`
`I85
`
`review articles
`
`Plasma or serum in therapeutic drug monitoring and clinical toxicology
`
`DONALD R.A. UGES
`
`Introduction
`
`In therapeutic drug monitoring and clinical toxi-
`cology most determinations of the concentration of a
`substance are carried out in blood or blood com-
`
`ponents. Whole blood can only be analysed if the
`blood has been sampled in collecting tubes con-
`taining an anticoagulant. When such samples are
`centrifuged, plasma is obtained. Therefore. plasma
`always contains one or more anticoagulants. Hep-
`arin sodium (3-100 Ulml) is most widely used for this
`purpose, but citrate, edetate, oxalate and fluoride
`are also suitable additives. Serum is the clear liquid
`that separates from blood when blood is allowed to
`clot completely and then centrifuged. Therefore
`serum contains no fibrogen and no anticoagulants.
`The choice between whole blood on the one hand
`
`and either plasma or serum on the other is clear.
`Whole blood concentrations are only measured if the
`compound is concentrated in the erythrocytes (e.g.
`lead, cyanide, mercury, carbon monoxide. chlor-
`thalidone),"3 if there is a fluctuating erythrocyte-
`plasma ratio (ciclosporin A),“ or because of the risk
`of loss during storage or centrifugation.‘
`The clinical effect of several compounds correlates
`better with the concentration in a tissue compart-
`ment than with the serum or plasma concentration.
`Considering erythrocytes as a readily available tis-
`sue, whole blood should be the matrix of choice in
`therapeutic drug monitoring ' and toxicology.”
`Sometimes lymphocyte concentration can predict
`the therapeutic effect better (epirubicin, mitox-
`antrone). 9 In nearly all other cases, plasma or
`serum concentrations are measured.
`
`Is there any difference between serum and plasma
`concentration?
`
`The expression ‘plasma concentration’ is a part of
`the title of many articles in the literature. Reading
`these articles, however,
`it appears that serum
`samples have often been taken. The authors very
`seldom present
`their arguments for
`the choice
`between plasma or serum.
`Some of the advantages of plasma over serum
`are:
`
`— large volume;
`— no delayed clotting;
`— less risk of haemolysis;
`— the sample is often suitable for both whole blood
`and plasma monitoring.
`Some of the disadvantages of plasma over serum
`are:
`
`— the (unknown) influence of the anticoagulant on
`the assay, the protein binding and the stability of
`the sample;
`— the (unknown) influence of additives or impurities
`in the anticoagulant on the assay and the concen-
`tration;
`— the risk ofthe formation of small clots (incomplete
`mixing, instability);
`— dilution of the sample;
`— sampling is more expensive;
`— choice of anticoagulant can be confusing.
`
`ADVANTAGES OF PLASMA OVER SERUM
`
`Larger available volume
`If blood is allowed to clot and is then centrifuged,
`
`formation of small clots and dilution of the sample.
`
`Keywords
`Anticoagulants
`Blood coagulation
`Blood preservation
`Plasma
`
`Protein binding
`Serum
`
`Therapeutic drug monitoring
`Toxicology
`
`‘Laboratory for Clinical and
`Forensic Toxicology and Drug
`Monitoring, Department of
`Pharmacy, University Hospital
`Groningen, P.O. Box 30.001,
`9700 RB Groningen.
`
`Uges DRA. Plasma or serum in therapeutic drug monitoring and clinical
`toxicology. Pharm Weekbl [Sci] I988;I0:185-8.
`
`Abstract
`
`The relative merits of plasma and serum in blood analysis are reviewed.
`The expression ‘plasma concentration’ is often used in the literature,
`although serum samples have been taken. In most cases serum and plasma
`concentrations of analytes are the same. The choice depends mostly on
`the policy of the hospital or the availability of the test tubes in the ward.
`Some of the advantages of plasma over serum are large volume, no
`delayed clotting, less risk of haemolysts. In addition, the sample 15 often
`suitable for both whole blood and plasma monitoring. Some of the
`disadvantages of plasma over serum are the (unknown) influence of the
`anticoagulant on the assay, on the protein binding and on the stability of
`the sample, the (unknown) influence of additives or impurities in the
`anticoagulants on the assay and on the concentration, the risk of the
`
`_
`
`
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 1
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 1
`
`
`
`186
`
`Vol. 10 - 1988 Pharmaceutisch Weekblad Scientific Edition
`
`about 30 to 50% of the original volume is collected as
`serum (upper layer). A slightly larger sample (in
`general 50%), known as plasma, can be obtained by
`centrifugation of a blood sample collected in a tube
`containing an anticoagulant. Serum and plasma are
`not significantly different with respect
`to drug
`analysis.
`that serum is devoid of the
`Despite the fact
`proteins associated with the clotting process, the
`most important proteins (quantitatively speaking),
`such as the albumins and globulins, are present in
`similar amounts (approximately 2% wtlvol) in both
`serum and plasma. Thus plasma is in general pre-
`ferred_ because of its greater yield from blood
`samples. The greater yield the greater the amount of
`drug and the fewer the problems with sensitivity, or
`with sampling neonates.” "
`
`Delay through time needed for clotting
`Clotting can sometimes be of long duration. This is
`especially true when samples are taken in polypro-
`pylene test tubes, in which case clotting can continue
`even after centrifugation.
`
`Decreased risk of haemolysis
`The chlorthalidone concentration in erythrocytes
`is about 40 times as high as that in plasma.“ One
`percent of haemolysis (which cannot easily be seen)
`will increase the apparent plasma concentration by
`25%. This problem is even more acute when serum
`samples are produced from whole blood.
`The separation by centrifugation should be carried
`out as quickly-as possible in order to prevent any
`effects due to lysis of the blood clot. " This phenom-
`enon is well known in the analysis of the endogenous
`compounds potassium and iron in serum.
`
`Suitability of the sample for both whole blood and
`plasma analysis
`If several analyses have to be carried out on the
`blood of a single patient, determinations can be
`carried out in whole blood as well as plasma from one
`blood sample containing a suitable anticoagulant. In
`bedside chemistry whole blood is preferred to
`plasma or serum, because this avoids a centri-
`fugation step.”
`
`DISADVANTAGES or PLASMA
`
`IN RELATION
`
`ro SERUM
`
`The influence of the anticoagulant on the assay
`Walters and Roberts showed that heparin at a
`concentration of 100 U per ml interfered with the
`gentamicin EMIT3 assay, but not with the TD?
`FPIA assay.” This effect is thought to be due to the
`inhibition of enzyme activity rather than disturbance
`of the antigen-antibody reaction. Blood collected
`into heparinized tubes usually contains approxi-
`mately 30 U of heparin per ml and at this level no
`interference with the EMIT® assay occurs. How-
`
`ever, plasma derived from specimens collected from
`arterial and venous blood is likely to yield erroneous
`results.
`
`The formation of a gentamicin—heparin complex
`has been demonstrated by Myers etal. “ Cipolle et al.
`concluded that heparinized tubes should not be used
`when collecting blood samples for assays where
`inactivation of gentamicin would give false values.”
`Edetate was for gentamicin a suitable alternative.
`On the other hand, the Dutch foundation ‘Quality
`control
`clinical drug analysis
`and toxicology‘
`(KKGT) found with its quality control programme
`that edetate interfered strongly with the FPIA assay
`and not with the EMIT‘@ assay of carbamazepine and
`with the E.MI'I® assay and not with FPIA assay of
`valproic acid. The reason of this interference is
`unknown (Dijkhuis IC, personal communication,
`1988).
`
`The influence of the anticoagulant on protein
`binding
`A discrepancy in the measurement of lidocaine in
`heparinized blood and in serum could be expected,
`as heparin increases the free fraction of lidocaine.”
`Lidocaine is subsequently redistributed from the
`plasma into red blood cells.
`The percentage of free ibuprofen in heparinized
`plasma is significantly higher than in non-heparin-
`ized plasma. 7 Although small doses of heparin can
`affect drug binding, the extent and variability of the
`effect depends on the biological activity of the
`heparin, and varies with the manufacturer and
`batch, time of sampling, and food intake . 18 Although
`the intravenous administration of heparin results in a
`high lipolytic activity in vivo and consequently
`increases the concentrations of non-esterified fatty
`acids in vitro, it is unlikely that this will happen after
`adding heparin to the tube.
`Non-esterified fatty acid can displace numerous
`drugs from their binding sites in plasma. When the
`non-esterified fatty acids concentration increases in
`vltro. phenytoin will be displaced from the binding
`sites. The free phenytoin then shifts from the plasma
`into the erythrocytes, resulting in a lower phenytoin
`plasma/level (up to 20% below its original value!).'9
`The same phenomenon was found with amitrip-
`tyline, imipramine and maprotiline.5 The recovery
`of chlorproniazine added to heparinized plasma in
`vitro was reduced compared with that obtained with
`water or oxalated plasma."
`
`The influence of the anticoagulant on the stability
`of the sample
`When heparin deteriorates or adsorbs onto the
`wall of the collecting tube, clotting will still occur in a
`clear plasma sample. The pH of the sample can be
`changed by using sodium citrate, or oxalate as
`anticoagulant. This change of pH may influence the
`stability of the drug in the sample (e.g. atracurium).
`Sodium fluoride can be used as anticoagulant (cal-
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 2
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 2
`
`
`
`
`
`187
`Vol. 10 - 1988 Pharmaceutisch Weekbiad Scientific Edition
`
`cium binder), and as a preservative (antimicrobial)
`of ethanol. Blood samples that are to be tested for
`ethanol must contain 10 mg sodium fluoride per ml
`during storage.” 22
`
`The influence of additives or impurities" in the
`anticoaguiant on the assay
`Benzyl alcohol, which may be added to heparin as
`a preservative, can disturb the fluorimetric assay of
`corticosteroids.” Serum zinc concentrations are 16%
`higher than plasma concentrations. This phenom-
`enon was explained by a release of zinc from
`platelets during coagulation of the blood samples.“
`Keyzer et at. , however, found no difference between
`the serum and plasma zinc levels of fifty volun-
`teers.”
`
`The influence of additives or impurities in the
`anticoaguiant on the concentration
`Anticoagulants can contain impurities which are
`precisely the compounds to be determined, e.g. lead,
`aluminium, copper, fluoride.
`
`The risk of clot formation because of poor mixing
`or poor stability
`Unfortunately, blood samples in heparinized
`tubes are sometimes not very well mixed. In such
`cases a partially clotted sample will arrive in the
`laboratory.
`
`Improper dilution of the sample
`We have had a number of incidents in which blood
`
`samples of about 5 ml were diluted with about 1 ml of
`a solution of anticoagulant (20% dilutionl). Differ-
`ences in serum and plasma transcobalamin II levels in
`the literature have been explained by dilution by
`EDTA or sodium fluoride solutions.’
`
`Cost
`
`Sample tubes with heparin are approximately 10%
`more expensive than non-heparinized tubes.
`
`The choice of anticoagulant
`There are several anticoagulants, each of which
`must be used in a different concentration. Examples
`are sodium heparin,
`lithium heparin, potassium
`edetate, sodium edetate, sodium citrate, and sodium
`fluoride. The differences in the influence of these
`
`anticoagulants on the various assays are largerly
`unknown.
`Differences were demonstrated between serum
`
`and plasma levels of endogenous compounds, which
`may be of sufficient magnitude to alter clinical
`interpretation of some results when using different
`radioassay procedures and different anticoagu-
`lants.37 These differences are in general larger with
`immunoassays than with chromatographic methods.
`Cholinesterase activity can be determined in serum
`or heparinized plasma. Sodium fluoride and citrate
`should not be used as anticoagulants because they
`
`depress cholinesterase activity as measured by sev-
`eral methods.”
`
`Anticoagulants used in sampling plasma can con-
`found diffusion assays by causing alterations in the
`agar gel matrix. Therefore, in the microbiological
`measurement of, for example, erythromycin, cytara-
`bine or 5-fluorouracil the use of serum is recommen-
`ded.“
`
`Conclusion
`
`In conclusion, we can say that in most cases serum
`and plasma analyte concentrations are the same. The
`choice depends mostly on the established practice in
`the hospital or the kind of available test tubes in the
`ward. In special cases such as the determination of
`free drug levels, however,
`there are important
`differences. It is advisable, therefore, that during
`kinetic studies of drugs, the similarity of serum and
`plasma levels for the compound being studied be
`proven. Furthermore, to avoid confusion, authors
`should be careful and precise in their presentations
`so that readers are aware that blood, plasma or
`serum is the sample under discussion.
`
`References
`
`' Casarett Ll, Doull J. Toxicology. The basic science of
`poisons. New York: Macmillan Publishing Company,
`1975-
`‘ Moffat AC, ed. Clarke's isolation and identification of
`drugs. 2nd ed. London: The Pharmaceutical Press,
`1986.
`" Beenrlann B, Groschinsky-Grind M. Clinical pharmaco-
`kinetics of diuretics. Clin Pharmacokinet r98o;5:22r-45.
`‘Keown PA, Stiller CR, Stawecki M, McMichael J,
`J-Iowson W. Pharmacokinetics and interactions of ciclo-
`sporin. In: Schindler R. ed. Ciclosporin in autoimmune
`diseases. Berlin: Springer-Verlag, 198539-42.
`5 Uges DRA, Van der Paauw HJW. Alcohol, de bloed—
`proef en de contra—expertise [Alcohol, blood test and
`counterexpertise]. Ned Forens Tijdschr 1982;2:7-9.
`" Maguire KP, Burrows GD, Norman JR, Scoggings BA.
`Blood plasma distribution ratios of psychotropic drugs.
`Clin Chem 1980;216:1624-5.
`1 Linnoila M, Dorrity F, Jobson K. Plasma and erythro-
`cyte levels of tricyclic antidepressants in depressed
`patients. Am J Psychiatry1978;135:557-61.
`" De Vries EGE, Greidanus J, Mulder NH, et al. A phase
`Ipharmacoltinetic study with 21-day continuous infusion
`of epirubicin. J Clin Oncol 1987;5:1445-51.
`" De Vries EGE, Greidanusl, Mulder NH, et al. A phase
`1 and pharmacokinetic study with 21-day continuous
`infusion of mitoxantrone (MK). Am Soc Clin Oncol
`1988;7:58.
`'° Smits RV, Stewart .lT. Textbook of biopharmaceutic
`analysis. Philadelphia: Lea & Febiger, 1g8r:3-I6.
`" Toseland PA. Samples and sampling. In: Moffat AC,
`ed. Clarke's isolation and identification of drugs. 2nd
`ed. London: The Pharmaceutical Press, 19862111-7.
`'2 I-lichs JA,
`Iosefsohn M. Another physician’s office
`analyser:
`the Abbott vision evaluated. Clin Chem
`1987;33:817-9-
`" Walters MI, Roberts WH. Gentamicinfheparin inter-
`actions: effects on two immunoassays and on protein
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 3
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 3
`
`
`
`
`
`188
`
`Vol. to - I988 Pharmaceutisch Weekbiad Scientific Edition
`
`binding. Ther Drug Monitor 1984;6:t99~2o2.
`“‘ Meyers DR, DeFehr J, Bennett WM, Porter GA, Olsen
`GD. Gentamicin binding to serum and plasma proteins.
`Clin Pharn-1acolTher t978;23:356—6o.
`'5 Cipolle RJ, Zaske DE, Crossley K. Gentamicinftobra-
`mycin. Therapeutic use and serum concentration moni-
`toring. In: Taylor WJ, Finn LA, eds. Individualizing
`drug therapy. 1. New York: Frank Cross Townsend Inc. ,
`1931138.
`“ Dube LM, Ho-Ngoc A, Davies RF, Beaulands DS,
`Mousseau N, McGilveray IJ . Influence of protamine on
`lieparine induced increases of lidocaine free fraction.
`Res Commun Chem Pathol Pharmacol 1983;42:401-
`15.
`” Pearce GA, Brown KF. Heat inhibition of in vitro
`lipolysis and "C ibuprofen protein binding in plasma
`from heparinized uraemic subjects. Life Sci 1983;
`33:1457—66.
`'3-Naranjo CA, Sellers EM, Khouw V, Alexander P, Fan
`T, Shaw I. Variability in heparin effect on serum drug
`binding. Clin Pharmacol Ther 1980;223:545-5o.
`"Schulz P, Abang A, Giacomini JC, Blaschke TF,
`Giacomini KM. Effect of heparin on the red blood
`cell-to-plasma concentration ratio of diphenylhydantoin
`and prazosin. Ther Drug Monitor 1983;5:497-9.
`‘° Whelpton R. Tricyclic antidepressants and neuroleptics.
`In: Curry AS, ed. Antalytical methods in human
`toxicology. Part I. Weinheim: Verlag Chemie, 1985:
`139-58-
`" Flanagan R], Widdop B. Clinical toxicology. In: Curry
`AS, ed. Analytical methods in human toxicology. Part I.
`Weinheirn: Verlag Chemie, 19:85:37-66.
`
`1’ Froentjes W, Schute JB, Strengers Th, Venvey AMA.
`Analytisch—chemische aspecten van de wijziging van de
`Wegenverkeerswet (11) [Analytical and clinical aspects
`of the change in the Traffic Act 01)]. Phartn Weekbl
`1976;111:349-62.
`33 Werk EE jr. Interference of heparin containing benzyl
`alcohol
`in the fluorometric determination of plasma
`corticosteroids. J Clin Endocrinol 1967;27:I35o.
`" Foley B, Johnson SA, Hackley B, Smith JC jr, Halsted
`JA. Zinc content of human platelets. Proc Soc Exp Biol
`Med 1968;128:265-9.
`3’ Keyzer JJ, Oosting E, Wolthers BG, Muskiet FAJ,
`Hindi-iks FR, Van der SlikW. Zincin plasma and serum:
`influence of contamination due to the collection tubes.
`
`Pharm Weekbl [Sci] 198332248-51.
`2“ Carmel R. Vitamin B12-binding proteins in serum and
`plasma in various disorders. Efiects of anticoagulants.
`Am J Pathol 1978;69:3I9-25.
`1'’ Kubasik'NP, Sine HE. Results for serum and plasma
`compared in
`15
`selected radioassays. Clin Chem
`1973:2413 137-9-
`” Huizinga JR, Gips CI-I. Evaluation of the UV-34o
`spectrophotometric determination for pseudocholin-
`esterase activity (EC 3.1.1.8) in human serum. J Clin
`Chem Clin Biochem 1987;25:I61-5.
`2" S:-nits RV, Stewart JT. Microbiologic assay methods. In:
`Textbook of biopharmaceutic analysis. Philadelphia:
`Lea & Febiger, 19811249-59.
`
`Received December 1987.
`Revised April 1988.
`Accepted May 1988.
`
`
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 4
`
`MYLAN PHARMS. INC. EXHIBIT 1055 PAGE 4