186 CLINICAL IMMUNOLOGY Newsletter Vol. 1 1, No. 12, 1991 luteinizing hormone. FASEB J 3:A1474. 16. Weigent DA, Baxter JB, Wear WE, et al.: Production of immunoreactive growth hormone by mononuclear leukocytes. FASEB J 2:2812-8, 1988. 17. Kao T-L, Keenan BS, Meyer WJ: Char- acterization of human growth hormone secreted by lymphocytes. Endocrine Soci- ety Annual Meeting Abstract, 1990. 18. Hiestand PC, Mekler P, Nordmann R, et al.: Prolactin as a modulator of lympho- cyte responsiveness provides a possible Wieconoln Proficiency Program for Lyme Disease: A Program for Improving Detection of Antibodies to Be555rli4 JpurfpJe f ri Lori L. Bakken, Steven M. Callister, Kay L. Case, and Ronald F. Schell Wisconsin State Laboratory of Hygiene, Madison, Wisconsin, and Gundersen Medical Foundation and Lutheran Hospital, La Crosse, Wisconsin mechanism of action for cyclosporin. Proc Natl Acad Sci USA 83:2599-603, 1986. 19. Marwick AJ, Lolait SJ, Funder JW: Im- munoreactive arginine vasopressin in the rat thymus. Endocrinology 119:1690-6, 1986. B orrelia burgdorferi, the spirochete that causes Lyme disease, can p infect many nssues of the body including the skin, heart, nervous system, and joints.17 Because of this, Lyme dis- ease symptoms can resemble those of other illnesses such as influenza, aseptic meningitis, multiple sclerosis, or rheuma- toid arthritis. This ability to cause symp- toms similar to other diseases can make the accurate diagnosis of Lyme disease difficult. Therefore, the physician relies heavily on the ability of the laboratory to detect accurately antibodies to B. burg- dorferi. Serologic testing is currently the only practical method for establishing a Lyme disease diagnosis. Infected humans pro- duce IgM and IgG antibodies that recog- nize B. burgdorferi antigens, lgM antibodies generally peak three to six weeks following infection, while IgG an- tibodies peak one to three weeks later. The concentration of IgG antibody may then remain elevated for weeks to years following infection. 4' 16 Western immunoblotting is used by some laboratories to detect antibodies to individual antigenic components of B, burgdorferi. However, its use as a prac- tical clinical laboratory test remains un- clear. For example, some investigators have shown that this test is superior to the indirect ELISA for determining early Lyme disease, 7 while others have demon- strated that it is not the method of choice 0197-1859/91/$0.00 + 2.20 for diagnosing early Lyme disease. 2'9 Until these concerns are resolved, this procedure appears to have limited useful- ness in most routine diagnostic laborato- ries. Currently, indirect immunofluorescent assays (IFAs) and enzyme-linked immu- nosorbent assays (ELISAs) are the most commonly used methods for identifying and quantifying the presence of antibodies to B. burgdorferi. Some investigators have reported that ELISA is more sensi- tive than IFA, 7 while others report that IFA is as predictive as ELISA if the labo- ratory has experience and conservative cutoff titers are used.12'13 Regardless, both systems remain somewhat less than perfect because of both common and unique problems associated with each. Whole borrelial cells are invariably used as the antigen in IFA. However, this is where similarities among IFA test procedures end since neither the antigens nor the methods have been standardized among laboratories. For example, diag- nostic titers considered indicative of B. burgdorferi exposure vary among labora- tories depending in part on their individ- ual criterion for intensity of fluorescence. This causes significant cutoff points to vary. Using ELISA, which generally utilizes either whole cells or sonicated cells as antigen, sera are considered diagnostically positive if they elicit absorbances 2.5 to 3 standard deviations above the mean ab- I © 1991 Elsevier Science Publishing Co., Inc. sorbance of a group of healthy controls. This technique has also not been stan- dardized among laboratories and wide variations currently exist. The cultures of B. burgdorferi used as antigen may also cause differences among test procedures. We have observed up to 16-fold differences in reported IgM anti- body levels of individual patient serum when performing IFA with diagnostic slides prepared using B. burgdorferi cul- tured in separate BSK medium containing different lots of fraction V bovine serum albumin. 3 It is easy to envision the same effect occurring with ELISA. Another variable affecting serologic testing may be use of other B. burgdor- feri isolates as antigen. Most laboratories and commercial companies use the origi- nal Shelter Island, NY, USA, isolate B31 (ATCC 35210). However, other isolates are also used. Greene et al. 6 showed that immunoblot patterns of sera from dogs naturally infected with B. burgdorferi were changed depending on which isolate of B. burgdorferi was used. We have also shown that different isolates signifi- cantly change immunoblot patterns of sera from humans with Lyme disease. 18 In addition, Schwan et al. 14 demonstrated that noninfectious B. burgdorferi isolate B31 (multiple passage) had a reduced immunoblot compared to an infectious isolate. These observations provide evi- dence that a serum may be falsely labeled negative if reactivity varies depending on I