Comparison of a new enzymatic assay for serum homocysteine on Toshiba TBA-c16000 against an immunoassay on Abbott Architect

Figures & data

Table 1. Precisions of Hcy assays.

Level (µmol/L)	Mean (µmol/L)	SD (µmol/L)	Repeatability, CV (%)	Total imprecision, CV (%)
BSBE enzymatic assay
Level 1 (12)	13.09	0.42	3.00	3.20
Level 2 (29.5)	34.59	0.54	1.50	1.60
Architect immunoassay
Level 1 (8.18)	8.07	0.43	4.50	5.40
Level 2 (14.47)	13.86	0.67	3.80	5.00
Level 3 (28.43)	27.88	1.09	3.90	3.90

Figure 1. Comparison of Architect immunoassay and BSBE enzymatic assay by Passing–Bablok regression analysis (A) and Bland–Altman analysis (B).

Figure 2. Linearity test on five different levels of Hcy were measure by BSBE enzymatic assay. The linearity of the assay was 0.5–53.3 µmol/L.

Table 2. Recovery rate.

Observed Hcy concentration (µmol/L)	Expected Hcy concentration (µmol/L)	Recovery (%)	Difference (%)	Bias (%)	Result
53.3	53.3	100	0.0	0.0	Acceptable
40.1	40.5	100.9	0.4	0.9	Acceptable
26.9	27.5	102.2	0.6	2.2	Acceptable
13.7	12.8	93.6	−0.9	−6.4	Acceptable
1.2	1.2	100	0.0	0.0	Acceptable

Acceptable range: recovery (%) ≤ ±10% or R² ≥ 0.975.

Table 3. Accuracy test results from the insert provided by the manufacturer.

	Test 1	Test 2	Test 3
NIST SRM 1950 (8.5 µmol/L)	8.66	8.58	8.57
	8.63	8.60	8.62
	8.58	8.61	8.59
Average (µmol/L)	8.62	8.60	8.59
Bias (%)	1.45	1.14	1.10
Hcy control 1 (12.3 µmol/L)	12.61	12.80	12.76
	12.76	12.60	12.71
	12.65	12.92	12.86
Average (µmol/L)	12.67	12.77	12.78
Bias (%)	3.04	3.85	3.88
Hcy control 2 (31.5 µmol/L)	31.71	31.95	32.21
	31.40	31.79	32.26
	31.33	32.04	32.53
Average (µmol/L)	31.48	31.93	32.33
Bias (%)	−0.06	1.35	2.65

Acceptance criteria was bias ≤ ±10%.