PRODUCTION OF ANTIPNEUMOCOCCIC SERUM 561 



Aside from the resultant loss of antibody and difficulty in effect- 

 ing solution occasioned by denaturation of globulins, there is the 

 possibility that the chemical or physical alteration may be re- 

 sponsible for the chills following therapeutic administration of 

 concentrated serums. 



Studies designed to determine the cause of chill-reactions and 

 methods for eliminating chill-producing principles in antipneumo- 

 coccic serum have followed several directions. The reactions are 

 not confined to the use of antipneumococcic serum, according to 

 Rhoades. 1135 Larson and Fahr (1925) 79 ° observed marked chills in 

 three of seven patients treated with antipneumococcic rabbit se- 

 rum. Differences between immune horse and rabbit serums have 

 been studied by Goodner and Horsfall (1935), 537 and by Goodner, 

 Horsfall and Bauer (1936), 639 while Chow (1936) 224 recently re- 

 ported that the major portion of antibody in antipneumococcic 

 rabbit serum occurs in a different globulin fraction from that in 

 horse serum. Whether these differences may explain in part the 

 high incidence of chills following therapeutic use of untreated rab- 

 bit serums remains unknown. 



A different approach to the problem was reported by Kolmer 

 and Matsumoto (1920), 744 who studied normal and immune horse 

 serum for the content of hemagglutinins and hemolysins to ascer- 

 tain whether intravascular agglutination or hemolysis might play 

 a part in chill reactions incident to serum therapy in pneumonia. 

 Although it was found that practically all horse serum contains 

 agglutinins for human erythrocytes, the authors believed that the 

 intravenous administration of horse serum probably does not in- 

 troduce sufficient amounts of hemagglutinin and hemolysin to pro- 

 duce ill effects referable to agglutination and hemolysis. Bullowa, 

 Rosenbluth, and Merkin (1927), 184 as a result of an incidental 

 observation that a chill-producing serum agglutinated red cells in 

 vitro whereas a chill-free serum did not, restudied this phase of the 

 question. No constant relation was found between erythrocyte ag- 

 glutination in vitro and serum reactions in man, thus confirming 



