Comparative assssment of three commercially available methods for measurement of serum apolipoproteins A-1 and B

ABSTRACT

The aim of the study was to establish a reliable and cost-effective method of assessing apolipoprotein levels in a Kuwaiti teaching hospital Lipid Clinic population. Plasma levels of apo A-I and B were assayed in 88 subjects by each of the following 3 immunological methods: [1] Behring's nephelometry with dedicated reagents; [2] turbidimetry on a Hitachi 911 Autoanalyser using reagents from Boehringer; and [3] manual turbidimetry with Randox reagents. For both apo A-I and B, the levels obtained were consistently highest for method 1, lowest for method 3, and intermediate for method 2. With regards to apo A-I, the differences were statistically significant for 2 and 3 vs. I [p <0.001], but not between 2 and 3. For apo B, values obtained for methods I and 2 were similar statistically but significantly higher than those from method 3 [p < 0.01]. The correlation between values for apo A-I and apo B obtained by the three methods demonstrated a similar trend for apo A-I, the correlation coefficients were greatest [r 0.75] for method 1/2, and less for 1/3 [r 0.64] and 2/3 [r 0.62]; for apo B, the correlation for 1/2 [r 0.74] was greater than for 1/3 [r 0.43] and 2/3 [r 0.69] [all p <0.001]. Altman and Bland plots indicated -20% bias for apo A-I and apo B between methods I and 2, much smaller than the -20%->100% range for 1/3 and 2/3 for both apoproteins. Similarly concordance for diagnostic values of apo A-I [<1.20 g/l] and apo B [>1.20 g/l] were best for methods 1 and 2. In economic terms, assays for either apo A-I or apo B cost about $3.00 for methods 1 and 2, with a similar technician time/sample of about 2 minutes; for 3, the cost is less [about $2.25], but technician time/sample was about doubled. The study suggests that measured apo A-I and B levels may differ depending on the type of commercial kits used, making it imperative for each laboratory to establish its own reference range. Caution should be exercised in inter-laboratory comparisons of data on apo A-I and B, particularly as applied to patient care