Potassium ethylenediaminetetraacetic acid (kEDTA) sample cross-contamination: prevalence, consequences, identification, mechanisms, prevention and mitigation
Introduction
Ethylenediaminetetraacetic acid (EDTA) is used as a blood sample tube anticoagulant, most notably for the full blood count (FBC) as it also preserves cellular components (1). EDTA contains four carboxylic acid groups and two amine groups, each with a lone pair of electrons that can form coordination bonds with metal cations, making it a hexadentate ligand. It can, therefore, form six coordination bonds when complexing the metal cations. EDTA complexes with calcium ions in blood samples, preventing them from contributing to the coagulation cascade (1). Sample tubes for FBC most commonly use potassium salts of EDTA, either K2EDTA or K3EDTA, and potassium EDTA (kEDTA) may also be combined with sodium fluoride in tubes designed for the measurement of glucose (2). Each carboxylic acid group can independently ionise in aqueous solution (dissociate to release a proton, leaving behind a negatively charged -COO− group). K2EDTA and K3EDTA are formed when two or three of these carboxylic acid groups form salts with K+ ions. In vitro, EDTA contamination of lithium heparin, gel containing, plain tubes or sodium citrate containing tubes results in spurious electrolyte and other analyte results leading, if unrecognised, to clinical risk.
Methods of identifying publications for review
PubMed was used to search Medline with the following strategy “((clinical laboratory) AND (edta)) AND (contamination)” to identify relevant publications. Abstracts were screened for relevance by N.L. and S.M. The references for selected papers were also screened to find further relevant studies and publications.
Prevalence of EDTA contamination
Most laboratories identify kEDTA contaminated serum samples using surrogate markers, typically hyperkalaemia, hypocalcaemia, hypomagnesaemia and hypophosphatasia (2-7). A national survey on specimen contamination in the United Kingdom reported a median 76 EDTA contaminated samples per annum per laboratory [median 19 per quarter; standard deviation (SD): 57.6] (8). The authors reported a higher prevalence of EDTA contamination in laboratories receiving Sarstedt tubes compared to those receiving Becton-Dickinson (BD) tubes (8). In a multi-centre study, Cornes et al. also reported laboratories receiving Sarstedt tubes had a higher prevalence of EDTA contamination than those that receiving Greiner or BD tubes (5.7%, 1.4%, and 5.0%, respectively) (9). The reasons for this remain unexplained but it may reflect variations in phlebotomy practice rather than tube type.
Most studies where EDTA was measured reported a higher prevalence of EDTA contamination likely because low levels of clinically significant kEDTA contamination can only be confidently detected by measurement of serum EDTA (7). White et al. tested all 4,789 samples received for routine biochemistry analysis over 7 days for EDTA contamination, defined as an EDTA concentration ≥0.1 mmol/L, and found it present in 22 of them, giving a point prevalence in their laboratory of 0.46% of biochemistry samples (10). Hawkins et al. took a random selection of 1054 serum samples received for biochemistry over a 6-week period which also had an EDTA sample collected at the same draw (5). Detectable EDTA was present in 240 (23%) of 1,054 samples, where the lower limit of detection was 0.02 mmol/L. A total of 20 (2%) samples out of 1,054 serum samples had EDTA concentration deemed significant, which for this study was ≥0.10 mmol/L. As the study sample selection was not clear, it is difficult to draw conclusions on prevalence in their laboratory. In a single centre study using Sarstedt Monovette® tubes, Cornes et al. reported a kEDTA contamination rate of 24% (28 out of 117) in hyperkalaemic samples with potassium concentrations ≥6.0 mmol/L based on an EDTA concentration ≥0.1 mmol/L. Of the 27 patients re-bled, serum potassium was within the reference range in all, verifying previous kEDTA sample contamination (7). A five laboratory multicentre study over 1 month reported a prevalence of EDTA contamination (≥0.2 mmol/L) of 4.1% (range, 1.2% to 6.7%) in hyperkalaemic samples (serum potassium ≥6.0 mmol/L) (9). In a more recent study of hyperkalaemic samples (serum potassium ≥6.0 mmol/L), Kalaria and colleagues reported a median baseline prevalence of 5.6 cases [interquartile range (IQR) of 3.1 to 9.2] of EDTA contamination (≥0.15 mmol/L) per 10,000 samples received for urea and electrolyte (U&E) quantification per week over a 52-week period (11). Different studies have used EDTA concentrations varying from 0.10 to 0.20 mmol/L as the lower limit of “significant” kEDTA contamination. There may, therefore, be differences in reported prevalence of contamination, depending on the cut-off chosen. Like any analyte, however, there may also be variation in performance of the EDTA assay between platforms and from one centre to another, which may add variation to the reported prevalence of EDTA contamination.
It is not widely appreciated that EDTA (not necessarily as a potassium salt) is ubiquitous and is often present in water and internal quality control (IQC) material and may be detectable in blood (12). This may explain the high prevalence of very low levels of “EDTA contamination” in the absence of spurious electrolytes and the differences in EDTA cut-offs. For example, Hawkins showed that there were many more samples with an EDTA concentration above the lower limit of quantification in his study, which was ≥0.02 mmol/L (17.1% to 38.5%), compared with those which had significant contamination, defined here as EDTA concentration ≥0.10 mmol/L (0 to 7.2%) (5). The higher number of samples with EDTA ≥0.02 and <0.10 mmol/L is possibly due to the low-level environmental presence of EDTA.
Effects of EDTA contamination
Potassium
The hallmark of kEDTA contamination is spurious hyperkalaemia because of direct addition of potassium to the serum or plasma from the kEDTA salts (2-4,7,9,13-15). Severe hyperkalaemia is a cause of arrhythmias and cardiac arrest and therefore a medical emergency. If spurious hyperkalaemia is not recognised as such, it can lead to the patient requiring an unnecessarily urgent repeat test, or worse yet, the patient may be started on emergency treatments with the intention to lower serum potassium, and because they are inappropriate, may lead to harm. Additionally, kEDTA contamination may mask hypokalaemia, delaying treatment and/or investigation of what may also be a life-threatening electrolyte abnormality (6).
The extent of spurious hyperkalaemia caused by kEDTA is dependent on type of contaminant. Each K2EDTA and K3EDTA molecule has respectively two and three dissociable potassium ions and therefore each 1.0 mmol/L increase in EDTA would be expected to increase serum potassium by 2.0 and 3.0 mmol/L when K2EDTA and K3EDTA are completely dissociated. Kalaria et al. proposed using this positive stoichiometric relationship of increase in serum potassium with EDTA concentration for independent verification of an EDTA assay. This is of importance because external quality assurance schemes or inter-laboratory comparison are not available for EDTA assay in view of a small number of laboratories employing the assay (15).
Divalent cations
EDTA, as a chelator, binds and complexes divalent cations. Thus, contamination of serum or plasma samples meant for the quantification of these may result in falsely low results, depending on the method of quantification (15,16). Routine automated biochemistry analysers use colorimetric methods for the quantification of divalent ions in blood, especially the common ones such as calcium, magnesium, iron and in some centres zinc. Colorimetric methods are generally affected by EDTA contamination. Methods, however, using dissociation of polyatomic ions into constituent ions such that the divalent cation is no longer bound by the EDTA are not affected. These include inductively coupled plasma atomic emission spectroscopy (ICP-AES), inductively coupled plasma-mass spectrometry (ICP-MS), and atomic absorption spectrophotometry (AAS) (2,15,16). Paradoxically, ion chromatography-ICP-MS (IC-ICP-MS) may be affected because chelated and free ions may have been separated prior to dissociation by ICP, dependant on protocol (15,16).
Spurious hypocalcaemia is well documented as a known effect of EDTA contamination in colorimetric methods (2,4,6,7,15-17). These include the commonly used methods on automated analysers that use arsenazo III and O-cresolphthalein complexone (OCPC). EDTA can complex calcium ions with greater affinity than these two chromophores, reducing the amount of calcium available for chromophore binding in contaminated samples. The linear negative stoichiometric relationship between EDTA in serum and apparent (colorimetrically measured) calcium concentration was proposed for identification of gross EDTA contamination by Davidson (2). Davidson later used the regression of calcium results by AAS and OCPC methods in the EDTA contaminated serum samples for assessment of accuracy of the automated EDTA assay (15,16).
The measurement of ionised calcium using ion selective electrodes is subject to negative interference with EDTA. This was actually used to accurately determine the amount of EDTA added to calcium containing solutions in the past (18).
Apparent magnesium concentrations determined using colorimetric methods, such as formazan dye, xylidyl blue or calmagite may be spuriously lower in an EDTA contaminated sample. Indirect enzymatic methods of magnesium quantification, such as Abbott’s method of isocitrate dehydrogenase to convert D-isocitric acid and nicotinamide adenine dinucleotide phosphate (NADP) to 2-oxologlutarate, CO2 and reduced NADP (NADPH), a process that relies on magnesium ions as a cofactor, are also expected to give spuriously low results with EDTA contamination (Abbott Laboratories, Abbott Park, IL, USA). The effect may be method dependent. Kalaria et al. demonstrated the decrease in magnesium with EDTA contamination did not exceed its reference change value even at an EDTA concentration of 5.60 mmol/L in an Abbott Architect® automated method which at the time used arsenazo dye with a calcium chelator (15). However, Davidson found significant decrease in serum magnesium in EDTA contaminated samples in the xylidyl blue method on the Roche platform (2) as did Lima-Oliveira et al. but the method used was not stated (17).
Quantification of iron by Abbott’s colorimetric ferene-S method can be negatively affected by EDTA contamination (15). However, Davidson reported an interesting situation where Roche’s ferrozine method for iron quantification was not affected while its unbound iron binding capacity (UIBC) method, which also utilised ferrozine, showed positive interference with EDTA (2). He hypothesised this was due to the difference in pH of the reactions- the iron method was conducted at pH 2, where the EDTA’s affinity for ferrous ions was 101 L/mol, while the UIBC reaction was carried out at pH 9.1, where EDTA’s affinity for ferrous ions was 1013 L/mol. Walmsley et al. in 1992 tested three different commercial platforms’ performance of measuring iron using ferrozine in plasma from volunteers collected into EDTA FBC tubes (19). The Technicon “SMAC” analyser, Boehringer kit on a Hitatchi 717 and a Roche Cobas method and platform were tested. The first two showed significant interference by the EDTA, with iron being completely unmeasurable in these samples. The Roche method, however, gave results for iron in the EDTA samples which were on average 6.2 µmol/L lower than the paired serum samples collected from the same volunteers. The authors hypothesised that this was because of the higher ferrozine concentration in the Roche method reaction mixture than the others. They were able to almost abolish the EDTA interference in the Roche method by adding excess zinc sulphate, because the EDTA preferentially binds zinc ions instead of the ferrous ions. The addition of zinc did not mitigate the interference by EDTA for the Boehringer method, though adding excess ferrozine did. The authors did have an additional issue, however, which was that the EDTA FBC tubes were contaminated with iron (19).
Zinc quantification by a colorimetric method using 2-(5-bromo-2-pyridylazo)-5-(n-n-propyl-n-3-sulfopropylamino) phenol, also known as 5-Br-PAPS, used in a kit from Wako chemicals (Fujifilm Wako Chemicals GmbH, Neuss, Germany) is affected by EDTA contamination of serum as reported by Cornes et al. and Sharratt et al. (6,7). In fact, because of EDTA’s higher affinity for zinc compared to many other divalent cations, hypozincaemia by colorimetric assay has been reported as the most sensitive surrogate marker of EDTA contamination when investigating hyperkalaemic samples for EDTA contamination (6,7,20). However, most laboratories now use ICP-MS instead of colorimetric methods for zinc determination. Though data on this is not available, zinc determination by ICP-MS should not be affected by EDTA contamination as explained earlier.
Enzyme activity
Alkaline phosphatase (ALP) is a metalloenzyme with a zinc ion as part of its structure and its action also depends on magnesium ion as a cofactor. Exposure of ALP to EDTA results in a continued decrease in ALP activity with time, due to continued irreversible stripping of zinc ions from the enzyme structure (2,21). ALP inhibition is competitive at lower EDTA concentration and in presence of excess substrate, however, the inhibition becomes irreversible with time and with EDTA concentration >1.0 mmol/L (22,23).
The International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) ALP method used by most manufacturers depends on the patient’s ALP converting p-nitrophenyl phosphate and water to p-nitrophenol and phosphate and measuring the increase of absorbance of the p-nitrophenol. This method has excess zinc and magnesium ions added to the reaction mixture (24). ALP activity decreases in EDTA contaminated serum samples (2,7,15). However, significant hyperkalaemia and hypocalcaemia occurs at much lower EDTA concentration compared to the much higher EDTA concentration required for significant decrease in ALP activity by the IFCC method and therefore ALP is of limited value as a surrogate marker of EDTA contamination (2,7,15). In a study by Kalaria et al. analysing samples within 8 hours serum ALP did not change with EDTA concentrations of 0.89 mmol/L or lower on Abbott Architect p-nitrophenyl phosphate with chelated metal iron buffer method. However, ALP in pooled serum decreased by 7.1% and 54.1% at EDTA concentrations of 1.86 and 3.66 mmol/L, respectively (15). ALP inhibition could also be dependent on the length of time in presence of EDTA. ALP activity decreased by 14.8% at 20 minutes and 70.5% at 72 hours in the presence of 5.0 mmol/L EDTA in a study by Davidson (2). Tissue of origin of ALP may also determine susceptibility to EDTA inhibition (21).
Amylase, a metalloprotein, has a calcium ion as part of its structure and, like ALP, amylase activity is also inhibited by EDTA in a time-dependent manner (2). Amylase activity did not change at 20 minutes but decreased by 34.3% after 72 hours in the presence of 5.0 mmol/L EDTA in a study by Davidson (2).
The effect of EDTA on lactate dehydrogenase (LDH) appears complex and is likely method dependent. Davidson in 2002 showed that EDTA slowed down the time-dependent deterioration in LDH, aspartate aminotransferase (AST), alanine aminotransferase (ALT), and creatinine kinase (CK) activity in stored samples (2). The LDH measurement in this study was done using a Roche method which at the time relied on the conversion of pyruvate to L-lactate by LDH in the sample. It was hypothesised that heavy metals in serum can lead to oxidation of the LDH, AST, ALT and CK enzymes with time and EDTA chelation of metals slowed this down. This has been shown experimentally before (25). Lima-Oliveira et al., however, showed an immediate negative interference of LDH with EDTA using a Roche method in 2014 (17). At some point Roche changed their method for LDH to measuring the conversion of L-lactate to pyruvate (Roche Diagnostics, Indianapolis, IN, USA). Kalaria et al. did not detect a significant difference due to EDTA contamination in LDH activity using Abbott’s method utilising the conversion of L-lactate to pyruvate (15).
Bicarbonate, measured on a Roche method that utilised phosphoenolpyruvate carboxylase (PEPC) to convert bicarbonate and phosphoenolpyruvate (PEP) to oxaloacetate, decreased in presence of EDTA (2). Magnesium ion is a cofactor for PEPC and the method used only had 0.7 mmol/L magnesium in the reagent. Methods with higher magnesium concentration in the reagent may be less susceptible.
Lima-Oliveira et al. found a negative interference with the quantification of phosphate using an undescribed Roche method (17). Roche platform currently uses the molybdate method, in which ammonium molybdate reacts with phosphate to form ammonium phosphomolybdate, whose absorbance can be measured at 340 nm (Roche Diagnostics). EDTA can interfere with the formation of phosphomolybdate and thus EDTA plasma has not been recommended for determination of phosphate (26). However, the current Roche phosphate assay kit actually states that EDTA plasma can be used (Roche Diagnostics). Kalaria et al. did not find interference with phosphate measurement using Abbott molybdate method in EDTA contaminated samples (15).
Dissociation-enhanced lanthanide fluorescent immunoassay (DELFIA)TM
DELFIATM technology uses a lanthanide label to generate a fluorescent signal during quantification of an analyte. The lanthanide is usually attached to an antibody. Once binding is complete an enhancing solution is added that leads to release of the lanthanide and its chelation by another molecule. This new chelated lanthanide is fluorescent, and the intensity of the fluorescence is measured for quantitation of the analyte. Holtkamp et al. published in 2008 showing that 27 out of 138 EDTA contaminated newborn screening (NBS) bloodspots had falsely elevated 17-hydroxyprogesterone using a DELFIATM method from PerkinElmer (27). Their laboratory also measured thyroid stimulating hormone (TSH) using a PerkinElmer DELFIATM method and experimental contamination of bloodspots to higher concentrations than those seen in real life led to falsely low TSH results. EDTA in the sample can chelate the lanthanides and interfere with analyte quantitation (28).
Coagulation
Calcium ions are required for the coagulation cascade. EDTA’s anticoagulant activity is due to chelation of calcium ions. Coagulation studies in the blood sciences laboratory are usually performed on plasma collected into citrate containing tubes. The citrate chelates calcium ions as well to keep the blood anticoagulated until it is ready to be tested. After separation of the plasma, a fixed amount of calcium ions is added to allow activation of the coagulation cascade. This process relies on a fixed amount of citrate in the solution. EDTA contamination, by adding more calcium ion chelation capacity, slows the coagulation cascade, giving spuriously prolonged results for both prothrombin time and activated partial thromboplastin time (29,30). It can also lead to erroneous positive results for factor V and factor VIII inhibitor testing (30). In an Australian external quality assurance programme samples from kEDTA plasma were mis-identified by 68% of 42 laboratories as representing patients with factor V and/or factor VIII inhibitor and only one laboratory correctly identified kEDTA as the source of abnormal results (31).
Sources of contamination (Table 1)
Table 1
Analyte | Mechanism(s) of interference by EDTA | Bias | Reference |
---|---|---|---|
Potassium | Additional potassium released from K2EDTA or K3EDTA is measured | Positive | (2-4,6,7,9,11,17) |
Divalent cations (calcium, magnesium, iron, zinc) | EDTA chelates the free ions, preventing them from taking part in reactions used for measurement | Negative | (2,4,6,7,15-19) |
ALP activity | EDTA chelates zinc ions from the ALP structure in a time-dependent method | Negative | (2,21-23) |
Amylase activity | EDTA chelates the calcium ions from the amylase structure in a time-dependent method | Negative | (2) |
LDH activity | EDTA may slow down the time-dependent deterioration in LDH activity due to chelating heavy metals in the serum/plasma that cause oxidation of the enzyme. EDTA may also cause an immediate reduction in LDH enzyme activity via an unknown mechanism | None or negative | (2,17) |
Bicarbonate | Methods that utilise phosphoenol pyruvate carboxylase have the magnesium ions chelated from this enzyme | Negative | (2) |
Phosphate | Methods that utilise the formation of ammonium phosphomolybdate suffer interference as EDTA prevents this formation, but it does not happen universally | Negative | (15,17) |
17-hydroxyprogesterone and TSH in blood spots by DELFIATM methods | EDTA chelates the lanthanides used and interferes with their fluorescence | Positive | (27) |
Prothrombin time and activated partial thromboplastin time | EDTA chelates calcium and prevents its participation in the coagulation cascade | Positive | (29-31) |
Factor V and factor VIII inhibitors | EDTA chelates calcium and prevents its participation in the coagulation cascade | Positive | (30,31) |
kEDTA, potassium ethylenediaminetetraacetic acid; EDTA, ethylenediaminetetraacetic acid; ALP, alkaline phosphatase; LDH, lactate dehydrogenase; TSH, thyroid stimulating hormone; DELFIATM, dissociation-enhanced lanthanide fluorescent immunoassay.
EDTA contaminates blood samples when it is introduced from one tube in which it is an acceptable anticoagulant into another tube in which it is not intended to be. This typically occurs at the phlebotomy. The process by which this occurs has been a focus of study in the last two decades. It has been repeatedly shown that “closed” phlebotomy techniques, in which the blood tubes are never uncapped but are instead attached to an appropriate sampling needle or device during venepuncture, do not result in contamination with EDTA, even if order of draw is reversed, that is the EDTA sample is drawn before other sample types (32-36). Studies using Monovette® (34) and Vacuette® (32) systems illustrated that wrong “order of draw” did not result in cross-contamination. Although not studied, poor phlebotomy practice by collecting blood with sample tubes held upside down may cause needle tip kEDTA contamination and lead to spurious results. This may explain why guidelines still advise that blood is drawn into plain tubes before additive containing sample tubes (37,38).
Asif et al. attempted to determine the cause of EDTA contamination in an observational study in 2019 (39). The laboratory screened all samples with potassium ≥6.0 mmol/L by reflex direct measurement of EDTA and identified 96 EDTA contaminated serum samples over a 4-month period. The authors were able to identify and speak with 64 of the healthcare professionals who collected these samples and in 52 cases they recalled taking the sample in question. All 52 recalled opening the blood tubes to put blood into them, rather than following the recommended “closed” sampling technique. This strongly supports the notion that EDTA contamination does not occur during “closed” phlebotomy and almost exclusively occurs using open phlebotomy techniques. This emphasises the importance of order of tube fill during open phlebotomy. In vitro experiments, furthermore, have demonstrated that kEDTA contamination may occur either due to syringe needle/top contamination when delivering blood into EDTA sample tubes before clot activator gel or lithium heparin tubes or the direct transfer of blood from kEDTA containing tubes to other tubes (39).
Detection of EDTA contamination in the medical laboratory
Indirect methods
A UK survey of medical laboratories in 2014 found that 25 out of 26 laboratories who responded to a question about how they detect EDTA contamination in biochemistry samples said they did so by indirect methods. This involved looking for a pattern of raised potassium, low calcium, low magnesium and low ALP activity (8). The remaining single laboratory tested for EDTA concentration directly when potassium concentrations were raised. Observational studies have shown that kEDTA contamination significant enough to cause spurious results is often missed when indirect markers are relied upon. Cornes et al. showed that up to 50% of significant kEDTA contamination may be missed if indirect methods are relied upon for identification when investigating hyperkalemic samples (6,7,9). The screening ability of calcium, zinc, magnesium and ALP for detection of kEDTA contamination in hyperkalaemic serum samples was assessed. Using the lower reference limit of each, calcium, zinc, magnesium and ALP had sensitivities of 80%, 100%, 80%, and 29% and specificities of 92%, 63%, 95%, and 99% respectively in samples with potassium ≥6.0 mmol/L (7). This study showed that of these markers, only serum zinc using colorimetric methods has the potential for detecting low levels of kEDTA contamination causing clinically significant spurious hyperkalaemia, but these colorimetric zinc assays have largely been superseded by ICP-MS methods. Therefore, none of the routinely available indirect markers for detection of EDTA contamination can reliably identify clinically significant spurious hyperkalaemia due low levels of kEDTA contamination. This is likely true in coagulation studies as well for low level contamination with kEDTA, though we found no reported studies for coagulation testing at different levels of kEDTA contamination. However, indirect methods such as potassium and calcium will identify gross kEDTA contamination where kEDTA plasma is erroneously submitted for analysis of biochemistry testing and for coagulation and clotting factors analysis (30,40).
Direct measurement of EDTA
Davidson in 2007 published the validation of a method for direct measurement of EDTA in serum on an automated biochemistry analyser to test for contamination (16). This colorimeteric assay on a Roche Cobas platform relied on EDTA’s ability to abstract copper ions from pyridylazonaphthol-copper (PAN-Cu), turning it from violet to yellow. The decrease in absorbance of the PAN-Cu was monitored at 546 and 660 nm. This method was also used by many other authors who have studied EDTA contamination (6,7,9,17). White and Hawkins ran the same method on a Beckman automated platform (5,10). Chadwick et al. validated the same method on Abbott Architect® platform (20) and it has since been used in other studies (11,15). Most studies used EDTA cut-offs of 0.10 to 0.20 mmol/L for significant contamination, though Hawkins also reported all detected EDTA (>0.02 mmol/L) in his report (5).
Holtkamp et al. used a tandem mass spectrometry (MS/MS) method on NBS blood spots to detect significant EDTA contamination (27). Derivatisation by butylation created a product that could be detected on MS/MS with mass to charge transition of 517.1 to 188.1. They found that EDTA concentrations of 2.6 g/L were sufficient to cause a blood spot of 17-hydroxyprogesterone concentration equating to 2.5 nmol/L to be as spuriously high as 168 nmol/L.
Scholes et al. in 2015 tested the QUANTOFIX EDTA qualitative test strips (41). These test strips, originally designed to detect EDTA contamination of drinking water, could detect very high concentrations of EDTA, consistent with plasma taken from an EDTA containing tube, but were not sensitive enough to detect EDTA at the lower concentrations seen with contamination of biochemistry serum samples that can still significantly impact results. These strips detect EDTA contamination by colour change of xylenol orange complex when bismuth from the complex is chelated by EDTA in the sample. The user visually determines the end point colour and matches it to a semi-quantitative scale. The authors concluded the strips were not suitable for routine use in a biochemistry laboratory due to the poor sensitivity. They also found that if the strips were dipped into serum samples, rather than the serum being pipetted on to the strips, it would result in contamination of the serum sample with bismuth and citrate and could lead to other analytical issues (41).
The advantage of direct measurement is clear when compared to the indirect methods. EDTA measurement may be automatically reflexed in real-time in samples suspected to have EDTA contamination, for example those with hyperkalaemia, hypocalcaemia, hypomagnesaemia or hypophosphatasia. This mitigates the risk of releasing spurious results from contaminated samples as well as risk and anxiety of not identifying the contaminated samples or unnecessary blocking of genuine high potassium or low calcium results from non-contaminated samples. If these analytes are not requested, it is possible that EDTA contamination may be missed, for example in the case of a spuriously low amylase. The EDTA test, however, may be added on by a laboratory medical specialist if EDTA contamination affecting less commonly requested analytes are suspected.
Despite the benefits of direct EDTA measurement, most laboratories use indirect methods of detection (8) perhaps because they believe it is not a problem or that there are unable to do anything about it. Underlying this may include lack of commercially manufactured assay kits for measurement for EDTA, unfamiliarity with the EDTA assay amongst laboratorians, unawareness of increased sensitivity for the detection of contamination, or concerns about increased associated costs and effort. In cases where laboratories do not have access to direct EDTA measurement and have to rely on indirect markers, they should be aware of the risk of missing low levels of kEDTA contamination causing clinically significant spurious hyperkalaemia.
Feedback to health care professionals
Davidson et al. in 2007 wrote that despite efforts to continually educate staff, kEDTA contamination rates were rising, coincidental with the switch to using the Sarstedt Monovette® tubes in his centre (16). Mitigation of kEDTA contamination is aimed at improving phlebotomy with emphasis on use of closed venepuncture systems and “order of tube fill” with open phlebotomy. Indeed, trained phlebotomists have much lower rates of kEDTA contamination compared to other of healthcare professionals (5,35,36).
Kalaria et al. studied the effect of automated explicit laboratory feedback on the frequency of EDTA contamination (11). The laboratory routinely screened all samples with potassium concentration ≥6.0 mmol/L. The feedback used in this study is shown in Table 2.
Table 2
EDTA concentration (mmol/L) | Feedback issued on laboratory report |
---|---|
<0.15 | • No feedback |
• Results reported as measured | |
≥0.15 to <0.20 | • “Possible potassium EDTA contamination, please send repeat specimen. EDTA contamination occurs with open phlebotomy. Recommend closed phlebotomy (vacutainer) or with open phlebotomy fill gold top tube before purple and grey top tubes” |
• Results reported as measured | |
≥0.20 | • “Potassium EDTA contamination, suggest urgent repeat. EDTA contamination occurs with open phlebotomy. Recommend closed phlebotomy (vacutainer) or with open phlebotomy fill gold top tube before purple and grey top tubes” |
• Results from analytes potentially impacted by kEDTA removed from report |
EDTA, ethylenediaminetetraacetic acid; kEDTA, potassium ethylenediaminetetraacetic acid.
The prevalence of EDTA contamination was measured for 52 weeks before and 43 weeks after introduction of the comments, with a 4-week washout period in between. The introduction of these comments reduced the prevalence of EDTA contaminated serum samples from a median (IQR) of 5.6 (3.1–9.2) per 10,000 samples for U&Es per week to 2.3 (1.1–4.4) per 10,000 samples per week (P<0.001); a 56% reduction in prevalence.
Reporting of analytes in EDTA contaminated samples
Most laboratories do not report any results from EDTA contaminated samples when identified (8) but the prevalence of EDTA contamination makes this overly restrictive. Kalaria et al. argued for selective reporting of analytes from EDTA contaminated samples in a relatively recent study. They proposed only those analytes whose measurement is affected by EDTA contamination be removed from the report instead of not reporting all results (15). EDTA tubes are spray dried and therefore have no dilution effect on samples. The spurious results are specific for those analytes where EDTA interacts with the analyte or measurement method. By selective reporting of non-affected analytes, the laboratory may spare repeat phlebotomy for some patients because the clinician may get all the information required from the reported results, for example, creatinine and estimated glomerular filtration rate (eGFR) from U&Es profile may have been what the clinician wanted. However, caution must be taken if this approach is adopted because EDTA affects some analytes based on measurement methods and therefore, as discussed before, data from literature may not be applicable to all measurement methods for an analyte or when methods get updated. If a laboratory is to adopt selective reporting of analytes from EDTA contaminated samples, verification will be required for all the analytes they intend to report in the presence of significant EDTA contamination and reverification if a method is changed with time or by change of analytical platform.
Current practice in blood sciences, The Royal Wolverhampton NHS Trust
In our biochemistry laboratory, measurement of EDTA using the automated PAN-Cu method is reflexed automatically on all samples which have a measured potassium of ≥6.0 mmol/L and if normokalaemic an adjusted calcium of ≤1.80 mmol/L. EDTA measurement can also be added to samples at the discretion of the medical or scientific biochemist if other analyte abnormalities consistent with kEDTA contamination, such as a low ALP activity, low magnesium or low iron concentration, are seen. We follow a simplified reporting protocol adapted from Kalaria et al. (11). For samples with EDTA <0.20 mmol/L, we report all results as measured. For samples with EDTA ≥0.20 mmol/L, we do not report potentially affected analytes (Table 1) with an educational feedback comment stating there is significant EDTA contamination (Table 2).
Conclusions
Spurious results due to kEDTA contamination are common and of these, spurious hyperkalaemia is the most concerning and may adversely affect patient care if unidentified. Gross contamination is easy to identify with hyperkalaemia, hypocalcaemia, hypomagnesaemia and low ALP activity. Low levels of kEDTA contamination, however, causing clinically significant spurious hyperkalaemia may only be reliably identified by measurement of EDTA since indirect markers are neither sensitive nor specific. Automated reflex measurement of EDTA in potentially contaminated samples is therefore the definitive method of identifying EDTA contamination and is suitable for routine laboratories.
Contrary to published phlebotomy guidelines, reverse order of draw using closed phlebotomy systems does not cause kEDTA sample cross-contamination. KEDTA contamination exclusively occurs with syringe-tip or needle-tip contamination in open phlebotomy or occasionally by decanting blood from EDTA containing tubes to other tubes. Prevention is aimed at improving phlebotomy by encouraging use of closed phlebotomy systems and promoting correct “order of sample tube fill” with open phlebotomy methods. Selected reporting of analytes in EDTA contaminated samples unaffected by EDTA may mitigate the effects of EDTA contamination (Table 3).
Table 3
Section | Summary |
---|---|
Prevalence of significant contamination | 0.056% to 0.46% of chemistry samples (10,11) |
1.2–24% of hyperkalemic samples (7,9) | |
Effects of significant contamination (method dependent) | Spuriously raised: potassium, UIBC, TSH by DELFIATM, 17-hydroxyprogesterone by DELFIATM, PT, APTT (2-4,6,7,9,13-15,27,29,30) |
Spuriously lowered: calcium, magnesium, zinc, iron, ALP, LDH, amylase, bicarbonate, phosphate (2,4,6,7,15-17,19,20,22,23,27) | |
Source of significant contamination | “Open” phlebotomy techniques (39) |
Detection of significant contamination | Indirect marker: checking other analytes that are also affected by EDTA contamination—more widely available but not sensitive to lower-level significant contamination (6-9) |
Direct measurement: EDTA measured directly using PAN-Cu or MS/MS—less widely available but more sensitive and able to detect all levels of significant contamination (5-7,9,10,15,17,20,27) | |
Feedback when significant contamination detected | Adding explicit educational comments on reports of samples with significant contamination may reduce rates of future contamination (11) |
Selective reporting of results in presence of significant contamination | Some analytes in a sample significantly contaminated by kEDTA may still be safely reported if the laboratory has validated that EDTA does not interfere with measurement of those analytes (15) |
UIBC, unbound iron binding capacity; TSH, thyroid stimulating hormone; DELFIATM, dissociation-enhanced lanthanide fluorescent immunoassay; PT, prothrombin time; APTT, activated partial thromboplastin time; ALP, alkaline phosphatase; LDH, lactate dehydrogenase; PAN-Cu, pyridylazonaphthol-copper; MS/MS, tandem mass spectrometry; EDTA, ethylenediaminetetraacetic acid; kEDTA, potassium ethylenediaminetetraacetic acid.
Acknowledgments
Funding: None.
Footnote
Peer Review File: Available at https://jlpm.amegroups.org/article/view/10.21037/jlpm-24-42/prf
Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jlpm.amegroups.org/article/view/10.21037/jlpm-24-42/coif). R.G. serves as an unpaid Associate Editor-in-Chief of Journal of Laboratory and Precision Medicine from January 2024 to December 2025. The other authors have no conflicts of interest to declare.
Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.
References
- Banfi G, Salvagno GL, Lippi G. The role of ethylenediamine tetraacetic acid (EDTA) as in vitro anticoagulant for diagnostic purposes. Clin Chem Lab Med 2007;45:565-76. [Crossref] [PubMed]
- Davidson DF. Effects of contamination of blood specimens with liquid potassium-EDTA anticoagulant. Ann Clin Biochem 2002;39:273-80. [Crossref] [PubMed]
- Fitzpatrick MF, Newell J, Grimes H, et al. Spurious increase in plasma potassium concentration and reduction in plasma calcium due to in vitro contamination with liquid potassium edetic acid at phlebotomy. J Clin Pathol 1987;40:588. [Crossref] [PubMed]
- Stevens RJ, Laina V, Wong CK, et al. A case of hyperkalaemia associated with hypocalcaemia. Emerg Med J 2008;25:379-80. [Crossref] [PubMed]
- Hawkins RC. EDTA contamination of serum samples: common but not necessarily significant. Ann Clin Biochem 2011;48:478. [Crossref] [PubMed]
- Sharratt CL, Gilbert CJ, Cornes MC, et al. EDTA sample contamination is common and often undetected, putting patients at unnecessary risk of harm. Int J Clin Pract 2009;63:1259-62. [Crossref] [PubMed]
- Cornes MP, Ford C, Gama R. Spurious hyperkalaemia due to EDTA contamination: common and not always easy to identify. Ann Clin Biochem 2008;45:601-3. [Crossref] [PubMed]
- Logie JJ, Chaloner C. A national survey of specimen contamination in the UK. Ann Clin Biochem 2019;56:219-27. [Crossref] [PubMed]
- Cornes MP, Davidson F, Darwin L, et al. Multi-centre observational study of spurious hyperkalaemia due to EDTA contamination. Clin Lab 2010;56:597-9. [PubMed]
- White G. Serum ethylenediaminetetraacetic acid concentrations in routine samples submitted for biochemical analysis. Ann Clin Biochem 2010;47:485-6. [Crossref] [PubMed]
- Kalaria T, Ford C, Gama R. Effect of automated feedback laboratory comments promoting good phlebotomy practice on the frequency of ethylenediaminetetraacetic acid (EDTA) contamination of serum samples. Ann Clin Biochem 2023;60:63-7. [Crossref] [PubMed]
- Sheppard RL, Henion J. Quantitative capillary electrophoresis/ion spray tandem mass spectrometry determination of EDTA in human plasma and urine. Anal Chem 1997;69:2901-7. [Crossref] [PubMed]
- Imafuku Y, Meguro S, Kanno K, et al. The effect of EDTA contaminated in sera on laboratory data. Clin Chim Acta 2002;325:105-11. [Crossref] [PubMed]
- Naguib MT, Evans N. Combined false hyperkalemia and hypocalcemia due to specimen contamination during routine phlebotomy. South Med J 2002;95:1218-20. [Crossref] [PubMed]
- Kalaria T, Ford C, Gama R. Managing ethylenediaminetetraacetic acid (EDTA) interference in EDTA contaminated samples - selectivity in reporting analytes. Ann Clin Biochem 2023;60:92-9. [Crossref] [PubMed]
- Davidson DF. EDTA analysis on the Roche MODULAR analyser. Ann Clin Biochem 2007;44:294-6. [Crossref] [PubMed]
- Lima-Oliveira G, Salvagno GL, Danese E, et al. Contamination of lithium heparin blood by K2-ethylenediaminetetraacetic acid (EDTA): an experimental evaluation. Biochem Med (Zagreb) 2014;24:359-67. [Crossref] [PubMed]
- Tackett SL. Automatic titration of calcium with EDTA using a calcium-selective electrode Anal Chem 1969;41:1703-5. [J]. [Crossref]
- Walmsley TA, George PM, Fowler RT. Colorimetric measurement of iron in plasma samples anticoagulated with EDTA. J Clin Pathol 1992;45:151-4. [Crossref] [PubMed]
- Chadwick K, Whitehead SJ, Ford C, et al. kEDTA Sample Contamination: A Reappraisal. J Appl Lab Med 2019;3:925-35. [Crossref] [PubMed]
- Conyers RA, Birkett DJ, Neale FC, et al. The action of EDTA on human alkaline phosphatases. Biochim Biophys Acta 1967;139:363-71. [Crossref] [PubMed]
- Marcos J, Townshend A. Studies on the inhibition of immobilised alkaline phosphatase by metal ions and EDTA in a flow-injection system. Anal Chim Acta 1994;299:129-36. [Crossref]
- Whisnant AR, Gilman SD. Studies of reversible inhibition, irreversible inhibition, and activation of alkaline phosphatase by capillary electrophoresis. Anal Biochem 2002;307:226-34. [Crossref] [PubMed]
- Schumann G, Klauke R, Canalias F, et al. IFCC primary reference procedures for the measurement of catalytic activity concentrations of enzymes at 37 C. Part 9: Reference procedure for the measurement of catalytic concentration of alkaline phosphatase: International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) Scientific Division, Committee on Reference Systems of Enzymes (C-RSE) 1. Clin Chem Lab Med 2011;49:1439-46. [Crossref] [PubMed]
- Breccia JD, Andersson MM, Hatti-Kaul R. The role of poly(ethyleneimine) in stabilization against metal-catalyzed oxidation of proteins: a case study with lactate dehydrogenase. Biochim Biophys Acta 2002;1570:165-73. [Crossref] [PubMed]
- Endres DB, Rude RK. Chapter 38: Disorders of Bone. In: Burtis CA, Ashwood ER, Bruns DE, et al. editors. Tietz Fundamentals of Clinical Chemistry. 6th ed. Philadelphia: Saunders; 2008.
- Holtkamp U, Klein J, Sander J, et al. EDTA in dried blood spots leads to false results in neonatal endocrinologic screening. Clin Chem 2008;54:602-5. [Crossref] [PubMed]
- Bishop SC, Winefield R, Anbanandam A, et al. Aqueous synthesis of a small-molecule lanthanide chelator amenable to copper-free click chemistry. PLoS One 2019;14:e0209726. [Crossref] [PubMed]
- Banković Radovanović P, Živković Mikulčić T. Simović Medica J. Unexpected abnormal coagulation test results in a 2-year-old child: A case report. Biochem Med (Zagreb) 2020;30:011002. [Crossref] [PubMed]
- Adcock DM, Favaloro EJ. Pearls and pitfalls in factor inhibitor assays. Int J Lab Hematol 2015;37:52-60. [Crossref] [PubMed]
- Favaloro EJ, Bonar R, Duncan E, et al. Identification of factor inhibitors by diagnostic haemostasis laboratories: a large multi-centre evaluation. Thromb Haemost 2006;96:73-8. [Crossref] [PubMed]
- Ercan Ş, Ramadan B, Gerenli O. Order of draw of blood samples affect potassium results without K-EDTA contamination during routine workflow. Biochem Med (Zagreb) 2021;31:020704. [Crossref] [PubMed]
- Cornes MR, Sulaiman RA, Whitehead SJ, et al. Incorrect order of draw of blood samples does not cause potassium EDTA sample contamination. Br J Biomed Sci 2012;69:136-8. [Crossref] [PubMed]
- Sulaiman RA, Cornes MP, Whitehead SJ, et al. Effect of order of draw of blood samples during phlebotomy on routine biochemistry results. J Clin Pathol 2011;64:1019-20. [Crossref] [PubMed]
- Cadamuro J, Felder TK, Oberkofler H, et al. Relevance of EDTA carryover during blood collection. Clin Chem Lab Med 2015;53:1271-8. [Crossref] [PubMed]
- Bazzano G, Galazzi A, Giusti GD, et al. The Order of Draw during Blood Collection: A Systematic Literature Review. Int J Environ Res Public Health 2021;18:1568. [Crossref] [PubMed]
- Clinical & Laboratory Standards Institute. GP41. Collection of Diagnostic Venous Blood Specimens. 7th ed. 2017.
- Ialongo C, Bernardini S. Phlebotomy, a bridge between laboratory and patient. Biochem Med (Zagreb) 2016;26:17-33. [Crossref] [PubMed]
- Asif U, Whitehead SJ, Ford C, et al. Preanalytical potassium EDTA sample contamination: open versus closed phlebotomy systems. Ann Clin Biochem 2019;56:711-4. [Crossref] [PubMed]
- Lippi G, Salvagno GL, Adcock DM, et al. Right or wrong sample received for coagulation testing? Tentative algorithms for detection of an incorrect type of sample. Int J Lab Hematol 2010;32:132-8. [Crossref] [PubMed]
- Scholes KL, La'ulu SL, Ence AT, et al. Evaluation of test strips for the rapid identification of ethylenediaminetetraacetic acid (EDTA) specimens. Lab Med 2015;46:97-108. [Crossref] [PubMed]
Cite this article as: Lorde N, Mahapatra S, Kalaria T, Gama R. Potassium ethylenediaminetetraacetic acid (kEDTA) sample cross-contamination: prevalence, consequences, identification, mechanisms, prevention and mitigation. J Lab Precis Med 2024;9:34.