The Journal of TRAUMA威 Injury, Infection, and Critical Care

Time to Laparotomy for Intra-abdominal Bleeding from

Trauma Does Affect Survival for Delays Up to 90 Minutes
John R. Clarke, MD, Stanley Z. Trooskin, MD, Prashant J. Doshi, MS, Lloyd Greenwald, PhD, and
Charles J. Mode, PhD

Objective: We examined the relation-
ship between survival and time in the
emergency department (ED) before lapa-
rotomy for hypotensive patients bleeding
from abdominal injuries.
Methods: Patients in the Pennsylva-
nia Trauma Systems Foundation trauma
registry with isolated abdominal vascular,
solid organ, or wall injuries grade 3 to 6
and hypotension were identified. Deaths
were predicted from the prehospital time,
systolic blood pressure (SBP) on ED ad-
mission, and time in the ED before either
laparotomy or ED death.
Results: Two-hundred forty-three
patients met the criteria. SBP ranged
from 30 to 90 mm Hg. Time to the ED
ranged from 7 to 185 minutes. Time in the
ED ranged from 7 to 915 minutes. Overall,
98 patients died (40%). The risk ratio for
the SBP increased, as expected, as SBP
dropped. The risk ratio for time spent in
the ED before laparotomy increased until
90 minutes, then significantly decreased
below all earlier values. Logistic regres-
sion on the 165 patients spending 90 min-
utes or less in the ED showed that the
probability of death increased with time in
the ED. The increase was as much as
0.35% per minute.
Conclusion: Among patients in a
trauma registry who were hypotensive on
arrival in the ED and had major injuries
isolated to the abdomen requiring emer-
gency laparotomy, the probability of
death showed a relationship to both the
extent of hypotension and the length of
time in the ED for patients who were in
the ED for 90 minutes or less. The prob-
ability of death increased approximately
1% for each 3 minutes in the ED.
Key Words: Abdominal injuries,
Emergencies, Emergency service, Hemor-
rhage, Hemorrhagic shock, Hypotension,
Laparotomies, Resuscitation, Risk ratio,
Survival analysis, Time factors, Trauma
centers, Trauma registries.
J Trauma. 2002;52:420 –425.

I
s time in the emergency department (ED) associated with
improved outcomes for hypotensive injured patients need-
ing emergency surgery or is the delay in definitive control
of internal bleeding associated with worsened outcomes? We
could not find the answer to that question in the medical
literature. This study examined the effect of time in the ED on
hypotensive patients needing definitive operative control of
bleeding.
MATERIALS AND METHODS
Information from the Pennsylvania Trauma Systems
Foundation (PTSF) trauma registry was used for the study.
The PTSF trauma registry contains information on all injured
patients cared for at Pennsylvania trauma centers who die, are
transferred to other trauma centers, or have a hospital stay of
more than 2 days; patients with isolated hip fractures are
Submitted for publication October 1, 2001.
Accepted for publication November 27, 2001.
From the Department of Surgery, MCP-Hahnemann University (J.R.C.,
S.Z.T.), the Department of Computer and Information Science, University of
Pennsylvania (J.R.C.), and the Department of Mathematics and Computer Sci-
ence, Drexel University (P.J.D., L.G., C.J.M.), Philadelphia, Pennsylvania.
Funded, in part, by a Synergies Grant from Drexel University.
This work was scheduled for presentation at the 61st Annual Meeting
of the American Association for the Surgery of Trauma, which was canceled
because of the terrorist attacks of September 11, 2001.
Address for reprints: John R. Clarke, MD, Department of Surgery,
MCP-Hahnemann University, 3300 Henry Avenue, Philadelphia, PA 19129;
email: jclarke@gradient.cis.upenn.edu.
excluded.1,2 The registry began enrolling patients in October
1986.
The PTSF trauma registry includes the following infor-
mation used for this study: the time of injury, the time the
ambulance was dispatched, the time the patient arrived at the
hospital’s ED, the patient’s systolic blood pressure (SBP) on
arrival in the ED, the patient’s diagnoses, preexisting condi-
tions, the time the patient left the ED, the time the patient
arrived in the operating room (OR), the operative procedures
performed, and the patient’s outcome (lived or died). It did
not include information about fluid given to the patient or the
systolic blood pressure on departure from the ED or arrival in
the OR.
The MCP-Hahnemann University Institutional Review
Board approved this study as exempt because the patient and
institutional identifications were coded and information could
not be identified by patient.
This study was restricted to patients who were brought
directly to the trauma centers from the injury scenes before
July 1999, were not transferred from the ED to another
hospital, and did not have confounding burn injuries (n ⫽
166,768). The cohort for this study was further restricted to
patients who had no documented confounding preexisting
conditions (n ⫽ 130,302). From this group of injured patients
without confounding conditions who were brought directly to
a trauma center, patients who met the following criteria were
identified: the patient’s SBP was greater than 0 mm Hg and
less than or equal to 90 mm Hg on arrival in the ED; the
patient either died in the ED or was transferred from the ED

420 March 2002


Table 1 Operations Included in the Study (with
ICD-9-CM Procedure Codes)
Laparotomy (ⱖ 54 and ⬍55) or
Adrenal surgery (ⱖ 7.0 and ⬍ 7.5) or
Vascular surgery (ⱖ 38 and ⬍ 40) or
Splenic surgery (ⱖ 41.1 and ⬍ 42) or
Gastrointestinal surgery (ⱖ 43 and ⬍ 49) or
Hepatopancreaticobiliary surgery (ⱖ 50 and ⬍ 53) or
Urinary surgery (ⱖ 55 and ⬍ 60) or
Female genital surgery (ⱖ 65 and ⬍ 67) or (ⱖ 68 and ⬍ 70) or
Obstetric abdominal surgery (ⱖ 74 and ⬍ 76)
ICD-9-CM, International Classification of Diseases, Ninth Revi-
sion, Clinical Modification.
to the OR for a laparotomy (Table 1); the patient had an
abdominal vascular, solid organ, or wall injury with an Ab-
breviated Injury Scale score3 in the range of 3 to 6 (Table 2);
the patient had no injuries with an Abbreviated Injury Scale
score greater than 2 in any other body region except for a
lacerated diaphragm (440604.3) or an open, displaced, or
comminuted pelvic fracture (852604.3); and either the time of
injury or ambulance dispatch and the time of arrival to the ED
and the time of departure from the ED or death in the ED or
arrival to the OR were recorded.
For each patient, the SBP on arrival to the ED and the
outcome were recorded and the following were calculated:
the elapsed time to the ED (prehospital time) was calculated
from either the dispatch time, when known, or the injury time
(in that order of preference) and the time of arrival to the ED.
The time of dispatch was preferred, for consistency, because
the registry permitted registrars to enter the time of dispatch
as an undeclared default for the time of injury when the time
of injury was otherwise unknown. The elapsed time in the ED
was calculated from the time of arrival to the ED and the time
Table 2 Diagnoses Included in the Study (with AIS
Numerical Codes and Severity Scores)
Penetrating abdominal injury with blood loss (516006.3) or
Abdominal vascular injury (ⱖ 540204.4 and ⱕ 521606.4, but not
520402.3, 520602.3, or 521402.3) or
Adrenal avulsion (540226.3) or
Renal contusion with expansion (541614.3) or
Renal laceration with vessel (541626.4 or 541628.5) or
Liver contusion (541814.3) or
Liver laceration (541824.3, 541826.4, 541828.5, or 541830.6) or
Mesenteric laceration (542024.3 or 542026.4) or
Omental laceration (542624.3) or
Pancreatic laceration with vessel or duct injury (ⱖ 542824.3 and
ⱕ 542832.5) or
Placental abruption (543400.3 or 543402.4) or
Retroperitoneal hematoma (543800.3) or
Splenic contusion (544214.3) or
Splenic laceration (544224.3, 544226.4, or 544228.5) or
Laceration of the pregnant uterus (ⱖ 545226.3 and ⱕ 545246.5) or
Pelvic fracture with vascular injury or hematoma (852606.4,
852608.4, or 852610.5)
Volume 52 • Number 3
Time to Laparotomy Does Affect Survival
of arrival in the operating room or the time of departure from
the ED (in that order of preference).
We were aware that short times to care could either
decrease mortality rates or imply patients at high risk of death
and that long times to care could either increase mortality
rates or imply patients at low risk of death. Therefore, times
to care might not predict outcome monotonically. Because of
the potential fluctuations in the direction of the relationship
between time and outcome, the data were analyzed using
interval risk ratios rather than regression formulas.
A risk ratio (RR) of death associated with an attribute is
the ratio of two percentages: the percentage of individuals
with the attribute among those who died compared with the
percentage of individuals with the attribute among those who
lived (RR is equal to percent with attribute among deaths
divided by percent with attribute among survivors).4 At-
tributes with risk ratios greater than 1.0 are associated with an
increased risk of death; those with risk ratios less than 1.0 are
associated with a decreased risk. Risk ratios are considered
significant if their 95% confidence intervals are either com-
pletely above or completely below 1.0.
The interval RR for death was calculated for the time to
the ED, the time in the ED, the total time (time to ED ⫹ time
in ED), and the SBP on arrival to the ED. We did use logistic
regression to model predictions of outcome using the contin-
uous variables of time and SBP within the time intervals that
were found to have significant risk ratios.
RESULTS
Two hundred fifty patients met the predetermined selec-
tion criteria. Of those, two were excluded because the time to
the ED seemed extraordinary: one patient was recorded as
having an injury time 8 hours 1 minute before ED arrival with
a SBP of 58 mm Hg, and the other patient was recorded as
having an injury time 7 days 20 hours 16 minutes before ED
arrival with a SBP of 80 mm Hg. Five other patients were
excluded because the recorded time in the ED exceeded 24
hours.
In the final cohort of 243 patients studied, the SBP on
arrival to the ED ranged from 30 to 90 mm Hg, with 200
patients having a SBP less than 90 mm Hg. Elapsed time to
the ED ranged from 7 to 185 minutes, with a median of 43
minutes. Time in the ED ranged from 7 to 915 minutes, with
a median of 55 minutes; 201 patients were within the PTSF
120-minute quality assurance (QA) standard (83%; 95% con-
fidence interval [CI], 77– 88%). Total time from ambulance
dispatch or injury to arrival in the operating room or depar-
ture from the ED ranged from 28 to 938 minutes, with a
median of 110 minutes. Overall, 98 patients died (40%; 95%
CI, 34 – 46%); 4 died in the ED.
As expected, the risk of death was significantly influ-
enced by the systolic blood pressure on arrival in the ED.
Among these patients with systolic blood pressures of 90 mm
Hg or less, the risk of death was significantly higher in those
patients whose SBP was less than 60 mm Hg and signifi-
421
 The Journal of TRAUMA威 Injury, Infection, and Critical Care

cantly lower in those patients whose SBP was 80 mm Hg or
higher (Table 3).
The risk of death was not significantly influenced by the
time between dispatch or injury and arrival at the ED (Table
4). The risk was increased between 30 and 60 minutes, but
not significantly, suggesting the possibility of an increased
risk with delays in transport, followed by a self-selection of
survivors with long delays.
The risk of death increased with time spent in the ED,
becoming significant for the interval 61 to 90 minutes, then
significantly decreased with stays beyond 90 minutes (Table
5). The pattern suggests increased risk with delays in the ED,
followed by a bias toward stable patients with further delays.
The risk of death increased with total time (time to the ED
and in ED), becoming significantly higher for the interval 61 to
90 minutes, then decreased beyond 120 minutes, becoming
significantly lower beyond 240 minutes (Table 6). The results
for the total time seem to follow the patterns of its components.
We made a logistic regression model of the probability
of death for those patients with a significant increased risk of
death with time spent in the ED, namely, those whose stay
was 90 minutes or less. For those patients (n ⫽ 165), a
logistic regression was performed on the basis of the SBP on
arrival in the ED in millimeters of mercury, the prehospital
time (PHT) in minutes, and the time in the ED (EDT) in
minutes: ln[p/(1 ⫺ p)] ⫽ 3.36166 ⫺ SBP ⫻ 0.05276 ⫺ PHT
⫻ 0.00032 ⫹ EDT ⫻ 0.01461.
The 95% CI on the coefficient for time in the ED
(0.01222–1.19553) confirms a significant positive correlation
between the time in the ED and the probability of death for
patients departing the ED within 90 minutes of arrival.
The highest average impact of time spent in the ED was
an average increase in the probability of death of 0.0035
(0.35%) per minute in the ED, or approximately 1 in 300
patients per minute. It occurred with the shortest prehospital
time (7 minutes) and a neutral SBP of 78 mm Hg. The lowest

Table 3 Risk Ratios for Systolic Blood Pressure on Arrival in the ED

SBP Died
Live Die Total RR –2 SD ⫹2 SD
(mm Hg) (%)

30–39 1 3 4 4.439 0.468 42.057 75.00

40–49 3 3 6 1.480 0.305 7.181 50.00
50–59 8 13 21 2.404 1.035 5.584 61.90*
60–69 19 18 37 1.402 0.776 2.533 48.65
70–79 26 28 54 1.593 0.998 2.545 51.85
80–89 55 23 78 0.619 0.409 0.936 29.49**
90 33 10 43 0.448 0.232 0.867 23.26**
All patients 145 98 243 40.33
* Significantly higher than the overall average (p ⬍ 0.05).
** Significantly lower than the overall average (p ⬍ 0.05).

Table 4 Risk Ratios for Time to Arrival to the ED

Died
Minutes to ED Live Die Total RR –2 SD ⫹2 SD
(%)

1–30 44 23 67 0.773 0.501 1.194 34.33

31–60 63 54 117 1.268 0.980 1.641 46.15
61–90 32 18 50 0.832 0.496 1.396 36.00
91–185 6 3 9 0.740 0.189 2.888 33.33
All patients 145 98 243 40.33

Table 5 Risk Ratios for Time in the ED

Died
Minutes in ED Live Die Total RR –2 SD ⫹2 SD
(%)

1–30 27 25 52 1.370 0.848 2.214 48.08

31–60 40 38 78 1.406 0.978 2.020 48.72
61–90 15 20 35 1.973 1.063 3.662 57.14*
91–120 30 6 36 0.296 0.128 0.684 16.67**
121–910 33 9 42 0.404 0.202 0.805 21.43**
All patients 145 98 243 40.33
* Significantly higher than the overall average (p ⬍ 0.05).
** Significantly lower than the overall average (p ⬍ 0.05).

422 March 2002

 Time to Laparotomy Does Affect Survival

Table 6 Risk Ratios for Total Time to the OR

Died
Minutes to ED Live Die Total RR –2 SD ⫹2 SD
(%)

1–30 3 2 5 0.986 0.168 5.795 40.00

31–60 21 10 31 0.705 0.347 1.430 32.26
61–90 21 35 56 2.466 1.531 3.971 62.50*
91–120 26 23 49 1.309 0.795 2.156 46.94
121–150 21 13 34 0.916 0.482 1.741 38.24
151–180 17 6 23 0.522 0.213 1.278 26.09
181–210 13 5 18 0.569 0.210 1.545 27.78
211–240 7 3 10 0.634 0.168 2.393 30.00
241–938 16 1 17 0.092 0.012 0.686 5.88**
All patients 145 98 243 40.30
* Significantly higher than the overall average (p ⬍ 0.05).
** Significantly lower than the overall average (p ⬍ 0.05).

average impact was an average increase in the probability of
death of 0.0011 (0.11%) per minute in the ED, or approxi-
mately 1 in 900 patients per minute. It occurred with the
longest prehospital time (185 minutes) and the lowest SBP of
30 mm Hg, which was the main predictor of outcome.
DISCUSSION
The intent of the study was to detect the effect of time in
the ED on hypotensive patients needing definitive operative
control of bleeding. We chose systolic blood pressure as the
best available surrogate for shock. We included 90 mm Hg to
ensure inclusiveness and maximize the size of the cohort. We
focused on patients whose hypotension was associated with,
and presumably caused by, major intra-abdominal injuries
that produce blood loss, because laparotomy was by far the
most common emergency operation on hypotensive patients
and times of operations were well documented in the registry.
External blood loss from simple lacerations was not recorded
and is an unmeasured variable. However, hypotensive pa-
tients with major injuries that are likely to cause internal
bleeding and are isolated to the abdomen are candidates for
emergency laparotomy, even if they have additional external
bleeding from simple lacerations that can be controlled in the
ED. By focusing on hypotensive patients whose diagnoses
warranted emergency laparotomy, rather than just patients
who had emergency laparotomies, we captured four patients
who died before getting to the OR. We did not include other
patients who received operations belatedly or not at all, on the
premise that their surgeons identified information, not re-
corded, that implied a low risk of death from delay or non-
operative management. The significant decrease in death rate
with operative delays of more than 90 minutes in the ED and
more than 4 hours overall from injury support that conjecture.
The cohort might be biased by the requirement that times
of dispatch or injury be available, but the results would be
obscured by the biases of missing data with the alternative.
The other information used was consistently available. All
information was carefully checked for accuracy and consis-
tency; corrections were made when indicated, most notably
for failures to change the date for time intervals that crossed
midnight. Frequency patterns for the times suggested that
they were sometimes rounded to the nearest 5 minutes. The
cohort might also be biased by the failure to identify, and
therefore document, preexisting conditions or other major
injuries in patients who received expedited assessment in the
ED and died quickly because of their obviously unstable
condition. Information about fluid resuscitation and SBP on
arrival in the operating room or departure from the ED would
have been useful, but was not available. The information
about prehospital blood pressure readings was available, but
was only recorded in about half of the cohort. We did not
include the arrival time of the attending surgeon; the presence
of the attending surgeon could have been a causal factor or
just an indicator of other factors driving the outcome.
The median time in the ED (55 minutes) was comparable to
times reported by others. For patients departing for emergency
trauma laparotomies, Henderson et al. reported a median time in
the ED of 54 minutes at their institution in the United Kingdom,
versus 115 minutes in the U.K. national database.5 For all emer-
gency trauma surgery, McNicholl and Dearden reported 117 and
111 minutes for each of 2 years at their institution, again in the
United Kingdom.6 Lowe et al., in the United States, reported 136
minutes between arrival in the ED and arrival in the OR for all
emergency trauma surgery.7 Khetarpal et al., also in the United
States, compared two institutions, one with and one without an
attending trauma surgeon present for all trauma resuscitations.8
They reported 102 and 107 minutes from ED admission to
incision, respectively, for patients with blunt trauma and 50 and
66 minutes, respectively, for patients with penetrating trauma.
Porter and Ursic, in the United States, reported 44 minutes to the
OR for all emergency trauma surgery when attending trauma
surgeons were in the resuscitation room and 109 minutes when
they were not.9
The total time from ambulance dispatch or injury until
arrival in the operating room or departure from the ED (110
minutes) compared favorably to Henderson et al.’s median total
time of 127 minutes for emergency trauma laparotomies at their
institution and 161 minutes for the U.K. national database.5

Volume 52 • Number 3 423

 The Journal of TRAUMA威 Injury, Infection, and Critical Care
The results suggest that both of our competing conjec-
tures about the relationship of death to the promptness of
operative control of bleeding in hypotensive patients were
true. Risk of death was higher for those patients departing the
ED within the first 90 minutes of arrival compared with those
with longer stays, as detected by the interval risk ratios,
suggesting greater urgency for patients recognized as unsta-
ble on the basis of indicators not captured in the registry. The
probability of death also increased as the time in the ED
increased over 90 minutes, as documented by the logistic
regression of the patients within that subgroup.
The contribution of time in the ED to the probability of
death in hypotensive patients needing definitive operative
control of bleeding ranged from approximately 1 death per
300 patients per minute to 1 death per 900 patients per
minute, depending on the contribution of prehospital time and
blood pressure. According to our analysis, a cohort of 100
hypotensive patients needing emergency surgery to control
intra-abdominal bleeding will experience one additional
death every 3 to 9 minutes in the ED. In the worst-case
scenario, an additional 5% of the cohort will die for every
extra 15 minutes in the ED.
Whether the time in the ED implied avoidable delay could
not be determined in this study. Alternative hypotheses include
patient factors, such as obesity or lack of cooperation because of
intoxication, and provider issues that would independently affect
the time needed to give care and the outcome. The absence of
major injuries to other body areas among the patients in this
cohort should minimize variability in the amount of ED care
needed. The other studies of time to the OR for emergency
trauma surgery suggest that delays are a factor. Henderson et al.
reviewed the case records of 21 patients whose total time to OR
was in the upper quartile and in whom delays were identified;
five had avoidable delays in the ED, because of either prolonged
resuscitation or inappropriate investigations.5 McNicholl and
Dearden noted sequential management, limited resources, and
failure to alert the trauma team as causes of delay, suggesting
improvement by earlier decisions, communication alerts, and
parallel management.6 Lowe et al. reported that their trauma
resuscitations averaged 24 minutes, with another 112 minutes
spent before getting to the OR.7 Porter and Ursic noted statisti-
cally significant differences in times to the OR within an insti-
tution, depending on whether the attending trauma surgeon was,
or was not, in the resuscitation room.9 Khetarpal et al. noted
statistically significant differences in times from ED admission
to incision, for patients with penetrating trauma only, comparing
institutions with and without an attending trauma surgeon
present for all trauma resuscitations.8
The time in the ED had greater impact than the prehos-
pital time. The reasons were not obvious. Hypotheses include
the biases of including only patients with documented pre-
hospital times and self-selection of survivors with delays in
extrication and transport time. Alternatively, ED providers
may have used their time to actively resuscitate patients to
near normal blood pressure, causing rebleeding (“popping the
424
clot”) before definitive operative management of the injury,
as postulated by Bickell et al.10
A PTSF QA filter has been any trauma patient who had
a laparotomy more than 2 hours after arrival to the ED,
particularly if the patient was hypotensive. Although 9 deaths
occurred among the 42 patients in this study who met the
PTSF QA criterion for review, the mortality rate (21%) was
significantly less than for those who had ED stays of less than
90 minutes. Given the increased probability of death with
time in the ED for stays of less than 90 minutes, review of ED
care for possible delays should be considered for any hypo-
tensive patients needing emergency laparotomy who dies.
CONCLUSION
In a study of 243 patients in the Pennsylvania Trauma
Systems Foundation trauma registry who were hypotensive on
arrival in the ED and had major injuries isolated to the abdomen
requiring emergency laparotomy, the probability of death
showed a relationship to both the extent of hypotension and the
length of time in the emergency department for those patients
who were in the ED for 90 minutes or less. The probability of
death increased as much as 1% for each 3 minutes in the ED.
The ED care of hypotensive patients with abdominal bleeding
needing emergency laparotomies for definitive operative control
of the bleeding should be expedited as efficiently as possible.
ACKNOWLEDGMENTS
We thank Mary Ann Spott, Associate Director of the Pennsylvania
Trauma Systems Foundation, and Lizann Scott, Medical College of Penn-
sylvania Hospital Trauma Registrar, for their assistance.
REFERENCES
1. Gillott AR, Thomas TM, Forrester C. Development of a statewide
trauma registry. J Trauma. 1989;29:1667–1672.
2. Forrester CB, McMinn DL. Anatomy of a statewide trauma registry.
Top Health Rec Manage. 1990;11:34 – 42.
3. American Association of Automotive Medicine. The Abbreviated
Injury Scale. Des Plaines, IL: American Association of Automotive
Medicine; 1990.
4. Simel DL, Samsa GP, Matchar DB. Likelihood ratios with
confidence: sample size estimation for diagnostic test studies. J Clin
Epidemiol. 1991;44:763–770.
5. Henderson KI, Coats TJ, Hassan TB, Brohi K. Audit of time to
emergency trauma laparotomy. Br J Surg. 2000;87:472– 476.
6. McNicholl BP, Dearden CH. Delays in care of the critically injured.
Br J Surg. 1992;79:171–173.
7. Lowe DK, Hedges JR, Marby DW, Mendelson D. An assessment of
time following trauma resuscitation: the transitional evaluation and
monitoring phase. J Trauma. 1991;31:1265–1269.
8. Khetarpal S, Steinbrunn BS, McGonigal MD, et al. Trauma faculty
and trauma team activation: impact on trauma system function and
patient outcome. J Trauma. 1999;47:576 –581.
9. Porter JM, Ursic C. Trauma attending in the resuscitation room: does
it affect outcome? Am Surg. 2001;67:611– 614.
10. Bickell WH, Wall MJ, Pepe PE, et al. Immediate versus delayed
fluid resuscitation for hypotensive patients with penetrating torso
injuries. N Engl J Med. 1994;331:1105–1109.
March 2002

EDITORIAL COMMENT
This very well done study concludes that within the first
90 minutes after arrival, mortality of hypotensive patients
increased with each passing minute spent in the ED before
laparotomy. The review suggests that there may be opportu-
nities for improvement in the care of these patients by exam-
ining the first 90 minutes of their care for avoidable delays.
This is an important finding, since most performance im-
provement indicators designed to detect avoidable delays
examine patients who spend longer than 90 minutes in the ED
before laparotomy. This study suggests that the mortality in
that group is low and that our efforts might be better spent
looking at the group of patients who go to the OR in less than
90 minutes.
While I believe that this is an intriguing possibility that
warrants further study, there are not sufficient data available
in this study to prove it. This study suffers from the same
limitations that all studies using only registry data suffer
from. That is, there is a limited data set. There is no way of
evaluating why a patient was in the ED for a certain amount
of time or what diagnostic tests or therapeutic procedures
were performed during that time. These data are only avail-
able by examining individual charts, and that was not done in
this study. Furthermore, this study looks retrospectively at
Volume 52 • Number 3
Time to Laparotomy Does Affect Survival
patients who went to the OR for intra-abdominal bleeding and
had no other bleeding source. However, when a patient ar-
rives hypotensive in the ER, it is often not obvious from
where he or she is bleeding. It takes time to find out. Factors
such as how the patient responded to resuscitation, whether
they needed intubation, how much blood was transfused, and
which tests were done to identify the source of hemorrhage
are not available for evaluation.
Additionally, I believe that the data may indicate that
there is no detrimental time effect until 60 minutes has
passed. The risk ratio from 31 to 60 minutes is almost iden-
tical to that from 0 to 30 minutes. It is only the 61-to-90-
minutes group that seems to have a higher mortality. Should
we focus our efforts there?
Despite these limitations I believe this is an extremely
important study. The facts support the idea that prospective
studies should be performed to determine whether active
efforts to minimize ED time for hypotensive patients will
improve outcome.
H. Gill Cryer, MD
Division of General Surgery
UCLA School of Medicine
Los Angeles, California
425