The results of skin-prick tests to four concentrations of venom (0.1, 1, 10 and 1000 μg/ml) carried out on two occasions were analysed in relation to the history of adverse reactions to stings and to the level of venom-specific IgE antibody in serum, in forty-two subjects allergic to insect stings (sixteen to bee and twenty-six to wasp). Fifty control subjects (some of whom had never been stung by bee or wasp) with no history of adverse reaction to stings were also studied. No subject gave a positive skin-test reaction to 0.1 μg/ml, and small numbers reacted to either 1 or 10 μg/ml. The lowest concentration of venom to which most subjects had a positive skin test was 100 μg/ml. Our data suggest that in wasp-allergic patients a positive skin test to 100 μg/ml is normally significant (reflecting the presence of specific IgE), whereas in bee-allergic patients a skin test reaction to 100 μg/ml is usually non-specific for the following reasons.
(i) In the allergic patients when skin tests were repeated, a reaction to 100 μg/ml bee venom often became negative (in six of eight), whereas for wasp venom the reaction became more positive (at 10 or 1 μg/ml) in seven of eight patients. Whilst this might reflect lack of reproducibility, the consistent direction of change for either bee or wasp venom suggests responses to this concentration of these venoms may have different interpretations.
(ii) In bee-allergic patients, where positive skin tests to 100 μg/ml wasp venom were found they were repeatable in all patients, and occurred only in patients who had been stung by a wasp. In wasp-allergic patients, skin test reactions to 100 μg/ml bee venom were not reproducible between studies in any patient and often occurred in subjects never stung by a bee.
(iii) A better correlation between skin test and RAST occurred for wasp venom when a skin-test reaction to 100 μg/ml was included as a positive (rather than reactions to 1 or 10 μg/ml only). For bee venom the correlation did not improve when skin-test reactions to 100 μg/ml were included.
(iv) In the control group, skin-test reactions to 100 μg/ml bee venom were often false positives (seven of eight had never been stung by a bee). Four out of fifty controls reacted to wasp venom 100 μg/ml, but three of these had received a wasp sting.
These findings suggest that in routine clinical practice skin tests should be carried out over the range 1–10 μg/ml in bee-allergic subjects but in wasp-allergic patients 100 μg/ml should also be included. Extracts of 0.1 μg/ml need not be used.