The relationship of immunosuppression to cytomegalovirus infection.

It has been pointed out that immunosuppression with cytotoxic, nonsteroidal drugs is associated with the development of cytomeglovirus (CMV) infection after renal transplantation, and that without such immunosuppression infection may not occur (3). What is not known is whether immunosuppression alone, without other factors peculiar to the transplant recipient, can cause CMV infection. One approach to this question is to study patients receiving such therapy for reasons other than to prevent allograft rejection. Patients with rheumatologic disorders were chosen for this purpose because there is no known association between these diseases and CMV. This investigation may conveniently be divided into two parts. The first was a prevalence survey conducted among patients attending a rheumatology outpatient clinic during a 7-month period. Blood and urine samples were obtained on a single occasion from essentially all clinic attendees. Isolation of CMV was attempted from buffy coat and urine specimens, and complement-fixing (CF) antibody was measured on each serum. Table 1 is a summary of the serologic results among 131 rheumatology clinic patients. The geometric mean titer and the fraction of rheumatology patients with a CF titer of > 1:4 are compared to a control group of 211 unselected blood donors. As would be expected in an adult population, both the geometric mean titer and the fraction who were seropositive increased with age up to 55 years. Above 55, CF titers appeared to level off or even decrease slightly. As shown in Table 1, the geometric mean titers of females are greater than those of males, both in rheumatology patients and in blood donors. This relationship between sex and titer is statistically significant in both of these two populations. Therefore, when the influence of other factors on CMV serology was examined, appropriate statistical methods to adjust for age and sex differences were used. In each instance the effect of a particular factor was examined with respect both to the observed proportion of relevant persons who had antibody and to the magnitude of the titer among those who were seropositive. In order to adjust for the effects of varying age and sex distributions when testing for significant differences between population subgroups in the observed proportion of individuals with antibody, Mantel-Haenszel summary chi-square statistics (4) were employed. Analysis of variance techniques were used to identify those variables which are significantly related to the magnitude of titer among seropositive individuals. By analysis of