An Algorithm to Aid in the Investigation of Thalassemia

Trait in Multicultural Populations

Thomas L. Kiss, MD, FRCPC; Mahmoud A. M. Ali, PhD, FRCPC; Mitchell Levine, MD, MSc, FRCPC; John D. Lafferty, ART

Context.The differentiation between iron deficiency

and a thalassemia syndrome is an important consideration
in the investigation of microcytic anemia.
Objective.An established statistical method was used
to demonstrate the importance of considering ethnic background in combination with mean cell volume (MCV) in
the investigation of b-thalassemia trait in a multicultural
urban population.
Design.Posttest probabilities for b-thalassemia trait
were calculated using likelihood ratios for various microcytic MCV ranges in conjunction with published pretest
probabilities for b-thalassemia trait based on ethnic background.
Setting.Regional hemoglobinopathy laboratory, St Josephs Hospital, Hamilton, Ontario, Canada.
Patients.Patient data were derived from a previously

published study. The original study cohort consisted of 789

patients aged 18 years or older who had an MCV less than
80 fL and were referred for routine complete blood count
during a 6-month period.
Main Outcome Measures.Posttest probabilities.
Results.Simplified tables for the determination of posttest probabilities for b-thalassemia trait in individual patients based on ethnic background and MCV are provided.
An algorithm to assist in determining when thalassemia investigations are indicated is presented.
Conclusions.A high index of suspicion based on ethnic
background and low MCV can provide increased sensitivity
and specificity for the detection of thalassemia trait in centers with multicultural populations similar to the study
population.
(Arch Pathol Lab Med. 2000;124:13201323)

ficiency with accuracy requires the use of specialized laboratory procedures, as well as a knowledgeable interpretation of the results.2 Investigators have long sought a reliable mathematical index based on routine complete
blood count results to help differentiate between microcytosis due to iron deficiency and that due to a thalassemia syndrome. Previous work conducted in this laboratory3 established that the use of the mean cell volume
(MCV) alone was as reliable or more reliable than any
other published index. This study demonstrated that the
use of an MCV less than 72 fL as a predictor of thalassemia trait resulted in a sensitivity of 0.88 and a specificity
of 0.84.3 The clinical usefulness of this approach, however,
is limited by the fact that 12% of patients with thalassemia
minor will be missed, while 16% without thalassemia minor, a potentially large number of individuals, would be
processed for unnecessary tests. These factors raise questions regarding the cost-effectiveness of such a strategy in
multicultural populations in which the percentage of individuals with an ethnic background associated with thalassemia is relatively small.4 This situation exists in many
jurisdictions and suggests that a more specific approach
to the detection of thalassemia trait is indicated here.5 The
purposes of this article are to demonstrate the importance
of considering ethnic background in combination with
MCV in the investigation of b-thalassemia trait and to present an algorithm to assist in determining when specific
thalassemia investigations are indicated in multicultural
populations.
METHODS

he thalassemia syndromes are a group of genetic abnormalities that cause deficient production of the aor b-globin chains of the hemoglobin (Hb) molecule. These
abnormalities are prevalent in patients with Mediterranean, African, Middle Eastern, Asian Indian, Chinese, and
Southeast Asian ancestry. In the heterozygous form (thalassemia trait), the syndromes present as a thalassemia minor phenotype characterized by microcytosis with varying
degrees of anemia. In the homozygous or compound heterozygous state (thalassemia major), the abnormalities
cause severe anemia, resulting in fetal death or a lifelong
dependency on transfusion. When investigating a microcytic anemia, the differentiation between thalassemia trait
and iron deficiency is an important consideration. This differentiation takes on added importance in patients who
are pregnant or planning a pregnancy, since thalassemia
trait is the carrier form for the more severe thalassemia
major phenotypes. Table 1 lists common hemoglobinopathies that present as thalassemia minor in the heterozygous state and as thalassemia major in the homozygous
or compound heterozygous state.1
To differentiate between thalassemia trait and iron deAccepted for publication March 31, 2000.
From the University Health Network/The Princess Margaret Hospital,
Toronto, Ontario, Canada (Dr Kiss); St Josephs Hospital, Hamilton, Ontario, Canada (Dr Ali and Mr Lafferty); and Father Sean OSullivan Research Centre, St Josephs Hospital, Hamilton, Ontario, Canada (Dr
Levine).
Reprints: John Lafferty, ART, Hemoglobinopathy Laboratory, St Josephs Hospital, 50 Charlton Ave E, Hamilton, Ontario, Canada L8N
4A6.
1320 Arch Pathol Lab MedVol 124, September 2000

Data used come from a previous study conducted in Hamilton,

Ontario, Canada. This region has a broad-based multicultural
MCV and Ethnic Background in ThalassemiaKiss et al

Table 1. Common Hemoglobinopathies, Which Present as Thalassemia Minor in the Heterozygous State and as
Thalassemia Major in Homozygous or Compound Heterozygous States
Thalassemia Trait
Genotype

Thalassemia Major
Genotype

a-Thalassemia

22/aa

22/22

b-Thalassemia
Hemoglobin E
Hemoglobin Lepore

b0/b
bE/b
bLepore/b

Hemoglobinopathy

b0/bLepore

b0/b0
b0/bE
or bLepore/bLepore

Table 2. Formulae Used for the Calculation of

Posttest Probabilities for b-Thalassemia Trait*
b-Thalassemia
Trait Present

b-Thalassemia
Trait Absent

Within mean cell

volume range
A
B
Outside mean cell
volume range
C
D
* Positive Likelihood Ratio 5 [A/(A 1 C)]/[B/(B 1 D)].
Pretest Odds 5 Pretest Probability/1 2 Pretest Probability.
Posttest Odds 5 Pretest Odds 3 Positive Likelihood Ratio.
Posttest Probability 5 Posttest Odds/1 1 Posttest Odds.

Table 3. Positive Likelihood Ratios for b-Thalassemia

Trait at Various Mean Cell Volume (MCV) Ranges
Calculated From the Previous Study Data
MCV Range, fL

Positive
Likelihood Ratio

5559.9
6064.9
6569.9
7074.9
7579.9

54.19
18.79
3.61
0.34
0.09

population; the majority of residents are of British, Northern European, North American Aboriginal, and Northern Asian ethnic
background. The prevalence of b-thalassemia trait in these ethnic
groups is low. An important proportion of the population (23%),
however, is of ethnic backgrounds in which b-thalassemia trait
is common. These populations include people of Italian (12%),
Portuguese (2.8%), other Mediterranean (1.9%), Chinese (1.6%),
Asian Indian (1.3%), African (1.1%), Southeast Asian (1.1%), and
Greek (1%) descent.6 The original study consisted of 789 patients
18 years of age or older with an MCV less than 80 fL who were
referred for routine complete blood count during a 6-month period. Seventy-seven patients were excluded because of insufficient
sample volumes. Forty of the remaining 712 patients were diagnosed with thalassemia minor, 31 with b-thalassemia trait, 8 with
a-thalassemia trait, and 1 with HbE trait.3 This article deals with
the b-thalassemia trait data only, as insufficient numbers of patients with other hemoglobinopathies were identified in the original study. Positive likelihood ratios for b-thalassemia trait were
calculated for various MCV ranges from the previous study data.
Relative to any range of MCV values, a group of subjects may be
divided into 4 mutually exclusive subgroups, A through D (Table
2). Likelihood that a patient with b-thalassemia trait is within
that MCV range is A/(A 1 C), and likelihood that a patient without b-thalassemia trait is within that MCV range is B/(B 1 D).
The positive likelihood ratio for that range is then the ratio of
these 2 ratios, the first equation of Table 2. This ratio is independent of the prevalence of b-thalassemia trait in the group. Subsequent equations in Table 2 combined this neutral ratio with the
appropriate prevalence factor for b-thalassemia trait based on
ethnic background to produce the final probability of b-thalassemia trait for that ethnic group within a given MCV range.
Arch Pathol Lab MedVol 124, September 2000

Thalassemia Major
Syndrome

Hydrops fetalis
b-Thalassemia major
b-Thalassemia major
b-Thalassemia major

Clinical Significance

Fetal death, difficult

pregnancy
Transfusion dependent
Transfusion dependent
Transfusion dependent

Table 4. Prevalence (Pretest Probability) of bThalassemia Trait in Different Populations

Population

Prevalence, %

African American
Algeria
Bulgaria
Northern
Southern
China
Canton Area
Cyprus
Ghana
Greece
Arta District
Crete
Mainland
Rhodes
India
Ahom
Assam
Bangalen
Bhanishali
Lohana
Saraswat NW
Saraswat West
Sindhi near Bombay
Italy
North and Central
Po delta
Sardinia
Sardinia (Southern)
Southern (includes Sicily)
Kurdish
Jewish population
Lebanon
Liberia
Malta
Nigeria
Pacific Islands
Lower range
Upper range
Pakistan
Karachi
Patham
Thailand
Northern
Overall
Turkey
Southwest

1.4
3.0
0.1
9.0
2.0
15.0
1.5
10.0
7.0
7.0
25.0
1.0
5.0
3.7
15.0
13.6
4.4
3.5
8.0
1.3
13.0
12.6
28.5
10.0
20.0
4.0
9.0
3.5
0.8
2.0
20.0
1.5
4.0
10.0
4.8
1.7

These equations represent the application of conditional (Bayesian) probability. Prevalence data for b-thalassemia trait were obtained from a published compilation of studies.7 All formulae
used for these calculations are outlined in Table 2.8

RESULTS
The positive likelihood ratios calculated for various MCV
ranges are listed in Table 3. Pretest probabilities derived for
a variety of ethnic backgrounds are shown in Table 4. They
MCV and Ethnic Background in ThalassemiaKiss et al 1321

Table 5. Posttest Probability of b-Thalassemia Trait*

Mean Cell
Volume,
fL

Pretest Probability
1.0%

2.5%

5.0%

7.5%

10.0%

12.5%

15.0%

17.5%

20.0%

22.5%

25.0%

27.5%

30.0%

5559.9
35.4
58.2
74.0
6064.9
16.0
32.5
49.7
6569.9
3.5
8.5
16.0
7074.9
0.3
0.9
1.8
7579.9
0.1
0.2
0.5
* All values reported as percentages.

81.5
60.4
22.6
2.7
0.7

85.8
67.6
28.6
3.6
1.0

88.6
72.9
34.0
4.6
1.3

90.5
76.8
38.9
5.7
1.6

92.0
79.9
43.4
6.7
1.9

93.1
82.4
47.4
7.8
2.2

94.0
84.5
51.2
9.0
2.5

94.8
86.2
54.6
10.2
2.9

95.4
87.7
57.8
11.4
3.3

95.9
89.0
60.7
12.7
3.7

An algorithm for the investigation of microcytosis utilizing posttest probabilities that are
based on ethnic background and mean cell
volume (MCV). 1In microcytic patients who
are pregnant, the requirement for rapid turnaround of test results may indicate the need
for a thalassemia investigation regardless of
posttest probability. 2Thalassemia investigation includes hemoglobin (Hb) A2 level, Hb
electrophoresis, HbF level, HbH inclusion
body screen, and serum ferritin. Interpretation
of these tests is reviewed elsewhere.14

include Italy (various regions), Greece (various regions),

Malta, Cyprus, Bulgaria (north and south), Turkey, Lebanon, India (various regions), Pakistan (various regions),
China (Kanton), Pacific Islands, Algeria, Liberia, Ghana,
Nigeria, and Thailand, as well as the probabilities for Kurdish Jews and African Americans. These probabilities include
many of the ethnic backgrounds encountered in the study
population. The posttest probabilities calculated using the
pretest probability, likelihood ratio, and MCV are listed in
Table 5. For the purpose of simplification, MCV values are
divided into ranges of 4.9 fL, while the pretest probabilities
are rounded to the nearest 2.5%. The posttest probability
of b-thalassemia trait for an individual patient can be established from Table 5 using the patients pretest probability, based on ethnic background (from Table 4) and the patients MCV. For example, an Italian patient with a Sardinian
background and an MCV of 63 fL has a pretest probability
of 12.6% and a posttest probability of 72.9%. With an MCV
of 76 fL, the same patient would have a greatly reduced
posttest probability of 1.3%. The Figure presents an algorithm for the investigation of microcytosis utilizing posttest
probabilities based on ethnic background and MCV.
COMMENT
Microcytic anemias are among the most common types
of anemia encountered by physicians in general hospitals
and outpatient clinics.9 The differential diagnosis includes
1322 Arch Pathol Lab MedVol 124, September 2000

iron deficiency, thalassemia minor resulting from a- or bthalassemia trait or from variant hemoglobins such as
HbE and Hb Lepore, anemia of inflammation, congenital
sideroblastic anemia, and lead poisoning.10,11 The 2 most
common causes of microcytic anemia are iron deficiency
and thalassemia minor.12 It is imperative to distinguish
between these 2 diagnoses appropriately. Iron deficiency
requires a diagnostic workup to determine the underlying
cause of the deficiency, medical intervention to correct it,
and replacement iron therapy. Thalassemia minor, on the
other hand, does not require the aforementioned medical
intervention but has implications for prenatal diagnosis
and genetic counseling.1 The experience of this laboratory
is that an investigation for thalassemia trait is often not
considered when investigating microcytosis in the multicultural population of Hamilton. This can lead to unnecessary medical intervention for iron deficiency and prevent access to prenatal diagnostic services for high-risk
patients who are pregnant or contemplating pregnancy.
Table 5 demonstrates that the probability of b-thalassemia
trait increases with the patients pretest probability, based
on ethnicity and the degree of microcytosis. The probability of b-thalassemia trait in patients with a modest pretest probability of 5% increases to almost 50% with an
MCV of 60 to 64.9 fL. In some patients from high-risk
ethnic backgrounds, the presence of an MCV of 55 to 59.9
fL results in a posttest probability of greater than 90%.
MCV and Ethnic Background in ThalassemiaKiss et al

Although posttest probabilities for other thalassemia syndromes were not calculated, the b-thalassemia trait posttest probabilities demonstrate compelling evidence of the
importance of considering a patients ethnic background
in the investigation of microcytosis. The absence of an ethnic background in which thalassemia is common does not
exclude thalassemia but makes other causes of microcytosis more likely. The combination of a low MCV and an
ethnic background in which thalassemia trait is prevalent
strongly indicates the need for specific thalassemia testing.
This is especially true in patients who are pregnant or are
planning a pregnancy, as these patients are candidates for
nondirective genetic counseling if they are at risk of conceiving an infant with thalassemia major.
In summary, failure to investigate for thalassemia trait
in high-risk patients can lead to unnecessary and potentially harmful investigation and therapy for iron deficiency.13 In patients who are pregnant or are contemplating
pregnancy, a failure to investigate for thalassemia trait can
prevent access to prenatal diagnostic services for thalassemia major. In Hamilton, the majority of the population
consists of ethnic backgrounds in which thalassemia trait
is uncommon, and an investigation for thalassemia trait
in these patients reduces effective laboratory utilization.
Many cities have multicultural populations similar to the
population of Hamilton, where an important minority of
the population includes ethnic backgrounds in which thalassemia is prevalent. The statistical case presented here
demonstrates that a high index of suspicion based on ethnic background and low MCV is a relevant approach to
the effective investigation of thalassemia trait in these centers. Use of the algorithm presented in the Figure should

Arch Pathol Lab MedVol 124, September 2000

provide increased sensitivity and specificity for the detection of thalassemia trait in these jurisdictions.
This study was funded by the Department of Laboratory Medicine, St. Josephs Hospital, Hamilton, Ontario, Canada.
References
1. Weatherall DJ. The thalassemias. In: Williams Hematology. 5th ed. New
York, NY: McGraw-Hill Inc; 1995:581615.
2. Lee GR. Microcytosis and the anemias associated with impaired hemoglobin synthesis. In: Wintrobes Clinical Hematology. 9th ed. Philadelphia, Pa: Lea
& Febiger; 1993:791807.
3. Lafferty J, Crowther MA, Ali MA, Levine M. The evaluation of various mathematical RBC indices and their efficacy in discriminating between thalassemic
and non-thalassemic microcytosis. Am J Clin Pathol. 1996;106:201205.
4. Le Gales C, Galacteros F. Economic analysis of neonatal screening for drepanocytosis in metropolitan France. Epidemiol Sante Publique. 1994;42:478492.
5. Sasi K, Sanderson D, Eydoux P, et al. Prenatal diagnosis for inborn errors of
metabolism and haemoglobinopathies: the Montreal Childrens Hospital experience. Prenat Diagn. 1997;17:681685.
6. Dickson L, Heale J, Chambers L. Determinants of health: language and
ethnicity. In: Fact Book on Health in Hamilton Wentworth. 3rd ed. Hamilton,
Ontario, Canada: McMaster University Faculty of Health Sciences; 1994:121
122.
7. Weatherall DJ, Clegg JB. The distribution and incidence of beta thalassemia
genes in different populations. In: Thalassemia Syndromes. 3rd ed. London, England: Blackwell Scientific Publications; 1981:295309.
8. Fletcher RH, Fletcher SW, Wagner EH. Diagnosis. In: Clinical Epidemiology:
The Essentials. 3rd ed. Baltimore, Md: Williams & Wilkins; 1996:4374.
9. Cash JM, Sears DA. The anemia of chronic disease: spectrum of associated
diseases in a series of unselected hospitalized patients. Am J Med. 1989;87:638
644.
10. Hoffbrand AV, Pettit JE. Erythropoiesis and anemia. In: Essential Hematology. 3rd ed. Oxford, England: Blackwell Science; 1993:2728.
11. Wintrobe MM, Lukens JM, Lee GR. The approach to the patient with anemia. In: Wintrobes Clinical Hematology. 9th ed. Philadelphia, Pa: Lea & Febiger;
1993:715744.
12. Ravel R. Basic hematologic tests and classification of anemia. In: Clinical
Laboratory Medicine. 6th ed. St Louis, Mo: Mosby; 1995:921.
13. Ali M. Hemoglobinopathies in the Hamilton region, II: thalassemia trait
and iron therapy. Can Med Assoc J. 1975;32:701702.
14. Lafferty J. The laboratory diagnosis of a thalassemia. Can J Med Lab Sci.
1998;60:183189.

MCV and Ethnic Background in ThalassemiaKiss et al 1323