How allergic are “allergic shiners”?

To the Editor:

Chen et al1 are to be commended for their careful study of so-called “allergic shiners.” This discoloration of the skin below the eyes is often taken to be evidence that the patient, especially a child, must be allergic or atopic, typically meaning that they must have allergic rhinitis. Although the authors have clearly demonstrated that shiners are more common, darker, and larger in children with allergic rhinitis that in normal healthy children, this fact would seem to be of limited usefulness. As the authors state, the phenomenon is “believed to be caused by venous stasis resulting from nasal congestion.” Thus, when presented with a patient with shiners, one may conclude that this is likely secondary to chronic nasal congestion. However, this does not necessarily mean that this nasal congestion is in turn the result of allergic rhinitis. A more appropriate control group would have been children referred for evaluation of chronic rhinitis who were determined not to be allergic, so-called nonallergic rhinitis. Even in the control group of children who had “never been given a diagnosis of allergy, had never presented with any allergic symptom, and had no underlying disease,” 24% had “shiners with significant darkness” (compared with 76% in the children with allergic rhinitis) and 15% had “shiners of significant size” (compared with 61% of children with allergic rhinitis). These percentages are likely to be even higher in children with nonallergic rhinitis than in these healthy children. If shiners were more common (or larger or darker) in children with allergic as opposed to nonallergic rhinitis, then they could be thought of as a useful clinical tool to predict nasal allergy; however, the current study has not shown this to be the case. Given that “allergic” shiners are relatively common even in healthy children, and the probability that they are even more common in children with nonallergic rhinitis than healthy children, using them to predict nasal allergies is not likely to be very accurate.