Review Article

Annals of Clinical Biochemistry

2016, Vol. 53(5) 527–538
! The Author(s) 2016
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0004563216643557
acb.sagepub.com

Serum indices: managing assay interference

Christopher-John L Farrell and Andrew C Carter

Abstract
Clinical laboratories frequently encounter samples showing significant haemolysis, icterus or lipaemia. Technical
advances, utilizing spectrophotometric measurements on automated chemistry analysers, allow rapid and accurate
identification of such samples. However, accurate quantification of haemolysis, icterus and lipaemia interference is of
limited value if laboratories do not set rational alert limits, based on sound interference testing experiments.
Furthermore, in the context of increasing consolidation of laboratories and the formation of laboratory networks,
there is an increasing requirement for harmonization of the handling of haemolysis, icterus and lipaemia-affected
samples across different analytical platforms. Harmonization may be best achieved by considering both the analytical
aspects of index measurement and the possible variations in the effects of haemolysis, icterus and lipaemia interfer-
ences on assays from different manufacturers. Initial verification studies, followed up with ongoing quality control
testing, can help a laboratory ensure the accuracy of haemolysis, icterus and lipaemia index results, as well as assist in
managing any biases in index results from analysers from different manufacturers. Similarities, and variations, in the
effect of haemolysis, icterus and lipaemia interference in assays from different manufacturers can often be predicted
from the mechanism of interference. Nevertheless, interference testing is required to confirm expected similarities or
to quantify differences. It is important that laboratories are familiar with a number of interference testing protocols and
the particular strengths and weaknesses of each. A rigorous approach to all aspects of haemolysis, icterus and lipaemia
interference testing allows the analytical progress in index measurement to be translated into improved patient care.

Keywords
Haemolysis, icterus, lipaemia, interference, index, indices
Accepted: 16th March 2016

Introduction
Interference with biochemistry assays due to haemoly-
sis, icterus or lipaemia (HIL) is a common problem in
clinical laboratory practice. Traditionally, the identifi-
cation and quantification of HIL interference has been
performed by visually inspecting for changes in the
colour and/or clarity of serum or plasma. Any redness,
yellowness or turbidity was subjectively classified into a
semi-quantitative grade or HIL ‘index’. However, the
manufacturers of modern biochemistry analysers have
developed instruments that are able to objectively
quantify haemoglobin, bilirubin and turbidity in
serum or plasma samples by the use of multiple spec-
trophotometric measurements. Modern analysers are
therefore able to identify HIL interference much more
accurately than the traditional approach.1
Both commercial assay manufacturers and clinical
laboratories have been slow to update their handling
Laverty Pathology, North Ryde, NSW, Australia
Corresponding author:
Christopher-John Lancaster Farrell, Laverty Pathology, 60 Waterloo
Road, North Ryde, NSW 2113, Australia.
Email: Chris.Farrell@Laverty.com.au
528
of HIL interference to take advantage of this technical
improvement. Indeed, some instrument manufacturers
have their instruments report semi-quantitative HIL
index results to accommodate laboratory scientists
who are comfortable with this traditional style of
reporting. Additionally, the acceptability criterion
used by manufacturers to decide the level at which an
assay shows significant interference is often a trad-
itional generic value (frequently  10%) rather than
the more rational analyte-specific approach advised
by the Clinical and Laboratory Standards Institute
(CLSI).2 Generally, manufacturers will routinely only
provide limited information regarding the details of the
experimental design and results used to establish the
recommendations regarding HIL interference with
their assays. Therefore, a laboratory may have some
difficulty in judging the quality of the information pro-
vided by the manufacturer or deriving its own conclu-
sions from the manufacturer’s data, for instance, by
applying a different acceptability criterion.
Clinical laboratories, too, have been slow to update
their management of HIL interference. It is uncommon
for laboratories to perform their own interference
testing, and the manufacturers’ recommendations for
interference alert limits are adopted by 95% of labora-
tories.3 Laboratories, individually or as networks, are
frequently using biochemistry analysers from different
manufacturers. In this context, the indiscriminate adop-
tion of manufacturers’ recommendations precludes a
rational, harmonized approach to HIL interference.
By modernizing their approach to HIL interference
and placing a greater emphasis on the HIL indices as
analytical results in themselves, laboratory managers
will be able to identify interference more accurately
and harmonize the handling of HIL-affected samples
across a laboratory.
Haemolysis
Haemolysis has been traditionally considered the most
common pre-analytical error affecting biochemistry
results.4 Recent data, however, suggest that modern
sample collection, transport and processing procedures
have made haemolysis a less frequent issue in routine
practice. Analysis of data from 75 laboratories during
2013 found a median rate of haemolysed samples of
0.44%, with a 25th percentile of 0.12% and a 75th per-
centile of 0.85%.5 The incidence with which labora-
tories receive haemolysed samples can vary markedly
depending on where and how their samples are col-
lected. For instance, haemolysis in samples received
from emergency departments has been reported to be
as high as 31%.6 While such data focus on venous sam-
ples, other sample types may also be affected by haem-
olysis at a significant rate. For example, 4% of arterial
Annals of Clinical Biochemistry 53(5)
blood gas samples in a tertiary referral hospital showed
significant haemolysis.7
Haemolysis results from disruption of the membranes
of red blood cells (RBCs). RBCs are robust cells, able to
withstand significant deformational stresses in the micro-
circulation; however, they are more sensitive to tangen-
tial stress.8 Shearing forces beyond 300 Pa cause lysis of
normal RBCs.9 In certain patients, this threshold may be
lowered. Increased mechanical fragility of red cells has
been described in diabetes, multiple sclerosis, as well as
in postmenopausal women.10–12
When blood passes through devices greater than
capillary size, as during blood collection, the most
important factor determining whether haemolysis
occurs is shearing stress, not at the interface of the
blood and solid surfaces, but within bulk blood
undergoing turbulent flow.13 This effect may underlie
many of the causes of haemolysis given in Table 1.
A number of factors related to sample collection
have been reported to be associated with haemolysis,
but may not be causative. For instance, increased rates
of haemolysis have been observed in clinical specimens
when the tourniquet compression time was pro-
longed.21 However, under experimental conditions,
when tourniquet application time is varied during mul-
tiple collections in individual patients, no such effect is
seen.22 Factors such as tourniquet application time
may correlate with difficulty of venesection, where tur-
bulent blood flow is more likely.
Haemolysis may interfere with biochemistry results
through a number of mechanisms. These may be
broadly categorized into factors that change the con-
centration of the analyte in the sample and those that
interfere with analyte measurement. The release of the
RBCs’ intracellular contents will increase the concen-
trations of analytes present in high concentrations within
the cells (such as potassium, aspartate aminotransferase
Table 1. Some of the reported causes of haemolysis.14–20
Collection into syringe with excessive suction applied to plunger
or aspiration against resistance
Collection through an intravenous cannula, especially if:
The cannula is partially obstructed
Blood froths due to a loose connection in the collection
assembly
Syringe transfer into sample tube, particularly if force is used
‘Traumatic’ collection
Site of collection other than antecubital fossa
Use of smaller gauge needles
Errors in handling, including:
Freezing the sample
Vigorously shaking the sample tube
Extracorporeal circulation
In-vivo haemolysis
Farrell and Carter
(AST) and lactate dehydrogenase) and decrease the con-
centrations of analytes present in low concentrations
within the cells. For the latter effect to be clinically sig-
nificant, the analyte must also be controlled within tight
physiological limits. Sodium and chloride are examples of
analytes affected in this manner.23 The release of haemo-
globin, or other intracellular components, may also cause
degradation of the analyte in the sample. Bilirubin is
degraded via the pseudoperoxidase activity of haemoglo-
bin, while intracellular proteases may break down insulin
and troponin T.24–26
Haemolysis may interfere with analyte measurement
via diverse mechanisms. Oxygenated and deoxygenated
haemoglobin have slightly different absorbance spectra
but show maximal absorbance around 415 nm, with
significant absorbance between 320 and 450 nm and
540 and 580 nm.27 Colorimetric assays that use absorb-
ance measurements in one or more of these ranges are
therefore susceptible to interference. Examples include
iron, lipase, albumin and g-glutamyl transferase.23,28
Alkaline phosphatase assays are susceptible to negative
interference because alkali denaturation of haemoglo-
bin may cause a negative offset in absorbance read-
ings.29 Haemolysis may also cause the release of
enzymes that participate in analytical reactions; for
instance, adenylate cyclase causes a positive interfer-
ence with creatine kinase assays.30 It is possible that
multiple mechanisms may influence particular results
in haemolysed samples, especially when the haemolysis
is marked.
Icterus
Icterus refers to the yellow colour resulting from
increased sample bilirubin concentration.2 Raised bili-
rubin concentrations most frequently result from liver
disease; however, some assays may be sensitive to the
relatively mild hyperbilirubinaemia that occurs in the
context of Gilbert’s syndrome or in vivo haemolysis.
Icterus exerts effects on chemistry tests primarily
through spectrophotometric and chemical interfer-
ences. Bilirubin absorbs light between 400 and
540 nm, with a peak around 460 nm.31,32 Colorimetric
assays taking primary or secondary absorbance meas-
urements at these wavelengths may be affected, such as
phosphate and Jaffe creatinine assays.33,34 In Jaffe cre-
atinine assays, the formation of the coloured creatinine-
picrate complex is measured by absorbance readings
taken around 500 nm. Under the alkaline conditions
of the reaction, bilirubin is oxidized to biliverdin,
which causes negative interference by decreasing
absorbance of the sample at 500 nm. A number of stra-
tegies have been used to minimize this interference.
Potassium ferricyanide may be added to oxidize the
bilirubin to biliverdin prior to initiating the Jaffe
529
reaction.35 Another strategy is so-called ‘rate-blanking’,
where the measured rate of colour change in the sample
after addition of sodium hydroxide is used as a correc-
tion factor.36
Bilirubin’s antioxidant properties gives it the ability
to react with hydrogen peroxide,37 which is an add-
itional mechanism of icteric interference. Hydrogen
peroxide is an intermediate in some common chemistry
assays (including cholesterol, triglycerides, uric acid
and enzymatic creatinine), which are therefore suscep-
tible to icteric interference.34,38 Assays may employ a
high-efficiency hydrogen peroxide acceptor and/or
ferrocyanide to try to minimize this interference.34
The unconjugated and conjugated forms of bilirubin
may exert different effects on certain assays.
Conjugated bilirubin has been found to cause the
greater degree of interference with most assays.39
However, this is not always the case; indeed, in some
instances, the unconjugated and conjugated forms of
bilirubin have been shown to cause interference in
opposite directions. The Abbott Aeroset and Roche
Modular triglyceride assays have demonstrated positive
interference by unconjugated bilirubin and negative
interference by conjugated bilirubin.39
Lipaemia
Lipaemia refers to the sample turbidity resulting from
the light scattering effect of a high concentration of
lipoprotein particles. The size of the particles in solu-
tion is a major factor determining the extent of light
scatter. Being the largest lipoprotein particles, with
diameters up to 1000 nm, the accumulation of chylo-
microns is the primary cause of lipaemia; however,
the large (60–200 nm diameter) and intermediate (35–
60 nm diameter) subclasses of VLDL may also contrib-
ute to sample turbidity.40 The most common cause of
lipaemia is a recent fatty meal.41 Lipaemia may also be
produced by high alcohol intake, diabetes mellitus,
renal impairment, total parenteral nutrition, certain
medications and hereditary conditions.42 Lipaemia
has been reported to be the most common interferent
in outpatient populations, occurring in up to 7.4% of
samples.43 Clinical laboratories may encounter an
increasing incidence of lipaemic samples resulting
from greater use of non-fasting collections. For
instance, some newer guidelines for cardiovascular dis-
ease risk assessment now permit the use of non-fasting
lipid measurements, while the acceptance of HbA1c for
diabetes diagnosis may decrease the use of fasting glu-
cose measurements.44,45
Lipaemia can cause interference in biochemistry
results through a variety of mechanisms. Light scatter-
ing, differential partitioning of analyte between the
polar and aqueous phases of the sample, and
530
Absorbance
300 350 400 450
Wavelength (nm)
Figure 1. Visual spectra of haemoglobin ( ), bilirubin ( ) and lipaemia ( ).2,27,32
interaction of the lipoprotein particles with assay
reagents may all cause interference in results.
Lipaemia causes light scattering across the visual spec-
trum (300 to 700 nm) and consequently exerts profound
effects on turbidimetric and nephelometric assays. The
scattering of light shows no specific peaks, but it does
increase as the wavelength decreases. Colorimetric
assays taking absorbance readings at the shorter wave-
lengths of the visual spectrum are therefore most sus-
ceptible to interference. As a result, assays that utilize
changes in NAD(P)H concentration, which is measured
around 340 nm, are susceptible to lipaemia interference.
The presence of lipoprotein particles with large lipid
cores may cause interference because of differential par-
titioning of analytes between the polar and non-polar
phases of the sample. Pseudohyponatraemia is a well-
recognized effect resulting from sodium remaining in
the aqueous phase of the sample. If the sample is
diluted prior to analysis, as in indirect ion-selective elec-
trode and flame photometry methods, an assumption is
made regarding the proportion of the sample that is
aqueous. This assumption does not hold for lipaemic
samples, and the consequence is falsely low results.
Chloride and potassium results are affected by the
same mechanism.46 Differential partitioning may also
cause erroneous results because of the tendency for
chylomicrons to settle to the top of a sample that is
left to stand. The chylomicrons will carry with them
polar analytes, such as steroid hormones and hydropho-
bic drugs. The results of testing such samples will vary
depending on the level of the tube from which the ana-
lyser aspirates (normally near the top of the tube).41 In
addition, there is potential for lipoprotein particles to
cause interference by adsorbing hydrophobic reagents
Annals of Clinical Biochemistry 53(5)
500 550 600 650 700
or reaction products or by non-specific interference
with antigen–antibody binding in immunoassays.47,48
Measurement of the serum indices
All current automated chemistry analysers use the same
principle for HIL index measurements. Multiple spec-
trophotometric absorbance readings are taken, gener-
ally after dilution of the sample. In the simplest
approach, three bichromatic absorbance measurements
are taken. Because of their spectral overlap, correction
calculations are required to account for the contribu-
tion of multiple HIL parameters to absorbance at some
wavelengths. Figure 1 shows the absorbance spectra of
haemoglobin, bilirubin and lipaemia. Lipaemia can
generally be measured without overlap from haemoglo-
bin or bilirubin by taking a primary wavelength of
650 nm or greater, with the secondary measurement at
a longer wavelength. Measurements for haemoglobin,
whether taken around 570 or 415 nm, must take into
account the contribution of lipaemia to absorbance.
This is achieved by applying a correction factor to the
result based on the measured lipaemia of the sample.
Measurements for bilirubin must employ correction
factors to account for both lipaemia and haemoglobin
absorbance. Some instruments take measurements at
more than three pairs of wavelength to calculate the
serum indices. In these instances, the correction equa-
tions are more complicated, but generally involve sum-
ming the difference of each pair of wavelengths
measured and multiplying the sum by a propriety
factor.49 Table 2 details the characteristics of HIL
index measurement for a number of the automated
chemistry platforms in routine use.
 Farrell and Carter
Table 2. Manufacturer information for haemolytic, icteric and lipaemic indices.2,41,50–77

Wavelengths measured (nm) Sample Dilution Manufacturer’s acceptance criterion

volume volume
Lipaemia Haemolysis Icterus (L) Diluent (L) Reporting Sodium CK Albumin

Abbott Architect 500/524; 572/604; 500/524; 5.3 Salinea 200 Semi-quant No No No

572/604; 628/660 572/604; interpretationb interpretationb interpretationb
572/804 628/660
Beckman Coulter AU 660/800 410/480; 480/570; 1.6/2.0 Saline 150 Semi-quant Not stated <10% <10%
600/800 600/800
Beckman Coulter 340, 410, 340, 410, 340, 410, 14 Tris buffer 200 Semi-quant 42 mmol/L or 2% 410 IU/L or 7% 44 g/L or 6%
DxC/Synchron 470, 600, 470, 600, 470, 600,
670 670 670
Ortho Vitros 700 522–750 507–776 35c Nil 0 Quant <4.3 mmol/L <38 IU/L <2 g/L
Roche Integra 659/800 583/629 480/512 5 Saline 125 Quant 410% 410% 410%
Roche Modular 660/700 570/600 480/505 8 Salinea 200 Quant 410% 410% 410%
Siemens Advia 658/694 571/596 478/505 5 Salinea 100 Semi-quant <10% <10% <10%
or quant
Siemens Dimension 700 405/700 452/700 10 Water 165 Semi-quant <10% <10% <10%
Vista
a
Reagent 1 of certain assays may also be used. AST or ALT reagents are recommended.
b
Manufacturer provides the results of interference testing without interpreting these against a particular acceptance criterion.
c
Testing does not consume sample (measurement performed in tip of sample probe).

531
532
The measurement of serum indices is not immune to
interference itself. The correction equations may not
fully account for spectral overlap of HIL absorbance,
particularly in samples showing marked HIL.2,94 There
are limited data published on these effects, and the
magnitude and/or direction of bias is likely to be dif-
ferent for different instruments because of variations in
the wavelengths measured and algorithms used.2 Not
all manufacturers provide information relating to these
effects. Where information is provided, it ranges from
general: ‘samples containing extremely high bilirubin
concentrations or samples showing gross haemolysis
may produce non-specific lipaemic flags’95 to more spe-
cific statements: ‘40 mg/dL bilirubin can have a nega-
tive effect on the haemolytic index. 500 mg/dL intralipid
can have a positive effect on the haemolytic index’.96
Additionally, the presence of other coloured com-
pounds in the sample may interfere with HIL index
results. For instance, Rose Bengal, used in the treat-
ment of melanoma, has been shown to cause positive
interference in the haemolytic index on some instru-
ments97, while Patent Blue V dye, used to identify sen-
tinel lymph nodes during cancer surgery, may cause
positive interference with the lipaemic index and nega-
tive interference with the haemolytic and icteric indi-
ces.98 High concentrations of EDTA, as may occur in
under-filled EDTA tubes, may also cause positive inter-
ference in haemolytic index results.99
Raised lipaemic index results may occur in samples
containing high concentrations of immunoglobulins. A
study examining 202 visually clear samples with extre-
mely high lipaemic index results found all samples to
have significantly elevated immunoglobulins: 87%
monoclonal or biclonal gammopathy and 13% poly-
clonal gammopathy.50 Conversely, however, only
approximately 25% of samples with a monoclonal pro-
tein give a raised lipaemic index.50 The high lipaemic
index only occurs when the protein precipitates prior to
measurement, either during its interaction with the dilu-
ent used for HIL analysis or because it acts as a cryo-
globulin.100 Susceptibility to this effect varies between
different biochemistry analysers.50
The icteric and lipaemic indices have corresponding
analytes in routine biochemical analysis, bilirubin and
triglycerides, respectively. There is excellent correlation
between the icteric index and the formal laboratory
measurement of bilirubin. Indeed, it has been proposed
that the icteric index is used to identify patients with
hyperbilirubinaemia who have not had liver function
tests requested.101 Furthermore, discrepancies between
a patient’s bilirubin result and icteric index may be used
to identify interference in the bilirubin assay. For exam-
ple, paraprotein interference may cause marked false
elevations in bilirubin measured by diazo methods,
but not affect the icteric index.102
Annals of Clinical Biochemistry 53(5)
In contrast, there is a much looser correlation
between the lipaemic index and triglyceride concentra-
tion in patient samples, with correlation coefficients
estimated to be 0.49–0.88.103,104 This poor correlation
is expected because of the large variation in the size and
triglyceride content between, and within, lipoprotein
classes. However, for assays where the interference is
mediated by the turbidity of the sample, the lipaemic
index will more accurately indicate the potential for
interference. The complementary information provided
by triglyceride concentration and lipaemic index,
reported as the logarithm of triglyceride concentration
to lipaemic index ratio, has been proposed as a useful
parameter for discriminating between certain lipopro-
tein disorders, identifying abnormalities of glycerol
metabolism and detecting pre-analytical errors due to
a non-fasting state.105 Additionally, interference in tri-
glyceride assays, as may occur because of high glycerol
concentrations, has been identified when patients have
a markedly elevated triglyceride result, but a normal
lipaemic index.106
There may be a role for information technology
solutions in highlighting potential inference in HIL
index results to laboratory staff. For instance, where
it has been established by a laboratory that a lipaemic
index greater than a certain value may produce a false-
high haemolysis index result, a comment may appear
warning laboratory staff of this effect. Similarly, there
may be value in the laboratory information system
(LIS) monitoring and flagging discrepancies between
triglyceride results and lipaemic indices as well as bili-
rubin results and icteric indices. This may help identify
instances of interference either in the analytes or their
corresponding indices. The latter effect may be particu-
larly helpful because some laboratories may have set up
their LIS to automatically validate results if the analyte
results are normal. Even in laboratories where results
are manually validated, it may be easy for staff to fail to
notice a high lipaemic flag when all the results of the
reported analytes are normal. In such an instance, some
form of LIS notification to staff may allow the identi-
fication of an undiagnosed paraprotein.
The haemolytic index does not have a corresponding
analyte among routine testing panels. However, there
may be value in considering use of the haemolytic index
beyond interference testing. Plasma-free haemoglobin
is a useful parameter in the monitoring of intravascular
haemolysis and has also been found to be a prognostic
marker, superior to procalcitonin and clinical scoring
systems, in the context of severe sepsis.107,108
Accurate HIL index results are important because
they influence the interpretation of laboratory testing.
It is therefore appropriate for laboratories to perform
some basic checks to ensure the accuracy of HIL index
results. The CLSI recommends that laboratories
Farrell and Carter
consider verifying an instrument’s index results prior to
use and perform ongoing quality control (QC).2
Verification checks suggested by the CLSI include com-
paring haemolytic and icteric index results to the direct
analysis of haemoglobin (such as by the cyanomethe-
moglobin method) and both total and direct bilirubin.
An alternative verification method proposed by the
CLSI is to simply add haemolysate, bilirubin or intra-
lipid to patient pools in a range of concentrations
expected to generate HIL flags.2 Although not explicitly
a component of the CLSI verification guide, labora-
tories should also consider evaluating the precision of
an instrument’s index results prior to routine use. This
would also help the verification fulfil the International
Organization for Standardization 151890 document.109
Basic QC material can be prepared by freezing aliquots
of a patient pool that has had haemolysate or intralipid
added. QC for the icteric index can be easily prepared
with commercial bilirubin QC material.2 Such checks
can help to ensure that HIL indices are performing as
expected over time and that there is acceptable agree-
ment in HIL index results between the different instru-
ments used by the laboratory.
Interference testing
It is important that laboratory managers are aware of
the various approaches to testing the susceptibility of
assays to HIL interference. The limitations of the inter-
ference information supplied by assay manufacturers,
as well as the need for many laboratories to manage
HIL interference on different chemistry analysers, make
performing inhouse interference testing all the more
imperative. Evaluation of HIL interference involves
using an experimental protocol that best replicates the
interference effects as they occur in patient samples and
interpreting the results against rational criteria.
Two basic experimental designs are used to evaluate
HIL interference. The potentially interfering substance
may be spiked into samples and the results compared
with unspiked samples. Alternatively, patient samples
containing the interferent may be obtained and results
from the assay being evaluated compared with results
from a highly specific assay not susceptible to interfer-
ence. The latter approach has the advantage of showing
the interference effects as they occur in patient samples.
However, this design is less commonly used because of
the requirement to identify numerous patient samples
demonstrating a wide range of haemolysis, icterus and
lipaemia, plus the requirement for a highly specific meas-
urement procedure. Nevertheless, as mass spectrometry
techniques become increasingly used in routine labora-
tories, this experimental design may become more feas-
ible. CLSI guidelines providing detailed instructions on
performing interference testing are available.78
533
There are a number of considerations when planning
HIL interference testing using protocols based on spik-
ing-in the interferent. In the case of haemolysis, it is
important to add RBC contents to samples rather
than simply haemoglobin. Haemolysates may be pre-
pared by subjecting whole blood to mechanical trauma,
osmotic shock or freezing. The use of mechanical
trauma, which can be achieved by passing a sample
multiple times through a fine-gauge needle, will most
closely mimic the mechanism of most cases of ex vivo
haemolysis.79 A shortcoming of this technique is the
difficulty of producing the desired increments in the
haemolytic index. Osmotic shock of washed RBCs, fol-
lowed by freezing, is the procedure recommended by
the CLSI.78 An additional consideration for analytes,
such as insulin and troponin T, where the mechanism of
haemolysis interference is degradation by intracellular
enzymes, is the duration for which the sample is
exposed to intracellular contents. For these analytes,
it may be important to expose samples to intracellular
enzymes for a similar amount of time to that expected
for routine clinical specimens.
Testing for icteric interference can be done by the
addition of bilirubin. However, it is important that
interference from both the unconjugated and conju-
gated bilirubin forms is tested. Certain assays have
shown interference with one, but not the other, of
these forms;80 other assays have shown positive inter-
ference with one form and negative interference with
the other.39 Interference studies are commonly per-
formed using ditaurobilirubin as a stable alternative
to the monoglucuronide and diglucuronide conjugate
bilirubin forms found in vivo. However, concerns have
been raised that ditaurobilirubin may not always pro-
duce the same interference effects as the glucuronide
conjugate forms.2
Interference from lipaemia is frequently assessed by
the addition of the soy-based emulsion, intralipid, to
samples. However, there are limitations to the use of
intralipid. It consists of phospholipid-rich liposomes
and triglyceride-rich artificial chylomicrons 200–
600 nm in size (mean 345 nm).81 Therefore, intralipid
particles are larger than VLDL particles and have a
smaller range of sizes than chylomicrons.
Additionally, the refractive index of intralipid differs
from lipoprotein particles.81 Consequently, samples
spiked with intralipid may fail to show the same effects
as lipaemic patient specimens. Indeed, lipaemic patient
specimens, but not samples spiked with intralipid, have
demonstrated negative interference for caeruloplasmin,
creatinine, haptoglobin and prealbumin.82,104
Lipaemia interference testing may employ a vari-
ation to the approach of using patient samples.
Rather than comparing results to an interference-free
method, the lipoproteins may be removed and the
534 Annals of Clinical Biochemistry 53(5)

analysis repeated by the routine assay. Lipoprotein taken for setting analyte-specific criteria. Some authors
removal may be achieved by ultracentrifugation, high- have used the criterion that the interference does not
speed centrifugation or treatment of the sample with exceed the total allowable limit of error for the analyte
the commercially available non-ionic cyclodextrin poly- specified by professional bodies, such as the Clinical
mer, lipoclear. Caution needs to be exercised when Laboratory Improvement Amendments (CLIA) or the
using lipoclear, as the agent itself may interfere with Royal College of Pathologists of Australasia.88,89
certain assays, such as sodium and albumin.83 Even The CLSI has taken a more stringent approach. It
with an ideal experimental protocol, however, there considers an interference to be acceptable when the
may be some patient-to-patient variation in assay inter- result is within a total error specification, considering
ference, related to variations in lipoprotein compos- the bias and imprecision of the assay under investigation
ition. Larger lipoprotein particles may exert a greater and the physiological variability of the analyte, in add-
effect on assays affected by differential partitioning, ition to the effect of the interferent. Therefore, the allow-
while the degree of turbidity is largely influenced by able degree of interference is the residual error after the
the number of particles present.84 bias, imprecision and physiological variability have been
It is recommended that when a spiking protocol is subtracted (as variances) from the total allowable error.78
used for interference experiments, at least two different Another approach is to derive an acceptability limit
analyte concentrations are tested.2 Although HIL inter- for interference directly from biological and/or analyt-
ference is often independent of analyte concentration, it ical variation data. A number of equations have been
is not always the case. Haemolysis interference with rou- used based on different considerations. The equation
tine bilirubin and troponin assays has been shown to I < 1.96  (CVA2 þ CVI2)1/2 specifies that the interfer-
vary with analyte concentration.85,86 The CLSI provides ence (I) is less than the central 95% of the variation
recommendations regarding the analyte concentrations in the result caused by the analytical (CVA) and intra-
at which interference testing should be performed.78 In individual biological variation (CVI).49,90 Others have
addition, there may be merit in performing interference used the equation I < 0.375  (CVI2 þ CVG2)1/2, where
testing with each sample type used by the laboratory. CVG is the between-individual biological variation.39
For instance, differences in the effect of haemolysis on This equation was derived using Gaussian statistics
potassium assays have been demonstrated in venous considering the proportion of a population with results
versus capillary samples.87 outside the reference interval (ideally 5%). It identifies
The setting of acceptability limits is of great signifi- an interference as significant if it increases the propor-
cance in the interpretation of results of interference tion of abnormal results to 6.7% or more.91 Others
studies. Manufacturers often use 10% as their criter- have derived the following equation, by assuming the
ion of acceptability for their assay performance claims quality goals of CVA < ½  CVI and minimal system-
with regard to interferences (Table 2). However, this atic error (SE) are met: I < CVI – 1.96  CVA þ SE.92
generic criterion may not be appropriate for all ana- Figure 2 illustrates how the use of different acceptabil-
lytes, and a more rational approach is to set analyte- ity criteria may influence the interpretation of the
specific criteria. Several different approaches have been results from an interference testing experiment.

(a) 160 (b) 250

155 225

200
Sodium result (mmol/L)

150
Creane kinase (U/L)

175
145
150
140
125
135
100

130 75

125 50
0 500 1000 1500 2000 0 250 500 750 1000 1250
Haemolyc index (mg/dL) Haemolyc index (mg/dL)
Sodium 10% Limits CLIA Limits Biological variaon limits CK 10% Limit CLIA Limit Biological variaon limit

Figure 2. The effect of using different acceptability criteria when setting HIL interference alert limits. The interferographs show data
for sodium (a) and creatine kinase (b) from an experiment in which haemolysate was spiked-in to pooled patient serum.94 A number of
examples of acceptability limits are plotted: 10%, allowable limits of error from CLIA, and allowable limits for interference based on
biological variation, using the equation I < 1.96  (CVA2 þ CVI2)1/2.
Farrell and Carter
In practice, a laboratory may consider a hierarchy of
HIL index limits, employing up to three levels. First, a
‘grey zone’ limit may be set. A grey zone index may
occur in one of two contexts: (i) the index result is
above the highest HIL level tested for the analyte
and, therefore, no interference data exist for the result
or (ii) the index result lies above the highest level tested
with no significant interference but below the lowest
level tested with unacceptable interference.2 Second,
the ‘alert index’ is the lowest tested concentration of
haemoglobin, bilirubin or lipaemia at which the analyte
result is altered to a clinically significant degree. Index
results above the alert index will therefore generate an
interference flag.
Some laboratories may choose to report some results
where the HIL indices are above alert levels along with
an accompanying comment explaining the likely effect
of interference on the result.93 Such laboratories may
then consider a third decision level, the ‘hold index’,
above which the interference is present to such a
degree that the result provides no useful information
to the clinician and should not be reported.
Laboratories may also consider modifying the alert
limits depending on analyte concentration, recognizing
that a particular degree of interference may be more or
less significant at different analyte concentrations.2
Managing indices across
different platforms
Managers of laboratories using biochemistry analysers
from multiple manufacturers face an additional layer of
complexity when developing a rational approach to
HIL interference. A similar challenge is faced by
laboratory networks, in which a variety of analytical
platforms are used, when they attempt to harmonize
their handling of HIL interference. In such scenarios,
consideration first needs to be given to how HIL index
results from the different platforms compare and,
second, to how HIL interference affects results from
each analyser.
All current biochemistry analysers use the same prin-
ciple for HIL index measurement. However, bias in
index results from different instruments may emerge
because of differences in the diluent and dilution
factor used, wavelengths measured and the correction
equations employed.110 Despite the ubiquitousness of
HIL testing, there have been few published studies
comparing HIL results from different platforms. The
WEQAS preanalytical proficiency programme has
reported some data on agreement of serum indices.111
The haemolytic index results for a single sample were
tabulated from a total of 144 laboratories. The mean
haemolytic index results for the different instruments
showed good agreement: Abbott Architect 168 mg/dL,
535
Roche Cobas 199 mg/dL, Ortho Vitros 173 mg/dL,
while mean results (n ¼ 57) for a sample spiked with
conjugated bilirubin spiked to a concentration of
320 mol/L showed greater variability between the
methods: Abbott Architect 292 mol/L, Roche Cobas
476 mol/L, Roche Modular 440 mol/L, Siemens
Dimension 684 mol/L. There was a similar amount
of intermethod variability for a sample spiked with
intralipid, mean results (n ¼ 111): Abbott Architect
371 mg/dL, Roche Cobas 527 mg/dL, Ortho Vitros
310 mg/dL, Siemens Advia 483 mg/dL.
The haemolytic index results for a number of meth-
ods have also been compared to a cyanomethaemoglo-
bin reference method. Lippi et al.110 concluded that the
agreement between haemolytic index results from the
Roche Modular, Roche Integra, Siemens Dimension
RxL, Siemens Advia 2400, Siemens Advia 1800,
Olympus AU 680 and Coulter DxC 800 with the refer-
ence method was ‘satisfactory’.110 Their data, however,
did show that the Siemens Advia 2400 and 1800 sys-
tems had a proportional positive bias of approximately
40% compared with the reference method, while the
other quantitative methods (Modular and Integra)
showed close agreement with the reference method.
However, when H index results were normalized to
the instrument-specific alert value (as defined by the
manufacturer or local user), the differences were greatly
reduced; for example, the normalized Advia 2400
results were all within 10% of the reference method.
This illustrated that difference in H index results were
being compensated for by adjustment of the alert limit
for each instrument. The ‘normalization’ process used
in the study simply involved dividing all results from
each instrument by a particular factor, therefore factor-
ing results from different platforms would appear to be
an alternative approach to attempt to harmonize index
results.
The ability of a laboratory to harmonize HIL
index results is greatly limited if one or more of the
analytical platforms reports HIL index results semi-
quantitatively. In this instance, it may be difficult to
alter the limits used to set the semi-quantitative
ranges. Nevertheless, it may still be possible to harmon-
ize reporting among the different systems by instructing
the LIS to report the quantitative index results in a
semi-quantitative manner, using the same index
ranges as used on the platform reporting semi-
quantitatively. Harmonizing index results to a semi-
quantitative method in this manner restricts the ability
to set precise alert limits for each assay. It also limits
the ability to distinguish degrees of HIL above the high-
est semi-quantitative category and selectively report
assay results accordingly.
There may also be differences in HIL index results
from the same analyser at different laboratories because
536
of the use of different method parameters. For example,
some instruments may be set-up to measure indices
after dilution in saline or in a particular assay reagent,
as detailed in Table 2. The Siemens Advia may be set-
up to use saline or alanine aminotransferase (ALT)
reagent 1 as the diluent for HIL analysis. However,
comparison of the use of these two diluents for HIL
analysis on the same instrument showed that the
haemolytic index was about 15% lower when the
ALT reagent was used, compared with the saline as
diluent. The icteric and lipaemic indices showed no sig-
nificant differences between results obtained using the
two diluents.94
The second consideration when managing the indi-
ces across different analytical platforms is whether
assays from different manufacturers are similarly
affected by HIL interference. For analytes where the
mechanism of interference is a change in the concentra-
tion of analyte in the sample, such as the effect of haem-
olysis on sodium and potassium results, the effect of
HIL is expected to be similar on all platforms.
Indeed, this has been found to be the case on the
Advia and Modular instruments for sodium and potas-
sium.94 For assays where the mechanism of interference
is interruption of the measurement process, there is a
much greater potential for dissimilar effects on different
instruments. Icterus, for instance, has been shown to
have no significant effect on some phosphate assays
(Ortho Vitros 250/950 and Hitachi 700/900 platforms),
positive interference with other assays (Technicon
Axon) and negative interference with yet other assays
(Roche Integra and Kone 30i/60i platforms).112
Therefore, there is much greater scope for harmonizing
the handling of HIL-affected samples on assays where
the mechanism is change in the concentration of ana-
lyte in the sample, as opposed to where there is inter-
ference in the measurement process itself.
Conclusions
HIL interference is a common problem in routine
laboratory practice. Recent advances by instrument
manufacturers have enabled the use of rapid and accur-
ate assessment of HIL by spectrophotometric analysis.
However, laboratory practice has generally failed to
take full advantage of the technical progress. The infor-
mation provided by manufacturers regarding HIL
interference with their assays may have limitations;
for instance, inappropriate criteria are often used for
identifying significant interference. Nevertheless,
laboratories frequently adopt the recommendations of
manufacturers for HIL. It is imperative that labora-
tories are able to critically evaluate the information
supplied by manufacturers regarding HIL interference
and perform their own interference studies where this
Annals of Clinical Biochemistry 53(5)
information is inadequate. Laboratory quality can be
improved by treating HIL indices as another quantita-
tive assay. HIL index verification and internal quality
control can ensure the ongoing accuracy of results.
Method comparison studies may also be appropriate
if laboratories, or laboratory networks, use chemistry
analysers from different manufacturers. On the basis of
such studies, factoring HIL results may be considered
to harmonize results from different analysers. Ongoing
method comparison checks may then ensure consist-
ency in how samples affected by HIL interference are
handled by the laboratory, independent of the analyser
on which the sample happens to be measured. A rigor-
ous approach to handling samples demonstrating sig-
nificant HIL will allow the advances achieved by
spectrophotometric measurement of these interferences
to be maximized and, consequently, patient safety
improved.
Acknowledgements
This article was prepared at the invitation of the Clinical Sciences Reviews
Committee of the Association for Clinical Biochemistry and Laboratory
Medicine.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the
research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or
publication of this article.
Ethical approval
Not applicable.
Guarantor
CF.
Contributorship
CF researched literature and wrote the first draft of the manuscript. All authors
planned, reviewed and edited the manuscript and approved the final version.
References
1. Simundic A, Nikolac N, Ivankovic V, et al. Comparison of visual vs. auto-
mated detection of lipemic, icteric and hemolyzed specimens: can we rely on
a human eye? Clin Chem Lab Med 2009; 47: 1361–1365.
2. Smith M, Chan Y, Dolci A, et al. Hemolysis, icterus, and lipaemia/turbidity
indices as indicators of interference in clinical laboratory analysis; approved
guideline. Wayne, PA, USA: Clinical and Laboratory Standards Institute;
2012. Document C56-A.
3. Aslan B, Stemp J, Catomeris P, et al. Clinical chemistry sample interferences
reporting patterns in Ontario laboratories. Clin Chem 2012; 58(Suppl):
A23–A24.
4. Chawla R, Goswami B, Tayal D, et al. Identification of the types of pre-
analytical errors in the clinical chemistry laboratory: 1-year study at G.B.
Pant Hospital. Lab Med 2010; 41: 89–92.
5. Plebani M, Sciacovelli L, Aita A, et al. Performance criteria and quality
indicators for the pre-analytical phase. Clin Chem Lab Med 2015; 53:
943–948.
Farrell and Carter
6. Soderberg J, Jonsson P, Wallin O, et al. Haemolysis index – an estimate of
preanalytical quality in primary health care. Clin Chem Lab Med 2009; 47:
940–944.
7. Salvagno G, Lippi G, Gelati M, et al. Hemolysis, lipaemia and icterus in
specimens for arterial blood gas analysis. Clin Biochem 2012; 48: 143–153.
8. Rand R and Burton A. Mechanical properties of the red cell membrane. I.
Membrane stiffness and intracellular pressure. Biophys J 1964; 4: 115–135.
9. Baskurt O. Red blood cell mechanical stability. Engineering 2012; 5: 8–10.
10. Lippi G, Mercandanti M and Aloe RTG. Erythrocyte mechanical fragility
is increased in patients with type 2 diabetes. Eur J Intern Med 2012; 23:
150–153.
11. Schauf C, Frischer H and Davis F. Mechanical fragility of erythrocytes in
multiple sclerosis. Neurology 1980; 30: 323–325.
12. Raval J, Water J, Seltsam A, et al. Menopausal status affects the suscep-
tibility of stored RBCs to mechanical stress. Vox Sang 2011; 100: 418–421.
13. Nakahara T and Yoshida F. Mechanical effects on rates of hemolysis. J
Biomed Mater Res 1986; 20: 363–374.
14. Baer D, Ernst D, Willeford S, et al. Investigating elevated potassium
values. MLO Med Lab Obs 2006; 2438: 30–31.
15. Carraro P, Servidio G and Plebani M. Hemolyzed specimens: a reason for
rejection or a clinical challenge? Clin Chem 2000; 43: 306–307.
16. Berg JE, Ahee P and Berg JD. Variation in phlebotomy techniques in
emergency medicine and the incidence of haemolysed samples. Ann Clin
Biochem 2011; 48: 562–565.
17. Lippi G, Blanckaert N, Bonini P, et al. Haemolysis: an overview of the
leading cause of unsuitable specimens in clinical laboratories. Clin Chem
Lab Med 2008; 46: 764–772.
18. Dugan L, Leech L, Speroni K, et al. Factors affecting hemolysis rates in
blood samples drawn from newly placed IV sites in the emergency depart-
ment. J Emerg Nurs 2005; 31: 338–345.
19. Heyer N, Derzon J, Winges L, et al. Effectiveness of practices to reduce
blood sample hemolysis in EDs: a laboratory medicine best practices sys-
tematic review and meta-analysis. Clin Biochem 2012; 45: 1012–1032.
20. Shah K, Idrovo J, Nicastro J, et al. A retrospective analysis of the incidence
of hemolysis in type and screen specimens from trauma patients. Int
J Angiol 2009; 18: 182–183.
21. Burns E and Yoshikawa N. Hemolysis in serum samples drawn by emer-
gency department personnel versus laboratory phlebotomists. Lab Med
2002; 33: 378–380.
22. Statland B, Bokelund H and Winkel P. Factors contributing to intra-indi-
vidual variation of serum constituents: 4. Effects of posture and tourniquet
application on variation of serum constituents in healthy subjects. Clin
Chem 1974; 20: 1513–1519.
23. Lippi G, Salvagno G, Montagnana M, et al. Influence of hemolysis on
routine clinical chemistry testing. Clin Chem Lab Med 2006; 44: 311–316.
24. Shull B, Lees H and Li P. Mechanism of interference by hemoglobin in the
determination of total bilirubin. I. Method of Malloy-Evelyn. Clin Chem
1980; 26: 22–25.
25. Chevenne D, Letailleur A, Trivin F, et al. Effect of hemolysis on the con-
centration of insulin in serum determined by RIA and IRMA. Clin Chem
1998; 44: 354–356.
26. Sodi R, Darn S, Davison A, et al. Mechanism of interference by haemolysis
in the cardiac troponin T immunoassay. Ann Clin Biochem 2006; 43: 49–56.
27. Reddy P, Bowie L and Callistein S. Binding of nitric oxide to thiols and
hemes in hemoglobin H: implications for alpha-thalassemia and hyperten-
sion. Clin Chem 1997; 43: 1442–1447.
28. Grafmeyer D, Bondon M, Manchon M, et al. The influence of bilirubin,
haemolysis and turbidity on 20 analytical tests performed on automatic
analysers. Results of an interlaboratory study. Eur J Clin Chem Clin
Biochem 1995; 33: 31–52.
29. Chance J, Norris E and Kroll M. Mechanism of interference of a polymer-
ized hemoglobin blood substitute in an alkaline phosphatase method. Clin
Chem 2000; 46: 1331–1337.
30. Kroll M and Elin R. Interference with clinical laboratory analyses. Clin
Chem 1994; 40: 1995–2005.
31. Berde C, Hudson B, Simoni R, et al. Human serum albumin. Spectroscopic
studies of binding and proximity relationships for fatty acids and bilirubin.
J Biol Chem 1979; 254: 391–400.
32. Lamola A, Bhutani V, Wong R, et al. The effect of hematocrit on the
efficacy of phototherapy for neonatal jaundice. Pediatr Res 2013; 74:
54–60.
33. Duncanson G and Worth H. Pseudohypophosphataemia as a result of
bilirubin interference. Ann Clin Biochem 1990; 27: 253–257.
537
34. Peake M and Whiting M. Measurement of serum creatinine - current status
and future goals. Clin Biochem Rev 2006; 27: 173–184.
35. O’Leary N, Pembroke A and Duggan PF. A simplified procedure for
eliminating the negative interference of bilirubin in the Jaffe reaction for
creatinine. Clin Chem 1992; 38: 1749–1751.
36. Hoffman RJ and Feld RD. A modified procedure for creatinine (CRT
which eliminates bilirubin (BILT) interference on the Hitachi 737
(Abstract). Clin Chem 1988; 34: 1313.
37. Kroll M and McCudden C. Endogenous interferences in clinical laboratory
tests, 1st ed. Berlin: Walter de Gruyter, 2013.
38. Spain M and Wu A. Bilirubin interference with determination of uric acid,
cholesterol, and triglycerides in commercial peroxidase-coupled assays, and
the effect of ferricyanide. Clin Chem 1986; 32: 518–521.
39. Uldall A, Loikkanen M, Olafsdottir E, et al. Nordic interference study:
effects of haemoglobin and bilirubin on some common serum analysis.
Report, Nordic Committee for External Quality Assurance Programmes
in Laboratory Medicine, 2002.
40. Kroll M. Evaluating interference caused by lipemia. Clin Chem 2004; 50:
1968–1969.
41. Nikolac N. Lipemia: causes, interference mechanisms, detection and man-
agement. Biochem Med (Zagreb) 2014; 24: 57–67.
42. Calmarza P and Cordero J. Lipemia interferences in routine clinical bio-
chemistry tests. Biochem Med (Zagreb) 2011; 21: 160–166.
43. Ryder K, Glick M and Glick S. Incidence and amount of turbidity,
hemolysis, and icterus in serum from outpatients. Lab Med 1991; 22:
415–418.
44. National Institute for Health and Care Excellence. Cardiovascular disease:
risk assessment and reduction, including lipid modification. NICE Guidelines
CG181, 2014.
45. World Health Organization. Use of glycated haemoglobin (HbA1c) in the
diagnosis of diabetes mellitus: abbreviated report of a WHO consultation.
WHO/NHM/CHP/CPM/11.1, January 2011.
46. Dimeski G, Mollee and Carter A. Effects of hyperlipidemia on plasma
sodium, potassium, and chloride measurements by an indirect ion-selective
electrode measuring system. Clin Chem 2006; 52: 155–156.
47. Creer M and Ladenson J. Analytical errors due to lipemia. Lab Med 1983;
14: 351–355.
48. Schiettecatte J, Anckaert E and Smitz J. Interferences in immunoassays.
In: Chiu NHL (ed.) Advances in immunoassay technology. Rijeka, Croatia:
Intech, 2012, pp.45–62.
49. Dolci A and Panteghini M. Harmonization of automated hemolysis index
assessment and use: is it possible? Clin Chim Acta 2014; 432: 38–43.
50. Fliser E, Jerkovic K, Vidovic T, et al. Investigation of unusual high serum
indices for lipemia in clear serum samples on Siemens analyers dimension.
Biochem Med (Zagreb) 2012; 22: 352–362.
51. McTaggart M, Cannon L and Kearney E. Effect on of Patent Blue V due
on sample interference indices on the Abbott Architect. Ann Clin Biochem
2012; 49: 510–511.
52. Abbott Laboratories. Architect sample interference indices, saline protocol
(product information). Abbott Park, IL, USA: Abbott Laboratories, 2007.
53. Abbott Laboratories. Architect: creatine kinase (product information).
Abbott Park, IL, USA: Abbott Laboratories, 2012.
54. Abbott Laboratories. Architect: albumin BCG (product information).
Abbott Park, IL, USA: Abbott Laboratories, 2012.
55. Beckman Coulter Inc. AU analyser: ISE reagent/standards (product infor-
mation). Brea, CA, USA: Beckman Coulter Inc, 2013.
56. Beckman Coulter Inc. AU analyser: albumin (product information). Brea,
CA, USA: Beckman Coulter Inc, 2009.
57. Beckman Coulter Inc. AU analyser: CK (NAC) (product information).
Brea, CA, USA: Beckman Coulter Inc, 2012.
58. Beckman Coulter Inc. Synchron systems: serum index (product informa-
tion). Brea, CA, USA: Beckman Coulter Inc, 2014.
59. Beckman Coulter Inc. Synchron systems: creatine kinase N-acetyl-L-
cysteine (product information). Brea, CA, USA: Beckman Coulter Inc,
2015.
60. Beckman Coulter Inc. Synchron systems: sodium (product information).
Brea, CA, USA: Beckman Coulter Inc, 2015.
61. Beckman Coulter Inc. Synchron systems: albumin (product information).
Brea, CA, USA: Beckman Coulter Inc, 2015.
62. Ortho-Clinical Diagnostics Inc. Vitros chemistry products Na slides (prod-
uct information). High Wycombe, UK: Ortho-Clinical Diagnostics Inc,
2010.
538
63. Ortho-Clinical Diagnostics Inc. Vitros chemistry products CK slides (product
information). High Wycombe, UK: Ortho-Clinical Diagnostics Inc., 2004.
64. Ortho-Clinical Diagnostics Inc. Vitros chemistry products ALB slides (prod-
uct information). High Wycombe, UK: Ortho-Clinical Diagnostics Inc.,
2004.
65. Roche Diagnostics. Cobas Integra 400/700/800: serum index gen 2 (product
information). Mannheim: Roche Diagnostics GmbH, 2006.
66. Roche Diagnostics. Cobas Integra 400/800: ISE indirect (product informa-
tion). Mannheim: Roche Diagnostics GmbH, 2010.
67. Roche Diagnostics. Cobas Integra 400/800: creatine kinase liquid reagent
(product information). Mannheim: Roche Diagnostics GmbH, 2011.
68. Roche Diagnostics. Cobas Integra 400/700/800: albumin gen 2 (product
information). Mannheim: Roche Diagnostics GmbH, 2007.
69. Roche Diagnostics. Modular analytics SWA: D/P modular – serum index
principles (user guide). Mannheim: Roche Diagnostics GmbH. Ref COBI-
CD ver 2.0, 2004.
70. Roche Diagnostics. Cobas: ISE indirect NA-K-Cl for gen 2 (product infor-
mation). Mannheim: Roche Diagnostics GmbH, 2014.
71. Roche Diagnostics. Cobas: CK (product information). Mannheim: Roche
Diagnostics GmbH, 2014.
72. Roche Diagnostics. Cobas: albumin (product information). Mannheim:
Roche Diagnostics GmbH, 2013.
73. Siemens Healthcare Diagnostics Inc. Advia Chemistry Systems: new serum
indices normalized for Advia Chemistry Systems (customer bulletin).
Tarrytown, NY, USA: Siemens Healthcare Diagnostics, 2008.
74. Siemens Healthcare Diagnostics Inc. Advia Chemistry Systems: sodium
(product information). Tarrytown, NY, USA: Siemens Healthcare
Diagnostics, 2015.
75. Siemens Healthcare Diagnostics Inc. Advia Chemistry Systems: creatine
kinase (CKNAC) (product information). Tarrytown, NY, USA: Siemens
Healthcare Diagnostics, 2011.
76. Siemens Healthcare Diagnostics Inc. Advia Chemistry Systems: albumin
(product information). Tarrytown, NY, USA: Siemens Healthcare
Diagnostics, 2011.
77. Smith M. Dimension Vista System: sample quality assessment for hemolysis,
icterus and lipemia (technical bulletin). Tarrytown, NY, USA: Siemens
Healthcare Diagnostics, 2012.
78. McEnroe R, Burritt M, Powers D, et al. Interference testing in clinical
chemistry; approved guideline, 2nd ed. Wayne, PA, USA: Clinical and
Laboratory Standards Institute, 2005. [Document EP7-A2].
79. Dimeski G. Effects of hemolysis on the Roche ammonia method for
Hitachi analyzers. Clin Chem 2004; 42: 976–977.
80. Alvarex F, Whalen K and Scott M. Conjugated, but not unconjugated,
bilirubin negatively interferes in Hitachi 747 assay of inorganic phos-
phorus. Clin Chem 1993; 39: 2345–2346.
81. Wabel C. Influence of lecithin on structure and stability of parenteral fat
emulsions. PhD Thesis. Nuremberg, Germany, 1998.
82. Hortin G and Goolsby K. Lipemia interference with a rate-blanked cre-
atinine method. Clin Chem 1997; 43: 408–410.
83. Roberts C and Cotten S. Cyclodextrin removal of lipemic interference: an
attractive alternative to ultracentrifugation for satellite laboratories. Arch
Pathol Lab Med 2013; 137: 1027–1028.
84. Kricka LJ. Optical techniques. In: Burtis CA, Ashwood ER and Bruns DE
(eds) Tietz textbook of clinical chemistry and molecular diagnostics, 4th ed.
St Louis: Elsevier Saunders, 2006, pp.61–91.
85. Kroll M, Ruddel M, Blank D, et al. A model for assessing interference. Clin
Chem 1987; 33: 1121–1123.
86. Florkowski C and Wallace Jand George P. The effect of hemolysis on
current troponin assays – a confounding preanalytical variable? Clin
Chem 2010; 56: 1195–1197.
87. Oostendorp M, van Solinge W and Kemperman H. Potassium but not
lactate dehydrogenase elevation due to in vitro hemolysis is higher in capil-
lary than in venous blood samples. Arch Pathol Lab Med 2012; 136:
1262–1265.
88. Koseoglu M, Hur A, Atay A, et al. Effects of hemolysis interference on
routine biochemistry parameters. Biochem Med (Zagreb) 2011; 21: 79–85.
89. Jones G. interferences – are some methods better than others? http://
www.sydpath.stvincents.com.au/other/AACB2003.htm (2005, accessed 31
July 2015).
Annals of Clinical Biochemistry 53(5)
90. Fraser C. Test result variation and the quality of evidence-based clinical
guidelines. Clin Chim Acta 2004; 346: 19–24.
91. Fraser C. Biological variation: from principles to practice. 1st ed.
Washington, DC: American Association for Clinical Chemistry, 2001.
92. Fuentes-Arderiu X and Fraser C. Analytical goals for interference. Ann
Clin Biochem 1991; 28: 393–395.
93. Carraro P. Potassium report of hemolyzed serum samples. Clin Chem Lab
Med 2008; 46: 425.
94. Farrell C. Implementing HIL on ADVIA automation,
www.healthcare.siemens.com.au/siemens_hwem-hwem_ssxa_websites-
context-root/wcm/idc/groups/public/@au/@lab/documents/download/
mda1/mde1/edisp/implementing-hil-on-advia-automation-dr-chris-far-
rell-01959230.pdf (2014, accessed 15 July 2015).
95. Siemens Healthcare Diagnostics Inc. Setting up a dedicated serum indices
method (customer bulletin). Tarrytown, NY, USA: Siemens Healthcare
Diagnostics, 2008.
96. Beckman Coulter Inc. LIH assay (product information). Brea, CA, USA:
Beckman Coulter Inc, 2012.
97. Dimeski G, Jones B and Ungerer J. Interference from Rose Bengal with
total bilirubin measurement. Clin Chem 2009; 55: 1040–1041.
98. Derby D and Broomhead C. Interference with serum indices measure-
ment, but not chemical analysis, on the Roche Modular by Patent Blue
V. Ann Clin Biochem 2008; 45: 289–292.
99. Unger J, Filippi G and Patsch W. Measurements of free hemoglobin and
hemolysis index: EDTA- or lithium-heparinate plasma? Clin Chem 2007;
53: 1717–1718.
100. Monk C, Wallage M, Wassell J, et al. A monoclonal protein identified by
an anomalous lipaemia index. Ann Clin Biochem 2009; 46: 250–252.
101. Salina M, Lopez-Garrigos M, Lugo J, et al. Diagnostic accuracy of icteric
index to detect abnormal total bilirubin values. J Cin Pathol 2012; 65:
928–933.
102. Pantanowitz L, Horowitz G, Upalakalin J, et al. Artifactual hyperbilir-
ubinemia due to paraprotein interference. Arch Pathol Lab Med 2003; 127:
55–59.
103. Twomey P, Don-Wauchope A and McCullough D. Unreliability of tri-
glyceride measurement to predict turbidity induced interference. J Clin
Pathol 2003; 56: 861–862.
104. Bornhorst J, Roberts R and Roberts W. Assay-specific differences in
lipemic interferences in native and intralipid-supplemented samples. Clin
Chem 2004; 50: 2197–2201.
105. De Haene H, Taes Y, Christophe A, et al. Comparison of triglyceride
concentration with lipaemic index in disorders of triglyceride and glyercol
metabolism. Clin Chem Lab Med 2006; 44: 220–222.
106. Speeckaert M, Segers H, Van Biesen W, et al. An unusual case of (pseu-
do)hypertriglyceridaemia. NDT Plus 2010; 3: 570–572.
107. Schrier S, Mentzer W and Tirnauer J. Approach to the diagnosis of hemo-
lytic anemia in the adult. UpToDate, www.uptodate.com/contents/
approach-to-the-diagnosis-of-hemolytic-anemia-in-the-
adult?source¼search_result&search¼Approachþtoþdiagnosisþofþhae-
molyticþanemiaþinþtheþadult&selectedTitle¼1%7E150 (2015, accessed
21 July 2015).
108. Adamzik M, Hamburger T, Petrat F, et al. Free hemoglobin concentra-
tion in severe sepsis: methods of measurement and prediction of outcome.
Crit Care 2012; 16: R125.
109. National Association of Testing Authorities, Australia. Interpretation of
NPAAC Requirements and ISO 15189: medical testing field application
document. Australia: Author, 2013.
110. Lippi G, Salvagno G, Blanckaert N, et al. Multicenter evaluation of the
hemolysis index in automated clinical chemistry systems. Clin Chem Lab
Med 2009; 47: 934–939.
111. Davies G. WEQAS., www.weqas.com/download/2014-scientific-session-
eqa-of-serum-indices/ (2015, accessed 26 July 2015).
112. Nordic Committee for External Quality Assurance Programmes in
Laboratory Medicine. Nordic interference study, http://doc.noklus.no/
handler.ashx?r=nkk&id=Prosjekter%20NQLM_Bilirubin_2001_
20011130.pdf (2001, accessed 5 February 2016).