RESUMO
Matching serum and oral fluid (saliva) samples were collected from 369 subjects in Tunisia in 2002, from a city in the north and a rural district in the south. Rubella-specific IgG was detected in sera by commercial ELISA (Dade Behring) and in matching oral fluids by two methods, a previously described IgG-capture ELISA (GACELISA) [J. Clin. Microbiol. 37 (1999) 391] and the Dade Behring ELISA with the assay protocol modified for use with oral fluids. Total IgG concentration of oral fluids was also measured. Rubella-specific IgG was detected in 289 (78.3%) sera overall. Differences in the age distribution of the study population in the north and south led to a higher prevalence being found in the north (86.2%) than in the south (71.8%). This difference was reflected in the oral fluid rubella-specific IgG results. With GACELISA, rubella-specific IgG was detected in 79.4% of oral fluids from the north and 69.7% from the south and with the modified Dade Behring assay, in 81.4% of oral fluids from the north and in 64.9% from the south. The sensitivity and specificity of GACELISA in comparison to results from the matching sera were 92.4 and 93.2%, respectively. The sensitivity and specificity of the modified Dade Behring ELISA were 89.8 and 92.0%, respectively. Total IgG concentration in oral fluid showed a weak correlation (r=0.19) with the modified Dade Behring results and with samples where total IgG was >7.5 mg/l, the sensitivity and specificity were 94.4 and 90.0%, respectively. Twenty-nine oral fluids, which gave false negative rubella-specific IgG results with the modified Dade Behring ELISA, had a lower mean total IgG concentration than 256 oral fluids which gave concordant positive results (7.0mg/l versus 15.8 mg/l, P<0.001). The study validated the modified Dade Behring ELISA, providing an alternative to the GACELISA for assessing levels of rubella immunity for population studies using oral fluid samples.