Pseudohyperkalemia in capillary blood samples; haemolysis not the only culprit

In vitro haemolysis is the most important pre-analytical cause for interference
in clinical laboratory blood tests1. Several parameters, like lactate
dehydrogenase (LDH), Aspartate aminotransferase (ASAT) and potassium are
spuriously elevated due to in vitro haemlysis.

Potassium is the major intracellular cation with roughly a 20-30 fold higher
concentration inside the cell in comparison to the extracellular compartment.
The sodium potassium ATP-ase is the major cellular enzyme responsible for
maintaining this gradient. Due to this large difference in potassium
concentration erythrolysis can have a substantial effect on the plasma
potassium concentration (pseudohyperkalemia) in the clinical laboratory. Since
both hyperkalemia and hypokalemia can cause dangerous cardiac arrest both
pseudohyperkalemia and pseudonormokalemia (masking real hypokalemia), may cause
dangerous over- or under treatment respectively. This phenomenon is poorly
understood by physicians, nurses and even laboratory technicians2.

Many modern laboratory analyzers have automated capability of measuring the
free hemoglobine concentration spectrophotometrically, the haemolyses-index
(H-index). The way the H-index is reported differs per manufacturer. For
example Beckman analysers report the H-index in a scale from 1 to 10 whereby
each level corresponds with a haemoglobin concentration range. Others (Roche
Diagnostics, Cobas) report the H-index as a concentration in µmol/l. Based on
the SKML external quality assessment of 2011, the Cobas Roche analyser is the
most common platform for determining potassium concentrations in Dutch
hospitals. Moreover a recent multicenter evaluation has shown that the H-index
on the Cobas Roche analyzer is highly reproducible between laboratories.

There is ambiguity in dealing with in vitro haemolysis when reporting potassium
results. Some laboratories add a qualitative remark (e.g. haemolysed sample) to
the laboratory potassium result which varies with the H-index magnitude. When
the H-index is beyond a predetermined threshold a *sample recollection* request
is added as a comment and no potassium result is reported. On the other hand
our laboratory and others use the H-index to correct potassium results with a
correction factor. This correction factor is based on a whole blood in vitro
lysate. This lysate is prepared by pooling whole blood which is then haemolysed
by a freeze-thaw cycle and centrifuged. The supernatant is serially diluted in
a non-haemolysed patient plasma sample. This will give a linear relation
between the change in potassium concentration and the H-index. With regression
analysis a slope is calculated and thus a correction factor determined. The
reported potassium is calculated as such; K+reported = K+measured - (H-index
*correction factor).

K+ = mmol/l
H-index = µmol/l hemoglobine
Correction factor = mmol K+/µmol hemoglobine

Using correction factors for estimating potassium concentration is
controversial. One of the main reasons is the broad range of the correction
factors proposed in literature. A significant part of this variability is
attributable to the inter-individual variation in intracellular erythrocyte
haemoglobin concentration. Another important argument against correction is
that in vivo haemolysis (e.g. artificial heart valves, DIS, AIHA) will be seen
as an in vitro haemolysed sample and unjustly corrected.

The golden standard sample for laboratory blood diagnostics is a venous sample.
In neonates capillary sampling is a very common technique to collect blood for
routine laboratory diagnostics, including the potassium status. Capillary
sampling is also performed in adults when venous sampling is problematic such
as in obese patients and patients with fear of needles. In certain laboratories
capillary samples are corrected for haemolysis in the method described above.
There is limited literature available which investigates potassium correction
factors and pseudohyperkalemia in capillary blood samples. It was even
addressed as one of the shortcomings in a recent study.

Our hypothesis is that pseudohyperkalemia in capillary samples is caused by:
1. Haemolysis. In lege artis capillary sampling mild pressure (squeezing) is
used as the driving force for the expulsion of blood. This mild pressure will
not only expel blood from the puncture opening but also induce haemolysis.
2. Tissue fluids. The puncture will cause tissue damage and thus leakage of
intracellular fluids (high in potassium) which contaminate the sample. This
contamination will have a large effect on capillary potassium concentration due
to the small sample volume. With this leakage, unlike erythrolysis, no
haemoglobin is released and therefore this cannot be corrected with a H-index.
It is generally considered that this first drop contains the vast majority of
tissue damage fluids. This contamination is neglectable in a venous blood
sample due to the large volume drawn.
3. Shear stress. This is a phenomenon whereby the erythrocyte membrane is
deformed but stays intact. Due to this deformation small electrolytes, unlike
large proteins like haemoglobin, can pass through the cellular membrane. The
mild force used in capillary sampling to expel blood out of the puncture site
will cause shear stress in the erythrocytes.

In other words we believe that there are multiple sources of potassium, other
than haemolysis, in capillary samples. Given the fact that these sources have
no relation with a change in haemoglobin concentration, the H-index cannot be
used to correct potassium results.

Primary objectives:
• Establish that capillary plasma samples are unsuitable to evaluate the
potassium status in patients.
• Verify that pseudohyperkalemia in capillary samples is not only caused by
haemolysis but also due to other potassium sources e.g. tissue damage, shear
stress. If true, this will make the use potassium correction factors, based on
in-vitro haemolysis alone, invalid for capillary blood samples.


Secondary objective:
• Verify that the contribution of potassium from sources other than haemolysis
in pseudohyperkalemia is more substantial in incorrect capillary sampling than
with lege artis capillary sampling.

Cross-sectional study in twenty adult patients. From each volunteer we will
take four capillary samples.

The first sample will be performed correctly (lege artis).
A second capillary sample will be collected from another finger but now the
first drop will not be wiped away.
A third capillary sample will be collected from another finger using a milking
technique.
A fourth capillary sample will be collected from another finger wherby the
first drop is not wiped away and a milking technique is applied.


Venous plasma (400 µl) will be taken from the vacutainer which will already be
collected for routine laboratory diagnostics. This means no extra phlebotomy
and no extra venous blood sample will be necessary from the volunteers.

All blood samples will be obtained by trained and certified phlebotomists from
the Sint Franciscus Gasthuis. Sampling volunteers will be completed in a time
frame of one month.

Burden and risk associated with a capillary blood sample out of a finger are
neglectable. Futher, we like to stress that incorrect capillary sampling (not
wiping first drop and/or milking the finger) will not increase the burden or
risk for the patient.