Antinuclear Antibodies

To the Editor.—  I should like to comment on the disturbing consequences of employing methanol-fixed, paraffin-embedded sections of bovine and human epidermis for antinuclear antibody testing as described in the November 1972 Archives by Guss and Ugel.¹ My objections to this technique are threefold:

1. Loss of Diagnostic Potential of Nuclear Immunofluorescent Patterns.—  By the authors own admission, only one nuclear immunofluorescent pattern is distinguishable with this technique (nucleolar). According to Burnham² and T. K. Burnham, MD, BS, and P. W. Bank (unpublished data), 15 patterns have now been identified and reported. Several of these patterns are of major diagnostic significance, such as the peripheral pattern (Fig 1), which is seen mainly in severe systemic lupus erythematosus (SLE), and the homogeneous leukocyte specific antinuclear antibody (Fig 2),^2,3 which is seen mainly in rheumatoid arthritis and SLE. The latter antibody reacts only with granulocytic nuclei and is therefore not demonstrable on epithelial