Unreliability of local reactions to serum as tests for general sensitivity.

The problem of sensitivity tests for serum can be stated to be as follows: Is it possible, by observing the local reaction which follows an intradermal or subcutaneous injection of a small amount of serum, to determine beforehand whether or not a patient will develop anaphylactic shock, and/or severe serum sickness, when given a full therapeutic or prophylactic dose? To be of practical value, the local reaction should show within at least one hour of the test dose, otherwise the doctor would have to postpone the main dose far too long. We have long maintained that local reactions are unreliable guides of general sensitivity (Laurent and Parish, 1952). A recent paper, however, entitled "An Analysis of Tetanus Prophylaxis in 3,455 Cases" by Binns (1961), though it is a useful collection of data, contains some statements and conclusions in opposition to our views, and they should not be permitted to pass unchallenged. The author claims, on the basis of 9 months' experience at the Leeds Public Dispensary, that the local reaction which follows a subcutaneous injection of 0-1 ml. serum is "important and that due cognizance should be taken of it". He has found a "relationship" between the local reaction and the general reaction, and he implies-though he does not say so-that tests for a local reaction could be of value in detecting the presence of general serum-sensitivity. According to Binns's own figures, 73 patients had a local reaction within an hour, and of these only three developed either anaphylactic shock or serious serum sickness; presumably the other seventy had average or mild serum sickness or had none at all. Local reaction appeared within 4 days in 133 patients, and of these eight developed anaphylactic shock or serious serum sickness; presumably the other 125 had average or mild serum sickness or had none at all. It follows, therefore, that when the tests are read after one hour, a positive result signifies hypersensitivity in only 4 per cent.; when read after 4 days, it signifies hypersensitivity in only 6 per cent. Binns remarks that "the timing of the lQcal reaction seemed to make little difference to the risk of developing a general reaction subsequently". Our comment on this statement is that time is necessarily at the crux of the problem. How can the late appearance of a local reaction (say after 24 hours or more) be helpful as a guide to procedure …