The study included 76 women (mean age 30.8 years, range 19 to 41 years) who had had repeated abortions or intrauterine death of the fetus. Blood samples were obtained at the time of abortion or within 3 months of fetal death. The study participants had been referred to the Department of Gynecology and Obstetrics, Rigshospitalet, Copenhagen, Denmark, between 1 September 1991 and 31 October 1992 as part of a larger study of pregnant women with repeated primary or secondary abortions or repeated intrauterine fetal deaths. Serum specimens were tested for antibodies to N. caninum and T. gondii as described below.

The absorbence values for the human serum samples were 0.10 to 1.24 absorbence units, whereas the mean value for the presumed N. caninum-negative human control serum was 0.26 (0.13 to 0.56). The mean absorbence values for the high-positive and low-positive control pig sera were 1.73 (1.54 to 1.93) and 0.87 (0.85 to 1.07), respectively. As no true N. caninum-negative or -positive human sera were available, serum specimens with absorbencies 0.50 (n = 12) were selected for further investigation (Table).

Figure

Figure. Western blot of Toxoplasma gondii or Neospora caninum antigen. Analysis was performed essentially as described by Sharma et al. (12) by using tachyzoites from in vitro culture of the...

None of the 12 specimens showed specific fluorescence in the indirect fluorescence antibody test (IFAT) at dilution 1:640 with N. caninum tachyzoites that had been cultivated in vitro (6). (Sera had been diluted in twofold serial dilutions from 1:20 in phosphate-buffered saline.) Of the 12, only 3 had T. gondii antibodies. The reactivities in the N. caninum enzyme-linked immunosorbent assay (ELISA) were not associated with the presence of T. gondii antibodies (Table). Only 1 of the 12 human serum specimens tested showed reactivity against the N. caninum antigen by Western blot analysis. This specimen, number 279, recognized an antigen with apparent molecular weight of 60 kDa (Figure, lane 11 and 12). This antigen was not recognized by the N. caninum-positive pig sera, and serum 279 did not recognize any of the low-molecular weight antigens recognized by the N. caninum-positive pig sera. Three serum specimens reacted with the T. gondii antigen; all were T. gondii-positive in the dye test.

Because of the biologic similarities between N. caninum and the human pathogen T. gondii, it has been speculated that N. caninum could be transmissible to humans. Since repeated abortions and stillbirths are common manifestations of neosporosis in cattle (4), women with a history of repeated abortions seemed an obvious category to investigate for human N. caninum infection. However, in this study of serum samples from women with repeated abortions, no evidence of N. caninum infection was detected.

The assays we used were based on methods used for T. gondii analyses; we used the same conjugates and serum dilutions found optimal in these analyses. The N. caninum immunostimulating complex antigen has a high specificity (14) and has been used for serologic investigations in different animal species (8,9,15). It was therefore anticipated that it would be applicable in a human system as well. However, because we could not define a proper cut-off for the assay, we further investigated the serum samples with the highest ELISA absorbence values by IFAT, regarded as the reference test for N. caninum antibodies in different species (4), and Western blot. None of the human sera investigated showed any reactivity in IFAT. Only one of the specimens reacted with the N. caninum antigen in the Western blot. However, because it only reacted with a band not recognized by sera from the infected pigs, the reaction was considered unspecific, and cross-reactivity between T. gondii and N. caninum was not found.

That we found no evidence of N. caninum infection in women who had repeated spontaneous abortions does not rule out the possibility that the infection might occur in humans. The predominant effects of neosporosis in dogs are primarily progressive neurologic signs including paralysis. It might, therefore, be worthwhile to examine human patients with clinical symptoms other than abortions, e.g., neurologic disorders of unknown etiology. Furthermore, the possible presence of N. caninum in patients with weakened immune systems should be considered. Researchers might continue the search for N. caninum by using serologic tests, as we did, or, alternatively, by using material collected at biopsy or autopsy for polymerase chain reaction or immunohistochemical analysis.