Phlebotomy top gun: Measuring potassium in capillary blood samples

Expires: February 1, 2024

  • Presenter

    Brad Karon, M.D., Ph.D.
    Division Chair, Clinical Core Laboratory Services
    Professor of Laboratory Medicine and Pathology
    Mayo Clinic, Rochester, Minnesota


    Contact us:

    Transcript and references

    Speaker Introduction

    Every year at our phlebotomy conference, I present a talk called “Phlebotomy Top Gun.” The format for “Phlebotomy Top Gun” is a case-based presentation. I solicit from conference attendees ideas for cases based upon issues or questions you have related to phlebotomy. I present these as case-based scenarios, and using an audience response voting mechanism, the attendees of the conference vote on the action or answer they feel is appropriate for each case. I then present the evidence and data from the literature and Mayo collective experience related to the topic.

    At the end of the case, the attendees vote again. For each case, I can see whether I’ve been able to change anyone’s mind on the question at issue by presenting the data and information relevant to the topic. This is an actual case from a previous year’s phlebotomy conference “Phlebotomy Top Gun” presentation, derived from a question about the validity of measuring potassium from capillary puncture blood samples.


    I have no disclosures related to today’s talk.

    Regarding capillary compared to venous potassium (K) measurement, which of the following are true statements?

    Based upon the question submitted by a phlebotomy conference attendee two years ago, I presented this question to the conference audience. Regarding capillary compared to venous potassium (K) measurement, which of the following are true statements:

    1. Capillary samples are, on average, more hemolyzed than venous samples; therefore, potassium is systematically higher in capillary samples.
    2. For a given amount of free hemoglobin (hemolysis), the increase in potassium is higher in a capillary sample compared to a venous sample.
    3. Experienced techs or phlebotomists can collect capillary samples with no more free hemoglobin (hemolysis) than a venous sample.
    4. Both 1 and 2 are true.
    5. All of the above responses are true.

    At the conference, we would use a live audience feedback system to poll the audience and see what answers audience members believed to be correct about this issue. For those of you watching now, which do you think is the correct answer?

    Capillary blood

    What I do next with each case in the “Phlebotomy Top Gun” presentation is go through data and evidence, either internal or external to the Mayo Clinic practice, that I can find related to the question. At the end of each case, I re-poll the audience to see if anybody’s mind has been changed.

    To remind our audience today, capillary “blood” is really a mixture of arterial blood, venous blood, interstitial fluid (the fluid that bathes cells within our tissues), and fluid released from cells during the capillary puncture itself. That combination of interstitial fluid and cellular fluid is often called “tissue fluid.” While concentrations of most chemical analytes are identical or very similar between arterial and venous blood, it is tissue fluid that can vary dramatically in the concentration of various analytes we measure in the laboratory. To make matters more complicated, each capillary puncture will have a different amount of tissue fluid collected into the sample.

    The importance of this variable amount of tissue fluid collected in a capillary sampling differs by analyte and depends upon how different cellular and interstitial concentrations of the measured analyte are compared to venous or arterial blood. For example, there is no hemoglobin in interstitial fluid, so variable amounts of interstitial fluid will cause variability in capillary hemoglobin measurement compared to venous or plasma hemoglobin. After meals, interstitial fluid glucose concentration changes more slowly than venous or arterial glucose, and this complicates interpretation of capillary glucose in the non-fasting state. So, what about capillary sampling for potassium measurement? Let’s explore this question further, remembering that we break more cells during capillary puncture than during venipuncture.

    Studies of capillary K1

    On this slide, I provide a reference to an older study comparing potassium values from infants with simultaneous blood collection from an in-dwelling arterial catheter compared to a capillary heel stick puncture. The capillary samples yielded potassium measurements that were on average 1.2 millimolar higher than the paired arterial serum sample, consistent with the results of several other studies that demonstrated that capillary potassium measurements do run higher than values obtained from venous or arterial samples. But the question is, “Why is this?”

    Studies of hemolysis in capillary puncture2

    This next study, also a bit older, compared the average amount of free hemoglobin, from hemolysis during capillary blood collection between infants and children of varying ages. For comparison, the average amount of free hemoglobin in a venous specimen collected from an adult is less than 50 mg/dL. As you can see, capillary punctures performed on infants in the first two weeks of life resulted in, on average, almost 400 mg/dL of free hemoglobin in these samples. By two years of life, the average free hemoglobin level decreases to around 150 mg/dL. So, capillary puncture in very young infants will have more free hemoglobin than capillary puncture in children older than two years of age.

    Studies of hemolysis in capillary puncture3

    This next study was performed at Mayo Clinic around ten years ago. We wanted to know whether phlebotomist training impacted the amount of hemolysis, or free hemoglobin, in capillary puncture samples from newborns. We looked at three groups of capillary puncture samples. Group One consisted of capillary punctures performed by an outpatient phlebotomy group, while Group Two were samples collected by the very same group of outpatient phlebotomists after specific training or retraining (in this case) on capillary puncture in infants. Group Three consisted of samples collected by an inpatient pediatric phlebotomy group that has more experience and does more newborn capillary collections. As you can see, the bottom line of our study is that experience in terms of number of capillary punctures or re-training staff has no effect on the hemolysis levels or amount of free hemoglobin in capillary puncture samples from healthy newborns.

    Effect of hemolysis on capillary K4

    The last formal study that I’ll discuss is perhaps the most interesting. This study compared the amount of free hemoglobin, this time measured as a semi-quantitative index, to average potassium concentrations in both capillary and venous samples. This study compared average potassium levels in thousands of capillary and venous samples. The average potassium capillary and venous samples that had no hemolysis was the same as we would expect, at 4.0 millimolar. A small amount of free hemoglobin in the venous sample, the semi-quantitative H indices of 1 or 2, resulted in modest increases in potassium to 4.2 to 4.5 millimolar, respectively.

    In contrast, the same amount of free hemoglobin increased potassium in capillary samples to 4.5 and 5.0 millimolar. In other words, the same amount of hemolysis increased potassium concentration more in capillary samples compared to venous samples.

    Capillary K, blood gas samples

    This last slide describes our internal validation study of capillary blood gas testing at Mayo. We collected capillary blood gas specimens from 22 healthy infants, after heal warming and with hand transport of the capillary blood gas specimens to the testing lab.

    The average potassium level in the samples was 5.7 millimolar, so our study was consistent with others showing on average higher potassium levels in capillary samples compared to the venous samples, which have an average around 4.0. A look at the individual values for the 22 infants demonstrated potassium in these normal healthy infants of anywhere between 4 and 10 millimolar.

    Based upon this data showing wide variability in potassium in healthy infants, we decided not to report potassium values in our capillary blood gas samples from newborns.

    Capillary K conclusions

    To summarize the data and studies, capillary samples contain more free hemoglobin, higher levels of hemolysis, than venous or arterial blood samples. One study did find that for a given amount of free hemolysis or free hemoglobin, potassium increases more in capillary samples compared to venous samples. I believe that potassium measurement in capillary blood samples should only be done in laboratories that have the ability to measure free hemoglobin with every measurement and have a good process to withhold results when there is too much free hemoglobin to produce reliable potassium results. Testing laboratories might also consider implementing a stricter or different threshold for the amount of hemolysis or free hemoglobin in capillary samples.

    Regarding capillary compared to venous potassium (K) measurement, which of the following are true statements?

    At this point during the Top Gun phlebotomy presentation at our phlebotomy conference, I would re-poll the audience to see if the attendees had changed their answers based upon the information I had provided. I would also reveal what I think is the correct answer — in this case, I believe the correct answer is #4, the first two statements are true. Capillary samples have more free hemoglobin than venous samples, and there is some evidence that a given amount of hemolysis in a capillary sample increases potassium more than the same amount of hemolysis in a venous sample. However, there is no evidence that phlebotomist training or experience can impact hemolysis levels in capillary punctures.


    I have provided for you references to the studies I mentioned here today.

    1. Patel H, Ryan SW, McLain B. Sources of error in neonatal blood sampling. Arch Dis Child. 1988;63:752-753.
    2. Meites S, Lin SS, Thompson C. Studies on the quality of specimens obtained by skin puncture of children. Clin Chem. 1981;27(6)875-878.
    3. Algerciras-Schimnich A, Cook WJ, Milz TC, et al. Evaluation of hemoglobin interference in capillary heel-stick samples collected for determination of neonatal bilirubin. Clin Biochem. 2007;40:1311-1316.
    4. Oostendorp M, van Solinge WW, Kemperman H. Potassium but not locate dehydrogenase elevation due to in vitro hemolysis is higher in capillary than in venous blood samples. Arch Pathol Lab Med. 2012;136:1262-1265.

    Phlebotomy conference

    Thank you for your time and attention. If you liked today’s presentation, or even if you didn’t, please do consider attending the upcoming Mayo Clinic Laboratories phlebotomy conference, “Phlebotomy 2021: Empathy, Engagement, and Excellence in Patient Care,” which will take place virtually on April 15 and 16, 2021. As in previous years, the conference will feature presentations from expert speakers covering a variety of topics related to phlebotomy skills, phlebotomy management, and other topics. Learning will take place via virtual large groups and virtual small breakout group sessions. Virtual tours of Mayo Clinic facilities will also be offered during the conference. I will also be presenting an all new “Phlebotomy Top Gun” presentation based on this year’s attendees’ interests, issues, and concerns as expressed during the registration process. More information on the phlebotomy conference program can be found on the Mayo Clinic Laboratories website.

    Thank you for listening and have a wonderful day.

    MCL Education (@mmledu)

    MCL Education

    This post was developed by our Education and Technical Publications Team.