Volume of procedures and outcome of treatment

I have agreed with Bosset et al that the operative mortality of 10%
for oesophagectomy reported by the Medical Research Council (MRC)
Oesophageal Cancer Working Party is unacceptably high. I have also
observed that the operative mortality of 12.7% reported in the General
Infirmary in Leeds is equally unacceptable (1). The operative mortality
in Bosset et al’s hands (3.2%) and Orringer’s (4%) and Skinner’s (3.9%)
latest reports of their data was very much lower (2). These two surgeons
are amongst the best known oesophageal surgeons in the country. Upon
further review of the literature I find I overlooked that Skinner had an
operative mortality of 11% when he reported the results of his first 100
cases of en bloc oesophagectomy , an equally unacceptable figure (3). It
is of concern that even if none of the additional 28 patients he reported
in his series had not died his operative mortality should have been only
as low as 8.5% and not as low as 3.9%. How then did he come up with the
sanitised figure of 3.9% especially as Altorki is a co-author on both
publications? Might these be the cases from New York Hospital performed by
the third co-author, Giradi, in his latest report? If so he must have
discarded all or some of the patients he operated upon when he was
chairman of surgery at the University of Chicago.

Skinner, a product of the Halstead surgical tradition at The Johns
Hopkins, has published 220 papers in scientific journals, many of them on
oesophageal diseases. In the case of oesophagectomy for cancer his
contention is that mediastinal lymph node removal improves outcome, a
debatable point especially as his operative mortality is so unacceptably
high (4). He is nevertheless one of the most prominent surgeons in the US
surgery and has been especially active in the surgical and cardiothoracic
surgical societies (4,5). As president of New York Hospital he has been
instrumental in their establishing a Paris affiliation, apparently with
the intention of extending US healthcare into Europe. It is not surprising
that he has been actively involved in healthcare reform in the US,
seemingly playing an important part in establishing the political
correctness of surgical practice (6) . He and his peers who have both
general surgery and subspecalty boards are firm believers in the need for
board certification to be granted privileges in a private hospital and for
subspecialty boards to be granted privileges in subspecialty surgery,
certainly in the US. Recertification of these privileges is another issue
with which he is particularly familiar (7). That position, if carried to
its logical conclusion and includes in addition to cardiothoracic surgery
colo-rectal surgery, paediatric surgery, orthopaedic surgery and other
established or candidate subspecialties, such as vascular surgery.
Vascular surgery, previously firmly within general surgery, is sitting on
the fence on this issue but seems likely to jump into bed with the sub-
specialists.

I have had some experience with the political implications these
political dealings. As chief of the surgical services at the Brooklyn VA
for a brief period I inherited a string of misadventures. The most
striking concerned an army veteran in his early sixties who had had weeks
before I arrived a curative resection for premalignant changes in a
Barrett’s oesophagus. The operation had been performed by a chief
resident in the last year of his general surgical training assisted by a
thoracic surgeon in private practice who was one a several covering the
service on a rotation. The operation had been a disaster. The oesophagus
has been replaced with a colonic interposition that had been made far too
long. The top end had been anastomosed in end-to-side fashion with the
formation of what was in effect a large pharyngeal pouch. When the
patient was not able to swallow they investigated him supposedly with
upper endoscopy performed by the gastroenterologists. I was informed by
the chief resident that they had found a “cesspool” in the stomach
indicative of gastric stasis. A barium swallow performed later revealed a
long ischaemic stricture extending throughout the upper two thirds of the
interposed colon.

I asked the chief resident why they had not used the stomach instead
of the colon. “Because his stomach was irradiated for a lymphoma” he
replied, a statement which I was never able to confirm. He added that the
pyloroplasty had been performed in two layers, an operation abandoned
decades ago because of the stasis is caused. I then asked the surgical
resident who had assisted him and why he had made a two layered
pyloroplasty. “Oh,” said the thoracic surgeon “ I always use a stapler” ,
not in my book a good way of making a pyloroplasty but a single layered
one. It transpired that the chairman of surgery at the affiliated
university from which the residents came had decreed that chief residents
were to be given responsible for the care of all general surgical
operations including oesophageal resections. The thoracic surgeons, who
believed the oesophagus was their territory, responded by working to rule
and doing exactly as the resident asked even when they knew it to be wrong
and appeared even to have attempted to mislead them into making obvious
mistakes. Their intention appears to have been to demonstrate to the
chairman of the surgical training program that chief residents were not
qualified to perform oesophageal surgery even in their final months of
their training and needed to complete a thoracic fellowship before being
qualified to do so.

In attempting to resolve the problem for this poor man, nurses
supportive of the thoracic surgeons, compounded his problems by also
working to rule and neglecting basic nursing care including getting him
out of bed and responding to his repeated requests for assistance unless
there was a doctors’ order written for it. On one occasion that I had
written a very specific order on fluid management then nurses appeared to
have made deliberate errors in his fluid input-output chart to make the
order impossible to follow. The harder I rode heard to insure that proper
care was provided the more disruptive the nursing staff became ultimately
leaving for the medical wards. In the two oesophageal operations I
performed transhiatally with the residents, contrary to Skinner’s views,
and followed very closely myself there were even more blantant efforts
seemingly by nurses to interfere with management.

I ordered an angiogram to look at the blood supply of his stomach.
This showed that the gastroepiploic vessels had been divided thus
eliminating the stomach as a possible replacement for the ischaemic
oesophagus. As he needed a free graft of small intestine to re-establish
gastrointestinal continuity I approached the Sloane Kettering trained
“cancer surgeon” who had evidently been trained in microvascular
techniques for head and neck surgery. He said the VA did not have the
necessary instruments and expressed the opinion that the patient should be
referred elsewhere.

I decided to take the radiographs to Skinner, who worked in a
compewting hospital but was the most appropriate person in the community,
and see if he would assume responsibility for the resolution of his
problems. By this time it had become clear that management of patients in
the VA which was already very bad, not just in the Brooklyn VA but in all
Vas studied in a national audit, had become so dangerously compromised
that I contemplated closing the operating rooms. Having presented the
case to Skinner and a plastic surgeon of his he did not offer to assume
responsibility. Instead he declared that his plastic surgeon would be
happy to come to the VA to construct a free graft after we had formed a
tunnel under the presternal skin with a skin expander. We agreed that
revision of the pylorus was necessary. In short Skinner had made it clear
that he believed that oesophageal cancer should be taken care of thoracic
surgeons, those who were double boarded, and that if general surgeons
wanted to operate on the oesophagus they had to take care of their own
problems. That did not phase me, but my readiness to assume responsibility
appears to have phased them.

I decided to revise the pyloroplasty first converting it into a
Finney, which I did without complication, before tackling the oesophagus.
I did not form a tunnel under the skin as suggested for upon reflection I
thought it was not only unnecessary but an undesirable permanent solution
in a patient who had been cured of his malignant disease and should have
many years left to live. I thought it best to remove the ischaemic colon
for it was likely to create problems in the future, probably through the
left chest and neck, and place the small bowel free graft in the chest. It
was my intention to monitor the viability of the free graft with
tonometers as I had successfully done with another free graft I had asked
the plastic surgeons to construct for me on another patient. This would
have given me the ability to reoperate the moment the blood supply to the
graft became compromised were that to happen in the postoperative course.
I never got the chance for having advised the chief of staff to close the
operating rooms I was asked to resign. When I refused because they had not
addressed my concerns I had expressed to the hospital director about the
deliberate interference of patient care I was relieved of my duties,
brought before a tribunal and fired for “upsetting too many people”. In
short it appears to have been decided that it was not politically correct
for general surgeons to be doing oesophageal surgery. At least the
hospital director was good enough to write me a letter after my dismissal
thanking me for what I had done for her patients.

Skinner remains actively involved in surgical care, evidently for
political reasons (3). His case is that nodal resection, which requires a
formal thoracotomy, improves life survival relative to transhiatal or
blunt oesophagectomy which he believes does not. Orringer, who appears to
have changed the name of the noperation he does from blunt to transhatal
oesophagectomy, appears also to have changed the name for political
reasons. Indeed he makes the copint that he does not do it bluntly anymore
but does the dissection under direct vision. I, on the other hand,
definitely do it bluntly having on one occasion pulled the oesophagus out
with a varicose veins stripper trying a technique that had been reported
in the literature. I presented the case at grand rounds which is attended
by the thoracic surgeons. Perhaps the point has been made is that doing it
bluntly is not an adequate cancer operation, a point Skinner has held for
many years (8). My response to that having had no operative deaths and few
complications is “first do no harm”.

In reviewing the evolution of Orringer’s “transhiatal oesophagectomy
it is evident that his early experience was traumatic to say the least (9-
16). Certainly he encounted a multitude of complications that I never
enountered in my modest series which is more comparable to his earlier
reports in numbers. The complciations that he has repeatedly seen which I
have never seen include entry into the pleural cavity, recurrent nerve
damage, uncontrollable mediastinal heaemorrhage, and mediastinal and
subphrenic abscesses. It is also evident from his review that he gives the
appearance in his latest reports that these complications were uncommon
and unconsequential which they most certainly were not. Indeed his latest
reports are reminiscent of Skinner’s apparent sanitisation of his
operative mortality.

If Orringer and Skinner are the best cardiothoracic surgeons have to
offer I dread to think what kind of results cardiac surgeons who do the
occasional oesophagectomy might be getting especially as the majority of
the operation is done in the abdomen and neck not the chest, regions in
which they have little operative experience. The only conclusion worth
drawing is that subspecialisation has not been in the best interests of
patient care. There are many ways of skinning a cat and mandating one way
or another for political reasons is counterproductive and even
destructive.

1. Fiddian-Green RG. Training is more important than volume
bmj.com/cgi/eletters/325/7368/787#26226, 12 Oct 2002

2. Altorki NK, Skinner DB. En bloc esophagectomy: the first 100
patients.
Hepatogastroenterology. 1990 Aug;37(4):360-3

3. Altorki NK, Girardi L, Skinner DB. En bloc esophagectomy improves
survival for stage III esophageal cancer. J Thorac Cardiovasc Surg. 1997
Dec;114(6):948-55

4. Skinner DB. Presidential address: Society of University Surgeons.
Recruitment and retention of academic surgeons.
Surgery. 1979 Jul;86(1):1-12.
5. Skinner DB. Shaping the revolution: thoracic surgeons and something
more. Presidential address.
J Thorac Cardiovasc Surg. 1997 Nov;114(5):699-706.

6. Skinner DB.Implications of United States healthcare reform for
European cardiothoracic surgery.
Eur J Cardiothorac Surg. 1997 Apr;11(4):599-603.
7. Skinner DB. Recertification: is it worthwhile?
Bull Am Coll Surg. 1981 Nov;66(11):9-11.

8. Skinner DB. Invited letter concerning: how extensive should lymph
node dissection be for cancer of the thoracic esophagus? J Thorac
Cardiovasc Surg. 1994 Apr;107(4):1154-5.

9. Orringer MB, Marshall B, Iannettoni MD. Transhiatal esophagectomy
for treatment of benign and malignant esophageal disease. World J Surg.
2001 Feb;25(2):196-203. Review
10. Orringer MB, Marshall B, Stirling MC. Transhiatal esophagectomy for
benign and malignant disease.
J Thorac Cardiovasc Surg. 1993 Feb;105(2):265-76; discussion 276-7.
11. Orringer MB. Transhiatal esophagectomy without thoracotomy for
carcinoma of the thoracic esophagus. Ann Surg. 1984 Sep;200(3):282-8.
12. Orringer MB. Substernal gastric bypass of the excluded esophagus--
results of an ill-advised operation. Surgery. 1984 Sep;96(3):467-70.
13. Orringer MB. Technical aids in performing transhiatal esophagectomy
without thoracotomy.
Ann Thorac Surg. 1984 Aug;38(2):128-32.
14. Orringer MB, Sloan H. Substernal gastric bypass of the excluded
thoracic esophagus for palliation of esophageal carcinoma. J Thorac
Cardiovasc Surg. 1975 Nov;70(5):836-51.

15. Orringer MB, Sloan H. Esophagectomy without thoracotomy.
J Thorac Cardiovasc Surg. 1978 Nov;76(5):643-54.
16. Orringer MB, Skinner DB. Unusual presentations of primary and
secondary esophageal malignancies.
Ann Thorac Surg. 1971 Apr;11(4):305-14.

It is not surprising that there is a degree of correlation between
volume and operative mortality for as I recall from reading his book,
Social Transformation of American Medicine, Starr observed that 300 beds
was the minimum number of hospital beds required to cover all services
effectively. The correlation between volume and operative mortality for
colon cancer in the State of Maryland, for example, almost certainly
included community hospitals many of which have fewer than 300 beds (1).
If one looks solely at those hospitals with a critical mass, those with
more than 300 beds, there may be no meaningful correlation between a
surgeon’s caseload and operative mortality. Consider oesophageal surgery.

At the General Infirmary at Leeds, the largest hospital in the UK,
the outcome for 204 patients having oesophageal surgery was 12.7% (2). The
operative mortality for oesophagectomy reported by the Medical Research
Council (MRC) Oesophageal Cancer Working Party was 10% (3). Bosset et al,
who report an operative mortality of 3.2%, consider 10% unacceptably high
. Bosset at al’s operative mortality compares favourably with that
reported by Orringer and Skinner in the US (4,5). Orringer et al have
reported the largest series of transhiatal oesophagectomies in the world,
1085 cases. Their operative mortality is 4%. Skinner, reputably one or the
top two or three oesophageal surgeons in the US, report an operative
mortality of 3.9% in his group’s experiences with 128 en bloc
oesophagectomies with removal of the mediastinal nodes. Mariette et al in
France report an operative mortality of 4.7% for oesophagogastrectomy
performed for 126 patients with adenocarcinoma of the cardia (6). Thus the
operative mortality for oesophagectomies performed in credible UK
institutions of appropriate size is some two to four times higher than
that in centres of excellence in the US and France, unacceptably high.

The principle cause of adverse outcome for all major surgical
operations studied appears to by an inadequacy of gut mucosal
mitochondrial oxidative phosphorylation, identified from the presence of a
gastric intramucosal acidosis, and its putative consequences (7). The
volume of blood transfused is positively associated with an adverse
outcome. This may be because red blood cell transfusions have an equivocal
effect on the adequacy of mitochondrial oxidative phosphorylation
oxygenation and may be more likely to induce or compound the severity of
an inadequacy of oxidative phosphorylation especially if the blood is more
than fifteen days old (8). It may do so by releasing cytokines which
uncouple oxidative phosphorylation. An impairment in oxygen uptake and
utilization, which uncoupling would cause, is certainly the most
difficult of all physiological abnormalities to treat in the critically
ill.

The degree and duration of shock, defined as a gastric intramucosal
acidosis, and the volume of blood transfused must reflect to some degree
the clinical skills of the anaesthetist and technical skills of the
operating surgeon. I suspect that an unacceptable degree and/or duration
of shock and/or an unacceptably large volume of blood transfused might be
the principle causes of the unacceptably high operative mortality for
oesophagectomies in the UK. The implication is that surgeons performing
oesphagectomies in the UK are poorly trained.

Bosset et further observe that an operative mortality as low as 1%
has been achieved in other centres (3). This compares with my more modest
experience primarily working in the same centre as Orringer. If one of my
patients died I guess that my operative mortality might have been between
1% and 3% depending on whether oesophagogastrectomies were included or
not. To the best of my knowledge, however, none of my patients died
within 30 days of either an oesophagectomy or oesophagogastrectomy. If so
my operative mortality was 0% and a Chi squared or Fisher’s exact test is
likely to have shown a statistically significant difference in operative
mortalities in my and Orringer’s hands. (9). Why might my operative
mortality have been lower than Orringer’s?

Our patients came from the same pool and were cared for by the same
nurses, anesthesiologists, and residents and we had the same subspecialty
support. If anything Orringer’s residents were better trained for most of
the thoracic fellows had completed their general surgery training in our
program. Some 5% of Orringer’s patients had carcinomas in the upper third
of the oesophagus whereas none of mine did but location of tumour has not
shown to be an independent predictor of operative mortality. His series
might have included tumours that were more advanced than mine but that too
has not been shown to be an independent predictor of operative mortality.
Carcinomas in the lower third were the predominant tumours in both our
hands, and the transhiatal oesophagetomy the preferred operation.
Intrathoracic anastomoses and colonic interpositions were uncommon
operations in both series. The average blood loss in Orringer’s series was
689mls. My loss must have been very similar blood transfusion having been
rarely being indicated. Uncontrolled haemorrhage, chylous leaks, and
tracheal injuries were uncommon complications in Orringer’s large series.
I do not recall ever having had such complications but the numbers are too
small for any meaningful statistical comparison.

There were two differences in our practices that might account for
the putative difference in operative mortalities. Whereas Orringer
routinely inserted feeding jejunostomies I never did. I believe the
practice of placing jejunostomy feeding tubes in all patients may be a
practice that Orringer and possibly even Skinner acquired at Johns
Hopkins. (Halstead is the father of the Johns Hopkins surgical program). I
was not trained to place jejunostomy feeding tubes. A review of
jejunostomy feeding tubes placed in the institution, University of
Michigan Medical Center, performed by a resident revealed that their
placement was associated with a surprisingly high morbidity and even
mortality.

The other difference is that I tended to give all my patients having
a major operation a gavage with Golytley the afternoon before their
operations to cleanse their gut (9). This is not normal practice for upper
gastrointestinal or oesophageal surgery and might not have been either
Orringer’s or Skinner’s. Recent experience with colonic lavage and
diverting loop ileostomy in the management of fulminant amoebic colitis
implicates the translocation of colonic endotoxin, its release of
cytokines and possible uncoupling of oxidative phosphorylation in the
pathogenesis of adverse outcomes from either haemodynamically compensated
or uncompensated shock (10). The practice has reduced the mortality from
fulminant amoebic colitis from some 60% to 80% to 5%. The benefical
effects appears to be primarily due to the prevention of the translocation
of endotoxin and accompanying cytokine release. I have proposed,
therefore, that my practice of cleansing the bowel with Golytely before
major surgery might have been responsible for my achieving a very low
mortality in all my serious elective cases (9).

In hospitals with more than 300 and fewer than 600 or 700 beds
operative mortality would seem, therefore, to be primarily determined by
the technical training surgeons have had and the perioperative management
they provide. If so should achieving a much lower operative mortality
rather than hiring more radiotherapists and oncologists be the primary
goal in improving long term survival from oesophageal cancer? Reducing the
operative mortality from 12.7% or 10% to 1% would confer a long term
survival benefit of equivalent magnitude. Eliminating the need for blood
transfusions, which may halve long term outcome from cancer surgery, would
confer a survival benefit of 100% (7). The combination of the two would
confer a survival benefit in excess of 100%, far in excess of any
survival benefit ever achieved in the management of oesophageal cancers
with adjuvant radiotherapy and/or chemotherapy.

1. Harmon JW, Tang DG, Gordon TA, Bowman HM, Choti MA, Kaufman HS,
Bender JS, Duncan MD, Magnuson TH, Lillemoe KD, Cameron JL. Hospital
volume can serve as a surrogate for surgeon volume for achieving excellent
outcomes in colorectal resection. Ann Surg. 1999 Sep;230(3):404-11;
discussion 411-3.
2. Zafirellis KD, Fountoulakis A, Dolan K, Dexter SP, Martin IG, Sue-Ling
HM. Evaluation of POSSUM in patients with oesophageal cancer undergoing
resection. Br J Surg. 2002 Sep;89(9):1150-5.
3. Jean-François Bosset, Mariette Mercier, Jean-Pierre Triboulet, Thierry
Conroy, Jean-François Seitz Surgical resection with and without
chemotherapy in oesophageal cancer Volume 360, Number 9340 12 October 2002
4. Orringer MB, Marshall B, Iannettoni MD. Transhiatal esophagectomy for
treatment of benign and malignant esophageal disease. World J Surg. 2001
Feb;25(2):196-203. Review.
5. Altorki NK, Girardi L, Skinner DB. En bloc esophagectomy improves
survival for stage III esophageal cancer. J Thorac Cardiovasc Surg. 1997
Dec;114(6):948-55
6. Mariette C, Castel B, Toursel H, Fabre S, Balon JM, Triboulet JP.
Surgical management of and long-term survival after adenocarcinoma of the
cardia. Br J Surg. 2002 Sep;89(9):1156-63.
7. Fiddian-Green RG. Failures of surgical care and outcome from cancer
surgery bmj.com/cgi/eletters/320/7239/895#7229, 31 Mar 2000
8. Fiddian-Green RG. Mitochondrial considerations
bmj.com/cgi/eletters/325/7367/735/a#26019, 4 Oct 2002
9. Fiddian-Green RG. Gastric lavage for major operations?
bmj.com/cgi/eletters/325/7366/674/b#25849, 27 Sep 2002
10. Fiddian-Green RG. Colonic lavage for severe haemorrhagic shock?
bmj.com/cgi/eletters/325/7366/674/b#25839, 27 Sep 2002