PCR assay for diagnosis of human brucellosis.

We have read with interest the recent article by Morata et al. reporting their study of two of the possible factors which can interfere with specific DNA amplification in a peripheralblood PCR assay used for the diagnosis of human brucellosis (2). The study deals with the optimization of a method described in a previous work by these authors (4), in which they report a sensitivity of 100% and a specificity of 98% for rapid diagnosis of human brucellosis by PCR. Taking into account the clinical importance of this finding, we have tried to reproduce these results with 10 patients diagnosed with acute brucellosis and 5 healthy subjects. The diagnostic criteria were the isolation of a Brucella sp. from blood (n 5 6) or the finding of a $1/160 standard tube agglutination titer of antibodies to Brucella, in association with compatible clinical findings (n 5 4). Isolation of DNA from blood samples and DNA amplification were carried out by following the technique described by Morata et al. (2). All PCRs were carried out in duplicate. The results are shown in Table 1. The sensitivity was 50% (confidence interval [CI], 95%; 18.7 to 81.3%), and the specificity was 60% (CI, 95%; 14.7 to 94.7%). Curiously, of the samples from the six patients with brucellosis from which Brucella melitensis was isolated by blood culture, four were PCR negative. The existence of false negatives for patients with positive blood cultures could be due to the presence of polymerase inhibitors. Many substances have been suggested to be amplification inhibitors, including hemoglobin, urine, heparin, phenol, and sodium dodecyl sulfate. The technique described by Morata et al. seems to overcome inhibition caused by some of them, such as hemoglobin and heme compound derivatives. Nevertheless, we have not been able to reproduce the results of these authors, although we carefully followed the exact indications that are published in their article. Other hypotheses to account for those false-negative PCR results are the presence of a number of organisms below the detection limit and the degradation of target DNA in the samples (5). False positives for healthy subjects could be due to (i) asymptomatic infections not detectable by conventional diagnosis procedures (blood cultures and serology) or (ii) crossreactions with DNA from other bacteria, as described by Romero et al. (5). Those authors have developed a PCR assay with primers F4 and F5, derived from the 16S rRNA sequence of Brucella abortus. Romero et al. carried out a very wide specificity study, and only Ochrobactrum anthropi biotype D yielded a PCR product of the expected size, suggesting a close relationship between Brucella spp. and O. anthropi biotype D. When applying these methods to human samples (6), these authors found problems with false positives. The presence of O. anthropi DNA in human blood could be detected by this method. Even though the natural habitat of this bacterium is unknown, it has been isolated from different kinds of samples, among these, clinical samples from immunodepressed patients as well as from healthy individuals (6). Baily et al. (1) did not test whether the B4-B5 primers do amplify the DNA of O. anthropi. If this hypothesis is true, it could explain our false positive. Lastly, the differences between our results and those of Morata et al. could be due to the difficulty in reproducing PCR results in distinct laboratories, as described by Noordhoek et al. (3).