The question you pose is one that comes up often in clinical practice. There are clearly circumstances in which the skin test and serum IgE level do not "agree" with one another, and in these instances, one must make a clinical judgement on the risk of challenge in conjunction with the patient's/parents' wishes.
First, there is not a wealth of knowledge on the value of peanut butter skin testing, and I can't offer you any personal experience as to its validity. Below is one article that suggests that the specificity is 0.82, which is pretty good. However, newer data on component testing suggest that the specificity of these tests may even better than that (1).
In this case, I would suggest re-sending the peanut serum testing (to whole peanut and components). It's possible that this single result was a lab error (a distinct possibility given the history and skin test results). If the peanut and component serum testing again come back negative, I would offer the challenge, with the caveat that there is still a risk of reaction, as would not dismiss the skin test results.
Jay A. Lieberman, Susanne Glaumann, Sofia Batelson, Magnus P. Borres, Hugh A. Sampson, Caroline Nilsson
The Utility of Peanut Components in the Diagnosis of IgE-Mediated Peanut Allergy Among Distinct Populations
The Journal of Allergy and Clinical Immunology: In Practice, Vol. 1, Issue 1, p75–82, 2013
Pediatr Allergy Immunol. 2007 May;18(3):231-9.
Combining skin prick, immediate skin application and specific-IgE testing in the diagnosis of peanut allergy in children.
Wainstein BK1, Yee A, Jelley D, Ziegler M, Ziegler JB.
Previous studies have suggested various diagnostic cut-offs of allergy tests for the diagnosis of clinical peanut allergy in children. There are few data relating to the use of combinations of these tests in children. We aimed to determine the validity of previously reported diagnostic cut-off levels ofpeanut allergen skin tests and peanut specific-immunoglobulin (Ig) E, as well as the usefulness of combinations of these, for predicting clinical peanutallergy in our Allergy Clinic. Children attending the Allergy Clinic with a positive peanut skin prick test (SPT; n = 84) were included in the study. Immediate skin application food tests (I-SAFT) using 1 g of peanut butter (positive if any wheals were detected at 15 min), peanut specific-IgE levels and open-label peanut food challenges were performed. Fifty-two of 85 peanut challenges were positive. Skin prick test specificity was 67% at >or=8 mm and 100% at >or=15 mm. The I-SAFT was 82% specific. A peanut specific-IgE level of 0.37 kU/l was 98% sensitive but 33% specific. A level of 10 kU/l was 100% specific. Combinations of a SPT of >or=8 mm with a positive I-SAFT and a peanut specific-IgE >or=0.37 kU/l were 88% specific with a sensitivity of 38%. Using challenge outcomes as the standard, available in vitro and in vivo diagnostic tests for peanut allergy have poor sensitivity and specificity and combining them does not significantly improve their clinical usefulness. Previously described diagnostic cut-off levels do not have general applicability. Allergy practitioners may need to interpret results of allergy tests in the context of their own practices.
Phil Lieberman MD