No abstract

SUMMARYThe results of an Interregional quality assurance scheme for tests in the diagnosis of megaloblastic anaemia were reviewed to assess the methods used. Serum folate assays showed great variation between methods, partly due to limitations in assessment by external quality assurance. Red cell folate assays yielded widely different results and much imprecision due both to the differences in preparation of the haemolysate and to the problems inherent in radioassay of a mixture of folate compounds. Intrinsic factor antibody tests showed appreciable variation in sensitivity. There was considerable inconsistence in the detection of polymorph nuclear hypersegmentation.Regional trials in haematology provide laboratories with external quality assurance. We have organised a scheme for vitamin assays for some years,' adding two other tests more recently. The number of participants, however, permits not only an assessment of individual performance but also of the methods in use. The results of the trial give an indication of current practice. TRIALS Vitamin assaysTwo samples are issued 11 times yearly for serum vitamin B12 and folate assays with a whole blood specimen in CPD-adenine for red cell folate assays. An arbitrary haematocrit of 0 45 and serum folate of 1-0 gg/l were given with the whole blood sample. A secondary liquid reference serum, standardised against the British standard for human serum vitamin B12, assigned a value of 355 (338-372) ng/l by our radioisotopic non-commercial assay, was distributed as an unknown serum in a recent trial.

SUMMARY Five commercial kits for the determination of folate and six kits for the determination of vitamin B12 were investigated. Their performance has been compared with microbiological methods for the two vitamins and with a non-commercial radioisotopic method for B12. The results show the importance of the determination of the reference range for an individual laboratory for each method. The precision of the kits varied appreciably, as did their performance using specimens from patients with different haematological disorders. In particular, certain kits failed to detect all patients with pernicious anaemia. The relative accuracy of the kits was assessed. Various factors which should be taken into account in the final selection of a satisfactory kit are discussed.

SUMMARY The sensitivity of methods to detect antibodies to intrinsic factor was assessed. Five sera of known antibody content were tested in 31 laboratories and 30 sera from patients with pernicious anaemia were tested in one laboratory. Five non-commercial methods and two kits for type I antibodies and one non-commercial method for types I and II antibodies are in current use. Differences in sensitivity of the non-commercial methods for type I antibodies related more to the antigen:antibody ratio in the test system than to the method itself. A radioimmune assay for types I and II antibodies showed the best sensitivity but that of an enzyme linked immunosorbent assay (ELISA) method was poor Type I antibodies (IFA) to intrinsic factor, which block the binding of cobalamin to intrinsic factor, have been recorded in the 31% to 76% ofpatients with pernicious anaemia.' This large variation may be due to case selection or to differences in technique used, or both. Because the detection ofIFA may avoid the need for further investigation in the diagnosis of pernicious anaemia we compared the sensitivity of some of the methods in current use, including two which detect type II antibodies (inhibitors of the attachment of intrinsic factor to the ileal mucosa) at the same time. Tests for type II antibodies alone are not often used because the antibodies are thought to be less common than type I antibodies, though this may not be so.2 Material and methodsSera were collected from patients with pernicious anaemia and from a healthy control. These were coded and sent to laboratories which undertake IFA testing and which had agreed to test samples for us. Two sera from patients with pernicious anaemia containing, respectively, 6 and 2 units IFA/ml were issued first and on a second occasion, patients' serum containing 10 units IFA/ml together with this serum diluted by a factor of 8 and 10 in the normal serum, and the normal serum were sent.

Fourteen commercial kits for serum vitamin B12 assay have been assessed. The analytical performance and clinical correlation were used to rank the methods employed for the extraction of the vitamin from its binders and for the separation of 'bound' from 'unbound' counts. The introduction of extraction without boiling and the separation by attachment of the vitamin B12 binder to a solid matrix have not been shown to be generally better or worse than conventional methods.