ABSTRACT
Objectives
Neonatal patients present a challenge to the clinical laboratory because of their low blood volume. The Sysmex XN‐series features a predilution (PD) mode allowing hematologic measurements with only 20–50 μL of blood. We verified the PD mode for analysis of selected hematologic parameters in 50 μL microsamples.
Methods
Microsamples were prepared from adult EDTA blood. White blood cell count (leukocytes, neutrophils, and lymphocytes) and red blood cell parameters (erythrocytes, hemoglobin, hematocrit, mean cell volume, and reticulocytes) were evaluated. The imprecision of the PD mode was evaluated over 3 days, and the accuracy was assessed by method comparison with the standard whole blood mode. The effect of capillary sampling, microsample stability during storage, and pneumatic tube transportation was evaluated. Finally, the bias was reproduced in a small sample of pediatric patients.
Results
For white blood cells, the bias was ≤ 5.4% (95% CI: 4.5–6.2) and imprecision was below 3.5% (except at the lowest levels). Capillary sampling had little effect on PD analytical performance (bias ≤ 0.7% and imprecision ≤ 4.7%) and white blood cell counts were stable for 7 h at room temperature and after pneumatic tube transportation. The red blood cell parameters generally exceeded the allowable bias and imprecision. The bias of the pediatric samples remained within the 95% PI for the method comparison studies.
Conclusion
The PD mode has acceptable analytical performance and preanalytical stability for white blood cell counts but not for red blood cell parameters. It may offer a low‐volume alternative for hematologic monitoring.
Keywords: biochemistry, method comparison, neonatal hematology, precision, red blood cells, white blood cells
1. Introduction
Neonatal patients present a special challenge to the clinical laboratory due to their low blood volume, which precludes repeated, normal volume blood draws. This is handled by maintaining strict indications for laboratory analysis and use of low‐volume methods such as capillary pipettes and raised‐bottom tubes (RBT). Special precaution is needed when frequent and repeated sampling is indicated. One such example is the monitoring of thrombocyte count in neonatal patients to uncover the potential development or progression of severe, life‐threatening disease [1]. We recently published data demonstrating that the predilution (PD) mode of Sysmex XN‐9000 is feasible for automated thrombocyte counting in low‐volume capillary microsamples and that the introduction of the PD mode led to analytical and clinical improvements at our hospital [2].
Following the implementation of the PD mode for thrombocyte counting in capillary microsamples in our laboratory, a question arises whether this method has the potential to improve other aspects of hematologic monitoring in neonatal patients or other patients refraining from ordinary venous blood sampling. Thus, this report elaborates further on the potential of the Sysmex PD mode for hematologic measurements in capillary microsamples. It verifies eight additional key hematologic parameters (white blood cell [WBC] count, neutrophil and lymphocyte counts, red blood cell [RBC] count, hemoglobin, hematocrit, mean cell volume [MCV], and reticulocyte count) on the PD mode of our Sysmex XN9100 according to standard procedures in our accredited laboratory (DS/EN ISO18159). We also examine the effects of critical preanalytical parameters on PD mode performance.
2. Materials and Methods
2.1. Blood Samples
The blood samples for this study were collected and analyzed at the Department of Clinical Biochemistry, Aalborg University Hospital between September 2022 and June 2023. The samples were venous blood or capillary samples from different sources. (1) Venous blood samples were collected anonymously from adult volunteers by standard venipuncture of the antecubital vein into 2 mL K2‐EDTA anticoagulated tubes (Vacuette, Greiner Bio‐One International GmbH). (2) Excess venous blood was leftover K2‐EDTA anticoagulated venous blood samples in our routine hematologic laboratory. We used this material to prepare (3) microsamples by soaking blood from the EDTA tubes into a 50 μL EDTA coated pipette (VITREX Medical A/S, Herlev, Denmark) and dissolved in 300 μL CellPack DCL (Sysmex Europe, Norderstedt, Germany). (4) Capillary microsamples were collected from the fourth finger of adult volunteers into 50 μL EDTA anticoagulated pipettes after skin piercing of the index finger with a contact activated lancet (Becton, Dickinson and Company Limited, Dublin, Ireland) and subsequently dissolved in CellPack DCL as described above. (5) Pediatric capillary samples were collected anonymously (only age registered) from leftover material in our routine laboratory. This material was from EDTA‐anticoagulated RBT (Microtainer, Becton Dickinson). This blood is sampled through a skin incision of the lateral heel with a standard lancet (VITREX Steriheel Baby, VITREX Medical) into the RBT tube (250 μL). (6) Pediatric microsamples were prepared from the RBT tubes as outlined above (no. 3).
2.2. Evaluation of Bias
The accuracy of the PD method was evaluated by its bias against WB measurements on the same Sysmex XN‐9100 instrument. Samples were collected and analyzed over several days between November 2022 and June 2023. For each parameter, at least 155 samples covering the measurement range were identified. Of these, at least 20 samples were below a predefined limit for low level: WBC < 3 × 109/L; neutrophils < 1.5 × 109/L; lymphocytes < 0.5 × 109/L; erythrocytes < 3.0 × 1012/L; hemoglobin < 5.5 mmol/L; hematocrit < 0.30 vol frac; MCV < 85 fL; and reticulocytes < 30 × 109/L. The samples were analyzed once in WB mode, and microsamples were subsequently prepared and measured in PD mode.
2.3. Evaluation of Precision
The precision was evaluated at three levels over 3 days. Two EDTA tubes at each level (n = 6, in total) were collected among routine samples in our laboratory on three different days (n = 18, in total). Two microsamples were prepared from each EDTA tube, and each of the microsamples was analyzed in PD mode in duplicate. The precision was evaluated at two levels: (1) between the microsample duplicates (CVintra sample) to provide a measure of pure analytical variation and (2) between the first of the microsample duplicates (CVinter sample) to incorporate the preanalytical variation of handling and dilution of the capillary samples.
2.4. Preanalytical Factors
The effect of capillary sampling on PD performance was evaluated in a separate experiment. To evaluate the bias, a paired venous blood and capillary microsample was collected from 10 adult volunteers, measured as singles in WB and PD mode, respectively. To evaluate the imprecision, paired capillary microsamples were collected from the right and left fourth fingers of these 10 adult volunteers and measured as singles in PD mode.
For stability testing, capillary microsamples were collected from two adult persons at 5 different days (n = 10). The samples were measured as singles in PD mode after 0‐, 4‐, or 7‐h incubation at room temperature (21°–23°).
The effect of transportation through our pneumatic tube system (Tempus 600, SARSTEDT) was evaluated in 10 samples over 2 consecutive days. A venous blood sample and a capillary microsample were simultaneously collected from random, adult patients. The samples were sent by the pneumatic tube system and analyzed in WB and PD mode, respectively, within 60 min of sampling.
2.5. Pediatric Samples
To evaluate the PD mode in pediatric samples, excess material from RBT tubes of 29 random pediatric patients was collected in our routine laboratory. As per routine in our laboratory, a full hematologic profile is performed on pediatric RBT tubes. A microsample was prepared from each RBT sample and analyzed in single in PD mode. The present project is a quality improvement project and therefore exempt from formal ethical approval. It has been presented to the local ethical committee for confirmation of this exemption (No.: 20220031).
2.6. Hematologic Analysis
All measurements were performed using a Sysmex XN‐9100 hematologic analyzer validated and accredited for routine clinical use. Our Sysmex line includes five XN analyzers, of which the PD channel is calibrated in two. The performance of the PD channel is monitored by daily comparison with the WB mode in the same XN module using patient samples with high and low platelet counts. For the presented experiments, WB measurements were performed among routine samples by loading on the standard conveyor belt and PD analysis by hand‐held front loading. All laboratory work was conducted by certified laboratory technicians with special competencies in hematologic laboratory work.
2.7. Statistics
Data are presented as mean with standard deviation (SD), range, and absolute numbers. The relative bias of the PD mode is expressed as the mean of PD/WB with a 95% confidence interval (95% CI) and a 95% prediction interval (95% PI) at the full measurement range and in the predefined low level. The data are presented in Bland–Altmann plots with WB levels on the X‐axis and the relative bias of the PD measurement (PD/WB) on the Y‐axis. The association between PD and WB measurements were also evaluated by Passing–Bablok regression with bootstrapping 95% CI and Lin's concordance correlation coefficient. The maximum allowable bias was calculated from intra‐ and interindividual biological variation as recommended by EFLM and Westgaard [0.25 × ((CVIntra)2 + (CVInter)2)0.5] [3, 4]. Estimates for biological variation were retrieved from the EFLM database [3]. The estimated maximum allowable biases are: WBC = 4.9%, neutrophils = 6.9%, lymphocytes = 6.3%, erythrocytes = 1.8%, hemoglobin = 1.6%, hematocrit = 1.5%, MCV = 0.9%, and reticulocytes = 7.2%.
The precision of the PD mode was evaluated as the coefficient of variation (CV) between microsamples and between microsample duplicate measurements at three predefined levels (low, intermediate, and high). The maximum allowable imprecision was calculated from biological variation as recommended by EFLM and Westgaard (0.5 × CVintra) and the estimates for biological variation were retrieved from the EFLM database [3, 4]. The estimated maximum allowable imprecisions are WBC = 5.4%, neutrophils = 7.0%, lymphocytes = 5.4%, erythrocytes = 1.3%, hemoglobin = 1.4%, hematocrit = 1.4%, MCV = 0.4%, and reticulocytes = 4.9%.
The effect of capillary sampling on analytical performance of the PD mode was evaluated by the bias of venous vs. capillary sampling and the imprecision of finger‐to‐finger sampling. The maximum allowable bias was estimated as above. The finger‐to‐finger imprecision was expressed as mean CV and evaluated against the maximum allowable imprecision calculated above.
For the stability studies, the maximum allowable deviation for individual measurements was defined as the total analytical error [bias + 1.65 × imprecision] whereas the allowable deviation of the mean from initial values was the bias. The total analytical errors were WBC = 13.9%, neutrophils = 18.5%, lymphocytes = 15.2%, erythrocytes = 3.9%, hemoglobin = 3.9%, hematocrit = 3.8%, MCV = 1.6%, and reticulocytes = 15.3%. We compared levels of analytes in venous blood and capillary microsamples after transportation in our pneumatic tube system. The maximum allowable relative difference (PD/WB) was the bias as estimated above. For pediatric samples, the relative bias between the microsample measured in PD mode with the RBT sample measured in WB mode (PD/WB) was calculated. This bias was compared with the 95% PI of the method comparison studies. For all studies, an alpha level of 0.05 was applied, and the statistical tests and graphical presentations were performed in STATA version 17.0.
3. Results
3.1. Analytical Performance
Results of the method comparison are presented in Figure 1a,b, and the precision data are presented in Table 1. In general, the bias and imprecision of the WBCs were within the maximum allowable deviations, except at the lowest analyte levels (Figure 1a). The RBC parameters generally exceeded the maximum allowable biases and imprecision limits (Figure 1b).
FIGURE 1.
(a) Method comparison between predilution (PD) and whole blood (WB) modes of white blood cell counts. Bland–Altman plots show relative bias expressed as PD/WB ratio against level measured in WB mode. The full range (left panel) and low levels (middle panel) of each parameter are shown. Solid red line indicates the mean bias with its 95% confidence interval (CI) in red dashed lines. Black dashed lines indicate the 95% prediction interval (PI). The bias estimate with 95% CIs and 95% PIs are displayed in the lower right corner of each Bland–Altman plot along with number of measurements. For leukocyte and neutrophil counts, samples > 100 × 109/L are excluded for presentation purposes. Results of the Passing‐Bablok regression are presented in the right column. The fitted regression line and its 95% CI is shown. The regression equation (Beta) and Lin's concordance coefficient (Lin's) is displayed in the lower right corner. Maximum allowable biases: WBC = 4.9%, Neut = 6.9%, Lymp = 6.3%. (b) Method comparison between PD and WB modes of white blood cell counts. Bland–Altman plots show relative bias expressed as PD/WB ratio against the level measured in WB mode. The full range (left panel) and low levels (middle panel) of each parameter are shown. Solid red line indicates the mean bias with its 95% CI in red dashed lines. Black dashed lines indicate the 95% PI. The bias estimate with 95% CIs and 95% PIs are displayed in the lower right corner of each Bland–Altman plot along with the number of measurements. Results of the Passing–Bablok regression are presented in the right column. The fitted regression line and its 95% CI are shown. The regression equation (Beta) and Lin's concordance coefficient (Lin's) are displayed for each regression line. Maximum allowable bias: RBC = 1.8%, Hb = 1.6%, HCT = 1.5%, MCV = 0.9%, Ret = 7.2%.
TABLE 1.
Precision of predilution measurements in adult blood samples.
| Low | Medium | High | |
|---|---|---|---|
| White blood cells | |||
| CVintrasample | 6.1% (3.8–8.4) | 1.4% (1.2–1.7) | 1.0% (0.7–1.3) |
| CVintersample | 6.1% (4.4–7.8) | 1.7% (1.3–2.0) | 1.4% (0.8–2.0) |
| Range | 0.1–1.4 × 109/L | 4.7–8.0 × 109/L | 16.0–39.8 × 109/L |
| Neutrophils | |||
| CVintrasample | 16.9% (3.8–30.0) | 1.7% (1.4–1.9) | 1.0% (0.8–1.2) |
| CVintersample | 8.6% (0.5–16.7) | 1.9% (1.5–2.3) | 1.3% (0.7–1.9) |
| Range | 0.01–1.51 × 109/L | 4.03–7.91 × 109/L | 12.91–32.82 × 109/L |
| Lymphocytes | |||
| CVintrasample | 12.2% (2.5–21.9) | 2.6% (2.1–3.1) | 3.3% (2.5–4.1) |
| CVintersample | 10.2% (5.7–14.8) | 2.8% (2.0–3.6) | 3.5% (2.5–4.6) |
| Range | 0.13–0.52 × 109/L | 1.97–2.89 × 109/L | 4.50–22.70 × 109/L |
| Erythrocytes | |||
| CVintrasample | 1.0% (0.8–1.3) | 0.8% (0.6–0.9) | 1.0% (0.6–1.4) |
| CVintersample | 1.3% (0.9–1.8) | 1.3% (1.0–1.6) | 2.1% (1.0–3.3) |
| Range | 1.74–2.50 × 1012/L | 4.02–4.89 × 1012/L | 5.47–7.15 × 1012/L |
| Hemoglobin | |||
| CVintrasample | 0.9% (0.5–1.3) | 0.7% (0.6–0.9) | 0.7% (0.4–1.0) |
| CVintersample | 1.5% (0.8–2.0) | 1.2% (0.9–1.5) | 1.8% (1.1–2.5) |
| Range | 3.0–4.5 mmol/L | 7.0–9.0 mmol/L | 9.9–13.2 mmol/L |
| Hematocrite | |||
| CVintrasample | 1.2% (0.8–1.5) | 1.0% (0.9–1.2) | 1.1% (0.8–1.4) |
| CVintersample | 1.6% (1.0–2.3) | 1.3% (1.0–1.7) | 1.6% (0.9–2.3) |
| Range | 0.17–0.22 | 0.30–0.40 | 0.45–0.67 |
| Mean cell volume | |||
| CVintrasample | 0.7% (0.6–0.8) | 0.7% (0.6–0.8) | 0.8% (0.6–0.9) |
| CVintersample | 0.3% (0.2–0.3) | 0.2% (0.2–0.3) | 0.3% (0.2–0.5) |
| Range | 65–81 fL | 85–95 fL | 105–121 fL |
| Reticulocytes | |||
| CVintrasample | 10.7% (7.0–14.4) | 4.5% (3.9–5.0) | 2.8% (1.9–3.7) |
| CVintersample | 8.3% (5.1–11.5) | 5.0% (4.1–5.8) | 4.3% (2.1–6.4) |
| Range | 4–10 × 109/L | 50–100 × 109/L | 150–545 × 109/L |
Note: Precision of predilution (PD) measurements was evaluated at three levels over at least 3 days. Two microsamples were prepared from random adult blood sample tubes and measured in duplicate. The coefficient of variation (CV) within these microsamples (CVintrasample) and between microsamples (CVintersample) is presented with 95% confidence interval (CI) for each evaluated range of analyte.
3.2. Pre‐Analytical Variables
3.2.1. Effect of Capillary Sampling
To evaluate the effect of capillary sampling on PD performance, bias, and imprecision of capillary microsamples were evaluated and are presented in Table 2. The bias and imprecision of the WBCs were within the acceptance limits. The bias of the RBCs, except for reticulocytes, exceeded the acceptance criteria for bias, while the imprecision was generally acceptable.
TABLE 2.
The effect of capillary sampling on bias and imprecision of the PD mode.
| Bias (95% CI) | Imprecision (95% CI) | Tested range | |
|---|---|---|---|
| White blood cells | 0.7% (−4.0 to 5.5) | 3.3% (2.1–4.4) | 5.1–13.0 × 1109/L |
| Neutrophils | 0.3% (−4.2 to 4.7) | 2.9% (1.4–4.4) | 2.92–10.18 × 109/L |
| Lymphocytes | 0.7% (−5.3 to 6.8) | 4.7% (2.4–7.1) | 1.67–3.85 × 109/L |
| Erythrocytes | 3.0% (1.0 to 5.0) | 1.4% (0.5–2.2) | 3.89–5.46 × 1012/L |
| Hemoglobin | 3.2% (1.6 to 4.7) | 1.0% (0.6–1.4) | 7.2–10.6 mmol/L |
| Hematocrite | −1.7% (−3.3 to 1.0) | 1.4% (0.6–2.2) | 0.33–0.47 vol frac |
| Mean cell volume | −4.5% (−5.1 to −3.9) | 0.2% (0.1–0.3) | 82–90 fL |
| Reticulocytes | −7.1% (−12.8 to −1.4) | 4.4% (1.9–6.8) | 41–120 × 109/L |
Note: The impact of capillary sampling on predilution (PD) performance was evaluated by comparison of the indicated parameters. The bias between paired venous and capillary samples were evaluated. The imprecision was evaluated between paired capillary samples from the right and the left finger fourth finger. Results presented as relative bias with 95% confidence intervals.
3.2.2. Stability and Transportation
The results of the stability studies are presented in Table 3 and the effect of transportation by the pneumatic tube system in Table 4. The WBCs were stable for up to 7 h and remained within acceptance limits after pneumatic tube transportation. The RBCs were stable for 4 or 7 h, except MCV, which was increased. Only hematocrit levels remained within allowable limits after transportation by the pneumatic tube system.
TABLE 3.
Stability of adult microsamples at room temperature.
| 0 h | 4 h | 7 h | |
|---|---|---|---|
| White blood cells | |||
| Relative level | 100% | 98.8% (93.8 to 102.9) | 103.3% (100.7–106.0) |
| Range | 4.7–12.6 × 109/L | ||
| Neutrophils | |||
| Relative level | 100% | 98.8% (−93.2 to 103.6) | 102.6% (98.8–107.0) |
| Range | 2.24–8.65 × 109/L | ||
| Lymphocytes | |||
| Relative level | 100% | 99.6% (95.0 to 105.5) | 106.1 (101.6–110.5) |
| Range | 1.66–3.36 × 109/L | ||
| Erythrocytes | |||
| Relative level | 100% | 97.9% (96.0 to 99.7) | 99.7% (98.8–102.1) |
| Range | 4.74–5.88 × 1012/L | ||
| Hemoglobin | |||
| Relative level | 100% | 99.8% (103.3 to 100.9) | 100.5% (99.3–102.6) |
| Range | 8.5–11.0 mmol/L | ||
| Hematocrite | |||
| Relative level | 100% | 100.7% (98.8 to 102.7) | 104.2% (102.2–106.3) |
| Range | 0.41–0.50 vol frac | ||
| Mean cell volume | |||
| Mean ± SD | 85 ± 1 | 87 ± 1 | 88 ± 1 |
| Relative level | 100% | 102.9% (102.6 to 103.2) | 104.0% (103.7–104.4) |
| Range | 83–87 fL | ||
| Reticulocytes | |||
| Relative level | 100% | 93.9% (88.5 to 99.3) | 100.8% (97.0–104.6) |
| Range | 31–150 × 109/L | ||
Note: Microsamples from 10 random adult patients were stored at room temperature for 0, 4, and 7 h and measured in predilution (PD) mode. Results presented as relative levels with 95% confidence interval in parentheses. The absolute 0 h level also presented for each parameter. For individual samples, the maximum allowable deviation from 0 h values were the total analytical error whereas the maximum allowable error for the sample mean was the bias.
TABLE 4.
Effect of pneumatic tube transportation of adult capillary microsamples.
| Relative difference (PD/WB) | Tested range | |
|---|---|---|
| White blood cells | 4.1% (−2 to 10.1) | 2.9–8.2 × 109/L |
| Neutrophils | 2.5% (−6.3 to 11.2) | 1.15–5.25 × 109/L |
| Lymphocytes | 2.4% (6.0 to 10.8) | 0.38–2.29 × 109/L |
| Erythrocytes | 2.2% (−0.8 to 5.2) | 3.75–5.11 × 1012/L |
| Hemoglobin | 1.8% (−1.5 to 5.1) | 7.4–9.9 mmol/L |
| Hematocrite | −1.1% (−3.7 to 1.6) | 0.36–0.47 vol frac |
| Mean cell volume | −3.2% (−3.9 to −2.5) | 87–107 fL |
| Reticulocytes | −10.6% (−18.6 to −2.7) |
Note: The effect was evaluated in capillary microsamples from 10 random adult patients over 2 days. Paired venous blood samples and capillary microsamples were send by the pneumatic tube system and analyzed within 1 h of sample collection. The effect is expressed as mean predilution/whole blood (PD/WB) bias with 95% confidence interval in parentheses. The maximum allowable bias was estimated from the biological variation of each analyte.
3.2.3. Pediatric Samples
Twenty‐nine pediatric RBT samples were collected from random patients with an average age of 17 ± 23 months (range: 4–94 months). The bias between pediatric RBT tube and microsamples was WBC = 5.6% (95% CI: 4.2–7.1), neutrophils = 10.3% (95% CI: 7.5–13.0), lymphocytes = 1.6% (95% CI: 0.0–3.2), erythrocytes = 5.0% (95% CI: 4.0–6.0), hemoglobin = 4.3% (95% CI: 3.2–5.3), hematocrit = 3.5% (95% CI: 2.2–4.8), MCV = −1.5% (95% CI: −2.1 to −0.8), and reticulocytes = −5.3% (95% CI: −18.3 to 7.6). All these biases with their 95% CIs were within the 95% PIs estimated for adult samples (Figure 1).
4. Discussion
The present study verifies the PD mode for selected hematologic parameters on a Sysmex XN9000 analyzer. It demonstrates that the PD method has acceptable analytical performance for white cells and does not uncover critical preanalytical effects on these cells during the collection and processing of capillary microsamples. By contrast, red cell parameters exceeded the allowable biases and imprecisions and were susceptible to storage and transportation.
While the analytical performance for WBCs was generally acceptable, it exceeded the predefined maximum allowable limits at the lowest analyte levels. Increased imprecision at low levels is common, and the observed deviations are limited in absolute terms and have little clinical relevance. Furthermore, the observed imprecision and biases were smaller than those outlined in the Sysmex performance specifications [5]. For practical reasons, we performed our method comparison studies using microsamples prepared from venous blood samples rather than capillary sampling. This may have contributed to the high analytical performance, and differences between capillary and venous cell counts are well known [6, 7]. Therefore, subsequent experiments comparing capillary sampling with venous blood were performed in a small sample of subjects. This experiment confirmed an acceptable analytical performance of the PD mode also for capillary blood sampling. Finally, the WBC counts in capillary microsamples were resistant to 7 h of storage at room temperature and to transportation by the pneumatic tube system. This is in line with stability studies on plasma [8] and confirms the stability also in diluted samples of capillary blood. Collectively, our results demonstrate an acceptable analytical performance of the PD mode along with feasible preanalytical stability to allow its clinical use for measurement of leucocytes, neutrophils, and lymphocytes.
By contrast, the analytical performance of RBC parameters did not meet most of the biologically based performance specifications. Furthermore, all parameters except hematocrit were affected by transportation by the pneumatic tube system, and the stability of MCV and hematocrit decreased after 4 and 7 h. It may be argued that the benefits of reduced volume demand outweigh the reduction in analytical quality when using the PD method in the clinical treatment of neonatal patients. However, low‐volume methods with better performance characteristics for key parameters such as hemoglobin are available [9]. As such, the performance of the PD mode on the evaluated RBC parameters was inferior to established alternatives and therefore not currently feasible for clinical use.
We performed this study using adult blood for ethical reasons. Neonatal and adult blood may not be directly comparable as differences in cellular composition and rheologic properties may exist [10, 11]. To provide some indication of transferability of our results to pediatric patients, we performed studies on excess blood from pediatric RBT tubes. The WB measurement of the RBT blood was compared with PD measurements on microsamples prepared from the RBT blood. These results cannot be directly compared with our studies on adult blood because RBT blood is collected by capillary sampling. However, the bias and 95% CIs of the pediatric microsamples were within the 95% PIs established for the PD mode. This provides some indication that PD measurements of pediatric samples are not profoundly different from PD measurements of adult samples and thus do not pose special challenges.
The PD mode is an integrated channel on the Sysmex XN‐series that allows analysis of 1 + 6 prediluted blood [5]. The PD channel needs individual calibration and quality control for optimal performance. In our laboratory, this is done by daily comparison of each PD channel with the WB mode on the same instrument using selected patient samples. As such, maintaining the PD channel involves additional laboratory work. Moreover, the PD channel uses the general set of Extended rules, which may lead to issues not unveiled in our experiments. It remains unclear how, for example, rules ordering a smear will perform on prediluted samples, but this is beyond the scope of this report. Therefore, practical issues with the safe release of hematologic results from PD measurements need to be addressed before routine implementation.
The main limitation of this study is the use of adult rather than pediatric blood. While our studies did not uncover vast differences in pediatric samples, the risk of unforeseen peculiarities cannot be ruled out and caution should be exerted during initial use in neonatal patients. We did not attempt to establish a limit of detection for the parameters tested. The Sysmex performance specifications suggest linearity from zero for WBCs [5]. Our data suggest that reliable detection is not possible at very low levels and warrants the establishment of a detection limit or the setting of conservative arbitrary limits. We used 50 μL pipettes instead of the minimal 20 μL suggested in the Sysmex manual. This was done since it vastly improves handling and dissolving of the pipette content. It is therefore likely that a smaller pipette would tend to decrease analytical performance and increase preanalytical problems. We recommend that each laboratory develops its individual procedures. In this report, we evaluated selected parameters only. This was done for capacity reasons and the parameters judged most relevant were selected. The development of capillary microsampling is intended for routine monitoring and if advanced hematologic parameters would be needed in neonatal patients, ordinary volume blood sampling is advised.
Taken together, this report demonstrates that PD measurement of capillary microsamples can be an attractive alternative to large volume sampling for the monitoring of WBCs in neonatal patients. It demonstrates that the PD mode has sufficient analytical performance for measurement of leukocytes, neutrophils, and lymphocytes in routine use and uncovers that these parameters are robust to preanalytical handling and transportation.
Author Contributions
C.V.B.H. and A.R.H. concepted and designed the study. V.S.C. and K.R.J. acquired the data. C.V.B.H. and J.R.M. analyzed and interpreted the data. C.V.B.H. drafted the paper. All authors critically reviewed the draft, approved the final version, and accepts to be accountable for all aspects of the work.
Ethics Statement
The project is excepted from formal approval and has been presented to the local ethical committee for confirmation of this exception.
Consent
The authors have nothing to report.
Conflicts of Interest
The authors declare no conflicts of interest.
Hviid C. V. B., Christensen V. S., Jensen K. R., Møller J. R., and Hansen A. R., “Automated Red and White Blood Cell Counting in Capillary Microsamples by Sysmex‐XN Predilution Mode,” International Journal of Laboratory Hematology 47, no. 5 (2025): 808–816, 10.1111/ijlh.14478.
Funding: The authors received no specific funding for this work.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
References
- 1. Gunnink S. F., Vlug R., Fijnvandraat K., van der Bom J. G., Stanworth S. J., and Lopriore E., “Neonatal Thrombocytopenia: Etiology, Management and Outcome,” Expert Review of Hematology 7, no. 3 (2014): 387–395, 10.1586/17474086.2014.902301. [DOI] [PubMed] [Google Scholar]
- 2. Boesen C. V., Christensen V. S., Jensen K. R., Hansen A. R., and Hviid C. V. B., “An Automated Method for Thrombocyte Counting in Capillary Microsamples,” International Journal of Laboratory Hematology 46, no. 6 (2024): 1044–1051, 10.1111/ijlh.14354. [DOI] [PubMed] [Google Scholar]
- 3. The European Federation of Clinical and Laboratory Medicine (EFLM) , “EFLM Biological Variation Database,” accessed June 2023, https://biologicalvariation.eu/.
- 4. Westgard QC , “Desirable Biological Variation Specifications,” accessed June 2023, https://westgard.com.
- 5. Sysmex Corporation , XN Series Users Guide. Version 03/202. 2017‐2020 (Sysmex Corporation, 2020). [Google Scholar]
- 6. Hollis V. S., Holloway J. A., Harris S., Spencer D., van Berkel C., and Morgan H., “Comparison of Venous and Capillary Differential Leukocyte Counts Using a Standard Hematology Analyzer and a Novel Microfluidic Impedance Cytometer,” PLoS One 7, no. 9 (2012): e43702, 10.1371/journal.pone.0043702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Schalk E., Heim M. U., Koenigsmann M., and Jentsch‐Ullrich K., “Use of Capillary Blood Count Parameters in Adults,” Vox Sanguinis 93, no. 4 (2007): 348–353, 10.1111/j.1423-0410.2007.00978.x. [DOI] [PubMed] [Google Scholar]
- 8. Doeleman M. J. H., Esseveld A., Huisman A., de Roock S., and Tiel Groenestege W. M., “Stability and Comparison of Complete Blood Count Parameters Between Capillary and Venous Blood Samples,” International Journal of Laboratory Hematology 45, no. 5 (2023): 659–667, 10.1111/ijlh.14080. [DOI] [PubMed] [Google Scholar]
- 9. García‐Payá M. E., Fernández‐González M., Morote N. V., Cobos Á. B., Díaz M. C. T., and Hernández J. F. S., “Performance of the Radiometer ABL‐90 FLEX Blood Gas Analyzer Compared With Similar Analyzers,” Laboratory Medicine 44, no. 1 (2013): e52–e61, 10.1309/lm7boxdxqnq0fdul. [DOI] [Google Scholar]
- 10. Prabhu S. B., Rathore D. K., Nair D., et al., “Comparison of Human Neonatal and Adult Blood Leukocyte Subset Composition Phenotypes,” PLoS One 11, no. 9 (2016): e0162242, 10.1371/journal.pone.0162242. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Rampling M. W., Whittingstall P., Martin G., et al., “A Comparison of the Rheologic Properties of Neonatal and Adult Blood,” Pediatric Research 25, no. 5 (1989): 457–460, 10.1203/00006450-198905000-00006. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.