Abstract
We compared Mycobacterium tuberculosis sputum culture recovery and contamination rates between Lowenstein-Jensen medium (LJ) containing the following decontaminants and LJ alone: (i) PANTA (n = 299), (ii) Selectatab-MB (n = 299), and (iii) penicillin G (n = 234). The contamination rate for LJ alone was approximately 31%, versus 5.0% for PANTA-containing, 2% for Selectatab-containing, and 9% for penicillin-containing media (P < 0.001). M. tuberculosis isolation rates were 9.8%, 17%, 18%, and 12% for standard LJ, PANTA, Selectatab, and penicillin cultures, respectively.
TEXT
Tuberculosis (TB), caused by Mycobacterium tuberculosis, is a high-burden disease, with over 95% of TB cases occurring in developing countries (1). Culture remains the gold standard for detecting TB and drug sensitivity testing (2). However, the effectiveness of culture systems is greatly undermined by contamination with bacteria and fungi (3). Contamination reduces the proportion of interpretable results, thereby limiting the diagnostic value of culture systems. Contaminated cultures must be repeated, at additional cost to public health systems, which delays or ultimately prevents TB diagnosis.
Decontamination techniques have been implemented at all stages of specimen collection and processing. Antimicrobial and antifungal oral rinse solutions can be used prior to sputum expectoration (4), and decontamination using cetrimide, oxalic acid, and sulfuric acid can be used during sputum processing (5, 6). A variety of decontamination methods have been developed for various M. tuberculosis culture media. The liquid-culture mycobacterium growth indicator tube (MGIT) contains a rapid and sensitive culture medium with low contamination rates. However, its high cost is prohibitive for resource-limited settings (RLS). The Bactec 960 liquid culture system uses PANTA (polymyxin B, amphotericin B, nalidixic acid, trimethoprim, and azlocillin) to reduce contamination (7). A combination of vancomycin, amphotericin B, and nalidixic acid (VAN) has also been shown to minimize contamination in the Middlebrook 7H11 agar formulation (7).
Although Lowenstein-Jensen medium (LJ) remains the most common medium used in RLS (8), few studies have assessed the efficacy of antibiotics on LJ. We assessed the performance of PANTA (Becton, Dickinson, Franklin Lakes, NJ, USA), mycobacterium Selectatab (Selectatab-MB; Mast, Merseyside, United Kingdom), and penicillin G (Sigma-Aldrich, Gillingham, United Kingdom) in reducing contamination and improving the yield and recovery of M. tuberculosis in LJ. We also characterized a set of contaminants isolated from antibiotic-free LJ.
Sputum samples were collected as part of routine clinical care from patients with presumptive TB attending the outpatient department of Mbarara Regional Referral Hospital (MRRH) and the Immune Suppressed Syndrome (ISS) (HIV) clinic, Mbarara, Uganda, between October 2011 and January 2012. Sputum smears were prepared for microscopic examination using the auramine-LED fluorescence technique, and results were reported according to World Health Organization (WHO) guidelines (9, 10).
All laboratory procedures were performed at the laboratory of the Epicentre Research Centre in Mbarara, Uganda, which is monitored by the Institute of Tropical Medicine of Antwerp, Antwerp, Belgium. We first processed the sputum for culture using the N-acetyl-l-cysteine-sodium hydroxide method (11) and inoculated 200 μl of sediment into plain LJ or antibiotic-containing LJ tubes. Two distinct groups of specimens were evaluated by matched comparison of decontamination techniques. The first group of 299 sputum specimens was analyzed in triplicate with plain LJ, LJ with PANTA, and LJ with Selectatab-MB. A second group of 234 sputum specimens was analyzed in duplicate to compare plain LJ against LJ containing penicillin G. All media were prepared according to manufacturer guidelines. Penicillin-treated LJ was prepared to make a final concentration of 100 IU/ml before coagulating. The costs of PANTA, Selectatab-MB, and penicillin per test were approximately $0.80, $0.40, and $0.14, respectively.
Tubes were incubated at 37°C for a maximum of 8 weeks. Cultures showing no growth after 8 weeks of incubation were reported as negative. Liquefied or discolored (dark green) LJ media or LJ slants with colonies of non-acid-fast bacteria were considered contaminated. Internal quality control for microscopy and LJ was performed as previously described (12). For a subset of 20 contaminated LJ tubes, culture liquid was inoculated on blood agar and incubated at 37°C for 48 h. After growth was detected, an isolated colony was picked for further characterization using Gram staining and API Gallery (bioMérieux, Durham, NC, USA).
Data were double entered into Epidata (version 3.1; Epidata Association, Odense, Denmark) and analyzed using Stata SE version 12 software (StataCorp, College Station, TX, USA). We compared contamination rate, yield (proportion of M. tuberculosis culture-positive specimens among smear-positive specimens), and recovery (proportion of M. tuberculosis culture-positive specimens by sputum smear result) between plain LJ and antibiotic-containing LJ using McNemar's test for paired samples.
Contamination rates were 32.1% in plain LJ, 5.0% in LJ-PANTA, and 2.3% in LJ–Selectatab-MB among 299 matched specimens (Table 1). While both decontaminant-containing cultures had significantly lower contamination rates than plain LJ (P < 0.001), the difference in rates between the two antibiotic-containing media was not significantly different (P = 0.10). Among the 234 specimens cultured in both penicillin-treated medium and plain LJ, contamination rates were 8.9% and 30.7%, respectively (P < 0.001).
TABLE 1.
General M. tuberculosis yield and contamination ratio in PANTA-, Selectatab-MB-, and penicillin-treated LJ versus antibiotic-free LJ
| Antibiotic and medium |
M. tuberculosis-positive samples |
Contaminated samples |
||
|---|---|---|---|---|
| No. (%) | P | No. (%) | P | |
| PANTA | 0.38 | <0.0001 | ||
| LJ-PANTA (n = 299) | 52 (17.4) | 15 (5.0) | ||
| Plain LJ (n = 299) | 33 (11.0) | 96 (32.0) | ||
| Selectatab-MB | 0.38 | <0.0001 | ||
| LJ–Selectatab-MB (n = 299) | 55 (18.4) | 7 (2.3) | ||
| Plain LJ (n = 299) | 33 (11.0) | 96 (32.1) | ||
| Penicillin | 0.34 | <0.0001 | ||
| LJ-penicillin (n = 234) | 29 (12.4) | 21 (8.9) | ||
| Plain LJ (n = 234) | 23 (9.8) | 72 (30.8) | ||
TB-positive-culture rates were 11.0% (33/299), 17.4% (52/299), and 18.4% (55/299) for plain LJ, LJ-PANTA, and LJ–Selectatab-MB, respectively, with a significantly higher proportion in the decontaminated samples than in plain LJ. In group 2, positive-culture rates were 12.4% (29/234) and 9.8% (23/234) in penicillin-treated and plain LJ, respectively (P = 0.34) (Table 1).
Recovery differed by culture medium type. Among specimens graded 2+/3+ (corresponding to more than 10 acid-fast bacilli per field in at least 50 fields) by microscopy, recovery was 58.8% (20/34) in plain LJ, 88.2% (30/34) with PANTA (McNemar's exact P = 0.01), and 97.1% (33/34) with Selectatab-MB (McNemar's exact P < 0.001) (Table 2). Overall, recovery was 58.7% (27/46) in plain LJ, which increased to 87% (40/46) with PANTA (McNemar's exact P < 0.001) and to 93.5% (43/46) with Selectatab-MB (McNemar's exact P < 0.0001). The difference in recovery between the two treated-medium strategies was not statistically significant (McNemar's exact P = 0.25). The recovery rate in LJ with penicillin was 78.6%.
TABLE 2.
Contamination and Mycobacterium tuberculosis and recovery according to microscopy grading
| Microscopy gradea | Antibiotic | No. (%) |
|
|---|---|---|---|
| Contaminated cultures | M. tuberculosis-positive tubes | ||
| No AFB (n = 253) | PANTA | 10 (4.0) | 12 (4.7) |
| Selectatab-MB | 5 (2.0) | 12 (4.7) | |
| None | 78 (30.8) | 6 (2.4) | |
| Scanty/+1 (n = 12) | PANTA | 1 (8.3) | 10 (83.3) |
| Selectatab-MB | 1 (8.3) | 10 (83.3) | |
| None | 4 (33.3) | 7 (58.3) | |
| 2+/3+ (n = 34) | PANTA | 4 (11.8) | 30 (88.8) |
| Selectatab-MB | 1 (2.9) | 33 (97.1) | |
| None | 14 (41.2) | 20 (58.8) | |
Microscopy grading was done according to the WHO/IUATLD scale (1). AFB, acid-fast bacilli.
We evaluated the effectiveness of selected antibiotics in reducing contamination in LJ. Notably, contamination rates decreased from approximately 30% with plain LJ to ≤5% for LJ with the addition of PANTA or Selectatab-MB (Table 3). Notwithstanding the improved decontamination rates, positive culture rates increased from 11% to 17 to 18% for the decontaminated media, and yield increased from 56% to 89 to 97% for specimens that were graded 2/3+ by sputum microscopy. While penicillin G-containing media also demonstrated lower rates of contamination, yield and recovery improvements were not seen to the same degree. Selectatab-MB contains ticarcillin, a penicillin which acts on Gram-negative bacteria, particularly Pseudomonas aeruginosa. It is also one of the few antibiotics capable of treating Stenotrophomonas maltophilia, one of the major contaminants isolated in this study, and this is probably why Selectatab-MB performed better overall. In summary, the addition of PANTA and Selectatab-MB to standard LJ appears to reduce contamination rates without reducing the growth or recovery of TB in culture.
TABLE 3.
Characterization of contaminants isolated from antibiotic-free LJ tubes (n = 20)
| Gram reaction | Organism | No. (%) detected |
|---|---|---|
| Positive | Streptococcus spp. | 2 (16.7) |
| Staphylococcus spp. | 6 (49.9) | |
| Cellulomonas spp. | 2 (16.7) | |
| Fungi | 2 (16.7) | |
| Total | 12 (60) | |
| Negative | Stenotrophomonas spp. | 3 (37.5) |
| Aeromonas spp. | 2 (25) | |
| Serratia spp. | 1 (12.5) | |
| Brevibacterium spp. | 1 (12.5) | |
| Enterobacter spp. | 1 (12.5) | |
| Total | 8 (40) |
Our contamination rates with plain LJ were approximately 31%, greater than the recommended threshold of 5% for laboratories that receive freshly collected sputum samples or 5 to 10% for settings where samples take several days to reach the laboratory (9). This may be due to enrollment of patients with cough for more than 2 weeks and no other TB symptoms. Although LJ contains malachite green, which has antibiotic properties, several other groups have reported similarly high contamination rates (13). For example, in a recent study carried out at the Zambian National Reference Laboratory, an LJ contamination rate of 14.9% was reported (14). These data reinforce the need for improved decontamination methods for laboratories that rely on LJ culture for TB diagnosis and susceptibility testing.
Other strategies to reduce contamination rates have been reported, with varying results. In our setting, we have tested mineral water rinsing prior to sputum production but with limited effects (Y. Boum II, unpublished data). A recent study on the use of oral rinse solutions (chlorhexidine and nystatin) showed a lower contamination rate in samples collected from patients with presumptive TB, but this was also shown to affect the recovery of M. tuberculosis (4).
Conclusion.
Both Selectatab-MB and PANTA offer excellent solutions for reducing contamination in LJ. Due to its low cost, Selectatab-MB should be further explored for use as a medium additive to improve recovery rates and decrease contamination for LJ-based culture media in resource-limited settings.
ACKNOWLEDGMENTS
We thank the Mbarara Regional Referral Hospital administration for allowing us carry out this study in the hospital. We also thank the entire staff and administration of Epicentre, especially the TB Lab Team, for their participation at various stages of the study: Ivan T. Mugisha, Gorette Kyomugisha, Johnnie Mary Ngattu, Esther Turyashemererwa, Victoria Katawera, and Henry Munyambabazi. We are grateful to Armand Van Deun and Mourad Gumusboga from Institute of Tropical Medicine, Antwerp, Belgium, for the technical support provided to this study and to our laboratory. We are also grateful to FHI 360's Mario Chen, Laura Philips, Kenneth Schultz, and Suzanne Fischer. We thank Yolanda Muller, Juliet Mwanga-Amumpaire, and Patricia Kahn for their critical review of the manuscript.
Footnotes
Published ahead of print 30 April 2014
REFERENCES
- 1.Phanouvong SVH, Blanc L, Wondemagegnelm E, Rägo L, Moore T, Mboussa J, Norval P, Matji R. 2002. Operational guide for national tuberculosis control programmes on the introduction and use of fixed-dose combination drugs. World Health Organization, Geneva, Switzerland [Google Scholar]
- 2.Bello AK, Njoku CH. 2005. Tuberculosis: current trends in diagnosis and treatment. Niger. J. Clin. Pract. 8:118–124 [PubMed] [Google Scholar]
- 3.Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH. 1999. Manual of clinical microbiology, 9th ed. ASM Press, Washington, DC [Google Scholar]
- 4.Kalema N, Boon SD, Cattamanchi A, Davis JL, Andama A, Katagira W, Everett C, Walter N, Byanyima P, Kaswabuli S, Worodria W, Huang L. 2012. Oral antimicrobial rinse to reduce mycobacterial culture contamination among tuberculosis suspects in Uganda: a prospective study. PLoS One 7:e38888. 10.1371/journal.pone.0038888 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Gautam G, Singh TSK, Sharma TD, Ghimire M, Regmi SM. 2010. Decontamination of sputum sample for the isolation of mycobacterium tuberculosis. J. Gandaki Med. College 3:32–35 [Google Scholar]
- 6.Peres RL, Palaci M, Loureiro RB, Dietze R, Johnson JL, Maciel EL. 2011. Reduction of contamination of mycobacterial growth indicator tubes using increased PANTA concentration. Int. J. Tuberc. Lung Dis. 15:281–283 [PubMed] [Google Scholar]
- 7.Chang CL, Park TS, Oh SH, Kim HH, Lee EY, Son HC, Kim CM. 2002. Reduction of contamination of mycobacterial growth indicator tubes with a modified antimicrobial combination. J. Clin. Microbiol. 40:3845–3847. 10.1128/JCM.40.10.3845-3847.2002 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Satti L, Ikram A, Abbasi S, Butt T, Malik N, Mirza IA. 2010. Evaluation of BACTEC MGIT 960 system for recovery of Mycobacterium tuberculosis complex in Pakistan. Malaysian J. Microbiol. 6:203–208 http://web.usm.my/mjm/issues/vol6no2/research14.pdf [Google Scholar]
- 9.Narvaiz de Kantor I, Frieden TR, Laszlo A, Fabio L, Pierre-Yves N, Hans LR, Pedro V, Karin W. 1998. Laboratory services in tuberculosis control. WHO, Geneva, Switzerland [Google Scholar]
- 10.International Union Against Tuberculosis and Lung Disease. 2000. Sputum examination for tuberculosis by direct microscopy in low income countries. International Union Against Tuberculosis and Lung Disease, Paris, France [Google Scholar]
- 11.Sharma M, Misra RN, Gandham NR, Jadhav SV, Angadi K, Wilson V. 2012. Comparison of modified Petroff's and N-acetyl-l-cysteine-sodium hydroxide methods for sputum decontamination in tertiary care hospital in India. Med. J. DY Patil Univ. 5:97–100. 10.4103/0975-2870.103323 [DOI] [Google Scholar]
- 12.Boum Y, Orikiriza P, Rojas-Ponce G, Riera-Montes M, Atwine D, Nansumba M, Bazira J, Tuyakira E, De Beaudrap P, Bonnet M, Page AL. 2013. Use of colorimetric culture methods for detection of Mycobacterium tuberculosis complex isolates from sputum samples in resource-limited settings. J. Clin. Microbiol. 51:2273–2279. 10.1128/JCM.00749-13 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rishi S, Sinha P, Malhotra B, Pal N. 2007. A comparative study for the detection of Mycobacteria by BACTEC MGIT 960, Lowenstein Jensen media and direct AFB smear examination. Indian J. Med. Microbiol. 25:383–386. 10.4103/0255-0857.37344 [DOI] [PubMed] [Google Scholar]
- 14.Muyoyeta M, Schaap JA, De HP, Mwanza W, Muvwimi MW, Godfrey-Faussett P, Ayles H. 2009. Comparison of four culture systems for Mycobacterium tuberculosis in the Zambian National Reference Laboratory. Int. J. Tuberc. Lung Dis. 13:460–465 [PubMed] [Google Scholar]