The sensitivity of the QuantiFERON®-TB Gold Plus assay in Zambian adults with active tuberculosis

SUMMARY SETTING AND OBJECTIVE: To investigate the sensitivity of the new interferon-gamma release assay (IGRA), QuantiFERON®-TB Gold Plus (QFT-Plus), for active TB (used as a surrogate for latent tuberculous infection) in a Zambian TB clinic. DESIGN: Consecutive smear or Xpert® MTB/RIF-positive adult (age ⩾18 years) pulmonary TB patients were recruited between June 2015 and March 2016. Venous blood was tested using QFT-Plus. The sensitivity was defined as the number positive divided by the total number tested. Using logistic regression, factors associated with positive QFT-Plus results were explored. RESULTS: Of 108 patients (median age 32 years, interquartile range 27–38; 73% male; 63% human immunodeficiency virus [HIV] positive), 90 were QFT-Plus-positive, 11 were negative and seven had indeterminate results; sensitivity was 83% (95%CI 75–90). There was no difference in sensitivity by HIV status (HIV-positive 85%, 95%CI 75–93; n = 68 vs. HIV-negative 80%, 95%CI 64–91; n = 40; P = 0.59). In models adjusted for age alone, CD4 cell count <100 cells/μl (OR 0.15, 95%CI 0.02–0.96; P=0.05) and body mass index <18.5 kg/m2 (OR 0.27, 95%CI 0.08–0.91; P = 0.02) were associated with decreased odds of positive QFT-Plus results. CONCLUSION: Overall, the sensitivity of QFT-Plus is similar to that of the tuberculin skin test and other IGRAs. While overall sensitivity is not affected by HIV status, QFT-Plus sensitivity was lower among people living with HIV/acquired immune-deficiency syndrome with severe immunosuppression.

C O N C L U S I O N : Overall, the sensitivity of QFT-Plus is similar to that of the tuberculin skin test and other IGRAs. While overall sensitivity is not affected by HIV status, QFT-Plus sensitivity was lower among people living with HIV/acquired immune-deficiency syndrome with severe immunosuppression. K E Y W O R D S : interferon-gamma release assays; latent tuberculous infection; validity THE INCREASED RISK of active tuberculosis (TB) among people living with human immunodeficiency virus/acquired immune-deficiency syndrome (PLHIV) 1,2 can be substantially decreased by treating latent tuberculous infection (LTBI) with isoniazid preventive therapy (IPT). 3 However, there are numerous barriers to IPT use, 4,5 including the lack of a good diagnostic test for LTBI among PLHIV. 6 Two imperfect but widely used tests for LTBI that measure the host's cell-mediated immune response to Mycobacterium tuberculosis antigens are the tuberculin skin test (TST) and interferon-gamma (IFN-c) release assays (IGRAs) (QuantiFERON w -TB Gold In-Tube [QGIT], QIAGEN, Hilden, Germany; and the T-SPOT w .TB test, Oxford Immunotec, Abingdon, UK). 6,7 Both tests have poor and comparable sensitivity among PLHIV (TST sensitivity 45% and IGRA sensitivity 60-75%). [7][8][9] CD4þ T-lymphocytes play a critical role in the immunological control of M. tuberculosis through the secretion of IFN-c; 10 QGIT mainly evokes and measures this CD4-mediated response. 11 There is growing evidence to suggest that CD8þ T-lymphocytes also play a role in M. tuberculosis control through the production of IFN-c, 10,[12][13][14] and M. tuberculosisspecific CD8þ T-lymphocytes have been identified in subjects with LTBI and active TB. 14,15 The QGIT has thus now been modified to include two antigen tubes-TB1, which contains a long peptide cocktail aimed at stimulating CD4þ T-lymphocytes, and TB2, which contains an additional shorter peptide cocktail aimed at stimulating CD8þ T-lymphocytes in addition to CD4þ T-lymphocytes 16 -now called QuantiFER-ON w -TB Gold Plus (QFT-Plus). The antigen TB 7.7, found in QGIT, has been removed. 16 This approach, which negates the dependence on CD4-mediated responses alone, could potentially increase the test's validity for LTBI among PLHIV.
We conducted a study in Lusaka, Zambia, a setting with high TB and HIV prevalence, to determine the sensitivity of QFT-Plus. As a gold standard for LTBI does not exist, active TB was used as a surrogate, as patients with active TB are infected with M. tuberculosis.

Study population and procedures
The study was based in a government health facility. All patients with suspected pulmonary TB underwent sputum testing with either fluorescent microscopy or the Xpert w MTB/RIF assay (Cepheid, Sunnyvale, CA, USA); pulmonary TB was defined as at least one positive sputum sample on either smear or Xpert. Culture is not routinely performed. HIV counselling and testing according to national guidelines is routinely offered to all patients investigated and/or treated for TB. For HIV testing, sequential bloodbased rapid antibody tests with Determine w HIV-1/2 Antibody (Alere, Waltham, MA, USA) is used first, which, if reactive, is followed by the Uni-Gold TM Recombigen HIV-1/2 Antibody test (Trinity Biotech, Bray, Ireland). 17 Between June 2015 and March 2016, consecutive adult (age 718 years) pulmonary TB patients who were 62 days from starting anti-tuberculosis treatment were recruited for the study. Written voluntary informed consent was provided by all study participants. Information on demographics (age, sex), TB and HIV diagnosis and physical measurements (height and weight) were obtained. Body mass index (BMI), defined as (weight in kg)/(height in m) 2 , was classed into standard categories (,18.5 underweight, 18.5-24.9 normal weight, 25-29.9 overweight, 730 obese). 18 Venous blood for testing with QFT-Plus was collected into lithium heparin tubes, stored at room temperature and transported within 8 h to the laboratory for processing according to the manufacturer's specification. 16 Venous blood was also collected for CD4 cell count estimation if patients were HIV-co-infected. [IU]/ml); a positive result on either antigen tube, TB1 or TB2, was considered positive. The upper limit for quantifying IFN-c concentrations was 10 IU/ml. Tlymphocyte estimation (CD4 þ , CD8 þ and CD3 þ ) was performed within 48 h of blood collection using BD FACSCount reagents and a BD FACSCount flowcytometer (BD, Sparks, MD, USA), according to the manufacturer's specification. The upper limit for CD8 cell count measurement was 2000 cells/ll.

Patient follow-up
To explore whether negative/indeterminate QFT-Plus results were due to TB and/or HIV-associated immunosuppression, which should improve upon treatment, patients with negative or indeterminate QFT-Plus results at enrolment underwent repeat testing with QFT-Plus at~1-2 months post-enrolment.

Statistical analysis
The target sample size was 100 pulmonary TB patients, aiming to determine an overall QFT-Plus sensitivity of 74% (based on QGIT sensitivity in this population 11 ) with a precision of 69% assuming 95% confidence intervals (CIs). A higher sensitivity of 85% would give a precision of 67%.
Data were analysed using STATA, version 12.0 (StataCorp LP, College Station, TX, USA). Sensitivity was defined as the number of positive results divided by the total number tested at enrolment; the denominator included indeterminate results. To explore factors associated with positive QFT-Plus results, negative and indeterminate QFT-Plus categories were combined to give a binary variable. Logistic regression was used to determine odds ratios (ORs), 95%CIs and P values from the likelihood ratio test. Age was considered a forced variable a priori. Due to the small number of negative/indeterminate results, a multivariable analysis could not be undertaken. The analysis was therefore limited to a univariate analysis consisting of base models (each factor adjusted for age) alone. Results are presented following STARD (Standards for Reporting of Diagnostic Accuracy) guidelines. 19 Ethics approval Ethics approval for the study was obtained from the Biomedical Research Ethics Committees of the University of Zambia, Lusaka, Zambia, and the London School of Hygiene & Tropical Medicine, London, UK.

DISCUSSION
This is the first study to describe the sensitivity of the new QFT-Plus assay for TB in a population with a high prevalence of HIV. The assay's overall sensitivity was maintained, regardless of HIV status. However, sensitivity was lower among PLHIV with severe immunosuppression.
A multicentre study across six European hospitals found a QFT-Plus sensitivity of 86% (102/119; 95%CI 78-91%) for bacteriologically confirmed TB, 20 which is comparable to our findings. All four patients who were TB-HIV-co-infected had positive QFT-Plus results. 20 While these numbers are too small to draw any firm conclusions, they corroborate our findings of high QFT-Plus sensitivity among PLHIV. QFT-Plus specificity among low-risk subjects in Europe was 97% (95%CI 92-99). 20 Similar results have been reported from Germany and Japan, where head-to-head comparisons of QFT-Plus and QGIT among predominantly immunocompetent patients showed similar test sensitivities. 21,22 A study using the same methods as our study was conducted by Raby et al. at the same health facility in 2007 to determine QGIT and TST sensitivity in a similar population. 11 The proportion of TB-HIV-coinfected patients and the median CD4 cell counts among PLHIV were similar in both studies (Table 4). Overall QFT-Plus sensitivity in our study was similar to QGIT and TST sensitivity; however, unlike QGIT and TST, overall QFT-Plus sensitivity was not affected by HIV status. Studies of QGIT in a number of countries have shown a lower overall assay sensitivity among PLHIV; 9 our findings are thus contrary to these. While QFT-Plus sensitivity, like QGIT, was lower in PLHIV with CD4 cell counts ,100 cells/ll, the point estimate suggests that QFT-Plus sensitivity may be higher in those with CD4 cell counts ,100 cells/ll than with QGIT. However, the very small number in the strata precludes any firm conclusions being drawn. Larger studies with direct head-to-head comparisons of QFT-Plus and QGIT among severely immunosuppressed patients are needed to investigate this further.
Like QGIT, the QFT-Plus TB1 tube elicits a CD4mediated immune response alone. TB1 sensitivity was not affected by HIV status. However, it is not possible to directly infer QGIT sensitivity based on TB1 sensitivity, given the changes that have been made to the antigens and their formulation. Studies with direct head-to-head comparisons of QFT-Plus and QGIT in high HIV prevalence settings are needed to investigate this further. The higher QFT-Plus sensitivity among PLHIV could have important clinical and policy implications. Given the similar performance but higher costs of IGRAs compared to TST, the World Health Organization recommended the continued use of TST in low-and middle-income countries in 2011. 23 This stance may need to be re-evaluated in the light of our findings, taking into consideration the reproducibility of the results, direct head-to-head comparisons of QGIT and QFT-Plus, including in severely immunosuppressed patients, costs and logistics. In addition, an ideal test for LTBI should be able to predict LTBI patients who will progress to develop active TB. The predictive value of QFT-Plus for active TB therefore needs to be further investigated.
When compared with QGIT in the study by Raby et al., 11 the proportion with indeterminate results on QFT-Plus was lower (Table 4). Indeterminate results may be due to technical errors in sample handling and processing, or immunosuppression of lymphocyte cell lines due to advanced and severe HIV and/or TB. 7,16 Improvements in laboratory procedures over time may therefore partly explain the differences observed. Likewise, earlier diagnosis of HIV and/or TB would result in a shift from indeterminate to positive results. However, TB-HIV-co-infected patients had similar levels of HIV-associated immunosuppression in both studies, suggesting that this is less likely. Finally, targeting both CD8-and CD4-mediated responses by QFT-Plus can increase IFN-c levels sufficiently, even in the presence of a low IFN-c response to mitogen, as seen by the three results indeterminate on TB1 which were positive on TB2.
In contrast, negative results indicate an immune response to the positive control, without detectable Tlymphocyte responses to M. tuberculosis-specific anti-gens. With QFT-Plus, as both CD4-and CD8-mediated IFN-c responses are targeted, a decrease in negative results was anticipated. However, the proportions with negative QFT-Plus were similar to those of QGIT (Table 4). 11 The exact mechanism underlying these negative results are unclear. False-positive microbiological results may play a role; however, studies have shown that the specificity of a positive smear or Xpert for M. tuberculosis is high in high TB prevalence settings. [24][25][26] Representative data on the proportion of smear-or Xpert-positive pulmonary TB patients who are culture-positive for M. tuberculosis in our setting are not available. On univariate analysis, negative/ indeterminate QFT-Plus results were associated with a low BMI, which in the context of TB suggests advanced disease that may be associated with TB-specific immune paresis, 27 resulting in negative responses to M. tuberculosis-specific antigens but positive responses to mitogen. This hypothesis is corroborated by the finding that nearly 70% of negative results at enrolment had converted to positive at 1-2 months following treatment. Studies immunologically characterising negative responses are needed to confirm this. However, the IFN-c concentrations in a number of these samples were close to the manufacturer's threshold, and may represent fluctuations in values around the threshold. Repeat testing of the samples may therefore have yielded positive results.
Our study has several limitations. It was small and undertaken to explore QFT-Plus sensitivity in a high HIV prevalence setting; CIs around point estimates are therefore wide. Culture was not performed on TB patients. A direct head-to-head comparison of QGIT and QFT-Plus was not undertaken. The results cannot be generalised to children, as only adults were included. CONCLUSION QFT-Plus sensitivity was similar to QGIT and TST sensitivity for pulmonary TB, used as a surrogate for LTBI. While overall sensitivity was not affected by HIV status, the sensitivity was lower among PLHIV with severe immunosuppression. Work is needed to understand the assay's predictive value for active TB.
Acknowledgements QIAGEN provided the QuantiFERON w -TB Gold Plus kits free of charge for use in this study. They had no role in the study design, collection, analysis or interpretation of data, manuscript preparation or decision to publish findings. The views expressed are solely those of the study authors. Declaration of interests: All authors declare that QIAGEN provided non-financial support by providing the QuantiFERON w -TB Gold Plus kits free of charge to use in the study. HA, BK, KM, DKM, KC and SF received grants from National Institutes for Health, Bethesda, MD, USA, during the conduct of the study, which partially paid their salaries. SLB received a grant from the Wellcome Trust, London, UK, which paid her salary during the conduct of the study.
This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.