Clinical Chemistry 47:7
1204 –1211 (2001)
Discontinuation of the Bleeding Time Test without
Detectable Adverse Clinical Impact
Christopher M. Lehman,1,3* Robert C. Blaylock,1,3 Donald P. Alexander,4 and
George M. Rodgers1,2,3
Background: The bleeding time (BT) test predicts a
higher bleeding complication rate in populations at risk
for inherited or acquired platelet dysfunction, but it is
of limited assistance in evaluating individual patients.
There are no reports of clinical outcomes after discon-
tinuation of the BT test.
Methods: Interviews with a subset of the physicians
who had ordered the BT test before discontinuation of
the test were conducted. The total number of platelet-
aggregation tests, the mean number of monthly, unmod-
ified platelet units transfused, the incidence of kidney
biopsy complications, and the number of doses of
1-deamino-8-D-arginine vasopressin (DDAVP) adminis-
tered 5 months before and after discontinuation of the
BT test were compared. We recorded the rates of bleed-
ing complications in the Major Surgery Risk Pool dur-
ing the 12 months before and the 5 months after the
discontinuation of the BT test.
Results: Clinicians reported they did not significantly
change their preprocedural work-ups, postpone an in-
vasive procedure, experience an increase in bleeding
complications, or increase their use of blood products
after discontinuation of the BT test. Platelet-aggregation
tests (n ⴝ 9, before and after), platelet transfusions (P ⴝ
0.958), and DDAVP administration (before ⴝ 24; after ⴝ
10) did not increase after discontinuation of the BT test.
The rate of postprocedural bleeding complications did
not increase significantly in either Major Surgery Risk
Pool cases (<3␴ deviation from the mean rate) or in
patients undergoing renal biopsies (P ⴝ 0.225 for de-
crease in hematocrit; P ⴝ 1.000 for the percentage of
patients transfused) after discontinuation of the BT test.
1
Department of Pathology, Division of Clinical Pathology, 2 Department
of Medicine, Division of Hematology, and 4 Department of Pharmacy Services,
University of Utah Health Sciences Center, Salt Lake City, UT 84132.
3
ARUP Laboratories, Inc., 500 Chipeta Way, Salt Lake City, UT 84108.
*Address correspondence to this author at: Department of Pathology,
University of Utah Health Sciences Center, 5C124 SOM, 50 N. Medical Drive,
Salt Lake City, UT 84132. Fax 801-585-6666; e-mail lehmanc@arup-lab.com.
Received in revised form March 12, 2001; accepted April 5, 2001.
1204
Evidence-based
Laboratory Medicine
and Test Utilization
Conclusions: Our study failed to identify a clinically
significant, negative impact of discontinuing the BT
test.
© 2001 American Association for Clinical Chemistry
Excessive bleeding may complicate medical procedures
when a patient has an acquired or inherited disorder of
platelet function (1 ). Bleeding time (BT)5 tests were de-
veloped in the hope that quantifying the length of time a
patient bled after incising or puncturing the skin would
aid in the diagnosis of platelet dysfunction disorders and
would predict the risk of abnormal bleeding from internal
organs during and/or after an invasive procedure (2, 3 ).
Unfortunately, the BT test became popularized in clinical
practice without sufficient peer-reviewed evidence sup-
porting its use, probably under the assumption that an in
vivo test was needed and no other test was available. By
the early 1990s, sufficient data questioning the utility of
the BT test as a predictor of clinical bleeding had accu-
mulated to prompt the publication of two review articles
opposing routine preoperative use of the BT test, and
suggesting that the BT test might be of limited use under
any circumstances (4, 5 ). The College of American Pathol-
ogists and the American Society of Clinical Pathologists
later published a position article (6 ) concluding that the
BT test was not effective as a screening test in patients
lacking a history of excessive bleeding, including patients
who had recently taken aspirin or nonsteroidal antiin-
flammatory agents (NSAIDs). Nonetheless, the BT test is
still commonly recommended for use in assessing the risk
of bleeding in patients taking drugs or herbs or suffering
from inherited or acquired disorders known to adversely
affect platelet function (7–12 ).
In recent years, the frequency in ordering the BT test at
our institution had gradually decreased to the point that
its use was confined to patient populations with impend-
5
Nonstandard abbreviations: BT, bleeding time; NSAID, nonsteroidal
antiinflammatory agent; BUN, blood urea nitrogen; and DDAVP, 1-deamino-
8-d-arginine vasopressin.
 Clinical Chemistry 47, No. 7, 2001 1205

ing invasive procedures who were presumed to be at risk

for acquired platelet function disorders. The decreased
volume of testing made it difficult to train new personnel,
to document continuing competency of existing person-
nel, and to perform proficiency testing as required by
good laboratory practice and the Clinical Laboratory
Improvement Act of 1988. Consequently, our Hospital
Laboratory Committee, composed of diverse clinicians
and administrators, reevaluated the efficacy of the BT test
in contemporary clinical practice and decided to discon-
tinue offering the test to our clinicians. Although other
institutions have reportedly limited or eliminated the BT
test (13, 14 ), we are unaware of any reports of clinical
outcomes after discontinuation of the test. This report
summarizes our experiences after removal of the BT test
from the repertoire of clinical laboratory tests at a busy,
tertiary care, university medical center.

Materials and Methods

Clinicians with admitting privileges to the Hospital were
notified by mail of discontinuation of the BT test as
approved by a multidisciplinary Hospital Laboratory
Committee in January 1999. The memo included back-
ground information on the BT test, as well as a discussion
of its limitations and citations of the literature supporting
its discontinuation. An algorithm for evaluating the risk
of a bleeding disorder in patients with a personal or
family history of bleeding (Fig. 1) was also included in the
mailing, along with phone numbers of two physician
laboratorians who were available for consultation. The
University of Utah Health Sciences Center is a tertiary
care center with active adult Hematology/Oncology, Ne-
phrology, Transplant (renal, heart, heart/lung, kidney,
bone marrow), Gastroenterology and Neurosurgery pro-
grams. The Newborn Intensive Care and Well Baby Units
represent the only pediatric patient care units. The clinical
specialties represented on the Hospital Committee at that
time included Family Practice, Hematology, Pulmonary
Medicine, General Surgery, Obstetrics/Gynecology, An-
esthesiology, Pathology and Emergency Medicine.
Indications for platelet transfusions administered to
patients with normal platelet counts in the first 4.5 months
after discontinuation of the BT test were determined by
reviewing patient medical records. The mean number of
nonirradiated, nonleukocyte-reduced apheresis and
pooled platelet units transfused during the 5 months
(February 1999 through June 1999) after discontinuation
of the BT test was compared with the mean number
transfused in the 5 months immediately before discontin-
uation of the BT test and with the mean number trans-
fused during the same 5-month interval of the preceding
year (February 1998 through June 1998), using the Student
t-test (SPSS 9.0.1; SPSS). The unmodified platelet products
were chosen for analysis because they would be used for
nonimmunosuppressed patients, the primary patient
population to undergo BT testing at our institution. The
number of platelet-aggregation tests ordered by clinicians
Fig. 1. Algorithm for evaluating the risk of excessive bleeding in
patients with a personal or family history of bleeding, which was
distributed to clinicians before discontinuation of the BT test.
PT, prothrombin time; PTT, partial thromboplastin time.
during the same 5-month intervals described above were
also compared. To assess the potential effect of discontin-
uation of the BT test on surgical procedures, the summary
statistics for the overall rate of postprocedural hemor-
rhage or hematoma in patients in the Major Surgery Risk
Pool (University HealthSystem Consortium Clinical Per-
formance Measure 2947) for the 12 months before and the
5 months after discontinuation of the BT test were re-
viewed. The Major Surgery Risk Pool consists of 126
surgical procedures including gastrointestinal, cardiovas-
cular, oncologic, transplant, urologic, orthopedic, plastic,
gynecologic, and endocrine procedures. Bleeding compli-
cations are identified by reviewing bleeding-related
ICD-9 codes reported in association with Major Surgery
Risk Pool procedures. Complication rates are calculated
as the number of patients experiencing postprocedural
hemorrhage or hematoma divided by all eligible surgical
patients. The monthly complication rates were compared
through the use of a Shewart control chart for attributes
(15 ). Complication rates for the 12 months before and 5
months after discontinuation of the BT test were plotted
on a “fraction-defective” or “P” control chart with stan-
dard upper and lower control limits of ⫾ 3␴, based on a
1206 Lehman et al.: Bleeding Time Test
binomial distribution (SPSS 9.0.1). Complication rates
exceeding the mean rate by 3␴ are an indication of a
process that is statistically out of control (15 ).
To assess the possible impact of discontinuation of the
BT test on bleeding complications in a patient population
known to be predisposed to abnormal platelet function
and who undergo invasive diagnostic procedures, we
chose to evaluate the complication rate of kidney biopsy
procedures performed during the 5 months immediately
before and the 5 months immediately after discontinua-
tion of the BT test. Demographic characteristics (age, sex),
the origin of the biopsied kidney (native vs allograft), and
pertinent laboratory values [prebiopsy BT test, blood urea
nitrogen (BUN), creatinine, hematocrit, platelet count,
prothrombin time, partial thromboplastin time, and the
postbiopsy hematocrit] were summarized for each
5-month interval. The percentage of patients with abnor-
mal coagulation values (prothrombin time ⬎15.0 s; partial
thromboplastin time ⬎40 s), an abnormally low platelet
count (⬍140 ⫻ 109/L), or a hematocrit ⬍0.300 (30%) was
calculated. The incidence of red blood cell transfusions
secondary to postbiopsy bleeding, documented by ab-
dominal computed tomography, was also assessed for the
two intervals. Continuous variables were tested with the
Kolmogorov–Smirnov test to determine whether ob-
served values came from a gaussian distribution. All
variables followed a gaussian distribution except for
creatinine and BUN. Mean values were compared using
the two-sided Student t-test. Median values were com-
pared using the Mann–Whitney test. Frequencies were
compared using the Pearson ␹2 test unless the expected
value of any cell was ⬍5, in which case the Fisher’s exact
test was used. Potential correlations between variables
were evaluated using the Spearman’s ␳ coefficient. All
statistical tests were performed using SPSS 9.0.1. A prob-
ability of P ⬍0.05 was considered statistically significant.
Use of the BT test during a 5-month period preceding
discontinuation of the test was quantified and summa-
rized by the specialty and ordering physicians. Question-
naires for surgical and nonsurgical specialists were con-
structed and acted as the basis for interviews with a
subset of the physicians who had ordered the BT test
during the 5 months before discontinuation of the BT test
(Table 1). To assess the reliability of the clinicians’ per-
ceptions, and to assess the potential impact of a change in
use of hemostatic agents after discontinuation of the BT
test on bleeding complications, we reviewed the use of the
drug 1-deamino-8-d-arginine vasopressin (DDAVP) dur-
ing the 5 months immediately before and after discontin-
uation of the BT test. Administration of DDAVP repre-
sents the primary intervention taken by clinicians at our
institution to counteract platelet function defects (e.g.,
uremia, Von Willebrand’s disease) in patients who are at
risk and are scheduled to undergo invasive procedures, or
who are actively bleeding. The total number of patients
who received platelet-active doses of DDAVP (⬃0.3 ␮g/kg),
as well as the number of DDAVP-treated patients with
Table 1. Questionnaire used for clinician interviews.
Non-surgical Services
Has discontinuation of the BT test changed your practice with
respect to organ biopsy? Assessment of bleeding risk for other
procedures or operations?
Have you seen an increase in bleeding complications from organ
biopsy (or other procedures) since discontinuation of the BT test?
Subjective? Objective?
Have you changed your pre-biopsy assessment or treatment of
patients undergoing biopsy? If so, how?
Are biopsies being postponed, for lack of the BT result, until
pharmacologic, transfusion or dialysis therapy is completed? Is
this a change from prior practice?
Surgical Services
Has discontinuation of BT test changed your practice with respect to
pre-surgical assessment of bleeding risk? If so, how?
Postponement of surgery?
Discontinuation of drug affecting platelet function?
Dosing of drug to improve platelet function (e.g., steroids,
DDAVP)
Transfusion of platelets, RBC’s or cryoprecipitate prior to
surgery?
Change in Pre-surgical assessment of risk?
Platelet function assessment?
Assessment of prior history (personal or family) of bleeding?
Consult from Hematology or Transfusion Service.
Have you seen an increase in bleeding complications since
discontinuation of the BT test? Subjective? Objective?
Have you increased your use of blood products since
discontinuation of the BT test?
coincidental renal insufficiency, and the fraction of pa-
tients who received DDAVP before renal biopsy were
recorded for the two 5-month intervals.
This study was conducted under a protocol approved
by the University of Utah Institutional Review Board.
Interviews were conducted with the written consent of
the clinicians.
Results
The algorithm for evaluating the risk of excessive bleed-
ing in patients with a personal or family history of
bleeding is presented in Fig. 1. Note that, in the absence of
a carefully elicited bleeding history, no further work-up
was recommended. The number of procedures, and the
clinical services ordering them, in the 5 months before
discontinuation of the BT test, are summarized in Table 2.
The primary services ordering the BT test were Nephrol-
ogy, Obstetrics/Gynecology, Hematology (primarily one
clinician), Pulmonary (transplant), and Neurosurgery.
The predominance of subspecialty services requesting BT
tests reflects a relatively conservative use of the BT test.
Evaluation of platelet function in patients before kidney
biopsy (n ⫽ 19), before liver biopsy (n ⫽ 10), or with a
history of NSAID ingestion (n ⫽ 5) included ⬎40% of the
indications for a BT test. Before discontinuation of the BT
test, only one clinician (Neurosurgery) contacted the
Laboratory after the notification memo was distributed. A
 Clinical Chemistry 47, No. 7, 2001 1207

specifically requested a platelet transfusion to correct an

Table 2. Clinical services at the University of Utah Hospital
requesting the BT test.
Service Number of proceduresa (%)
Nephrology
Hematology
OB/Gyn
Pulmonary
Neurosurgery
Gastroenterology
Pediatrics
Othersb
Unspecified
Total
a
Number of BT tests performed in the 5 months immediately prior to
discontinuation of the test.
b
Otolaryngology (3), General Surgery (3), Emergency Medicine (3), Urology (2),
General Internal Medicine (2), Orthopedic Surgery (1), Neurology (1), Cardiology
(1), and Family Practice (1).
second clinician (Neurosurgery) contacted the Laboratory
⬃12 months after discontinuation, asking for a BT test. In
both cases, the concern was related to determining the
risk for excessive bleeding in patients taking NSAIDS.
These were the only inquiries regarding the BT test
fielded by the Laboratory in the 24 months since discon-
tinuation of the test.
An evaluation of the ordering frequency of platelet-
aggregation tests revealed no increase during the 5
months after discontinuation of the BT test compared
with either the 5-month period immediately before dis-
continuation or the same 5-month interval (February
through June) in 1998 (Table 3). Likewise, there was no
significant difference in the number of nonirradiated,
nonleukocyte-reduced platelet units transfused in the 5
months before discontinuation of the BT test vs the 5
months after, nor for the months of February through
June of 1998 vs 1999 (Table 3). A review of the indications
for platelet transfusions administered to patients with
platelet counts in the reference interval over the first 4.5
months after discontinuation of the BT test revealed only
cardiopulmonary bypass as an indication related to a
known, acquired platelet function defect. BT tests had not
historically been ordered for this group of patients before
discontinuation of the test. In addition, no clinicians have
acquired platelet function defect in a patient undergoing a
procedure.
The rate of postprocedural hemorrhage or hematoma
18 (22.5)
10 (12.5)
in patients in the Major Surgery Risk Pool has remained in
10 (12.5) statistical control (⬍3␴) in the 5 months since discontin-
6 (7.5) uation of the BT test (Fig. 2). In fact, the monthly variation
5 (6.3) in complication rates has remained within 2␴ of the mean.
4 (5) Therefore, there is no evidence that discontinuation of the
4 (5) BT test has led to an increase in bleeding complications. In
17 (21.2) addition, the Laboratory has not been notified of any
6 (7.5) cases of excessive bleeding in surgical patients where
80 discontinuation of the BT test was thought to play a
significant role in the clinical outcome.
To assess the possible impact of discontinuation of the
BT test on bleeding complications in a nonsurgical patient
population known to be predisposed to abnormal platelet
function, and who undergo invasive diagnostic proce-
dures, we evaluated the complication rate of kidney
biopsy procedures immediately before and after discon-
tinuation of the BT test. Demographic variables and
potentially significant risk factors were compared for
patients undergoing biopsy in the before and after
5-month intervals. BUN was higher in the patients biop-
sied after discontinuation of the BT test, but did not
correlate with the decrease in the postbiopsy hematocrit
(rs ⫽ 0.154; P ⫽ 0.224). On the other hand, significantly
more renal allografts were biopsied after the BT test was
no longer available (Table 4); and an analysis of the mean
decrease in hematocrit for allografts vs native kidneys, for
all eligible patients (n ⫽ 66), revealed a significantly larger
mean decrease in the hematocrit (P ⫽ 0.008) for allograft
kidneys (mean ⫾ SD ⫽ 0.96 ⫾ 2.31; n ⫽ 31) vs native
kidneys (mean ⫾ SD ⫽ 2.68 ⫾ 2.72; n ⫽ 35). This probably
accounts for the lower mean decrease in hematocrit
demonstrated by the patients biopsied after discontinua-
tion of the BT test, although the difference did not reach
statistical significance (Table 4). There was no significant
difference in the incidence of RBC transfusion after biopsy
between the two time intervals (Table 4). Three of the
transfused patients were female, and three had a native
kidney biopsied. In addition, three of the four patients
had prebiopsy prothrombin time and/or platelet count

Table 3. Clinician practice behavior before and after discontinuation of the BT test.
BT test available, BT test available, BT test unavailable,
2/98–6/98 9/98–1/99 2/99–6/99 P
Monthly platelet units transfused 44.8 ⫾ 14.8a 42.0 ⫾ 13.9a 41.6 ⫾ 8.6a 0.687b
0.958c
Total platelet-aggregation studies 17 9 9 NTd
Total patients receiving DDAVP NA 24 10 NT
Uremic patients receiving DDAVP NA 22 8 NT
a
Mean ⫾ SD.
b
Student t-test: 2/98 – 6/98 vs 2/99 – 6/99.
c
Student t-test: 9/98 –1/99 vs 2/99 – 6/99.
d
NT, not tested; NA, not assessed.
1208 Lehman et al.: Bleeding Time Test

[General/Trauma, Neurosurgery (2 ), Ear Nose and

Fig. 2. Rate of postprocedural hemorrhage or hematoma for patients in
the Major Surgery Risk Pool for the 12 months before and 5 months
after discontinuation of the BT test.
Data are plotted as a “P” control chart with standard upper and lower control
limits of ⫾ 3␴. The months during which the BT test was no longer available are
indicated by an asterisk and by BT Discontinued on the graph. F, observed
monthly rate; ⫺ ⫺ ⫺, standard upper and lower confidence limits; ——, mean
rate.
values that were abnormal. A BT test was not performed
on any of the four patients.
Because clinicians might not have actively reported
bleeding complications in patients thought to have had a
platelet-function disorder, 18 (35%) physicians who had
ordered a BT test in the 5 months before discontinuation
of the test were interviewed using the questionnaires
presented in Table 1. Of the 4 surgeons interviewed
Throat], none had changed his or her practice as a result
of discontinuation of the BT test. They reported that they
did not significantly change their preoperative work-ups
(with the exception of elimination of the BT test), post-
poned a surgical procedure, experienced an increase in
bleeding complications, or increased the use of blood
products. However, the concern about preoperative eval-
uation of neurosurgical patients taking NSAIDs or aspirin
was raised again.
The responses from the nonsurgical clinicians were
consistent with those of the surgical faculty. They re-
ported that they had not modified their practices or noted
an increase in bleeding complications, except for one
physician who would consider the use of DDAVP in lung
transplant patients requiring a lung biopsy when the
patient’s BUN was ⬎16.1 mmol/L. This clinician reported
one episode of excessive bleeding associated with a lung
biopsy in a transplant patient with uremia. The clinician
stated that a preprocedural BT test would normally have
been ordered, but would not have affected the timing of
the biopsy.
A review of the administration of therapeutic (i.e.,
platelet-active) doses of DDAVP to patients before and
after discontinuation of the BT test failed to show an
increase in the overall use (Table 3), the use in patients
with renal disease (Table 3), or the administration imme-
diately before renal biopsy (Table 4). This is consistent
with the clinicians’ perception of their practices. Admin-
istration of DDAVP represents the primary intervention
taken by clinicians at our institution to counteract platelet
function defects (e.g., uremia, Von Willebrand’s disease)
in patients who are at risk for bleeding because of

Table 4. Comparison of renal biopsy patient variables.

Variable BT test available 9/98–1/99 BT test unavailable 2/99–6/99 P
Age, years 38.3 ⫾ 16.8a (n ⫽ 37)b 39.8 ⫾ 19.9a (n ⫽ 44) 0.718c
Sex, % male 56.8 (n ⫽ 37) 63.6 (n ⫽ 44) 0.528d
Creatinine, ␮mol/L 180 (60–1010);e (n ⫽ 35) 180 (60–1370);e (n ⫽ 42) 0.556f
BUN, mmol/L 10.9 (3.2–39.6);e (n ⫽ 34) 14.6 (4.3–50.0);e (n ⫽ 41) 0.025f
Prebiopsy hematocrit 0.361 ⫾ 0.070a (n ⫽ 31) 0.330 ⫾ 0.067a (n ⫽ 36) 0.064c
Prebiopsy hematocrit ⬍0.300, % 25.8 (n ⫽ 31) 33.3 (n ⫽ 36) 0.502d
Coagulopathy, % 27.8 (n ⫽ 36) 24.4 (n ⫽ 41) 0.735d
Thrombocytopenic, % 14.7 (n ⫽ 34) 11.4 (n ⫽ 35) 0.734g
DDAVP,h % 10.8 (n ⫽ 37) 4.5 (n ⫽ 44) 0.404g
Allografts, % 27.5 (n ⫽ 37) 72.5 (n ⫽ 44) 0.001d
⌬ hematocrit 0.023 ⫾ 0.028a (n ⫽ 31) 0.015 ⫾ 0.025a (n ⫽ 35) 0.225c
Transfused, % 5.4 (n ⫽ 37) 4.5 (n ⫽ 44) 1.000g
a
Mean ⫾ SD.
b
All n values in the table represent the total number of patients available for analysis for the time interval.
c
Student t-test.
␹ test.
d 2

e
Median (range).
f
Mann-Whitney test.
g
Fisher’s exact test.
h
DDAVP administered prior to biopsy.
 Clinical Chemistry 47, No. 7, 2001
scheduled invasive procedures or who are actively bleed-
ing.
Discussion
The role of the BT test in defining the risk of pathologic
bleeding in “susceptible” populations remains controver-
sial. Most literature surveys that have objectively evalu-
ated the performance of the BT test have found no clinical
utility for this test (4 – 6 ). Our study failed to identify a
clinically significant, negative impact of discontinuing
availability of the test at our Medical Center. Objective
measures of complication rates validated the clinicians’
perceptions of unchanged patient outcomes. The rate of
postoperative bleeding in patients undergoing proce-
dures in the Major Surgery Risk Pool before and after
discontinuation of the BT test did not change. This sug-
gests that there was not a dramatic increase in bleeding
episodes attributable to unidentified platelet dysfunction.
However, a minor increase cannot be ruled out because
ICD-9 codes do not allow stratification of events by root
cause. Likewise, we were unable to identify an increase in
bleeding complications after renal biopsy, whether mea-
sured by a decrease in hematocrit or a requirement for
RBC transfusions. These findings must be tempered,
however, by the potential bias of an increased proportion
of allograft kidneys in the cohort of patients after discon-
tinuation of the BT test. It is interesting to note that our
finding of a smaller decrease in the postprocedural he-
matocrit in transplant patients is at odds with other
reports of greater arteriovenous fistula formation (16 ) and
a greater decrease in the hematocrit of biopsied transplant
kidneys (17 ). Transplantation may have been acting as a
surrogate for an unmeasured variable because renal trans-
plant patients did not differ significantly from nontrans-
plant patients for any other characteristic that was as-
sessed (data not shown).
Because clinical outcomes could have been influenced
by a change in clinician behavior after discontinuation of
the BT test, we attempted to assess potentially significant
changes in clinical practices. Interviews with physicians
who had previously ordered the BT test failed to identify
an impact of discontinuation of the test on their clinical
practice and/or patient outcomes. None believed he or
she had significantly altered preprocedural work-ups or
periprocedural transfusion practices as a result of discon-
tinuation of the BT test; and, with one exception, none
could recall a case where he or she felt a BT test might
have led to an improved outcome. This included 60% of
the practicing nephrologists, many of whom had fre-
quently used the BT test before kidney biopsy. In fact, the
Nephrology Division did not oppose discontinuation of
the BT test when they were surveyed before elimination of
the test. Their consensus opinion was that the question-
able utility of the test did not justify the necessary
maintenance. Consistent with the clinicians’ perceptions
of their practice, utilization of platelet-aggregation tests,
administration of DDAVP, and mean monthly platelet
1209
transfusions did not increase after discontinuation of the
BT test. In addition, an evaluation of the indications for
platelet transfusion in patients with normal counts at the
time of transfusion failed to identify increased platelet
utilization in patients at risk for acquired platelet dysfunc-
tion. Therefore, we found no evidence of significant
changes in clinician practice that could have biased our
assessment of patient outcomes before and after discon-
tinuation of the BT test.
Although our neurosurgeons expressed concerns
about their ability to identify patients at risk for bleeding
among those taking NSAIDs, they did not use platelet-
aggregation tests as a substitute and did not prescribe
platelet transfusions for patients on NSAIDs. The lack of
reported bleeding complications in their patients may
suggest that the actual risk attributable to platelet dys-
function is less than that perceived. The degree of risk of
bleeding complications in patients undergoing invasive
procedures who have ingested aspirin or NSAIDs is
somewhat controversial, even in neurosurgical patients,
where even a small bleed may be catastrophic (18 –21 ). A
review of the literature revealed very few case reports of
presumed NSAID-induced bleeding episodes in patients
undergoing neurological procedures, and no true pro-
spective studies could be found (22–25 ). In fact, one
author recently concluded that there was insufficient
evidence of increased risk of bleeding complications in
patients who received aspirin therapy before central neu-
ral blockade to recommend a routine preprocedural BT
test (20 ). In addition, two experimental studies have
demonstrated that the skin template BT is, in fact, not
predictive of visceral BTs in subjects treated with aspirin.
Treatment of healthy human volunteers with aspirin
prolonged the skin BT test, but did not prolong a gastric
mucosal BT test performed during endoscopy (26 ); and
an animal model of neurosurgical bleeding found that
aspirin prolonged the skin BT test, but not a brain BT test
induced by a standard incision (27 ).
Use of the BT test in clinical medicine relies on the
premise that clinically relevant bleeding episodes, during
or after invasive procedures, can be predicted a priori in
susceptible patients, leading to appropriate preventative
action. Whereas the BT test may predict a higher bleeding
complication rate in populations at risk, it is of limited
assistance in evaluating an individual patient, because of
low sensitivity and specificity. In fact, Lee and Lamila (13 )
suggest that more harm, in the form of delayed surgery or
administration of unnecessary or inappropriate blood
products, may be incurred as a result of false-positive BT
tests. The almost uniform lack of clinician concern about
discontinuation of the BT test at our institution suggests
that the test was not considered to be essential for
practice, and possibly, unnecessarily obstructive. Many
clinicians stated that they had ordered the test because it
was available or because it was how they were trained.
The apparent lack of clinically significant adverse
events attributable to discontinuation of the BT test would
1210 Lehman et al.: Bleeding Time Test
suggest the following: (a) in the absence of a personal or
family history of bleeding, excessive bleeding attributable
solely to platelet dysfunction must be relatively uncom-
mon [e.g., in 10 years, Lee and Lamila (13 ) detected only
two mild cases of von Willebrand’s disease through the
use of routine preoperative BT tests], and (b) the BT test
does not add significant positive predictive values to
routine coagulation parameters and a platelet count, even
in patients with the potential for undiagnosed, clinically
significant platelet dysfunction (e.g., renal failure, liver
failure). This may be attributable, in part, to relatively
recent changes in the clinical management of these pa-
tients. For example, patients suffering from chronic renal
failure commonly receive erythropoietin therapy, causing
an increase in their hematocrit, which is thought to
improve primary hemostasis (28 ). Administration of con-
jugated estrogens, DDAVP, or cryoprecipitate before pro-
cedures performed on uremic patients can correct a pro-
longed BT test, and presumably acquired platelet
dysfunction (29 –32 ), although a cause-and-effect relation-
ship with decreased hemorrhage has never been estab-
lished. For patients with hepatic failure, wider availability
of the transjugular (vs percutaneous) route for hepatic
biopsies in patients with coagulopathies and/or throm-
bocytopenia may have reduced the risk of serious bleed-
ing complications resulting from this procedure (33, 34 ).
Our assessment of the impact of discontinuation of the
BT test could have missed infrequent events of bleeding
that might have been avoided by performing the test. In
fact, one of the four patients who required RBC transfu-
sions after renal biopsy had normal coagulation values, as
well as a normal platelet count. A BT test was not
performed, although it was available at the time. It is
interesting to speculate whether performance of a BT test
would have impacted the clinical outcome. However,
because these episodes are rarely, if ever fatal (17, 35–39 ),
and because the clinical effectiveness of the interventions
routinely taken to correct an abnormal BT have never
been rigorously tested, the cost and significant effort
required to maintain competency in the face of dwindling
test numbers, and the repercussions of numerous false-
positive BT test results, are not clinically justified.
In conclusion, our study failed to identify a clinically
significant, negative impact of discontinuing availability
of the BT test at a busy, tertiary-care, university medical
center serving adult and neonatal patient populations.
The role for new in vitro methods for assessing platelet
function needs to be well defined, and the tests must be
clinically validated before clinical acceptance is attained
(40 ).
References
1. Bick RL. Platelet function defects associated with hemorrhage or
thrombosis [Review]. Med Clin North Am 1994;78:577– 607.
2. Duke WW. The relation of blood platelets to hemorrhagic disease.
Description of a method for determining the bleeding time and
coagulation time and report of three cases of hemorrhagic disease
relieved by transfusion. JAMA 1910;55:1185–92.
3. Ivy AC, Nelson D, Bucher G. The standardization of certain factors
in the cutaneous “venostasis” bleeding time technique. J Lab Clin
Med 1941;26:1812–22.
4. Rodgers RPC, Levin J. A critical reappraisal of the bleeding time.
Semin Thromb Hemost 1990;16:1–20.
5. Lind SE. The bleeding time does not predict surgical bleeding
[Review]. Blood 1991;77:2547–52.
6. Peterson P, Hayes TE, Arkin CF, Bovill EG, Fairweather RB, Rock
WA, et al. The preoperative bleeding time test lacks clinical benefit
[Review]. Arch Surg 1998;133:134 –9.
7. Fugh-Berman A. Herb-drug interactions [Review]. Lancet 2000;
355:134 – 8.
8. Mcdonald R. Bleeding time. Br J Anaesth 1992;69:329.
9. Sutor AH. The bleeding time in pediatrics [Review]. Semin Thromb
Hemost 1998;24:531– 43.
10. Mattix H, Singh AK. Is the bleeding time predictive of bleeding
prior to a percutaneous renal biopsy [Review]? Curr Opin Nephrol
Hypertens 1999;8:715– 8.
11. Boberg KM, Brosstad F, Egeland T, Egge T, Schrumpf E. Is a
prolonged bleeding time associated with an increased risk of
hemorrhage after liver biopsy? [Review]. Thromb Hemost 1999;
81:378 – 81.
12. Messmore HL, Godwin J. Medical assessment of bleeding in the
surgical patient [Review]. Med Clin North Am 1994;78:625–34.
13. Lee S, Lamila CM. Reducing the number of routine preoperative
screening bleeding time tests [Letter]. Arch Pathol Lab Med
1991;115:1088 –9.
14. Rodgers RPC, Levin J. Reducing the number of routine preopera-
tive screening bleeding time tests [Letter]. Arch Pathol Lab Med
1991;115:1089.
15. Swift JA. Introduction to modern statistical quality control and
management. Delray Beach, FL: St. Lucie Press, 1995:350pp.
16. Stiles KP, Yuan CM, Chung EM, Lyon RD, Lane JD, Abbott KC.
Renal biopsy in high-risk patients with medical diseases of the
kidney [Review]. Am J Kidney Dis 2000;36:419 –33.
17. Mendelssohn DC, Cole EH. Outcomes of percutaneous kidney
biopsy, including those of solitary native kidneys. Am J Kidney Dis
1995;26:580 –5.
18. Belisle S, Hardy JF. Hemorrhage and the use of blood products
after adult cardiac operations: myths and realities. Ann Thorac
Surg 1996;62:1908 –17.
19. Self DD, Bryson GL, Sullivan PJ. Risk factors for post-carotid
endarterectomy hematoma formation. Can J Anaesth 1999;46:
635– 40.
20. Wildsmith JAW, McClure JH. Aspirin, bleeding time and central
neural block [Letter]. Br J Anaesth 1993;70:112.
21. Ferraris VA, Gildengorin V. Predictors of excessive blood use after
coronary artery bypass grafting. A multivariate analysis. J Thorac
Cardiovasc Surg 1989;98:492–7.
22. Merriman E, Bell W, Long DM. Surgical postoperative bleeding
associated with aspirin ingestion. Report of two cases. J Neuro-
surg 1979;50:682– 4.
23. Heye, N. Is there a link between acute spinal epidural hematoma
and aspirin? Spine 1995;20:1931–2.
24. de Swiet M, Redman CW. Aspirin, extradural anaesthesia and the
MRC Collaborative Low-dose Aspirin Study in Pregnancy (CLASP).
Br J Anaesth 1992;69:109 –10.
25. Palmer JD, Sparrow OC, Iannotti F. Postoperative hematoma: a
5-year survey and identification of avoidable risk factors. Neuro-
surgery 1994;35:1061– 4.
26. O’Laughlin JC, Hoftiezer JW, Mahoney JP, Ivey KJ. Does aspirin
prolong bleeding from gastric biopsies in man? Gastrointest
Endosc 1981;27:1–5.
 Clinical Chemistry 47, No. 7, 2001
27. MacDonald JD, Remington BJ, Rodgers GM. The skin bleeding
time test as a predictor of brain bleeding time in a rat model.
Thromb Res 1994;76:535– 40.
28. Anand A, Feffer SE. Hematocrit and bleeding time: an update
[Review]. South Med J 1994;87:299 –301.
29. Boyd GL, Diethelm AG, Gelman S, Lagner R, Laskow D, Deierhoi
M, Barber WH. Correcting prolonged bleeding during renal trans-
plantation with estrogen or plasma. Arch Surg 1996;131:160 –5.
30. Weigert AL, Schafer AI. Uremic bleeding: pathogenesis and ther-
apy [Review]. Am J Med Sci 1998;316:94 –104.
31. Mannuccio PM, Remuzzi G, Pusineri F, Lombardi R, Valsecchi C,
Mecca G, Zimmerman TS. Deamino-8-D-arginine vasopressin
shortens the bleeding time in uremia. N Engl J Med 1983:308:
8 –12.
32. Janson PA, Jubelirer SJ, Weinstein MJ, Deykin D. Treatment of the
bleeding tendency in uremia with cryoprecipitate. N Engl J Med
1980;303:1318 –22.
33. Sawyer AM, McCormick PA, Tennyson GS, Chin J, Dick R, Scheuer
PJ, et al. A comparison of transjugular and plugged-percutaneous
liver biopsy in patients with impaired coagulation. J Hepatol
1993;17:81–5.
34. Bonnard P, Vitte R, Barbare J, Denis J, Stepani P, Di Martino V, et
1211
al. Is bleeding time measurement useful for choosing the liver
biopsy route? The results of a pragmatic, prospective multicentric
study in 219 patients [Letter]. J Clin Gastroenterol 1999;29:
347–9.
35. Mahal AS, Knauer MC, Gregory PB. Bleeding after liver biopsy.
West J Med 1981;134:11– 4.
36. Piccinino F, Sagnelli E, Pasquale G, Giusti G. Complications
following percutaneous liver biopsy. A multicenter retrospective
study on 68,276 biopsies. J Hepatol 1986;2:165–73.
37. Little AF, Ferris JV, Dodd GD, Baron RL. Image-guided percutane-
ous hepatic biopsy: effect of ascites on the complication rate.
Radiology 1996;199:79 – 83.
38. Benfield MR, Herrin J, Feld L, Rose S, Stablein D, Tejani A. Safety
of kidney biopsy in pediatric transplantation: a report of the
Controlled Clinical Trials in Pediatric Transplantation Trial of
Induction Therapy Study Group. Transplantation 1999;67:544 –7.
39. Hergesell O, Felten H, Andrassy K, Kuhn k, Ritz E. Safety of
ultrasound-guided percutaneous renal biopsy-retrospective analy-
sis of 1090 consecutive cases. Nephrol Dial Transplant 1998;
13:975–7.
40. Carr ME Jr. In vitro assessment of platelet function [Review].
Transfus Med Rev 1997;11:106 –15.