Spurious hyperkalaemia due to EDTA contamination: common and not always easy to identify

Abstract

Background

To study the detection and prevalence of spurious hyperkalaemia due to potassium ethylenediaminetetra-acetic acid (kEDTA) contamination.

Methods

In a one-month prospective study, serum EDTA, zinc, calcium, magnesium concentrations and alkaline phosphatase activity were measured in samples with serum potassium ≥6.0 mmol/L.

Results

Twenty-eight out of 117 hyperkalaemic samples were contaminated with EDTA. Only serum zinc values below the reference range had 100% sensitivity for indicating EDTA contamination, but even at an optimal specificity of 89% at least 12 potentially genuine hyperkalaemic samples would be rejected.

Conclusion

Spurious hyperkalaemia due to kEDTA contamination is common. Gross kEDTA contamination is obvious by marked unexpected hyperkalaemia, hypocalcaemia, hypomagnesaemia and hypozincaemia. Spurious hyperkalaemia due to low concentrations of kEDTA contamination can only be confidently detected by measurement of serum EDTA.

Introduction

Spurious hyperkalaemia due to preanalytical factors is well recognized and includes delayed separation, familial pseudohyperkalaemia, in vitro haemolysis, thrombocytosis, leucocytosis and contamination with potassium ethylenediaminetetra-acetic acid (kEDTA). kEDTA is widely used as a sample tube anticoagulant for laboratory assays. ^1,2 In vitro kEDTA contamination of lithium heparin, gel-containing or plain tubes results in artefactually high potassium concentrations and low magnesium, calcium, zinc and alkaline phosphatase values. ^2–4 Of these, spurious hyperkalaemia is the most problematic inevitably leading to patient inconvenience, but may also lead to mismanagement of the patient. ^3,5

There is little data in the literature relating to the prevalence of sample contamination with EDTA. As part of a service evaluation, we measured EDTA in hyperkalaemic samples to identify EDTA contamination and further analysed all the samples for calcium, zinc, magnesium and alkaline phosphatase to study whether these could act as surrogate markers to reliably identify EDTA contamination.

Methods

Over a one-month period, concentrations of EDTA, calcium, zinc and magnesium, and activity of alkaline phosphatase were measured in hyperkalaemic serum samples (potassium ≥6.0 mmol/L).

Blood samples for routine biochemistry are collected into Sarstedt serum/z4 gel tubes using the Sarstedt Safety–Monovette system (Sarstedt Safety Monovette serum Z/4 2, Aktiengesellschaft and Co, Nümbrecht, Germany). Exclusion criteria included samples which were haemolysed, ‘left-on cells’, of insufficient volume, unlabelled and mislabelled.

Analytes were measured using the Roche MODULAR^® analyser (Roche Diagnostics GmbH, Mannheim, Germany). Serum EDTA was measured as previously described. ⁶ The EDTA assay has a detection limit of 0.1 mmol/L with respective intra-assay and inter-assay % coefficients of variation (%CVs) of 3.2 and 6.7 at 0.25 mmol/L.

Potassium was measured by an indirect ion selective electrode. Magnesium, calcium and alkaline phosphatase were measured using colorimetric assays supplied as kits from Roche (Roche Diagnostics GmbH, Mannheim, Germany). Zinc was measured using a colorimetric kit assay supplied by Wako (Wako Chemicals GmbH, Neuss, Germany). All CVs for these tests were 1.5% or less excepting an inter-assay %CV of 4.2 for zinc at a concentration of 14.3 μmol/L.

Data were normally distributed. Results are expressed as means with ranges in parentheses. Unpaired t-test was to assess the significance of between-group variables. Pearson's linear correlation was used to measure the significance of association between EDTA concentrations and concentrations of calcium, zinc and magnesium as well as alkaline phosphatase activity. Data were analysed using the statistical package GraphPad Instat version 3.00 for Windows 95 (GraphPad Software, San Diego, USA).

Results

In the one-month period studied, out of 28,471 samples for ‘urea and electrolytes’ there were 117 samples with a serum potassium ≥6.0 mmol/L. Of these, 28 had an EDTA concentration of >0.1 mmol/L. Of these, 18 were immediately identified as and reported as possible EDTA contamination by laboratory staff prior to EDTA measurement, whereas 10 were initially reported as hyperkalaemia and were subsequently identified as contaminated with EDTA. Twenty-seven of these patients were re-tested and all had repeat serum potassium concentrations within the reference range. Twenty-two (79%) and six (21%) of contaminated samples originated from inpatients and patients in primary care, respectively.

Serum potassium (16.8 [6.8–45.6] mmol/L versus 6.8 [6.1–8.9] mmol/L; P = 0.0174) and EDTA (0.73 [0.35–1.76 mmol/L versus 0.27 [0.11–0.41] mmol/L; P = 0.0014) were significantly higher in the 18 samples immediately identified as contaminated compared with those later identified as EDTA contamination following additional EDTA measurements.

Serum values of EDTA correlated positively with those of potassium (r = 0.4933; P = 0.008) and negatively with those of calcium (r = −0.7164; P = <0.0001), zinc (r = −0.5701; P = 0.0029), magnesium (r = −0.7080; P = <0.0001), and alkaline phosphatase (r = 0.5321; P = 0.0043).

The diagnostic sensitivity and specificity of serum calcium, zinc, magnesium and alkaline phosphatase values in detecting possible EDTA contamination are shown in Table 1.

Table 1

Usefulness of surrogate markers at the lower limit of their reference ranges for detecting ethylenediaminetetra acetic acid contamination

Analyte	Cut-off value	Sensitivity (%)	Specificity (%)	Reference range
Calcium	2.10 mmol/L	86	92	2.10–2.65 mmol/L
Zinc	11 μmol/L	100	63	11–25 μmol/L
Magnesium	0.70 mmol/L	80	95	0.70–1.20 mmol/L
Alkaline phosphatase	38 IU/L	29	99	38–126 IU/L

Discussion

In our study, spurious hyperkalaemia due to kEDTA contamination was common, accounting for about 25% of all ‘hyperkalaemic’ samples. In the 27 patients re-tested, serum potassium was within the reference range confirming EDTA contamination.

In vitro kEDTA contamination usually occurs by one of three possible mechanisms. Direct transfer of blood from kEDTA-containing tubes to other tubes, backflow as a result of collecting blood into EDTA tubes before clot activator gel or lithium heparin tubes or due to syringe contamination when delivering blood into EDTA sample tubes before clot activator gel or lithium heparin tubes. Backflow is the regurgitation of blood from the evacuated blood collection tube back into the needle or vein. If blood is first collected from EDTA-containing tubes, the regurgitated blood may be contaminated with EDTA which is then transferred to the following sample tubes. If blood from a syringe is delivered into an EDTA sample tube, contamination of the syringe tip may result in droplet transfer of EDTA to subsequent sample tubes. ^2,3 Similar contamination may occur with Sarstedt glucose sample tubes (Sarstedt Safety Monovette glucose FNE/2.7, Aktiengesellschaft and Co, Nümbrecht, Germany), which contain sodium EDTA and potassium fluoride. It is also possible by using raised potassium as the sole discriminator for the application of EDTA measurement that EDTA contamination may be missed in samples from hypokalaemic patients.

Decanting of blood from kEDTA-containing tubes to other tubes results in grossly abnormal results indicating obvious contamination. More subtle changes, however, may occur with EDTA-contaminated blackflow and EDTA-contaminated syringes. These subtle changes may be difficult to detect; indeed in our study, 36% of EDTA-contaminated samples were initially undetected by laboratory staff and only identified by measurement of EDTA. EDTA sequesters divalent and trivalent metal ions. It is possible, therefore, that EDTA contamination could be identified by a decrease in concentrations of the commonly measured cations, namely zinc, magnesium and calcium as well as a decrease in the activity of alkaline phosphatase a zinc-containing metalloenzyme. The logarithmic of the stability constants for EDTA binding to calcium, zinc and magnesium are 10.61, 13.1 and 8.8, respectively. ^7,8 In detecting EDTA contamination, hypozincaemia would be expected to the most effective followed by hypocalcaemia and then hypomagnesaemia. Indeed, serum zinc was the best surrogate marker for detecting EDTA contamination. Only serum zinc values below the reference range had 100% diagnostic sensitivity for EDTA contamination, but even use at optimal diagnostic specificity (serum zinc < 7.8 μmol/L; 100% sensitivity; 89% specificity) would unacceptably result in potentially genuine hyperkalaemic samples being rejected. Therefore, kEDTA contamination can only be confidently detected by measurement of serum EDTA.

In addition to potassium, calcium, magnesium and alkaline phosphatase, kEDTA contamination may also affect the serum measurement of iron, unsaturated iron-binding capacity, bicarbonate, aspartate transferase, alanine transaminase, ammonia, copper, lactate dehydrogenase, creatine kinase and amylase. ^2,3

In our study, EDTA contamination was limited to samples from inpatients and general practice. It is interesting to note that EDTA contamination was not evident in samples from outpatients, which accounts for 20% of our workload and where blood samples are collected by trained phlebotomists. Continuing education, therefore, regarding correct order of draw of blood samples and avoidance of transferring blood between sample tubes will be reinforced (Pathology Handbook) and targeted to those involved in collection of blood specimens from patients in primary care (GP Newsletter) and hospital inpatients (Hospital Grand Round and Trust induction). The education programme will be then be re-audited to evaluate its efficacy.

In summary, spurious hyperkalaemia due to kEDTA contamination is common. Gross contamination is usually evident by unexpectedly high potassium concentrations with low serum calcium, magnesium and zinc concentrations. Spurious hyperkalaemia due to low concentrations of kEDTA contamination may be more difficult to identify and can only be confidently detected by measurement of serum EDTA. We have, therefore, introduced routine measurement of EDTA in all samples with a serum potassium ≥6.0 mmol/L.