Hyperkalaemia is defined as potassium levels greater than 5.5 mmol/L.¹ It can occur following dysfunction of the renin-angiotensin-aldosterone system due to kidney failure, adrenal hypofunction, or the use of certain drugs, thus resulting in decreased potassium elimination.^2,3 Other common causes of hyperkalaemia include increased release of potassium from cells, as seen in diabetic ketoacidosis due to decreased insulin levels, as well as cellular breakdown or necrosis, such as rhabdomyolysis, tumour lysis syndrome, haemolysis, trauma, and burns.^4,5 Hyperkalaemia is also not unusual in patients with thrombocytosis or leucocytosis.^5,6 Persistent hyperkalaemia can be life-threatening, causing cardiac arrhythmias with possible cardiac arrest or paralysis of the respiratory muscles even with slight deviations from normal levels (3.5 mmol/L – 5.0 mmol/L).^1,6 Thus, accurate laboratory potassium measurements are necessary for clinicians to make appropriate medical decisions.

Within the laboratory, the quality and confidence of a test result depends on the quality and functioning of the total testing process, which encompasses the pre-analytical, analytical and post-analytical stages.^7,8 This is important as 60% – 70% of medical decisions rely on laboratory investigations.⁹ Due to advancements in laboratory methods, instruments, and monitoring through internal quality control and external quality assurance programmes, fewer errors are typically attributed to the analytical phase.^8,10 The pre-analytical phase involves test requisition and sample labelling, collection, handling, transport, and processing. This phase accounts for up to 70% – 75% of all laboratory errors reported, most of which are due to human errors.^8,9 These errors can lead to samples that are unsuitable for analysis or unreportable results requiring rejection. Pre-analytical factors that may affect sample quality include haemolysed, clotted, or icteric samples, as well as samples that are mislabelled, unlabelled, or collected in inappropriate tubes. Insufficient sample volume not only limits the number of tests that can be done from the received sample but can also affect sample quality, as the blood-to-additive ratio may interfere with analysis and produce erroneous results.^8,9 Mild contamination by ethylenediaminetetraacetic acid (EDTA-K), a potassium-containing anticoagulant, may also cause subtle analyte changes that can be missed.¹¹ Prolonged centrifugation of samples during the pre-analytical phase can also interfere with sample quality where, for example, platelets can lyse, resulting in falsely elevated potassium levels.^12,13,14

Falsely elevated potassium levels or pseudohyperkalaemia occurs when generated results are not consistent with the clinical features of the patient.^4,12 Not only can pseudohyperkalaemia result in inappropriate patient management, but it can also mask hypokalaemia, leading to a missed diagnosis as concentrations may falsely present within a ‘normal’ reference interval.¹² The most commonly occurring source of pseudohyperkalaemia is haemolysis, which is reported to cause up to fivefold more rejections than any other reason.^7,13

A subset of haemolysed samples may result from endogenous causes such as haemolytic anaemia, immune reactions, toxin exposure, and haemodialysis treatment, or due to direct damage to red blood cells.^3,5 In vitro or exogenous haemolysis accounts for most haemolysed cases. Causes include poor sample acquisition technique, substandard handling and transport, prolonged storage time, and delayed processing.^15,16 Sample acquisition techniques that specifically affect potassium include prolonged tourniquet application and patient fist clenching, which can cause a 1 mmol/L – 2 mmol/L increase in potassium levels due to potassium efflux from cells during muscle contraction.^12,13,14 The use of needles with unsuitable diameters during sample collection can also result in haemolysis. In addition, cold temperatures can decrease the function of the sodium/potassium-adenosine triphosphase pump, resulting in the passive movement of potassium down the concentration gradient and out of red blood cells.¹²

Rejection of samples or test results inconveniences both patients and healthcare personnel as it delays urgent clinical decision-making. Specimen recollection may be required, leading to prolonged admission for the patient and unnecessary costs to the laboratory and the hospital or clinic.^10,17 Pre-analytical errors are estimated to cost approximately 0.23% – 1.02% of a hospital’s total operating budget in the United States.^9,10 While there are at present no local studies on the financial implications of hyperkalaemia on the health sector in South Africa, the prevalence of hyperkalaemia¹⁸ and its contribution to prolonged hospital stay have been previously described.¹⁹ Magwai and colleagues recently described a rejection rate of 1.4% in an academic laboratory in South Africa.²⁰

Although sample and result rejection may appear to be a small and insignificant factor in the total testing process, it is an important quality indicator for laboratories and can have great implications for the patient and health facility. This study thus aimed to assess the frequency and reasons for potassium test request rejection at the National Health Laboratory Service (NHLS), Tshwane Academic Division (TAD), South Africa, during the period from 01 January 2018 to 31 December 2018 and the financial impact of these rejections on this laboratory.

The potassium result rejection rate for the NHLS TAD Chemistry laboratory in 2018 was 15.1%, and there was a statistically significant difference between the reasons for potassium result rejection at the different facility types. Old sample was the main reason for rejections, comprising 41.4% of total rejections. This was the most common reason for rejection for both PHs and outpatient facilities. Owing to the location of some of these facilities, it is likely that the reasons for delayed analysis might be multifactorial. These are likely to include distance, long time lapse following collection, storage, and transportation conditions. Other potential reasons for delay in testing may include batch testing or factors related to the testing laboratory itself such as excessive workload, instrument issues, and sample sharing between departments, which may delay sample arrival to the chemistry laboratory as previously described in studies from China and the United States.^16,22

Old samples are rejected because of biochemical changes that occur as red blood cells break down during long storage, thus leading to erroneous results.¹⁵ Delays in sample analysis can affect accuracy and quality, and, as indicated in a study from the United States, samples stored at temperatures of 2 °C – 8 °C for longer than 24 h result in spuriously elevated potassium levels due to cessation of the sodium/potassium-adenosine triphosphatase pump.¹² A study from France has indicated that although potassium levels are stable at room temperature, samples should be analysed within 4 h – 6 h for accurate measurements.²³ Moreover, increased potassium levels have been reported in samples stored at room temperature within 1 h of collection in a study done in China in 2017.¹⁶ Another factor to consider when samples are sent to external or distant laboratories for testing is whether the samples were centrifuged at the collection site prior to transportation or only after they arrived at the testing laboratory. Delayed serum separation may cause potassium to leak out of the red blood cells, yielding hyperkalaemia. Ideally, according to a United Kingdom study done in 2003, samples should be centrifuged within 1 h of collection to prevent these falsely increased potassium levels.²⁴ Additionally, an important aspect to be aware of for potassium measurements is susceptibility to seasonal pseudohyperkalaemia, where potassium levels are elevated in the cooler winter months owing to the inhibition of the sodium/potassium-adenosine triphosphatase pump activity.^12,24

Interestingly, the main reason for result rejection in this study, old sample, is not the most common cause of rejection reported in previous studies done in Ethiopia, the United States, and Turkey between 2015 and 2016.^10,11,25,26 The major reasons for rejection reported in these studies include haemolysis, insufficient sample, incorrect sample or tube, clotted samples, contamination, and mislabelled or unlabelled samples.^26,27 This difference may, however, be explained by the fact that about 74% of all samples received at this laboratory were referred from off-site facilities, thus making delayed sample analysis a major factor causing rejection.

The second most common reason for potassium result rejection in this study was haemolysed samples (38.2%). This was the main reason for potassium result rejection at the Steve Biko Academic Hospital (42.02%), a facility that makes up a quarter of all samples received by the laboratory. At the peripheral facilities, approximately 50% of rejections were due to haemolysed samples, which is lower than previous reports from Australia and Turkey in 2010, where haemolysed samples account for about 60% of all rejections.^6,28 Findings from previous studies done in Italy in 2008 and Malaysia in 2019 suggest that haemolysis is the main cause of pseudohyperkalaemia.^7,13 Making medical decisions based on these erroneous results can lead to inappropriate patient care. This may occur when there is pseudohyperkalaemia or in instances when hypokalaemia is missed due to false potassium elevation in the sample.¹² A Malaysian study done in 2019 demonstrated that although haemolysis may occur in vivo, this only occurs in about 2% of cases.⁷ When in vivo haemolysis is suspected, additional information such as patient history, haptoglobin measurements, bilirubin levels, and red blood cell count are required to arrive at the diagnosis.⁷ In 2020, a study in Italy reported that in vitro or exogenous causes are more common and include poor sample collection techniques, poor sample handling and transport, prolonged storage time and inadequate storage temperature.²⁹ Personnel collecting samples should be trained and educated on the importance of correct collection techniques. A study investigating the impact of educational training among nurses in a hospital in Oman in 2017 showed about 75.9% improvement in sample quality after re-training.³⁰ The availability of specially trained phlebotomy personnel within a facility is important not just for proper blood sample collection, but also to ease the workload on the already overworked nurses and doctors, who may tend to rush sample collection due to the increased workload.³⁰ Haemolysed samples certainly pose a serious challenge for clinical laboratories.⁷

Contamination by EDTA-K anticoagulant as a cause of pseudohyperkalaemia accounted for 6.3% of the potassium rejections in this study and is common in clinical chemistry laboratories.³¹ As reported in studies from Thailand in 2019³² and India in 2020,³³ EDTA-K contamination may potentially lead to patient mismanagement. Thus, technique and order of draw need to be cautiously considered during sample collection to avoid contamination.³⁴

Rejection of test results can negatively impact patient care at the health facility and laboratory. Importantly, rejected results can delay critical and potentially life-saving clinical decisions. Based on our calculations, the NHLS TAD incurred a financial loss of almost R1 000 000.00 ZAR (approximately $56 000.00 USD) in a single year due to potassium test request rejections. An additional aspect to consider is the person-hours wasted due to these rejections. Rejection of 29 806 potassium test requests over a 12-month period equates to approximately 2500 rejections per month. If the processing time of each of these samples is 10 min, 25 000 min or 416 h per month are lost on wasted labour. This equates to approximately 17 days per month. Also, additional time is wasted as laboratory staff remove erroneous results and assign reasons for rejection. To solve this, the NHLS TAD could implement automatic cancellation by the laboratory information system of potassium requests that are older than 1 day. This would prevent the sample from being analysed and eliminate the subsequent wastage of financial resources and labour.