Diagnosis of disseminated mycobacterial infection: testing a simple and inexpensive method for use in developing countries.

Diagnosis of disseminated mycobacterial infection: testing a simple and inexpensive method for use in developing countries F.C.O. Fandinho,1 B. Grinsztejn,2 V.G. Veloso,2 M.C.S. Louren9o,3 E. Werneck-Barroso,4 E. J S.A. Nogueira,6 & L. de S. Fonseca7 With the development of the acquired immunodeficiency syndrome (AIDS) epidemic, the isolation of mycobacteria from blood has become a common problem for clinical laboratories. In this study two methods were used for the recovery of mycobacteria from blood specimens obtained from AIDS patients: (1) direct inoculation of a biphasic medium, and (2) a non-commercial lysis-centrifugation method. A total of 3 consecutive blood samples were taken at 15-minute intervals from each of 50 AIDS patients with clinical suspicion of disseminated mycobacterial disease. Mycobacterium growth was noted in 70/138 blood speci- mens from 30 (60%) patients. These cultures yielded Mycobacterium tuberculosis in 19 (63%) and Mycobacterium avium complex organisms in 11 (37%) patients. Cultures using the lysis-centrifugation method were positive in 54% of the patients while cultures using biphasic medium were positive in 44% (P > 0.05). The positivity for M. avium complex was higher with lysis-centrifugation (91%) than with biphasic medium (45.4%) (P < 0.05). However, the positivities for M. tuberculosis with the lysis-centrifugation method (89.5%) and direct inoculation in biphasic medium (100%) were similar (P > 0.05). The use ofa non- commercial lysis-centrifugation technique is inexpensive, reliable, and can be an alternative method for the diagnosis of mycobacteraemia in developing countries. Introduction Following the development of the acquired im- munodeficiency syndrome (AIDS) epidemic, an increasing number of secondary disseminated mycobacterial infections has prompted the search I Graduate Student, Microbiology Institute, Federal University of Rio de Janeiro and Department of Technological Development, Immunobiological Products Institute, Fiocruz, Rio de Janeiro, Brazil. 2 Physician, Evandro Chagas Hospital, Fiocruz, Rio de Janeiro, Brazil. 3 Chief, Bacteriology Laboratory, Evandro Chagas Hospital, Fiocruz, Rio de Janeiro, Brazil. 4 Associate Researcher, Evandro Chagas Hospital, Fiocruz, Rio de Janeiro, and Pulmonary physician, Pneumology and Physiol- ogy Institute, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil. 5 Physician, Servidores do Estado do Rio de Janeiro Hospital, Rio de Janeiro, Brazil. 6 Associate Professor, Department of Infectious Diseases, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil. 7 Associate Professor, Department of Medical Microbiology, Microbiology Institute, Federal University of Rio de Janeiro, Centre for Health Sciences, Bloco 1, liha do Fundao, 21.941-590, Rio de Janeiro, Brazil. Requests for reprints should be sent to this author. Reprint No. 5789 for a sensitive method of recovering mycobacteria from blood. The highest diagnostic yield for Myco- bacterium avium complex organisms is achieved by lysis-centrifugation culture methods (1, 2). Such sys- tems are commercially available and are commonly used in the USA and in Western Europe (3,4). Diag- nosis with the radiometric BACTEC culture me- dium is more rapid and has a similar sensitivity but requires a radioactive substrate and does not yield isolated colonies (5,6). Both methods are too expen- sive for routine use in developing countries. This prospective study compares the efficiency of two dif- ferent methods for detecting mycobacterial growth from the peripheral blood of AIDS patients: (1) direct inoculation onto a biphasic medium and (2) an inexpensive non-commercial lysis-centrifugation method. Methods Patients During the period December 1993 to August 1994, blood specimens were obtained from 50 AIDS pa- tients with suspected disseminated mycobacterial Bulletin of the World Health Organization, 1997, 75 (4): 361-366 © World Health Organization 1997 361 F.C.O. Fandinho et al. disease. The main characteristic used to select pa- tients was fever of unknown origin plus one of the following: hepatosplenomegaly, lung infiltrate, anae- mia, or elevated alkaline phosphatase. Furthermore, when possible, a CD4 lymphocyte T cell count (cells per il) was performed. Blood specimens A total of 3 blood samples were taken from each patient at 15-minute intervals by a trained phle- botomist. Each sample was collected in 2 Vacutainer tubes (Becton Dickinson, MD, USA) containing ethylenediaminetetraacetic acid (EDTA) as an anticoagulant. The temperature of the patients was measured at each sample. A total of 138 blood samples were obtained from the patients studied. Other specimens (e.g. sputum) were obtained for routine mycobacterial culture according to clinical indications. Procedure Upon arrival in the laboratory the contents of one Vacutainer tube were mixed with 30ml of 0.3% (w/v) sodium deoxicholate solution and incubated for 15min at room temperature. After centrifuga- tion for 25min at 3000g (4°C), the pellet was sus- pended in 1 ml of 0.2% bovine albumin saline (Fraction V, Sigma, St. Louis, MO, USA), and 0.2ml of the suspension was added to one tube with Middlebrook 7H9 broth and to two Lowenstein- Jensen slants. The contents of the second Vacutainer were di- rectly added to biphasic medium containing 5ml of Middlebrook 7H9 broth and 15ml of Middlebrook 7H10 agar. All mycobacterial isolates were identified using standard biochemical methods. Statistical analysis The x2 test (with Yates' correction for 2 x 2 tables) and Fisher's exact test were used to compare statis- tics between the patient groups. For continuous variables Student's t-test was used. In each case, a two-sided test with a 0.05 significance level was used. Results Mycobacterial growth was obtained from 70/138 specimens (from 30/50 (60%) patients). Positive cultures yielded M. tuberculosis in 19 (63%) and M. avium complex organisms in 11 (37%) patients. Only 4 patients with M. avium complex cultured from blood had mycobacteria isolated from cul- tures obtained from other clinical specimens. Of these, one patient had M. avium complex iso- lated from a liver sample (patient No. 3). Two patients with M. avium complex isolated from blood showed M. tuberculosis in cultures from either sputum (patient No. 46) or a lymph node biopsy (patient No. 1). One patient had Myco- bacterium gordonae isolated from faeces (patient No. 13). The majority of patients with positive blood cultures for M. tuberculosis had the same myco- bacteria species isolated from other clinical samples. However, 5 patients could be diagnosed with tuber- culosis only after the results of blood culture (pa- tients No. 6, 7, 18, 19, and 32). These results, includ- ing CD4 lymphocyte T cell counts, are depicted in Table 1. The mean of CD4 lymphocyte T cell counts for patients with positive blood cultures for M. avium complex was 15.9 cells per pl, while for patients with positive M. tuberculosis cultures the mean was 71.1 cells per ptl (range, 0-449 cells per p.l). A total of 19 (63%) patients with positive mycobacterial cultures were positive by both meth- ods used. With the lysis-centrifugation method blood cultures were positive for 54% (27/50) of patients, while with biphasic medium blood cul- tures were positive for 44% (22/50). A comparative analysis of the two methods showed no statistically significant difference (P > 0.05). The mean detection time for lysis-centrifugation and biphasic medium was 42.8 and 58.5 days, respectively. This prob- ably means that the lysis-centrifugation method recovered more viable mycobacteria from blood than did the biphasic medium, despite the lack of a statistically significant difference between them. A comparison of the results obtained with the Lowenstein-Jensen slants and the Middlebrook 7H9 broth used for the isolation of mycobacteria in the lysis-centrifugation method showed that for M. tuberculosis (34% and 42%, resp.) as well as for M. avium complex (45% and 36%, resp.) no statistically significant difference was observed (P > 0.05, results not shown). However, a striking difference was observed when we compared the two blood-culture methods according to the species of mycobacteria recovered (Table 2). The posi- tivity for M. avium complex was higher with lysis- centrifugation (10/11 = 91%) than with biphasic medium (5/11 = 45.4%) (P < 0.05), although the positivity for M. tuberculosis by the lysis-centrifuga- tion method (89.5%) and direct inoculation of bi- phasic medium (100%) was similar (P > 0.05) (Fig. 1). Samples from the same patient showed concor- dant results in 39/50 (78%) patients. WHO Bulletin OMS. Vol 75 1997362 Inexpensive method for diagnosis of mycobacterial infection Table 1: Mycobacterium culture results for the 50 AIDS patients studied Blood culture: Other specimen culture: Patient CD4 lymphocyte Biphasic Lysis-centrifugation Mycobacterium Collection Mycobacterium No. T cell counts medium method species identified sites species identified M. avium complex lymph node M. avium complex - M. avium complex liver M. tuberculosis BALa M. tuberculosis M. tuberculosis M. avium complex M. tuberculosis sputum M. avium complex - M. avium complex - M. avium complex faeces sputum M. tuberculosis M. tuberculosis sputum M. tuberculosis bone marrow sputum faeces M. tuberculosis bone marrow M. tuberculosis lymph node M. tuberculosis bone marrow M. tuberculosis sputum bone marrow lymph node M. tuberculosis M. tuberculosis liver sputum M. tuberculosis lymph node M. tuberculosis sputum M. tuberculosis sputum M. tuberculosis sputum M. tuberculosis sputum M. tuberculosis bone marrow sputum M. avium complex M. avium complex M. avium complex M. avium complex sputum sputum sputum sputum BAL lymph node M. tuberculosis M. avium complex M. tuberculosis M. tuberculosis M. gordonae M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis M. tuberculosis a BAL = bronchoalveolar lavage. b NP = not performed. WHO Bulletin OMS. Vol 75 1997 0 52 12 40 100 46 86 50 0 0 20 0 19 31 165 89 49 0 50 32 38 34 80 265 48 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 25 25 26 27 28 29 30 30 30 31 32 33 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Npb NP + + + ± + + + + + + + + + + ± + 151 15 59 33 135 + + + + ± + + + + + + + + + + + + + ± + + + + + 53 20 25 75 0 449 5 13 116 0 0 15 100 2 0 5 2 0 2 5 363 F.C.O. Fandinho et al. Table 2: Comparison of different culture media for the recovery of mycobacteria from the blood of AIDS patients studied M. tuberculosis: M. avium complex: Biphasic Biphasic LC-7H9a LC-LJb medium LC-7H9a LC-LJb medium Positive 16 13 17 8 10 5 Negative 3 6 0 3 1 6 Not performed 0 0 2 0 0 0 a Blood processed by lysis-centrifugation and added to Middlebrook 7H9 broth. b Blood processed by lysis-centrifugation and added to Lowenstein-Jensen slants. Fig. 1. Comparison of the positivity of the two culture methods studied, by species isolated. Biphasic medium, P < 0.05. 100 CO CD) QD 50- 0 0. Lysis-centrifugation Biphasic medium method * Mycobacterium tuberculosis a Mycobacterium avium complex Discussion Despite culture techniques making the detection of mycobacteria in blood easier (2, 7), few studies have focused on M. tuberculosis bacteraemia in AIDS patients (6,8, 9). This study suggests that M. tubercu- losis bacteraemia is common among AIDS patients in Brazil, and some workers have suggested that such bacteraemia has become more frequent (5). This study also suggests that blood culture is a valu- able diagnostic tool when tuberculosis is suspected, even when the culture of other specimens fails to yield the organism. Thus, our observations concur with those of Bouza et al. (10), but contrast with those of Shafer et al. (6), who did not find blood culture useful for the diagnosis of tuberculosis. M. avium complex is the most common bacterial infection in AIDS patients in the USA and Europe (3, 8), but its frequency in developing countries is almost unknown (11). Our results and those of Barreto et al (12) suggest that M. avium complex infection might be more common in AIDS patients in Brazil than is realized. The higher yield for isolation of M. avium com- plex with the lysis-centrifugation method arises be- cause M. avium complex organisms are not usually found free in plasma (1). The lysis of peripheral blood leukocytes releases intracellular mycobac- teria, which are concentrated by centrifugation, in- creasing the sensitivity of blood culture (4). As M. tuberculosis is a true pathogenic bacterium with greater capacity for causing disseminated disease, both methods were sensitive for detecting M. tuberculosis bacteraemia. Although Middlebrook 7H9 broth has generally been found to be more sensitive than tests on solid media, such as the Lowenstein-Jensen slants, the lack of significant dif- ference found here may be due to the relatively small sample size. The low CD4 lymphocyte T cell counts in patients with positive M. avium cultures from blood (15.9 cells per Fl) are concordant with the findings of some workers who estimate the risk for the occur- rence of M. avium bacteraemia to be much greater when the CD4 lymphocyte T cell count is less than 50 cells per 1t (13-15). However, the mean CD4 lymphocyte T cell count of 71.1 cells per 1t in pa- tients with M. tuberculosis cultured from blood was lower than most of those reported elsewhere (range, 23-742 cells per [tl) at the moment of tuberculosis diagnosis. Human immunodeficiency virus (HIV) positive patients with extrapulmonary tuberculosis have lower CD4 lymphocyte T cell counts (mean, 153 cells per [tl) than patients with pulmonary tuber- culosis only (16, 17). The characteristics used to se- lect patients for this study probably favoured the selection of more complicated cases of tuberculosis. This is supported by the high percentage of patients WHO Bulletin OMS. Vol 75 1997364 Inexpensive method for diagnosis of mycobacterial infection with disseminated tuberculosis (78%) in this study and by the fact that all patients had negative tuber- culin skin tests. All three consecutive blood samples from an individual showed concordant results in 39 patients (78%), suggesting that one blood sample is sufficient for the diagnosis of mycobacteraemia. Similar results were found in a retrospective study evaluating the usefulness of paired blood cultures, which demon- strated concordant results in 96% of pairs (18). None of our patients had fever at the moment of blood collection, which contrasts with the criteria for tak- ing blood samples for the diagnosis of bacteraemia caused by non-acid-fast organisms. In conclusion, this study shows that blood cul- ture can be performed using a non-commercial, sen- sitive, inexpensive method for the diagnosis of disseminated mycobacterial disease, and we recom- mend it for routine use in mycobacterial laboratories in developing countries; moreover, this study dem- onstrated a high frequency of disseminated M. avium complex infection in HIV-positive patients in Brazil. Acknowledgements This work was supported by CNPq and FINEP, Brazil. We thank M.G. Silva and O.M. Santos for their technical sup- port, and Dr A. Silva for her help with the Resume. Resume Diagnostic de la mycobacteriose generalisee: essai d'une methode simple et bon marche utilisable dans les pays en developpement Avec le developpement de l'6pid6mie de syndrome d'immunod6ficience acquise (SIDA), l'isolement des mycobact6ries dans le sang est devenu un probleme fr6quent pour les laboratoires d'analyses m6dicales. Deux methodes ont et6 utilisees dans la presente 6tude pour cultiver les mycobact6ries a partir des pr6levements de sang pratiqu6s chez les patients atteints de SIDA: 1) inoculation directe d'un milieu biphasique, et 2) m6thode non com- mercialis6e par lyse et centrifugation. Au total, trois 6chantillons de sang consecutifs ont ete pr6leves a 15 minutes d'intervalle chez chacun des 50 patients pr6sum6s cliniquement atteints de mycobact6riose gen6ralisee. Les mycobacteries ont cultive pour 70 des 138 prelevements de sang realis6s chez 30 patients (soit 60%). Ces cultures ont permis d'obtenir Mycobacterium tuberculosis chez 19 pa- tients (soit 63%), et des germes appartenant au complexe Mycobacterium avium chez 1 1 patients (soit 37%). Les cultures obtenues par lyse-centrifu- gation etaient positives chez 54% des patients, tandis que les cultures sur milieu biphasique 6taient positives chez 44% des patients (p > 0,05). Le nom- bre de cultures positives pour le complexe M. avium 6tait plus grand avec la lyse-centrifugation (91%) qu'avec le milieu biphasique (45,4%) (p < 0,05). Pour M. tuberculosis, la sensibilit6 de la lyse- centrifugation (89,5%) et celle de l'inoculation directe en milieu biphasique (100%) 6taient tou- tefois comparables (p > 0,05). La lyse-centrifuga- tion non commercialisee est une technique bon marche, fiable, et qui peut etre utilis6e comme m6thode de remplacement pour le diagnostic de la mycobact6ri6mie dans les pays en developpement. References 1. Meylan PR, Richman DD, Kornbluth RS. Charac- terization and growth in human macrophages of Mycobacterium avium complex strains isolated from the blood of patients with acquired immunodeficiency syndrome. Infection and immunity, 1990, 58: 2564- 2568. 2. Salfinger M et al. Comparison of three methods for recovery of Mycobacterium avium complex from blood specimens. Journal of clinical microbiology, 1988, 26: 1225-1226. 3. Inderlied CB, Kemper CA, Bermudez LE. The Mycobacterium avium complex. Clinical microbiology review, 1993, 6: 266-31 0. 4. Pitchenik AE, Fertel D. Tuberculosis and non- tuberculous mycobacterial diseases. Medical clinics of North America, 1992, 76: 121-171. 5. Baber TW, Craven DE, McCabe WR. Bacteremia due to Mycobacterium tuberculosis in patients with human immunodeficiency virus infection. Medicine, 1990, 69: 375-383. 6. Shafer RW et al. Frequency of Mycobacterium tuber- culosis bacteremia in patients with tuberculosis in an endemic area for AIDS. American review of respira- tory diseases, 1989, 140: 1611-1613. 7. Kiehn TE et al. Mycobacterium tuberculosis bacte- remia detected by the isolator lysis-centrifugation blood culture system. Journal of clinical microbiology, 1985, 21: 647-648. 8. Saltzman BA et al. Mycobacterium tuberculosis bact- eremia in the acquired immunodeficiency syndrome. Journal of the American Medical Association, 1986, 256: 390-391. 9. Truffot-Pernot C et al. Results of blood cultures for detection of mycobacteria in AIDS patients. Tubercle, 1989, 70: 187. 10. Bouza E et al. High prevalence of tuberculosis in AIDS patients in Spain. European journal of clinical microbiology and infectious disease, 1988, 7: 785-788. 11. Gilks CF et al. Disseminated Mycobacterium avium infection among HIV-infected patients in Kenya. Journal of acquired immune deficiency syndrome and human retrovirology, 1995, 8: 195-198. WHO Bulletin OMS. Vol 75 1997 365 F.C.O. Fandinho et al. 12. Barreto JA et al. Isolation of Mycobacterium avium complex from bone marrow aspirates of AIDS patients in Brazil. Joumal of infectious diseases, 1993, 168: 777-779. 13. Crowe SM et al. Predictive value of CD4 lymphocyte numbers for the development of opportunistic infec- tion and malignancies in HIV infected persons. Jour- nal ofacquired immune deficiency syndrome, 1991, 4: 770-776. 14. Nightingale SD et al. Incidence of Mycobacterium avium complex bacteremia in human immunode- ficiency virus-positive patients. Journal of infectious diseases, 1992, 165: 1082-1085. 15. Horsburg Jr CR. Current concepts: Mycobacterium avium complex infection in the acquired immuno- deficiency syndrome. New England joumal of medi- cine, 1991, 324: 1332-1338. 16. Theuer CP et al. Human immunodeficiency virus in- fection in tuberculosis patients. Journal of infectious diseases, 1990,162: 08-12. 17. Jones BE, Young SMM, Barnes PF. Relationship of the manifestations of tuberculosis to the CD4 counts in patients with human immunodeficiency virus infec- tion. American review of respiratory diseases, 1993, 148: 1292-1297. 18. Stone BL et al. Utility of paired blood culture and smears in diagnosis of disseminated Myco- bacterium avium complex infection in AIDS patients. Joumal of clinical microbiology, 1994, 32: 814- 842. 366 WHO Bulletin OMS. Vol 75 1997