Assuming this result holds... Perhaps anti-A antibodies that may be produced by non-A blood types have some cross-reactivity with the virus?