Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Dried blood spots (DBS) were introduced by Guthrie and Susie [1] in 1963 as a sampling technique for screening metabolic diseases of neonates in a large population. The applications of DBS have been reviewed recently by two research groups [2, 3]. DBS offers many advantages over conventional sampling techniques for blood, plasma, or serum collection [4]. These advantages can be summarized as less invasive sampling, simpler storage and handling, less infection risks through using pretreated sampling cards, and smaller sample volume [5]. In addition to newborn screening, DBS has also been applied in therapeutic drug monitoring and pharmacokinetics studies [6–37]. All these applications were based on the sole assumption in which the same size punches from dried blood spots on sampling paper/cards absorb the same volume of blood regardless of any differences among test subject genders, matrix lots, disease states, or blood viscosity. Sampling paper/cards are the ultimate carrier of the target analyte in blood. There have been a few types of sampling paper/cards available specifically designed for this purpose. Hundred percent cotton filter paper, marketed under the names of plain filter paper, Guthrie paper, Shleicher & Scheull 903, S & S903, Whatman 903, or No. 545 filter paper over different times or by different vendors, has been used successfully over the last few decades for newborn screening. However, inconsistent extraction was randomly observed between low and high concentration spots via quality controls, but also between fresh and aged blood spots [16, 25]. Paper pretreatment with chemicals [25] and via impregnation [36] provided a solution for unbalanced extraction recoveries. The paper pretreatment via impregnation of other chemicals is commercialized as multiple products such as Whatman FTA series for lysing cells, denaturing proteins, and preventing the growth of microorganisms. FTA series (i.e., FTA, FTA Elute, DMPK cards) was designed with proprietary formulations so that the paper/cards could be used in the bioanalysis of drug metabolism and pharmacokinetic samples for better accuracy over plain filter paper. Although other factors such as hematocrit [37], blood volume, and blood distribution have direct impacts on the accuracy and precision of bioanalytical methods on DBS, in this article we present only the potential impact of the sampling paper/cards as we have learned from implementing DBS-LC-MS in our laboratory.

2 Experiments

2.1 Chemicals

Six proprietary compounds were from Abbott Laboratories (North Chicago, IL, USA). Two stable isotope labeled compounds and additional proprietary compound from Abbott Laboratories were used as the internal standards. HPLC grade methanol and acetonitrile were from Sigma-Aldrich (St. Louis, MO, USA). A.C.S. grade ammonium formate and formic acid were also from Sigma-Aldrich. Water was purified through a Milli-Q water purifier from Millipore (Billerica, MA, USA). Human blood was purchased from Biological Specialty Corporation (Colmar, PA, USA).

2.2 Sampling Paper/Cards

Nine kinds of sampling paper/cards were used. FTA Elute (Micro), FTA DMPK-B, FTA DMPK-A, Indicating FTA (pink), FTA Plant, Blood Stain, Protein Saver 903 (previously Guthrie Card or Schleicher & Scheull Card in different times), and 31 ET were from Whatman (now GE Healthcare, Piscataway, NJ, USA) and VWR 237 filter paper from VWR International (West Chester, PA, USA).

2.3 Instruments

Harris Uni-Core 1.00, 3.00, 6.00 mm cutters and cutting mats were from Whatman. A Sil-10 HTc autosampler and two LC-10AD pumps were from Shimadzu (Kyoto, Japan). An API-4000 mass spectrometer with turbo ionspray ionization probe was from AB Sciex (Foster City, CA, USA). The software for data acquisition and processing was Analyst ver. 1.5 from Applied Biosystems as well. An automated liquid handler was MicroLab AT 2 Plus from Hamilton (Reno, NV, USA). A MixMate vortexer was from Eppendorf (Hauppauge, NY, USA). A CR412 centrifuge was from Jouan (now Thermo Fischer). Column selection valve was from Valco Instruments (Houston, TX, USA). An Atlantis dc18 column (100  ×  2.1 mm, 5 μm) was purchased from Waters (Milford, MA, USA).

2.4 Sample Preparation

Dried blood spot on different paper/cards were punched and dispensed into the wells of a 96-well plate, 20 μL of internal standard in approximately 50 % acetonitrile was added. Three hundred microliters of 70 % acetonitrile (v/v) was added and the plate was vortexed for 5 min at 800 rpm. Fifty microliters of 6 M ammonium formate was added and the plate was vortexed at 850 rpm for 10 min. The plate was then centrifuged at 3,500 rpm and 10ºC for 5 min. The upper layer was transferred to a new plate and the samples were evaporated by a nitrogen stream at room temperature. The dried samples were reconstituted with 30 % acetonitrile and 5 μL of reconstituted sample were injected.

2.5 Chromatography

Two mobile phases were used. Mobile phase A was 0.1 % formic acid and 5 % acetonitrile in water and mobile phase B was 0.1 % formic acid and 95 % acetonitrile in water. Chromatographic separation was achieved for all compounds using an 8.5 min gradient method. Initially, 7.5 % mobile phase B was used for system equilibration. After each injection, mobile phase B was run at 7.5 % for 1 min then it was linearly increased to 90 % over the course of 3 min, it was maintained at 90 % for 2.5 min and returned to 7.5 % in 0.1 min. The postgradient column equilibration was run for 1.9 min prior to the start of pretreatment for the next injection. Figure 1 shows a typical chromatogram of six compounds (Compounds A through F) and three internal standards eluted with the described chromatography conditions. Compound G was used as the internal standard for Compounds B, C, and E.

Fig. 1
figure 1

Chromatogram of six compounds (Compounds A–F) with two labeled internal standards (IS) and one nonlabeled IS. A-IS is the labeled IS for Compounds A and D and F-IS is the labeled IS for Compound F. Compound G is the IS for Compounds C and E

Full size image

2.6 Mass Spectrometry

LC-MS/MS detection was performed using an AB Sciex API 4000TM triple quadrupole mass spectrometer equipped with a TURBO VTM ionspray ionization source operated in positive ion mode. The software used for instrument control was Analyst version 1.5. The spray voltage was set at 5,500 V and the source temperature was 550 °C. Other ion source parameter settings were 10 for curtain gas, 65 for GS1, 65 for GS2. The collision gas setting was 4 and the entrance potential was 10 V. Other compound specific parameters such as the declustering potential (DP), the collision energy (CE), and the cell exit potential (CXP) varied depending on the analyte. The DP ranged from 50 to 110 V, the CE ranged from 23 to 80 eV, and the CXP ranged from 5 to 15 V. Acquisition was performed in scheduled MRM mode with a 60 s window around the retention time of each analyte and a target scan time of 0.2 s.

3 Results and Discussion

3.1 Water-Soluble Chemicals on the Cards

In order to estimate the amount of water-soluble components in the sampling paper/cards, two 6 mm disks from each type of paper were cut into preweighed test tubes and weighed using an analytical balance, the disks were wetted with 500 μL of distilled water, shaken well, and then weighed again. The water was carefully transferred off the washed paper disks into preweighed test tubes with a micropipette and weighed, then put under a stream of nitrogen gas for drying and reweighed when completely dried. The paper disks were also dried under nitrogen and reweighed. Table 1 shows the relative amounts of chemicals on the paper in percentage. The weight loss of the paper disks before and after water washing was confirmed by the weight of the water-soluble chemicals released into water which confirmed the quantitative transfer of the liquids between the vials. The ratio of weight of the chemicals released into water to the weight loss of the disk is presented as accuracy (%) in Table 1. FTA elute card appears to carry the most water-soluble chemicals (≥50 % of total card weight) among all sampling cards. FTA DMPK-A, FTA plant, and Indicating FTA cards appear to be the same or similar paper, but Indicating FTA paper carries a pink dye additive. Blood stain, 31ET, and VWR 237 paper/cards carry the least water-soluble chemicals. Actually, 31ET and VWR 237 are marketed as plain cotton fiber paper.

Table 1 Measured amount of water-soluble chemicals on sampling paper/cards
Full size table

3.2 Identities of Chemicals Impregnated on the Paper

Based on the amount of water-soluble chemicals on the paper/cards, three types of cards (FTA Elute, Indicating FTA, and VWR 237) were used to explore the chemical identifications. The possible chemicals on the cards were checked with mass spectrometry Q1 scanning in both positive and negative ionization modes. Two disks (6 mm i.d. each) free from biological matrix for FTA Elute, Indicating FTA and VWR 237 paper/cards were extracted. Two microliters of the reconstituted solution was injected into the liquid chromatograph system. More than ten chemicals were found to be present among pretreated paper/cards, but only a few typical chemicals such as tris-hydroxymethyl aminomethane and sodium dodecyl sulfate from Indicating FTA (DMPK-A), guanidine thiocyanate from FTA DMPK-B could be clearly identified by matching the m/z of their molecular ions. Figure 2a–e show the extracted ion chromatograms for the representative m/z of the various chemical components observed in the noted paper/cards. Figure 2a–c were from positive ionization scanning and Fig. 2d–e were from negative ionization scanning. VWR 237 paper appears to be significantly cleaner than FTA Elute and Indicating FTA paper/cards. According to the patent [38] covering the sampling cards, the coating materials on sampling cards may include impregnating agents such as polystyrene, a weak base, a chelating agent, an anionic detergent and optional uric acid/urate salt, protein denaturing agent, and a free radical trap, etc.

Fig. 2
figure 2

(a) XIC of +Q1: 292.8–293.3 Da as EDTA from Indicating FTA Paper; (b) XIC of +Q1: 121.5–122.0 Da as Tris from Indicating FTA Paper; (c) XIC of +Q1: 59.8–60.3 Da as guanidine thiocyanate from FTA Elute Paper; (d) XIC of −Q1: 57.75–58.25 Da as guanidine thiocyanate on FTA Elute; (e) XIC of −Q1: 264.431–264.931 Da as SDS on Indicating FTA

Full size image

Most of the chemicals found on the FTA Elute and Indicating FTA paper/cards are believed not to be friendly with LC-MS because these chemicals affect the accurate quantitation of target analyte in the manners of ionization suppression, enhancement or matrix effect when they are co-eluted with the target analytes or even distortion of the chromatography by prolonged retention or accumulation. Therefore, the impacts of these water-soluble chemicals on the quantitation of the target analyte must be carefully evaluated in the development of any DBS-LC-MS method, a general approach is to minimize or eliminate the co-elution of these compounds by separating these chemicals from the target analyte using liquid chromatography. Depending on the compounds used to impregnate the cards that may be easily achievable or may become a challenge as in the case of anionic surfactants with common reversed phase columns. The anionic surfactants tend to be highly retained and difficult to clear from the chromatographic columns.

3.3 Paper/Card and Compound Dependence of Analyte Distribution in Blood Spots

In order to evaluate the impacts of different paper/cards on the analytical results, calibration standards and quality controls (QCs) were prepared in fresh human blood containing six compounds (Compound A through F) with different physical and chemical properties. Log  D was considered as a critical measure of the compound properties [39]. The log  D (at pH 7.4) is shown in Table 2. The purpose of using various compounds was to explore the possible correlations among the compounds, paper/card types, matrix effect, and spotting process related factors. Accurately controlled volumes (5 or 15 μL) of both calibration standards and QCs were spotted onto the sampling paper/cards with calibrated positive displacement pipettes and the spotted cards were air-dried at room temperature, a few sets of aliquots of the standards and QCs in blood were also refrigerated for future use. When the evaluation experiment was conducted a set of blood calibration standards and QCs (5 μL each), and/or a set of DBS standards and QCs were extracted through a salting-out assisted liquid–liquid extraction as described in Sect. 2. For the evaluation of analyte distribution in the dried blood spots, QCs were quantified against blood standards. Although concentrations of DBS QCs measured against blood standards is only semiquantitative due to the sample type difference, this approach was believed to be sufficient to provide the relative difference in concentration of an analyte across one spot. Certain volumes of blood quality controls of six compounds (A–F) as listed in Table 2 were spotted on three types of sampling paper/cards (FTA DMPK-B, FTA Elute, and VWR 237), only the results from two paper/cards (FTA Elute and VWR 237) and two compounds (Compounds A and F) are presented, to show the impact of the paper/card types on the compound quantitation.

Table 2 Log  D (7.4) of the compounds
Full size table

3.3.1 Spotting Volume Impact on the Compound Quantitation on Dried Blood Spots on Different Paper/Cards

The impact of spotting volume has been evaluated. DBS cards were prepared by spotting increasing volumes of fresh blood QCs of six compounds from 5 to 50 μL and air-dried. Three millimeter (i.d.) punches were taken from the centers (visually located), extracted and quantified against calibration standards in fresh human blood. Three replicates were analyzed and the mean found concentration  ±  standard deviations for Compounds A and F are shown in Fig. 3a–d. The VWR 237 paper shows lower change in center spot concentration between the lowest and the highest spotting volumes than FTA Elute paper/cards. For example, the center concentration of Compound A for 50 μL spotting volume was approximately 120 % of that of the same compound for 5 μL spotting volume on the VWR 237 paper. On the other hand, the center concentration of Compound A for 50 μL spotting volume was approximately 170 % of that of the same compound for 5 μL spotting volume for the FTA Elute card. For Compound F with high log  D, similar trends were observed, and VWR 237 paper show more consistent mean concentration with larger standard deviation, while the FTA Elute cards show more change in center spot concentration across different volumes with smaller standard deviation among replicates. Based on the observed trends, a conclusion can be drawn that a certain measure of control needs to be utilized for the spotted volume. Depending on the acceptance criteria of the individual assay, more or less variation of the spotted volume may be acceptable. The most accurate concentration will still be obtained when an accurate volume is spotted.

Fig. 3
figure 3

Mean concentration (center punches) vs. spotting volume for (a) Compound A on VWR 237 paper; (b) Compound A on FTA Elute; (c) Compound F on VWR 237 paper; (d) Compound F on FTA Elute cards

Full size image

3.3.2 Paper/Card Dependent Distribution of Analytes on Dried Blood Spots

There are various factors that can influence the distribution of analytes in a dried blood spot. Water-soluble chemicals uniformly coated on DBS cards would redistribute when the blood was spotted. The redistribution of chemicals may depend on their properties, viscosity of blood, the volume spotted, and the technique used for spotting. Another factor is the viscosity of the blood. Viscosity is normally dependent on the blood composition (hematocrit, protein, lipid levels), and it can affect the physical spread of the blood spot in that the same volume of a less viscous blood will form a larger diameter spot than that of a more viscous blood sample. Viscosity, combined with the chemical redistribution on the sample cards, will increase the complexity of the analyte distribution.

In order to study the radial distribution of the six analytes on dried blood spots, smaller DBS punches (1 mm. i.d.) were extracted and quantified against blood standards. Six DBS 1 mm punches were taken from the centers of six separate spots and combined into one well of a 96-well plate. The next DBS punch was taken adjacent to the center one and each subsequent punch was taken adjacent to the previous one radially moving outward towards the edge as shown in Fig. 4. Additional punch was taken from just outside of the visual edge of the blood spot as a control sample to verify that analytes are not moving outside of the visual spot. All results from the control samples confirmed the absence of the tested compounds outside of the visual border of spots. Three types of paper/card and six compounds were used to evaluate the paper impact on the analyte distribution, but only FTA Elute and VWR 237 cards are presented here.

Fig. 4
figure 4

High resolution punches with 1 mm (i.d.) punches

Full size image

Figure 5 shows the radial distribution of analytes on FTA Elute cards and VWR 237 paper using two spotting volumes (15 and 50 μL). The general impression from the obtained results is that FTA Elute cards show a more significant change in concentration across the spot especially when crossing the line between the darker and lighter (halo) regions of the spot. The VWR 237 paper shows minimal change in distribution of the analytes across the spots. One observation relates to the distribution of the analytes within the dark center of the spots on the FTA Elute cards where the higher log  D compound F shows a gradual change upward and then downward in concentration from center to the edge of the dark circle as shown in Fig. 5b (for both 15 and 50 μL spotting volumes), while the low log  D compound A shows no obvious upward or downward change as shown in Fig. 5a. This was also seen for the rest of the compounds studied when looking at their log  D values. It is possible that this phenomenon may be due to the redistribution of the water-soluble chemical components of the paper, as previously described, as well as to the specific properties of the analytes.

Fig. 5
figure 5

Mean concentration vs. high resolution punch distance from the spot center for (a) Compound A on FTA Elute Cards and VWR 237 paper; (b) Compound F on FTA Elute Cards and VWR 237 paper

Full size image

3.3.3 Paper/Card Dependent Matrix Effect

Accurate quantitation of analytes requires consistent measurements independent from matrix lot variation. Matrix lot to lot variation becomes critical for successfully implementing DBS. In addition to the analyte distribution across dried blood spots, matrix effect was also evaluated. In addition to the blood pool used for ­standards and QCs preparation, two additional lots of blood were used to prepare two matrix effect (ME) samples at low and high concentrations. All ME samples, QCs and calibration standards in blood carrying six compounds were spotted on three types of paper/cards at the same time. The spotting volume was either 15 or 5 μL. The spot centers (3 mm i.d.) were punched for the samples prepared with 15 μL of blood but full spots (6 mm i.d.) were punched for the samples prepared with 5 μL of blood. The punches were extracted and analyzed with the procedure described earlier. QCs and ME samples were quantified against calibration standards that were prepared, spotted, and analyzed in the same fashion. Mean bias (%) was calculated for all evaluation sample lots (QCs or ME samples) and all concentration levels of each of the five compounds. Standard deviation error bars are used to reflect the variation of the individual bias among all concentration levels. Compounds A, B, C, E, and F on FTA Elute Cards and VWR 237 paper are presented in Fig. 6a–d showing the paper/card impact on the matrix effect. Compound D is not presented here because the analyte was eluted at the zone when severe compound-specific matrix ionization suppression was observed (chromatogram not shown). The acceptance criteria used was ±15 % for bias in matrix effect evaluation. Figure 6a, b indicate that FTA Elute cards have significant matrix effect compared to VWR 237 paper spot center punches for all analytes tested. However, when using accurate low volume spotting and a full punch, the overall matrix effects generated on FTA Elute cards and VWR 237 paper are similar. In other words, plain paper with either partial or full punch of the spot generates less matrix effect than treated paper/cards.

Fig. 6
figure 6figure 6

(a) Mean bias (%) of ME samples from 15 μL spotting on FTA Elute cards and partial punches. (b) Mean bias (%) of ME samples from 15 μL spotting on VWR 237 paper and partial punches. (c) Mean bias (%) of ME samples from 15 μL spotting on FTA Elute cards and full punches. (d) Mean bias (%) of ME samples from 15 μL spotting on VWR 237 paper and full punches

Full size image

4 Conclusion

The paper/cards used for dried blood spot sampling can affect the accurate quantitative determination of analytes for pharmacokinetic evaluations. Various factors were evaluated using six analytes over a wide log  D range. These were related to impregnated chemicals on the cards, variable volume, viscosity, and compound dependent distribution of the analytes. The widely used treated paper/cards generally showed a more complex and significant impact than the plain filter paper/cards. Chemicals on the treated paper/cards in conjunction with the specific properties of the analytes may cause variations in the analyte distribution that could provide a challenge, so evaluations of the sampling paper/card should be carefully conducted before a DBS method can be applied for accurate measurement of analytes in DBS-LC-MS or if pharmacokinetics is the end point of a study. The level of acceptable variation needs to be evaluated and decided before implementing such a method for quantitative work. Spotting on uniform plain filter paper was shown to reduce the variability with regard to matrix effects, thus resulting in a more accurate analyte determination. However, the most accurate matrix and paper/card independent values were derived from accurate volume spotting with full cut of the spots. In addition to the factors presented, hematocrit effect or any other components that lead to varying blood viscosity, the paper/card type, media evenness, and paper lot-to-lot variation should be considered as critical impact factors.