Which test could not be performed on a serum sample

Clin J Am Soc Nephrol. 2016 Sep 7; 11(9): 1675–1679.

Abstract

Plasma samples collected in tubes containing separator gels have replaced serum samples for most chemistry tests in many hospital and commercial laboratories. Use of plasma samples for blood tests in the dialysis population eliminates delays in sample processing while waiting for clotting to complete, laboratory technical issues associated with fibrin formation, repeat sample collection, and patient care issues caused by delay of results because of incompletely clotted specimens. Additionally, a larger volume of plasma is produced than serum for the same amount of blood collected. Plasma samples are also acceptable for most chemical tests involved in the care of patients with ESRD. This information becomes very important when United States regulatory requirements for ESRD inadvertently limit the type of sample that can be used for government reporting, quality assessment, and value–based payment initiatives. In this narrative, we summarize the renal community experience and how the subsequent resolution of the acceptability of phosphorus levels measured from serum and plasma samples may have significant implications in the country’s continued development of a value–based Medicare ESRD Quality Incentive Program.

Keywords: CMS, National Quality Forum, Hematologic Tests, Humans, Kidney Failure, Chronic, Medicare, Phosphorus, Plasma, Quality Improvement, renal dialysis

Introduction

For most chemical analytes of interest in testing blood samples, serum was the substrate of choice for many years. With the advent of collection tubes with separating gels, tubes could be transported over long distances and sampled directly without removing the serum samples from the collection tubes.

Serum samples present challenges when obtained from patients with ESRD treated by maintenance dialysis. One reason is a higher rate of therapy with anticoagulants, in addition to pathologic clotting, which has been linked to uremia itself (1). Samples from these patients with ESRD require long periods of time to clot completely. If samples are held beyond the recommended 30–60 minutes to allow complete clotting to occur, analytes, such as potassium and phosphorus, will be affected. Thus, samples processed in the traditional manner may be centrifuged even if they have not clotted completely. Therefore, the supernatant serum contains fibrinogen and other clotting factors that allow for continued thrombogenesis after separation.

When an incompletely clotted sample is tested in the laboratory without further processing, it may present as clotted serum that must be broken up and recentrifuged, or it may contain small clots that may clog up the fluidics of the laboratory instruments unless they have clot detection capability. The clots may cause erroneous results for the sample with the clots as well as the otherwise appropriately prepared samples tested after it. Troubleshooting and instrument service may be required before these samples can be tested again. Furthermore, in cases of erroneous results, the laboratory needs to track the samples to the right patients and communicate with the appropriate persons in each of their outpatient facilities to have the blood samples redrawn.

In addition to the burden that improperly clotted samples pose on the personnel time and resources within the laboratory, the need to redraw blood samples adds to the blood loss from patients with ESRD, particularly those treated by maintenance hemodialysis. In turn, it may require greater use of medications to enhance erythropoiesis for clinicians to maintain acceptable hemoglobin levels for these patients on dialysis. Furthermore, the dialysis facility staff not only need to spend extra time redrawing, recentrifuging, and resending the samples, they also face pressure from quality audits that routinely look for chemistry test results for monthly data when test results are unavailable because of sample issues. Significantly, the patient’s condition at the time of the collection of the original sample may be changed when the sample is recollected. If there is a 1-week delay between collection of the original sample and reporting of the redrawn sample, it could be a critical lapse of time. If the physician is not careful or does not recognize that the sample had to be recollected, it could lead to misinterpretation of results and potentially inappropriate changes in treatment. To our knowledge, no study has been done to detect if minor incidents of improper clotting could be contributory to the overall variability of the chemistry results obtained.

Recently, many hospital and commercial laboratories have converted to heparin plasma samples collected in gel separator tubes for most of these tests. Heparin plasma samples from anticoagulated patients do not have the clotting issues of serum samples. Waiting for the right time to centrifuge (i.e., after the specimen clots) is not required with heparin plasma samples. There is an additional advantage to using heparin plasma samples in that the volume of plasma produced is 15%–20% higher than serum from the same volume of blood—therefore, reduced extraction blood volumes could suffice. The differences between results obtained from heparin plasma and serum samples have been judged to be acceptably small for most chemical analytes of interest in testing the dialysis population; however, we found no report of comparability of results from samples taken from patients on dialysis (2–7). Although laboratory medicine, particularly for the in-hospital laboratory, has moved on to accept the use of many chemical analytic results from heparin plasma samples interchangeably with those traditionally measured from serum, this change has occurred under the radar of most nephrologists and the renal community at large, including regulators from the Centers for Medicare and Medicaid Services (CMS).

Clinicians evaluate laboratory values, such as calcium and phosphorus results reported from hospital or outpatient dialysis laboratories interchangeably from patients with ESRD, without necessarily realizing the contribution of serum versus plasma to the overall assay variability. However, this practice works clinically, because the interlaboratory variability turns out to be larger than the serum versus plasma variability. In the case example of calcium and phosphorus that we use to illustrate the point, the CMS had inadvertently locked regulatory clinical performance measure requirements into accepting only serum calcium and phosphorus results for its ESRD Quality Incentive Program (QIP) (8). However, the published rules also allowed for the use of calcium and phosphorus results obtained from settings other than outpatient dialysis, unknowingly including plasma–based test results from hospitals and other theaters of care. This type of inconsistency led the authors to work with the renal community in petitioning the CMS to revise the laboratory requirement for serum calcium and phosphorus in the proposed QIP rule during 2012.

In response, the CMS referred the renal community to its work with the National Quality Forum (NQF), providing an arena for both scientific and policy discussions. In particular, the NQF 0255 clinical performance measure for phosphorus (a QIP reporting measure) was amenable to an ad hoc technical review, becoming a test bed for this exercise. In December of 2012, we proposed and presented supporting information establishing that changes were necessary to accommodate industry–accepted standard measurements of both serum and plasma phosphorus. The comparability of predialysis phosphorus tests performed on paired serum and heparin plasma samples from 101 patients of dialysis as part of internal testing at Spectra Laboratories was presented. To illustrate, the side by side distribution of phosphorus results is shown in Figure 1A, and the correlation is shown in Figure 1B. We further sought to affirm the general principle that regulations on the basis of the NQF standards should avoid unnecessarily indicating a preference for a particular testing method without appropriate consultation with field experts and the stakeholders within the renal community. By December of 2013, after the discussion and deliberation had occurred, the NQF expert panel recommended that measure 0255 be amended to include either serum or heparin plasma phosphorus testing (i.e., remove serum or add heparin plasma to the measure specifications). They further noted that the key issue addressed in measure 0255 is whether phosphorus levels are monitored in patients with ESRD, and either serum or heparin plasma testing would be appropriate. In the PY2017–2018 proposed rule released in 2014, the CMS announced changes in the QIP that will eventually allow for the use of serum or plasma phosphorus by the year 2016 (9).

Relationship between simultaneously drawn serum and plasma phosphorus in 101 dialysis patients. (A) Histograms of serum and heparinized plasma phosphorus results (n=101 paired samples). (B) Scatterplot of heparinized plasma results versus serum results for phosphorus (n=101 paired samples). Standard error of estimate of estimate (sy/x) =0.16 mg/dl; r=1.00.

Although heparin plasma is the preferred sample for many chemistry analytes, with serum and plasma phosphorus results being practically interchangeable, heparin is not acceptable for all analytes. A World Health Organization (WHO) document states (10) that heparin plasma samples are recommended for several analytes, because the constituents in plasma are better at reflecting the pathologic situation of a patient than those in serum: alkaline phosphatase, aspartate aminotransferase (AST), hemoglobin (plasma), ionized calcium, lactate dehydrogenase, phosphorus (inorganic), potassium, protein (total), and thyroxine.

The same WHO document also states that heparin plasma samples “can be used without changes of result” for many other analytes on the basis of acceptability of differences observed in comparison of results of testing of plasma samples and serum samples versus criteria on the basis of biologic variation (10). Analytes listed include alanine aminotransferase, albumin (when bichromatic assay is used), amylase, bicarbonate, bilirubin, brain natriuretic peptide, calcitonin, calcium, chloride, cholesterol, cholesterol (HDL), creatinine, creatinine kinase, d-dimer, erythropoietin, ferritin, folate, fructosamine, γ-glutamyl transferase, haptoglobin, homocysteine, insulin, iron, lipase, magnesium, osmolality, sodium, thyroxine, free thyroxine, tri-iodothryonine, triglycerides, urea, uric acid, vitamin B12, and zinc.

The WHO document further states that heparin plasma samples either are not recommended or are not recommended for all analytical methods for the following analytes: C-reactive protein (plasma not recommended for all methods), fibrin degradation products, cholesterol (LDL; not recommended for all methods), glucose, lipoprotein electrophoresis, parathyrin (PTH; EDTA plasma recommended; serum or heparin plasma acceptable for some methodologies), and somatotropin growth hormone (10). Of note, we have been using EDTA plasma samples for intact PTH since 1995. EDTA plasma samples are not generally used in the chemistry laboratory because of interference with calcium, electrolytes, iron, magnesium, and alkaline phosphatase assays as noted in the WHO document.

Our internal testing on paired samples to confirm the acceptability of heparin plasma samples after overnight transportation for analytes of interest in testing patients on dialysis (in addition to phosphorus results submitted to the NQF) is summarized in Tables 1 and ​2. Table 1 provides information for analytes for which plasma samples were judged acceptable in the WHO document (10) report. Table 2 provides information on analytes for which plasma either was not acceptable or acceptability information for plasma was not provided in the WHO document for some analytical methods. Tables 1 and ​2 include maximum desirable differences on the basis of biologic variation. Given in the work by Ricos et al. (11) as percentages, these were converted to concentration units at the mean serum concentrations to provide a basis for judging the observed differences.

Table 1.

Plasma-serum correlations from Spectra Laboratories internal testing for analytes for which heparin plasma samples are acceptable according to World Health Organization (10)

Table 2.

Plasma-serum correlations from Spectra Laboratories internal testing for analytes for which heparin plasma samples are not acceptable according to World Health Organization (10) (glucose) or acceptability information was given only for specific instrument systems (LDL cholesterol and C-reactive protein)

For nearly all of the analytes in Tables 1 and ​2, observed differences were less than maximum desirable differences. Although differences observed for several tests exceeded the criteria for acceptability and the observed differences were statistically significant, the heparinized plasma sample may be preferred, because the serum sample has been interfered with by the coagulation process (10). Glucose is the exception as explained below.

Some calcium is consumed by the clotting process; thus, the concentration of calcium in plasma is slightly higher than that in serum. The difference observed in this study of 0.13 mg/dl is of the order of +0.1 mg/dl, the decimal place to which calcium results are usually reported. The 95% confidence interval of the average difference is 0.03 to 0.16 mg/dl, which includes the 0.07 mg/dl criterion for acceptability; thus, it is uncertain whether the difference is statistically or clinically significant. In an unpublished study submitted to the CMS that was performed by eight large laboratories serving approximately four of five patients with ESRD in 2008, the absolute average difference (by laboratory) from the grand mean of the group was 0.09 mg/dl, which is less than the plasma versus serum difference that we observed. In that same study, the maximum difference from the group mean by any one laboratory was 0.36 mg/dl, which exceeds most reported plasma versus serum differences (2–4,7). Of note, consistent with the WHO recommendations, the plasma calcium concentration may actually better represent the in vivo condition of the patient.

Phosphorus is released during the clotting process; thus, the concentration of phosphorus is higher in serum than in plasma. The difference observed in this study was insignificant (−0.01 mg/dl), but differences as large as −0.2 mg/dl have been reported (3). The histogram shown in Figure 1A and the scatterplot shown in Figure 1B show the agreement between the individual results from the two sample types. In the 2008 study cited above, the mean absolute difference from the grand mean of the eight laboratories was 0.09 mg/dl, and the maximum difference by any one laboratory was 0.20 mg/dl. Again, the interlaboratory differences are similar to if not greater than the differences caused by sample type.

The 0.20 g/dl difference between plasma and serum results for total protein is caused by fibrinogen in plasma, and it is similar to previously reported differences (3,4,7). This gap is often not clinically meaningful.

The difference for sodium (−0.4 mmol/L) is <1 mmol/L, the decimal place to which sodium is usually reported. As with calcium, the 95% confidence interval of the average difference (−0.6 to −0.2 mmol/L) contains the 0.3-mmol/L criterion for acceptability. Reported plasma − serum differences for sodium ranging from −2.6 to 1.3 mmol/L have been considered to be negligible (2–4,6). The 2008 study did not include sodium.

Heparin activates AST. Plasma is stated to be the preferred sample for AST measurement in the WHO document. Serum and plasma results should not be used interchangeably (5). However, using either one test sample type singularly with attention to trends is our recommended approach.

Finally, glucose is consumed by platelets in plasma, causing a negative bias versus serum. The short time between sample collection and analysis minimizes this effect in the hospital laboratory. In the internal testing that we did on samples transported overnight, individual sample biases ranged from −30 to +7 mg/dl and were independent of concentration. The magnitude of this negative bias may be sufficient to affect interpretation, and therefore, results should be reviewed with caution.

In summary, for most chemistry tests performed on patients on dialysis, heparin plasma samples collected in plasma separator tubes can be used in lieu of serum samples collected in serum separator tubes without clinically significant effect. Three notable exceptions are intact PTH, where EDTA plasma is recommended; AST, where it is recommended that either plasma or serum results be used for trending; and interpretation of plasma glucose, which tended to have a negative bias (i.e., lower values than from serum samples), where clinical indications may require serum measurements or other methods of glucose-level monitoring. From a nephrologist’s perspective, with the caveat of the exceptions noted above, the results for common chemistry analytes from serum and heparin plasma may be used interchangeably in the clinic—with the same understanding that, although within-laboratory variability often is within desirable biases, one still needs to consider the potential influence of interlaboratory variability when comparing test results from multiple sources.

When results from blood tests are also used for regulatory purposes—particularly for quality improvement and specifically when linked to reimbursement, such as within the ESRD QIP—an important consideration includes determination of the necessity for limiting the type of assay or sample used for testing. For example, although restriction to serum samples versus heparin plasma samples is not necessary for phosphorus, considerations for albumin measurement may include expert deliberations on whether restrictions on the methodology used for measurement (e.g., bromcresol green versus other methods) is appropriate. The drive toward a common electronic health record that spans across theaters of care requires recognition of the comparability of many aspects of care, including laboratory testing. In addition to potential clinical implications for individual patient care, population-based initiatives as well as health care facility–based quality and safety initiatives could also be affected by comparability issues. Hence, in the development of value–based health care system, where ESRD care seems to be at the forefront (12), a broad–based collaborative consultation process involving multispecialty teams that occasionally may require clinical laboratory experts will be helpful moving forward.

Disclosures

All authors are employees of Spectra Laboratories except for E.L., who was an employee of Fresenius Medical Care at the time that the manuscript was written. Spectra Laboratories is a division of Fresenius Medical Care.

Acknowledgments

The authors acknowledge the technical assistance of Ruth Paje, Ruby Kaamino, and Rowena Suizo, employees of Spectra Laboratories at the time of this work. No other financial support was received.

Footnotes

Published online ahead of print. Publication date available at www.cjasn.org.

References

1. Lutz J, Menke J, Sollinger D, Schinzel H, Thürmel K: Haemostasis in chronic kidney disease. Nephrol Dial Transplant 29: 29–40, 2014 [PubMed] [Google Scholar]

2. Lum G, Gambino SR: A comparison of serum versus heparinized plasma for routine chemistry tests. Am J Clin Pathol 61: 108–113, 1974 [PubMed] [Google Scholar]

3. Doumas BT, Hause LL, Simuncak DM, Breitenfeld D: Differences between values for plasma and serum in tests performed in the Ektachem 700 XR Analyzer, and evaluation of “plasma separator tubes (PST).” Clin Chem 35: 151–153, 1989 [PubMed] [Google Scholar]

4. Miles RR, Roberts RF, Putnam AR, Roberts WL: Comparison of serum and heparinized plasma samples for measurement of chemistry analytes. Clin Chem 50: 1704–1706, 2004 [PubMed] [Google Scholar]

5. Wei Y, Zhang C, Yang X, Ji M: The feasibility of using lithium-heparin plasma from a gel separator tube as a substitute for serum in clinical biochemical tests. Lab Med 41: 215–219, 2010 [Google Scholar]

6. Er TK, Tsai LY, Jong YJ, Chen BH: Selected analyte values in serum versus heparinized plasma using the SYNCHRON LX PRO assay methods/instrument. Lab Med 37: 731–732, 2006 [Google Scholar]

7. Ciuti R, Rinaldi G: Serum and plasma compared for use in 19 common chemical tests performed in the Hitachi 737 analyzer. Clin Chem 35: 1562–1563, 1989 [PubMed] [Google Scholar]

8. Centers for Medicare & Medicaid Services (CMS), HHS: Medicare program; end-stage renal disease prospective payment system, quality incentive program, and durable medical equipment, prosthetics, orthotics, and supplies. Fed Regist 78: 72155–72253, 2013 [PubMed] [Google Scholar]

9. Centers for Medicare & Medicaid Services (CMS), HHS: Medicare program; End-Stage Renal Disease prospective payment system, quality incentive program, and Durable Medical Equipment, Prosthetics, Orthotics, and Supplies. Final rule. Fed Regist 79: 66119–66265, 2014 [PubMed] [Google Scholar]

10. Banfi G, Bauer K, Brand W, Buchberger M, Deom A, Ehret W, Engel WD, da Fonseca-Wollheim F, Fraser CG, Friemert VJ, Golf S, Gross H, Gruder WG, Gunzer G, Hagemann P, Heil W, Henny J, Hinzmann R, Hyltoft Persen P, Hoffman G, Kallner A, Karallus A, Kitta H, Klahr D, Kolpe D, Kukuk J, Kunert-Latus T, Lammers M, Leppänen EA, Mikulcik P, Narayanan S, Neumaier M, Peça Amaral Gomes MA, Probst R, Schmitt Y, Sonntag O, Töpfer G, Weisheit R, Wisser H, Zawta B, Zinck R: Use of Anticoagulants in Diagnostic Laboratory Investigations and Stability of Blood, Plasma and Serum Samples. WHO/DIL/LAB/99.1 Rev 2, Geneva, Switzerland, World Health Organization, 2002 [Google Scholar]

11. Ricós C, Alvarez V, Cava F, García-Lario JV, Hernández A, Jiménez CV, Minchinela J, Perich C, Simón M: Current databases on biologic variation: Pros, cons and progress. Scand J Clin Lab Invest 59: 491–500, 1999 [PubMed] [Google Scholar]

12. VanLare JM, Conway PH: Value-based purchasing--national programs to move from volume to value. N Engl J Med 367: 292–295, 2012 [PubMed] [Google Scholar]

Articles from Clinical Journal of the American Society of Nephrology : CJASN are provided here courtesy of American Society of Nephrology