Lab errors: if you can't catch them, avoid them (Proceedings)

Introduction

Quality Assurance in the clinical laboratory consists of an important chain of events that can be divided into three general phases: the pre-analytical phase, the analytical phase, and the post-analytical phase. These phases are further broken down into several links; the strength of each and every link is as important as the next regarding maintaining the quality of the chain as a whole and the ultimate impact on patient care.

The pre-analytical phase includes the multiple steps prior to the actual sample testing. Aspects relate to the order (right test and time), patient (preparation, identification) and sample (collection and handling). As the name implies, the analytical phase relates to the steps directly involved in generating a test result including the instrument performance, the method, manual technique, reagents, etc. The post-analytical phase involves all the steps occurring after the generation of the test result including reporting, data management, interpretation, diagnosis and treatment. Since the pre-analytical phase is by most accounts the most common source of errors related to laboratory testing, this session will focus on some of the more common causes and effects of these errors, and how to avoid them.

The pre-analytical phase

The pre-analytical phase involves many links in the quality assurance chain of events in laboratory testing. Insuring the appropriate test is requested at the appropriate time are links beyond the scope of this session. Patient identification is a huge source of error in the human testing world. Because of this, large hospitals go to great expense to automate this step with the use of bar-coded identification wristbands and this bar code follows the patient's sample throughout the testing process. This is not practical and is perhaps less of an issue in the smaller clinic. Needless to say, appropriate patient identification and accurate legible sample identification are important steps.

Patient preparation

Lipemia is one interferent that can sometimes be avoided with appropriate fasting of the patient prior to sample collection. Owners should be reminded to fast their pets prior to blood sampling whenever possible. However some metabolic conditions as hypothyroidism, diabetes mellitus and Cushing's disease may induce lipemic samples even in fasted states. Lipemia creates a turbid sample that may interfere with spectrophotometric methods as with hemoglobin concentration and many of the biochemical determinations. It also often induces hemolysis so the sample resembles tomato soup. Lipemia will falsely increase the hemoglobin concentration and since the calculation for MCHC uses this value, it will also be increased. One can easily check for lipemia by examination of the microhematocrit tube after spinning. Biochemical instruments can minimize these effects with sample blanking and lipemia is less of an interferent in systems using dry-slide technology. Lipemia and paraproteinemia can create pseudohyponatremia via volume exclusion in systems that dilute samples prior to measurements by ion-selectrode technology (ISE). Such systems are typically located in larger referral laboratories and are known as "indirect ISE" methods. The smaller systems that test undiluted samples with ISE technology, as i-STAT and IRMA, are not subject to this issue. For this reasons, comparisons between these two system types should be limited to non-lipemic samples and samples from myeloma patients should be excluded.

Sample collection

There are several components of the phlebotomy technique that require consideration. First is the anticoagulant or tube type and draw order. Those trained in blood collection should already be aware that the specimen of choice for hematology studies in domestic animals is EDTA (lavender top) because of preservation of cellular morphology and stain-ability; for chemistries it is typically no anticoagulant (red top; clot tube), for coagulation is citrate (blue top) and for blood gases and electrolytes is heparin (heparinized anaerobic syringe and green top, respectively). Notable hematology exceptions are discussed in the exotic session "Scratching the Surface".

One may not realize why these are important and what exceptions can be made in the event that not all desired samples are able to be collected. EDTA is comprised of potassium ethylenediaminetetraacetic acid and may be obtained in liquid (K3-EDTA) or dried solution (K2-EDTA) forms. It acts by chelating calcium, along with other divalent cations as zinc and magnesium, to prevent the natural clotting process. Chelation of these cations removes them from the test plasma resulting in falsely reduced calcium, zinc and magnesium concentrations. Since magnesium (or in some cases zinc) is required for the chemical reaction for determining alkaline phosphatase activity, ALP results will also be falsely decreased. In addition, since EDTA is formulated with potassium (K+), potassium results will be falsely elevated. If one uses EDTA whole blood or plasma for these tests, the results obtained will be incompatible with life and therefore obviously erroneous. If, however, a clot tube is simply contaminated from carryover effects of filling the EDTA tube first, the effects can be more subtle and may lead to erroneous interpretations.

Since the goal is to prevent clotting of samples with anticoagulant, a syringe is often used with an attached needle to fill the EDTA first and clot tube second; after all, who cares if the clot tube clots in the meantime, right? Wrong. For tubes commonly collected for the clinical laboratory testing, the tube order recommended by the National Committee for Clinical Laboratory Standards (NCCLS) is Blue (citrate), Red (clot), green (heparin), and lavender (EDTA). This avoids contaminating the serum with EDTA or heparin. The citrate should be filled first to insure clotting is prevented.

It is possible, however, to use some tubes in lieu of others in select cases. For example, heparin plasma may be used for many biochemical tests as well as hematology. Drawbacks of using heparinized samples for hematology include less than ideal staining, shortened preservation of morphology, and automated differentials may not be valid. Fresh films should always be made, and doing so followed by immediate testing can minimize some of these effects. Alternatively, EDTA plasma can often be used for specific biochemical tests as glucose, urea, enzymes, albumin and creatinine. The use of alternative samples should be verified with the technical service support personnel for the specific equipment in use.

A common question is "How much sample do you need to run X"? The volume of sample required may be very small compared to how full the tube should be to provide accurate results. It is important that the blue top tubes collected for coagulation studies are appropriately filled to maintain the 1 part citrate to 9 parts blood ratio. In other words, these should be at least 90% full, but not overfilled. If using a butterfly to fill the coagulation tube by vacuum (i.e. vacutainer), collect a discard tube first to eliminate the effects of collecting air from the tubing. Significantly under-filled tubes can lead to erroneously prolonged coagulation times and referral labs will likely reject the sample. Samples collected in EDTA and heparin should be at least half-filled to minimize effects of excess anticoagulant and filling to the intended volume is ideal. This is particularly true for liquid EDTA tubes used for hematology. The highly concentrated EDTA has a shrinkage effect on erythrocytes which is most noticeable on manual PCV determinations; this is ameliorated by the isotonic diluents used in most analyzers so automated hematocrit determinations are altered to a lesser degree. In addition, the liquid EDTA will actually dilute the sample so, for example, the total effect on a tube filled only 1/4-full may have a normal PCV reduced by ~11% from the actual value. This change is enough to make a normal patient appear anemic. Another point to consider is the effect of excess EDTA on plasma proteins determined with a refractometer. The EDTA in a 2ml tube containing 7.5% EDTA will have an apparent protein content of ~7.5 g/dl, with no protein actually in it. Larger tubes may contain 15% EDTA with an apparent protein content of ~15 g/dl; that level can significantly increase the apparent protein content if under filled. The effect is even greater on transudates (low protein cavity fluids). Tubes containing dried K2-EDTA are more expensive but recommended to avoid these phenomena and micro collection tubes (i.e. for 200-500ul) should be kept on hand for the very small or compromised patient.

Heparin is an acidic solution and the oxygen content is in equilibrium with room air. As long as the concentration of heparin is kept below 200 IU/ml blood, blood gas values will not be affected. Anticoagulation is achieved by adding 0.2 ml heparin (1000IIU/ml) in 5 ml blood for a final heparin concentration of 40 IU/ml. If more than .5 ml heparin is added to 5 ml blood, pCO2 and the calculated values using pCO2 (bicarbonate and base excess) can be significantly reduced but pH and pO2 are not. Five ml of blood is significantly more blood than required by current analyzers for testing and as a result, the temptation is to collect smaller sample volumes. If collecting samples in a 2-ml syringe with only the 0.1 ml dead space filled with heparin, completely filling the syringe obtains a final concentration of 50 IU/ml heparin and should obtain accurate blood gas results. However, at these volumes, there may be >0.15 negative bias in ionized calcium values due to the effect calcium binding by heparin and from dilution. For ionized calcium determination, it is best to test immediately with non-anticoagulated blood (as with an i-Stat), use appropriate heparin/blood ratios or use syringes containing dried balanced heparin that allow for collecting smaller volumes (PICO syringe, Radiometer America).

Sample handling

Once the tube is collected, inappropriate handling of the sample prior to testing can be a large source of erroneous results. Regardless of the anticoagulant used, once the tube is filled, it should be mixed by twisting the wrist at least 8-10 times. Simply filling the tube is not sufficient mixing and clots may form. Large clots are often detected by gross observation of the tube. Detection of smaller clots may require stirring the sample with one or two wooden applicator sticks. The clots will adhere to the stick for easy observation. Microclots may only be detectable microscopically. Obtaining a new sample is recommended as the presence of clots can adversely affect hematology and biochemistry results as well as plug analyzers leading to wasted test units or valuable time spent cleaning and troubleshooting. Clots are composed of platelets, effectively reducing or completely removing them from the sample resulting in falsely low platelet counts. Microclots may be mistakenly counted as leukocytes by automated analyzers, and usually as lymphocytes. Their presence may provide a classic pattern on the histogram (failure of left leukocyte/lymphocyte peak to approach baseline and forming a continuum with the platelet curve) or cytogram (dots spraying up and out from left corner) produced by impedance or light scatter cell counters, respectively. Larger clots may also incorporate enough leukocytes or erythrocytes to alter the accuracy of these results as well.

As mentioned previously, blood films should always be made fresh. The preservative action of EDTA only maintains cellular morphology for about 24 hours and only for 4 hours when examining for infectious agents as Mycoplasma sp. If shipping to a referral lab is necessary, protecting the blood from extreme heat or cold is also important to avoid hemolysis and altered cellular morphology. Since cellular swelling occurs with prolonged storage and can falsely elevate the MCV and PCV, overnight shipping is recommended.

If the tube for hematology or whole blood chemistry testing (acid-base, metabolites) is allowed to set for any length of time, settling occurs. This occurs most quickly in sick animals and horses. It may be grossly apparent in the top of a racked sample tube with an air pocket. Remixing is essential for accurate results. Depending on how settled and where the aspiration needle is placed in the sample, results may be falsely increased or decreased. Typically erythrocytes settle more quickly than leukocytes and platelets. Settling of a sample is less apparent in anaerobic samples collected for arterial blood gas evaluation that by necessity lack an air pocket. In these samples, inhomogeneous samples can lead to spurious results. Since there is no air pocket, mixing must occur by wrist-twisting 5-10 times followed by rolling the syringe between the palms for 5-10 seconds.

Samples collected for blood gas and ionized calcium determinations must be collected anaerobically and any air bubbles present should removed from the sample within 30 seconds. Failure to do so will result in falsely high pO2 readings and falsely low i-Ca concentrations. Since air can pass through plastic syringes, samples stored longer than 30 minutes should be collected in glass and placed in an ice-water bath. If plastic is used, the samples should be stored at room temperature and tested within 30 minutes of collection.

When serum is needed for biochemical testing, be sure to allow sufficient clotting time prior to centrifugation. Fifteen to twenty minutes is required for most species and thirty minutes or more may be necessary for some species as cattle or patients with clotting problems. Check for good clotting my holding the tube upside down with the cap on and giving it a shake. If the clot holds in place, it is probably ready to spin. If the serum forms a gel after centrifugation, insufficient clotting has occurred and fibrin clots may continue to form even after the serum had been harvested. These clots can plug up cartridges and analyzers, rendering the cartridge useless, generate erroneous results or creating instrument malfunction that requires technical troubleshooting and cleaning. For this reason, it is often desirable to use heparin plasma for biochemical testing; the clotting step is eliminated and turn-around-time is reduced accordingly. Be sure to verify the ability to use heparin plasma with the company's technical services department if it is not stated in the operator manual.

Once the tube has been centrifuged, insure that the serum or plasma is free of erythrocytes and harvest the serum or plasma as soon as possible. If the sample appears pink or red and hazy or cloudy, spin the separated sample a second time and pour over or pipette into a clean tube. Erythrocytes continue to metabolize glucose at a rate of ~10%/hour at room temperature and certain infectious agents may metabolize glucose even more quickly. RBC constituents, especially K+ in large animals, dogs with high reticulocyte counts, and Japanese breeds known to have high red cell potassium content as Akitas, may leach out or be released after hemolysis. Note that serum will have slightly higher K+ concentrations than plasma due to its release from platelets during clotting. The higher the platelet count, the higher the difference. Leukocyte concentrations > 100k/ul may also contribute to artifactually high K+ concentration for similar reasons. The turbidity created by samples containing retained erythrocytes may also interfere with spectrophotometric methods. Other interferences may exist and the extent may vary between instruments.

Appropriate patient preparation, sample collection and handling can significantly reduce errors and the frustrations experienced in the clinical laboratory. The more common sources of these errors have been discussed. It is highly recommended that at least one dedicated individual be responsible for thorough knowledge of these concepts as they relate to the instrument(s) in use so that the quality assurance chain remains intact.

Suggested reading

Lassen, DL, Weiser MG, Laboratory Technology for Veterinary Medicine in Thrall's Veterinary Hematology and Clinical Chemistry, Lippincott, Williams, and Wilkins, 2004

Weiser MG, Vap L, Perspectives and Advances in In-Clinic Laboratory Diagnostic Capabilities: Hematology and Clinical Chemistry in Vet Clinics of North America, Small Animal Practice, March 2007

Allison RW, Meinkoth JH, Sample Collection and Handling: Getting Accurate Results in Vet Clinics of North America, Small Animal Practice, March 2007

Other References

Higgins C, Ionized Calcium, AcuteCareTesting.org, Radiometer July 2007

NCCLS Ionized calcium determinations: Precollection variables, Specimen Choice, Collection and Handling: Approved guideline (2nd ed) NCCLS document C31-A2 2001 NCCLS Pennsylvania USA.

Narayanan S, Preanalytical issues related to blood sample mixing, AcuteCareTesting.org, Radiometer. Oct. 2005.

Wennecke G, Useful tips to avoid preanalytical errors in blood gas testing: electrolytes, AcuteCareTesting.org, Radiometer, Oct. 2003.