.
I remove wisdom teeth in my office under IV anesthesia. We are rural and the nearest "hospital" is 10 miles up the road. The next nearest "hospital" is an hour away. If we need a true place of higher level of care, that is a 2 and a half hour drive. Considering that I have almost zero external support and that 99.99% of the time what I do can hardly be considered life saving, I have a very, very low threshold for canceling cases where the patient is not NPO for at least 6 hours. But I also have the luxury of saying to them that we can proceed with local anesthesia only or we can re-schedule in just a week or two when you follow directions.
Considering another scenario, a patient might be scheduled for tumor removal in a large center with oodles of support. Perhaps that patient had their latte 3 hours before their surgery. Perhaps postponement might make a huge difference in their outcome. That might be a circumstance where the anesthesiologist would recommend proceeding. There has to be nuance in the decision to proceed or not.
Interesting. As a former lactation consultant, I take an interest because we lobbied hard and mostly successfully for counting breastmilk as a clear liquid for purposes of fasting guidelines. Given the comment below and the article, clearly there is high variability not just in what patients' stomachs actually contain just prior to surgery, but in whether the anesthesiologist even knows what's in there, and also in surgeons'/ anesthesiologists' tolerance for reported or discovered intake. Those intersecting major areas of uncertainty would all have to be sternly controlled for in a large RTC. I'd rather read (1) a comparison of aspiration events in emergency surgeries where fasting could not take place (subdivided by time of day or some such rough proxy for likelihood of a recent full meal), versus aspiration events in planned surgeries where fasting orders were given; and (2) a stringent analysis of all aspiration events over a period of years.
Here’s what the medical literature shows regarding (1) comparisons of aspiration rates in emergency surgeries (where fasting may not occur) versus elective surgeries with fasting orders, and (2) large‑scale analyses of aspiration events over time.
⸻
1. Emergency surgery without fasting vs elective surgery with fasting
• A 2020 retrospective review at a single center in acute care surgery (appendectomy/laparotomy in 2016–17) compared unfasted patients (<8 h NPO) to those fasted (>8 h). There were zero aspiration events in either group—aspiration risk was statistically equivalent (relative risk ≈ 1.0) .
• In procedural/emergency sedation settings (emergency department sedation in adults and children), systematic reviews found only one aspiration case out of over 22,000 sedations, and non‑fasting status was not linked to increased risk .
• In children, a German multicenter prospective observational study (12,093 cases) found regurgitation in 0.26%, suspected aspiration in 0.08%, confirmed aspiration in 0.03%. Emergent status increased risk (odds ratio ≈ 2.8), but overall confirmed pulmonary aspiration remained rare and resolved quickly .
Summary: High‑quality observational data suggest that non‑fasted patients undergoing emergency or urgent surgery do not experience a statistically higher risk of aspiration than fasted patients in elective surgery—particularly when modern airway protections like rapid sequence induction are used.
⸻
2. Rigorous analyses of aspiration events over time
• A new meta‑analysis published June 2025 by UCLA examined 17 studies (801 in experimental arms, 990 controls) on fasting duration and clinical aspiration. Aspiration was rare—0.50% vs 0.71%—and no significant difference was found between fasting regimens and controls. Trial sequential analysis suggested further studies are unlikely to change that conclusion .
• Older reviews estimate aspiration incidence in elective general anesthesia at roughly 1.4–6 per 100,000 cases (~0.0014–0.006%) .
• Large ED procedural sedation series (JAMA Pediatrics, >2,600 pediatric cases) reported no clinically apparent aspiration events and found no association between fasting times and adverse events like vomiting or other sedation complications .
⸻
🔍 Answer to your two questions
(1) Comparison of aspiration events: Emergency (non‑fasted) vs elective (fasted)
• While we lack large randomized controlled trials directly assigning fasting vs non‑fasting in emergency surgery, several retrospective and observational studies show no increased aspiration in non‑fasted emergency cases when proper anesthesia protocols (rapid sequence induction) are used.
• Emergency pediatric cases do show a modest increase in regurgitation or suspected aspiration, but confirmed aspiration remains rare and usually benign  .
• Time‑of‑day stratification (e.g. likelihood of having eaten recently) is rarely detailed in published datasets; up‑to‑date emergency vs elective comparisons typically focus on emergent status rather than specific timing or meal presence.
(2) Period‑wide analyses of aspiration events
• Meta‑analysis covering multiple studies up to 2023 found aspiration is very rare (~0.5–0.7%) in clinical settings and not impacted by different fasting regimens, based on 17 RCTs/observational arms .
• Broader epidemiological data across decades consistently puts aspiration rates in elective GA around 1–6 per 100,000 cases (~0.001–0.006%)  .
⸻
🚑 Additional context: Why emergency cases may not show elevated risk
• In emergency surgery, anesthesiologists typically use rapid sequence induction (RSI) with a cuffed endotracheal tube—this protects the airway even when the stomach is full  .
• The risk of delaying a life‑saving surgery far outweighs the low risk of aspiration. Safety is optimized by airway techniques, vigilant suctioning, and early intervention if aspiration occurs —not by waiting for gastric emptying  .
⸻
✅ Limitations
• No large RCTs directly randomizing fasting vs non‑fasting in emergency surgery, for obvious ethical reasons.
• Time‑of‑day or recent full‑meal versus fasted subgroups are rarely captured in uniform data.
• Most existing evidence comes from observational retrospective or prospective cohort studies, not experimental trials—but these are well‑controlled and consistent across settings.
⸻
###📊 Summary Table
Study Type Setting Fasting Status Aspiration Incidence Key Findings
Retrospective cohort (n=100) Acute-care appendectomy/laparotomy < 8 h vs > 8 h 0 events in either group No increased risk in non‑fasted emergencies 
Pediatric prospective observational (n≈12k) Mixed planned/emergency Varied fasting 0.03% confirmed aspiration Emergent status increased minor events, not aspiration  
Emergency procedural sedation reviews ED adults & children Mostly non‑fasted ~0.005% (1 of ~22k) Aspiration extremely rare despite NPO violation  
Systematic review/meta‑analysis (17 studies) Various elective/emergency settings Reduced fasting 0.5% vs 0.71% No difference between fasting regimens 
⸻
🧠 Bottom line
• No randomized trials directly compare aspiration rates in emergencies stratified by fasting, but multiple robust observational studies show no significant difference in aspiration risk for emergency non‑fasted patients versus elective fasted patients, when proper airway protection is applied.
• Longitudinal and meta‑analytic data confirm fatal or clinically meaningful aspiration is very rare, and strict fasting beyond guideline recommendations does not appear to reduce aspiration rates.
⸻
If you’d like targeted data — for example, aspiration rates by time‑of‑day or meal timing in emergency settings — these likely aren’t captured in the published literature, but ongoing or future large registries using real‑time gastric ultrasound or detailed fasting logs might enable that. Let me know if you’re interested in designing such a study or exploring institutional data.
Thank you! That was very informative.
Typos happen. Add an "s" to "Apiration."
As an Emergency Physician for almost 30 years, I have heard both sides of this discussion for decades. We don’t have the liberty of patient fasting and intubate ad nauseum in the most dire of circumstances. I have seen the anesthesiologist cancel surgery generously in the silliest of circumstances for a glass of water. A full meal and such I agree with you only for your control and comfort. But, I must say after hundreds of thousands patient encounters, intubating through the brutane era to the rapid sequence with video. After performing my own real time case study analysis for 30 years, I have never had an aspiration. This includes blood, food, gsw, and any foreign object you can conjure up. I find their position amusing at best. My 2 cents.
A proper study of the risks would thus logically require assessing both the risk from aspiration, AND the risk from canceling or postponing an urgent major surgery.
Thank you for your writing. Appreciate all contributions to sensible medicine. Love the post.