CD4 response of QuantiFERON-TB Gold Plus for positive consistency of latent tuberculosis infection in patients on dialysis

A significantly negative reversion in the QuantiFERON-TB Gold In-tube (QFT-GIT) test is reported in patients on dialysis, which makes the results unreliable. The CD4 and CD8 responses of the QFT-Gold plus (QFT-Plus) may have better positive consistency, but this needs to be investigated. We enrolled dialysis patients with baseline positive QFT-GIT0 results and conducted two rounds of follow-up paired QFT-GIT1&2 and QFT-Plus1&2 tests at an interval of 6 months. The positive consistency, concordance, and discordance of the QFT results were analyzed. A total of 236 patients on dialysis were screened, and 73 participants with positive QFT-GIT0 results were enrolled. The baseline QFT-GIT0 response was higher in the 1st QFT-Plus1(+) group than in the QFT-Plus1(−) group, but insignificantly different between the 1st QFT-GIT1(+) and QFT-GIT1(−) groups. The two assays had good correlation when concurrently tested. Fifty-three subjects completed a second round of the QFT-GIT2 and QFT-Plus2. Persistent positivity was higher with the QFT-Plus2 (81.8%) than with the QFT-GIT2 (58.8%, p = 0.040). The QFT-GIT1 and QFT-Plus1 CD4 responses were higher in patients with persistent positivity than in those with negative reversion, whereas the difference of the QFT-Plus TB1 and TB2 data, representative of the CD8 response, were similar between positive persistence and negative reversion. In conclusion, the QFT-Plus provides more reliable positive consistency than does the QFT-GIT. The CD4 interferon-γ response might play a role in maintaining positivity of LTBI.

Tuberculosis (TB) remains one of the most common infectious diseases in the world. According to the World Health Organization, around 10.0 million TB cases and an estimated 1.2 million TB deaths occurred in HIVnegative people in 2018 1 . In fact, an estimated one quarter of the world's people carry Mycobacterium tuberculosis (M.tb) 2 . When elderly and immunocompromised patients develop active TB, mortality and morbidity are still high 3,4 . In the future of the management of active TB, latent TB infection (LTBI) intervention is one strategy for preventing reactivation and reducing transmission 5,6 .
Among the high-risk populations for TB reactivation, patients on dialysis have a higher incidence of TB (10-25 times higher) than the general population 7,8 . Notably, dialysis patients have high mortality once active TB develops due to the atypical manifestation and delayed diagnosis 9,10 . Therefore, in addition to those with close contact with TB, patients receiving dialysis are also highly prioritized for LTBI intervention in Taiwan 5 .
However, LTBI is diagnosed indirectly by immune assays such as the tuberculin skin test and interferongamma release assays (IGRAs), which, due to immune depression, sometimes yield false negative results in patients receiving dialysis 11,12 . Although our prior study using QuantiFERON-TB Gold In-tube (QFT-GIT) demonstrated an acceptable indeterminate rate of < 5% in dialysis patients 12 , the negative reversion rate at 6-month follow-up was 44-48% 13,14 , which was higher than the 33-35% in contacts or healthcare workers 15,16 . Notably,

Results
Participant enrollment. A total of 236 patients on maintenance dialysis at hospital 1 and 234 patients of hospital 2 were screened for LTBI (Fig. 1). Among them, 108 patients (44 [18.6%] from hospital 1 and 64 [27.4%] from hospital 2) had positive QFT-GIT results. Thirty-five of these patients rejected the invitation to participate in the study. Thereafter, 73 subjects were enrolled in the first round of testing of the QFT-GIT and QFT-Plus together. Around 6 months later, at the follow-up testing of the QFT-GIT and QFT-Plus, twenty participants withdrew from the study, leaving fifty-three participants in the second round.
As shown in Table 2, for the IFN-γ response of the first follow-up test of the QFT-GIT (QFT-GIT 1 ), the TB1 tube in the first test of the QFT-Plus (T1 of QFT-Plus 1 ) and the TB2 tube in the first test of the QFT-Plus 1 (T2 of QFT-Plus 1 ) had good correlation with the IFN-γ response of the QFT-GIT 0 (Spearman correlation r: 0.520, 0.611, and 0.611, respectively [all p < 0.001]). However, the QFT-GIT 0 had no significant correlation with the IFN-γ level of the TB2 tube minus the TB1 tube of the QFT-Plus (T2-T1 of the QFT-Plus 1 ) (Spearman correlation coefficient r: 0.085, p = 0.477).

Discordance and concordance of first follow-up QFT-Plus and QFT-GIT tests.
Regarding the QFT 1 response, the QFT-GIT 1 had good correlation with T1 of the QFT-Plus 1 and T2 of the QFT-Plus 1 (r = 0.879 and 0.895, respectively, both p < 0.001) ( Table 2). But T2-T1 of the QFT-Plus 1 had non-significant correlation with the QFT-GIT 1 (r = 0.150, p = 0.206). For binary results (positive or negative), the concordance rate of the Demographic differences between persistent positivity and negative reversion patients. For the patients with persistent positive results on two subsequent QFT 2 tests, no significant differences from their corresponding negative reversion groups existed in age, dialysis type, diabetes mellitus, BCG vaccination scar or LTBI treatment rate (Table 4). However, more men had persistent positive QFT-Plus results than negative reversion (77.8% vs. 33.3%, p = 0.053). The baseline test of the QFT-GIT 0 (2.47 ± 0.55 vs. 1.20 ± 0.81, p < 0.001) Table 1  www.nature.com/scientificreports/ (Table 4) and that of the QFT-GIT 1 (3.68 ± 2.47 vs. 0.98 ± 0.81, p < 0.001) were higher in patients with persistent positivity than in those with negative reversion of QFT-GIT 2 . The T1 and T2 responses of the QFT-Plus 1 were also higher in those with persistent positivity than in the negative reversion group. But the T2-T1 levels of subsequent QFT-Plus 1 tests were not significantly different between the persistent positivity and negative reversion groups.

Discussions
To examine the positive consistency of the new QFT-Plus test, we screened 236 patients on dialysis. Among them, 73 were enrolled and 53 completed the second round of follow-up paired QFT-GIT and QFT-Plus tests in the present study. The correlation between the two tests when they were administered concurrently was excellent (Pearson's correlation > 0.9). The negative reversion rate of the QFT-GIT was as high as 41.2% at the 6-month Table 3. The intra-test discordance and concordance of serial QuantiFERON-TB Gold Plus (QFT-Plus) and QuantiFERON-TB Gold In-tube (QFT-GIT) tests, respectively. *p = 0.040 by comparing persistent positivity between the 1st and 2nd QFT-GIT(+) versus the 1st and 2nd QFT-Plus(+) using Pearson chi-squared test.  21,22 . Negative reversion of the immune assay for M.tb antigens may indicate regression of the immune response, especially in immunocompromised patients with underlying comorbidities 13,15 . In our previous study, QFT-GIT negative reversion was found to be as high as 45.9% in patients on dialysis 13 , which is similar to that in the present study (41.2% by QFT-GIT). This inconsistent finding is concerning for LTBI screening among high-risk populations who have no clear history of TB exposure that could serve as a timepoint for commencement of LTBI survey. Furthermore, random LTBI testing to screen the dialysis population with QFT-GIT may produce false negative results if the immune response has fallen below the cut-off value of the immune test. Hence, optimal LTBI screening tests with high consistency would provide more reliable information for LTBI management in high-risk groups.
The positive consistency of the QFT-Plus at follow-up was higher (81.8%) than that of the QFT-GIT (58.8%), which might be explained by the greater sensitivity of the QFT-Plus in detecting LTBI consistently. For example, in the first round of the QFT-GIT 1 (−) tests, the QFT-Plus 1 detected specific cases with a high response (T1: 0.98 IU/ml [ Table S3]) beyond the detection grey zone (< 0.8 IU/ml) proposed by Metcalfe et al 23 . In addition, the difference in QFT-GIT 0 between the QFT-Plus 1 (+) and (−) results was larger than that between the QFT-GIT 0 (+) and (−) results, indicating that the QFT-Plus might have a better discriminating ability. In addition, CD8 antigens were added to the QFT-Plus to augment CD8 effector or memory responses in individuals with defective or low CD4 responses. However, the CD8 response was not significantly different between the persistent positivity and negative reversion groups. By contrast, the CD4 (TB1 tube) or CD4 plus CD8 (TB2 tube) responses were both well correlated with persistent positivity (Table 4). Therefore, the present study indicates that remote or chronic M.tb infection might be more highly correlated with the CD4 response than with the CD8 response 24 . In this cohort of patients on dialysis, the CD4 antigens of the QFT-Plus, including ESAT 6 and CFP10 peptides but no TB7.7, as compared with the QFT-GIT, might be the dominant and consistent inducers of the memory CD4 T cell response. Our study demonstrates a more reliable response of M.tb-specific CD4 T cell response over time for the QFT-Plus than for the QFT-GIT. Even though both tests measure the CD4 response, it was reported that the QFT-Plus might be more useful in detecting LTBI in elderly 18 and immunocompromised patients 25 . However, determining the exact mechanism of why the QFT-Plus can maintain the positive consistency well will require further study and large-scale validation.
It should be noted that the response decreased from the first to the second round, regardless of which QFT assay was used, indicating a decaying immune response to M.tb antigens. This finding is consistent with previous   www.nature.com/scientificreports/ reports that high initial values might predict positive consistency 13 , but so far, the results have been too inconclusive to titrate a cut-off value. Because follow-up is time-consuming, our previous report also showed that additional inflammatory markers at the baseline could be used to discriminate persistent positivity from negative reversion, at a possible cost of more than one test 26 . The present study indicates that an initial positive QFT-Plus result has a likelihood of being consistent at 6-month follow-up of higher than an 80%, so this test could be considered as a one-time test.
Regarding the results of the QFT-GIT 0 , QFT-GIT 1 and QFT-Plus 1 , respectively, the positive consistency rates were both around two-thirds, with no inter-assay difference (67% vs. 64%). Because of a lack of baseline QFT-Plus data, the QFT-GIT 0 might be specific only for QFT-GIT 1 follow-up. This might indicate that initial screening is a critical point in the positive judgement and has the same importance as its detection consistency. Another issue is that we recruited several patients with old TB cases, who might not have been candidates for LTBI screening. We only aimed to test the positive consistency by IGRA assay. When the results were re-analyzed for the patients without a history of old TB, the positive consistency of the QFT-Plus was still higher than that of the QFT-GIT.
This study had several limitations. First, the sample size in the present study was not large. In addition, the study design was not a randomized comparison. However, the two assays were compared using the same blood samples so as to avoid inter-subject variation. Third, the subjects were recruited from a set of patients with positive IGRA tests, so the results could not be used to interpret positive conversion from negative results. Fourth, we did not enroll patients with tuberculosis or household contacts. Further application of the CD8 response in the QFT-Plus might have a potential for disease status discrimination 27 . No participants had a history of recent TB contact, which could have affected the performance of the QFT-Plus 28 , especially the CD8 response 29,30 .
In conclusion, the present study demonstrates that the QFT-Plus had a higher positive result consistency (81.8%) at a 6-month follow-up as compared with the QFT-GIT. The inter-assay correlation was remarkably high. Another finding was that the CD4 response of the QFT-Plus might be important for positive consistency of LTBI. Although larger studies are needed to further validate the findings, we conclude that the new QFT-Plus has more reliable positive results than those of the QFT-GIT and is therefore more suitable for TB screening of dialysis patients.

Methods
Participants. Subjects were recruited from two tertiary-care medical centers in northern Taiwan in 2016-2019 under IRB approval (Nos. 201603066DIPB and 108067-F). Written informed consent was obtained. Patients on maintenance dialysis underwent LTBI screening according to the routine health policy of one hospital and by invitation in a research survey at the other. The LTBI screening included chest plain film and the QFT-GIT. If radiography indicated possible TB, three sets of sputum cultures for M.tb were obtained. Patients on dialysis with positive QFT-GIT results but no evidence of active TB were included. Patients with diagnoses of human immunodeficiency virus infection or active TB were excluded. Identifiers of participants were removed in accordance with the Health Insurance Portability and Accountability Act (HIPAA) Privacy Rule. We confirmed that all experiments were performed in accordance with relevant guidelines and regulations.
Study design. After a participant had been enrolled, the QFT-GIT and QFT-Plus (QIAGEN, Germany) were performed concurrently on the same blood sample according to the manufacturer's instructions 17,31 . The interferon-gamma (IFN-γ) level of the post-reaction supernatant was then measured with an enzyme-linked immuno-sorbent assay. The IFN-γ response of the QFT kits was the IFN-γlevel in the TB antigen tube minus that in the negative-control tube in the QFT-GIT and the TB1 or TB2 tube minus the negative-control tube in the QFT-Plus. Results were interpreted as positive, negative, or indeterminate 32,33 . Around 6 months later, the QFT-GIT and QFT-Plus were followed up together on the same participants and on the same samples of whole blood. Regarding the primary end-point, we compared the proportions of persistent positive results (consistency) of the different QFT kits. Data collection. We recorded participant's demographics, including age, gender, and comorbidities, as well as dialysis type (hemodialysis or peritoneal dialysis). Standardized case report forms were used and completed with default options. The Bacillus Calmette Guérin (BCG) vaccination scar was checked. The IFN-γ levels of the QFT-GIT and QFT-Plus were recorded.

Statistical analyses.
Pearson's chi-squared test or Fisher's exact test was used to compare categorical variables, and continuous variables were compared using the Student's t test or the Mann-Whitney U test, where appropriate. Pearson's correlation analysis was used to assess relationships between the first and second pairs of the QFT-GIT and QFT-Plus assays. All analyses were performed in SPSS software (version 19.0 for Windows, SPSS Inc.).

Consent for publication
Written informed consent was obtained from the participants at the time of enrollment.