usefulness of such a technique for confirming cure. The numbers involved in this technique were too small to recommend it as a method of follow-up, and are only part of an ongoing study. Similar findings are needed for firm conclusions to be drawn. Although the sensitivity of the serological methods using agglutination assay was also
obvious limitation in distinguishing established cure from active infection, as the antibody titers were persistently high in the three samples submitted in convalescent phases during the study period. This result is in agreement with the findings of a previous report,12 and confirms the fact that serological follow-up of patients by agglutination tests does not always correlate suitably with the clinical outcome. If only agglutination tests had been considered in the three patients, it would have been difficult to establish cure unless several samples some few weeks apart had been taken to demonstrate declining antibody titer. The sensitivity of blood culture was only 57% (8/14) and there was a delay of about one week before results were issued, while PCR-based assays were able to give the results on the same day.
Although the apparent specificity of serological tests is superior to the PCR (97% vs. 88%), the corrected specificity of PCR, as calculated after exclusion of the occupational group, was higher (100%). No amplification signal was detected in the samples from febrile illnesses group despite the fact that there were four different microbial isolates in some of them, including Salmonella, which is known to share antigenic components with Brucella.3 While the serological test showed one false- positive case among the febrile group, representing 97% specificity, other serological tests that measure antibodies against specific Brucella proteins such as ELISA were found not to cause the problem of cross-reaction with other gram-negative bacteria. 12
The use of newer serological ELISA techniques was not included in our comparative study. These tests are generally more sensitive, as they are based on primary interaction, provide a profile of immunoglobulin classes for the diagnosis of acute and chronic brucellosis,13 might be useful markers for active brucellosis,12 and may readily
lipopolysaccharide antigen and monoclonal antibody.14 Nevertheless, agglutination tests remain the standard methods against which other tests must be compared. 15
For statistical analysis, we excluded the occupationally exposed group because it is unlikely that these tests are requested in asymptomatic individuals. We believe that the ability of PCR technique to detect four bands in otherwise asymptomatic but heavily exposed persons is a reflection of the extreme sensitivity of this technique, which is known to compromise the specificity of any assay.16 The clinical significance of positive results for such a highly sensitive method remains unclear.17 Queipo-Ortuno et al.4 did not find a single positive case in their control group, perhaps
DIAGNOSIS OF ACUTE HUMAN BRUCELLOSIS
due to the difference in exposure and protective measures in the two groups. The intermittent exposure of these people might have led to a localized infection or represent a very early stage of the disease not sufficient to produce detectable antibodies,16 however, the possibility of inadequately treated or self-limited past infection18 cannot be excluded. In either case, follow-up of these patients, both clinically and by serology or culture, is highly recommended for early diagnosis and treatment. We have advised these patients to report any symptoms and to repeat serology every 2-3 months.
The results of statistically significant tests applying hypothesis testing19 are also shown in Table 2. Both the sensitivity and specificity of PCR was significantly greater than that of culture and serology, respectively (P<0.001).
Because Brucella spp. are intracellular parasites, relapse is not an uncommon complication. The diagnosis of relapse is even more difficult than the diagnosis of initial infection, as the patient may remain seropositive for a year or longer,3 and negative blood culture is not uncommon in chronic infection.2 In this report, there was a clear band in the single patient presenting with symptoms of suspected relapse after five months of completing treatment. His blood culture was negative and serology showed persistent high titer antibrucella antibodies. This was a finding in a single case of relapse, and we are in the process of collecting more cases before valid conclusions can be made about the significance of PCR in diagnosing relapse.
In conclusion, the PCR-based assay has several advantages over the current microbiological methods for the diagnosis of brucellosis, including speed, safety, high sensitivity and specificity, therefore, it should be considered. For evaluation of asymptomatic, occupationally exposed persons, the traditional methods might be superior, especially if the cost is taken into consideration. Nevertheless, the information provided by PCR should be considered complementary to the results of conventional methods for the time being. Further researches on a large number of similar cases are necessary to establish the suitability of each method.
We are grateful to Mr. Dennis Dasal, Miss Nehad Al- Mubarak, and Mrs. Seham Kullab for technical assistance. We are also indebted to Dr. Zaheer Khan and Dr. Salem Khalil for collecting the study data and providing some reagents. We are also thankful to Mr. Hassan A. Al-Basri for technical and administrative help.
Volk WA, Gebhardt BM, Hammerskjold M-L, Kadner RJ. Brucella, Yersinia, Francisella and Pasteurella. In: Essentials of medical microbiology. 5th edition. Philadelphia: Lippincott-Raven, 1996;383- 95.
Annals of Saudi Medicine, Vol 20, Nos 3-4, 2000