depending on the lipid composition, making it difficult to predict the extent or even the direction of interference for dissimilar matrices.
9.2 Estimated Impacts Based on Interference Studies
9.2.1 Interference by Light Scattering
Most of the existing data on turbidity interference is derived from the use of non- endogenous lipid emulsion. It is the industry standard to use the soybean-derived lipid emulsion Intralipid® to estimate lipemia/turbidity in interference studies [5].
Thus, most manufacturers’ package inserts effectively indicate how Intralipid® affects a given test. This is not entirely without merit, as there are a number of patients who receive this as nutritional supplement. However, when the cause of turbidity is not Intralipid®, this information may be misleading. Direct comparison of Intralipid® and endogenous lipemia has shown significant differences between the extent of bias [1].
In one study, pooled lipemic samples were compared against Intralipid® spiked into the same samples. Each pool was assayed for triglycerides, α-1 antitrypsin, cerulo- plasmin, haptoglobin, transthyretin, and albumin on a turbidimetric chemistry ana- lyzer. While Intralipid® had no significant effect on any of the measured analytes, endogenous lipids showed substantial negative bias for transthyretin (up to −52 %), ceruloplasmin (up to −38 %). It was thought that the mechanism of interference was reduction of the absorbance due to interaction between the endogenous lipids and the reagents in the tests. However, the complex composition of endogenous lipids makes it possible that other mechanisms, such as lipid partitioning or volume displacement (see Section 9.2.2 below), were also involved. The same study also compared detection of interference using the ‘L-index’. The detection of interference was largely similar between Intralipid® and endogenous lipids as was correlation with triglycerides.
However, a few individual samples showed marked differences between triglycerides and the ‘L-index’, supporting abnormally triglycerides-rich or triglycerides-poor lipo- proteins. The implications are that turbidity detection indices may not always corre- late with triglycerides. The more significant message was that Intralipid® cannot be used to reliably predict the effect of interference for individual patient results.
Other studies have also identified differential interference between lipid emul- sion and endogenous lipemia [6]. Endogenous lipemia has been shown to affect cre- atinine measurement using the Jaffe method. The Jaffe method is notoriously sensi- tive to interferences in general and so it is not surprising that turbidity due to lipemia interferes with this method. What is surprising, is that lipemic interference can vary tremendously in different patients. In this study [6], the effect of lipemia on creatinine values ranged from none to −100 % (absurd values were found with some samples, e.g. −354 μmol/L). As noted in the aforementioned study [1], lipid emulsions and tri- glycerides concentration were not predictive of the extent of interference. Another
important finding was that the age of the lipids affected the extent of the interference;
overnight storage eliminated the interference. This means that even studies that use endogenous lipids may not identify clinically significant bias if the materials used are not identical to those encountered clinically. Because of most of these factors, it can be conclude that some analytical errors due to hyperlipidemia may go unrecognized even despite the best attempts of the laboratory to detect turbidity and avoid report- ing unacceptably biased results. While there is no easy solution to these problems, it is worthwhile being aware of these potential pitfalls, such that they may be addressed on a case by case basis when consulted regarding results that are inconsistent with clinical findings.
9.2.2 Interference by Volume Displacement
The second most common type of interference due to lipemia is volume displace- ment. The clinical importance of turbidity interference is highlighted in numerous case reports of pseudohyponatremia [7–10]. For clarity, pseudohyponatremia in this chapter refers to factitiously low sodium due to hyperlipidemia (as opposed to low sodium due to cellular shifts in a hyperosmolar state). Pseudohyponatremia effec- tively amounts to short sampling (not pipetting enough volume), as a high concen- tration of lipids displaces water. Electrolyte measurements that depend on pipetting aqueous volume accurately are subject to error in the presence of high lipid (or protein) concentrations. In a classic pseudohyponatremia scenario [10], a patient with hyper- lipidemia, for example in pancreatitis, has a measurement of sodium by indirect ISE.
The high lipid concentration results in falsely low sodium values and the patient may be erroneously treated with hypertonic saline. Because of the erroneous lab value, the treatment results in iatrogenic hypernatremia and hyperosmolarity putting the patient at risk for cerebral dysfunction. While this phenomenon is well recognized in the literature, there are frequently cautionary reports published throughout the medical literature when the phenomenon goes unrecognized. Pseudohyponatremia due to turbidity is serious enough that even mild hyponatremia can prompt unneces- sary treatment and cause significant morbidity and mortality [7].
Besides pseudohyponatremia, it is reported that hyperlipidemia can affect other commonly measured electrolytes, such as potassium, and chloride among others (Tab. 9.1) [11]. In one study, sodium and chloride were decreased by ~1 mmol/L and potassium was decreased by 0.04 mmol/L for every 10 mmol/L increase in total lipid concentration [12]. Because of the apparent linear relationship they observed, the authors derived correction formula to correct for lipemia. While the findings of this study do highlight the effect of lipemia on the accuracy of electrolytes, it does bring to light an important concern. The practice of correcting laboratory results with inter- ference is in general strongly discouraged [13]. The heterogeneous lipid composition,
9.2 Estimated Impacts Based on Interference Studies 97
variable mechanisms of interference, and biologically variability between patients make it risky to try and correct results for a single patient.
Analyte Direction of Effecta Interferent Tested
α-1 antitrypsin increase/no effect Intralipid®, patient specimes
Albumin increase Intralipid®, patient specimes
ALT increase Intralipid®
AST increase Intralipid®
Calcium increase Intralipid®, patient specimes
Ceruloplasmin decrease Intralipid®, patient specimes
Chloride decrease Intralipid®
Conjugated Bilirubin decrease Intralipid®
Creatinine (Jaffe Method) no effect/decrease Intralipid®, patient specimes
CRP decrease Intralipid®
GGT decrease Intralipid®
HDL-C decrease Intralipid®
IgG increase Intralipid®
IgM decrease Intralipid®
Magnesium increase Intralipid®
Phenytoin decrease Intralipid®
Potassium decrease Intralipid®
Progesterone decrease Intralipid®
Salicylate decrease Intralipid®
Sodium decrease Intralipid®
Testosterone decrease Intralipid®
Theophylline decrease Intralipid®
Total Bilirubin decrease Intralipid®
Transferrin decrease Intralipid®, patient specimes
Transthyretin decrease Intralipid®, patient specimes
Uric Acid decrease Intralipid®
Vancomycin decrease Intralipid®
a Extent and direction of interference are instrument dependent; list is not exhaustive Tab. 9.1: Partial list of analytes affected by turbidity.
An approach to both identification and elimination of turbidity interference due to lipemia is ultracentrifugation. Ultracentrifugation refers to applying centrifugal force in excess of 30,000⋅g to fractionate the sample; high-speed centrifugation may also be used where samples are spun repeatedly in microcentrifuges, which achieve
~ 10,000–15,000⋅g. In principle, the sample is fractionated by density. Low density particles, such as lipids, rise to the top and heavy particles, such as cells, fibrinogen, or cellular debris sediment to the bottom. In one study, which used ultracentrifugation to identify analytical interference due to lipemia, clinically significant bias was found for creatinine, phosphorus, and liver function tests (ALT, GGT) [14]. Lipemic samples from 110 patients were ultracentrifuged at 40,000⋅g for 18 hrs and the results before
and after centrifugation were compared. Predictably, triglycerides and cholesterol showed the largest changes of 21 % and 7.5 % respectively; lipid particles are of course rich in these substances. What was not expected, however, was the high number of statistically significant differences found for other analytes, including phosphorus (5 %), creatinine (4 %), GGT (3 %), urea (2 %), iron (2 %), ALP (2 %), calcium (1.6 %) among others. Using error goals based on biological variation, the authors concluded that the effect of lipemia on phosphorus, creatinine, total protein and calcium was clinically significant. While the 18 hrs centrifugation times are unlikely to be repli- cated in clinical practice, it does show that lipemia affects a wide array of analytes.
Ultra-centrifugation is therefore a viable option for determining the effect of turbidity on analytical tests experimentally.
Unfortunately, ultracentrifugation is not a cure-all for lipemic specimens. Some analytes are themselves affected by the ultracentrifugation process, which can lead to inaccurate results. One widely-cited example of this is with coagulation testing (PT, APTT) [15], where ultracentrifugation causes sedimentation of fibrinogen or factor VIII when bound to von Willebrand factor. Loss of these key complexes results in no detectable clot with some non-mechanical methods. Thus, the solution (ultracen- trifugation) can be worse than the problem (lipemia) if one is not careful. The effect of lipemia on coagulation tests is itself method dependent, such that ultracentrifugation is not needed when using some methods. Newer analyzers with photo-optical clot detection have been shown to be unaffected by lipemia [16]. Because of this method- dependence, laboratorians need to be aware of how lipemia affects their own in-house methodology. When significant interference is found, it is essential to experimentally confirm that the process of removing interference (e.g. ultracentrifugation) does not cause another analytical error. When properly validated, ultracentrifugation serves as a useful tool for clinical laboratories to handle lipemic specimens.
Another issue for laboratorians to be aware of is that turbidity indices do not detect all lipid interferences that cause volume displacement. It is well known that factitiously low electrolytes may be found in patients with lipoprotein X. Lipoprotein X is an abnormal lipoprotein that may appear in patients with cholestasis. However, samples with high concentrations of lipoprotein X may remain clear and thus unde- tectable by turbidity indices. In one case study [17], the inability to immediately iden- tify this interference delayed appropriate treatment of a patient with severe obstruc- tive cholestasis secondary to pancreatic cancer; sodium was reported as 20 mmol/L lower than the correct value and potassium was reported up to 1 mmol/L lower than the accurate value. The solution to volume displacement due to lipoprotein X is actu- ally the same as turbidity due to lipemia, where direct ISEs and ultracentrifugation ameliorate the problem. Astute clinical awareness of apparent discrepancies between laboratory results and clinical findings remains essential for detection of all types of interference.