Simultaneous point‐of‐care testing of blood lipid profile and glucose: Performance evaluation of the GCare Lipid Analyzer

Abstract Background Point‐of‐care (POC) testing provides quick results and includes tests for blood glucose and lipid profiles. We evaluated the newly developed POC device, the GCare Lipid Analyzer, which is used to measure glucose, total cholesterol (TC), triglyceride (TG), and high‐density lipoprotein cholesterol (HDL‐C) levels. Methods Venous and capillary blood samples were collected from patients who visited Korea University Guro Hospital. The results obtained using the GCare Lipid Analyzer were compared with those obtained using the TBA 2000FR chemistry analyzer and YSI 2300 STAT Plus analyzer. The glucose system evaluation process was based on the International Organization for Standardization 15197:2013 guidelines. Results The correlation coefficients (R) for TC, TG, and HDL‐C were 0.965, 0.969, and 0.943 in capillary blood and 0.969, 0.990, and 0.956 in venous blood, respectively. The total errors for TC, TG, and HDL‐C of the lipid profile using venous blood were all acceptable at 6.6%, 9.3%, and 11.6%, respectively. For glucose concentrations <100 mg/dl, 96.1% of the measured glucose levels were within ±15 mg/dl in venous samples and 100% were within ±15 mg/dl in capillary samples. For glucose concentrations ≥100 mg/dl, 100% and 99.5% of the measured glucose levels were within 15% for venous and capillary blood, respectively. Conclusion The performance of the GCare Lipid Analyzer is acceptable for both blood glucose and lipid profile testing, indicating that it is reliable for use in patients with diabetic dyslipidemia.

The decision to start treatment for dyslipidemia is based on the analysis of lipid fractions including TC, HDL-C, LDL-C, and non-HDL-C. 8 This baseline lipid evaluation is performed periodically after initiating pharmacological interventions such as statins. Patients are required to visit the hospital laboratory to provide blood samples for lipid profile testing before their outpatient clinic visit. The use of point-of-care (POC) devices in this situation is expected to alleviate this inconvenience. Several hand-held portable POC devices are currently available to measure lipid and glucose levels in the blood, such as the Accutrend Plus, Bene Check Plus, CardioChek PA, Veri-Q, 3 in 1, and elemark™, as well as the compact desktop analyzer, the Cholestech LDX ® . These devices can measure lipid profiles and ratios in whole blood, plasma, or serum using reflectance or biosensor technology and feature disposable strips, rotors, or cassettes. 9 However, most patients with diabetes are unfamiliar with systems for self-monitoring lipid profiles, unlike commonly used self-monitoring blood glucose devices.
Here, we present a new laboratory information system developed in South Korea, a connectable hand-held glucose and lipid (TC, TG, HDL-C, calculated LDL-C) monitoring system named the GCare Lipid Analyzer (Green Cross Medical Science, Yongin, Korea).
Considering that this is the first study on the performance of this POC device, we evaluated the glucometer's performance in accordance with the International Organization for Standardization (ISO) 15197:2013 guidelines and assessed lipid measurements for precision, accuracy, and correlation with values obtained using a TBA2000FR chemistry analyzer.

| Subjects
We recruited two groups of adult volunteers (age range, 19-80 years) among patients who visited Korea University Guro Hospital for their prescribed blood tests. For the evaluation of the lipid panel, an additional 2 ml EDTA venous blood and 150 µL of capillary blood were collected from the study subjects (n = 136). Specimens in which the hematocrit (Hct) range did not fall within 25%-60% were excluded. An additional 10 ml of venous blood (into an EDTA tube) and 150.5 µl of capillary blood were collected from the group (n = 100) and tested for blood glucose concentration, complete blood count, and blood type. Specimens that did not meet the Hct range of 15%-65% were excluded from the evaluation.

| GCare Lipid Analyzer
GCare Lipid Analyzer (Green Cross Medical Science) is a compact, hand-held combined lipid and glucose POC device. Both capillary and venous blood samples are available for testing. The device is similar to most other blood glucometers, except that it features two slots that enable users to easily select tests according to their needs. The inlet at the bottom accommodates the lipid test strip, and the smaller inlet on the left side is for the blood glucose test strip ( Figure 1). The GCare Lipid Profile Test Strip was used, which incorporated an enzymatic colorimetric method, to measure TC, TG, and HDL-C. The LDL-C value is calculated from the TC, TG, and HDL-C values measured by the device, according to the Friedewald formula (TC -HDL-C -TG/5) when the TG value is <350 mg/dl. The measurement time for the lipid profiles is 180 s, and the required sample volume is 40 µl. Each measured lipid parameter can be checked on the screen of the device by clicking the arrow button on the device serially. For the glucose level testing, the GCare Glucose Test Strip is used which is based on the glucose dehydrogenase flavin adenine dinucleotide system. The sample volume required for measurement is 0.5 µl and the measuring time is 5 s.

| Study design for the GCare Lipid Analyzer using GCare Lipid Profile Test Strip
For the precision evaluation of the lipid panel (TC, TG, and HDL-C), three concentrations (low, medium, and high) of venous blood were measured 10 times using three lots according to the Clinical  and Laboratory Standards Institute (CLSI) EP5-A3. 10 Analytical precision was assessed by calculating each measured value per lot as mean, standard deviation (SD), and coefficient of variation (CV, %).
Standardization is important for the measurement of lipoproteins because treatment decision points have been established by the expert consensus of the National Cholesterol Education Program (NCEP). 11 The NCEP states that the analytical performance criteria in terms of total error (%) for TC, TG, and HDL-C are ≤8.9%, ≤15%, and ≤13%, respectively, while the certification criteria for imprecision (CV, %) of TC, TG, and HDL-C are ≤3%, ≤5%, and ≤4%, respectively. [12][13][14][15] Regarding the standardization criteria of the Centers for Disease and CV are the same as those prescribed by the NCEP.
The measurement procedure comparison study was conducted according to CLSI document EP9. 16 Each measurement was tested in duplicate and was compared with the reference value obtained

| Study design for the GCare Lipid Analyzer using GCare Glucose Test Strip
The blood glucose meter evaluations were performed in accordance with the ISO15197:2013 guidelines. 18

| Statistical analysis
Data recording and processing were performed using Microsoft

| Precision profile of the GCare Lipid Analyzer
The precision profiles of the GCare Lipid Analyzer are presented in Table S1. The NCEP-recommended performance criteria for precision were met in three lots and at low to high TG concentrations (≤5%). TC showed an acceptable CV (≤3%) only at high concentrations in lots 1 and 3, while other CV values did not exceed 5%. The HDL-C measurements meet the performance criteria (CV ≤4%) at low and high level in lot 2 and at high level in lot 3, while lot 1 failed to meet the criteria in all levels.
The grand average, pooled variance, pooled SD, and pooled CV for each glucose concentration were calculated using the results from all three reagent lots. Pooled SD for levels 1 and 2 were 2.9 mg/ dl and 3.0 mg/dl, respectively. Pooled CV values for levels 3, 4, and 5 were 3.0%, 2.6%, and 1.9%, respectively.

| Comparison of measurements of the GCare Lipid Analyzer and laboratory measurements including user performance evaluation and sample type comparison
The measurements of TC, TG, and HDL-C in the GCare Lipid Analyzer were compared with those of the Toshiba laboratory method as a reference. Bias was estimated for the GCare Lipid Analyzer and the laboratory method using paired results. The mean bias values for TC, TG, and HDL-C were 0.560, 3.391, and 1.121, respectively, in capillary blood and −0.527, 2.385, and 0.221 in venous blood, respectively. The greatest bias was observed for TG in both types of specimens. The Bland-Altman plots for each lipid test in capillary and venous blood, including the 95% limit of agreement, are displayed in Figure 2.
The Passing-Bablok regression analysis revealed a good to excellent correlation between the GCare Lipid Analyzer and the laboratory method (Figure 3, Table S2). The correlation coefficients (R) for TC, TG, and HDL-C were 0.965, 0.968, and 0.943 in capillary blood, respectively, while the R values for venous blood were 0.969, 0.990, and 0.956, respectively. All of the R values were >0.950, reflecting excellent correlations, except that for HDL-C in the capillary blood specimen, which was 0.943.
The correlation between the results of the capillary blood from the user and those acquired through the laboratory method was acceptable at R = 0.957 for TC, 0.991 for TG, and 0.936 for HDL-C. The mean bias was also within the criteria suggested in the NCEP guidelines:

| Analytical performance of GCare Lipid Analyzer according to NCEP recommendations
Combining the lowest to highest CV as the overall analytical imprecision (CVa, %) and the mean bias calculated above, the total error was calculated for each test analyte ( Table 2). The total error for the venous blood specimens was calculated as bias +1.96 × CVa. 19 The total errors for TC, TG, and HDL-C of the lipid system using venous blood were all acceptable at 6.6%, 9.3%, and 11.6%, respectively.
The overall analytical imprecision for TC and HDL-C exceeded 3% and 4%, respectively.

| Clinical agreement according to ATP III risk categories
Each result measured by the GCare Lipid Analyzer was categorized based on the ATP III guidelines. Whether the category to which the value of GCare Lipid Analyzer belonged to the category to which the reference Toshiba value belonged was calculated as a percentage. "Agreement" was defined as the two results being in the same category, while "disagreement" was not in the same category, such as a difference of one or more category. The categorized results for the capillary and venous blood samples are displayed in Table 3. The agreement (%) was >90% for all categories. All cases of disagreement occurred in only one category, belonging to those below or above.
The greatest disagreement in venous blood was observed for TC (8.6%). All misclassified cases with a TC <200 mg/dl had slightly overestimated TC levels; thus, they belonged to the upper categories of ≥200 mg/dl and <240 mg/dl. In the 10  disagreement rate for TG in the capillary blood (8.5%) was greater than that in the venous blood (4.5%). Among the 10 discordant cases of capillary blood in the TG level range of ≥150 to <200 mg/dl, six were misclassified as <150 mg/dl and four were in the ≥200 mg/dl group, revealing no significant systematic trend. In HDL-C, 20.6%-26.5% of cases with a level <40 mg/dl were falsely higher and misclassified into the above categories (>40 and >60 mg/dl).

| Evaluation of interfering substances in the GCare Lipid Analyzer using the GCare Glucose Test Strip
Potential substances that can influence glucose levels, including Hct, were evaluated. The difference between the average measured value at each Hct level and the average measured value at the mid-level Hct for glucose concentrations <100 mg/dl was less than ±10.0 mg/dl, and less than ±10.0% for glucose concentrations >100 mg/dl ( numbers in parentheses. The system accuracy results for venous and capillary blood specimens are summarized in Table 4 and Figure 4.
For glucose concentrations <100 mg/dl, 96.1% and 100% of the measured glucose levels were within ±15 mg/dl for the venous and capillary blood samples, respectively. For glucose concentrations ≥100 mg/dl, 100% and 99.5% of the measured glucose levels were within 15% for the venous and capillary blood samples, respectively.
The CEG analysis ( Figure S1) revealed that all measured glucose levels were located within Zones A and B for the venous and capillary blood specimens.

| User performance evaluation of the GCare Lipid Analyzer using GCare glucose test strip
A total of 100 volunteers with diabetes representing different ages, genders, and education levels were recruited to evaluate the GCare Lipid Analyzer using GCare Glucose Test Strip (Table S6). The blood glucose levels in the capillary blood measured by the test subjects were compared to the reference values obtained using the YSI 2300 STAT Plus analyzer. All glucose levels measured by the users were within ±15 mg/dl and ±15% when compared with the reference values (Table 4, Figure 4C). The CEG analysis demonstrated that all the results (100%) obtained by the users from capillary blood were within Zone A ( Figure S1C).

| DISCUSS ION
The lipid or lipoprotein test results form the basis for treatment policies in guidelines of dyslipidemia. According to the 2019 European guideline, 20 the treatment of hypertriglyceridemia is recommended at values of TG >200 mg/dl and HDL-C <40 mg/dl in patients at high cardiovascular risk. The therapeutic goal of LDL-C is subdivided according to risk status of each patients: <100 mg/dl for those at moderate risk, <70 mg/dl for those at high cardiovascular risk, <55 mg/dl for those at very high cardiovascular risk, and even lower, <40 mg/dl, for those who had a second vascular event during treatment of maximum dose of statins. As several meta-analyses revealed that an increased HDL-C did not reduce the risk of CVD or mortality, 21 Table 3).
The threshold for treating dyslipidemia is a TG level of 200 mg/dl, as mentioned above. This underestimation can be problematic for users performing capillary blood measurements at home. However, since low HLD-C levels (<40 mg/dl) are also included in the criteria for determining the treatment policy, the possibility of simultaneous underestimation of TG and false HDL-C seems low.  In this study, the GCare Lipid Analyzer demonstrated good clinical agreement with the reference values for TC, TG, and HDL-C using capillary and venous blood. The analytical performance based on the NCEP criteria was also acceptable. The GCare Glucose Test Strip meets the requirements for system accuracy indicated in the ISO 15197:2013 guidelines. The acceptable results of the user performance evaluation suggest that the device is reliable when used by non-experts. Thus, the GCare Lipid Analyzer is expected to benefit patients by facilitating the convenient monitoring of blood glucose and lipid levels.

CO N FLI C T O F I NTE R E S T
The author(s) declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

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.