Diagnosis of recent and relapsed cases of human brucellosis by PCR assay
© Nimri; licensee BioMed Central Ltd. 2003
Received: 10 February 2003
Accepted: 28 April 2003
Published: 28 April 2003
Brucellosis affects human populations in many developing countries including the Middle East, and Latin America where it is still endemic. It has been prevalent in Jordan for years, where 7842 cases of human brucellosis were registered at the Ministry of Health during 10 year-period. This study was initiated by the recent increase in the number of human cases diagnosed in a rural area in the Northern Jordan to help assess the status of the disease in that area. For this purpose blood specimens from brucellosis suspected cases were tested by serology, culture and PCR.
Peripheral blood specimens from 50 healthy control subjects and 165 seropositive patients having compatible signs and symptoms that were clinically diagnosed to have brucellosis were tested by blood culture, and by PCR. The PCR assay used genus-specific primers from the conserved region of the 16S rRNA sequence, which showed high specificity for the Brucella spp.
Diagnosis of Brucella was established by PCR in 120 cases (72.7%). All of them were seropositive and 20 were positive by culture. Forty-eight of 58 (82.8%) of the relapsed cases two months after completing the treatment with an increase in the previous serological titers were positive by PCR. The assay has 85.7% positive predicative value, 100% sensitivity and specificity since it correctly identified all cases that were positive by blood cultures, 95.8% by serology and none of the control group was positive.
Results showed that PCR assay can be applied with serology for the diagnosis of brucellosis suspected cases and relapses regardless of the duration or type of the disease without relying on the blood cultures, especially in chronic cases.
Brucellosis continues to be an important zoonosis of economic significance and human suffering. It affects human populations in many developing countries including the Middle East, and Latin America where it is still endemic . The World Health Organization reports an annual incidence of human brucellosis of less than one to 78 cases per 100,000 population in the Middle East, with six countries reporting an annual total incidence of more than 90,000 cases . Four Brucella spp. namely, Brucella abortus, B. canis, B. suis and especially B. melitensis are able to cause human infections . Brucellosis is usually transmitted to humans by ingestion of unpasteurized dairy products or by direct contact with infected animals. Occupational disease is contracted by exposure of abattoir workers and veterinarians to infected animals especially aborted fetuses, fluids, membranes or urine. In developed countries, the incidence of human brucellosis has declined in the last 50 years as a result of infection control measures, and in these countries most cases result from occupational disease, travel-acquired infections, or accidental laboratory exposure through contaminated aerosols . According to the World Health Organization, half a million new human cases each year are reported worldwide . These numbers greatly underestimate the true incidence of human disease  because the clinical picture of human brucellosis is extremely variable , and because it is under declared to local authorities and misdiagnosed by clinicians.
Brucella spp. are able to cause prolonged morbidity due to the capability of this bacteria to evade the host defense mechanisms by surviving as intracellular organisms. The diagnosis of the disease can be challenging and is frequently delayed or missed because the clinical picture may mimic other infectious and noninfectious conditions [6, 7]. Diagnosis can be established by laboratory methods such as serology and blood cultures. Prolonged incubation period, special growth media, and subcultures are required for the isolation of these fastidious, slow growing bacteria. However, cultures are not always positive when other tests are positive . Automated systems have been reported to detect more than 95% of Brucella melitensis-positive cultures within seven days of incubation . The technology is lacking in developing countries or rural areas where the disease is prevalent and diagnoses rely mainly on serology. Many serological tests have been used for the diagnosis of human brucellosis such as agglutination tests, indirect immunofluorescence, ELISA. The most commonly used tests are the serum agglutination test (SAT), the Coombs anti-Brucella test, the Rose Bengal test, and complement fixation . Each test has its own disadvantages, and the presence of antibodies doesn't always mean an active case of brucellosis, since humans from endemic areas often show weak serological responses . As for other fastidious pathogens, amplification of DNA by PCR offers an alternative way of diagnosis of brucellosis.
Brucellosis has been prevalent in Jordan fo r years, where 7842 cases of human brucellosis were registered at the Ministry of Health during 10 year-period, from January 1988 to 1997 . Ovine and caprine brucellosis caused by Brucella melitensis have been present in sheep and goat flocks in the bedouin sector for years. The number of cases has increased considerably in the mid to late 1980s. The increase might partially be attributed to improved diagnostic methods.
This study was initiated by the recent increase in the number of human cases diagnosed in the health centers in a rural area in the Northern Jordan Badia to help assess the status of the disease in that area. For this purpose blood specimens from brucellosis suspected cases were cultured and tested with PCR assay using a genus-specific primer pair derived from the conserved region of the16S rRNA sequence which showed high specificity for the Brucella spp. . The results are compared with that of the blood cultures and the Rose Bengal agglutination test.
A total of 215 peripheral blood specimens was collected from patients in four comprehensive health centers over a period of 18 months. Of these blood specimens, 165 were from patients who presented with clinical signs compatible with brucellosis. Clinical diagnosis was made by the physicians in the Badia of Jordan, a rural area in the north east of Jordan inhabited by recently settled Bedouins. These centers receive patients from 40 villages with a population exceeding 12,000 inhabitants. The diagnosis of brucellosis is established based on the presence of compatible signs and symptoms with the demonstration of specific antibodies at significant titer or seroconversion. The physicians are aware of the zoonotic nature of the disease because of the continuous reporting of the disease in sheep, goats and in humans in this area. Sixty (36.4%) of the samples were collected after adequate antibiotic treatment was started.
A questionnaire was completed for each subject at the time of specimen collection to record demographic and other relevant information such as the clinical history, symptoms and physical signs, contact with animals, drinking unpasteurized milk, and homemade dairy products.
The duration of symptoms prior to diagnosis was 1 to 16 weeks. Cases with clinical symptoms less than two months were considered as acute cases, those that lasted more than 6 months before treatment was initiated were considered as chronic cases. A relapse was considered to be either a positive blood culture two months to one year after completing the treatment or the reappearance of compatible symptoms not otherwise explained together with an increase in the previous serological titers. All patients had fever during the course of the disease with other symptoms such as muscle pain, wrist arthritis.
Treatment of recent cases is by a combination of a daily intramuscular injection of one gram of streptomycin for 14 days and doxycycline for a month. In the chronic cases, a combination of doxycycline and rifampin is given to the patient for one month according to the internationally accepted treatment regimens . The treatment was repeated when the symptoms persist and antibody titer was still high two weeks after concluding the treatment.
Samples were obtained from 50 subjects, composed of patients randomly selected among patients from the same region attending the same health centers. None of these patients was currently diagnosed or had a history of brucellosis.
The study was approval by the University ethics committee.
The serological tests were carried out by the laboratories of the health centers. The serologic diagnosis was established by Rose Bengal agglutination test. A titer of 1/160 was considered positive.
Five to ten ml peripheral blood samples were withdrawn and immediately inoculated under aseptic conditions in broth media (Bloodgrow®, Medical Wire & Equipment C. Ltd, Corsham, Wiltshire, England) or diphasic blood culture bottles (Hemoline performance diphasique bioMerieux, Marcy l'Etoile, France). Cultures were incubated at 37°C for 30 days in the presence of 5% CO2 and were periodically checked for growth. Subcultures on blood agar plates were performed in a blind manner at 10, 20, 30 days, they were recorded as negative after the last negative subculture. Brucella spp. were identified using standard methods .
DNA was extracted from blood specimens using a commercial purification system (Wizard Genomic DNA Purification Kit, Promega, Madison, WI) according to the manufacturer's instructions for DNA purification from blood. Final pellets were resuspended in 50 l of TE (10 mM Tris, I mM EDTA, pH 7.2).
The Brucella DNA-Detect PCR Kit (Vita-Tech International Inc., CAN), including the reagents and oligonucleotid primers designed for the direct amplification of the genus Brucella, was used for the detection of Brucella spp. in blood samples. The primers used in the kit were genus-specific primer pair designed to amplify a highly conserved region within the 16S rRNA of the genus Brucella .
The size of the amplification products was a 905 fragment of the rRNA gene. The reaction mixture contained 5 μl of 10 × PCR buffer, 1 μl of the primer mix, 2 μl of dNTP mix (5 mM each), 3 μl of 25 mM MgCl2, 1 unit of Taq DNA polymerase, 5 μl of sample DNA in a volume of 50 μl. The same mixtures were used with the positive control provided by the kit. The mixture with no DNA sample, and DNA free water were used as negative controls to monitor contamination.
The reaction was performed in a thermal cycler (Gene Amp PCR System 9700, Perkin Elmer, Norwalk, Con.). The cycling conditions were an initial denaturation at 95°C for 5 min, template denaturation at 95°C for 30s, annealing at 54°C for 90s, and primer extension at 72°C for 90s for a total of 35 cycles, with a final extension at 72°C for 6 min. A sample was considered positive when the size of the DNA band matched with that of the po sitive control (905 bp). All standard precautions recommended for prevention of contamination with DNA and amplicons were undertaken. Twelve microliters of the PCR product were run by electrophoresis in a 2% agarose gel in 1 × TBE buffer (Promega, Madison, Wis.), and gels were stained with ethidium bromide (2 μg/ml).
PCR amplification products were detected by visualization of the bands under UV light.
The means, ranges, percentages of positive samples, specificity, sensitivity and positive predictive value of PCR were calculated.
The mean age of the patients was 46 years (range, 6 to 86 years). Of these patients, 58 (35.2%) were males and 107 (64.8%) were females. The information recorded in the questionnaire indicated that, out of 165 patients included in this study, 66 (40%) had contact with animals (e.g., sheep and goats) either raised nearby where they live or at work (e.g., farmers). Seventy-four (44.9%) acquired their infections by consuming unpasteurized milk or homemade dairy products (e.g., soft cheese), 32 (19.4%) acquired their infection possibly from contact with animals as mentioned above, 24 (14.5%) acquired their infection possibly from one of these two sources.
The other 35 (21.2%) patients acquired their infection from unknown sources, since they either drink pasteurized milk, do not drink milk, or drink powder milk, and had no homemade cheese. Two of the patients drink camel milk.
Epidemiological, clinical, and serological results of 165 cases of human brucellosis.
No. of patients studied
Mean age in years (range)
Clinical and laboratory diagnosis:
Mean duration of the chronic cases in years (range)
No. of patients with fever
No. of patients with muscle pain and anorexia
No. of relapsed patients
No. of positives in Rose Bengal agglutination test (1/160)
No. of patients with positive blood cultures
No. of patients with positive PCR
No. of relapsed patients with positive PCR
One hundred and forty (84.9%) of these patients were positive by Rose Bengal agglutination test with titers ranging between 1/1 60 and 1/1280 units. The 25 clinical cases who were seronegative were not included in the group of certainly infected patients.
The number of patients who were tested for Brucella antibodies in that area during the study was 1050, they had symptoms consistent with brucellosis, but only 452 (43%) were positive at the time they were tested.
Brucella was isolated in blood cultures from 20 of the 165 (12%) blood specimens who didn't receive antibiotic treatment at the time the specimen was colleted. Isolates were identified by standard methods as B. melitensis. These samples were positive by both PCR and serology.
One hundred and twenty (72.7%) of the 165 samples were positive by PCR. These patients had symptoms for more than one month t o several years and were all positive by serology; recent infections of less than one month were negative by PCR. The DNA bands that appeared in the acute cases were brighter than those of the chronic cases.
Comparison of the PCR results with blood culture and serology for the165 cases
Result of tests
Seroa, Cult b
Forty-eight (82.8%) of the 58 relapsed cases were positive by PCR and by serology. The positive predicative value (e.g., t he proportion of true positive among all positives) of the assay was 85.7% based on the serology results.
None of the 50 subjects was positive by serology or by PCR.
The sensitivity of the PCR assay is 85.7% compared to serology and the specificity is 100% since all the control subjects were negative.
Although brucellosis is a notifiable disease in many countries, official figures do not fully reflect the number of cases reported annually, and the true incidence has been estimated to be between 10 and 25 times higher than what the reported figures indicate .
The increase in the incidence of human cases was linked to the lambing season in Jordan [10, 14]; these reports are in agreement with the findings of the current study. Bedouins are at higher risks of contracting brucellosis, due to the lifestyle, environmental and social conditions of this population .
Diagnosis of human brucellosis based on the clinical picture alone is not definitive since the symptoms mimic other diseases and is fairly nonspecific. The isolation of the organism in culture or detection by molecular methods is necessary to confirm the clinical diagnosis. There are few studies concerning the use of PCR technique in animals [16, 17], the studies of human cases are limited and the number of patients included was small [3, 18, 19]. The present work studied the potential use of a single-step PCR assay as a rapid test for the diagnosis of acute, chronic infections, and a relapse in previously treated patients. Patients were also studied by means of blood cultures, Rose Bengal agglutination. In many laboratories the serological diagnosis of human as well as animal brucellosis is based on a Rose Bengal test, which was improved from being a screening test to be a titrable one .
One of the genetic targets frequently used for strain identification and strain phylogeny is the rRNA operon, particularly the 16S rRNA gene used in this study. These genes are highly conserved and diverge very slowly. The DNA sequences from separate species within a genus will differ by only a few percent. Sequence identity among 16S rRNA sequences is typically interpreted as indicating a single species . A PCR assay with primers derived from the 16S rRNA sequence of Brucella abortus was developed . The specificity and high sensitivity of this assay provide a valuable tool for the diagnosis of brucellosis.
The PCR results confirmed the clinical diagnosis in 120 (72.7%) patients, the results obtained by serology in 120/140 (85.7%) and by blood cultures in 20/20 (100%). It correctly diagnosed 48/58 (82.8%) of the relapse cases who were identified based on the symptoms and the increase in the antibody titer after completing the treatment. It can reduce the time needed for blood cultures and its limitations caused by the low number or bacteria in the blood especially in chronic and focal type of the disease . The assay was able to detect Brucella DNA in cases of acute, chronic and relapsed brucellosis. The DNA bands in chronic cases were weak most probably due to low numbers of Brucella in the peripheral blood. The same results were reported by a previous study in serum samples obtained from staff of cattle breeding centers .
Twenty of the forty-five cases that were negative by PCR were recently diagnosed cases of brucellosis (< one month). These cases were negative both by serology and by blood cultures, indicating that they are true negatives. The other 25 samples were negative by the three methods, which might be due to the absence of infection, the possibility of temporary or episodic lack of bacteremia in brucellosis patients, the stage or nature of the disease, or the treatment that lowers the number of bacterial cells in the peripheral blood beyond detection. The therapy can modify PCR, culture, and serology results and therefore these cases should constitute a category different from those where samples were taken before therapy.
Several factors were reported to inhibit PCR in a blood specimen such as the high concentrations of leukocytes DNA and heme compounds .
The sensitivity of the PCR assay was 100% based on the results of the 20 patients who were positive by both serology and blood culture and 85.7% based on the number of patients who were positive by serology alone. The specificity of the assay was 100% since none of the 50 control blood samples was positive by PCR. The threshold sensitivity for the PCR assay was reported to be from 80 ng to 0.08 fg  as determined by testing serial dilutions of B. abortus 2308 DNA detected aft er 40 cycles of amplification. Similar findings with respect to the amount of Brucella DNA detected by PCR have been obtained by using primer sequences from the genes encoding a 43-kDa outer membrane protein  and a 31-kDa B. abortus antigen . It was suggested that 60 to 100 fg of bacterial DNA is equivalent to 20 cells . The importance of this finding is based on the fact that low numbers of circulating Brucellae are present in focal and chronic forms of the disease that are difficult to detect by culture and leads to a suboptimal recovery rate of Brucella from blood.
The PCR test proved t o be useful in the early detection of relapses as was reported by another . study . The diagnosis of these relapses is difficult by conventional methods Relapses after concluding a correct treatment are known to occur in brucellosis [3, 27]. This problem is related to the ability of Brucella spp. to evade some of the basic mechanisms of the host's immune system . The relapse rate of 32.5% of the pat ients after a combined treatment could be explained in part by lack of compliance in some patients to the full course of treatment. This is common in such rural communities where Bedouins show up for treatment several weeks after having symptoms and discontinue treatment once the fever subsides. It might be also explained by the fact that Brucella organisms localize in the reticuloendothelial system after surviving the intracellular mechanisms by phagocytes and polymorphonuclear leukocytes [3, 28]. A combination regimen of two or three drugs is more effective than a single-drug regimen. The relapse rate may be as high as 50% in cases of single-drug regimens [28, 29]. Longer courses of therapy are required to cure relapses and focal forms of the disease such as meningitis and ost eomyelitis. The po sitive PCR results obtained in the 48 (82.8%) of the relapse cases in this study is especially important indicating that the assay co uld be a useful tool to confirm a relapse in cases of a treated brucellosis. However, due to the extremely high sensitivity of the technique, the ability to amplify the DNA of non-viable, non-culturable Brucella DNA from dead or phagocytized cells especially the remains of DNA present in the circulating mononuclear cells in certain patients who have concluded successfultreatment  should be considered when interpreting the results.
The criteria to indicate a cure of brucellosis is not definite, since negative blood cultures do not exclude the presence of the disease. Another study reported recovering the bacteria in 53.4% to 95% blood cultures of patients but successful isolation of the organism decreased over time . Other specimens from bone marrow and liver tissue or lymph nodes may improve recovery rates , but these specimens are invasive and are not always feasible. Antibodies may remain elevated for a long time after conclusion of the treatment , which was noticed in some o f the chronic cases in this study that were negative by PCR.
The results of this study are in agreement with several previous studies [8, 18, 22, 26] that reported on the usefulness of PCR and recommended its use for the diagnosis of initial episodes, focal complications, post treatment follow-up, and relapses.
The peripheral-blood-based PCR assay described is highly sensitive and specific, easy to perform, and could provide results to a clinician in less than 8 h. In addition, it avoids the risks to laboratory personnel associated with handling the microorganism
This study received financial support from the Jordan Badia Research and Development Pro gram, The Higher Council of Science and Technology, Grant # 129/99.
The author thanks the employees in the Safawi Center, the physicians and laboratory employees in the Badia Health centers for their cooperation during the study.
- Alton G: Brucella melitensis. In Networking in Brucellosis Research. Edited by: Frank JF. 1991, Japan, United Nations University Press, 205-216.Google Scholar
- World Health Organization: Sixth report of the Joint FAO/WHO Expert Committee on Brucellosis. WHO Technical report series Rep. 740. World Health Organization, Geneva, Switzerland. 1986Google Scholar
- Young EH: An overview of human brucellosis. Clin Infect Dis. 1995, 21: 283-290.View ArticlePubMedGoogle Scholar
- World Health Organization: Fact sheet N173. World Health Organization, Geneva, Switzerland. 1997Google Scholar
- Zerva L, Bourantas K, Mitka S, Kansouzidou A, Legakis NJ: Serum is the preferred clinical specimen for diagnosis of human brucellosis by PCR. J Clin Microbiol. 2001, 39: 1661-1663. 10.1128/JCM.39.4.1661-1664.2001.View ArticlePubMedPubMed CentralGoogle Scholar
- Araj GF: Human brucellosis: a classical infectious disease with persistent diagnostic challenges. Clin Lab Sci. 1999, 12: 207-212.PubMedGoogle Scholar
- Yagupsky P: Detection of Brucellae in blood cultures. J Clin Microbiol. 1999, 37: 3437-3442.PubMedPubMed CentralGoogle Scholar
- Romero C, Gamazo C, Pardo M, Lo'pez-gon I: Specific detection of Brucella DNA by PCR. J Clin Microbiol. 1995, 33: 615-617.PubMedPubMed CentralGoogle Scholar
- Orduña A, Almaraz A, Prado A, et al.: Evaluation of an immunocapture-agglutination test (Brucellacapt) for serodiagnosis of human brucellosis. J Clin Microbiol. 2000, 38: 4000-4005.PubMedPubMed CentralGoogle Scholar
- Abu Shaqra QM: Epidemiological aspects of brucellosis in Jordan. Eur J Epidemiol. 2000, 16: 581-584. 10.1023/A:1007688925027.View ArticlePubMedGoogle Scholar
- Ariza J, Gudiol F, Pallare's R, Fernandez-Viladrich P, Rufi G, Corrediora J, Miratvilles MR: Treatment of human brucellosis with doxycycline plus rifampin and doxycycline plus streptomycin, A randomized double-blind study. Ann Intern Med. 1992, 117: 25-30.View ArticlePubMedGoogle Scholar
- Shapiro DS, Wong JD: Brucella. In Manual of Clinical Microbiology. Edited by: Murray PR. 1999, USA, ASM Press, 625-631.Google Scholar
- Corbel MJ: Brucellosis: an overview. Emerg Infect Dis. 1997, 3: 213-221.View ArticlePubMedPubMed CentralGoogle Scholar
- Dajani YF, Masoud AA, Barakat HF: Epidemiology and diagnosis of human brucellosis in Jordan. J Trop Med Hyg. 1989, 92: 209-214.PubMedGoogle Scholar
- Levy A, Fraser D, Vardi H, Dagan R: Hospitalizations for infectious diseases in Jewish and Bedouin children in southern Israel. Eur J Epidemiol. 1998, 14: 179-186. 10.1023/A:1007439908351.View ArticlePubMedGoogle Scholar
- Leal-Klevezas DS, Martinez-Vazquez IO, Lopez-Merino A, Martinez-Soriano JP: Single-step PCR for detection of Brucella spp. from blood and milk of infected animals. J Clin Microbiol. 1995, 33: 3087-3090.PubMedPubMed CentralGoogle Scholar
- Siftuentes-Rincon AM, Revol A, Barrera-Saldana HA: Detection and differentiation of the six Brucella species by polymerase chain reaction. Mol Med. 1997, 3: 734-739.Google Scholar
- Isabel MA, Queipo-ortun O, Morata P, Oco N, Manchado PP, Colmenero JD: Rapid diagnosis of human brucellosis by peripheral-blood PCR assay. J Clin Microbiol. 1997, 35: 2927-2930.Google Scholar
- Morata P, Queipo-Ortuno MI, Colmenero JD: Strategy for optimizing DNA amplification in a peripheral blood PCR assay used for the diagnosis of human brucellosis. J Clin Microbiol. 1998, 36: 2443-2446.PubMedPubMed CentralGoogle Scholar
- Altwegg M, Bohl E: Evaluation of a rapid, reliable, and inexpensive screening test for the serological diagnosis of human brucellosis. Zentralbl Bakteriol Mikrobiol Hyg. 1985, 260: 65-70.Google Scholar
- Bricker BJ, Ewalt DR, MacMillan AP, Foster G, Brew S: Molecular characterization of Brucella strains isolated from marine mammals. J Clin Microbiol. 2000, 38: 1258-1262.PubMedPubMed CentralGoogle Scholar
- Morata P, Queipo-Ortuño MI, Reguera JM, Miralles F, Lopez-Gonzalez JJ, Colmenero JD: Diagnostic yield of a PCR assay in focal complications of brucellosis. J Clin Microbiol. 2001, 39: 3743-3746. 10.1128/JCM.39.10.3743-3746.2001.View ArticlePubMedPubMed CentralGoogle Scholar
- Shestopalov MI, Kalinovskii AI, Balakhonov SV, Kapustin IM, Golubinskii EP: Analysis of PCR-diagnosis of brucellosis in human. Mol Gen Mikrobiol Virusol. 1999, 3: 12-17.PubMedGoogle Scholar
- Fekete A, Bantle JA, Halling SM, Sanborn MR: Preliminary development of a diagnostic test for Brucella using polymerase chain reaction. J Appl Bacteriol. 1990, 69: 216-227.View ArticlePubMedGoogle Scholar
- Baily GG, Krahn JB, Drasar BS, Stoker NG: Detection of Brucella melitensis and Brucella abortus by DNA amplification. J Trop Med Hyg. 1992, 95: 271-275.PubMedGoogle Scholar
- Morata P, Queipo-Ortuño MI, Reguera JM, García-Ordoñez MA, Pichardo C, Colmenero JD: Post treatment follow-up of brucellosis by PCR assay. J Clin Microbiol. 1999, 37: 4163-4166.PubMedPubMed CentralGoogle Scholar
- Ariza J, Corrediora J, Pallare's R, Fernandez-Viladrich P, Rufi G, Pujol M, Gudiol F: Characteristics of and risk factors for a relapse of brucellosis in humans. Clin Infect Dis. 1995, 20: 1241-1249.View ArticlePubMedGoogle Scholar
- Solera J, Martinez-Alfaro E, Espinosa A: Recognition and optimum treatment of brucellosis. Drugs. 1997, 53: 245-256.View ArticlePubMedGoogle Scholar
- Anonymous : Brucellosis, occupational health-zoonotic disease fact sheet # 14. [http://www.ksu.edu/research/animal/occhs/fact14.htm]
- Ariza J, Pellicer T, Pallare's R, Foz A, Gudiol F: Specific antibody profile in human brucellosis. Clin Infect Dis. 1992, 14: 131-140.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1471-2334/3/5/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article: verbatim copying and redistribution of this article are permitted in all media for any purpose, provided this notice is preserved along with the article's original URL.