| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
|
|
|
|||||||
|
|||||||||
Electronic Letters to:
|
Electronic letters published:
![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
A correction
|
|
|||
|
Hugh Mitchell, Principal Biochemist Charing Cross Hospital
Send letter to journal:
h.mitchell{at}ic.ac.uk Hugh Mitchell
|
Dear Sir, As a co-author on this paper, it concerns me to find that there have been changes and additions made to the draft I had signed. Particularly, I fundamentally disagree with the conclusion in the paper's last paragraph as I believe that competitive assays are needed for hCG testing of patients with trophoblastic disease. At this hospital, which is a trophoblastic disease centre for the UK, we use a rabbit polyclonal in-house automated RIA which has a cross-reaction of <0.25% with LH and between-assay CV of 10%. Using the same antiserum since 1979 we have assayed more than 2.3 million samples and the assay has been used in the treatment of some 2000 patients with gestational trophoblastic disease and some 20,000 patients with hydatidiform mole. I do not share the first author's opinion about the "negative aspects" of RIA, although I will concede that monthly iodination of the labelled hCG is inconvenient. I believe that the work originally published by Dr Cole (1) was essentially correct and that only assays based on a one-site competitive principle are able to measure all of the forms of hCG produced in disease. Two-site sandwich assays are too specific, and in my opinion (2) they all (including DPC) have "blind spots" which miss the detection of some of the variants found in cancer (data supporting this will be published shortly). In my opinion, manufacturers should be persuaded to develop competitive assays using modern tracers in order to provide the broad-spectrum hCG assays necessary for oncology. At the moment we have a very unsatisfactory situation in that manufacturers cannot endorse their hCG assays for anything other than pregnancy, knowing full well that most customers are using them for oncology. Since table 5 consists of samples which are said to be falsely positive on the AxSYM assay, Dr Cole makes the assumption that the three raised RIA results are likewise falsely elevated, and thereby that the DPC assay is superior to RIA. Only one of these three samples was actually tested on the DPC assay, and one of these was also positive on one other immunometric assay (excluding Abbott). The negative conclusions about RIA on this basis are, I suggest, unfounded. No two RIA's are the same and to generalise about them on the basis of results from one is as untenable as it would be to generalise about immunometric assays on the basis of one such assay. In addition, the processes which lead to false positives in sandwich assays need not necessarily apply to competitive assays. Further illustration of the differing ability of sandwich assays to detect hCG in disease is shown by the last sample in table 4. The assays detect from as little as 3.2% (ACS) to 50.5% (Baxter) of the RIA value, indicating the broad range of hCG variants which may be produced. References 1. Cole LA, Kohorn EI, Kim GS. Detecting and monitoring trophoblastic disease. J Reprod Med 1994; 39: 193-200. 2. Hugh Mitchell. Analysis of hCG: clinical applications and assay requirements [letter]. Ann Clin Biochem. 1999; 36 ( Pt 2) Hugh Mitchell (Principal Biochemist) Department of Medical Oncology, Charing Cross Hospital, LONDON W6 8RF, UK. h.mitchell@ic.ac.uk |
|||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |