Complete blood count (CBC) and leukocyte differential count (LDC) are the most basic and widely used tests for the diagnosis, treatment, and monitoring of various diseases, including hematological disorders. Various automated hematology analyzers capable of reporting quick and accurate results have been developed and used in clinical practice [1]. The DxH 900 (Beckman Coulter Inc., Brea, CA, USA) is the most recently developed quantitative, multiparameter automated hematology analyzer for in vitro diagnostic use and screening of patient populations in clinical laboratories. It identifies and enumerates 29 parameters, including monocyte distribution width (MDW), which is a novel marker for the early diagnosis of sepsis [2-4]. The DxH 900 analyzer functions are based on the Coulter Principle, which accurately counts and measures the sizes of cells by measuring changes in electrical resistance in white blood cells (WBC), red blood cells (RBC), and platelets in a conductive liquid passed through a small aperture. The measurement of nucleated red blood cells (NRBC), reticulocyte percentage, MDW, and differential analysis of WBC occur in the VCS module, suggesting that the measurements for volume, conductivity, and light scatter are collected at multiple angles. In this study, we evaluated the analytical performance of the parameters of the DxH 900.

1. Sample collection

This study was approved by the Institutional Review Board of Inje University Haeundae Paik Hospital, Busan (approval number: HPIRB 2018-12-003), and was conducted in accordance with the Declaration of Helsinki. The requirement for informed consent was waived owing to the anonymity of the collected data and the use of residual samples. Residual samples from patients with both normal and abnormal hematological profiles referred to the central laboratory were used for comparison. Samples were collected in K2-EDTA tubes (Becton Dickinson, Plymouth, UK) from April to May 2019, stored at room temperature (20–26℃), and analyzed by a dedicated technologist within 2 hours of collection. Commercial controls (Beckman Coulter Inc.) were included in the samples to ensure precision, carryover, and linearity.

2. Analyzer

The DxH 900 analyzer uses photometric measurements to measure hemoglobin (Hb) levels. The mean cell volume, red cell distribution width and standard deviation (SD) are derived from the RBC histogram, and the mean platelet volume is derived from the platelet histogram. Other parameters such as hematocrit (Hct), mean corpuscular hemoglobin (MCH), MCH concentration (MCHC), immature reticulocyte fraction, mean reticulocyte volume, and absolute counts of neutrophils, lymphocytes, monocytes, eosinophils, basophils, NRBC, and reticulocytes are also calculated.

3. Precision

Within-run and between-run precision analyses for the DxH 900 analyzer were performed using normal and abnormal high- and low-level quality control (QC) materials (6C Plus Control level 1–3). Within-run precision was analyzed based on 20 measurements, and between-run precision analysis was performed once daily for 10 consecutive days. The coefficient of variation (CV) was expressed as a percentage and was calculated from the mean and SD for each level of the QC material, which was compared with the manufacturer’s protocol and current state-of-the-art precision limits [5].

4. Carryover

Carryover was measured for WBC, RBC, Hb, and platelets using the Coulter LIN-X Linearity Control (Beckman Coulter) according to the Clinical and Laboratory Standards Institute (CLSI) H26-A2 guidelines [6]. Carryover was assessed by analyzing high-level samples 3 times (H1, H2, and H3), followed by three consecutive analyses of low-level samples (L1, L2, and L3). Carryover was calculated using the following formula:

Carryover (%)=(L1–L3)/(H3–L3)×100.

5. Linearity

Linearity analysis for WBC, RBC, Hb, and platelets was performed using Coulter LIN-X Linearity Control in duplicate, according to the CLSI H26-A2 guidelines [6]. The Pearson’s correlation coefficients (r) between the expected and measured values were calculated.

6. Comparison

The correlation between the measurements obtained from the DxH 900 analyzer were evaluated with the measurements obtained from the XE-2100 (Sysmex Corp., Kobe, Japan), which was in use in the laboratory. Seventy samples for measurements within the analytical measurement range, according to the CLSI EP09C-ED3 guidelines [7]. The correlation coefficient (r), slope, and intercept were calculated for WBC, RBC, Hb, Hct, platelets, and 5-part LDC.

1. Precision

The DxH 900 analyzer demonstrated good precision for the measurement of normal and abnormal high- and low-level QC materials for all parameters except basophils (Table 1). Despite the lack of current state-of-the-art limits for MDW, all the measured %CVs for MDW were <3%, which was much lower than the manufacturer’s limit of 10%.

Table 1 . Within-run and between-run precision analyses for normal and abnormal high-level and low-level quality control material.

Variable	Normal control			Abnormal high-level control			Abnormal low-level control		Manufacturer’s limit, CV (%)	State-of-the-art, CV (%)
	Mean	CV (%)		Mean	CV (%)		Mean	CV (%)
Within-run
WBC (×109/L)	8.7	0.97		21.1	0.91		3.2	1.21	3	2.5/1.5/2.5
Neutrophils (×109/L)	4.78	1.49		13.61	1.59		1.45	2.09	3.5	2.5
Eosinophils (×109/L)	0.63	8.36		1.12	8.4		0.14	7.64	13.5	10
Basophils (×109/L)	0.003	39.19		0.015	40.64		0.001	56.42	-	20
Lymphocytes (×109/L)	2.49	2.48		3.15	3.46		1.37	1.1	5	3.5
Monocytes (×109/L)	0.77	3.71		3.16	2.88		0.29	4.38	10	8.5
RBC (×1012/L)	3.96	1.3		5.26	0.51		1.77	0.5	1.5	1.1
Hb (g/dL)	11.7	0.29		15.4	0.35		4.7	0.37	1.5	0.9
MCV (fL)	88.03	0.21		89.05	0.19		80.99	0.17	1	0.6
RDW (%)	15.4	0.53		15.97	0.74		15.25	0.97	2.5	2.0
Platelets (×109/L)	206	1.21		402	1.88		69	2.15	3.5	3.0
MPV (fL)	9.76	0.82		10.08	0.57		9.46	0.55	2.5	2.5
MDW (units)	37.97	2.3		37.44	2.44		36.73	2.58	10	-
Between-run
WBC (×109/L)	8.8	1.3		21.3	1.05		3.3	1.28	3	2.5/1.5/2.5
Neutrophils (×109/L)	4.9	1.9		13.8	1.51		1.5	1.91	3.5	2.5
Eosinophils (×109/L)	0.52	5.22		1.15	5.28		0.14	6.55	13.5	10
Basophils (×109/L)	0.005	98.07		0.019	37.7		0.001	148.79	-	20
Lymphocytes (×109/L)	2.5	2.53		3.1	3.49		1.4	2.28	5	3.5
Monocytes (×109/L)	0.79	4.42		3.3	2.95		0.29	4.31	10	8.5
RBC (×1012/L)	3.94	0.86		5.32	1.09		1.79	1	1.5	1.1
Hb (g/dL)	11.8	0.67		15.6	0.51		4.8	0.35	1.5	1.0
MCV (fL)	88.05	0.32		89.07	0.28		81.01	0.31	1	0.8
RDW (%)	15.44	1.04		16.03	1.04		15.34	1.18	2.5	2.0
Platelets (×109/L)	208	1.91		404	1.49		70	1.87	3.5	3.0
MPV (fL)	9.79	0.56		10.07	0.59		9.47	0.66	2.5	2.5
MDW (units)	37	2.76		38.3	2.45		36.7	2.41	10	-

Abbreviations: CV, coefficient of variation; WBC, white blood cells; RBC, red blood cells; Hb, hemoglobin; MCV, mean corpuscular volume; MDW, monocyte distribution width; MPV, mean platelet volume; RDW, red cell distribution width..

2. Carryover and linearity

The percentage of carryover was –0.08% for WBC, 0.00% for RBC, 0.00% for Hb, and –0.04% for platelets. These carryover values were not clinically significant. However, the values were within the manufacturer provided limit of 0.05% for WBC and RBC, and 1.0% for Hb and platelets. The r values for WBC, RBC, Hb, and platelet counts were close to 1.0, indicating excellent linearity (Table 2).

Table 2 . Linearity of WBC, RBC, Hb, and platelet counts obtained using the DxH 900.

Variable	Range	Correlation coefficient (r)	Intercept	Slope
WBC (×109/L)	0.25–372.90	0.9999	–0.1453	0.9986
RBC (×1012/L)	0.03–8.24	1	–0.01	0.9973
Hb (g/dL)	0.2–24.7	0.9999	–0.0751	0.9946
Platelets (×109/L)	3–2,655	0.9997	5.269	0.9902

Abbreviations: WBC, white blood cells; RBC, red blood cells; Hb, hemoglobin..

3. Comparison

The results of the comparison between two hematology analyzers are shown in Table 3. Fig. 1 shows the correlations between the measurements performed using the DxH 900 and XE-2100 analyzers. The results obtained for WBC, RBC, Hb, Hct, platelet, and 5-part LDC measurements from the DxH 900 analyzer were comparable to measurements obtained from the XE-2100 analyzer. All parameters, except the percentage of basophils (r=0.553) and monocytes (r=0.932), showed a strong correlation, with r>0.95. The DxH 900 provided higher values than the XE-2100 for these two parameters. The mean (95% confidential interval) of measurement bias from the DxH 900 over the XE-2100 represented by Bland–Altman difference plots was –3.9% (–4.5 to –3.3) for WBC, –3.2% (–3.5 to –2.9) for RBC, –3.0% (–3.2 to –2.8) for Hb, –2.4% (–2.9 to –1.9) for Hct, –0.1% (–1.5 to 1.3) for platelets, 1.1% (0.6 to 1.6) for neutrophils, –0.5% (–5.3 to 4.3) for eosinophils, 147.1% (111.7 to 182.5) for basophils, –5.1% (–6.9 to –3.3) for lymphocytes, and 5.8% (2.1 to 9.5) for monocytes.

Table 3 . Comparison of major complete blood count parameters between DxH 900 and XE-2100 (N=70).

Variable	Range		Correlation coefficient (r)	Slope	Intercept
	DxH 900	XE-2100
WBC (×109/L)	2.7–20.1	2.76–20.34	0.998	0.9652	–0.0301
RBC (×1012/L)	1.7–5.85	1.81–5.99	0.998	0.9896	–0.0843
Hb (g/dL)	5.5–17.8	5.8–18.7	0.999	0.9443	0.3159
Hct (%)	16.9–53.9	18.1–54.8	0.995	0.9968	–0.7592
Platelets (×109/L)	23–692	22–641	0.989	1.0264	–6.7465
Neutrophils (%)	25.1–88.5	24.5–87.9	0.995	0.9947	0.9373
Eosinophils (%)	0–3.4	0–3.7	0.983	0.956	0.0227
Basophils (%)	0.2–0.8	0–1.3	0.553	0.6317	0.4518
Lymphocytes (%)	2.5–55.8	3.1–59.4	0.993	0.9808	–0.682
Monocytes (%)	2.5–23.9	3.4–25.9	0.932	0.9444	0.8861

Abbreviations: WBC, white blood cells; RBC, red blood cells; Hb, hemoglobin; Hct, hematocrit..

Figure 1. Correlations between the values of white blood cells (WBC) (A), red blood cells (RBC) (B), hemoglobin (Hb) (C), hematocrit (Hct) (D), platelets (E), and 5-part WBC differential results (F–J) estimated using the UniCel DxH 900 (y axis) and XE-2100 (x axis). The R value represents Pearson’s correlation coefficient.

In this study, the analytical performance of the DxH 900 analyzer was evaluated. The precision (%CV) of all parameters, except for basophils, was less than the limits provided by the manufacturer and current state-of-the-art. The %CV for basophils was between 37.70%–148.79% with normal, high-level, and low-level QC material in within-run and between-run precision analyses, and exceeded the 20% limit considered as state-of-the-art currently. This could be due to the very low concentration of basophils in the QC material, which was cross-checked by reanalyzing using other automated hematology analyzers [8,9].

MDW is a relatively new laboratory parameter that has been studied as a potential early indicator of sepsis in emergency departments (ED) [2,4]. The DxH 900 can measure MDW as part of the standard CBC without requiring extra time. So MDW can provide quick results to aid prompt decision-making in the ED. In addition, all %CV values for MDW in this study were less than 3%, demonstrating good precision, which is similar to the values of 1.99%–3.54% reported by Martinez-Iribarren et al. [10]. Therefore, MDW can be a useful biomarker for the early detection of sepsis in the ED.

The DxH 900 analyzer showed no significant carryover effect or linearity bias in WBC, RBC, Hb, or platelet counts. Martinez-Iribarren et al. [10] reported that the DxH 900 displayed good linearity for low WBC and platelet counts. These results could be obtained in the absence of using specific modes for low WBC counts as required by other automated hematology analyzers.

The correlation between the DxH 900 and XE-2100 CBC, and 5-part LDC values was high, except for the percentage of basophils (r=0.553). The measurements for basophils were obtained by most automated hematology analyzers were poorly correlated, which is attributable to the low proportion of basophils in the total WBC count [8,10-12].

There are some limitations for this study. First, there were no specific exclusion criteria for specimens with an abnormal hematological profile. Second, we did not perform a microscopic blood film examination and 5-part LDC in our samples.

In conclusion, our findings show that the DxH 900 analyzer has good analytical performance and good agreement with Sysmex XE-2100 results. The DxH 900 analyzer also showed good analytical performance in measuring MDW, a novel parameter for the early diagnosis of sepsis.