Nathan Geffen, TAC
Current State of TB Testing
Current diagnostics for active TB are limited by cost, complexity, long diagnostic time, poor sensitivity or poor specificity. A point of care, affordable, easy-to-use, highly sensitive and specific test for active TB, analogous to HIV rapid tests, is urgently needed. Sputum microscopy is the standard technique used to diagnose TB in many developing country health facilities but has poor sensitivity especially in immune-compromised patients. It also requires a skilled health worker or technician.
The closest to a current gold standard for an active TB diagnosis is to perform both a liquid and solid culture test. These, however, have median turnaround times measurable in weeks.
Nucleic acid amplification techniques such as the FAST Plaque TB assay and PCR give results much faster but still take days and are much less sensitive and specific. They are also expensive. All these methods are compromised by the fact that some patients struggle to provide quality sputum samples.
Cepheid Gene Xpert
Mark Perkins of FIND presented data at the International TB conference in Cancun on a new nucleic acid amplification diagnostic technology for TB, the Cepheid Gene Xpert. The key data he presented has subsequently been published in the Journal of Clinical Microbiology by a group of researchers from FIND; New Jersey Medical School; Cepheid; Pham Ngoc Thach Hospital, Ho Chi Minh City and Makerere University, Kampala. [1, 2]
The Clinical Microbiology paper describes the Cepheid Gene Xpert as an integrated hands-free sputum-processing and real-time PCR system with rapid on-demand, near-patient technology, to simultaneously detect M. tuberculosis and rifampicin resistance. The device consists of a computer installed with Cepheid’s proprietary software and a machine –the smallest of which is about the size of a desktop computer– that takes cartridges loaded with sputum and reagents. The cartridges consist of a syringe barrel, a sonicator dome, a reverse-transcriptase PCR tube and a rotary valve.
The molecular beacon-based PCR assay that the system uses was shown to have high sensitivity and specificity on MDR isolates from the US, Spain, Mexico and India but without the machine is complex and labour-intensive to carry out. Rifampicin resistance is detected by mutations in the 81-base pair rifampicin resistance-determining region of the rpoB gene, which according to research cited by the authors occurs in 95 to 98% of rifampicin-resistant strains and are usually absent in rifampicin-susceptible ones. The authors indicate that rifampicin resistance is a good proxy for isoniazid resistance (and consequently MDR-TB, which is defined as resistance to both drugs).
The cartridges contain liquid sample-processing, PCR buffers and freeze-dried (lyophilized) PCR reagents. To prepare a cartridge, (1) PCR buffer is added to the sputum sample (ratio of 2 to 1), (2) the mixture is shaken for 5 seconds, (3) then allowed to stand for ten to 15 minutes, (4) 2 to 3 ml are transferred to the cartridge and (5) the cartridge is loaded into the machine, at which point the test can begin. It takes approximately two hours to give a result. Up to four cartridges can be loaded into the smallest machine.
The researchers report results of several tests to measure the machine’s accuracy. Only some are described here. They found that the device has a limit of detection (LOD, defined as the number at which there is a 95% probability of a positive result) in clinical samples of 131.0 colony forming units (CFU) per ml of sputum (95% CI: 106.2-176.4). By way of comparison, the gold standard BACTEC 960 MGIT culture test has a LOD of 10-100 CFU/ml and fluorescent microscopy has a LOD of 10,000 CFU/ml. 
Sputum samples from Vietnamese patients were used to validate the clinical efficacy of the device. The assay detected TB bacteria in all 29 smear-positive samples (95% CI: 85.4-100%) and 38 of 53 smear-negative samples that were culture-positive (71.7%; 95% CI: 57.4-82.8%).
To validate rifampicin resistance detection, 64 Ugandan cases were studied, in which the patients were being retreated for TB and therefore had a higher probability of rifampicin resistance. The device detected resistance in all nine known rifampicin resistance cases, giving a sensitivity of 100% (95% CI: 63.0-100%) and in 1/55 rifampicin-susceptible cases giving a specificity of 98.2%. (95% CI: 89.0-99.9%). The authors sequenced the rpoB gene of this discordant isolate and found a mutation commonly associated with rifampicin resistance. They therefore suggest that the Gene Xpert gave the correct result and the current standard rifampicin resistance test gave the wrong one. They point out that similar errors have been reported in other studies and furthermore that in this case the isolate was known to be resistant to both isoniazid and ethambutol, increasing the likelihood of rifampicin resistance. Furthermore, the device detected M. tuberculosis in 63 of the 64 sputum samples from culture-positive Ugandan patients (98.4%; 95% CI: 90.5 to 99.9%). Twenty laboratory control sputum samples from patients not suspected of having tuberculosis were negative.
The researchers measured the time required to analyse one and eight sputum samples, beginning at the moment that a sputum sample was given to a laboratory technician. The time-to-result for one sputum sample processed alone was just under two hours. The time-to-result for eight samples processed together was two hours. Total hands-on time was about five minutes.
Perkins indicated that sensitivity and specificity are greatly improved if more than one cartridge is used per patient instead of one, but this is not discussed in the Clinical Microbiology article.
Cepheid has not responded to a question about the cost of the device and its reagents. Based on informal discussions it appears the cheapest complete machine costs between US$25,000 and US$30,000 and handles four cartridges. The cartridges are $22 each. If three cartridges are needed per patient, then the cost per test is $66. This all needs to be confirmed, however. Cepheid’s website indicates that developing country public health systems will qualify for substantial discounts.
A largely hands-off automated test that produces highly sensitive and specific results in less than two hours would greatly improve patient-care. However it remains unclear that the Gene Xpert can be widely used in resource-poor settings, including most of Southern Africa.
Opinion about the device is cautious, with concern about its cost and lack of validation.
The machine is being tested in a National Health Laboratory Systems laboratory at the Site B Community Health Centre in Khayelitsha, Cape Town, with the assistance of Medecins Sans Frontieres. While this is a resource-poor setting, the health centre is relatively well resourced compared to the kinds of health facilities usually found in this kind of neigbourhood. The staff at this facility found the device convenient and easy to use, especially compared to smear microscopy. It is conceivable that at facilities with this kind of infrastructure, the machine will have a useful place.
There are several caveats about the machine:
- Unless the prices of both the machine and the cartridges come down dramatically, this device will not be rolled out beyond advanced laboratories and academic hospitals. If the machine proves useful in practice, there is potential for an activist-driven campaign to drive prices down, with the argument for price reductions perhaps strengthened by the fact that substantial public money was used to develop the device. The authors of the Clinical Microbiology article indicate that cost and cost-effectiveness studies are planned.
- It is unclear how complex the process of preparing the cartridge is and how much training is required for health workers to operate it, though the Khayelitsha technician operating the machine found it easy and quick to use. It will be important to validate the specificity and sensitivity of the device in pilot operational settings before it is made generally available.
- The authors point out that freezing may alter sputum viscosity and improve nucleic acid recovery from mycobacteria. They therefore state that the results need to be confirmed in larger prospective studies with fresh samples. (They did however document similar LODs regardless of whether the assay was performed with fresh or frozen sputum samples in analytic studies using spiked sputum.)
- The clinical test sample sizes are quite small with consequent large confidence intervals. Further validation, preferably from an operational setting such as Khayelitsha, is needed.
- It is unclear if Cepheid is or has obtained CE mark (demonstrating that it has met European Union consumer safety, health or environmental requirements) or FDA approval for the device.
- If the device does prove itself in practice a further question that will arise is to what extent rifampicin resistance should be used as a proxy for assuming isoniazid resistance and consequently MDR TB. Should isoniazid-resistance testing still be done if rifampicin resistance is detected? If so, what clinical decisions must be taken while waiting for results?
- Even in the best-case scenario, ie the Gene Xpert proves itself to be a robust highly specific and sensitive TB diagnostic tool in operational settings and its price is made affordable, we remain short of the holy grail point-of-care test for active TB. Nevertheless, it is possible that this device will have a useful place in the current TB diagnostic arsenal.
Thank you to Julian Duncan, Helen Lee, Gregg Gonsalves, Cheryl McDermid, Helen Cox and the Khayelitsha NHLS staff.
1. Perkins M. Xpert MTB/RIF: Automated molecular detection of TB and MDR at point of treatment. Presentation at 40th Union World Conference on Lung Health, December 2010.
2. Helb D et al. Rapid Detection of Mycobacterium tuberculosis and rifampin Resistance by Use of On-Demand, Near-Patient Technology. Journal of Clinical Microbiology, Jan. 2010, p. 229–237.
3. Roscigno G. Drug resistant TB and new diagnostics for people living with HIV: emerging results from FIND. 5th IAS Conference on HIV Pathogenesis, Treatment and Prevention, 19-22 July, Cape Town.