4. [may obscure classification attempts. high IgG phase 1 may be predictive for chronic Q 7CKA fever, but also that high IgG phase 2 may aid in detecting such putative chronic cases. Key words: were measured by IFA [7, 12]. The IFA utilized for Q fever produces semi-quantitative data. 7CKA Immunofluorescence in serial dilutions of the test sera are visually compared to a standard and the titre is usually scored as a dilution factor (IFA, Focus Diagnostics, USA) [5, 13]. As dilutions increase twofold, any observed antibody titre is known up to a single dilution step of magnitude 2. For any quantitative interpretation, the reported measurements must be translated to antibody concentrations. Table 1 shows examples of the interpretation of IFA data. Any observed titre is usually usually an interval-censored observation. Note that concentrations may be too low to read (<1:32) or sera may not have been diluted sufficiently to allow measurement to within a single dilution step. Table 1. Quantitative interpretation of immunofluorescence assay data: example of numerous censored observations post-infection is usually described by a longitudinal function at different times and were fixed (assumed not to vary between patients), and were random, as was the initial antibody titre and normal, and standard deviations and gamma distributed. A full account of the longitudinal model has been published [16, 17]. Using Markov Chain Monte Carlo (MCMC) methods a Monte Carlo sample is usually obtained of the individual parameters to the combination likelihood is usually (6) where is the proportion positive samples. The two fitted components allow quantification of specificity and sensitivity [20]. For half-times a similar likelihood function can be constructed. When described as a binary distribution combination, any 7CKA titre can be assigned a probability of belonging to either subpopulation. Using the ratio (7) individual classification can be done, assigning odds against time following symptom onset in 344 patients measured by immunofluorescence assay. Data from your same patient are connected. Symbols show censoring: circles at geometric mean when both an upper and lower level have been observed. Triangles show absence of either a lower bound (downward sign) or Rabbit polyclonal to PPAN an upper bound (upward symbol). There was no obvious difference between any of the antibodies measured in males or females, nor could an age pattern be established (observe Appendix Fig. A3). Characteristics of the antibody response The modelled antibody responses showed considerable heterogeneity. Both the magnitude and the shape of the serum antibody response varied strongly in individual patients. IgM and IgG against phase 2 tended to reach higher levels than the corresponding phase 1 responses, while IgG antibodies tended to be more prolonged than IgM. Peak titres of 7CKA phase 2 antibodies were higher, almost by an order of magnitude compared to phase 1 antibodies. Estimated decay rates were smallest (slowest decay) in IgG phase 1, and more or less the same in all other antibody responses. Due to the low decay rates, patients with high estimated peak titres tend to keep these high titres for a prolonged period, for more than a 12 months after diagnosis of acute Q fever. Correlation coefficients of these characteristics are given in the Appendix. As different antibody classes have been fitted independently, correlation has not been included into the longitudinal models. However, by using the parameter estimates of the individually fitted responses any correlation in the observed data is usually conserved in the fitted responses. Apart from unfavorable correlation of time to peak and peak.