The purpose of this investigation was to compare the levels of serum IgG2, the frequency of detection of Gm(23)-negative allotype and frequency of detection of FcgammaRIIa and FcgammaRIIIb receptor haplotypes in 32 refractory, 54 successfully treated and 27 periodontally healthy individuals. Refractory subjects showed mean full mouth attachment loss and/or >3 sites with attachment loss >2.5 mm within 1 year after both scaling and root planing, and surgery plus systemically administered tetracycline. Successfully treated subjects showed mean attachment level gain and no sites with attachment loss >2.5 mm 1 year post-therapy. Periodontally healthy subjects exhibited no pocket depth or attachment level >3 mm, and no evidence of progressing disease during 1 year of monitoring. Blood was obtained from each subject at baseline. Serum IgG2 and Gm(23) allotype were determined using radial immunodiffusion. DNA was extracted from whole blood and the FcgammaR genotypes determined using PCR and allele specific oligonucleotide probes. Significance of differences among clinical groups were sought using the Kruskal-Wallis or chi-square tests. Associations between 2 or more variables were tested using regression analysis. Refractory subjects exhibited higher mean attachment loss and pocket depth than successfully treated or periodontally healthy subjects. Smoking status did not differ significantly among groups. No significant differences in serum IgG2 levels and frequency of detection of Gm(23)-negative allotype were observed among the clinical groups. Serum IgG2 level was positively associated with the number of serum antibody responses to subgingival species (r=0.51, p<0.001). Subjects with the Gm(23)-negative allotype exhibited lower mean levels of serum IgG2 (3.06+/-0.3 versus 3.9+/-0.2, p<0.01) and mean number of serum antibodies to subgingival species (17.7+/-1.7 versus 23.3+/-1.4, p<0.05) than allotype positive individuals. No significant differences in FcgammaR haplotype distribution were observed among the 3 clinical groups. Associations of serum IgG2 level, Gm(23) allotype, FcgammaRIIa and FcgammaRIIIb receptor haplotypes and smoking status were weakly related or not related to clinical status. This lack of relationship may have been due to a reality of no relationship, or the inadvertent pooling of subjects where these factors were of primary importance with subjects in whom these factors played a less important role.