Development of a sensitive HPLC‐MS/MS method for 25‐hydroxyvitamin D2 and D3 measurement in capillary blood

Abstract Background Measurement of 25‐hydroxyvitamin D [25(OH)D)] levels is important. The current method requires a relatively large volume of serum. To minimize the amount of serum needed, we established a high‐performance liquid chromatography‐tandem mass spectrometry (HPLC‐MS/MS) method to measure 25(OH)D in capillary serum. Methods Venous blood and fingertip blood were collected from 90 participants. Volumes of 100 µL of venous serum and 20 µL of capillary serum were collected. The serum samples were pretreated by protein removal, extraction and concentration, and an HPLC‐MS/MS method based on chromatographic separation and multi reactive ion monitoring was conducted. The intra‐ and inter‐batch variation coefficients were less than 10% for both 25‐hydroxyvitamin D3[25(OH)D3] and 25‐hydroxyvitamin D2[25(OH)D2)]. For venous specimens, the accuracies were 3.87% and 4.91%, respectively. For capillary specimens, the accuracies were 1.65% and 5.32%, respectively. Results The limit of detection (LOD) of 25(OH)D3 was 0.01 ng/mL, and the LOD of 25(OH)D2 was 0.05 ng/mL. The results showed that the mean concentration of 25(OH)D in venous blood was 22.56 ± 9.50 ng/mL, while the mean concentration of 25(OH)D in capillary blood was 18.14 ± 7.86 ng/mL. Furthermore, the adjusted capillary blood 25(OH)D level was 22.99 ± 10.24 ng/mL by the correction formula in our study. Similarly, the mean concentration of 25(OH)D3 in capillary blood was 17.98 ± 7.98 ng/mL. The adjusted capillary blood 25(OH)D3 level was 22.85 ± 10.42 ng/mL. No difference in the content of 25(OH)D or 25(OH)D3 was found between venous serum and corrected capillary serum. The correlation coefficients between venous and corrected capillary concentrations of 25(OH)D and 25(OH)D3 were 0.7941 and 0.8103, respectively, and the areas under the receiver operating characteristic curve were 0.9367 and 0.9565, respectively. Conclusions This capillary blood method requires minimal sample preparation and is suitable for routine use in the 25(OH)D detection.


| INTRODUC TI ON
Vitamin D is a fat-soluble vitamin that functions as a steroid hormone. It plays a crucial role in mineral homeostasis and skeletal health. 1 Vitamin D deficiency is increasing in the general population and is considered an important public health problem with serious effects. 2,3 It is necessary to implement widespread vitamin D testing among the public. In particular, key populations should be screened for vitamin D deficiency or insufficiency. However, the current classical method of HPLC-MS/MS requires a relatively large quantity of serum samples. Additionally, conventional arterial or venous blood collection methods are invasive and could potentially cause pain, needle stick injuries, and contamination. 13 In particular, it is difficult to extract venous blood from children. Thus, it is urgent to develop a method that requires only trace blood to detect. At present, blood glucose, 14 viruses, blood lead, 15 and other analytes can be detected from microblood in routine medical practice. Capillary blood analysis is relatively easy to perform and requires only a small volume of blood (<100 μL) obtained via a finger stick. Second, the test is much less costly than a venous blood test. Third, there are no environmental contamination problems associated with the Eppendorf (EP) tubes, even when tests are performed with small capillary blood specimens. 16 To address the shortcomings of venous blood testing, this study gathered 90 matched pairs of capillary and venous blood specimens collected on the same day during routine visits to a pediatric healthcare clinic and used them to establish a new HPLC-MS/MS method for quantitative determination of 25(OH)D in capillary blood.   The quality control range was ‾x ± 2s. The Westgard multirule quality control method was used to determine whether the quality of the results was acceptable. At least two quality control measurements were made in each batch of samples. The intra-and inter-batch variation coefficients were less than 5% for both 25(OH)

| Detection analysis of capillary serum 25(OH)D
The detection instrument was a high-performance liquid chromatography-tandem mass spectrometry apparatus (AB Sciex  (Table 2). Mass spectrometer settings were as follows: The temperature of the ion source is 400℃, the spray pressure is 70 psi, the air curtain pressure is 20 psi, and the ionization voltage is 4500 V.
The relative standard deviation was less than 15%. The content of 25(OH)D 3 and 25(OH)D 2 in quality control samples and samples to be tested was calculated according to the standard curve equation. The quality control range was ‾x ± 2s. The Westgard multirule quality control method was used to determine whether the quality of the results was acceptable. At least two quality control measurements were made from each batch of samples. The intra-and inter-batch variation coefficients were less than 10% for both 25(OH)

| Method comparison
The two HPLC-MS/MS methods were applied to 90 serum samples from patients (36 boys and 54 girls aged 3-10 years) who requested a 25(OH)D test in our clinical laboratory. Bland-Altman plots were used to identify the mean bias (the average of the difference between the measurements obtained from the two assays was also compared), and a 95% limit of agreement was observed between methods. Agreement between the assays regarding 25(OH)D status was assessed using Cohen's kappa (agreement: <0.4, poor; 0.4-0.75, fair to good; >0.75, excellent).
P<.05 was considered statistically significant. All statistical analyses were performed using EmpowerStats and GraphPad statistical software.

| Analytical performance evaluation of the HPLC-MS/MS methodology
The  2.5 ng/mL to 160.0 ng/mL, the linearity is good, and the correlation coefficient R 2 > 0.99. In the determination of 25(OH)D 3 , three concentrations of 10 ng/mL, 40 ng/mL, and 160 ng/mL were set for quality control. In the determination of 25 (OH) D 2 , three concentrations of 5, 10, and 40 ng/mL were set for quality control.

| Validation results of the HPLC-MS/MS assay method of venous and capillary specimens
The recovery (%) of the accuracy of the two compounds was in a reasonable range of 85%~115%. The accuracies were 3.87% and 4.91%, respectively (Table 3).

| 25(OH)D levels in venous and capillary specimens
The  (Table 5). Since 92% of samples contained less than 0.1 ng/mL 25(OH)D 2 , we did not include that analyte in the subsequent statistical analyses.

| Comparing the results of 25(OH)D and 25(OH) D 3 after correction
There was no difference in the mean concentration of 25(OH)D or 25(OH)D 3 between venous blood and corrected capillary blood among the three groups (Table 6). Accordingly, the data from the showed that the mean concentration of 25(OH)D in venous blood was 22.56 ± 9.50 ng/mL, which is consistent with values reported in the literature. 17 The mean concentration of 25(OH)D in fingertip blood was 17.98 ± 7.98 ng/mL, which was slightly lower than that in venous blood. This result was not surprising in light of a previous study that compared the concentrations of 12 chemical constituents between capillary serum or plasma and venous serum. 18 That study concluded that there were differences in the concentrations of potassium, total protein, and calcium between venous serum and capillary serum or plasma. The probable explanation for the lower concentrations in skin-puncture blood specimens is that they are mixed with interstitial fluid. These differences should be considered when comparing these types of specimens. Therefore,  screening young children in the field. Second, data on some important confounders, such as diet and 25(OH)D supplementation, were unavailable. However, considering the sensitivity of this method, the influence of diet and 25(OH)D supplementation should be limited.
Third, the retest validity and intertester reliability of our 25(OH)D assessment were not evaluated, which may obscure the measurement error.

| CON CLUS ION
A highly sensitive method for the determination of 25(OH)D in fingertip blood by HPLC-MS/MS was established. The method has few interfering factors, and the results are accurate and reliable. It is suitable for large-scale determination of 25(OH)D status for populationwide nutrition surveys. In the context of a public health program to screen 25(OH)D status, the finger preparation protocol would add convenience due to the reduced cost, shortened collection times, and increased safety for collection-site personnel.

CO N FLI C T O F I NTE R E S T
The authors have no financial or personal conflicts of interest to declare.

AUTH O R S ' CO NTR I B UTI O N S
XJ contributed significantly to the analysis and wrote the manu-