SEARCH

SEARCH BY CITATION

Keywords:

  • Active disease;
  • interferon gamma release assay;
  • Quantiferon Gold;
  • low incident setting;
  • tuberculosis

Abstract

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References

The role of interferon-γ release assays in the diagnosis of active tuberculosis disease is uncertain, and recent guidelines do not support their routine use. We reviewed the clinical records of 415 patients who had a QuantiFERON-TB Gold In-Tube assay between 29 June 2005 and 28 October 2010 to determine its performance in the diagnosis of active tuberculosis disease in a low prevalence setting, specifically in human immunodeficiency virus (HIV) -positive and HIV-negative patients, those of UK and non-UK origin, and those with pulmonary and extrapulmonary disease. For the diagnosis of active tuberculosis disease the overall sensitivity of QuantiFERON-TB Gold In-Tube assay was 71.4% (95% CI 59.3–81.1), specificity was 81.0% (95% CI 75.5–85.6) and negative predictive value was 92.6% (95% CI 88.2–95.5). No significant difference in sensitivity was seen in culture-positive and culture-negative tuberculosis, in pulmonary and extrapulmonary disease, or with HIV infection. Specificity and negative predictive value were significantly higher in patients of UK origin compared with those of non-UK origin (89.3% (95% CI 83.3–93.3) and 97.1% (95% CI 92.7–98.9) versus 66.3% (95% CI 55.6–75.5) and 83.3% (95% CI 72.6–90.4)). Our study suggests that there may be a role for interferon-γ release assays in excluding active tuberculosis disease, particularly extrapulmonary disease, in patients originating from areas of low tuberculosis incidence, with a negative test highly predictive of a lack of active tuberculosis disease in this group. We cannot support the use of these assays in the diagnosis of active tuberculosis infection in patients from areas of higher incidence.


Background

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References

Interferon-γ release assays (IGRA) are increasingly used in the diagnosis of latent tuberculosis (TB) infection [1, 2]. Their role in the diagnosis of active TB disease is less clear, with recent UK guidelines not supporting their use as a routine diagnostic tool [3, 4]. Nevertheless, despite apparent uncertainty in the interpretation of their results, their use continues in the diagnosis of active TB disease in clinical practice.

Recent meta-analyses of IGRA in the diagnosis of active TB disease in a variety of geographical and clinical settings report sensitivity of 80–85% and specificity of 79–84% [5, 6]. Higher sensitivity is reported in human immunodeficiency virus (HIV) -negative versus HIV-positive patients [7]. Negative predictive values (NPV) has been reported from 67.6 to 87% [8, 9].

Newcastle upon Tyne is in north-east England, a region with a low reported TB incidence of 5.4 cases per 100 000 population in 2010 (http://www.hpa.org.uk/web/HPAweb&HPAwebStandard/HPAweb_C/1287147464622). We undertook the following study to assess the use of the QuantiFERON-TB Gold In-Tube assay (QFT-GIT, Cellestis, Chadstone, VA, Australia) within our department, and to determine its sensitivity, specificity and NPV in the diagnosis of active TB disease in a low prevalence setting, specifically comparing HIV-positive and HIV-negative patients, those of UK and non-UK origin, and those with pulmonary (PTB) and extrapulmonary (EPTB) disease.

Methods

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References

Clinical records of patients who had QFT-GIT requested through our unit between 29 June 2005 and 28 October 2010 were reviewed. Data collected included the indication for QFT-GIT assay, final diagnosis, site of TB infection, TB culture and histology reports, sex, age, nationality, HIV status, and CD4 count if HIV positive.

Cases were assigned into five predefined diagnostic categories, similar to those used in previous studies [9]:

  1. Culture-confirmed TB disease with supporting clinical and radiological findings;
  2. Culture-negative highly probable cases with clinical and radiological features highly suggestive of TB disease, a decision made to treat by a clinician with an appropriate response to therapy, and supportive histology if available, including cases where culture was not done;
  3. Clinically indeterminate cases where not possible to confirm or exclude a diagnosis of TB disease;
  4. Symptomatic patients where active TB disease excluded, all microbiological samples smear and culture negative for TB, and an alternative diagnosis identified;
  5. Asymptomatic latent infection with QFT-GIT requested for screening purposes only.

Duplicate tests, patients on whom insufficient clinical data were available, clinically indeterminate (category 3), and asymptomatic cases (category 5) were excluded. Analysis was undertaken comparing patients with active TB disease (categories 1 and 2) with symptomatic patients with an alternative diagnosis (category 4) acting as controls. Sensitivity, specificity and NPV were calculated for QFT-GIT overall, in HIV-positive versus HIV-negative patients, those of UK versus non-UK origin, and those with PTB versus EPTB. The 95% CI were calculated using the Wilson score method. Indeterminate results were treated as false-negative when calculating sensitivity and false-positive when calculating specificity and NPV.

QFT-GIT and TB culture were performed at the Newcastle-upon-Tyne Health Protection Agency Regional Centre for Mycobacteriology; QFT-GIT in accordance with manufacturer's guidelines [10], and TB culture using an automated liquid mycobacterial culture system (BacT/ALERT 3D, bioMérieux, Marcy l'Etoile, France) supported by Löwenstein–Jensen solid medium.

Acid-fast microscopy was performed on all samples submitted for mycobacterial culture with PCR performed selectively according to national guidelines [1]. As microscopy is not specific for Mycobacterium tuberculosis and the number of samples on which PCR was performed is low, comparative data against these parameters would not be statistically valid. Hence, IGRA results are compared with mycobacterial culture, the reference standard for the laboratory diagnosis of TB, and to clinically diagnosed highly probable cases of TB disease. Tuberculin skin tests were not routinely performed and therefore not included in the analysis.

Results

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References

From 415 QFT-GIT requested, 120 were excluded from further analysis (28 insufficient data, 19 duplicate tests, 20 clinically indeterminate cases, 53 asymptomatic screening cases). From the remaining 295 symptomatic patients 42 had culture-positive TB, 21 had culture-negative TB and 232 had an alternative diagnosis.

Overall, 53% were male patients; median age was 40 years; 55% were of UK origin, 20% were from Africa, 19% were from the Indian subcontinent, 3% were from the Middle East, and 2% and 1% were from other Asian and other European countries, respectively; 17% were HIV-positive with a median CD4 count of 329 cells/μL (range 28–1060) (Table 1). Forty-six (73%) of those with TB had EPTB, 9 (14%) had PTB, 7 (11%) had mixed PTB and EPTB and 1 (2%) had miliary TB.

Table 1. Demographic and clinical characteristics
 AllActive TBAlternative diagnosis
  1. HIV, human immunodeficiency virus; TB, tuberculosis.

n (%)29563 (21)232 (79)
Male sex (%)156 (53)39 (62)117 (50)
Age, median years (range)40 (16–90)33 (17–90)42 (16–82)
Origin (%)
UK163 (55)14 (22)149 (64)
Africa59 (20)19 (30)40 (17)
Indian subcontinent55 (19)23 (37)32 (14)
Middle East8 (3)4 (6)4 (2)
Other Asia7 (2)3 (5)4 (2)
Other European3 (1)3 (1)

HIV +ve (%)

Median CD4 (range)

51 (17)

329 (28–1060)

14 (22)

261 (93–891)

37 (16)

350 (28–1060)

For the diagnosis of active TB disease the overall sensitivity of QFT-GIT was 71.4% (95% CI 59.3–81.1), specificity was 81.0% (95% CI 75.5–85.6) and NPV was 92.6% (95% CI 88.2–95.5). No significant difference in sensitivity was seen in culture-positive TB (76.2% (95% CI 61.5–86.5)) compared with culture-negative TB (61.9% (95% CI 40.9–79.3)) or in PTB (75.0% (95% CI 50.5–89.8)) compared with EPTB (71.7% (95% CI 58.4–82.0)). Specificity and NPV were significantly higher in patients of UK origin compared with those of non-UK origin (89.3% (95% CI 83.3–93.3) and 97.1% (95% CI 92.7–98.9) versus 66.3% (95% CI 55.6–75.5) and 83.3% (95% CI 72.6–90.4)) (Table 2).

Table 2. Overall analysis
Patient group QFT-GIT resultTBNon-TBSensitivitySpecificity (95% CI)NPV (95% CI)
  1. EPTB, extrapulmonary tuberculosis; NPV, negative predictive value; PTB, pulmonary tuberculosis; QFT-GIT, QuantiFERON-TB Gold In-Tube assay.

  2. a

    Includes seven with PTB and EPTB.

  3. b

    Includes eight atypical mycobacterial infections.

All patients63232b71.4% (59.3–81.1)81.0% (75.5–85.6)92.6% (88.2–95.5)
Positive4530
Negative15188
Indeterminate314
Culture positive4276.2% (61.5–86.5)
Positive32
Negative8
Indeterminate2
Culture negative2161.9% (40.9–79.3)
Positive13
Negative7
Indeterminate1
PTB16a75.0% (50.5–89.8)
Positive12
Negative2
Indeterminate2
EPTB53a71.7% (58.4–82.0)
Positive38
Negative13
Indeterminate2
UK-born1414957.1% (32.6–78.6)89.3% (83.3–93.3)97.1% (92.7–98.9)
Positive86
Negative4133
Indeterminate210
Non-UK born498375.5% (61.9–85.4)66.3% (55.6–75.5)83.3% (72.6–90.4)
Positive3724
Negative1155
Indeterminate14

In all HIV-infected patients the sensitivity of QFT-GIT was 71.4% (95% CI 45.4–88.3), specificity was 81.1% (95% CI 65.8–90.5) and NPV was 90.9% (95% CI 76.4–96.9) for the diagnosis of active TB disease. There was no significant difference in the performance of the assay in patients with CD4 counts of <200 cells/μL (Table 3).

Table 3. HIV-positive patients
Patient population QFT-GIT resultTBNon-TBSensitivity (95% CI)Specificity (95% CI)NPV (95% CI)
  1. HIV, human immunodeficiency virus; NPV, negative predictive value; QFT-GIT, QuantiFERON-TB Gold In-Tube assay; TB, tuberculosis.

All HIV143771.4% (45.4–88.3)81.1% (65.8–90.5)90.9% (76.4–96.9)
Positive105
Negative330
Indeterminate12
CD4 < 20061083.3% (43.7–97.0)70.0% (39.7–89.2)87.5% (52.9–97.8)
Positive52
Negative17
Indeterminate01

Discussion

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References

This study confirms the relatively low overall sensitivity and specificity of QFT-GIT for the diagnosis of active TB disease seen in previous studies [5, 6] and reinforces recommendations that IGRA should not be used routinely in the diagnosis of active TB disease [2-4]. There was no significant difference in the performance of the assay when comparing culture-positive and culture-negative and pulmonary and extrapulmonary disease, nor when comparing HIV-positive and HIV-negative individuals.

Of note, specificity and NPV were significantly higher in patients of UK origin compared with those of non-UK origin (89.3% and 97.1% versus 66.3% and 83.3%). Although a greater proportion of patients of UK origin had PTB compared with those of non-UK origin (55 versus 16%), overall TB culture positivity was greater in patients of non-UK origin (75 versus 64%). IGRA has been shown to have improved specificity in patients with EPTB compared with PTB with similar sensitivity in each group [11]. It therefore seems likely that the difference in specificity in the two groups in our study reflects a higher prevalence of latent TB infection in patients of non-UK origin rather than differences in disease site. The high NPV in UK-born patients reflects not only the low incidence in this group, but also a significant number of IGRA requested in patients with non-specific symptoms unlikely to have TB infection.

Previous studies of IGRA in areas of low TB incidence have focused largely on the diagnosis of latent TB infection and there are few studies looking at the diagnosis of active disease in this setting. Winqvist et al. [8] compared the performance of QuantiFERON-TB Gold in immigrants from countries of high and medium TB incidence with native Swedes and found the performance to be similar; 62.9% in this cohort had PTB and IGRA performance was noted again to be better in EPTB. Detjen et al. [12] reported excellent performance of QFT-GIT in the diagnosis of active TB disease in children in Germany but did not compare those of German and non-German origin. Vesenbeckh et al. [13] demonstrated a sensitivity of QFT-GIT for the diagnosis of active TB disease in Germany of 82%; 53/61 had PTB and there were again no data on patient country of origin.

This is contrasted by Metcalfe et al. [14] who reported sensitivity of QFT-GIT of 69% (95% CI 52–83) and specificity of 52% (95% CI 41–63) for the diagnosis of active PTB disease in low and middle income countries with high TB incidence. In South Africa, Ling et al. [15] demonstrated a sensitivity of 76% (95% CI 68–83) and specificity 42% (95% CI 36–49) for QFT-GIT for the diagnosis of active TB disease, independent of HIV status and CD4 count.

Our study suggests that there may be a role for IGRA in excluding active TB disease in patients originating from areas of low TB incidence, with a negative test being highly predictive of a lack of active TB disease in this group. As a higher proportion of these patients had PTB, when microbiological diagnosis should be more straightforward, it seems sensible to restrict this role to patients from areas of low TB incidence where EPTB is suspected. This is supported by previous studies demonstrating improved performance of IGRA in EPTB and in areas of low TB incidence. Based on this study we cannot support the use of IGRA in the diagnosis of active TB infection from areas of higher incidence.

The study is limited by being retrospective, with relatively small numbers of patients allowing only limited subgroup analysis, and with one observer there is the potential for bias. As a consequence of local arrangements, the respiratory department in our centre sees a larger proportion of PTB presenting, hence the high proportion of patients with EPTB seen in this cohort.

In conclusion, this study demonstrates significantly higher specificity and NPV of QFT-GIT for the diagnosis of active TB disease in patients from areas of low TB incidence compared with those from areas of higher TB incidence. Although there may be a limited role for IGRA in excluding active TB disease, particularly EPTB, in this group, attempts at clinical, radiological and microbiological diagnosis remain paramount, even when a rapidly available blood test appears tempting.

Transparency Declaration

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References

These data have previously been presented as an oral presentation at the 17th Annual Conference of the British HIV Association (BHIVA) 2011, and as a poster at the Federation of Infection Societies Scientific Conference 2011.

References

  1. Top of page
  2. Abstract
  3. Background
  4. Methods
  5. Results
  6. Discussion
  7. Transparency Declaration
  8. References
  • 1
    National Collaborating Centre for Chronic Conditions (UK), Centre for Clinical Practice at NICE (UK). Tuberculosis: clinical diagnosis and management of tuberculosis, and measures for its prevention and control. London: National Institute for Health and Clinical Excellence (UK); 2011.
  • 2
    Brouqui P, Delaigue S, Parola P, Drancourt M. Interferon-γ release assay: use and misuse. Clin Microbiol Infect 2012; 18: 10531054. doi: 10.1111/1469-0691.12025.
  • 3
    Health Protection Agency Position Statement on the use of Interferon Gamma Release Assay (IGRA) tests for tuberculosis (TB) (Draft Interim HPA Guidance), Oct 2007.
  • 4
    Pozniak AL, Coyne KM, Miller RF et al. British HIV Association guidelines for the treatment of TB/HIV coinfection 2011. HIV Med 2011; 12: 517524.
  • 5
    Sester M, Sotgiu G, Lange C et al. Interferon-γ release assays for the diagnosis of active tuberculosis: a systematic review and meta-analysis. Eur Respir J 2011; 37: 100111.
  • 6
    Dai Y, Feng Y, Xu R, Xu W, Lu W, Wang J. Evaluation of interferon-γ release assays for the diagnosis of tuberculosis: an updated meta-analysis. Eur J Clin Microbiol Infect Dis 2012; 31: 31273137.
  • 7
    Chen J, Zhang R, Wang J et al. Interferon-γ release assays for the diagnosis of active tuberculosis in HIV-infected patients: a systematic review and meta-analysis. PLoS ONE 2011; 6: e26827.
  • 8
    Winqvist N, Björkman P, Norén A, Miörner H. Use of a T cell interferon γ release assay in the investigation for suspected active tuberculosis in a low prevalence area. BMC Infect Dis 2009; 9: 105.
  • 9
    Dosanjh DP, Hinks TS, Innes JA et al. Improved diagnostic evaluation of suspected tuberculosis in routine practice. Ann Intern Med 2008; 148: 325336.
  • 10
    QuantiFERON-TB Gold (In-Tube Method) Package Insert. Australia: Cellestis.
  • 11
    Feng Y, Diao N, Shao L et al. Interferon-γ release assay performance in pulmonary and extrapulmonary tuberculosis. PLoS ONE 2012; 7: e32652.
  • 12
    Detjen AK, Keil T, Roll S et al. Interferon-γ release assays improve the diagnosis of tuberculosis and nontuberculous mycobacterial disease in children in a country with a low incidence of tuberculosis. Clin Infect Dis 2007; 45: 322328.
  • 13
    Vesenbeckh SM, Schönfeld N, Mauch H et al. The use of interferon γ release assays in the diagnosis of active tuberculosis. Tuberc Res Treat 2012; 2012: 768723.
  • 14
    Metcalfe JZ, Everett CK, Steingart KR et al. Interferon-γ release assays for active pulmonary tuberculosis diagnosis in adults in low- and middle-income countries: systematic review and meta-analysis. J Infect Dis 2011; 4: S1120S1129.
  • 15
    Ling DI, Pai M, Davids V et al. Are interferon-γ release assays useful for diagnosing active tuberculosis in a high-burden setting? Eur Respir J 2011; 38: 649656.