Skip to main content
Premium Trial:

Request an Annual Quote

Pooling Samples Leads to 30 Percent Cost Savings for Xpert TB Testing


NEW YORK (GenomeWeb) – Even with concessionary pricing and international investment, the cost of Cepheid's GeneXpert MTB/RIF for Mycobacterium tuberculosis and rifampin resistance testing sometimes strains impoverished areas.

In an effort to further mitigate the cost of testing, an international team has shown that testing pools of four patient samples per $9.98 cartridge, followed by individual testing if a pool comes up positive, saved about 30 percent on cartridge costs and around 60 percent of operator time. The team published its study last week in the Journal of Clinical Microbiology.

"Although the GeneXpert is subsidized by many countries, it is very pricey — nearly $10 per test — and quite often they have stock outs," Luis Cuevas, a professor of international public health and epidemiology at the Liverpool School of Tropical Medicine and an author on the study, told GenomeWeb in an interview.

Pooling is frequently used to screen samples in blood banks, and has been shown effective for acute HIV testing. Cuevas said he believes this is the first study using pooled sputum on the GeneXpert.

In their study, the investigators assayed 738 sputum samples from adults in Nigeria's capital, Abuja, suspected of having pulmonary TB.

These were grouped into 183 pools of four, plus two pools containing three samples each. Samples were also run individually, and the individual result was the comparator.

The researchers found 81 of the pools were positive for TB, 101 were negative and three reported a failed result even after retesting. In all, the pooled approach showed 99 percent agreement with individual testing, the study said.

The savings from pooled testing has a sweet spot that is related to prevalence, Cuevas explained. "If you find, let's say, 90 percent of your pools are positive, then you have to retest all of them basically, and you wouldn't save anything."

For the JCM study, some of the samples were collected in a clinic, and others were from a community canvas of slums and rural areas. These two groups had different prevalence of infection, and the estimated savings was four times higher for the low-prevalence community-derived samples.

The study also modeled different pool sizes for the two groups. The larger the pool size, the more likely a pool will be positive and require individual testing, essentially duplicating the efforts.

Larger pool sizes may also dilute the positive sample. The study suggested five falsely negative pools may have been due to low bacillary load in one patient sample. However, Cuevas said the newer Xpert TB Ultra assays, due out in 2016, are expected to be more sensitive, and are reported to have a multi-copy TB target to boost sensitivity by about 10- to 15-fold.

The main drawback to the pooled testing, however, is that it has the potential to cause laboratory-acquired TB infection.

"The more you manipulate sputum, the higher the risk of infection and cross-contamination," Cuevas noted. "You need to do this in a laboratory that has good quality infection control and a staff that is prepared to do this carefully."

The study was done in a research lab at the Zankli Medical Center in Abuja, and Cuevas said the method is probably not appropriate for use "below the district level."

He further explained that the total savings in this study, $2,295, would equate to 230 additional test cartridges. Governments pay staff salaries, while international aid groups often pay for consumables, so this savings would likely be reinvested in tests.

But there aren't enough tests currently to be used on all patients. "At the moment the algorithms say that this test should only be used for people with HIV, or people who have repeatedly tested negative for other tests, and so on, so if you had the possibility to test more patients with the same number of cartridges, then that will spread more thinly," he said.

Pooled testing may not ultimately reduce Cepheid's sales, either. "The number of cartridges [sold] is growing exponentially, they don't have enough capacity to produce as many as needed, and less than 10 percent of people with TB are tested with these cartridges; anything that facilitates using them more efficiently is good," Cuevas said.

Newer, more sensitive tests, may also make this method appropriate for pediatric and extrapulmonary TB testing, where the Xpert MTB/RIF has faltered somewhat.

Although the lab in Abuja liked the method and thought it would help them to use their resources better, testing policies have not yet changed, Cuevas said. This would require an official recommendation from the World Health Organization.

"That is usually done when there are five or six studies that have repeated the same results in several locations," Cuevas said.

He cited the example of adjusting a protocol for the number of sputum samples required for microscopy from three to two. A total of 7,000 patients in seven countries had to be evaluated to sufficiently prove to the WHO that a third sample was superfluous. 

"Hopefully more people will hear about this study, repeat it, and then in three or four years time it could be adopted as a recommendation," Cuevas said.

Finally, only 16 percent of patients in the study had TB, and Cuevas noted healthcare beyond TB is important to address, too. Chronic cough in those cases could be caused by recurrent pneumonia, cystic fibrosis, asthma, or even exposures from cooking with charcoal and wood.

"In some places where we work, when we tell people, 'You have tuberculosis, don't worry you will be treated,' some of them feel relieved because you have given them a diagnosis and said you will be cured," Cuevas explained. "If you say, 'You don't have TB but we don’t know what you have,' they still have the problem."