0.99). Comparison of results obtained with the RA-1000. (y) and the colorimetric method. (x) of Glynn and Kendal. (1), corrected for salicylate. (2), for. 23 patients'.
mented serum samples in the range 0-900 mg/L (0-5.95 mmolIL) was 101% (range 94-104%). Precision of 20 determinations at 200 mg/L (1.3 mmol/L) was 0.8% within-batch and 2.6% between-batch, and at 350 mg/L (2.31 mmolfL) was 1.6% and 3.3%, respectively. Regression analysis of the data produced by 18 different control sara with known acetaminophen concentrations gave the equation y = 1 .034x 0.361 mg/L (y RA-1000 result, x = weighed-in acetaminophen value, and r = 0.99). Comparison of results obtained with the RA-1000 (y) and the colorimetric method (x) of Glynn and Kendal (1), corrected for salicylate (2), for 23 patients’ samples gave the regression equation: y = 1.109x 27.7 mg/L (r = 0.99). Settings for the RA-1000 were: -
=
-
0.0 900 200 1.0
2
Range
6
Range
Filter
6
Std.
Delay
3 00 70 19
Slope
2 30 7
2nd reagent delay Cal. factor Reagent blank
lpe % sample
vol
% reagent vol 2nd reagent vol A2 delay Units Unit factor
RBL RBL
low high value
0.0 0.02
Intercept Ep. limit
1.0
3 00
1. Calcd. by the J instrument
0.0 0.5
low high
References 1. Glynn JP,
Kendal
SE.
Paracetainol
Lancet
measurement.
1975;i:1147-8.
2. Dinwoodie AJ. Interference in an emergency acetaminophen method and its modification. Clin Biochem 1978;11:131-2.
High-Dose
“Hook”
Somatotropin
Effect in Measurement
of
by Two-Site immunoradiometrlc
P. Garcia-Webb, F. E. Watson, of Clin. Biochem., University Nedlands WA 6009, Australia)
and
Nola
of Western
Assay,
Whiteside
(Dept.
Australia,
We measured a somatotropin concentration of 45 nulliunmts/L in serum from an acromegalic patient recently transferred from another hospital. The previous day their result had been 1400. On dilution, our sample gave a dilution-corrected result of 1110 milli-int. unitsfL. Because we use a two-site immunoradiometric assay (hGH RIA 100; Pharmacia, Uppsala, Sweden), whereas the other hospital uses an in-house double-antibody radioimmunoassay, we mt.
considered the “hook” effect. We extended
possibility
of the
error’s
being
due to the
the standard curve (usually 0.5-150 milliat our institution) to 3500 milli-int. unitsfL by diluting a wto somatotropin standard (66/217) in kit sample diluent. The plot of somatotropin concentration vs bound counts divided by total counts (BIT) was curvilinear. BIT increased with somatotropin concentrations from 0 to 100 milli-int. unitsfL, was approximately flat from 100 to 400, and decreased when concentrations exceeded 400 milli-int. units/L. Consequently, BIT values of 25-40% corresponded to somatotropin concentrations of 25-85 or 500-1500 milliint. unitsfL. For somatotropin concentrations of 100-400 milli-int. unitsfL, a small change in counts bound caused a
int. units/L
large
2102
difference
CLINICAL
in the
apparent
CHEMISTRY,
result.
Vol. 32, No. 11, 1986
Results of immunoradiometric assays (antibody is in excess) are read from the ascending part of the curve. A large increase in antigen alters this antibody excess and
results
in a curve
that decreases as antigen concentration effect, e.g., in measurements of ferritin and prolactin in serum. Perhaps because results of >1000 milli-int. unitsfL in acromegaly are rare, the Pharmacia kit insert does not mention that the “hook” effect might occur. To investigate acromegaly, we now measure the sample twice: undiluted and diluted twofold. A result of
