J Clin Pathol 1992;45:701-703

The cause of turbidity in lyophilised plasmas andits effects on coagulation tests

C F Hirst, L Poller

AbstractAims: To investigate the cause of turbidityin reconstituted lyophilised plasmas andto determine its effect on coagulometers.Methods: The turbidities of 20 normalplasmas and 16 reconstituted lyophilisedplasmas were determined by comparing a1 in 4 dilution in distilled water with astandard suspension in an AmincoFluorocolorimeter (American InstrumentCo) in nephelometric mode. The tur-bidities of five other plasmas were deter-mined before and after lyophilisation. Theturbid components of fresh and recon-stituted lyophilised plasmas were studiedusing electron microscopy. The effects ofturbidity on five types of coagulometerwere determined by adding varying con-centrations of a turbidity enhancingmaterial.Results: Reconstituted lyophilised plas-mas were more turbid than normal plas-mas, because of agglomerated liposomes.Serum depleted of chylomicrons and verylow density lipoproteins was not renderedmore turbid by lyophilisation. Three outof five types of automated coagulometertested gave activated partial thrombo-plastin times which were appreciablyaffected by plasma turbidity. One of theinstruments was unable to detect a clot ina moderately turbid plasma. A secondinstrument gave results which were sig-nificantly affected by turbidity. Turbidityof the substrate plasma did not affectspecific factor VIII assays in two types ofcoagulometer.Conclusions: Lyophilisation of plasmainduces turbidity due to the agglomera-tion of lipids. Such turbidity can affect theresults of coagulation tests. Suppliers oflyophilised plasmas should be aware ofthis problem.

United KingdomReference Laboratoryfor AnticoagulantReagents and Control,Withington Hospital,Manchester M20 8LRC F HirstL PollerCorrespondence to:Professor L Poller

Accepted for publication20 December 1991

Plasma is naturally turbid. Turbidity of serumis caused by the presence of lipid containingcomplexes, chylomicrons, and very low densitylipoproteins (VLDL).' The turbidity of plasmais probably attributable to the same factors.Reconstituted lyophilised plasmas are invari-ably more turbid than the plasmas from whichthey are prepared. Such turbidity may beundesirable if the plasmas are to be used incoagulometers based on nephelometric or

absorptiometric principles.

MethodsPLASMA TURBIDITY MEASUREMENTSThe technique was based on a method for themeasurement of water turbidity, described bythe American WaterWorks Association and theAmerican Public Health Association.2 Hydra-zine sulphate ([NH2]2H2SO4) (Sigma) 1-0 gand hexamethylenetetramine (Sigma) 10-0 gwere each dissolved in 100 ml of distilledwater. Five ml of each solution were mixed,diluted to 100 ml with distilled water, andincubated at 26°C for 24 hours. This solutiongradually develops a stable turbidity due to theformation of formazin polymer particles. Avalue of 400 nephelometric turbidity units(NTU) was ascribed to this suspension. Fromthis a series of working standards from 0-200NTU were prepared. The turbidities of 20normal plasmas and 16 reconstituted lyophi-lised plasmas were determined by comparing a1 in 4 dilution in distilled water with a standardsuspension in an Aminco Fluorocolorimeter(American Instrument Company) in nephelo-metric mode. The 16 lyophilised plasmascomprised seven obtained commercially andnine lyophilised in our laboratory. The turbid-ities of five further plasmas were determinedbefore and after lyophilisation.

ELECTRON MICROSCOPYElectron microscopic studies were performedon fresh and reconstituted lyophilised plasmas.One millilitre of each plasma was added to1-0 ml of distilled water and 3 0 ml of 30 g/lgluteraldehyde added. These preparations werecentrifuged at 65 000 x g for one hour. Ali-quots of the upper turbid layers were applied topialoform (Agar Scientific) coated electronmicroscopy grids and dried in air. They werethen examined in a Philips EM420 transmi-sion electron microscope (magnificationbetween x 10 000 and x 105 000).

TURBIDITY OF SERA DEPLETED OF LIPOPROTEINS

Four serum samples of varying lipid contentwere obtained from the Department ofBiochemistry at Withington Hospital, Man-chester. These comprised lipid rich, chylomi-cron and VLDL depleted, chylomican VLDLand low density lipoprotein (LDL) depleted,and lipid free. A fifth serum, devoid ofchylomicrons alone, was prepared by cen-trifugation of lipid rich serum at 3000 x g andremoval of the upper layer. The turbidity ofthese sera was determined before and afterlyophilisation.

701

on Novem

ber 17, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.8.701 on 1 A

ugust 1992. Dow

nloaded from

Hirst, Poller

EFFECT OF PLASMA TURBIDITY ON COAGULOMETERS

The turbidity of several aliquots of one plasmawas varied by adding standard volumes of thestandard formazin suspension. Activated par-tial thromboplastin times (APTT) by theManchester APTT method3 were performedon five types of caogulometer: Amelung KC 10(Baxter), IL 300R (Instrumentation Labo-ratory), IL 810 (Instrumentation Laboratory),MLA Electra 700 (Lab Medics) and Coaga-mate x 2 (Organon Teknika). Factor VIIIassays were also carried out on the plasmaseries using the two different models of coagul-ometer from Instrumentation Laboratory (IL300R and IL 810). Formazin suspension wasadded to a relatively clear factor VIII deficientplasma to test the effect of substrate plasmaturbidity in an APTT based on stage factorassay.

ResultsThe mean (SD) turbidities of fresh and recon-stituted lyophilised plasmas were 344 (126)and 1179 (56) NTU, respectively. The mean(SD) of the commercially obtained and locallyprepared reconstituted lyophilised plasmaswere 1533 (637) and 904 (351), respectively.The turbidities of fresh and reconstitutedplasmas were compared by unpaired t test andfound to be significantly different (p < 0-05).The mean (SD) turbidity of the five plasmastested before and after lyophilisation were 416(181) and 715 (280), respectively. These werecompared by paired t test and found to besignificantly different (p < 0 05).

Electron microscopic examination showedthat fresh and reconstituted lyophilised plas-mas were significantly different. Fresh plasmacontained lipid rich particles of between 150and 300 nm in diameter. Reconstitutedlyophilised plasma contained large lipid richagglomerates of up to 4 pm diameter, whichshowed structural liposome rearrangement.The commercially obtained plasmas containedthe largest agglomerates. Electron micrographsof fresh and reconstituted commerciallyobtained lyophilised plasmas are shown in thefigure.The turbidity values of the sera series

depleted of liposomes are shown in table 1.Chylomicrons andVLDL were responsible forthe increased turbidity of lyophilised plasmas.The APTT values of aliquots of a single

plasma with turbidity values artificially adjus-ted to between 264 and 3484 NTU, asmeasured on five coagulometers, are shown intable 2. The Amelung KC10, IL 300R, andMLA Electra 700 seemed to show no turbidity

i induced effects. The IL 810 was unable todetect the clot in a sample with a turbidity ofmore than 1216 NTU. The Coagamate x 2

Electron micrographs of the turbid fraction of natural (A) and reconstituted lyophilised(B) plasmas. Scale bar = 600 nm.

gave an APTT value which increased withturbidity. These relations were compared byregression analysis and the significance of theobserved correlations was tested by the F test.A significant correlation (p < 0-05) betweenturbidity and APTT was found in the Coaga-mate x 2 and the IL 300R. The tight precision

702

on Novem

ber 17, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.8.701 on 1 A

ugust 1992. Dow

nloaded from

The cause of turbidity in lyophilised plasnmas and its effects on coagulation tests

Table I Turbidity values before and after lyophilisation of a series of sera of differinglipoprotein content

IPre-lyophilisation Post-lyophilisationSerum type turbidity (NTU) turbidity (NTU)

Serum 500 720Serum depleted of chylomicrons 440 720Serum depleted of chylomicrons and VLDL 500 760Serum depleted of chyclomicrons VLDL and LDL 200 280Lipid free serum 200 220

Table 2 APTT values (seconds) of a single plasma "spiked" with varying levels ofturbidity tested on five coagulometers

Turbidity Amelung Organon MLA(NTU) KCIO IL 810 IL 300R Coagamate x 2 Electra 700

264 35 0 37-8 29-8 26-3 330411 34-3 39-8 29 8 25-8 31-6571 34-5 39-4 30-0 27-5 32-0896 35-7 38-6 29 8 28-5 33 0969 35-5 41-0 29 7 29-3 32-51067 34-4 41-3 30 0 29-2 32-61216 35-0 CE 30-2 30-0 32-41560 34-9 CE 30-0 29-4 32-41738 35-2 CE 30 5 29-6 32-12082 35-2 CE 30 4 29-6 32-62329 34-6 CE 31-1 30 3 32-22525 35-3 CE 30 5 30 3 33-02671 35-2 CE 30 7 30 4 32-63007 35 0 CE 30 9 30-1 32-43289 34 9 CE 30-6 310 32-53484 34-4 CE 30-2 31-2 32-3

CE = Coagulation error

(CV 1 4%) of the latter instrument allowed a

very small turbidity induced effect to be noted.There was no significant correlation betweenturbidity and AP'1T in the other analysers(p > 0 05).

Factor VIII assays performed on both typesof IL coagulometer provided no significantcorrelation between plasma turbidity and fac-tor concentration (data not shown) (factorassays were performed on these instrumentsbecause they use different clot detection prin-ciples when performing factor assays). Increas-ing the turbidity of a factor VIII deficientplasma from 520 to 1520 and 3520 NTU didnot significantly affect the results obtainedwhen assaying factorVIII concentrations of 20and 100 IU/dl on either model of IL coagul-ometer.

DiscussionLyophilisation significantly increased plasmaturbidity. Electron microscopy showed that theturbidity was due to structural rearrangementand agglomeration of the lipid vehicles inplasma. The different liposomes in plamacontain different quantitites of cholesterol,triglyceride, and phospholipid. The latter,being present in the liposome surface with

apolipoprotein, give the liposomes increasedwater solubility. The observed effects wereprobably due to liposome membrane fusing aspreviously described.4 6 Lyophilised materialhas been found to be unreliable in the assay ofcholesterol7 and apolipoproteins.8 This may bedue to the changed characteristics of lyophi-lised liposomes.The turbidity measurements of sera

depleted in liposomes showed that lyophilisa-tion induced turbidity is caused mainly bychylomicrons and VLDL.The observed effects of plasma turbidity on

coagulometers is important. The mean turbid-ity of lyophilised plasmas was only slightly lessthan the value which adversely affected the IL810 (1179 v 1216 NTU). This isntrument hasbeen superseded but many remain in use. TheCoagamate x 2 and the IL 300R gave APTTvalues which varied with turbidity. The highprecision of the IL 300R allows a small upwarddrift to be detected but this is not of any realimportance. The results with the Coagamatex 2 gave more cause for concern. A five seconddifference is evident between the least andmost turbid plasmas. Users of this instrumentwould obtain a result on a reconstitutedlyophilised plasma which could be significantlyaffected by its turbidity. Lipaemic samplescould also give the same effect. Such plasmashave been shown to have turbidities of up to6800 NTU.

In conclusion, turbidity induced by lyophili-sation is caused by lipids and can affect theresults of coagulation tests. Suppliers of lyophi-lised plasmas should be aware of this prob-lem.

We thank Dr A Curry for assistance with the electronmicro-graphy and to Dr A Duffy for providing the lipid depletedsera.

1 Gray CH, Howorth PJN. Clinical chemical pathology. Lon-don: Edward Arnold, 1979:135-49.

2 American Water Works Association and American PublicHealth Association. Standard methods for the examination ofwater and waste water. Washington, DC: APHA,1985:133-40.

3 Thomson JM, Poller L. The activated partial thromboplas-tin time. In: Thomson JM, ed. Blood coagulation andhaemostasis: A practical guide. Edinburgh: ChurchillLivingstone, 1985:301-39.

4 Strauss G, Schurtenberger P, Hauzer H. The interaction ofsaccharides with lipid bilayer vesicles; stabilisation duringfreeze-thawing and freeze-drying. Biochim Biophys Acta1986;858:169-80.

5 Womersley CM, Uster PS, Rudolph AS, Crowe JH.Inhibition of dehydration induced fusion between liposo-mal membranes by carbohydrates as measured by fluore-sence energy transfer. Cryo Biol 1986;23:245-55.

6 Crowe JH, Crowe M. Factors affecting the stability of dryliposomes. Biochim Biophys Acta 1988;939:327-34.

7 Kroll MH, Chester R. Effect of serum lyophilisation on therate constants of enzymic methods for measuring choles-terol. Clin Chem 1990;36:534-7.

8 Marcovina SM, Adolphson JL, Parlavecchia M, Albers nJ.Effects of lyophilisation of serum on the measurement ofapolipoproteins A-I and B. Clin Chem 1990;36:366-9.

703

on Novem

ber 17, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.8.701 on 1 A

ugust 1992. Dow

nloaded from