The ability of antimalarials to moderate severe disease activity in systemic lupus erythematosus (SLE) is plausible but undemonstrated. We evaluated the long-term effectiveness of maintaining treatment with hydroxychloroquine sulphate (HCQ) to prevent major flares in quiescent SLE. Forty-seven patients with quiescent SLE who had been randomized to take HCQ (n = 25) or placebo (n = 22) as part of a 24-week withdrawal trial were evaluated for an additional 3 years. The primary outcome was time to a major flare of SLE which resulted in either the institution of or an increase in the current dosage of prednisone of 10 mg/day or more, or institution of therapy with immunosuppressive agents. Secondary outcomes included the specific subtype of these major flares (glomerulonephritis, vasculitis or other) and hospitalization for an exacerbation of SLE. An intent-to-treat analysis was conducted. Over the 42 months of study, 11 of 22 (50%) patients randomized initially to placebo, and seven of 25 (28%) patients randomized to continue treatment experienced a major flare. The relative risk of major flare for those randomized to continue HCQ compared with controls was 0.43 (95% CI: 0.17, 1.12). The relative risks for subtypes of flares were 0.26 (95% CI: 0.03, 2.54) for nephritis, 0.51 (95% CI: 0.09, 3.08) for vasculitis and 0.65 (95% CI: 0.17, 2.41) for flares characterized by other symptoms. The relative risk of hospitalization for major flare for patients randomized to continue hydroxychloroquine was 0.58 (95% CI: 0.13, 2.60). While the results are not statistically significant, they are compatible with the clinical belief that HCQ has a long-term protective effect against major disease flares in SLE and suggest that on average, HCQ use reduces major flares by 57% (95% CI: 83% reduction to 12% increase).

A previous report described the development of the immune response in rabbits to streptococcal group-specific carbohydrates following intravenous immunization with streptococcal vaccines (1). I n a continuation of these studies, certain rabbits were noted to develop an unusually high level of streptococcal group-specific antibodies. Examination of the zone electrophoresis patterns of these immune sera revealed a remarkably sharp and narrow band within the ,y-globulin region. This report is concerned with the immunologic, chemical, and physical properties of the specific 3,o-globulin isolated from these sera. The findings suggest a selective manufacture in certain rabbits of specific immune 3,G-globulin which exhibits a restricted range of electrophoretic properties compared to that of the normal complement of q-globulin. Materials and MethodsStreptococd.--Group A-variant strain A486 was obtained from Dr. R. C. Lancefield of the Rockefeller University.RabMts.--7-1b. New Zealand red rabbits were supplied by Baumgartd Bunny Farm, Marion, Iowa.Streptococcal Vaccine.--The vaccine was prepared from Group A-variant streptococci, strain A486. The bacteria were grown in 1000 ml of DiSco Todd-Hewitt broth, collected by centrifugation, and washed three times with saline. The collected bacteria were resuspended in 10 ml of saline, adjusted to pH 2.0, which contained 1 mg/ml of pepsin, and were incubated for 2 hr at 37°C. The treated bacteria were washed two times with saline and finally resuspended in 50 ml of formalized saline. The vaccine was stored at 4°C.

Tumors of the plasma cell series often produce large amounts of proteins which may be found in serum in the form of 7S T-globulin, 19S T-gloublin, /3~,-globulin, or Bence Jones protein, and in the urine primarily as Bence Jones protein. Although instances of heterogeneous myeloma proteins have been described, both in man (1-3) and in transplantable plasma cell tumors in mice (4-7), these proteins are usually extremely homogeneous when compared with their normal counterparts which are highly heterogeneous in a number of physicochemical properties. They lend themselves to isolation in relatively large amounts and high degree of purity, thus being well suited for physicochemical and serological studies. Although these proteins fall into four distinct classes, they are related in structure as evidenced by serological cross-reactions, and all cross-react with normal 7S "},-globulin (1, 8-13). In addition, each appears to have a counterpart in normal serum where the 7S "y-, 19S "},-, and /3~A-classes of immunoglobulins are now clearly recognized. Current evidence indicates that they represent three classes of antibodies (14, 15).In recent years, a number of genetically determined factors have been described in human 7S T-globulin. Of the seven factors dearly defined, Gin(a) (16), Gin(b) (17), Gin(x) (18), and Gin(r) (19) are determined by genes at one locus. In whites, factors Gin(a) and Gm(b) are produced by genes which behave as alternate alleles (17), whereas both factors in Negroes appear to be produced by a single allele called Gm ab (20). Two alternate alleles at another locus determine the factors Inv(a) and Inv(b) (21)(22). The Gm and Inv loci are independent in population studies (21, 23); the frequencies of the different Inv types make linkage studies difficult, and a slight linkage between the two loci has so far not been excluded. The seventh factor, Gin-like (24), occurs in Negroes and is rare in most other ethnic groups (25). So far, only one Gm-like(+) individual has been found in whites among thousands of individuals tested (26). The relationship between the Gm locus and the locus of Gin-like has not

Gm(a) and Gm(b) factors are present in 7S gamma-globulin molecules and absent in 19S gamma-globulins, beta(2A)-globulins and Bence-Jones proteins, whereas the Inv(b) factor was demonstrated in all four kinds of proteins. The Inv type was identical in isolated 7S and 19S gamma-globulins of six normal sera. After papain splitting of 7S gamma-globulin, Gm determining sites were present only in the fast (F) split product and Inv determining sites were present only in the slow (S).