CLIN. CHEM.

38/4, 575-677 (1992)

Effect of In Vitro Hemolysis on 25 Common BiochemicalTests

Dogan Y#{252}cel
and Kiara Dalva

Clinical chemists frequently encounter hemolyzed sam- tients in the coronary care, intensive care, and hemodi-
ples. Our study examines the effects of hemolysison the alysis units of our hospital. With this distribution of
results of 25 common biochemical tests. We collected 60 patients, we hoped to have results over the entire range
15-mL blood samples from inpatients and outpatients and for each analyte tested. We drew the samples through
mechanically hemolyzed 10 mL of the samples in a wide-bore needles (18 gauge) with minimal stasis into
two-stepprocedure.We classified serum from these glass tubes. We centrifuged 5 mL of each sample within
samplesas beingnonhemolyzed,moderatelyhemolyzed, 30 mm of collection, at 1500 x g for 10 mm, thus
or severely hemolyzed and then performed 25 common obtaining 2.0-mL samples of nonhemolyzed serum. We
biochemical tests. Statistical analysis of the results hemolyzed the remaining 10 mL of each sample by
showed that hemolysis had the greatest effect on the stirring with a metallic bar, 5 mL for 1 mm and 5 mL for
lactate dehydrogenase, acid phosphatase, and potas- 2 mm (mechanical trauma), and we obtained 2.0-mL
sium tests. samples of moderately and severely hemolyzed serum
after centrifuging as before. This method of cell lysis
Additional Keyphrases: sample handllng ‘ lactate dehydroge- was chosen because it is analogous to the mechanical
nase - add phosphatase - potassium - vanatlon, source of disruption of erythrocytes that frequently occurs during
faulty blood collection and sample preparation.
Hemolysis is common in blood specimens, occurring Because the concentration of free Hb in serum is used
when blood contacts foreign surfaces. The use of partly as a measure of hemolysis, we measured free Hb in all
hemolyzed serum may be unavoidable, especially dur- specimens spectrophotometrically (7) and grouped the
ing cardiac surgery or cardiac bypass, when there is samples as nonhemolyzed (group I), moderately hemo-
transfusion or extracorporeal circulation. Hemolysis lyzed (group II), and severely hemolyzed (group ifi) (n =
can result from improper drawing and handling of 60 each).
specimens and can occur during centrifugation and We used a Dacos analyzer with Dart reagents (all
separation procedures. Therefore, clinical chemists from Coulter Electronics, Inc., Hialeah, FL) reagents to
must know whether hemolysis affects the analyses or- measure aspartate aminotransferase (EC 2.6.1.1), ala-
dered by clinicians. Many laboratories reject all hemo- nine aminotransferase (EC 2.6.1.2), urea, creatinine,
lyzed specimens without considering whether this ap- cholesterol, high-density lipoprotein cholesterol (after
proach is justified. precipitation with magnesium phosphotungstate), tri-
Hemolysis in serum is visible when the hemoglobin glyceride, total and direct bilirubin, sodium, and potas-
(Jib) concentration is >3.1 &mol/L (1,2). Hemolysis has sium. We used an Encore analyzer (Baker Instruments
little effect on constituents that are present at lower Co., Allentown, PA) to measure glucose (glucose oxi-
concentrations in erythrocytes than in plasma, but a dase/peroxidase/phenol/ammnoantipyrine chromogenic
marked effect may be observed for constituents that are system; equilibrium), albumin (dye binding with brom-
present at a higher concentration in erythrocytes than cresol green; kinetic), uric acid (uricase/peroxidase/phe-
in plasma. Hb may also interfere directly in the colori- nol/ammoantipyrine chromogenic system; equilibrium),
metric determination of constituents (2-4). alkaline phosphatase (EC 3.1.3.1; 4-nitrophenyl phos-
The effects of hemolysis were thoroughly studied by phate as substrate), and lactate dehydrogenase (EC
several investigators who simulated hemolysis by add- 1.1.1.27; pyruvate-+lactate reaction; ultraviolet, kinet-
ing lysed erythrocytes to plasma pooled from only a few ic). We used a Vitatron SPS spectrophotometer (Vital
subjects (2-6). Here we investigate the effects of hemol- Scientific, Dieren, The Netherlands) to measure calcium
ysis by correlating the results of biochemical tests with (cresolphthalemn complexone; equilibrium), magnesium
concentrations of free Hb in specimens hemolyzed by (with Titan yellow after deproteinization; equilibrium),
mechanical trauma. We think that this type of correla- inorganic phosphate (with acid molybdate after depro-
tion gives data that are more useful for evaluating teiiuzation; equilibrium), amylase (EC 3.2.1.1; amylo-
whether tests are affected by hemolysis. clastic method), lipase (EC 3.1.1.3; olive oil as substrate;
turbidimetric), acid pho8phatase (EC 3.1.3.2; 4-nitro-
Materials and Methods phenyl phosphate as substrate) and its prostatic isoen-
To study the effect of in vitro hemolysis on whole zyme (with tartrate inhibition), and Hb. We measured
blood, we obtained 60 15-mL blood samples from pa- chloride with mercurimetric titration.

Results
Clinical Biochemistry Laboratoiy, YOkaek Ihtisas Hastanesi
(High Specialization Hospital), Sihhiye, Ankara 06100, Thrkey. The mean (and SD) free Jib concentrations for groups
Received September 9, 1991; accepted February 10, 1992. 1,11,and ifi were 0.744(0.496), 6.743(2.155), and 17.608

CLINICAL CHEMISTRY, Vol.38, No.4, 1992 575

(4.247) pmol/L, respectively. Differences of the means tions (total, conjugated, and nonconjugated). In classical
between groups I and II and groups II and ifi were diazotization procedures, Hb in the sample competes
statistically significant (P <0.001). with nitrite for sulfanilic acid.
Differences between groups I and II and between Our results for measurements of the concentrations of
groups I and Ill were evaluated for each test. Only sodium, chloride, calcium, inorganic phosphate, and
lactate dehydrogenase, potassium, acid phosphatase, urea were unaffected by hemolysis, even in the severely
and prostatic acid phosphatase were significantly in- hemolyzed group. Because Jib may mask the endpoint of
creased (P <0.05) by hemolysis in group II. The increase the titration procedure for chloride, positive errors can
in concentrations of these constituents was more signif- result (9). These positive errors may have been compen-
icant for group ifi (P <0.001). Cholesterol concentra- sated for by the dilutional effect of hemolysis. Although
tions also increased significantly in group ifi (0.01 <P the concentration of inorganic phosphate is lower in
<0.05). erythrocytes than in serum, we found a slight increase
We obtained slightly decreased values with hemolysis that was due to hemolysis. Erythrocytes contain a large
for glucose, uric acid, alkaline phosphatase, lipase, and concentration of organic phosphate esters, and the con-
total and direct bilirubin and slightly increased values centration of inorganic phosphate in hemolyzed samples
for aspartate aminotransferase, alanine aminotrans- increases as these esters are hydrolyzed by serum phos-
ferase, high-density lipoprotein cholesterol, triglycer- phatases. Our results show that the dilutional effect of
ides, and inorganic phosphate; however, these changes hemolysis on this assay is unimportant and may be
were not statistically significant (P >0.05). Values for neglected. Concentrations of creatinine, total protein,
total protein, albumin, urea, creatinine, sodium, chlo- albumin, and magnesium, which are greater in eryth-
ride, calcium, magnesium, and amylase did not change. rocytes than in serum, were also unaffected by hemoly-
For each blood sample, we plotted the change in sis. These results agree well with the results from
analyte concentration between the hemolyzed sample previous studies (2-4).
and the nonhemolyzed samples vs the change in Jib Caraway (1) reported that erythrocytes contain -160-
concentration (n = 120). Correlation coefficients be- fold as much lactate dehydrogenase, 68-fold as much
tween these changes were zero for the following tests: acid phosphatase, 40-fold as much aspartate ami-
total protein, albumin, urea, creatinine, high-density notransferase, and 6.7-fold as much alanmne ami-
lipoprotein cholesterol, triglycerides, total and direct notransferase as does plasma. However, of the enzymes
bilirubin, sodium, chloride, calcium, magnesium, and studied, only lactate dehydrogenase and acid phospha-
amylase. The correlation coefficients were not zero for tase showed significant increases because of hemolysis.
the following tests: lactate dehydrogenase (r = 0.789), In contrast to previous studies (3, 4), we found that
potassium (r = 0.774), acid phosphatase (r = 0.627), prostatic acid phosphatase (tartrate-i#{252}hibitedacid phos-
prostatic acid phosphatase (r = 0.539), aspartate ami- phatase) was also affected by hemolysis; perhaps tar-
notransferase (r = 0.543), uric acid (r = 0.374), glucose trate could not inhibit the increase in enzyme activity
(r = 0.373), cholesterol (r = 0.340), alanine aminotrans- resulting from hemolysis. Also in contrast to previous
ferase (r = 0.249), inorganic phosphate (r = 0.213), and studies (3-5), we saw a negligible increase in aspartate
lipase (r = 0.202). aininotransferase. We found lipase to be inhibited by
Discussion hemolysis, but hemolysis had no influence on lipase
activity measured turbidiinetrically (10). The apparent
Our results show that lactate dehydrogenase, acid
decrease in lipase activity with hemolysis noted here
phosphatase, and potassium are significantly affected by
originated from the increase of the absorbance of the
hemolysis,which agrees well with previous investiga- control.
tions (1-4, 6). In contrast to these studies, we could not
find any statistically significant difference caused by
In conclusion, although the effects of hemolysis are
hemolysis for aspartate and alanine aminotransferases,
method-dependent (4), most common biochemical tests
inorganic phosphate, glucose, bilirubins, total protein,
are unaffected by hemolrsis. However, even when sam-
and albumin values.
ples are only slightly hemolyzed, results for lactate
We found a slight decrease in glucose and uric acid.
dehydrogenase, acid phosphatase, prostatic phospha-.
This effect could be expected in reactions involving the
tase, and potassium analyses must be rejected. Al-
hydrogen peroxide/peroxidase/phenoJaminoantipyrine
though the amount of free Jib can be measured and a
system as chromogen, because of premature decomposi-
calculation made to correct test results affected by
tion of hydrogen peroxide by Jib (8). However, we did
hemolysis (1, 3), we find this approach unsuitable.
not find this effect for cholesterol, high-density lipopro-
tein cholesterol, and triglycerides, which are assayed References
with the same chromogemc system. On the contrary, we 1. Caraway WT. Chemical and diagnostic specificity of laboratory
found an increase in cholesterol when there was severe tests. Am J Clin Pathol 1961;37:445-64.
hemolysis. This increase resulted directly from the Jib 2. Sonntag 0. Haemolysis as an interference factor in clinical
and can be compensated for by using an appropriate chemistry. J Clin Chem Clin Biochem 1986;24:127-39.
3. Biydon WG, RobertaLB. The efibet of haemolyais on the determi-
sample blank. nation of plasma constituents. Clin Chim Acta 1972;41:435-8.
We observed a slight decrease in all bilirubin frac- 4. Frank JJ, Berinea EW, Bickel MJ, Watkins BF. Effect of in

576 CLINICALCHEMISTRY,Vol.38, No.4, 1992

vitro hemolysis on chemical values for serum. Clin Chem 1978; Tiets NW, ed. Textbook of clinical chemistry. Philadelphia: WB
24:1966-70. Saunders, 1986:1495-588.
5. Laessig RH, Hassemer DJ, Paskey TA, Schwartz TH. The 8. Perlstein MT, Thibert RJ, Zak B. Bilirubun and hemoglobin
effects of 0.1 and 1.0 percent erythrocytes and hemolysis on serum alterations in several hydrogen peroxide generating procedures.
chemistry values. Am J Clin Pathol 1976;66:639-44. Microchem J 197823:13-27.
6. Randall UG, Garcia-Webb P, Beilby JP. Interference by hae- 9. Tietz NW, Pruden EL, Siggaard-Anderaen 0. Electrolytes. Op.
molysis, icterus and lipaemia in assays on the Beckman Synchron cit. (ref. 7):1172-91.
CX5 and methods for correction. Ann Cliii Biochem 199027:345- 10. Kreutzer HH, Penmngs AW, Punt JMHM, Verduin PA. Fur-
52. ther studies on the determination of lipase activity. Cliii Chim
7. Fairbanks VF, Klee GO. Biochemical aspects of hematology. In: Acta 1975;60:273-9.

CLIN.CHEM.38/4, 577-580 (1992)

Genetic Variants of a1-Antitrypsin and Hemoglobin Typed by Isoelectric Focusing in

Preselected Narrow pH Gradients with PhastSystemTM
Jan-Olof Jeppsson1 and Roland Einarsson2

In this method for automatically running and staining integrated in PhastSystemm (Pharmacia LKB Biotech-
isoelectric focusing (IEF) gels, pre-made dehydrated poly- nology, Uppsala, Sweden), a dedicated system for hori-
acrylamide gels were rehydrated before assays run with zontal electrophoresis and JEF in small gels, with auto-
the PhastSystem (Pharmacia LKB Biotechnology). The mated fixation, staining, and destaining (1,2). Here we
typing of genetic variants of hemoglobin and a1-antit- describe an IEF method performed with PhastSystem
rypsin in narrow pH gradients (pH 6.7-7.7 and 4.2-4.9, with a new dry polyacrylamide gel, which is soaked in a
respectively) was simple, convenient, and reproducible. narrow-pH gradient solvent before use in analysis for
The clinically important variants of a1-antitrypsin (ZZ and genetic variants of a1-antitrypsin and hemoglobin.
SZ) were identified from serum or dried blood on filter
paper. The fast screening of abnormal hemoglobin sam- MaterIals and Methods
ples (HbS) for cases in clinical medicine was easily
Blood and serum samples for typing genetic variants
performed. The total analysis time for the phenotyping
of a1-antitrypsin and hemoglobin were obtained from
with conventional protein staining was -60 mm.
the Department of Clinical Chemistry, General Hospi-
tal, MalmS, Sweden.
Additional Keyphra.es: screening - sickle cell disease
electroplioresis polyaciylamide gel . gycohemogkbin
PhastSystemand Gels
Isoelectric focusing (JEF) of proteins in polyacryl-
The apparatus comprises electrophoresis and auto-
amide gels has been used in various applications in mated staining and destaining units, which can be
clinical and forensic medicine to reveal detailed protein programmed and operated independently (1,2).
microheterogeneity not obtained with other techniques. Dehydrated polyacrylamide gels were rehydrated be-
fore use by applying a droplet of ampholyte solution on
Considerable time, effort, and skill are required to
a plastic surface and either placing the dry gel upside-
achieve acceptable and reproducible results with cur-
rent techniques for JEF, which restricts the use of the down on the droplet for 30 mm or using a reswelling
technique in routine clinical settings. Many develop- casette (Pharmacia LKB Biotechnology). For typing of
ments have been made in these areas, including the use a1-antitrypsin, we applied 1.3 mL of Pharmalyte, 4.2-
4.9, diluted 16-fold with water onto the plastic surface;
of ultrathin gels to shorten the distance for heat trans-
portation so that the voltage can be increased (for for typing of hemoglobin, we applied 1.3 mL of Pharma-
greater resolution), introduction of tailor-made ampho- lyte, 6.7-7.7, also diluted 16-fold with water. A spacer
lines with narrow pH gradients, and plastic backings to molecule (-alanine, 17 mmol/L) was added in some
simplify the handling of the gels during staining and experiments to better separate the glycohemoglobin
fraction from nonglycated hemoglobin (3).
drying.
Most of these new ideas for electrophoresis have been
IEF of a1 -Antitrypsin
Samples were prepared either from dried blood on
‘Department of Clinical Chemistry, General Hospital, Lund filter paper or from serum, essentially as previously
University, 5-21401 MahnO, Sweden. described (4, 5). Briefly, dried blood on filter paper
2Pharmacia Diagnostics AB, S-751 82 Uppeala, Sweden (author
for correspondence). (4-mm-diameter disc) was eluted in 15 pL of 1 mol/L
Received May 21, 1991; accepted February 3, 1992. glycine, pH 7.4, containing cysteine hydrochloride, 0.5

CLINICAL CHEMISTRY, Vol. 38, No. 4, 1992 577