Am J Clin Pathol 2008;130:127-130 127127 DOI: 10.1309/E5R9P5YPHXFE3198 127

© American Society for Clinical Pathology

Hematopathology / Westergren vs Centrifugation esr

Differences in Erythrocyte Sedimentation Rates Using the Westergren Method and a Centrifugation Method

Suresh G. Shelat, MD, PhD,1,2 Deborah Chacosky, MS, MT(ASCP)SH,1 and Susan Shibutani, MT(ASCP)1

Key Words: Sickle cell disease; Erythrocyte sedimentation rate; ESR

DOI: 10.1309/E5R9P5YPHXFE3198

A b s t r a c tThe erythrocyte sedimentation rate (ESR) is a

nonspecific screening test to assess elevations of acute phase proteins that occur in various acute and chronic diseases. The recommended method is based on the Westergren method. Other technologies to measure ESR have been developed, including centrifugation-based methods. We compared a centrifugation method (ESR STAT PLUS, HemaTechnologies, Lebanon, NJ) with the current Westergren method (Sediplast ESR system, Polymedco, Cortlandt Manor, NY) at 3 ESR levels (0-20, 21-60, and >60 mm/h) in a pediatric population and found correlation coefficients of r = 0.63, r = 0.83, and r = 0.43 respectively. The centrifugation ESR exceeded the Westergren ESR by a small but significant amount in the 0-20 mm/h range, where most normal ranges fall. The 2 methods were also compared in blood samples from patients with sickle cell disease, who tend to have lower ESRs owing to impaired rouleau formation. In the sickle cell group (n = 39; r = 0.87), the centrifugation ESR consistently exceeded (97% of time) the Westergren ESR. This study identifies the differences in ESR results using 2 methods in general and specific pediatric populations.

The erythrocyte sedimentation rate (ESR) is a straight-forward, inexpensive, but nonspecific screening test to assess an inflammatory or acute phase response.1 It has been used in the diagnosis or management of temporal arteritis, polymyal-gia rheumatica, and osteomyelitis and in tracking the course of various disease states, including tuberculosis, rheumatoid arthritis, and Hodgkin disease.2 The reference method to determine the ESR is based on the Westergren method,3,4 which determines the sedimentation of erythrocytes after 1 hour in a vertically mounted tube of defined length and bore size. However, other technologies have been developed to improve on the Westergren method (eg, smaller sample volume, less manual manipulations, shorter testing times, automation, and interfacing with the laboratory information system). This study was performed to compare the perfor-mance of the current Westergren method with a method based on centrifugation that required a smaller sample volume and shorter testing time.

Materials and Methods

General Patient StudyThe study was performed on samples sent to our labora-

tory for determination of ESR. All samples (n = 124) were run in tandem on the Sediplast ESR system (Polymedco, Cortlandt Manor, NY), which is based on the Westergren method, and the ESR STAT PLUS (HemaTechnologies, Lebanon, NJ), which is based on a centrifugation method. Hereafter, the Sediplast ESR will be referred to as the Westergren method and the ESR STAT PLUS as the centrifugation method.

128 Am J Clin Pathol 2008;130:127-130128 DOI: 10.1309/E5R9P5YPHXFE3198

© American Society for Clinical Pathology

Shelat et al / Westergren vs Centrifugation esr

Sickle Cell Study

Sickle cell patient blood was obtained from 39 specimens sent to the hemoglobin laboratory for hemoglobin identifica-tion and quantification. All samples were analyzed in tandem by the Westergren and centrifugation methods. Hemoglobin S quantification was performed by high-pressure liquid chroma-tography (Bio-Rad Variant II, Bio-Rad, Hercules, CA).

Westergren MethodThe procedure was performed according to the manufac-

turer’s instructions, using the manufacturer’s vials, graduated tubes, leveling rack, and vertical tube holder. Briefly, 0.8 mL of EDTA-anticoagulated blood was pipetted into a separate prefilled plastic vial containing 0.2 mL of 3.8% sodium citrate diluent and inverted 10 times to mix. Next, a plastic tube with graduated markings (0-150 mm) was inserted through the pierceable stopper to the bottom of the vial. This combina-tion of “tube in vial” was then placed vertically in a hole in a perfectly flat holder. The system is self-zeroing, such that the column of RBCs begins at the 0-mm mark at the top of the tube. After 60 minutes, the ESR was read as the millimeter mark where the erythrocyte-plasma interface appeared. Daily quality control methods were performed.3 The coefficient of variation was 43% for a low ESR sample (13 replicates; mean, 1.84 mm/h; SD, 0.8 mm/h) and 12.8% for a higher ESR sample (13 replicates; mean, 55 mm/h; SD, 7 mm/h).

Centrifugation MethodThe procedure was done according to the manufacturer’s

instructions (version 3.4, revised April 1, 2006). Briefly, EDTA-anticoagulated blood was mixed by inversion at least 15 times before 25 µL was drawn (by capillary action) into a heparin-coated, self-seal capillary analysis tube (provided by the manufacturer). This procedure was repeated imme-diately for a duplicate tube placed on a second ESR STAT PLUS instrument. The sample tube was placed in the ESR STAT PLUS centrifuge and spun at 1,500 to 2,000 rpm for 3 minutes. According to the manufacturer’s operator manual,

an infrared laser tracks the erythrocyte-plasma interface and makes multiple measurements, from which the linear por-tion of the sedimentation curve is identified and used by the software algorithm to determine the ESR result (mm/h). Daily quality control procedures were performed according to the manufacturer’s instructions. The coefficient of variation for a low ESR sample was 12.2% (9 replicates; mean, 6.8 mm/h; SD, 0.83 mm/h) and 2.57% for a higher ESR sample (9 repli-cates; mean, 94 mm/h; SD, 2.4 mm/h).

Statistical AnalysisStatistical analysis was performed by using GraphPad

Prism (version 5.0; GraphPad, San Diego, CA). In both studies, the 2 ESR STAT PLUS instruments had a close correlation with each other (r = 0.98; y = 0.98x + 1.42; r2 = 0.96). Therefore, the average of the 2 ESR STAT PLUS readings was used in the statistical comparison with the Westergren method.

Results

In the general patient study, a total of 124 samples were analyzed in tandem by the Westergren and centrifugation methods. zTable 1z shows the results of the correlation and regression analysis, the mean differences of paired ESR results, and results of the paired t test between the 2 methods. The analysis was performed in toto (n = 124) and also subdi-vided among 3 ESR ranges (0-20, 21-60, and >60 mm/h) zFig-ure 1z. The correlation was best in the 21-60 mm/h group (r = 0.83) and least in the more than 60 mm/h group (r = 0.43).

There was a statistically significant difference between the methods in the 0-20 mm/h range, in which the centrifu-gation ESR exceeded the Westergren ESR by an average of 3.3 mm/h, a result that was extremely unlikely (P < .0001) to occur by random sampling (Table 1). The Bland-Altman plot zFigure 2z graphically demonstrates the difference in ESR (Westergren minus centrifugation) at different ESR values.

In the sickle cell study, a total of 39 samples from hemo-

zTable 1zComparison of ESR Results Between the Westergren and Centrifugation Methods

No. of Correlation Slope y-Intercept Mean Difference in Paired t Samples Coefficient (95% CI) r2 (95% CI) (95% CI) Paired ESR* (95% CI) Test†

ESR (mm/h)‡ 0-20 54 0.63 (0.43 to 0.77) 0.40 0.88 (0.58 to 1.19) 4.22 (1.13 to 7.31) –3.30 (–5.08 to –1.52) t53 = 3.72, P < .0001 21-60 33 0.83 (0.69 to 0.91) 0.69 0.99 (0.75 to 1.2) –1.94 (–10.91 to 7.01) 2.19 (–0.72 to 5.11) t32 = 1.53, P = .134 >60 37 0.43 (0.13 to 0.65) 0.18 0.48 (0.14 to 0.82) 39.49 (11.54 to 67.45) 1.74 (–4.68 to 8.17) t36 = 0.55, P = .585Overall study 124 0.92 (0.89 to 0.94) 0.85 0.89 (0.83 to 0.96) 4.16 (0.98 to 7.33) –0.33 (–2.50 to 1.83) t123 = 0.302, P = .762Sickle study 39 0.87 (0.78 to 0.93) 0.77 1.21 (0.99 to 1.43) 6.99 (4.15 to 9.83) –9.01 (–11.07 to –6.95) t38 = 8.87, P < .0001

CI, confidence interval; ESR, erythrocyte sedimentation rate; r2, coefficient of determination.* Westergren minus centrifugation.† P values <.05 were considered statistically significant.‡ According to the Westergren method.

Am J Clin Pathol 2008;130:127-130 129129 DOI: 10.1309/E5R9P5YPHXFE3198 129

© American Society for Clinical Pathology

Hematopathology / original artiCle

globin SS patients were analyzed in tandem by the Westergren and centrifugation methods zFigure 3Az. Shown in Table 1 are the results of the correlation and regression analysis, the mean differences of paired ESR results, and results of the paired t test between the 2 methods. In 38 (97%) of 39 samples, the ESR values from the centrifugation method exceeded those of the Westergren method by an average of 9.01 mm/h zFigure 3Bz. It is extremely unlikely (P < .0001) that a difference of this size occurred by random sampling (Table 1). There was no discern-ible relationship between level of hemoglobin S and the magni-tude of the difference between the 2 ESR results zFigure 3Cz.

Discussion

To our knowledge, this is the first study compar-ing the ESR STAT PLUS centrifugation method with the Westergren method. The latter is the recommended stan-dard testing method by the International Committee for Standardization in Haematology.3 There are many factors that result in an increase in the ESR, including increasing age, female sex, amount of circulating fibrinogen levels,5 and degree of anemia. Conversely, the ESR may be low with polycythemia, hypofibrinogenemia, or erythrocyte abnormali-

ties, such as spherocytes or sickle cells.2 In this head-to-head study, the preanalytic factors were identical because identical blood samples were analyzed in tandem. Therefore, the ESR differences reflect analytic differences between the methods.

The centrifugation method uses multiple optical read-ings of the erythrocyte-plasma interface to determine the ESR. In the general patient study, the centrifugation ESR exceeded the Westergren ESR by a small (95% confidence interval, 1.52-5.08 mm/h) but statistically significant amount

0

10

20

0 10 20

Westergren ESR (mm/h)

Cen

trif

ug

atio

n E

SR

(m

m/h

)

30 40 50

30

40

50

0

20

40

60

0 20 40

Westergren ESR (mm/h)

Cen

trif

ug

atio

n E

SR

(m

m/h

)

60 80 100 120 140

80

100

120

140

0

20

40

60

0 20 40

Westergren ESR (mm/h)

Cen

trif

ug

atio

n E

SR

(m

m/h

)

60 80 100 120 140

80

100

120

140

0

20

0 20

Westergren ESR (mm/h)C

entr

ifu

gat

ion

ES

R (

mm

/h)

40 60 80

40

60

80

A B C

DzFigure 1z Correlation in general patient study between samples at an erythrocyte sedimentation rate (ESR) of <20 mm/h (A), 21-60 mm/h (B), >60 mm/h (C), and overall (D) (n = 124). The Pearson r, sample size, slope, and y-intercept results are shown in Table 1. All samples were run in tandem by the Westergren and centrifugation methods.

–50

–40

–30

–20

3 4

Dif

fere

nce

in E

SR

(Wes

terg

ren

– C

entr

ifu

gat

ion

) (m

m/h

)

6 8 11 16 23

Average ESR (mm/h)

32 45 64 91 128–10

0

10

20

30

40

50

zFigure 2z Bland-Altman plot, general patient study. The x-axis shows the average erythrocyte sedimentation rate (ESR) between the Westergren and centrifugation methods; it is plotted in log2 to allow visualization of data points at the low end.

130 Am J Clin Pathol 2008;130:127-130130 DOI: 10.1309/E5R9P5YPHXFE3198

© American Society for Clinical Pathology

Shelat et al / Westergren vs Centrifugation esr

in the lower range (0-20 mm/h), where most normal ranges fall (Table 1). This difference in ESR was even greater in the sickle cell group (95% confidence interval, 6.95-11.07 mm/h), in which 97% of the samples tested demonstrated a higher centrifugation ESR than Westergren ESR. The discrepancy did not seem to correlate with the level of hemoglobin S in the sample. The sickle cell study showed that there are certain patient populations in which a system-atic bias can occur between the Westergren method and the centrifugation method. It is unclear if this bias also occurs in other patient populations in which the ESR test may be ordered. Patients with sickle cell disease often demonstrate a low ESR owing to hindrance of rouleau formation by the sickled morphologic features.6 However, the ESR may be an important test to identify subsets of patients with sickle cell disease at risk of bacterial infection.7

The ESR STAT PLUS centrifugation-based ESR has the potential benefit of requiring a smaller volume than other Westergren method ESRs and can produce a result in a shorter time. However, there are concerns about its cor-relation with the reference Westergren method, especially in the 0-20 mm/h range, where most normal ranges fall. In addition, this centrifugation method requires a minimum of 15 mixing intervals, followed by a 5-minute limit (maxi-mum) before drawing into the capillary tube. If this process is not strictly followed, the process must be repeated with another capillary tube. Failure to do so can result in higher intra-assay variation. Because human error can occur in any busy hematology laboratory, it is critical that procedures be as straightforward as possible to achieve accurate ESR results.

From the 1Children’s Hospital of Philadelphia and 2Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia.

Address reprint requests to Dr Shelat: Hematology Laboratory, Children’s Hospital of Philadelphia, Clinical Laboratories, 5 Main, 34th St and Civic Center Blvd, Philadelphia, PA 19104.

Acknowledgments: We thank the hematology laboratory technologists at Children’s Hospital of Philadelphia for help in conducting this study.

References 1. Sox HC Jr, Liang MH. The erythrocyte sedimentation rate:

guidelines for rational use. Ann Intern Med. 1986;104:515-523.

2. Saadeh C. The erythrocyte sedimentation rate: old and new clinical applications. South Med J. 1998;91:220-225.

3. International Council for Standardization in Haematology Expert Panel on Blood Rheology. ICSH recommendations for measurement of erythrocyte sedimentation rate. J Clin Pathol. 1993;46:198-203.

4. National Committee for Clinical Laboratory Standards 2000. Reference and Selected Procedures for the Erythrocyte Sedimentation Rate (ESR) Test (H2-A4). Wayne, PA: NCCLS; 2000.

5. Bain BJ. Some influences on the ESR and the fibrinogen level in healthy subjects. Clin Lab Haematol. 1983;5:45-54.

6. Winsor T, Burch E. Rate of sedimentation of erythrocytes in sickle cell anemia. Arch Intern Med. 1944;73:41-52.

7. Robins EB, Khan AJ, Atrak T, et al. Erythrocyte sedimentation rate: a valuable test in infants and children with sickle cell disease. Clin Pediatr (Phila). 1993;32:681-683.

0

20

10

0 10 20Westergren ESR (mm/h)

Cen

trif

ug

atio

n E

SR

(m

m/h

)

30 40 50

30

40

50

20 40 60 80 100

% Hemoglobin S

Dif

fere

nce

in E

SR

(Wes

terg

ren

– C

entr

ifu

gat

ion

) (m

m/h

)

–30

–25

–10

–15

–10

–5

0

5

–30

–20

10 20

Average ESR

Dif

fere

nce

in E

SR

(Wes

terg

ren

– C

entr

ifu

gat

ion

) (m

m/h

)

30 40 50

–10

0

10

A B C

zFigure 3z Comparison of Westergren and centrifugation erythrocyte sedimentation rate (ESR) results in sickle cell samples. A, Correlation of ESR results. All 39 samples were run in tandem by the Westergren and centrifugation methods. The results are summarized in Table 1 (“Sickle study”). B and C, Bland-Altman plots. The x-axis shows the average ESR between the Westergren and centrifugation methods (B) and percentage of hemoglobin S in the sample (C).