Download Is Autopsy Useful? - Société Française de Médecine Légale

BRIEF COMMUNICATION
Postoperative Death and Malpractice Suits:
Is Autopsy Useful?
Philippe Juvin, MD*, Frédéric Teissière,
and Michel Durigon, MD†
MD†,
Fabrice Brion,
MD†,
Jean-Marie Desmonts,
MD*,
*Service d’Anesthésie et de Réanimation, Hôpital Bichat, Paris; and †Service de Médecine Légale, Hôpital Raymond
Poincaré, Garches, France
T
he number of autopsies has declined in many
hospitals in the last few decades (1). One factor
in this trend is probably the greater confidence
with which physicians can establish a premortem diagnosis. However, when a malpractice suit is filed
after a postoperative death, the plaintiff often considers the medical chart data inadequate or untrustworthy. In this situation, an autopsy could have confirmed
the cause of death and eventually determine whether
a fault occurred. The medical value of such autopsies
has not been assessed. The present study was designed to determine whether autopsies performed in
the case of postoperative death with suspicion of malpractice help identify the cause of death.
Methods
This retrospective medical chart review study included all adult (⬎18 yr old) patients who died after a
surgical procedure between 1993 and 1998 in three
districts of France (Yvelines, Eure et Loir, and Val
d’Oise, i.e., approximately 2.8 million inhabitants) and
those who had an autopsy performed after the family
filed a malpractice suit. All autopsies were performed
within 3 days after the death, according to the same
procedure, in a single forensic institute. The organs
were examined in situ, then removed, examined and
dissected ex situ, and subjected to a histological
examination.
The circumstances and the cause of death indicated
by the physician in charge of the patient (primary
diagnosis) were abstracted from the patients’ medical
charts, as were the results of the autopsy (autopsy
diagnosis). The primary and the autopsy diagnoses
were compared by a panel of three certified physicians
Accepted for publication April 26, 2000.
Address correspondence and reprint requests to Philippe Juvin,
Service d’Anesthésie et de Réanimation, Hôpital Bichat, 46 Rue
Henri Huchard, 75018 Paris, France. Address e-mail to pjuvin
@free.fr.
344
Anesth Analg 2000;91:344–6
(one anesthesiologist, one forensic physician, and one
surgical intensivist). When the diagnoses were different, the three physicians reviewed the data to confirm
or refute the autopsy diagnosis. When the autopsy
failed to identify the cause of death, it was determined
whether a different autopsy technique would have
been more informative.
Results
Fifteen autopsies were performed during the study.
Mean time (range) from surgery to death was 4 days
(0 –10 days) (Day 0 ⫽ day of surgery). The clinical
history and primary and autopsy diagnoses of the
patients are reported in Table 1.
A primary diagnosis was recorded in 11 cases
(Cases 1 to 11). Among these primary diagnoses, 3
were confirmed (Cases 1–3) and 7 (Cases 4 –10) were
invalidated by the autopsy. Among these 7 cases, the
autopsy corrected the primary diagnosis in 5 cases
(Cases 4 to 8) and disproved the primary diagnosis
without finding the cause of death in Cases 9 and 10.
An autopsy diagnosis was obtained in 11 cases (all the
cases except Cases 9, 10, 11, and 15). Among the 4
cases with no primary diagnosis, the autopsy identified the cause of death in 3 cases (Cases 12–14). The
technique used during the autopsy was not appropriate to perform a diagnosis in 2 cases (Cases 11 and 15).
Discussion
This report demonstrates that most early postmortem
autopsies performed on patients who died postoperatively provide new and often unexpected information of great assistance in identifying the cause of
death.
Our first result was that most of the diagnoses performed by the physician in charge of the patient at the
time of death were wrong. Our second result was that
the autopsy yielded new information (providing or
©2000 by the International Anesthesia Research Society
0003-2999/00
ANESTH ANALG
2000;91:344 –6
BRIEF COMMUNICATION
JUVIN ET AL.
POSTOPERATIVE DEATH AND AUTOPSY
345
Table 1. Patient Characteristics with Primary and Autopsy Diagnoses
Case
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Clinical history
Medical chart with a primary diagnosis of the death
Cause of death confirmed by the autopsy
Male, 70 yr old. Printzmetal angina. Two-vessel
coronary bypass procedure. Died on D1.
Male, 57 yr old. Uninvestigated chest pain 3 wk
earlier. Surgery for fracture of the humerus.
Arterial hypotension intraoperatively and acute
heart failure postoperatively. Died on the same
day with shock
Female, 35 yr old. Cardiorespiratory arrest during
labor.
Cause of death invalidated by the autopsy
Cause of death identified by autopsy
Male, 77 yr old. Coronary heart disease and left
ventricular failure. Transurethral resection of
the prostate. Seemed to have shown evidence
of chest pain intraoperatively. Was not
investigated postoperatively. Found dead in his
bed on D1.
Female, 92 yr old. Patient on aspirin therapy.
Uninvestigated fall 2 mo before surgery for an
acetabular fracture. Found dead on D2.
Female, 30 yr old. Laparoscopic surgery for
uterine myoma. Circulatory collapse followed
by cardiorespiratory arrest during the
procedure. Died a few hours later.
Female, 20 yr old. Substance abuse.
Appendectomy. Found dead on D10.
Male, 73 yr old. Surgery for an acetabular
fracture. Inadequate perioperative
anticoagulant therapy. Died suddenly on D1.
Cause of death not identified by autopsy
Female, 59 yr old. Surgery for L4-5 disk
herniation. Circulatory collapse followed by
circulatory arrest when turned to the prone
position. Glasgow 3 coma after resuscitation
and until death on D15.
Male, 84 yr old. Colectomy. Fell from his bed on
D7. Found dead on D8.
Female, 47 yr old. Liposuction. Was given 2 g of
amoxicillin intraoperatively. Left the operating
room at 5 pm. At 10 pm, loss of consciousness
followed by circulatory arrest during an
amoxicillin injection. Died on D1.
Medical chart with no primary diagnosis of the death
Cause of death identified by autopsy
Male, 78 yr old. Total gastrectomy for cancer.
Fever on D4. Negative laparotomy on D5. Died
from septic shock on D10.
Male, 87 yr old. Continuity reconstitution after
removal of the transverse colon. Shock at the
end of the procedure. Death on D3 with MOF.
Male, 35 yr old. Surgery for diastasis of the rectus
abdominis muscles. Intestinal obstruction
requiring resection of necrotized bowel loops
on D8. Died on D11 with MOF.
No cause of death identified by autopsy
Male, 68 yr old. History of diabetes, asthma, and
coronary heart disease. Surgery for a fracture of
a toe. Unexplained agitation the next day
requiring injection of droperidol. Found dead
15 min after the injection.
D ⫽ postoperative day, MOF ⫽ multiple organ failure.
Primary diagnosis
Autopsy diagnosis
Myocardial infarction
Myocardial infarction
Myocardial infarction
Myocardial infarction
Amniotic fluid embolism
Amniotic fluid embolism
Myocardial infarction
Old myocardial infarction, with
no evidence of recent cardiac
ischemia. Pleural effusion (1.3
L) caused by decompensated
heart failure and responsible for
the death
Intracerebral bleeding (right
fronto-parietal area and
brainstem)
Congenital abnormality of the
coronary arteries explaining the
recent ischemic lesions
responsible for the death
Massive pulmonary embolism
Pulmonary embolism
Pulmonary embolism
Overdose
Pulmonary embolism
Cancer of the esophagus. Massive
bleeding caused by fistulization
of the esophagus into the aorta
Myocardial infarction
No myocardial infarction. No
findings explaining the death
Cerebral hematoma
Anaphylactic shock
No cerebral hematoma. No
findings explaining the death
No findings explaining the death
None (septic shock of
unknown cause)
Intraperitoneal abscess opposite
the esoduodenal suture
None (MOF of unknown
cause)
Large intraperitoneal hematoma
(⬎5 L)
None (septic shock of
unknown cause)
Necrosis of the entire small bowel.
Massive bleeding from a gastric
ulcer
None
No findings explaining the death
346
BRIEF COMMUNICATION JUVIN ET AL.
POSTOPERATIVE DEATH AND AUTOPSY
correcting the diagnosis) in two of three of the cases.
These proportions are at variance with those of earlier
studies performed in intensive care units (2,3), in
which most primary diagnoses were confirmed by the
autopsy. However, there was no suspicion of malpractice in these previous studies. These discrepancies
may be explained by the fact that autopsies performed
in case of suspicion of malpractice may be more likely
to disprove the primary diagnosis than autopsies
overall. It is indeed reasonable to assume that family
members who did file charges may have done so
because they noticed something that raised their suspicions. It follows that autopsies may have been performed more often in cases characterized by egregious
mistakes; in this situation, the yield of the autopsies
would be high, as was the case in our study. This
recruitment bias in our study is similar to that in the
American Society of Anesthesiologists Closed Claim
Project.
In agreement with previous studies (3,4), the new
information provided by the autopsy could have influenced treatment decisions in some cases. In Cases
12 (unrecognized intraperitoneal abscess), 13 (unrecognized intraperitoneal hematoma), and 14 (unrecognized necrosis of the small bowel), early surgery could
have modified the prognosis. In Case 4, early recognition of decompensated heart failure could have allowed an appropriate treatment. In Case 5, the discontinuation of aspirin therapy after a fall could have
reduced the intracerebral bleeding. In Case 6, an intensive perioperative management of the coronary
disease could have prevented the myocardial infarction. In Case 7, an anticoagulant therapy could have
changed the prognosis. However, Patients 6 and 7
appear to have suffered from rare events, and the
desirable changes in management are not supported
by data on cost-effectiveness. In Case 8, outcome
would not have been different even if the fistula had
been identified. Thus, among the eight cases (Cases
4 – 8 and 12–14) in which the autopsy provided a diagnosis which had not been suspected premortem,
changes in perioperative management would have
modified the prognosis in five cases (Cases 4, 5, and
12–14), suggesting that faulty or negligent practice
may have occurred. More generally, one cannot be
certain that that knowing the true diagnosis would
have prevented the death. Nevertheless, it is clear that
failure to make the diagnosis has reduced the likelihood of patient survival.
In two cases, the autopsy technique was not appropriate for performing a diagnosis. In Case 11, the
primary diagnosis was anaphylactic shock, and the
postmortem specimens for anaphylaxis mediator testing were not obtained. In Case 15, death occurred after
ANESTH ANALG
2000;91:344 –6
injection of a neuroleptic agent, but specimens were
not obtained to look for a drug overdosage or an
anaphylactic reaction. These cases suggest that good
practice recommendations should be developed to determine the specific characteristics of the autopsy procedure to be used in the event of a perioperative death.
Limitations of this report include the small sample
size and the fact that autopsies were performed a few
days after the death. In the hypothesis of a late claim,
months or years after death, the yield of the autopsies
should be reevaluated because of the putrefaction
phenomena. However, the conclusions concerning the
weak value of the primary diagnosis remain valid. In
any case, our report was representative of present and
actual early postmortem claims after postoperative
death in a population of 2.8 million inhabitants.
This report demonstrates the extremely high yield
of early postmortem autopsies performed in the case
of postoperative death with suspicion of malpractice.
They frequently identified undetected complications,
including surgical complications and disease processes. They could also suggest faulty or negligent
practice that would otherwise go unrecognized. As
reported in the present study, the autopsy can help the
defense of the patient’s family when a complication,
particularly a surgical one, was missed (5), or the
defense of unjustly accused physicians when the autopsy findings refute doubts about the appropriateness of their practice. Autopsy can also help conflict
solving: when the family suspects that malpractice has
occurred, they are unlikely to believe the explanations
provided by the physician to justify his management
of the case, and in this situation, an autopsy can restore trust and resolve conflicts by providing data
viewed as “objective” by the family. In this situation,
performing an autopsy can avoid litigation. Finally,
the autopsy can help identify inappropriate and hazardous practices with the goal of improving the safety
of future patients. For all these reasons, anesthesiologists should request autopsy revival (1).
References
1. Marwick C. Pathologists request autopsy revival. JAMA 1995;
273:1889 –91.
2. Goldman L, Sayson R, Robbins S, et al. The value of the autopsy
in three medical eras. New Engl J Med 1983;308:1000 –5.
3. Blosser SA, Zimmerman HE, Stauffer JL. Do autopsies of critically ill patients reveal important findings that were clinically
undetected? Crit Care Med 1998;26:1332– 6.
4. Battle RM, Pathak D, Humble CG, et al. Factors influencing
discrepancies between premortem and postmortem diagnoses.
JAMA 1987;258:339 – 44.
5. Charatan F. Medical errors kill almost 100,000 Americans a year
[letter]. BMJ 1999;319:1519.