From Transnasal Esophagoscopy:

Jamie Koufman, M.D., Director, Voice Institute of New York

On August 28th, 2014, something happened during Joan Rivers’ endoscopy that took her life, and there is now clear evidence that many things went wrong that took Joan Rivers’ life. For sure, she met her demise in the endoscopy suite at Yorkville Endoscopy. The nation wants to know if there are lessons to be learned from her death that might have implications for all of us. There are.

Because I am an expert in both laryngology1and upper endoscopy,2 I feel obliged to raise some fundamental questions about Ms. Rivers’ death and to try and provide some insight, if not some answers. When the media first reported that Joan Rivers had had a throat procedure in a GI (gastrointestinal) endoscopy suite (in an ambulatory surgery center), I was shocked. That made no sense to me. Without knowing the exact details of Joan Rivers’ lethal procedure, I have good reason to believe that: (1) for several reasons, Ms. Rivers should never have been at Yorkville Endoscopy in the first place; (2) she should not have had an upper endoscopy (EGD) or any kind of throat (vocal cord) procedure performed; and (3) when Joan Rivers crashed, the medical team should have been able to resuscitate her.

Why Joan Rivers should never have been at Yorkville: I watched Joan Rivers on television and I personally met Joan. As a laryngologist, it was obvious from the sound of her voice that she had reflux laryngitis (acid reflux) and polypoid corditis. (The latter is a term for swollen, floppy vocal cords, which are commonly seen in older women with reflux and also in long-time cigarette smokers).  Indeed, during a conversation with Joan, I told her, “From the sound of your voice, I think that you have acid reflux and small vocal cord polyps. Why don’t you come see me; I’m a voice and reflux expert?” Smiling, she responded that she loved her throat doctor.

Why did Joan’s beloved ENT doctor consult a gastroenterologist? Why did she need to have an endoscopy at a GI facility? Joan’s low-pitched, raspy voice was fine or maybe her voice was getting huskier; but polypoid corditis should almost never be biopsied unless there is a very strong reason for doing so.

Laryngologists routinely use a technology called videostroboscopy for examination of the larynx (voice box) that also allows visualization of the vocal fold vibrations. If any kind of change or lesion grows on the vocal cords, the strobe examination should help determine if a biopsy is needed. Then, the biopsy should be done by an expert laryngologist using an operating microscope with microscopic precision using special instruments and in a hospital operating room capable of managing any complication. Vocal cords should never be examined or biopsied using GI equipment or in a GI endoscopy suite without absolute control of the airway.

For my patients with reflux laryngitis — also called LPR (laryngopharyngeal reflux), airway reflux, throat reflux, and silent reflux (all terms that I coined) — I never consult a gastroenterologist. The in-office laryngeal examination and specific (throat) reflux testing,not sedated GI endoscopy, is how the diagnosis of reflux laryngitis is made. (GIs do not examine the throat or do such throat-reflux testing.)


While all patients with reflux, including reflux laryngitis, should at some point undergo esophageal examinations to rule out esophageal cancer or its precursor, Barretts esophagus, in my practice sedated GI endoscopy (aka EGD) is almost never the procedure of choice, because there is a less expensive, more accurate, and much safer alterative to EGD!

Almost fifteen years ago, I pioneered transnasal esophagoscopy (TNE), which is in every way superior to EGD. TNE is performed in my office with a completely awake and comfortable patient using an ultra-thin endoscope which is introduced through the nose. We do numb the nose with a topical medicine first, but there is no IV or medication involved. Transnasal esophagoscopy offers huge advantages over EGD.

TNE is profoundly safer than EGD, there have been no deaths or serious complications with TNE since its introduction fifteen years ago. By comparison, I estimate that since 1975, between 25,000 and 50,000 people have died having sedated GI endoscopy (EGD) performed.

Indeed, Joan Rivers was not a one-in-a-million sedated-endoscopy death; the data suggest that the death rate is about 1:7,143. Using that formula, if 18,000,000 sedated endoscopies are performed this year, it would be expected that almost 2,000 people would lose their lives this year as a result of EGD.

EGD is expensive. It is likely that America has expended nearly $1 trillion on sedated GI endoscopy since 1975. With in-office TNE, there is no facility fee, no anesthesia fee, and no lost time from work for the patient. When I have asked gastroenterologists why they do not perform TNE, the (honest) response it that it does not reimburse enough, that is, they make a lot more doing EGD.

TNE is more accurate than EGD. In a series of patients who consulted me after having received a diagnosis of Barretts esophagus, we found that fewer than one-third actually had it. Apparently the biopsies during EGD are often performed “too low.”

Gastroenterologists, in defending EGD, insist that their patients demand to be put to sleep. I don’t believe this for a minute. In my practice, 99% of my patients prefer TNE to EGD. Finally, EGD is wrong for even more reasons. The duodenum is rarely abnormal so it could be EG; and in addition, GIs almost routinely biopsy the stomach for no reason — virtually every reflux patient comes to me having had an EGD with “mild gastritis,” and already on a purple pill (a proton pump inhibitor) — the EGD, the gastritis, and the treatment are all wrong.

The above concerns lead me to believe that had Joan Rivers seen me, she would still be alive. Below for your perusal are some facts about sedated GI endoscopy, about the ambulatory surgery centers (ASCs) (aka “surgicenters”) where Joan Rivers met her demise, and about the massively expensive mismanagement of reflux. Draw your own conclusions.


Marketing of the modern flexible endoscope for colonoscopy and upper endoscopy — EGD (esophagogastroduodenoscopy) — by the Olympus Corporation began in about 1975.

In 1975, there were about 700 gastroenterologists in the United Sates; now, there are over 13,000. This represents an almost twenty-fold increase in GIs, much more growth than any other medical specialty.

This year, GI doctors’ mean income is $442,000, which ranks them third (of all medical specialties), after orthopedic surgeons and interventional cardiologists.

In 1975, there were no surgicenters like Yorkville Endoscopy; now, there are over 5,000; and 90% have physician ownership creating massive conflicts of interest.

The sedated GI endoscopy death rates have climbed since the 1970s, from about 5:100,000 to about 14:100,000 today. Increasing endoscopy death rates and GI physician income appear to positively correlate. See attached links: McLernon & Donnan 2007, deRoux & Sgarlato 2012.

GI doctors do not video-record sedated endoscopy examinations; although it would be very easy to do so. Indeed, GI endoscopy is one of the last “black box” specialties.

What goes on during an endoscopy is only known to the doctor, because the examination is not video-recorded. By comparison, I have video-documentation of every laryngeal and endoscopic examination that I have performed since 1978. GI doctors almost never record examinations, but had Dr. Cohen videorecorded Joan Rivers’ procedure; we would probably know what happened to her. So, why don’t GIs record examinations? The apparent answer is so that there can be no scrutiny; GI endoscopy is a “black box.” When I recently asked one prominent GI doctor why GIs didn’t record examinations, he replied, “No one I know does that. We could. But no one would want to be second-guessed on a review.” This implicitly suggests that there may be some shady stuff going on.

Since 1975, after almost 400 million sedated endoscopies and as many as 55,000 endoscopy-related deaths (398,000,000 procedures x 0.00014 (reported death rate) = 55,720 deaths). (BTW, I recognize that not all of those procedures were EGDs, so perhaps the actual EGD number of deaths is closer to 25,000 (1975 to the present).

On the GI watch, reflux has increased 400% and cancer of the esophagus more than 500% to become the fastest growing cancer in the United States. GI has NOT gotten it done and the GI model  of reflux, especially how they diagnose and treat it, is wrong. The GI heartburn-is-reflux-and-reflux-is-heartburn model needs to be replaced with a more accurate, effective, and less self-serving reflux paradigm. See Dropping Acid: The Reflux Diet Cookbook & Cure, for example.


Videorecording of all endoscopic procedures should be mandated by law. (CMS could make and enforce this requirement.) There should be routine outside scrutiny and monitoring of medical procedures.

Establish a central and transparent database for all medical M&M (mortality and morbidity). Accurate reporting would be under penalty of law with failure to accurately report potentially being punishable by imprisonment; this would require federal legislation.

Unaccountability is one of the biggest obstacles to cost containment and innovation in U.S. healthcare. The current practice of each medical specialty setting its own guidelines without outside scrutiny proliferates “black box” medical procedures such as EGD. This can be likened to allowing the fox to make the rules in the hen house. Accurate data is again the key to accurate data analysis. Specialties cannot be allowed to represent themselves in the performance  evaluation of their respective specialty.The conflicts of interest are too great. The fox cannot be allowed to make the rules for the hen house.

Other conflict of interest? Physicians must be made declare publicly all conflicts of interest, including ownershp of ASCs, relationships with pharmaceutical companies, and ownership of any publicly traded or privately owned company related in any way to medicine. (BTW, when a physician speaks at an academic meeting, s/he is supposed to do this.)

Sedated GI endoscopy (EGD) should be replaced by unsedated, transnasal esophagoscopy (TNE).


If what happened to Joan Rivers happened in an up-scale Manhattan endoscopy center at the hand of a well-known GI doctor, what are the chances that this might be happening elsewhere? Unfortunately, Joan Rivers’ death was not a one-in-a-million catastrophe. Last year, 18 million sedated endoscopies were performed. Since the reported endoscopy-related death rate is 0.00014, we would expect that there were 2,520 sedated-endoscopy (mostly unreported) deaths. Transnasal esophagoscopy (TNE) is a safer and less expensive technology that should replace EGD (sedated GI upper endoscopy) in an accelerated fashion.

Jamie Koufman, M.D. [New York]

1 Laryngologist: A physician, usually an otolaryngologist (ENT doctor), who specializes in problems of the voice and throat, including LPR, airway reflux. I am director of the Voice Institute of New York and I have been a pioneer and opinion-leader in laryngology andendoscopy for thirty-five years. I am the current president of the New York Laryngological Society and a past-president of the American Broncho-Esophagological Association.

2 Upper endoscopy: Examination of the esophagus and stomach, sometimes the upper bowel as well. Endoscopy refers to examination performed by looking with a special instrument, an endoscope, from the inside. I pioneered TNE (transnasal esophagoscopy). GI doctors’ form of sedated endoscopy is EGD (esophagogastroduodenoscopy).

Read more from From Transnasal Esophagoscopy….