Minimizing pseudohyperkalaemia in blood samples from general practice
General Practitioner
E-mail: edwardcarabine{at}doctors.org.uk
 |
Abstract |
|---|
 TOP Abstract BackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
A day in the life of a GP usually begins with the daily trawl through our email inbox to check the results of the blood tests. The assessment of renal function, with U + Es (i.e. urea, creatinine, sodium and potassium), will form a significant percentage of them. Severe hyperkalaemia is a life-threatening emergency as it may cause cardiac arrhythmias, leading to cardiac arrest. With these potentially catastrophic sequelae, urgent investigation and treatment of genuine hyperkalaemia is essential. Common causes of hyperkalaemia include renal failure, metabolic acidosis and drugs (e.g. potassium sparing diuretics, Angiotensin Converting Enzyme (ACE) inhibitors and excess potassium therapy). However, a well recognized cause of raised serum potassium is pseudohyperkalaemia.
|
Background |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
Pseudohyperkalaemia occurs when the serum potassium level has been artificially raised by the release of potassium, during the collection, storage or transport of specimens. Causes include the following:
- Difficulty in taking the sample
- The patient has a clenched fist when the sample is taken
- The sample was shaken or squirted through a needle into the collection tube
- Cooling
- Contamination with anticoagulant from another sample, for example potassium ethylenediaminetetraacetic acid (EDTA)
- Deterioration of the specimen due to length of storage
Blood sampling in primary care is obviously faced with difficulties as the point of sample collection (i.e. the practice) is often a long distance from the hospital laboratory where the sample is usually processed. Once taken, the blood samples face an uncertain wait for collection from the practice and transport to the laboratory, before processing can take place.
Up to 30% of blood samples from general practice have serum potassium concentrations reported as above the reference range (Johnston and Hawthrone, 1997). The phenomenon of spurious hyperkalaemia recorded in blood samples that have undergone haemolysis is well recognized (Kerrensen et al., 1986; Moore et al., 1989; Masters et al., 1996). Serum potassium levels are particularly sensitive to temperature. At warm temperatures (25–30°C), a fall in serum potassium concentration is seen, due to stimulation of the red cell membrane ATPase. This regulates the exchange of potassium for sodium ions between the potassium-rich intracellular environment and the relatively low potassium concentration in serum (Stuart and Smellie, 2007). Average laboratory results for serum potassium in samples from primary care have been reported to fall by up to 0.5 mmol/l during summer's high temperatures (Trull et al., 2004). Higher temperatures (in excess of 30°C) or long storage, such as storage overnight, may lead to haemolysis of the sample and large rises in serum potassium. Cold temperatures inhibit the membrane ATPase, leading to pseudohyperkalaemia in the lower ambient temperatures of winter and when samples have long collection runs to the laboratory (Webster et al., 1952). However, studies have shown that in specimens stored at 18°C for up to 16 hours, the potassium remains stable (Masters et al., 1996).
Pseudohyperkalaemia can cause major difficulties in primary care and is a source of avoidable emergency referral and even admission into hospital. For the GP, even in the context of normal renal function and absence of likely causes, there is still a difficult decision to be made when faced with a raised potassium. Do we put it down to problems with sample collection, transport and delays in processing? Do we repeat the test, with the obvious anxiety for the patient and use of resources? Do we need the patient seen that day in hospital to exclude serious causes and prevent arrhythmias? There is also variation between GPs as to how they deal with an abnormal result. Some ignore all reported serum potassium concentrations, whereas others pursue and abnormal result with vigour (Johnston and Hawthorne, 1997). It poses a particular problem in the context of out-of-hours services, when the lack of patient information may lead to unnecessary investigations and hospital admissions being arranged by on-call doctors. Hence, a spurious result is a source of unnecessary and significant stress for both the doctor and the patient.
|
Reason for choice of audit |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
My interest in this area stemmed from an incident that occurred while I was working for the local out-of-hours service as part of my GP training. I was working a ‘car session’ visiting patients at home who had been triaged at the call centre. Having just finished a visit, I received a phone call from one of the doctors at the base. They had received a call from the local laboratory with a potassium result of 6.8 mmol/l and the patient happened to be from my practice. They asked if I knew anything further about the patient as they had no background information on the patient and were finding it difficult to decide what to do regarding the result. Unfortunately, I had no additional knowledge on this patient. In the context of the patient's normal renal function and the fact that there had been a 9-hour delay between sample collection and processing, we felt it likely that the result was spurious. However, after much discussion, it was felt prudent to contact the patient by phone. She received a phone call at 11.30 p.m. informing her of the result, with the purpose of obtaining additional information about her medical history and drug regime. Following a sleepless night, the patient attended the surgery for a repeat test the following morning. She was understandably anxious and required an explanation about why this may have occurred. Thankfully, the repeat potassium was within the normal range and the patient and I were duly reassured.
Following this incident, I spoke to my trainer and the other partners in the practice and it emerged that this was not an uncommon scenario. They found themselves faced with high potassium levels on almost a daily basis. There was some variation in how each dealt with hyperkalaemia and each had appeared to develop their own cut-off, above which they would order a repeat test. All agreed that it was a source of significant anxiety for both doctors and patients. Some also expressed the opinion that we seemed to have ‘a lot of abnormal potassium results’ and it was simply blamed on the length of time it took to get the samples to the laboratory.
I decided to look into current best practice and compare it to our own with the hope of decreasing the incidence of spurious hyperkalaemia in our specimens. Clear guidance on the subject is difficult to find as there are no nationally or locally defined criteria that outline best practice for the collection, storage and transportation of blood samples taken in primary care. I contacted our local laboratory for guidance and spoke with the laboratory manager who agreed that this was a significant problem. He stated that the problem had been made worse recently as the specimen reception responsible for receiving, booking in and distributing the primary care samples had been removed from laboratory control in a cost-cutting measure. As a result, the reception was understaffed and was ill equipped to handle the large numbers of samples that arrive from primary care following the lunchtime collections from practices. Samples were lying in reception for anything up to 6 hours before they were centrifuged. Following centrifuge, samples become stable.
There were, however, some measures under practice control that could be optimized to limit the effect of the later delay. His advice, coupled with the information I gathered from the relevant literature, enabled me to formulate a best practice protocol for the surgery.
|
Criteria chosen |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
- All samples should be processed by the laboratory within 6 hours.
- All practical steps should be taken within the practice to exclude pseudohyperkalaemia as a source of raised potassium in blood samples.
These criteria were chosen following consultation with the local laboratory manager and review of the relevant literature. Best practice for hospital-taken samples was that all samples should be centrifuged within 4 hours and the local hospital laboratory discarded any samples that exceeded this time limit. It was felt that primary care-based samples should meet this same exacting requirement; however, in view of the additional time required for transport from point of sampling to the laboratory and the evidence available, 6 hours was picked as the gold standard. In order to enable the accurate interpretation of blood results, they should be free from inconsistencies or errors in collection, storage and transport to eliminate the potential for pseudohyperkalaemia.
|
Standards chosen |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
- Eighty per cent of samples taken should be processed by the laboratory within 6 hours.
- In the absence of a known cause of hyperkalaemia, 90% of potassium concentrations should fall within the normal range.
I discussed these standards with both my trainer and the laboratory manager. It was felt that it was unrealistic to expect all samples to be processed within 6 hours given the transit time and the fact that there is only one collection a day from the practice. Given the difficulties with transport and processing, it was agreed initially to aim for 80% of samples to be processed within 6 hours.
As it is unusual to encounter genuine hyperkalaemia in the absence of renal failure or causative medication, any raised potassium concentration in the absence of these has a high probability of being factitious. In view of this and as a way of quantifying the many factors that contribute to spurious hyperkalaemia, I looked at serum plasma concentrations in patients without any obvious cause of hyperkalaemia. In this setting, you would expect the vast majority of potassium concentrations to fall within the laboratory reference range. It is not reasonable to expect all results to fall within the range as there may be instances of rarer causes of hyperkalaemia. Following discussion with my trainer and the laboratory manager, it was agreed that 90% of results should fall within the reference range once renal failure and medication effects had been excluded.
|
|
Preparation and planning |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
This audit topic emerged from the incident described above. I felt it was a significant event and presented it as such at the practice meeting. Everyone agreed that the audit topic was relevant and could well highlight areas where improvements could be made within the practice. Following this discussion, I undertook an initial data collection which I presented to the practice multidisciplinary team meeting. In attendance were the practice manager, GP partners, GP registrars and the practice nurse. I also discussed the best practice for blood sample collection and storage with the practice phlebotomists and nurses, who are responsible for the majority of sample collection.
In addition, I performed a literature search via the Medline database, looking for relevant research and articles into best practice for collection, transport and storage of blood samples, especially those that were based in primary care. Further to this, I discussed the audit on two occasions with the laboratory manager responsible for the laboratory where the blood samples were being processed. An initial discussion took place prior to the first data collection to help set the criteria and standards for the search. Following this initial collection, I discussed my proposed changes with him prior to their implementation.
|
First data collection |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
I audited all the blood samples taken on the week commencing on the 26 March 2007 until the 30 March inclusive. I searched on VISION for all patients who attended an appointment with either the phlebotomists or practice nursing staff during this week. I then looked at the time the sample was collected and the time it was processed in the laboratory. The time that the sample was processed in the laboratory was obtained from the official results form received electronically from the lab. From this, I calculated the time between sample collection and sample processing.
I also looked at the potassium concentrations of these samples and compared them with the laboratory reference range (3.8–5.5 mmol/l). For those with a potassium concentrations above the reference range, I reviewed the patient notes on VISION and excluded any patients with evidence of renal failure (urea greater than 9 mmol/l or creatinine greater than 120 µmol/l) and those on drugs that raise potassium (ACE inhibitors, potassium-sparing diuretics and potassium salts). Samples that had haemolyzed prior to processing were included with the group of samples from patients with potassium concentrations above the reference range, as was one sample that was not processed by the lab due to an excessively long wait.
The results of the first round of data collection are summarized in Table 1. The mean time from sample collection to laboratory processing was 7 hours 33 minutes. When serum potassium was tested, the mean serum potassium concentration was 5.2 mmol/l.
|
Comparison with the standard set
There are a number of possible reasons why the standards were not met.
- Phlebotomists were using different gauge needles from the recommended 21 gauge (green needle) for some samples. Use of narrower needles carries a risk of hyperkalaemia secondary to damage to the red blood cells during sample collection. Haemolysis can also occur if syringes are used as the blood is squirted through the needle into the collection tube.
- The order in which the blood samples were taken varied from patient to patient. It is important that the biochemistry tube is taken first before the collection of other samples. This is to eliminate any contamination with potassium from anticoagulants used in other tubes, especially potassium EDTA for full blood counts. This can be transferred from one tube to the next on the tip of the needle, leading to artificially high potassium readings.
- Blood samples taken by phlebotomists and nurses were awaiting collection in the refrigerator. As previously discussed, samples should be stored at room temperature.
- Three blood samples were taken in the afternoon and thus had to wait overnight before they were collected and transported to the lab. As previously mentioned, a long storage time is associated with pseudohyperkalaemia.
- Following collection from the practice, which occurred between 12.30 p.m. and 1.30 p.m. during the week of data collection, some samples were not processed by the lab for many hours. The latest time of processing was 9.08 p.m. for a sample that was taken at 10.20 a.m. This 11-hour delay, of which 8 hours occurred after the sample had left the practice, would obviously impact on the result of the subsequent potassium concentration. The laboratory therefore cancelled the test and it had to be repeated. I tried to ascertain the reasons for this delay from the laboratory manager of the hospital trust where the laboratory is based. He thought that the hold up was due to a delay in the samples being booked into the system on their arrival at laboratory reception as a result of staffing issues.
|
Changes implemented |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
- Twenty-one gauge (green) needle and vacutainer system to be used for blood sampling.
- Patients not to clench fist when blood is taken.
- Transfer blood into collection tubes in the following order: biochemistry tube (i.e. potassium sample) first followed by the other tubes.
- Leave the samples taken within the practice at room temperature while awaiting collection.
- No samples for biochemical analysis to be taken in the afternoon. This will eliminate the problem of samples remaining in the practice overnight, while awaiting the once daily lunchtime collection.
All of the above were outlined in a practice protocol which was placed on the shared practice desktop for reference. In addition, a copy of this was attached to the two trolleys containing all of the necessary phlebotomy equipment to be used for quick reference.
Unfortunately, once the samples had left the practice itself, it was difficult to influence the process further as the process fell outside of our control. Following a series of productive discussions with the laboratory manager regarding the initial results, he emphasized to his staff the importance of prompt handling of specimens once they reached the laboratory.
Once the specimens have been collected from our practice by the courier, he visits several other practices before transporting the samples at the hospital laboratory at approximately 2 pm If samples are only being taken during the preceding morning, this should ensure that all samples are reaching the laboratory within 5 hours. This should be sufficient to ensure that more than 80% of samples are being processed within 6 hours of the collection time.
|
Second data collection |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
The second data collection was performed over the week commencing the 14 May 2007 until the 18 May inclusive. This was almost two months after the changes were implemented. Again, all blood samples taken by the phlebotomists and nurses were included, with the same exclusion criteria used when the potassium concentration was above the reference range.
Results are summarized in Table 1. On this occasion, the mean time from sample collection to laboratory processing was 8 hours 9 minutes. For those patients who had a sample taken for potassium measurement, the mean serum potassium concentration was 4.7 mmol/l.
Comparison with the first data collection and standards
The results from the second data collection, following the introduction of changes, were compared with the initial data collection and the standard. For the first standard, requiring 80% of blood samples to be processed within 6 hours of collection, there was no improvement whatsoever. The 4% of samples being processed during the required 6 hours falls well below the standard that we had set at 80%. There had been a significant improvement in the percentage of potassium concentrations falling outside the reference range by the time of the second data collection. In the second data collection, we exceeded the standard set with 96% of serum potassium levels falling within the normal range.
|
Conclusions and discussion |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
This audit emerged from a clinical problem that had been encountered by many members of the practice team and was a source of significant anxiety to staff and patients. Pseudohyperkalaemia arising from poor handling of blood samples collected by the surgery had been impacting on both clinical time and resources as patients with abnormal blood tests were contacted and new tests were ordered.
As the first data collection reveals, one-third of renal profiles ordered were coming back with a potassium above the reference range. Despite being unable to make any impact on the time taken for samples to be processed, the introduction of simple measures around the collection and storage of samples within the practice was able to decrease this to less than 4%. This will have a significant impact on both clinicians and patients as the changes make the chance of pseudohyperkalaemia less likely. Clinicians can therefore be more certain that elevated serum potassium is due to a genuine cause and investigate as appropriate. Of additional benefit, the better handling of samples will ensure that any hypokalaemia that may have been missed, due to artificial raising of the result to within the normal range, can now be recognized and treated.
We were unable to impact on time taken between collection and processing of the sample. The reason for this is that despite changes brought in to limit the length of time samples waited prior to collection in the practice (e.g. no samples being taken in the afternoon), the majority of the delay was occurring once the samples had left the practice and following their arrival at hospital. As such this was beyond our control and despite discussions with the parties involved, aforementioned staffing issues had not improved by the time of the second data collection. However, the results from the other part of this audit seems to support previous evidence that, at room temperature, potassium samples stored for up to 16 hours remain stable (Masters et al., 1996). If this is the case, then delay between collection and processing becomes much less of an issue, although limiting this remains best practice. In order to help with this, at the request of the local laboratory manager, a copy of this audit has been sent to him to showcase the delays in sample processing. He will use this as evidence to support his previous request for additional staff during the ‘lunchtime rush’.
This audit has highlighted the importance of careful sample collection and handling. The importance of correct storage cannot be overemphasized as this helps to limit the impact of the inevitable delay in sample processing that is associated with primary care blood sampling. The process has also illustrated how evidence-based medicine can influence quality of care and improve the experience for both staff and patients.
When I initially began to look at the problem, some colleagues felt that it was ‘the laboratory's problem’ and that nothing could be done ‘at our end’. The initial assumption that the spurious hyperkalaemia was solely down to a delay in processing has been proved incorrect. The process also shows the importance of significant event analysis, as in this case, investigation into a clinical problem has helped ensure that this is less likely to occur again.
The experience of our practice may demonstrate a lack of knowledge around the appropriate collection, storage and transport of samples from primary care to the laboratories, which are almost exclusively situated in secondary care settings. This could be a learning need for other practices. Clear advice regarding best practice in this area is required either nationally or locally to help educate GPs and limit the number of spuriously raised potassium results causing concern in the community.
As all members of staff are now trained and aware of the best practice for the collection and handling of blood samples, this audit should make a significant and enduring impact on the practice.
|
References |
|---|
|
TOPAbstractBackgroundReason for choice of...Criteria chosenStandards chosenPreparation and planningFirst data collectionChanges implementedSecond data collectionConclusions and discussionReferences |
|---|
-
Johnston JD, Hawthorne SW. How to minimise factitious hyperkalaemia in blood samples from general practice. British Medical Journal (1997) 314:1200.
Kerrensen L, Lins RL, Neels H, De Broe ME. Effects of needle size and storage temperatures on measurements of serum potassium. Clinical Chemistry (1986) 32:698–699.[Web of Science][Medline]
Masters PW, Lawson N, Marenah CB, Maile LJ. High ambient temperature: a spurious cause of hypokalaemia. British Medical Journal (1996) 312:1652–1653.
Moore D, Walker P, Ismail A. The alteration of serum potassium level during sample transit. The Practitioner (1989) 23:395–397.
Stuart W, Smellie A. Spurious hyperkalaemia. British Medical Journal (2007) 334:693–695.
Trull AK, Jackson C, Walsh S, Thornton A, Culank LS, McHugh J. The perennial problem with potassium. Annals of Clinical Biochemistry (2004) 41:47–52.
Webster JH, Neff J, Schiffiano SS, Richmond AM. Evaluation of serum potassium levels. American Journal of Clinical Pathology (1952) 22:833–842.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||