A Navy Veteran’s Death during a Routine Endoscopy: a $950,000 Federal Tort Claims Act Settlement Against the Atlanta VA
A Navy veteran in his late sixties was admitted to the Joseph Maxwell Cleland Atlanta VA Medical Center for abdominal pain and scheduled for an elective upper endoscopy the next morning. His pre-procedure note read like a warning label: obstructive sleep apnea, a body mass index around 38, congestive heart failure, atrial fibrillation, ASA physical status Class 3, and an airway the anesthesia screening described as possibly difficult. He was booked for moderate sedation. No anesthesiologist was scheduled to be in the room.
What the record shows about the morning he died
The sedation flowsheet is short. Over the course of the procedure, he received a total of 4 mg of midazolam, 100 mcg of fentanyl, and diphenhydramine. Those are not exotic doses for a low-risk patient. For a patient with OSA, a BMI of 38, and a documented difficult airway, they are the doses that require someone in the room whose full job is the airway.
At 11:22 a.m. his oxygen saturation was recorded at 68 percent.
He became cyanotic and unresponsive. Flumazenil, the benzodiazepine reversal agent that the facility’s own policy required be immediately available and administered when a patient decompensates on midazolam, was never given. Resuscitation continued for hours. He was pronounced dead at 2:35 p.m.
None of that came from an expert. It came from the record. The pattern in our sedation and rescue cases is the same one we saw in the $9.9 million radiology settlement out of Atlanta and the $3.25 million necrotizing fasciitis settlement: the flowsheet tells the story before anyone opines on it. What an expert does later is name the standards the record shows were missed. What the family needs first is somebody who can read a sedation record the way an anesthesiologist reads one.
How a routine endoscopy actually kills a patient
Endoscopy sedation kills through a small number of specific mechanisms, and the record almost always names which one.
Why aspiration is the most-cited mechanism in sedation deaths
Aspiration is the mechanism the case reports come back to. Incidence is roughly 1 in 3,000 to 10,000 procedures per World Journal of Gastrointestinal Endoscopy data, and the risk factors are the ones most patient-education pages skim past: obesity, gastroesophageal reflux disease, and diabetes. The debate that matters at the bedside is not moderate sedation versus deep sedation. It is monitored anesthesia care without a secured airway versus general anesthesia with a tracheal tube. Recent anesthesiology literature has associated MAC in higher-risk patients with a greater incidence of intraprocedural hypoxemia, hypercapnia, and apnea than GA with intubation. Gastric ultrasound before sedation and deliberate positioning are the specific preventions.
Why unmonitored respiratory depression is the mechanism no family sees coming
Midazolam and fentanyl each suppress respiration in a dose-dependent way. Given together, they suppress it more than either does alone. That is not a rare complication. It is the expected pharmacology, which is why continuous pulse oximetry and capnography are the monitoring standards, and why flumazenil for benzodiazepine reversal and naloxone for opioid reversal exist and are supposed to be in the room. A saturation reading of 68 percent is not a data point on a chart. It is a rescue emergency. In the Atlanta VA case, the drug that facility policy required be given in exactly that scenario was never administered. This is the mechanism that killed the veteran, and it is the mechanism families are least prepared to hear about, because it does not look dramatic on the way in.
The other two mechanisms deserve a sentence. Delayed recognition of a perforation or a bleed is the failure mode that showed up in the $3.25 million necrotizing fasciitis case in a different clinical setting: aggressive deterioration that the team in the room did not name until it was too late to change the outcome.
Hemodynamic collapse in an ASA 3 patient with heart failure and atrial fibrillation is the fourth. A benzodiazepine and opioid bolus that a healthy patient tolerates can drop the blood pressure of a comorbid patient below the point at which the coronary arteries and brain are perfused. When that window is missed, the pattern echoes the $9.9 million radiology settlement out of Atlanta, where a rescue window closed before anyone opened it.
Three failures a board-certified anesthesiology expert identified
A breach of a named, published standard is not a judgment call. Standards exist for exactly this reason, so that whether the care was reasonable does not depend on how the case is told afterward. In this case, a board-certified anesthesiology expert identified three.
Failure to obtain an anesthesiology consult. An ASA Class 3 patient with obstructive sleep apnea, morbid obesity, and a possibly difficult airway is not a moderate-sedation candidate under VA Directive 1073 (Anesthesia Services) or VA Directive 1157 (Moderate Sedation by Non-Anesthesia Providers), and the Society for Anesthesia and Sleep Medicine guidelines say the same. Those documents describe exactly this profile as one that requires anesthesiology involvement in the sedation plan before the procedure begins. The consult did not happen.
Failure to monitor for and prepare for an airway emergency. The American Society of Anesthesiologists 2018 practice guidelines for moderate procedural sedation and the American Society for Gastrointestinal Endoscopy 2018 guidelines both require continuous oxygenation and ventilation monitoring for high-risk sedation, and both require that airway rescue equipment and personnel with the training to use it be immediately available. A saturation of 68 percent is what those standards are designed to catch minutes earlier than this record caught it.
Failure to administer flumazenil. VA policy required flumazenil be immediately available and given to reverse the benzodiazepine when a patient decompensated on midazolam. The Medicare Local Coverage Determination for monitored anesthesia care describes the same expectation for high-risk sedation. It was not administered. There is a live academic debate about whether anesthesia involvement is automatically safer for a routine, low-risk endoscopy. This patient was not routine and not low-risk. The published standards written for his profile all pointed the same direction.
How often people actually die during endoscopy sedation
The honest answer is that the mortality rate depends on which patient is being counted.
The most-cited U.S. figure comes from a 2017 analysis in Gastrointestinal Endoscopy by Vargo and colleagues, which reviewed 1,388,235 endoscopy patients and identified 10 deaths, roughly 1 in 138,824. Six of those were classified as potentially sedation-associated. German data commonly cited alongside it puts sedation-associated mortality in ASA I to III patients at roughly 1 in 100,000 to 1 in 150,000.
The Leslie et al. cohort in the British Journal of Anaesthesia tells a different story because it counts a different patient. Across 2,132 sedation cases at nine hospitals, 30-day mortality was 1.2 percent overall, 0.2 percent for elective procedures and 6.0 percent for emergencies. Only one death occurred during the procedure itself. Median time to death was 11 days. More than one in twenty emergency patients in that dataset had a procedure in the last month of their lives.
The numbers differ because they measure different things. Vargo counts intraprocedural and immediately sedation-attributable deaths in a large mixed U.S. cohort. Leslie counts 30-day all-cause mortality in a sedation cohort skewed by emergency indications. Set against the U.S. denominator of roughly 6.1 million upper endoscopies performed annually, both are low absolute risks for a healthy patient scheduled electively.
The Navy veteran did not sit on the low end of any of those denominators. He carried four independent risk factors that push a patient toward the higher end of every published dataset. The number a family is quoted after a death, one in one hundred thousand, is almost always the elective, all-comers figure. For a patient like this one, it was never the right number to begin with.
Who should not be sedated the way most patients are
A high-risk patient sedated as if he were low-risk is not a rare bad outcome. It is a recognizable pattern. The published risk stratification is specific, and Leslie’s data quantifies it more cleanly than the patient-education pages that sit above it in the search results.
- Age. Odds ratios for unplanned intraoperative events: 1.97 for ages 40 to 59, 2.39 for 60 to 79, and 3.39 for 80 and above. Cox hazard ratio for 30-day death in the 80-plus group: 3.54.
- Obstructive sleep apnea. Upper airway collapse under sedation is the anticipated problem, not the unexpected one. It is the reason the SASM guidelines specifically address OSA sedation planning.
- Congestive heart failure and atrial fibrillation. Both change the hemodynamic tolerance for benzodiazepine and opioid boluses. Rate and pressure drops that a healthy heart absorbs, a failing heart does not.
- Body mass index at or above 35. Positioning, mask seal, and airway visualization all change. Aspiration risk climbs.
- ASA Class 3 or higher. By definition, severe systemic disease that is a substantive limitation on activity.
- “Possibly difficult airway” on a pre-procedure note. That phrase exists to trigger anesthesiology involvement in the sedation plan. It is not a note-taking convention.
The Navy veteran met four of those criteria before the procedure was scheduled, and his ASA classification captured the fifth. The published guidelines written for his profile called for an anesthesiology consult, an anesthesiology-directed sedation plan, and rescue readiness commensurate with the risk. When four criteria appear on a pre-procedure note and none of them changes the sedation plan, the plan is the problem before the drugs are drawn up.
Why a case against the VA is not a case against a hospital
Families of veterans are often told, correctly, that you cannot sue the federal government. That is the beginning of the sentence, not the end of it. The Federal Tort Claims Act is the exclusive route by which a private party can bring a claim against the United States for the negligent acts of its employees acting within the scope of employment, and VA physicians and nurses are federal employees.
An FTCA case does not start in court. It starts with an administrative claim submitted on Standard Form 95 to the agency responsible for the employee whose conduct is at issue. In a case against a VA medical center, that is the Department of Veterans Affairs. The SF-95 must state a sum certain and include the facts and injuries claimed. The agency then has six months to admit, deny, or investigate. A lawsuit can be filed only after a formal denial or after six months of agency inaction.
Two other differences matter. There is no jury trial in an FTCA case. Any trial is a bench trial in federal district court. And there are no punitive damages available under the FTCA, even in cases where a state-court plaintiff might otherwise seek them.
Most Georgia medical malpractice firms market state-court experience only. That is a real limitation for a veteran’s family, not a marketing detail. A case that has to be filed on the SF-95 before it can be filed anywhere else runs on procedural rules a state-court practice does not use every week.
How the FTCA deadline differs from Georgia’s two-year statute
Georgia state-court medical malpractice claims run on O.C.G.A. § 9-3-71, which sets a two-year period from the date of the injury and a five-year statute of repose. A complaint must be accompanied at filing by an expert affidavit under O.C.G.A. § 9-11-9.1, and the expert must practice or teach in the same specialty as the defendant provider. Non-economic damages are theoretically subject to the cap in O.C.G.A. § 51-13-1, though that cap has been under sustained constitutional challenge since Turner v. Medical Center of Central Georgia and is not settled law. Wrongful death claims are filed in the order set by O.C.G.A. § 51-4-2: spouse first, then children, then parents.
An FTCA case runs on none of that. The clock is 28 U.S.C. § 2401(b): the SF-95 must be presented to the agency within two years of the injury, and suit must be filed within six months of a formal denial. There is no 9-11-9.1 affidavit, no state cap, and no state hierarchy governing procedure, though state substantive law often supplies the measure of damages. Exceptions may apply in either system and depend on the facts. The single most important sentence for a veteran’s family to read is that the Georgia framework does not govern here. A lawyer who quotes O.C.G.A. § 9-3-71 in a VA case is not the lawyer that family needs.
What a Federal Tort Claims Act wrongful death recovery can provide for a veteran’s family
FTCA damages in a wrongful death case are not a payout. They are what a family would have had if the sedation plan had matched the patient. In a case governed by Georgia substantive law, the recoverable categories generally include medical expenses incurred before death, funeral and burial costs, the full value of the life of the deceased under Georgia’s wrongful death standard, and the loss of care, companionship, and support the surviving family will experience going forward. Punitive damages, which some state courts permit in cases of egregious misconduct, are not available under the FTCA.
When state substantive law is borrowed under the FTCA, the wrongful death filing hierarchy in O.C.G.A. § 51-4-2 still shapes who brings the claim: the surviving spouse, then the children, then the parents. The estate separately pursues the pre-death medical and funeral expenses through an administrator.
The Atlanta VA case resolved for $950,000 before trial. Past results do not guarantee future outcomes, and every FTCA case turns on its own facts, records, and expert review. Our firm handles medical malpractice on a contingency-fee basis, with the specific terms and expense treatment set out in a written fee agreement rather than promised on a webpage. A family considering whether to pursue a claim against a VA facility deserves both the honest recovery categories and the honest procedural rules before deciding.
Questions Georgia families ask about endoscopy sedation deaths
Can you die from an endoscopy?
Yes, though it is uncommon in low-risk patients. Sedation-associated mortality in ASA I to III patients has been estimated at roughly 1 in 100,000 to 1 in 150,000. The risk climbs with age, with sleep apnea, with cardiac disease, and with a documented difficult airway. Death is almost always mechanism-driven, not random, and the mechanism is usually visible in the sedation record.
What is the mortality rate for sedation during endoscopy?
It depends on the patient being counted. Vargo’s 2017 U.S. cohort put overall endoscopy mortality at roughly 1 in 138,824. Leslie’s Australian dataset reported 30-day mortality of 0.2 percent for elective sedation cases and 6.0 percent for emergencies. The difference is not a disagreement. It is what happens when different populations are measured.
How safe is sedation for an upper endoscopy?
Safe for most patients. The published guidelines describe a specific set of profiles that fall outside routine moderate sedation: patients with obstructive sleep apnea, ASA Class 3 or higher, morbid obesity, or a documented difficult airway. For those patients, the ASA 2018 and SASM guidelines call for anesthesiology involvement in the sedation plan. Whether that happened is one of the first questions a family should ask.
Is an upper endoscopy painful?
Under moderate sedation, the patient stays cooperative and may remember parts of the procedure. Discomfort is usually the gag reflex and the throat, not pain. Deep sedation and general anesthesia change that but come with different risks and staffing requirements.
Is endoscopy dangerous for elderly patients or those with heart conditions?
Age is an independent predictor of unplanned intraoperative events, with odds ratios of about 1.97 for ages 40 to 59, 2.39 for 60 to 79, and 3.39 for 80 and above. The Cox hazard ratio for 30-day mortality in the 80-plus group is 3.54. Congestive heart failure and atrial fibrillation reduce the hemodynamic reserve available to absorb a benzodiazepine or opioid bolus, which changes what a safe sedation plan looks like.
What should a family watch for after a loved one comes home from an endoscopy?
Chest pain, fever, black stools, severe abdominal pain, vomiting blood, or difficulty breathing. The Leslie dataset reported a median time to death of 11 days in the sedation deaths it identified, which means the risk window does not close when the patient leaves the procedure room.
How long does a family have to file a claim against a VA hospital?
Under 28 U.S.C. § 2401(b), the Standard Form 95 administrative claim must be presented to the VA within two years of the injury. Suit can be filed only after a formal denial or six months of agency inaction, and within six months of the denial. Exceptions may apply and depend on the facts. This is the deadline that most often costs families a claim they had, which is why it is the earliest thing to speak with a lawyer about.
If your family is asking why a routine endoscopy ended this way
If a loved one died during or after a sedation procedure at a VA facility, the question of whether a claim exists depends on the records, the standards that governed the sedation plan, and an expert review of both. Our firm has handled these cases, including the Atlanta VA case described above, and we work on a contingency-fee basis with the specific terms and expenses set out in a written fee agreement. If you would like help understanding what happened, or you are considering suing the VA for a sedation death in Atlanta, we are here to talk. You are welcome to request a confidential consultation or read more about our anesthesiology malpractice work.
This article is for general informational purposes only and is not legal advice. Every case depends on its own facts, medical records, and expert review. Reading this page does not create an attorney-client relationship.
