evaluation of the fibrinogen antigenic turbidimetric assay as a screening method for measurement of...

B R I E F C O MM U N I C A T I O N

Evaluation of the fibrinogen antigenic turbidimetric assay as ascreeningmethod formeasurement of fibrinogen concentrationin dogsPeter Ludwig, Ernst F. Leidinger, Emma H. Hooijberg

In Vitro Laboratory, Vienna, Austria

KeyWords

Citrate, EDTA, FIATA, heparin, particle-

enhanced, thermoprecipitation, vancomycin

Correspondence

Ernst Leidinger, In Vitro Laboratory, Rennweg

95, A-1030 Vienna, Austria

E-mail: [email protected]

DOI:10.1111/j.1939-165X.2012.00426.x

Background: A novel method for the rapid detection of fibrinogen concen-

tration in human plasma, the fibrinogen antigenic turbidimetric assay

(FIATA), is based on the precipitation of fibrinogen by vancomycin and a

resultant change in optical density.

Objective: The objectives of this study were to evaluate the FIATAmethod

for: (1) measuring fibrinogen concentration in canine plasma using speci-

mens collected in citrate, EDTA, and heparin, (2) species-specific calibra-

tion requirements, and (3) applicability for automation.

Methods: Standard curves were generated with both human and canine

fibrinogen standards in the FIATA, and a reference interval for fibrinogen

concentration was established using citrated plasma from healthy dogs

(n = 127). Using specimens collected from this population, results using

the FIATA were compared with a modified thermoprecipitation method,

and 24 of the FIATA samples were used for comparison with a particle-

enhanced turbidometric fibrinogen assay. The FIATA was also applied to

an automated chemistry analyzer using citrated plasma. Fibrinogen con-

centration was measured in EDTA and heparinized plasma in the manual

FIATA. Standards, methods, and anticoagulants were compared, and corre-

lation among these variables was evaluated.

Results: Significant differences between FIATA results using human and

canine standards and the manual and automated methods were not found.

For EDTA plasma, fibrinogen concentrations were not identical, but were

similar, to those for citrated plasma; heparinized plasma was not suitable

for measurement. Correlation between the thermoprecipitation method

and FIATA was weak. The reference interval for fibrinogen as measured by

the FIATA using citrated plasmawas 103–456 mg/dL.

Conclusions: The FIATA can be used as a screening method to measure

fibrinogen concentration in citrated or EDTA plasma from dogs.

Different methods are used for measuring fibrinogen

concentrations in animals, including turbidimetric

assays adapted from human medicine, coagulation

methods, such as the modified Clauss method,1 or a

thermoprecipitation, or Millar, method.2,3 As different

methods of measuring fibrinogen concentration yield

varying results in healthy animals,4 there is no defined

gold standard in veterinarymedicine.

Recently, a novel method, the fibrinogen anti-

genic turbidimetric assay (FIATA), for measuring total

fibrinogen, including inactive or dysfunctional fibrino-

gen or forms of soluble fibrin, was introduced to

human medicine.5 The method makes use of the mac-

rocyclic gylocopeptidic antibiotic, vancomycin, as a

precipitating agent. Vancomycin is known to inhibit

coagulation by specifically binding to fibrinogen and

factors V, VIII, and X.6 Fibrinogen is denatured and

precipitates, resulting in a concentration-dependent

increase in plasma turbidity. In a human study no sig-

nificant difference was found between FIATA and a

modified Clauss method for fibrinogen determina-

tion, and inter-assay and intra-assay coefficients of

Vet Clin Pathol 41/2 (2012) 243–248©2012 American Society for Veterinary Clinical Pathology 243

Veterinary Clinical Pathology ISSN 0275-6382

variation < 4% and a lower detection limit (analytical

sensitivity) of 100 mg/L were reported.5 Although it

has been shown that this technique provides reproduc-

ible and reliable results for human plasma, it is not

known if it is applicable to plasma for animals.

Although the principle of the test is simple, canine

fibrinogen might react differently than human fibrino-

gen upon contact with vancomycin. The objectives of

this study therefore were to evaluate the FIATA

method for: (1) measuring fibrinogen concentration in

canine plasma using specimens collected in citrate,

EDTA, and heparin, (2) species-specific calibration

requirements, and (3) applicability for automation.

Blood samples were collected from 127 clinically

healthy dogs presented to local veterinary practices for

routine examination. For each dog a complete CBC

and biochemical profile were performed and abnor-

malities were not detected. The population consisted of

35% intact males, 13% neutered males, 26% intact

females, and 25% spayed females. Of the 127 dogs,

18% were < 3 years of age, 30% were 3–9 years, and

52% were > 9 years. A wide range of pure-bred and

mixed breed dogs were represented. This population

was used to generate the FIATA reference interval.

Blood samples were collected into tubes (Vacuette,

Greiner Bio-One GmbH, Kremsmunster, Austria) con-

taining 3.2% sodium citrate (n = 127), potassium

EDTA (n = 64), and lithium heparin (n = 77) and

transported to the laboratory by courier service within

12 hour of collection. On receipt specimens were

checked visually for the presence of clots before hema-

tologic analysis or centrifugation. After centrifugation

for 5 minute at 3000g, plasma was visually inspected

for hemolysis or lipemia, aliquoted into plastic reaction

cups, and stored frozen at �40°C for up to 1 week.7

Owing to the presence of clots, hemolysis, or lipemia,

15 specimens collected in citrate and 13 collected in

EDTA were excluded from the study. Specimens were

thawed at 37°C for 5 minute prior to the analysis. For

the main part of the study, 103 citrated plasma, 64

EDTA plasma, and 77 heparinized plasma samples

were used; an additional 24 citrated plasma samples

were used to compare the FIATA with the particle-

enhanced turbidometric fibrinogen assay.

Reagents included a 5-level human fibrinogen

standard (Dade, Siemens Health Care Diagnostics Inc.,

Marburg, Germany), fibrinogen control (Dade Data Fi

abnormal fibrinogen control), total protein reagent

(TP Dimension; Siemens Health Care Diagnostics

Inc.), single-level canine fibrinogen standard (Dunn

Labortechnik GmbH, Asbach, Germany), bovine

serum albumin (BSA; Sigma-Aldrich, Vienna, Aus-

tria), phosphate-buffered solution (PBS, pH 7.40,

Sigma–Aldrich), and vancomycin hydrochloride

(Sigma–Aldrich). Equipment included microtiter

plates (Greiner Bio-One GmbH) and an ELx800 ELISA

reader (BioTek Instruments, Winooski, VT, USA).

Fibrinogen standards and controls were reconstituted

according to manufacturer instructions, aliquoted,

and stored at �40°C for up to 1 week. The FIATA

reagent consisted of vancomycin diluted to a working

concentration of 4.4 mmol/L using PBS and was used

within 4 hour.

The FIATA was performed as previously

described.5 Briefly, 25 lL plasma was added to 50 lLPBS in a microtiter plate and optical density (OD) was

measured at 405 nm; 50 lL FIATA reagent was then

added, the mixture was incubated for 2 minute at

room temperature (25°C), and OD measured again at

405 nm. Increase in absorbance (DOD) was calculated,

and the fibrinogen concentration was read from stan-

dard calibration curves.4 Two calibration curves were

established, one using the human fibrinogen standard

(FIATAhuman) and the other using the canine fibrino-

gen standard (FIATAcanine). Recalibration was per-

formed with every new batch of FIATA reagent; a 6%

solution of BSA in PBS was used as the zero standard

(blank). Automated FIATA was performed by adapting

the manual method to analyzer specifications using an

open channel on a Dimension RxL (Siemens Health

Care Diagnostics Inc.). Specifically, reagent in a vol-

ume of 180 lL and sample in a volume of 50 lL were

measured at a wave length of 405 nmwith a reference

filter of 700 nm. Automated fibrinogen measurement

by FIATA took < 2 minute. Citrated plasma was used

for FIATAhuman and for the automated version of FIAT-

A (FIATAautomated). To correct for dilution by sodium

citrate, fibrinogen concentrations were multiplied by

1.1. For EDTA plasma (FIATAEDTA) and heparinized

plasma (FIATAheparin), human fibrinogen standard was

used as calibrator.

The method used for comparison (fibrinogenDako)

was a particle-enhanced turbidimetric fibrinogen assay

(Dako Fibrinogen, Dako, Denmark) adapted to a Hit-

achi 911 chemistry analyzer (Roche, Vienna, Austria)

according to the manufacturer’s instructions. Citrated

plasma was used for this assay and for the modified

Millar thermoprecipitation method3 (fibrinogentherm)

in which 150 lL plasma was heated to 56°C in a water

bath for 3 minute and then centrifuged for 3 minute

at 3000g. Total protein concentration was measured in

the supernatant by an automated biuret method (TP

Dimension, Dimension RxL), and the difference

between this measurement and the total protein con-

centration of the unheated citrated plasma was calcu-

lated. Manual and automated FIATA measurements

Vet Clin Pathol 41/2 (2012) 243–248©2012 American Society for Veterinary Clinical Pathology244

FIATA assay for fibrinogen measurement Ludwig et al

were performed in duplicate with mean intra- and

inter-assay CVs of 4.6% for both assays.

Sample size was selected to result in statistical

power of > .8 whenever feasible. To estimate sample

size, SPSS v. 16.0.2 (SPSS GmbH, Munich, Germany),

Analyse-It v. 2.25 (Analyse-It Software Ltd., Leeds,

UK), and G*Power8 v. 2.1.2 (Kiel, Germany) were

used. Refval 4.09 (Oslo, Norway) was used for calculat-

ing reference intervals by the robust method. Descrip-

tive statistics were calculated and distribution of data

assessed for normality using the Shapiro–Wilk test. A

paired t-test was used to evaluate differences between

calibrators and anticoagulants. Significance was set at

P < .05. Bland–Altman plots were used to evaluate

agreement between methods and anticoagulants and

to determine bias. Passing–Bablok fit was used for

method comparison. In addition, Pearson’s correlation

was calculated to evaluate correlation among assays,

standards, and anticoagulants.

Data were either normally (FIATAhuman,

FIATAcanine, and FIATAEDTA) or log-normally (modi-

fied Millar method, fibrinogenDako, FIATAautomated)

distributed; mean and median values and confidence

intervals (CI) were calculated (Table 1). Applying the

paired t-test significant differences were not found

between FIATAhuman and FIATAcanine results

(P = .109; power .47) and between FIATAhuman and

FIATAautomated results (P = .052; power .99). Dif-

ferences were found between FIATAhuman and

FIATAEDTA results (P < .001; power .99), between

FIATAhuman and modified Millar method results

(P = .008; power .99), and between fibrinogenDako

and FIATAhuman results (P = .008, power .95).

Correlations between FIATAhuman and FIATAcanine

results (P < .01, r = .993, m = 69.3, b = 0.795; Fig-

ure 1) and between FIATAhuman and fibrinogenDako

results (P < .01, r = 1.0, m = 0.0, b = 1.10; Figure 2)

Table 1. Canine plasma fibrinogen concentrations (mg/dL) measured

using different methods, standards, and anticoagulants.

Method n Mean Median 95% CI

Particle-enhanced turbidimetric

assay

24 123 105 93–154

Modified Millar method 100 191 205 156–234

FIATAhuman 103 367 360 335–400

FIATAcanine 103 361 355 335–387

FIATAEDTA 64 316 316 274–359

FIATAautomated 92 323 315 280–373

CI, confidence interval; FIATAhuman, fibrinogen antigenic turbidimetric

assay using a human fibrinogen standard; FIATAcanine, FIATA using

a canine fibrinogen standard; FIATAEDTA, FIATA using EDTA plasma;

FIATAautomated, FIATA using automation. Except for FIATAEDTA, citrated

plasma was used in the assays.

Figure 1. Passing–Bablok scatterplot comparing calibration curves for

fibrinogen concentration (mg/dL) established using canine (FIATAcanin) or

human (FIATAhuman) calibrators in the fibrinogen antigenic turbidimetric

assay (FIATA). The thin solid line is the line of identity, the thick solid line

is the line of regression (15.04 + 0.97x), and the dotted lines are the 95%

confidence interval. Lines overlap in the plot.

Figure 2. Passing–Bablok scatterplot comparing fibrinogen concentra-

tions (mg/dL) obtained from a calibration curve established using human

calibrators in the fibrinogen antigenic turbidimetric assay (FIATAhuman)

and a particle-enhanced turbidimetric assay (Fibrinogen Dako) for 24

dogs. The thin solid line is the line of identity, the thick solid line is the line

of regression (�0.00 + 1.10x), and the dotted lines are the 95% confi-

dence interval. Lines overlap in the plot.


Ludwig et al FIATA assay for fibrinogen measurement

were excellent. Correlation between citrated plasma

(FIATAhuman) and EDTA plasma (FIATAEDTA, r = .884,

m = �65.6, b = 1.05; Figure 3) was good. Correlation

between FIATAhuman and the modified Millar method

results was weak (P < .01, r = .463, m = �149.4,

b = 1.06; Figure 4). When heparinized plasma was

used in the FIATA, results above the detection limit

were not obtained.

The reference interval (95th percentile) for fibrin-

ogen concentration measured in citrated plasma by

FIATA was calculated as 103 mg/dL (90% CI: 92–115)to 456 mg/dL (90% CI: 427–486) by a parametric

method after log-transformation. Significant differ-

ences in FIATA results between sexes or age groups

were not found using ANOVA. Construction of a

MEDx chart,10 together with specifications for biologic

variation of fibrinogen concentration (maximum

imprecision, 5.4%; maximum bias, 4.8%; total allow-

able error, 13.6%)11 resulted in acceptable perfor-

mance using a CV of 4.6% and a proportional bias of

1.1% (see regression line, Figure 5).

Use of human fibrinogen standards for FIATA cali-

bration seems to be appropriate, although the desired

power of 0.8 was not achieved. This is explained by the

excellent agreement between results obtained using

the human and canine standards for calibration, which

would necessitate a large sample size (n > 986) to

achieve a power > .8. Although use of species-specific

reagents is encouraged in veterinary medicine,12 avail-

ability of canine fibrinogen standards is problematic, as

there are few sources and their high costs do not seem

to warrant their use in light of the data presented here.

As the human fibrinogen standard was equivalent to

canine fibrinogen, andmore economical and available,

it is suitable for the FIATA assay.

Figure 3. Passing–Bablok scatterplot comparing fibrinogen concentra-

tions (mg/dL) measured using the fibrinogen antigenic turbidimetric

assay (FIATA) in EDTA (FIATAEDTA) or citrated (FIATAcitrate) plasma. The

thin solid line is the line of identity, the thick solid line is the line of regres-

sion (�65.55 + 1.05x), and the dotted lines are the 95% confidence

interval.

Figure 4. Passing–Bablok scatterplot comparing fibrinogen concen-

trations (mg/dL) measured in citrated plasma using the fibrinogen anti-

genic turbidimetric assay (FIATAcitrate) and the modified Millard method

(Fibrinogenthermo). The solid line is the line of regression (267 + 0.4982x),

the inner dotted lines are the 95% confidence interval, and the outer dot-

ted lines are the 95% confidence interval.

Figure 5. MEDx chart for evaluation of the performance of the fibrino-

gen antigenic turbidimetric assay (FIATA). Colored lines indicate criteria

for total allowable error: red, BIASmax + (1.65 * CVmax); orange, BIASmax +

(2 * CVmax); blue, BIASmax + (3 * CVmax); green, BIASmax + (4 * CVmax). The

red + indicates FIATA performance.



The EDTA slightly quenches the FIATA reaction

leading to significantly lower reported fibrinogen con-

centrations,5 and in our study there was a negative bias

of 15% when fibrinogen concentration was measured

using EDTA plasma compared with the concentration

obtained from citrated plasma. However, although sta-

tistically significant, the clinical impact of the lower

fibrinogen concentration measured in EDTA plasma

was minor; hence, EDTA could be used as an alterna-

tive to citrate, if needed. This fulfills an important pre-

requisite for the use of FIATA as a screening method,

as blood specimens collected in EDTA are typically sub-

mitted to referral laboratories for hematologic analysis.

Specimens collected in citrate are typically submitted

only when coagulation tests are indicated. Heparinized

plasma, however, is unsuitable for measuring fibrino-

gen by FIATA. Heparin has been implicated as a deacti-

vator of gentamicin through formation of an inactive

complex.13 It is possible that a similar mechanism

occurs with vancomycin; however, this has not been

investigated.

Moderate correlation between FIATA and theMil-

lar method implies a weak agreement between the 2

methods. Although still described in recent literature

and widely accepted,14,15 the modified Millar method

may be suboptimal in reliably measuring fibrinogen,

especially in the lower analytical range.16 Failure of

this method to report fibrinogen concentrations for 10

specimens in our study may have resulted from high

plasma viscosity or the presence of cryoglobulins.17,18

In contrast, vancomycin at the specified concentration

binds specifically to fibrinogen, and this diagnostic

approach is more specific than thermoprecipitation. In

addition, FIATA was highly reproducible with intra-

and inter-assay CVs of 4.6%, although at very high

fibrinogen concentrations (eg, > 800 mg/dL), CVs

increased up to 10%.

In the present study, FIATA was successfully

applied to an automated clinical chemistry analyzer;

application of the FIATA method was simple and

should be adaptable to many different automated sys-

tems with analyzer-specific modifications based on

requirements of the particular analyzer. The reagents

have a long shelf-life, and the calibration curves were

stable, making FIATA a suitable method for most

analyzers.

Study limitations included an incomplete method

comparison process, as recently described,19 with no

evaluation of linearity, interferences, and limits of

detections. In addition, < 40 specimens were available

for comparing FIATAhuman and fibrinogenDako. Despite

these limitations, we have demonstrated that FIATA is

a useful screening method for determining fibrinogen

concentration in canine plasma, as it is reliable, rapid,

automatable, and inexpensive. As fibrinogen can be

measured by FIATA in EDTA plasma, the assay can be

easily incorporated into routine laboratory panels to

screen for inflammation without considerably increas-

ing cost. The applicability of FIATA for measuring

fibrinogen in other species should be evaluated.

Acknowledgment

This study was conducted as a part of the requirements for a

Master thesis according to guidelines of the Faculty of Life

and Health Sciences at Ulster University.

Disclosure: The authors have indicated that they have

no affiliations or financial involvement with any organiza-

tion or entity with a financial interest in, or in financial

competition with, the subject matter or materials discussed

in this article.

References

1. Mackie J, Lawrie AS, Kitchen S, et al. A performance

evaluation of commercial fibrinogen reference prepara-

tions and assays for Clauss and PT-derived fibrinogen.

Thromb Haemost. 2002;87:997–1005.

2. Millar HR, Simpson JG, Stalker AL. An evaluation of

the heat precipitationmethod for plasma fibrinogen

estimation. J Clin Pathol. 1971;24:827–830.

3. Kaneko JJ. Serum proteins and dysproteinaemias. In:

Kaneko JJ, HarveyW, BrussML, eds. Clinical Biochemis-

try of Domestic Animals. 6th ed. San Diego, CA: Academic

Press; 2008:148–149.

4. Lutze G, Kutschmann K. Zur Problematik der Fibrino-

genbestimmung bei Rind, Schwein und Hund [Deter-

mination of fibrinogen in the dog, cattle and swine].Mh

Vet Med. 1988;44:28–30.

5. Stief TW. The fibrinogen antigenic turbidimetric assay

(FIATA): the v2x test – the corrected Chi-Square com-

parison against the control-mean. Clin Appl Thromb He-

most. 2007;13:73–100.

6. Stief TW, Kurz J, DossMO, Fareed J. Singlet oxygen

(1O2) inactivates fibrinogen, factor V, factor VIII, factor

X, and platelet aggregation of human blood. Thromb Res.

2000;97:473.

7. Alesci S, BorggrefeM, Dempfle CE. Effect of freezing

method and storage at -20 degrees C and -70 degrees C

on prothrombin time, aPTT and plasma fibrinogen lev-

els. Thromb Res. 2009;124:121–126.

8. G*Power 3. Available at: http://www.psycho.uni-dues-

seldorf.de/aap/projects/gpower. AccessedMay 16,

2011.

9. Solberg HE. RefVal: a program implementing the rec-

ommendations of the International Federation of


Ludwig et al FIATA assay for fibrinogen measurement

Clinical Chemistry on the statistical treatment of refer-

ence values. Comput Methods Programs Biomed. 1995;48:

247–256.

10. Westgard QC downloads. Available at: http://www.

westgard.com/downloads/cat_view/53-worksheets.

AccessedMay 16, 2011.

11. Ricos C, Alvarez V, Cava F, Garcia-Lario JV, Hernandez

A, Jimenez CV,Minchinela J, Perich C, SimonM.

Current databases on biologic variation: pros, cons and

progress. Scand J Clin Lab Invest. 1999;59:491–500.

12. Ceron JJ,Martinez-Subiela S. A seven point plan for

acute phase protein interpretation in companion ani-

mals. Vet J. 2008;177:6–7.

13. Moyer TP, Shaw LM. Therapeutic drugs and their

management. In: Burtis CA, Ashwood ER, Bruns DE,

eds. Tietz Textbook of Clinical Chemistry and Molecular

Diagnostics. 4th ed. St. Louis, MO: Elsevier Saunders;

2006:1237–1285.

14. Dintenfass L, Kammer S. Re-evaluation of the heat pre-

cipitationmethod for plasma fibrinogen estimation:

effect of abnormal proteins and plasma viscosity. J Clin

Pathol. 1976;29:130–134.

15. Kjelgard-HansenM. Comments onmeasurement of C-

reactive protein in dogs [Letter]. Vet Clin Pathol. 2010;

39:402–404.

16. Mischke R,Menzel D.Messung der Fibrinogenkonzen-

tration beim gesunden Hund: standardisierung,Metho-

denvergleich und Referenzwerte [Measurement of

fibrinogen concentration in the healthy dog. Standardi-

zation, method comparison, and reference values]. Zen-

tralbl Veterinarmed A. 1994;41:587–598.

17. Evans EW, Duncan JR. Proteins, lipids, and carbohy-

drates. In: Latimer KS,Mahaffey EA, Prasse KW, eds.

Duncan & Prasse’s Veterinary Laboratory Medicine. Clinical

Pathology. 4th ed. Ames, IA: Blackwell Publishing;

2003:162–192.

18. Stockham SL, ScottMA. Proteins. In: Stockham SL,

Scott MA, eds. Fundamentals of Clinical Pathology. 2nd

ed. Ames, IA: Blackwell Publishing; 2008:393–396.

19. Jensen AL, Kjelgaard-HansenM.Method comparison

in the clinical laboratory. Vet Clin Pathol. 2006;35:276–

286.



evaluation of the fibrinogen antigenic turbidimetric assay as a screening method for measurement of...

Documents