Reducing the effects of control materials based on interchangeability of estimates of day‐to‐day imprecision between commercial control materials and serum samples

Abstract Background Reduce the effects in the storage‐and‐thawing process of commercial control materials based on their interchangeability evaluation. Methods Seven assays—anti‐streptolysin O, complement 3, carcinoembryonic antigen, urea, ferritin, total bilirubin, and glucose—were selected. Commercial control materials and serum samples with similar concentrations were chosen as samples. The experiment was carried out in three stages. In the first stage, the assays with statistical differences in imprecision were screened. In the second stage, two specimens were sealed with parafilm and frozen at −80°C and thawed in the water bath, and the imprecision differences were compared again. Finally, the effective means to reduce the effects were included in the standard operating procedure to repeat confirmation. Results In the first stage, there was only a statistical difference (p < 0.05) in the imprecision of glucose and total bilirubin between two specimens, and the imprecision of control materials was higher than the serum samples. In the second stage, glucose imprecision was not statistically different (p > 0.05) and lower than in the first stage. In the third stage, the methods from the second stage were confirmed to be effective at reducing control material effects. Conclusion Finding variation factors and confirming and standardizing the measures will help lessen commercial control material effects.

control it within a reasonable range, in order to ensure that the patient results truly reflect the patient's status.
However, the imprecision detected in the internal quality control often includes the analytical variation of the detection system and the variation of the control material. The control material variation is equivalent to the interference signal, and the larger the proportion, the more difficult it is to accurately detect the analytic variation.
Only by reducing the control material variation as much as possible can the detection signal be amplified, so that the QC results can more truly reflect the analytic variation of the detection system and the control material can play a real role.
Many years ago, several authors asserted that there may be a lack of interchangeability between commercial control materials and serum samples regarding day-to-day imprecision. 3,4 These differences may come from the control material itself (ie, the matrix effect) or from variations in the control material processing (including storage and reconstitution) which does not exist in the operation process of serum samples. Once the control material is selected, the matrix effect cannot be amended. Therefore, the interchangeability of day-to-day imprecision for commercial control materials and serum samples as the standard is extremely important for evaluating how to reduce variations in control material processing. This standard can fundamentally evaluate whether the control material detection conforms to the specification; only when the imprecision between control materials and serum samples was interchangeable can all internal quality-control behaviors be considered effective for serum samples. To our knowledge, however, no study has used this criterion to evaluate how to reduce the difference in imprecision between control materials and serum samples; and no study has added the improvements to the laboratory standard operating procedure to confirm whether the improvement could be repeated.
In addition, if the noninterchangeability were found among control materials from different manufacturers or, worse still, among different lots of the same control material, monitoring day-today imprecision during long periods also would be very difficult. 4 Consequently, it is very important to strive to make the control materials have the imprecision interchangeable with the human samples.
In the present study, we compared the imprecision between commercial control materials and serum samples of seven assays of anti-streptolysin O (ASO), complement 3 (C3), carcinoembryonic antigen (CEA), urea (UREA), ferritin (FER), total bilirubin (TBIL), and glucose (GLU), and we tried to reduce the effects of commercial control materials by referring to the experience of reference-measurement research and discussing the feasibility of the method in the laboratory.

| Materials
Commercial control materials were purchased from Cliniqa Corp.

| Stage 1
Screening assays with a statistical difference in imprecision between commercial control materials and serum samples.

Sub-package and storage
According to the manufacturer's specifications, the Liquid QC ImmuTROL Serum Protein Control for the ASO and C3 assays was not aliquoted and stored at 2-8°C until measurement; the other control materials were reconstituted previously; then, each reconstituted control material and each serum sample were divided into 20 aliquots and stored at −20°C away from light until analysis.

Thawing
Each vial of Liquid QC ImmuTROL Serum Protein Control was mixed upside-down eight times before sampling to ensure homogeneity; then, the cap was immediately replaced and it was stored at 2-8°C.
The samples were sealed and left at room temperature (25 ± 5°C) for 15 minutes. Each vial of samples that was frozen at −20°C was thawed at room temperature (25 ± 5°C) for 15 minutes. All samples were thoroughly mixed with pipettes and measured within 10 minutes.

Analysis
The control materials were analyzed first. After each assay was in control, one measurement of each analyte was carried out in each of the serum samples within 2 hours by the same analyst. The specimens were analyzed for 20 consecutive days. When 20 replicated results for each analyte were obtained, the corresponding variances and coefficients of variation (CVs) representing imprecision were estimated. The imprecision for each assay between the control materials and the serum samples was compared, and the assays with statistical differences were selected for the second phase of the experiment.

| Stage 2
Re-comparing after improving the operational procedures of assays with differences in Stage 1.
Two serum samples were re-collected and pooled into a plain tube. After thoroughly mixing, each pool was aliquoted into 0.5-ml Eppendorf tubes. Then, each pool was composed of 40 aliquots, and the samples were randomly divided into two experimental groups with 20 aliquots in each group. The samples of Group 1 were stored, thawed, and measured as the first phase of the experiment. The samples of Group 2 were sealed with parafilm and stored at −80°C away from light until analysis. Thirty minutes before analysis, one aliquot of the samples of Group 2 was removed from −80°C, thawed in the water bath (25 ± 2°C) away from light for 10 minutes, mixed gently upside-down five times, left at room temperature (25 ± 5°C) away from light for 15 minutes. It was mixed gently upside-down again for five times and then measured within 10 minutes. 5-9 Commercial control materials were processed as serum samples. After each assay was in control each day, over the course of 20 working days, one measurement of each analyte was carried out in each of the control materials and serum samples simultaneously by the same analyst.

| Stage 3
Verify that the operational improvements in Stage 2 are reproducible.
Only control materials were analyzed, and the analyses were expanded to 21 analytes, and then, the difference in imprecision between the two specimen-processing methods for 20 days was compared.

| Statistical analysis
The mean, standard deviation (SD), and CV for each assay were calculated to compare the imprecision. When the ratio of mean-to-SD was less than 3, the SDs of the replicate analyses were compared by the Ftest, where F = (SD 1 ) 2 /(SD 2 ) 2 and SD 1 > SD 2 . Otherwise, the CVs were compared by a modification of the F-test, which has been designated the H-test, where H = (CV 1 ) 2 /(CV 2 ) 2 and CV 1 > CV 2 . The F Bilateral Boundary imprecision was compared with desirable analytical-quality specifications for imprecision upon biological variation. 10 The biological variation data preferentially used the latest data from the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM). 11

| RE SULTS
Each pair of variances was compared by the H-test, because the ratios of the means-to-SDs were greater than 3.
Results of the comparisons in the first stage are shown in Table 1.
There is no specification for imprecision for ASO, because it does not have biological variation data. Except C3, other assays met the desirable specifications for imprecision. There was only a statistical difference in the imprecision of GLU (Level 1) and TBIL (Level 2) between commercial control materials and serum samples. Figure 1 shows the arrangement of serum sample data and QC data of GLU (Level 1) and TBIL (Level 2) within 20 days. Each group of data fluctuated above and below the respective mean, and there was no obvious trend change.  In the third stage, the comparison results of the imprecision of two different processing methods for 21 assays are shown in Table 3. Among the 42 concentration levels of the 21 assays, 35 concentration levels showed that the control material imprecision under "sealed with parafilm, frozen at −80°C, and thawed in water bath" was less than under "stored at −20°C and thawed at room temperature"; of these, the differences were statistically significant in 10 concentration levels of eight assays, and the differences were all statistically significant in two concentration levels for creatinine (CREA) and lactate dehydrogenate 1 (LDH 1). In addition, among the assays with statistical difference in the above comparison, only under the condition, "stored at −20°C and thawed at room temperature," the analyses of LDH 1 (two concentration levels) and HDL cholesterol (Level 2) did not meet the desirable analytical-quality specifications for imprecision upon biological variation. There is no specification for imprecision for α-hydroxybutyrate dehydrogenase, because it does not have biological variation data.

| DISCUSS ION
The previous studies have mainly focused on the stability of serum samples and control materials or the interchangeability of day-today imprecision for them. [3][4][5][6][7][8][9] In this study, we screened out assays with differences in imprecision between the commercial control  materials and the serum samples, and made improvements based on the performance of these assays. After confirming the effect, the improvements were added to the laboratory standard operating procedure to confirm whether the improvements could be repeated.
In the first stage of this study, the imprecision between two specimen types was compared by using the daily operating procedures in our laboratory, of which the differences of GLU (Level 1) and TBIL (Level 2) were statistically significant, and the SDs obtained by analyses of commercial control materials were 1.9 times and 2.9 times that of serum samples, respectively. Although their impreci- This shows that there are many factors that affect quality-control efficiency, including the control materials themselves, operators, operating methods. The best way to improve quality-control efficiency may be to strictly control the operating process to reduce the variation in control material processing.
In addition to the matrix effect, the likely causes of the statistical difference in the imprecision between the commercial control materials and the serum samples include 3,4,6 (1) variations in the preparation and reconstitution of the control material (ie, the variations between bottles); (2) differences in the stability of the two samples, which affected factors that included sample moisture evaporation, storage temperature, the freeze-thaw process; (3) insufficient sample mixing.
In the second stage, we selected several of the above factors for standardized operation: low temperature, anti-evaporation, and gradient thawing (ie, the thawing method of Group 2 in Stage 2). Parafilm seal can prevent sample moisture evaporation, while low temperature and gradient thawing can reduce analyte damage during the preservation and thawing process; that is, the influence of "cold denaturation." The results showed that there was no statistical difference in the imprecision between the commercial control materials and the serum samples when the operation process was strictly controlled, which indicates that this method could effectively reduce the imprecision difference between two specimen types. At the same time, the results also showed that the uncontrolled specimens are more imprecise than strictly controlled specimens, and the measured data showed a decreasing trend, which was mainly affected by the so-called "cold denaturation." This result was consistent with the literature reports. [5][6][7][8]12 In order to confirm whether the operation can be extended to other assays, we incorporated the operation in the conclusion into the laboratory standard operating procedure and applied it to 21 routine assays. The results showed that there was a statistically significant decrease in the imprecision of eight assays, and the decrease in the imprecision of enzymes and micromole-level analytes was more obvious, and the imprecision of LDH 1 and HDL cholesterol (Level 2), which originally did not meet the specification, met the standard, which indicates the extendibility of the operation.
The limitation of this paper is that in the second stage, due to the difficulty of collecting serum samples, the comparison experiment of only the glucose analyses was carried out. In addition, in the control material operation, other standardized operations used in previous studies, such as adding samples with dilution dispenser and using water with different conductivity (whether <1 μs/cm), were not included because of the limitations of the experimental conditions.
Although the main result found in this study, "some measurands, especially glucose, are unstable if stored at −20°C instead of −80°C," it is well known, but because of the cost, customary and convenience, the laboratories generally keep the control materials at −20°C (recommended by the manufacturer) instead of −80°C. This paper explains its necessity from the perspective of improving the interchangeability of day-to-day imprecision for control materials and serum samples.
In summary, by comparing the analytical imprecision between control materials and serum samples, we can select a control material that has the imprecision interchangeable with the patient sample as F I G U R E 3 Z-value control chart of GLU and TBIL based on the mean of the control material and the SD of the serum sample. GLU, glucose; TBIL, total bilirubin much as possible. When selecting, attention should be paid to those assays with great coefficients of variation and poor interchangeability. If there are still assays with poor interchangeability with patient samples in the selected control materials, the method of strictly controlling the operation process in this study can be adopted to reduce the effects of control materials, so that the imprecision across control materials and patient serum samples can be interchangeable.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.