Comparison of test results obtained from lithium heparin gel tubes and serum gel tubes

Introduction

An estimated 70% of medical decisions are based on the results from laboratory testing [1]. Rapid turnaround of laboratory tests is therefore very important to achieve early diagnosis and treatment as well as to decrease emergency room or hospital in-patient services length of stay. However, a Q-Probes study [2] carried out in 2016 demonstrated that test turnaround times have not still satisfy the expectations of physicians. In this context, laboratory professionals frequently face the issues regarding turnaround time.

Serum is most often used sample type in testing clinical chemistry analytes. To obtain serum, blood samples should be waited at least 30 min for completely clotting before centrifugation [3], which may contribute the prolongation of the turnaround time of test results. However, clotting is not required for anticoagulated blood samples, hence specimens can be processed more quickly, shortening the turnaround time. In addition to this time-saving benefit, plasma has also some other advantages to serum including prevention of interference induced by fibrin clot on automated analyzer, shorter centrifugation duration and higher sample volume yield from a given volume of whole blood. Most significantly, a World Health Organization document [4] states that the constituents in plasma are more accurately reflecting the pathological situation of a patient than those in serum. There is nowadays a trend use of the plasma in place of the serum in clinical biochemistry [5].

In patients with end stage renal disorder treated by hemodialysis, clotting may not completely occur even if specimens for serum is hold on before centrifugation as recommended time, which cause various issues from instrumental errors to erroneous test results [6]. Incomplete clotting is due to anticoagulant therapy and uremia itself [7]. It is thought that the use of plasma instead of serum may be help overcoming the challenges related to blood clotting time in hemodialysis patients as well as in the patients receiving anticoagulant therapy [6].

In terms of standardization of laboratory, the use of only a sample type (serum or plasma) for measuring an analyte is desired, but this may not always practical [8]. For instance, if an analyte could be measured in both serum and plasma, plasma samples may be preferred in patients admitted emergency department, otherwise serum. However, it is thought that differences in sample matrix of serum and plasma can limits the use of the results obtained from two sample types as interchangeably [8].

In this study, it was aimed to investigate the differences between the results obtained from lithium heparin gel tube and serum gel tube for 28 clinical chemistry tests in healthy individuals and patients treated by hemodialysis.

Materials and methods

Results

The HI values were below than 50 in all samples, which no hemolysis interference for any test. The mean of results obtained from serum and plasma samples, the percent bias and the TE between serum and plasma results, the statistical significance and the TEa limits are shown in Table 1 for each analyte.

In the paired statistical analysis, it was found that there was significant difference between the results obtained from serum and plasma for ALP, direct bilirubin, calcium, CK, HDL-cholesterol, LDL-cholesterol, total cholesterol, GGT, glucose, iron, phosphorus, potassium, total protein, and uric acid (p<0.05).

The TE between two sample type was <5% for amylase, AST, calcium, total cholesterol, chloride, CK, glucose, HDL-cholesterol, iron, LDH, LDL-cholesterol, magnesium, sodium, total bilirubin, uric acid and urea; <7% for albumin, ALT, creatinine, CRP, lipase, phosphorus, potassium, total protein and triglyceride; <10% for ALP, direct bilirubin and GGT. For all analytes, the TE values were found to be within the TEa limits (Table 1).

Similarly, the TE values between the results from serum and plasma from patients treated by hemodialysis were determined to be inside the TEa limits for all analytes (Table 1).

The regression equation, slope and intercept values from Passing-Bablok regression analysis are shown Table 2. The regression plots are also presented in Figure 1. For all analytes, the 95% confidence intervals for the slope were found to be included the value of 1. It was therefore concluded that there was no proportional difference between the results from two sample types. Similarly, for all analytes except CRP, it was determined that the 95% confidence intervals for the intercept have contained the value of 0. Therefore, it was concluded that there was no proportional difference between the results from serum and plasma, except CRP.

Figure 1:

Passing-Bablok regression plots comparing test results obtained from serum and heparinized plasma. The straight line indicates the regression line, dashed lines around the solid line shows the confidence interval for the regression line.

Table 2 includes also correlation coefficients between the results obtained from serum and plasma. For all analytes, it was found that there was significant correlation between the two sample types.

Discussion

Nowadays, there is a trend that plasma might be alternative to serum due to some of its advantages. However, before introduced the use of plasma, it is suggested that plasma tube should be evaluated since tubes from different manufacturer might provide different laboratory results for clinical chemistry [15]. In this context, this is the first study providing data regarding the relation between heparinized plasma and serum samples by using the lithium heparin tubes with gel separator from Greiner Bio-one. In the present study, for 14 of 28 clinical chemistry tests analyzed, statistically significant differences between serum and heparinized plasma were observed, but not clinically significance.

Similar findings have been reported in some earlier studies, which investigated the agreement between the results from serum and heparinized plasma. In a study [16] compared the results obtained from serum and heparinized plasma for 16 analytes, Er et al. have found statistically significant but medically not differences between two samples type for potassium, CK and LDH. Clinically significant differences have been scarcely determined for only albumin and blood urea nitrogen, but have been ignored.

In another study [17], O’Keane and Cunningham have compared renal and lipid analyte profiles in plain serum tube, serum tube with gel and lithium heparin plasma tubes without gel. In their study, it has been determined that statistically significant difference between the results from the tube types was exist for most of the analytes, but clinical significance for only potassium. They have compared the values in plain serum tube and plasma tubes to identify clinical significance. However, in their study, when compared the results from plasma tube with those of serum tube with gel separator, it can be detected that there was no clinical significant difference for all tests measured.

In addition, in a study [18] which set out to determine reliability of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) method for LDH measurement in heparin plasma samples, Herzum et al. have noticed that LDH measurements is suitable in heparin plasma samples obtained from Sarstedt Monovette tubes. On the contrary, Bakker et al. [19] have reported that IFCC method for LDH measurements does not provide reliable results in heparin plasma samples due to significantly differences in duplicate measurements. However, in later a report [20], Bakker have declared that the duplicate errors arise from heparin plasma tube used in their study.

On others hand, in contrary to the findings from the current study, there are reports that have declared some differences between the results from serum and lithium heparin plasma for certain clinical chemistry analytes.

In a recent study [5], Arslan et al. have compared the values obtained from lithium heparin tube with mechanic separator and glass tube without additive for 22 analytes. They have found clinically significant difference between serum and plasma samples for AST, sodium, potassium, LDH, glucose and total protein.

In another study carried out by Wei et al. [21], the values of plasma samples in vacuum tubes with gel separator were compared with those of serum for 31 biochemical tests. For AST, ALP, LDH, potassium, phosphorus, glucose, and total protein, it has been suggested that plasma reference ranges should be established.

In a third study [22], Miles et al. have compared the results from serum and lithium heparinized plasma samples for 45 clinical chemistry analytes that measured on different two analyzers (Roche Modular P and Vitros 950). Clinically significant differences between serum and plasma samples have been identified on both analyzers for potassium and LDH. Moreover, in their study, clinically significant variations have been seen for aldolase, bile acids and angiotensin-converting enzyme measured on Roche analyzer, as well as total bilirubin analyzed on Vitros 950 analyzer.

These different findings from different studies can be explained in part by the variation in procedures used for the data analysis. The comparisons of results have only been statistically done in a study [21]. However, in a guideline [9] published by CLSI, to determine the relationship between measurements from different sample types by using patient samples, it has been recommended that laboratory professionals should be estimate bias and then compare it with the limits of acceptable bias. In this context, Arslan et al. have estimated bias values and compared with the allowable bias limits based on biological variation to identify clinically significant in their study [5]. However, for some tests, the limits of acceptable bias are very tight and impossible to achieve. For instance, the allowable bias limit based on biological variation for sodium and potassium is 0.23 and 1.81%, respectively [14]. Therefore, in the current study, TE was determined instead of bias and compared with TEa limits. The conclusion of a study might be differ related to the using of bias or TE in determination of clinically significant difference.

The TEa limits could be obtained from multiple sources including biological variation, external quality control program and local legal regulations [23]. The TEa limits from different references might be different each other. For some tests, the difference between TEa limits from multiple sources might be great enough to alter the conclusion of a study. For instance, for potassium, by using the TEa limits of 9% that defined by Republic of Turkey Ministry of Health [12], no clinically significant the difference between the results from serum and plasma was found in the present study. On other hand, if the TEa limits of 5.61% based on biological variation [14] had been selected, a different finding would have been obtained. Similarly, the TEa limits for total protein are suggested 15 and 3.63% by Republic of Turkey Ministry of Health and Ricos et al. [14] (based on biological variation), respectively.

Moreover, in combination with the variation in data analysis, the difference in the properties of analyzer used including cleaning of reaction cells sample and reagent volumes, or timing could also contribute to different findings from the studies.

In case the use of serum for clinical chemistry tests in hemodialysis patients, even if the recommends related to pre-centrifugation are followed, several issues regarding the prolonged clotting might be frequently encountered. In the present study, therefore, it was also evaluated the relationship between results from plasma and serum in hemodialysis patients and found that there was no clinically significant difference between the results from serum and plasma samples.

There are limited studies on comparing the results obtained from two sample types in patients treated by hemodialysis. In a commentary paper [6], Carey et al. have summarized their renal community experience regarding the acceptability of heparin plasma samples for 24 analytes in patients on hemodialysis. Authors have suggested that heparin plasma samples collected in tubes containing separator gels can be replaced serum samples for most chemistry tests without clinically significant effect, except parathormon, AST and glucose. In their study, the specimens have been measured after overnight transportation, which might lead to additional variation between results from serum and plasma, especially for glucose.

In another study [24], Meng and Krahn have shown that lithium heparin plasma can cause falsely low albumin values in hemodialysis patients when used a bromcresol green (BCG) method. Authors have proposed that the heparin contained in the collection tubes inhibits the bindings of BCG to albumin and thus lead to less colorimetric complex formation. On the contrary, it was found that there was no statistically and clinically significant difference for albumin values by BCG method in the current study. Although albumin reagent used in both studies have been produced by same manufacturer, the analytical platform used and calibrators used as well as the brand of tubes used were different between two studies. The combination of these variations might be reason of different findings.

This study has some limitations. First, serum and plasma samples were obtained from the tubes produced by one manufacturer. Also, testing was performed on only one analytical platform. It is therefore not known if the findings from the combination of the brand of tube used and the analyzer used in the present study is valid in other the corresponding combinations. Another limitation is that the effect of different centrifugation speed and duration, as well as of different blood collection technics, on the difference between the results from serum and plasma was not evaluated. Lastly, though intended to provide a wide concentration range for the analytes measured by including the specimens from hemodialysis patients to the study, this was not achieved for all of tests.

In conclusion, for 28 clinical chemistry test measured in the present study, the results from plasma samples from lithium heparin tubes with gel separator was found to be comparable to those of serum gel tubes, at least when performed the tubes and analyzer used in the study. It is thought that the use of heparinized plasma could reduce the time between blood collection and analysis, as well as eliminate the issues related clotting.