COMPLEMENT-FIXATION TEST IN TUBERCULOSIS 519 



yielding 76.7 per cent, positive results in 43 definite cases of phthisis, 

 80.7 per cent, in surgical tuberculosis, and 37.5 per cent, in glandular 

 tuberculosis. Of 87 normal individuals, only 3 gave positive reactions 

 (2 of these were lepers and 1 had Addison's disease.) Negative reactions 

 were obtained in 18 syphilitic patients. They look upon a positive 

 reaction as indicative of active tuberculosis. Stimson 1 (1915), who 

 gives a fairly exhaustive table of the recent literature, reports a small 

 number of cases in which a variety of antigens were used, but his re- 

 sults were inconclusive. Craig 2 (1915) reports the results of examination 

 of 166 cases of pulmonary tuberculosis in which he employed as antigen 

 an alcoholic extract of several strains of human tubercle bacilli which 

 had been grown on a liquid medium of alkaline broth containing egg; 

 96.2 per cent, positive results were obtained in active cases and 66.1 per 

 cent, positive in inactive cases. One hundred and fifty cases of syphilis 

 gave only 2 positive reactions, and these, on further examination, re- 

 vealed lesions in the lungs. One hundred other diseases examined all 

 gave negative results. In a later report Craig, using his same antigen 

 of a filtrate of an alcoholic extract of several strains of human tubercle 

 bacillus plus the culture-medium in which the bacilli had been grown, 

 has reported further favorable results in complement-fixation in tuber- 

 culosis. Miller and Zinsser 3 have described a simple antigen prepared 

 by grinding living or dead tubercle bacilli with dry table salt, and then 

 adding distilled water up to isotonicity. With this antigen they have 

 reported excellent results; Miller 4 found the reaction practically always 

 positive in active tuberculosis with negative reactions in non-tuberculous 

 syphilitic and normal persons. In my own studies in co-operation with 

 Montgomery, prepared after Miller's method, yielded positive reactions 

 only with the sera of tuberculous persons, but the percentage of positive 

 reactions was much less than reported. Corber, 5 employing an autolysate 

 of tubercle bacilli as antigen, found 30 per cent, positive reactions with 

 clinically definite cases of tuberculosis both active and inactive. Active 

 cases gave a higher percentage of positive results than inactive cases. 

 Eichhorn and Blumberg, 6 using Besredka's antigen and the sera of tuber- 

 culous persons and the lower animals, found the complement-fixation 

 test less reliable than the subcutaneous tuberculin test in the diagnosis 



1 Bull. Hyg. Lab. U. S. P. H. and M. H. S., 1915, No. 101, 7. 

 2 Amer. Jour. Med. Sci., 1915, 150, 781; Jour. Amer. Med. Assoc., 1917, Ixviii, 

 773. 



3 Proc. Soc. Exper. Biol. and Med., 1916, xiii, 134. 



4 Jour. Amer. Med. Assoc., 1916, Ixvii, 1519. 



5 Jour. Infect. Dis., 1916, 19, 315. 6 Jour. Agricult. Research., 1917, viii, 1. 



