Clinical Chemistry 47: 63-66, 2001;
(Clinical Chemistry. 2001;47:63-66.)
© 2001 American Association for Clinical Chemistry, Inc.
Effect of Urgent Clinician Notification of Low Hemoglobin Values
Paul Frooma,1,
Tayser Mahameed1,
Rosa Havis1 and
Mira Barak1
1
Central Laboratory of Haifa and Western Galilee, Clalit Health Services, Nesher, Israel.
a Address correspondence to this author at: Hematology Laboratories, Central Laboratory of Haifa and Western Galilee, Clalit Health Services, Nesher, Israel. Fax 972-4-8209094;
paulf{at}ioh.org.il.
 |
Abstract
|
|---|
Background: Urgent clinician notification of low hemoglobin values
is a common practice. The effect of such notification for samples
obtained during an office visit is uncertain.
Methods: We notified the attending physicians in 100 consecutive
cases of outpatients with hemoglobin <80 g/L. We reviewed the medical
charts of these patients with their personal physicians 1–2 months
later. We considered transfusions unnecessary only when given to a
clinically stable patient with probable iron deficiency anemia in the
absence of chronic diseases that produce the anemia.
Results: Overall, 47 (47%) of the patients were referred to the
emergency room (ER), and 31 of the 47 (66%) were transfused.
Increasing age and decreasing hemoglobin were associated with increased
referral of patients to the ER. Six of 31 transfusions were
unnecessary. An emergency transfusion was clearly indicated in only 3
of the 31 patients; the other patients were clinically stable. A
modified notification strategy excluding microcytic samples >50 g/L
would have detected these three patients. Over the follow-up period, no
patient not referred to the ER died from complications of anemia, nor
were any patients hospitalized with acute symptoms of low hemoglobin.
Conclusion: In some settings, urgent notification of medical
outpatient clinics of patients with hemoglobin <80 g/L may lead to
unnecessary ER visits.
|
Introduction
|
|---|
Red blood cell transfusions generally are recommended in
hospitalized patients if the hemoglobin is <60 g/L
(1)(2). Delay in appropriate use of red cell
transfusion might put the patient at risk for hypoxic complications.
Although the risk of acute events attributable to anemia probably
occurs only with a hemoglobin concentration <50 g/L
(2)(3), patients with ischemic heart disease
might be at risk with higher values. Because the clinical signs of
anemia are unreliable (4)(5), contacting the
physician immediately when the blood count is unexpectedly low could be
prudent. Immediate physician notification of a low hemoglobin obtained
during an office visit, however, might lead to adverse consequences.
Inappropriate notification squanders human resources and penalizes the
physician with an untimely information overload
(6)(7). A telephone call, in itself, might
suggest the need for immediate attention, leading to unwarranted
referrals to the emergency room (ER), inconvenience and anxiety for
patient and family, and even unnecessary blood transfusions
(2)(8). We are unaware of any outcome studies
that demonstrate either the benefits or risks of this practice.
|
Materials and Methods
|
|---|
Our clinical laboratory provides results of blood counts during
the day of collection (maximum delay is 9 h). The results arrive
by electronic mail to the clinics and are checked by the physician
within 1–3 days. Our laboratory’s policy is immediate notification of
results of the complete blood count if the hemoglobin value is <80 g/L
and no record of previous similar values are found in the laboratory
archives. We notified the attending physicians in 100 consecutive cases
of outpatients with hemoglobin <80 g/L and reviewed the medical charts
of these patients with them 1–2 months later. We considered
transfusions unnecessary only when given to a clinically stable patient
with probable iron deficiency anemia in the absence of chronic diseases
that produce the anemia.
Over a 3-month period, we performed 48 972 blood counts using one of
the following analyzers: CD3500 (Abbott), Sysmex9000 (Sysmex
Corporation), and H-2 (Miles Inc.). The performance of these analyzers
is harmonized, and we use internal and external (United Kingdom
National External Quality Assessment Scheme) quality-control programs.
All tests with a hemoglobin values <80 g/L (n = 183) were
repeated using another counter. If discordant results were obtained,
the test was repeated until a consistent result was obtained. Repeat
testing revealed one technical error, in which the hemoglobin
concentration was actually 121 g/L. The other 182 tests showed almost
identical results. In 81 cases, a previous test from the patient was
similar to the current test, and no attempt was made to contact the
physician. We attempted to notify the attending physician in 101
consecutive cases of patients with hemoglobin <80 g/L. One to 2 months
after the original telephone call, 100 patient medical charts were
reviewed. In one case, we were unable to contact the physician.
Patients were divided into groups by age and disease categories to
determine the frequency of unnecessary ER referrals and transfusions.
There were three age groups: children (<18 years), young adults
(18–49 years), and older adults (
50 years). Diseases were
categorized into five groups: cancer, myelodysplastic syndrome, chronic
renal failure, other diseases, and probable iron deficiency (mean
corpuscular volume 
70 fL and no reduction in either white blood cell
or platelet counts). After the review, we considered as
"necessary" those acute referrals in which the patient had an acute
bleeding episode, was unstable as judged by vital signs or symptoms
(e.g., chest pain), or had hemoglobin <50 g/L, where there might be an
increased risk for acute events (2)(3). Because
the indication for transfusion in patients with chronic disease is
subjective, we defined unnecessary transfusions as those given to a
clinically stable patient with probable iron deficiency in the absence
of chronic diseases that might be the cause of the anemia.
statistical analysis
The rate of anemia (hemoglobin <80 g/L) per 1000 tests was
calculated by dividing the number of cases of anemia for each age group
by the total tests done per age group. ANOVA for parametric data and
the 
2 test for nonparametric data were used to
determine statistical significance of differences in univariate
analyses. All significant variables were entered into a logistic
regression model. Variables with P values >0.2 were removed
and subsequently reentered sequentially to determine whether they
improved the model.
|
Results
|
|---|
Females predominated in those >17 years (Table 1
). The rates of anemia were highest in females of all ages and
in males 50 years. The overall rate was 3.7 cases of hemoglobin <80
g/L per 1000 tests, 1.2 per 1000 of cases with hemoglobin <70 g/L, and
only 0.2 per 1000 for hemoglobin <60 g/L. Only one patient had
hemoglobin <50 g/L, and he received a transfusion. Forty-seven percent
of the patients were referred to the ER. Referral rates were higher in
those with lower hemoglobin and in older patients (Table 2
). Hemoglobin of 70–79 g/L was associated with transfusion in
12 of 66 patients (18%) compared with 19 of 34 patients (56%) with
hemoglobin <70 g/L. Of 11 patients with hemoglobin <60 g/L, 10 were
sent to the ER, and all were transfused.
Most patients with cancer, myelodysplastic syndrome, chronic renal
failure, and other serious diseases were 50 years. In these patients,
referral rates to the ER were 47–71%, with a high rate of
transfusions (Table 3
). On the other hand, of patients with iron deficiency, only
33% were sent to the ER, and 14% were transfused. There were only two
patients with an acute bleed, and they were neither unstable in the
clinic nor when they reached the hospital. Of those referred to the ER,
66% (31 of 47) were transfused.
On logistic regression (Table 4
), both increasing age and decreasing hemoglobin influenced
decisions to refer patients to the ER. There was also a trend for more
ER referrals of males. None of the other variables improved the model.
Of those referred to the ER, the decision to transfuse was also
strongly influenced by age (odds ratio, 1.45 per 10 years; 95%
confidence interval, 1.13–1.86) and hemoglobin concentrations (0.14
per g/L; 95% confidence interval, 0.05–0.35).
analysis of the decision to transfuse
Of 11 patients under age 18, 3 received a transfusion. One
3-year-old child with chronic hemolytic anemia, usually transfused with
a hemoglobin 
60 g/L, received red cells with a hemoglobin of 69 g/L.
This was not an emergency, and a transfusion could have been given more
conveniently in an outpatient clinic. There were also two females 14
and 16 years of age who were clinically stable and transfused with
hemoglobin values of 66 and 60 g/L, low mean corpuscular volumes, and
red blood cell counts >3 x 1012/L. These
two patients were clinically stable and could have been given a trial
of oral iron.
Among patients 18–49 years of age, five females were transfused. In
four, the transfusion was not indicated because they had probable iron
deficiency anemia and could have been given a trial of oral iron. A
40-year-old female, with chronic renal failure and a hemoglobin of 53
g/L, also received a transfusion. She was clinically stable, and this
transfusion too was not an emergency.
In patients >50 years, two had acute bleeds: one had epistaxis, the
other an acute gastrointestinal bleed. The patient with the
gastrointestinal bleed and a hemoglobin of 76 g/L was transfused and
operated on, and a gastrointestinal non-Hodgkin lymphoma was
discovered. The patient with epistaxis had a hemoglobin of 65 g/L,
received a transfusion, and was hospitalized for closer follow-up.
These two patients might have benefited from the acute referral because
notification of the laboratory test result helped the physician correct
a clinical evaluation error. Another patient 78 years of age had a
hemoglobin of 49 g/L, and although the patient was clinically stable,
an emergency blood transfusion was indicated. It is uncertain whether
delayed referral would have led to a different outcome in the three
patients. All other patients transfused had chronic disease, no acute
symptoms, and no need for emergency transfusion, but probably required
transfusions on an outpatient basis.
follow-up
There were no deaths among patients not referred to the ER during
the 1- to 2-month follow-up period, and none were hospitalized with
acute symptoms related to low hemoglobin values. Two patients with
metastatic cancer died in the hospital, and one 87-year-old bedridden
patient died of unspecified causes; anemia did not contribute to these
three deaths. Those with iron deficiency received oral iron, and those
with chronic disease were transfused as indicated in outpatient
specialty clinics. Two females were referred for surgery to remove
uterine myomas.
|
Discussion
|
|---|
Our principal finding was that physician notification of
hemoglobin <80 g/L in an outpatient setting leads to unnecessary ER
visits, without any definite benefits.
In 1972, Lundberg (9) introduced the concept of critical
values to identify patients with "a pathophysiological state at such
variance with normal as to be life-threatening unless something is done
promptly and for which some corrective action could be taken". In
outpatients, another purpose of the critical value might be to identify
patients who require hospitalization to prevent a condition from
becoming life-threatening. Although no patients were identified
with a life-threatening condition, two patients had acute bleeding and
another had hemoglobin <50 g/L.
Clinical diagnoses of anemia often are incorrect because physical
examination alone is inaccurate. Pallor of the conjunctiva, tongue,
palm, or nail beds is only 66% sensitive and 68% specific in
distinguishing children with hemoglobin concentrations of 50–80 g/L
from those with higher values (5). Thus, the likelihood
ratio of a positive physical examination is only 2, and laboratory
confirmation is needed for accurate diagnosis of anemia.
In the absence of acute blood loss, the hemoglobin concentration alone
does not predict the need for transfusion. For those with chronic
anemia, response to pharmacological agents and the patient’s symptoms
need to be considered (2). From animal studies and anecdotal
patient reports, it is generally believed that there is an increased
risk of mortality in stable patients only when the hemoglobin is <50
g/L (2)(3), but in patients with coronary
stenosis, the hemoglobin concentration associated with an increased
risk might be as high as 75 g/L. In our study, only one patient had
hemoglobin <50 g/L, and none of the patients were clinically unstable.
Although transfusion in most of our patients might have been indicated,
values that trigger emergency transfusion and the low hemoglobin
critical limit value need not be identical (10). The
patients could have been given a transfusion as outpatients, and
perhaps some of these patients would have responded to pharmacological
agents.
There were at least six patients who were inappropriately transfused.
They had microcytic anemia, were clinically stable, and should have
been given iron replacement instead. Inappropriate use of transfusions
by hospital personnel has been reported (11)(12)(13). This is
attributable to widespread deficiencies in physicians’ knowledge of
transfusion indications (13) and insufficient use of
guidelines for red blood cell transfusions that have been shown to
decrease the number of transfusions by 10–30%
(11)(12). Once a patient has been referred for
emergency transfusion, it might be more difficult for the ER physician
to decide against such intervention.
An American national survey of trauma and medical centers found that
the range of hemoglobin critical limits was extremely wide (40–120
g/L), with a mean value of 66 g/L (7). The College of
American Pathologists Q-Probes Committee abandoned an attempt
to establish, by consensus, a national standard critical values list
(6), but suggested a critical value of <70 g/L as a
starting point for setting specific values. In a hospital setting,
surveyed physicians indicated that the hemoglobin critical limit could
be lowered from <80 g/L to <70 g/L (10). The uncomplicated
course of the 44 (of 66) patients in our study with hemoglobin
concentrations of 70–79 g/L who were not referred to the ER supports
the College of American Pathologists recommendations for outpatients.
If we had used that strategy, only the patient with an acute
gastrointestinal bleed would not have been hospitalized acutely.
An alternative strategy might be to leave the critical value of 80 g/L
but exclude those with blood counts suggestive of chronic iron
deficiency unless the hemoglobin value is <50 g/L. In our setting,
this would reduce the number of telephone calls for critical hemoglobin
values from 100 to 47 per 48 972 tests (1 per 1000) without missing
the two patients with the acute bleeding. For those with a hemoglobin
of 50–79 g/L, the telephone call also might include the caveat that,
if the patient is stable, without chest pain, and not bleeding acutely,
emergency intervention is not indicated. Perhaps fewer patients would
be referred to the ER if physicians were given this additional
information. Our findings should be extrapolated to other settings with
caution because variation in medical practice and patient behavior
might lead to different results.
|
References
|
|---|
-
Stehling LC, Doherty DC, Faust RJ, Greenburg AG, Harrison CR, Landers DF, et al. Practice guidelines for blood component therapy: a report by the American Society of Anesthesiologists Task Force on Blood Component Therapy. Anesthesiology 1996;84:732-747.[ISI][Medline]
[Order article via Infotrieve]
-
Simon TL, Alverson DC, AuBuchon J, Cooper S, DeChristopher PJ, Glen GC, et al. Practice parameter for the use of red blood cell transfusions: developed by the Red Blood Cell Administration Practice Guideline Development Task Force of the College of American Pathologists. Arch Pathol Lab Med 1998;122:130-138.[ISI][Medline]
[Order article via Infotrieve]
-
Viele MK, Weiskopf RB. What can we learn about the need for transfusion from patients who refuse blood? The experience with Jehovah’s Witnesses. Transfusion 1994;34:396-401.[ISI][Medline]
[Order article via Infotrieve]
-
Hetzel TM, Losek JD. Unrecognized severe anemia in children presenting with respiratory distress. Am J Emerg Med 1998;16:386-389.[Medline]
[Order article via Infotrieve]
-
Luby SP, Kazembe PN, Redd SC, Ziba C, Nwanyanwu OC, Hightower KAW, et al. Using clinical signs to diagnose anaemia in African children. Bull World Health Organ 1995;73:477-482.[ISI][Medline]
[Order article via Infotrieve]
-
Emancipator K. Critical values: ASCP practice parameter. Am J Clin Pathol 1997;108:247-253.[Medline]
[Order article via Infotrieve]
-
Kost GJ. Critical limits for urgent clinician notification at US medical centers. JAMA 1990;263:704-707.[Abstract]
-
Nicholls MD. Transfusion: morbidity and mortality. Anaesth Intensive Care 1993;21:15-19.[Medline]
[Order article via Infotrieve]
-
Lundberg GD. When to panic over abnormal values. Med Lab Observ 1972;4:47-54.
-
Lum G. Should the transfusion trigger and hemoglobin low critical limit be identical?. Ann Clin Lab Sci 1977;27:130-134.
-
Soumerai SB, Salem-Schatz S, Avorn J, Casteris CS, Ross-Degnan R, Popovsky MA. A controlled trial of educational outreach to improve blood transfusion practice. JAMA 1993;270:961-966.[Abstract]
-
Brandis K, Richards B, Ghent A, Weinstein S. A strategy to reduce inappropriate red blood cell transfusion. Med J Aust 1994;160:721-722.[ISI][Medline]
[Order article via Infotrieve]
-
Stehling L, Luban NLC, Anderson KC, Sayers MH, Attar LS, Leitman SF, et al. Guidelines for blood utilization review. Transfusion 1994;34:438-448.[ISI][Medline]
[Order article via Infotrieve]
Top
Abstract
Introduction
