Re-decontamination of liquid mycobacterial cultures: Additional Mycobacterium tuberculosis yield in the era of Xpert MTB/RIF Ultra in Cape Town, South Africa

A retrospective review of liquid mycobacterial cultures was performed at a laboratory in South Africa from 01 January 2018 to 31 December 2018 to assess the increased yield in detecting Mycobacterium tuberculosis complex following sample re-decontamination. Only 9 of 99 (9%) re-decontaminated samples were culture positive for M. tuberculosis complex. Xpert MTB/RIF Ultra, concurrently performed on 7 of the 9 samples, detected M. tuberculosis complex in all but 1 sample. Re-decontamination of non-sterile samples did not increase the M. tuberculosis complex yield enough to offset the financial costs and additional labour in a laboratory that utilises the Xpert MTB/RIF Ultra system as a first-line diagnostic modality.


Introduction
Tuberculosis is caused by Mycobacterium tuberculosis, which belongs to a group of closely related, slow-growing mycobacteria collectively referred to as the M. tuberculosis complex. In 2018, there were an estimated 10 million cases of tuberculosis globally, with seven million cases notified and 1.45 million tuberculosis-related deaths 1 . South Africa, in the same year, had an estimated 301 000 new tuberculosis cases 1 .
Detection of M. tuberculosis complex by culture method remains the gold standard for confirming tuberculosis disease. 2 It can detect as few as 10 M. tuberculosis complex colony-forming units (CFU) per mL in clinical specimens. 3 Liquid culture increases the number of bacteriologically confirmed cases of tuberculosis by 20% -30%, even when rapid, sensitive nucleic acid amplification tests such as the Xpert Ultra (Cepheid, Sunnyvale, California, United States) 3,4 are used. Culture is advantageous as it allows for phenotypic drug susceptibility testing of the isolates. Processing of specimens for culture is, however, costly, labour intensive, expertise requiring, and timeconsuming (requiring up to six weeks to process to completion) owing to the slow proliferation time of M. tuberculosis complex. 5,6,7 The Xpert Ultra assay is a closed, cartridge-based, nucleic acid real-time polymerase chain reaction system that allows for the simultaneous detection of M. tuberculosis complex and rifampicin susceptibility in under 2 h. It can be performed directly on clinical samples and has a 5 CFU/mL -25 CFU/mL limit of detection; but, it has a lower sensitivity than tuberculosis culture in samples from people living with HIV and on samples where no acid-fast bacilli (AFB) are visualised on microscopy (smear negative). 3,4 Clinical specimens from non-sterile sites that are submitted to the mycobacteriology laboratory may be contaminated by other more rapid-growing bacteria. 5,8 These specimens undergo a digestion-decontamination procedure as recommended by the World Health Organization. The N-acetyl-l-cysteine sodium hydroxide digestion-decontamination allows M. tuberculosis to be cultured in a liquid culture medium despite reducing the organism viability by between 20% and 30%. 3,6,8,9 A proportion of cultures will however remain contaminated despite standard decontamination procedures. An acceptable rate for contaminated cultures in liquid media is 5% -8%. 5,6,7,8,10 A retrospective review of liquid mycobacterial cultures was performed at a laboratory in South Africa from 01 January 2018 to 31 December 2018 to assess the increased yield in detecting Mycobacterium tuberculosis complex following sample re-decontamination. Only 9 of 99 (9%) re-decontaminated samples were culture positive for M. tuberculosis complex. Xpert MTB/RIF Ultra, concurrently performed on 7 of the 9 samples, detected M. tuberculosis complex in all but 1 sample. Re-decontamination of non-sterile samples did not increase the M. tuberculosis complex yield enough to offset the financial costs and additional labour in a laboratory that utilises the Xpert MTB/RIF Ultra system as a first-line diagnostic modality.

Re-decontamination of liquid mycobacterial cultures: Additional Mycobacterium tuberculosis yield in the era of Xpert MTB/RIF Ultra in Cape Town, South Africa
Read online: Scan this QR code with your smart phone or mobile device to read online.
Global recommendations for re-decontamination exist, 7 but no published data on an acceptable compliance level for redecontamination could be found. The World Health Organization recommends that re-decontamination of contaminated liquid culture be performed when the first of two submitted cultures is 'positive for M. tuberculosis and contaminated' and the specimen requires drug susceptibility testing, while the second culture is negative for M. tuberculosis, and if the mycobacterial protein antigen 64 result is indeterminate, owing to the presence of contaminants.
The aims and objectives of the study were to determine the contamination rate at the Tygerberg Hospital (TBH) mycobacteriology laboratory; assess the increase in M. tuberculosis complex yield following re-decontamination in samples that had undergone Xpert Ultra testing; to assess laboratory non-compliance rates with regard to the recommended sample re-decontamination protocol; and to perform a cost analysis of the re-decontamination of specimens.

Ethical considerations
Ethical approval for this laboratory-based study was obtained from the Health Research Ethics Committee, Stellenbosch University, South Africa, (project identification: 14939, HREC X20/04/016). Informed consent was waived by the Stellenbosch University ethics committee for this laboratorybased study. Data remained confidential throughout the study.

Study setting
This study was conducted at South Africa's National Health Laboratory Service (NHLS) Division of Medical Microbiology and Immunology mycobacteriology laboratory located in TBH.
Annually, the TBH mycobacteriology laboratory processes approximately 10 000 specimens for M. tuberculosis complex culture from TBH. All samples from non-sterile sites for M. tuberculosis complex culture undergo decontamination with 1.25% N-acetyl-l-cysteine sodium hydroxide and the decontaminated samples are processed according to the Becton-Dickinson Mycobacterial Growth Indicator Tube (MGIT) testing protocol. 5 Specimen re-decontamination is performed on specimens from anatomical sites that are not easily obtainable or contaminated specimens that are microscopy positive for AFB. Easily obtainable anatomical specimens from non-sterile sites, such as urine and sputum specimens, are not re-decontaminated. 11,12 Study design and data analysis
Three samples were eligible for re-decontamination but did not undergo re-decontamination: two were sputum samples, in which AFB was observed on the contaminated culture and had Xpert Ultra testing which detected rifampicin-susceptible M. tuberculosis complex, and one was a bronchoalveolar lavage sample in which a second sample was contaminated.
Non-compliance to re-decontamination was low, with 97.8% of samples correctly assessed for re-decontamination.
Of the 10 contaminated samples that were microscopy positive for AFB and that underwent re-decontamination, all were sputum samples; non-tuberculosis mycobacteria were isolated from three samples and M. tuberculosis was isolated from seven. Only six of the seven samples that had M. tuberculosis complex isolated on culture underwent Xpert Ultra testing and were all positive for M. tuberculosis complex.
In summary, only 9 of 99 (9%) re-decontaminated samples were culture positive for M. tuberculosis complex. On Xpert Ultra testing, 6 of the 9 (67%) tested positive for M. tuberculosis complex, 1 (11%) tested negative for M. tuberculosis complex, and 2 (22%) did not undergo Xpert Ultra testing (but were smear positive for AFB on an auramine stain).

Discussion
We found that 6% of all specimens undergoing mycobacterial culture at the TBH mycobacteriology laboratory were contaminated, which is in keeping with internationally accepted contamination standards of 5% -8% in liquid media. 5,6,7,8,10 Non-compliance to the recommended re-decontamination standard operating procedure was low at 2.2%. These three samples were not re-decontaminated as the Xpert Ultra had detected M. tuberculosis complex on two samples, and the third sample had additional specimens still incubating.
We found that of the seven re-decontaminated samples that were microscopy positive for AFB and culture positive for M. tuberculosis complex, six (86%) were also

Limitations
This was a laboratory-based study using routinely available data. The treatment status of patients who submitted samples could not be obtained, which may have impacted on M. tuberculosis complex yield following re-decontamination.
Owing to the small sample size eligible for re-decontamination, and varying laboratory decontamination protocols, the findings in this study may not allow generalisation of our findings to other centres.

Conclusion
The poor increase in yield of M. tuberculosis complex after redecontamination of samples reflects the efficiency of the South Africa tuberculosis-testing algorithm, which employs Xpert Ultra testing, that has low limit of detection. In our high-burden tuberculosis setting, routine re-decontamination is not cost-effective and not recommended in specimens that have undergone Xpert Ultra testing.