Building a Score to Discriminate Between Iron

Deficiency Anemia and Beta Thalassemia Trait

Author name: Fatimah Suliman Mohammad Aljebaly

Author Degree:Board Certified of Clinical Pathology Programat King Saud University

Name of the Institution: Department of Basic Medical Sciences, Unaizah Colloege of Medicine and

Medical Sciences, Qassim University, Saudi Arabia

Address of Author: Uthman Ibn Affan Rd, Unayzah, Al Qassim 56436, Saudi Arabia

Email: [email protected]

Telephone #/ Fax #:+966-554000222

1
Abstract:

Objective: Iron deficiency anemia (IDA) and thalassemia are the most common causes of anemia,

However the differentiation between Beta thalassemia trait (βTT) and IDA is not an easy process

and need more sophisticated procedure like hemoglobin electrophoresis, High Performance Liquid

Chromatography (HPLC), genetic and molecular studies, which are time and money consuming.

However, many equations are made using the hemoglobin, MCV, RDW, MCH and R.B.Cs count to

discriminate between IDA andβTT. An example of these equations are Metzer, Sirdah, Green and

king, Shine and Lal, and Ehsani also many other equations are present, but No one is superior to the

others, In our paper we are trying to build a score system to differentiate between IDA and βTT.

Methodology: We used the most 5 equations with high sensitivity and specificity and give a point

either 1 or zero for each result either IDA or βTT and if the final score is more than 3 it is most

properly IDA and if less than 3 it could be βTT.

Results: We applied this method for 50patientswhoarediagnosedasIDA orβTT and get the result

with good confidence interval (9.9 t0 21) and high precision.

Conclusion: We recommend using this score with the help of Artificial Intelligence (AI) as an easy,

cheap, fast, more specific and more sensitive tool to discriminate between IDA and βTT.

Key Words: Anemia, Beta, Deficiency, Iron, and Thalassemia Trait

2
 Introduction:

Iron deficiency anemia (IDA) is the most common cause of anemia (1), also Thalassemia is not are

type of anemia in the Middle East. (2) One variety of β Thalassemia, which is β Thalassemia trait

(βTT), has a blood picture similar in many aspects to Iron deficiency anemia. In both types of

anemia, we get low hemoglobin level (Hb), Low volume of red blood corpuscles (MCV), increased

degree of variation in cell volume (RDW) and low hemoglobin content (MCH). This makes the

differentiation between them a complicated process.

Although there is similarity in blood picture, the treatment is different and prescription of iron to β

Thalassemia trait may be not useful or even harmful. (7)

To differentiate between them, hemoglobin electrophoresis, High Performance Liquid

Chromatography (HPLC), genetic and molecular techniques are used. However, these methods are

expensive and time consuming. (8.9.10) In order to get rapid and easy methods to differentiate

between them many equations and formulae are represented using the R.B.Cs, MCV, MCH and RDW

to discriminate between β Thalassemia trait and Iron deficiency anemia.

The most popular known methods are shown in table (1):

Table (1)

Name Equation

Green and king MCV*MCVX RDW/Hb*100

Shine and Lal MCV*MCV*MCH/100

Mentzer index MCV/R.B.Cs

Ehsani MCV –(10*R.B.Cs)

Sirdah MCV– R.B.Cs –(3*Hb)

3
 Ricerca RDW/R.B.Cs

MDHL (MCH/MCV)*R.B.Cs

England and Fraser MCV-(5*Hb)-RBCs-3.4

RDWI MCV*RDW/R.B.Cs

No method has a sensitivity and specificity of 100%, but variable degrees of sensitivity and

specificity are present between these methods. (5) Thus, the aim of this study is to put a score

system. This score system is expected to have higher sensitivity and specificity.

This score system depends on the sum of the above methods and uses this score to discriminate

between βTT and IDA. We are going to use the higher equations for sensitivity and specificity as the

base of the score system, This score will be applied on 50 cases of IDA and BTT and will be tested to

see if the score can be used as a method to discriminate between iron deficiency anemia and beta

thalassemia trait.

We are going to use the following equations to build the score:

1- MCV/R.B.Cs

If > 13 equal 1 in the score system

If <13 equal 0 in the score system

2. MCV-(5*hb)-R.B.Cs–3.4

If>0 equal 1 in the score system

If<0 equal 0 in the score system

3. MCV- (10*R.B.Cs)

If>15 equal 1 in the score system

If<15 equal 0 in the score system

4
 4. MCV-R.B.Cs – (3XHb)

If >27equal 1 in the score system

If<27equal 0 in the score system

5. MCV*RDW/R.B.Cs

If>220 equal 1 in the score system

If<220 equal 0 in the score system

At the end, if the total score Number>3, the diagnosis will be IDA and if the total score<3, it will be βTT. So, we

are going to apply this hypothesis on 30 patients who are diagnosed with IDA and 20 patients who are

diagnosed with βTT and see if it is true or not.

Method

We retrospectively analyzed patients with microcytic anemia (meanage: 11-41years).Samples were

obtained from 50 patients with no clinical symptoms of acute or chronic inflammation, or infectious

diseases. We collected our data from laboratory in King Khalid University Hospital, in Riyadh.

Table 2

IDA group

Hb R.B.Cs MCV MCH RDW

1 6.8 3.65 58 17.2 16.9

2 7.5 3.82 60 19.6 19.3

3 8.8 4.01 64 21.1 18.2

4 7.2 3.71 56 19.4 17.2

5 10.6 4.10 73 25.8 19.4

6 8.7 3.02 63 28.8 18.3

5
7 7.0 3.78 60 18.5 16.7

8 9.1 3.51 64 25.9 17.0

9 7.7 3.92 59 19.6 18.8

10 9.3 3.92 66 23.7 17.7

11 10.9 4.10 72 26.5 16.6

12 8.9 3.12 67 28.5 19.6

13 7.3 3.45 55 21.1 19.5

14 9.5 3.56 68 26.6 18.1

15 10.0 4.20 71 23.8 17.1

16 7.6 3.32 54 22.8 18.7

17 10.2 4.00 72 25.5 17.3

18 10.7 4.52 74 25.1 17.0

19 9.0 3.32 62 27.1 17.4

20 7.9 3.12 59 25.3 19.1

21 8.7 3.91 63 22.2 17.9

22 9.6 3.78 67 25.3 18.4

23 9.7 3.99 66 24.3 19.1

24 7.1 3.01 57 23.5 19.0

25 10.5 4.32 75 24.3 16.8

26 9.3 3.78 67 24.6 17.9

27 7.4 3.21 58 23.0 19.4

28 10.7 4.12 73 25.9 16.9

6
 29 11.2 4.71 77 23.7 16.0

30 11.3 4.61 76 24.5 14.0

30 patients of them were diagnosed with IDA. Diagnosis was done after clinical and laboratory tests

including CBC, Serum Iron, TIBC, and Ferritin.

The above table shows various hematological parameters for a group of individuals with Iron

Deficiency Anemia (IDA). Hb: Hemoglobin (g/dL), R.B.Cs: Red Blood Cell Count (million/μL),

MCV: Mean Corpuscular Volume (fL), MCH: Mean Corpuscular Hemoglobin (pg), RDW: Red Cell

Distribution Width (%).

Table 3

The criteria of iron deficiency anemia

Serum iron for Men 75–150mcg/dL

Serum iron for Women 60–140mcg/dL

Total iron-binding capacity 250–450mcg/dL

Ferritin 30–300ng/mL

Transferrin saturation 20–50%

The above table provides normal reference ranges for various iron-related blood tests. These values are

used to assess an individual's iron status and help diagnose iron-deficiency anemia. Serum iron: This

measures the amount of iron circulating in the blood. (Men=75-150 mcg/dL, Women= 60-140

mcg/dL). Total iron-binding capacity (TIBC): This measures the maximum amount of iron that

transferrin, a protein in the blood, can bind to (Normal range: 250-450 mcg/dL). Ferritin: This is a

7
protein that stores iron in the body (Normal range: 30-300 ng/mL). Transferrin saturation: This is the

percentage of transferrin that is bound to iron (Normal range: 20-50%).

Table 4

βTT Group

Hb R.B.Cs MCV MCH RDW

1 7.0 4.34 52 16.1 19.1

2 10.5 5.30 63 19.1 14.2

3 8.7 4.72 58 18.4 15.7

4 8.1 4.56 56 17.7 18.7

5 7.5 4.23 53 17.7 17.3

6 9.4 4.98 59 18.8 14.2

7 8.2 4.12 57 19.9 15.3

8 9.6 4.87 60 19.7 15.9

9 7.9 4.83 53 16.3 17.0

10 10.6 5.23 62 20.2 13.5

11 9.9 4.89 59 20.2 14.2

12 7.2 3.97 55 18.1 18.3

13 9.0 4.73 58 19.0 16.2

14 8.5 4.75 52 17.8 17.6

15 10.3 5.26 61 19.5 13.7

16 7.9 4.57 54 17.2 18.7

17 7.6 4.14 59 18.3 19.3

8
18 11.0 5.50 62 20.0 13.6

19 8.6 4.54 51 18.9 17.1

20 9.2 4.97 59 18.5 16.1

9
The other 20 patients were diagnosed with βTT by doing CBC, Iron Profile and Hemoglobin

electrophoresis. The normal value of HBA2 is less than 3.2%, while 3.2% to 3.6% is considered

borderline, which warrants further investigations. Values are 3.6% to 7% are considered beta

thalassemia carriers. We build the score system by using the previously mentioned 5 formulas, all using

the 5 indices (Hb, R.B.Cs, MCV, MCH, and RDW) for evaluation.

The above table shows various hematological parameters for a group of individuals with Iron

Deficiency Anemia (βTT). Hb: Hemoglobin (g/dL), R.B.Cs: Red Blood Cell Count (million/μL),

MCV: Mean Corpuscular Volume (fL), MCH: Mean Corpuscular Hemoglobin (pg), RDW: Red Cell

Distribution Width (%).

Results:

In IDA patients, the mean value of Hb was 9.01±4.5, mean value of R.B.Cs 3.79±1.70 mean of MCV

65.2±23.77, MCH 23±11.60, and RDW 17.84 ±5.6. While in βTT, the mean value of Hb was 8.84±

4.00, mean of MCV57.15±12.00 and MCH 18.57±4.10 and RDW 16.29± 5.80.

Our results show that the mean Hb in IDA is slightly higher than βTT group, also other parameters of

MCV, MCH and RDW are higher in IDA than that of βTT the only exception is R.B.Cs count which

is more in βTT than IDA group. Table 5 and 6summarize the mean ±standard deviation of the various

hematological parameters obtained from individuals with IDA and βTT.

The data was then used to calculate the 5 ratios outlined in the introduction, and the outputs were

recorded accordingly to each patient. A binary distribution system was then used to assign whether a

case was an IDA or βTT diagnosis as per each ratio, where cases of IDA were assigned the value of 1

and BTT the value of 0.

The results were then used to create a score using the proposed scoring system, and each case was

assigned a test outcome. Example in Patient No.1 where Hb is 6.8, R.B.Cs 3.65, MCV58, MCH17.2

and RDW is 16.9. Calculation was done as follows:

1- MCV/R.B.Cs= if applied 58/3.65=15,>13=1in the score system.

10
 2- MCV-(Hb*5)-R.B.Cs -3.4 = if applied 58-(6.8X5)-3.65- 3.4 = 16.9 which is more than zero,

so the score is set to 1.

3- MCV-(10*R.B.Cs) = 58-(10*3.65) = 21.5so it is more than 15 so the score equals 1.

4- MCV-R.B.Cs-(3*Hb) = 58–3.65-(3*6.8) =33.9 which is morethan27so the score equals1.

5- MCV*6.9/R.B.Cs=58*16.9/3.65=268 which is more than 220 so the score equals1.

After summing all the scores above, the final score is 5. Using the same principle shown in the above

example, we applied the 5 equations to the 50 samples and get the result (table 7), the 30 patients.

Whose are diagnosed with IDA get a score more than 3 while in βTT group 15 patients are less than 3.

We statistically analyzed the result in the following section.

11
Data Analysis

The sample of 50 patients with confirmed cases of IDA and βTT were used in conducting this

analysis. The data consist of blood test results, with the descriptive statistics for each sample of

confirmed cases below (table 5 and 6).

Table 5

Descriptive statistics of IDA group

Measure Hb RBCs MCV MCH RDW

Mean 9.01 3.79 65.20 23.77 17.84

Mode 8.70 3.78 67.00 19.60 16.90

Median 9.05 3.80 65.00 24.30 17.90

Range 4.50 1.70 23.00 11.60 5.60

Standard Deviation 1.40 0.46 6.78 2.89 1.27

Variance 1.95 0.21 46.03 8.34 1.60

Above table shows descriptive statistics for a group of individuals with Iron Deficiency Anemia (IDA).

Hb (Hemoglobin): The average hemoglobin level is 9.01 g/dL, which is below the normal range,

indicating anemia. The standard deviation is 1.40 g/dL, indicating a moderate spread of hemoglobin

levels in the group. RBCs (Red Blood Cell Count): The average red blood cell count is 3.79

million/μL, which is also below the normal range, indicating anemia. The standard deviation is 0.46

million/μL, suggesting a relatively small variation in red blood cell counts. MCV (Mean Corpuscular

Volume): The average red blood cell size is 65.20 fL, which is below the normal range, indicating

microcytosis (small red blood cells). The standard deviation is 6.78 fL, suggesting a moderate variation

12
in red blood cell sizes. MCH (Mean Corpuscular Hemoglobin): The average amount of hemoglobin in

each red blood cell is 23.77 pg, which is within the normal range. However, when combined with the

low MCV, it indicates a hypochromic anemia (pale red blood cells). The standard deviation is 2.89 pg,

suggesting a moderate variation in hemoglobin content per red blood cell. RDW (Red Cell Distribution

Width): The variation in red blood cell size is 17.84%, which is within the normal range. This suggests

that the anemia is primarily microcytic rather than anisocytic (having a wide variation in cell size).

Mode: The most frequent value in the data.8.70 g/dL (Hb), 3.78 million/μL (RBCs), 67.00 fL (MCV),

19.60 pg (MCH), 16.90% (RDW). Median: The middle value in the data when arranged in order.9.05

g/dL (Hb), 3.80 million/μL (RBCs), 65.00 fL (MCV), 24.30 pg (MCH), 17.90% (RDW). Range: The

difference between the largest and smallest values.4.50 g/dL (Hb), 1.70 million/μL (RBCs), 23.00 fL

(MCV), 11.60 pg (MCH), 5.60% (RDW.) Variance: The square of the standard deviation.1.95 g/dL²

(Hb), 0.21 million/μL² (RBCs), 46.03 fL² (MCV), 8.34 pg² (MCH), 1.60%² (RDW). The descriptive

statistics indicate that the individuals in this group have iron-deficiency anemia, characterized by low

hemoglobin levels, microcytosis, and hypochromia.

Table 6

Descriptive statistics of βTT group

Measure Hb RBCs MCV MCH RDW

Mean 8.84 4.72 57.15 18.57 16.29

Mode 7.90 4.97 59.00 17.70 14.20

Median 8.65 4.74 58.00 18.65 16.15

Range 4.00 1.53 12.00 4.10 5.80

Standard Deviation 1.20 0.42 3.70 1.20 1.96

Variance 1.43 0.18 13.71 1.45 3.84

13
The above table shows descriptive statistics for a group of individuals, likely patients receiving Anti-

Tuberculosis Therapy (BTT). Hb (Hemoglobin): The average hemoglobin level is 8.84 g/dL, which is

slightly below the normal range, indicating mild anemia. The standard deviation is 1.20 g/dL,

indicating a moderate spread of hemoglobin levels in the group. RBCs (Red Blood Cell Count): The

average red blood cell count is 4.72 million/μL, which is within the normal range. The standard

deviation is 0.42 million/μL, suggesting a relatively small variation in red blood cell counts. MCV

(Mean Corpuscular Volume): The average red blood cell size is 57.15 fL, which is below the normal

range, indicating microcytosis (small red blood cells). The standard deviation is 3.70 fL, suggesting a

moderate variation in red blood cell sizes. MCH (Mean Corpuscular Hemoglobin): The average

amount of hemoglobin in each red blood cell is 18.57 pg, which is within the normal range. However,

when combined with the low MCV, it indicates a hypochromic anemia (pale red blood cells). The

standard deviation is 1.20 pg, suggesting a moderate variation in hemoglobin content per red blood

cell. RDW (Red Cell Distribution Width): The variation in red blood cell size is 16.29%, which is

within the normal range. This suggests that the anemia is primarily microcytic rather than anisocytic

(having a wide variation in cell size).Mode: The most frequent value in the data.7.90 g/dL (Hb), 4.97

million/μL (RBCs), 59.00 fL (MCV), 17.70 pg (MCH), 14.20% (RDW). Median: The middle value in

the data when arranged in order.8.65 g/dL (Hb), 4.74 million/μL (RBCs), 58.00 fL (MCV), 18.65 pg

(MCH), 16.15% (RDW). Range: The difference between the largest and smallest values.4.00 g/dL

(Hb), 1.53 million/μL (RBCs), 12.00 fL (MCV), 4.10 pg (MCH), 5.80% (RDW). Variance: The square

of the standard deviation.1.43 g/dL², 0.18 million/μL², 13.71 fL², 1.45 pg², 3.84%² (RDW). The

descriptive statistics indicate that the individuals in this BTT group have mild anemia with

microcytosis and hypochromia.

14
Table 7

2x2 Contingency table of Actual condition

Tested
Actual Positive Negative

Positive 30 0

Negative 2 18

The tested out comes were then compared to actual conditions. The table shows a 2x2

contingency table, the table compares the actual condition of individuals to the results of a diagnostic

test. Actual: This column represents the true condition of the individuals, whether they are positive or

negative for the condition being tested. Tested: This column represents the results of the diagnostic test,

whether it was positive or negative. 30 individuals who were actually positive for the condition were

correctly identified as positive by the test (True Positive). 0 individuals who were actually positive

were incorrectly identified as negative by the test (False Negative). 2 individuals who were actually

negative were incorrectly identified as positive by the test (False Positive). 18 individuals who were

actually negative were correctly identified as negative by the test (True Negative).

Table 8

F1 Score

Precision 0.94

Recall 1.00

F1 0.97

15
 To validate the results, two tests were conducted: the F1score, and the diagnostic odds ratio (DOR).

The F-Score uses the precision (ratio of true positives to all predicted positives) and the recall (ratio of

true positives to samples that were meant to be positive). The F1 score represents precision and recall

in one metric using the harmonic mean of both measures. As shown in above table, the results

displayed an F1 score of 0.97, which is on the higher end of precision, as a score of 1.0 indicates

perfect precision of the proposed scoring system.

The table presents the results of a diagnostic test, including the F1 score, precision, and recall.

These metrics are commonly used to evaluate the performance of classification models. Precision=

0.94, this indicates that 94% of the positive predictions made by the test were actually correct. In other

words, out of all the individuals the test predicted as positive, 94% truly had the condition. Recall=

1.00, this indicates that 100% of the individuals who actually had the condition were correctly

identified by the test. In other words, the test did not miss any positive cases. F1 Score= 0.97, the F1

score is a harmonic mean of precision and recall, providing a single metric that balances both

measures. In this case, the F1 score of 0.97 suggests that the test has good overall performance, with

both high precision and recall. The results suggest that the diagnostic test has high sensitivity (ability to

correctly identify positive cases) and specificity (ability to correctly identify negative cases). This is

indicated by the high recall and precision values, respectively. The F1 score further confirms the good

overall performance of the test.

To support our findings, the diagnostic odds ratio was also calculated. Due to having 0 false

negatives, a value of 0.5 was added to all figures in the contingency (table7) which the reasoning for

will be discussed further in the limitations of the data.

16
 Table 9

DOR Metrics

True Positive Rate 0.98

False Negative Rate 0.02

True Negative Rate 0.88

False Positive Rate 0.12

Positive Predictive Value 0.92

False Discovery Rate 0.08

Negative Predictive Value 0.97

False Omission Rate 0.03

The above metrics were calculated to aid with formulating the DOR. Diagnostic Odds Ratio

(DOR) the diagnostic odds ratio is metric used to evaluate the performance of diagnostic tests. It is

calculated as the ratio of the odds of a positive test result in individuals with the condition to the odds

of a positive test result in individuals without the condition. A high DOR indicates a strong association

between the test result and the presence or absence of the condition. Dividing the ratio of true positive

to false positive by the ratio of false positive to false negatives yields a DOR of 451, which can be

interpreted as the scoring system proposed is effective at 451:1. To test for significance, a 95%

confidence interval was calculated which yielded a confidence interval of 9,929 to 21.

The table 9 shows various performance metrics for a diagnostic test. True Positive Rate (TPR)

= 0.98, this represents the proportion of individuals who actually have the condition and were correctly

identified by the test (sensitivity). A high TPR indicates that the test is good at detecting individuals

with the condition. False Negative Rate (FNR) = 0.02, this represents the proportion of individuals who

actually have the condition but were incorrectly identified as negative by the test. A low FNR indicates

that the test is good at avoiding false negatives. True Negative Rate (TNR) = 0.88, this represents the
17
proportion of individuals who do not have the condition and were correctly identified as negative by

the test (specificity). A high TNR indicates that the test is good at avoiding false positives. False

Positive Rate (FPR) = 0.12, this represents the proportion of individuals who do not have the condition

but were incorrectly identified as positive by the test. A low FPR indicates that the test is good at

avoiding false positives. Positive Predictive Value (PPV) = 0.92, this represents the probability that an

individual who tests positive actually has the condition. A high PPV indicates that a positive test result

is a strong predictor of the condition. False Discovery Rate (FDR) = 0.08, this represents the proportion

of positive test results that are actually false.

A low FDR indicates that the test is good at avoiding false positives. Negative Predictive Value

(NPV) = 0.97, this represents the probability that an individual who tests negative does not have the

condition. A high NPV indicates that a negative test result is a strong predictor of not having the

condition. False Omission Rate (FOR) = 0.03. This represents the proportion of individuals who have

the condition but were incorrectly identified as negative by the test. A low FOR indicates that the test is

good at avoiding false negatives. These metrics suggest that the diagnostic test has good performance

in terms of sensitivity, specificity, and predictive values. It is able to accurately identify both

individuals with and without the condition, with relatively low rates of false positives and false

negatives.

18
Table 10

Results of the score system

Patient Hb RBCs MCV MCH RDW Type Score

1 6.8 3.65 58 17.2 16.9 IDA 5

2 7.5 3.82 60 19.6 19.3 IDA 5

3 8.8 4.01 64 21.1 18.2 IDA 5

4 7.2 3.71 56 19.4 17.2 IDA 5

5 10.6 4.1 73 25.8 19.4 IDA 5

6 8.7 3.02 63 28.8 18.3 IDA 5

7 7 3.78 60 18.5 16.7 IDA 5

8 9.1 3.51 64 25.9 17 IDA 5

9 7.7 3.92 59 19.6 18.8I IDA 5

10 9.3 3.92 66 23.7 17.7 IDA 5

11 10.9 4.1 72 26.5 16.6I IDA 5

12 8.9 3.12 67 28.5 19.6 IDA 5

13 7.3 3.45 55 21.1 19.5 IDA 5

14 9.5 3.56 68 26.6 18.1 IDA 5

15 10 4.2 71 23.8 17.1 IDA 5

16 7.6 3.32 54 22.8 18.7 IDA 5

17 10.2 4 72 25.5 17.3 IDA 5

18 10.7 4.52 74 25.1 17 IDA 5

19 9 3.32 62 27.1 17.4 IDA 5

20 7.9 3.12 59 25.3 19.1 IDA 5

21 8.7 3.91 63 22.2 17.9 IDA 5

22 9.6 3.78 67 25.3 18.4 IDA 5

24 7.1 3.01 57 23.5 19.0 IDA 5

25 10.5 4.32 75 24.3 16.8 IDA 5

26 9.3 3.78 67 24.6 17.9 IDA 5

27 7.4 3.21 58 23 19.4 IDA 5

28 10.7 4.12 73 25.9 16.9 IDA 5

29 11.2 4.71 77 23.7 16.0 IDA 5

30 11.3 4.61 76 24.5 14 IDA 5

31 7 4.34 52 16.1 19.1 βTT 2

32 10.5 5.3 63 19.1 14.2 βTT 1

33 8.7 4.72 58 18.4 15.7 βTT 2

34 8.1 4.56 56 17.7 18.7 βTT 3

35 7.5 4.23 53 17.7 17.3 βTT 1

36 9.4 4.97 59 18.8 14.2 βTT 1

37 8.2 4.12 57 19.9 15.3 βTT 4

38 9.6 4.87 60 19.7 15.9 βTT 1

39 7.9 4.83 53 16.3 17 βTT 1

40 10.6 5.23 62 20.2 13.5 βTT 1

41 9.9 4.89 59 20.2 14.2 βTT 1

42 7.2 3.97 55 18.1 18.3 βTT 5

43 9.0 4.73 58 19 16.2 βTT 1

44 8.5 4.75 52 17.8 17.6 βTT 1

45 10.3 5.26 61 19.5 13.7 βTT 1

46 7.9 4.57 54 17.2 18.7 βTT 2

47 7.6 4.14 59 18.3 19.3 βTT 5

20
48 11.o 5.5 62 20 13.6 βTT 0

49 8.6 4.54 51 18.9 17.1 βTT 1

50 9.2 4.97 59 18.5 16.1 βTT 1

21
Data Visualization

Figure 1

The bar graph Figure 1 compares the mean hemoglobin (Hb) levels between two groups: IDA

(Iron Deficiency Anemia) and BTT (Anti-Tuberculosis Therapy). Mean Hb in IDA Group = 9.007

g/dL. Mean Hb in BTT Group = 8.721 g/dL. The graph visually demonstrates that the mean

hemoglobin level is slightly higher in the IDA group compared to the BTT group. This suggests that

individuals in the IDA group may have a less severe anemia compared to those in the BTT group.

However, both groups have mean Hb levels below the normal range, indicating anemia in both.

Figure 2

22
 The bar graph Figure 2 compares the mean red blood cell count (RBCs) between two groups:

IDA (Iron Deficiency Anemia) and BTT (Anti-Tuberculosis Therapy). Mean RBCs in IDA Group =

3.786 million/μL. Mean RBCs in BTT Group = 4.725 million/μL. The graph visually demonstrates that

the mean red blood cell count is higher in the BTT group compared to the IDA group. This suggests

that individuals in the BTT group may have a higher number of red blood cells than those in the IDA

group.

Figure 3

The bar graph Figure 3 compares the mean corpuscular volume (MCV) between two groups:

IDA (Iron Deficiency Anemia) and BTT (Anti-Tuberculosis Therapy). Mean MCV in IDA Group =

65.2 fL. Mean MCV in BTT Group = 57.15 fL. The graph visually demonstrates that the mean MCV is

higher in the IDA group compared to the BTT group. This suggests that the red blood cells in the IDA

group are larger than those in the BTT group.

Figure 4

23
 The bar graph Figure 4 compares the mean corpuscular hemoglobin (MCH) between two

groups: IDA (Iron Deficiency Anemia) and BTT (Anti-Tuberculosis Therapy). Mean MCH in IDA

Group = 23.77 pg. Mean MCH in BTT Group = 18.57 pg. The graph visually demonstrates that the

mean MCH is higher in the IDA group compared to the BTT group. This suggests that the red blood

cells in the IDA group contain more hemoglobin than those in the BTT group.

Figure 5

24
 The bar graph Figure 5 compares the mean red cell distribution width (RDW) between two

groups: IDA (Iron Deficiency Anemia) and BTT (Anti-Tuberculosis Therapy). Mean RDW in IDA

Group = 17.85%. Mean RDW in BTT Group = 16.29%. The graph visually demonstrates that the mean

RDW is slightly higher in the IDA group compared to the BTT group. This suggests that there is a

slightly greater variation in red blood cell size in the IDA group.

Limitations:

25
 The sample size has proved so me limitations, mainly due to the lack of false negatives which

required an adjustment to the contingency table in order to make the DOR ratio meaningful (24).

Further, the confidence interval range can be perceived to be too wide to be representative.

However, the calculated DOR ratio after adjustment falls within the confidence interval, and since

the DOR ratio has no upper bounds, the result is still acceptable. To support the DOR ratio, the F1

score displayed 97% precision. Future studies can refine the results by incorporating a much larger

sample size covering a more diverse demographic.

Discussion:

IDA and βTT are the most common types of hypochromic microcytic anemia present in the

Middle East, and the blood picture is so similar that make the deferential between them not easy,

and to discriminate between the pass through many investigations including SerumIon, TIBC and

ferritin level also we have to measure HBA2 by HPLCor hemoglobin electro phoresis or even use

molecular technology. All these methods are expensive and time consuming. Through history many

scientists tried to use different equations depending on R.B.Cs, Hb, MCV, MCH and RDW.

No equation was better or more accurate than other there was variation in sensitivity and

specificity of these equation, this study propose to make a score to differentiate between IDA and

βTT this score depend on the sum of the most sensitive equations made though many years ago to

discriminate between IDA and βTT. If the result of the equation is showing that the result is giving

the diagnose of IDA, we give it a score of 1 and if the result of the equation is showing that the

patient is βTT we give it a score of 0. Then, we apply this to the five 5 equations and if the result is 3

or more so this is a case of IDA and if the result is less than 3 so it is βTT.

The results for all 30 patients of IDA were 4 or more. While in, βTT group one patient got the score of 4, one

patient had a score of 3 and two patients showing result of 5. This may be due to the combination of IDA and

βTT. However, it is recommended to repeat this score for these 4 patients after treatment of iron deficiency.

26
These results confirm that this method and scoring system are showing high sensitivity and specificity more

than any individual equation and we recommend using this score to differentiate between IDA and βTT. For

future improvements to this system, we hope to use Artificial Intelligence (AI) to make the result of the score

fast and easier.

Conclusion:

It was found statistically that the score system to discriminate IDA from βTT has a higher

sensitivity than the 5 equations when each equation is used alone, and we recommend using of this

score system to discriminate between IDA and βTT.

27
References:

1. D. A. Rathod, A.Kaur, V.Pateletal. “Usefulness of cell counter-based parameters and formulas

in detection of 𝛽-thalassemia trait in areas of high prevalence,” American Journal of Clinical

Pathology, vol.128, no.4, pp.585–589, 2007.

2. E.Urrechaga,L.Borque,and J.F.Escanero,“The role of auto-mated measurement of RBC sub

populations differential diagnosis of microcytic anemia and 𝛽-thalassemia screening,”

American Journal of Clinical Pathology, vol.135,no.3,pp.374–379,2011.

3. Zaini RG.Sickle-cell anemia and Consanguinity among the Saudi Arabian population. Arch

Med.2016;8 (3):3–15.

4. McLean E, Cogswell M, EgliI,etal. Worldwide prevalence of anaemia, WHO Vitamin and

Mineral Nutrition Information System, 1993-2005. Public Health Nutrition 2009; 12:444-54.

5. Mediterr J HematolInfectDiswww.mjhid.org 2015;7.

6. Olivieri NF.The beta-thalassemias.N Eng lJ Med.1999 Jul 08; 341(2):99-109.

7. Goonewardene M, Shehata M, Hamad A. Anaemia in pregnancy. Best Pract Res

ClinObstetGynaecol 2012;26:3-24.

8. Piperno A, Mariani R, Arosio C, et al. Haemochromatosis in patients with beta-thalassaemia

trait. Br J Haematol. 2000;111(3):908-914.

9. Zhao J., Li J., Lai Q., Yu Y. (2020). Combined use of gap-PCR and next-generation sequencing

improves thalassaemia carrier screening among premarital adults in China. J. Clin. Pathol.

10.1136/jclinpath2019-206339

10. Rund D. Thalassemia 2016: modern medicine battles an ancient disease. Am J Hematol 2016;

91: 15– 21

11. Hashemieh M, Timori Naghadeh H, Tabrizi Namini M, Neamatzadeh H, Hadipour Dehshal M.

The Iran Thalassemia prevention program: success or failure? Iran J Ped HematolOncol 2015;

28
 5: 161–6

12. W.C.MentzerJr.,“Differentiation of iron deficiency from thalassaemia trait,”The Lancet, vol.1,

no.7808, p.882,1973.

13. Shineand, S.Lal,“A strategy to detect 𝛽-thalassaemia minor,” The Lancet ,vol.1,

no.8013 ,pp.692–694, 1977.

14. J. M. England and P. M. Fraser, “Differentiation of iron deficiency from thalassaemia trait by

routine blood-count,” The Lancet”, vol.1, no.7801, pp.449–452, 1973.

15. P.C.Srivastava,“Differentiation of thalassemia minor from iron deficiency, ”The Lancet,” vol.2,

pp.154– 155, 1973.

16. R.Greenand, R.King, “A new red cell discriminant incorporating volume dispersion for

differentiating iron deficiency anemia from thalassemia minor, ”Blood Cells”, vol.15, no.3,

pp.481–495, 1989.

17. B.M.Ricerca, S.Storti, G.d’Onofrioetal., “Differentiation of iron deficiency from thalassaemia

trait: a new approach, ”Haematologica”, vol.72, no.5, pp.409–413, 1987.

18. M.Sirdah, I.Tarazi, E.AlNajjar, and R.Al Haddad, “Evaluation of the diagnostic reliability of

different RBC indices and formulas in the differentiation of the 𝛽-thalassaemia minor from iron

deficiency in Palestinian population,”International Journal of Laboratory Hematology,

vol.30 ,no.4, pp.324–330, 2008.

19. M. A. Ehsani, E. Shahgholi, M. S. Rahiminejad, F. Seighali,and A. Rashidi, “A new index for

discrimination between iron deficiency anemia and beta-thalassemia minor: results in 284

patients, ”Pakistan Journal of Biological Sciences, vol.12, no.5, pp.473–475,2009.

20. A. Telmissani, S. Khalil, and T. R. George, “Mean density of hemoglobin per liter of blood: a

new hematologic parameter with an inherent discriminant function,” Laboratory

Haematology”, vol.5,pp.149–152,1999.

29
 Appendix

Informed Consent

Title of the Study:

Building A Score to Discriminate Between Iron Deficiency Anemia and Beta Thalassemia Trait

Purpose of the Study:

This research aims to analyze and compare the hematological data of patients with microcytic anemia

to better differentiate between Iron Deficiency Anemia (IDA) and Beta Thalassemia Trait (βTT) , we are

trying to build a score system to differentiate between IDA and βTT. This will contribute to the

development of a score system for improving the diagnostic accuracy of these conditions.

Procedures:

The study will collect data from patients’ previous medical records stored in the laboratory database.

No direct contact with patients will be involved, and no additional testing will be required from

participants. All data will be used anonymously, ensuring the confidentiality of personal information.

Voluntary Participation:

Participation is voluntary, and patients may choose not to have their data included in the study

without any impact on their medical care. Since this is a retrospective study, informed consent is

assumed unless patients request otherwise.

Confidentiality:

All data will be de-identified, and no information will be used to reveal the identity of the participants.

Only the research team will have access to the anonymized data.

Risks and Benefits:

There are no direct risks to participants as no new testing or procedures will be performed. The

potential benefit is the improvement in the diagnostic process for microcytic anemia.

Contact Information:

30
For any queries or concerns regarding the study, you may contact [Fatimah Suliman] at [Email:

[email protected]]. Alternatively, you can reach out to the Ethics Committee at Qassim University

for more information.

Consent Statement:

I have read and understood the information provided above, and I agree to allow the researchers to

use my anonymized data for the purpose of this study.

o Yes

o No

Signature of Participant: Date

----------------- -----------------

Data Collection Tool

We collected our data from laboratory in King Khalid University Hospital, in Riyadh. 30 patients of

them were diagnosed with IDA. Diagnosis was done after clinical and laboratory tests including CBC,

Serum Iron, TIBC, and Ferritin.

Data Collection Source:

Data will be extracted from the laboratory records of King Khalid University Hospital, Riyadh. The

hematological parameters to be collected include:

 Hemoglobin (Hb, g/dL)

 Red Blood Cell Count (RBC, million/μL)

 Mean Corpuscular Volume (MCV, fL)

 Mean Corpuscular Hemoglobin (MCH, pg)

 Red Cell Distribution Width (RDW, %)

31
 Additional parameters: Serum Iron, Total Iron Binding Capacity (TIBC), Ferritin, and Transferrin

Saturation, where available.

32