Skip to main content
. 2021 Dec 1;55:89. doi: 10.11606/s1518-8787.2021055003125

Box 3. Description of the articles included in the integrative review on the effectiveness of the GeneXpert® MTB/RIF test in the diagnosis of pulmonary TB in people living with HIV/AIDS.

Authors Periodic Year/country Study design Objectives Total (n) Measuring instrument Achievements Highly rated items/ Number of Items (Limitations)
SCOTT, L. et al.6/ PLOS Medicine/ 2011/ South Africa Prospective study. Evaluate the performance of the Xpert MTB/RIF test, LightCycler Mycobacterium Detection (LCTB) and MTBDRplus (Hain Lifescience) test with smear microscopy and liquid culture in a single sputum sample. 311 individuals from a population with Prevalence of HIV. GeneXpert MTB/RIF, Ziehl-Neelsen smear microscopy, MGIT culture, MDRTBplus and LCTB.

  • GeneXpert – sensitivity among people with HIV 84% (95%CI 69–93), specificity 96% (95%CI 89–99), PPV 92% (95%CI 79–98), NPV 92% (95%CI 84–97).

  • Smear microscopy – sensitivity 54% (95%CI 38–69), specificity 100% (95%CI 95–100), PPV 100% (95%CI 85–100), NPV 80% (95%CI 70-87).

  • MDRTBplus – 70% sensitivity (95%CI 54–83), specificity 96% (95%CI 89–99), PPV 91% (95%CI 76–98), NPV 85% (95%CI 76–92).

  • LCTB – 70% sensitivity (95%CI 54–83), specificity 98% (95%CI 93–100), PPV 97% (95%CI 83–100), NPV 86% (95%CI 77–92).

  • 6/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing Xpert results;

  • - the study had 22.6% sample loss.
Dagnra et al.22/ New Microbes and New Infections/ 2015/Togo Did not quoted. Determine the prevalence of MDR-TB among previously treated HIV positive and negative patients and evaluate the performance of GeneXpert MTB/RIF in detecting MDR cases. 42 sputum samples. GeneXpert MTB/RIF, BACTEC MGIT 960 culture, and Ziehl-Neelsen staining microscopy. Only 31 (73.8%) of the 42 evaluated cases consented to be tested for HIV. 07 (22.6%) were HIV positive. All cases of coinfection were diagnosed by GeneXpert and two tested positive for resistance to rifampicin. There was excellent agreement between the sensitivity test and GeneXpert in the two people with MDR-TB.

  • 6/8 (2 N/A)

  • - not elucidated sampling process and reduced sample size, mainly regarding the number of co-infected;

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.
Rachow et al.23/ PLOS ONE/ 2011/ Tanzania Clinical validation study. Compare GeneXpert results to microscopy and culture in sputum samples from people suspected to have TB and to be infected or not with HIV. 292 patients, and 876 sputum samples. GeneXpert MTB/RIF, Ziehl-Neelsen staining microscopy and culture in LJ and BACTEC MGIT 960 methods.

  • Among 77 individuals undergoing empirical treatment for TB (smear microscopy and negative culture), 07 were positive for the GeneXpert test. They were living with HIV.

  • Sensitivity of the GeneXpert test in individuals with HIV: 82.0% (95%CI 68.6–91.4), smear microscopy: 52.0% (95%CI 37.4–66.3), culture (LJ): 68.0% (95%CI 53.3–80.5), and (MGIT): 76.0% (95%CI 61.8–86.9).

  • Sensitivity of both smear microscopy and culture (LJ): 78.0% (95%CI 64.0–88.5).

  • Specificity for the GeneXpert test was 98.0% (95%CI 89.4–100). Regarding the other tests, the specificity was 100% (95%CI 92.9–100).

 8/8 (2 N/A).
Sumangala et al.24/ Journal of International Medicine & Dentistry/ 2017/ India Cross-sectional study. Assessing the Effectiveness of the GeneXpert MTB/RIF Assay in the Early Diagnosis of Tuberculosis in People Living with HIV. 433 individuals and sputum samples. GeneXpert MTB/RIF and Fluorescent Microscopy. Among 433 sputum samples, 36 (8.3%) were positive for GeneXpert and 28 (6.4%) samples were positive for fluorescent microscopy. Five cases of resistance to rifampicin were detected by GeneXpert.

  • 5/8 (2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results;

  • - used sputum smear microscopy as the reference diagnostic method, having limitations to correctly classify a case of the disease.
Lawn1et al.25/ PLOS Medicine/ 2011/ África do Sul Prospective study. Evaluate the diagnostic accuracy of the GeneXpert MTB/RIF test in the diagnosis of TB in patients with advanced immunodeficiency and who have not started ART. 445 individuals and 839 sputum samples. GeneXpert MTB/RIF, microscopy, liquid culture by BACTEC MGIT 960.

  • Among 445 subjects participating in the study, 81 had a positive result for the culture.

  • GeneXpert – Sensitivity 73.3% (95%CI 61.9–82.9); specificity 99.2% (95%CI 97.7–99.8); PPV 94.8% (95%CI 85.6-98.9); NPV 94.8% (95%CI 92.1–96.8). GeneXpert had a low false-positive rate and detected all positive TB cases for both smear microscopy and culture tests.

  • Sputum smear microscopy – Sensitivity 28.0% (95%CI 18.2–39.6); specificity 100.0% (95%CI 98.9–100.0); 100% PPV (95%CI 83.9–100); NPV 87.3% (95%CI 83.7–90.3). GeneXpert detected 43.4% of smear-negative and culture-positive cases in a single sputum sample and 62.3% when two sputum samples were analyzed.

  • The test correctly identified rifampicin resistance in four patients who had MDR-TB, rendering three false positive drug resistance results.

 8/8 (2N/A).
Balcha et al.26/ PLOS ONE/ 2014/ Ethiopia Did not quoted. Evaluate the prevalence of TB, the diagnostic yield of GeneXpert MTB/RIF compared to sputum smear microscopy and culture, and the impact of GeneXpert results on the clinical treatment of HIV-positive adults eligible for ART. 812 individuals and 1,514 sputum samples. GeneXpert MTB/RIF, Ziehl-Neelsen stained microscopy of smear and liquid culture.

  • Among 137 bacteriologically confirmed TB patients, 123 (89.8%) were culture positive and 13 (10.2%) were negative. The GeneXpert test was positive for 94 individuals, which also detected 2 cases of resistance to rifampicin. The sputum smear microscopy was positive for 31 patients. Regarding smear microscopy, GeneXpert increased the detection of TB cases by 47.4% (64 cases). Moreover, GeneXpert detected 14 (10.2%) cases of culture-negative or contaminated TB.

  • Among 94 patients with sputum samples positive for GeneXpert, 30 (32%) were positive for sputum smear microscopy. Of the 122 culture-positive cases, 28 (23%) were smear-positive. Of 104 smear-negative cases, 54 (52%) were positive using both GeneXpert and culture, and 40 (38%) were positive using culture alone. All patients with positive smear microscopy were positive for GeneXpert or culture.

  • Moreover, 81 [66.4% (95%CI 57.2–74.6)] culture-positive TB cases were positive for GeneXpert.

  • GeneXpert – sensitivity 96.4% (95%CI 79.7–99.9) and 57.4% (95%CI 46.8–67.5) between positive and negative sputum smear microscopy cases, respectively; specificity 98.4% (95%CI 97.0–99.2); PPV 88.0% (95%CI 79.2–93.6) and NPV 94.2% (95%CI 92.1–95.8).

  • Smear microscopy – 23% sensitivity (95%CI 16.0–31.7).

 8/8 (2N/A).
Griesel et al.27/ Clinical Infectious Diseases/ 2018/ South Africa Prospective cohort study. Develop a clinical prediction rule for the diagnosis of tuberculosis and determine the diagnostic utility of the GeneXpert MTB/RIF assay in critically ill HIV-infected patients. 484 individuals. GeneXpert MTB/RIF, BACTEC MGIT 960 Culture and Fluorescence Smear Microscopy.

  • Considering the sputum culture, 53% (255/484) had a positive result. Among the participants, 56 (11.6%) underwent empirical treatment with negative sputum. Smear microscopy and GeneXpert, whence 22 (43.1%) had a positive result for the culture.

  • GeneXpert – sensitivity 86.3% (95%CI 81.5–90.3); specificity 96.1% (95%CI 92.6–98.2); PPV 96.1% (95%CI 92.7-98.2); NPV 86.2% (95%CI 81.4–90.2); PLR 21.9 (95%CI 11.5–41.6); NLR 0.14 (95%CI 0.11–0.19).

  • Sputum smear microscopy – sensitivity 57.0% (95%CI 50.7–63.2); specificity 98.7% (95%CI 96.2–97.7); PPV 98.0% (95%CI 94.2–99.6); NPV 67.2% (95%CI 61.9–72.2); PLR 43.3 (95%CI 14.0-134.0); NLR 0.44 (95%CI 0.38–0.50).

  • 7/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.
Akanbi et al.28/ BMC Pulmonary Medicine/ 2017/Nigeria Prospective cohort study. Evaluate the impact of GeneXpert MTB/RIF on diagnosis, treatment time, and treatment outcome among HIV-associated TB patients in Nigeria. 310 individuals and 620 sputum samples. GeneXpert MTB/RIF and smear microscopy. TB was diagnosed in 76 (24.5%) patients, of which 63 (82.9%) had bacteriological confirmation, 56 (73.7%) had a positive result for GeneXpert and 57 (75%) for sputum smear microscopy. In the concordance analysis, 50 patients (65.8%) were positive for both tests, 6 (7.9%) were positive for GeneXpert and negative for sputum smear microscopy, 7 (9.2) negative for GeneXpert and positive for sputum smear microscopy and 13 (17.1%) were negative for both exams.

  • 5/8(N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results;

  • - used sputum smear microscopy as a diagnostic method of comparison, which may not correctly classify a case of the disease.
O’Grady et al.29/ Clinical Infectious Diseases/ 2012/ Zambia Descriptive, prospective study. Evaluate the performance of the GeneXpert MTB/RIF Assay for the detection of pulmonary tuberculosis and MDR-TB in HIV-infected and non-HIV-infected adult patients. 643 individuals. GeneXpert MTB/RIF, smear microscopy, BACTEC MGIT 960 culture, and MGIT drug sensitivity test.

  • GeneXpert in HIV-positive patients – sensitivity 88.2% (95%CI 81.9–92.6); specificity 95.1% (95%CI 1.5–97.3); PPV 92.2% (95%CI 86.5-95.7); NPV 92.5% (95%CI 88.4-95.3); sensitivity when sputum smear microscopy was positive: 96.5% (95%CI 89.3–99.1); sensitivity when sputum smear microscopy was negative: 78.9% (95%CI 67.8–87.1).

  • GeneXpert in HIV negatives – sensitivity 74.3% (95%CI 56.4–87.0); specificity 96.3% (95%CI 91.7–98.5); PPV 81.3% (95%CI 63.0-92.1); NPV 94.5% (95%CI 89.5–97.3).

  • 6/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results;

  • - patients with contaminated and insufficient samples and patients with negative cultures in treatment were excluded (which represents a loss of 33.4% of eligible cases).
Theron et al.30/ American Journal of Respiratory and Critical Care Medicine/ 2011/ South Africa Did not quoted. Validate GeneXpert MTB/RIF test performance using a single cartridge and assess the impact of certain factors on test performance. 480 sputum samples taken from a high HIV prevalence population. GeneXpert MTB/RIF, fluorescence microscopy and BACTEC MGIT 960 culture.

  • GeneXpert in co-infected patients – 69.6% sensitivity (95%CI 55.2–80.1); sensitivity in people with negative sputum smear microscopy 47.3% (95%CI 29.2–67); specificity 91.7% (95%CI 83.8-95.9); NPV 84.6% (95%CI 75.9–90.6); PPV 82.1% (95%CI 67.3–91).

  • GeneXpert in non-coinfected patients – sensitivity 82.9% (95%CI 73.4–89.6); specificity 95.6 (95%CI 91.8–97.7); NPV 93.3% (95%CI 89.1–96.0); PPV 88.3 (95%CI 79.3–93.4).

  • Sputum smear microscopy in co-infected patients – sensitivity 50% (95%CI 36.1–63.9); specificity 98.9% (95%CI 94.6–99.8).

  • In patients with HIV and a CD4 count < 200 cells/ml, the combination of sputum smear microscopy and GeneXpert had a sensitivity of 69.6% (95%CI 49.3–84.4) and isolated microscopy of 39.1% (CI95 % 22.2–59.2).

  • 7/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.
Mbu et al.31/ PLOS ONE/ 2018/ Cameroon Did not quoted. Assess the burden of TB among people newly diagnosed with HIV and compare the performance of the WHO-recommended clinical screening algorithm and the methods of diagnosing TB in this population. 940 individuals. GeneXpert MTB/RIF, fluorescence microscopy and culture in LJ medium and BACTEC MGIT 960.

  • In the 131 samples with positive results for culture, the GeneXpert test was performed retrospectively in 108 (82%) of them.

  • GeneXpert – 68% sensitivity (95%CI 58–76); sensitivity in people with positive sputum smear microscopy 97%; sensitivity in people with negative sputum smear microscopy 51%.

  • Smear microscopy – sensitivity of direct smear microscopy 24% (95%CI 18–32); sensitivity of concentrated sputum smear microscopy 35% (95%CI 28–44).

  • Culture: sensitivity – solid 81% (95%CI 73–87); net 98% (95%CI 94–99).

  • 6/8(2N/A)

  • - the results of the reference test (sputum culture) were obtained before the knowledge of the Xpert results;

  • - the GeneXpert test was only performed on culture-positive samples, resulting in a significant sample loss.
Ssengooba et al.32/ PLOS ONE/ 2014/ Uganda Prospective study.

  • Compare the diagnostic gain of several complementary

  • GeneXpert strategies with that of an Xpert replacement strategy for the diagnosis of pulmonary TB among HIV-infected adults.

 424 individuals. GeneXpert MTB/RIF, smear microscopy and culture in solid and liquid media (MGIT and LJ).

  • 123 (29.0%) participants tested positive for MGIT culture.

  • Rifampicin resistance – 04 positive for GeneXpert and MGIT.

  • GeneXpert – sensitivity 76.4% (95%CI 67.9–83.6); sensitivity among participants with CD4 > 200 cells/mm3 91,7% (95%CI 73.0–98.9%); sensitivity among participants with CD4 ≤ 200 cells/mm3 73,2% (CI 63.2–81.7%). There was almost no difference in test sensitivity between participants with CD4 of 50 to 200 cells/mm3 and participants with CD4 < 50 cells/mm3 (68.9% vs 76.9%, p = 0.491).

  • LJ culture – sensitivity 81.3% (95%CI 73.2–87.7).

  • Ziehl Neelsen – sensitivity 31.7% (CI 23.6–40.7).

  • Direct fluorescence microscopy - sensitivity 35.0% (95%CI 26.5–44.0).

  • Concentrated fluorescent microscopy – sensitivity 43.9% (95%CI 34.9–53.1).

  • 6/8(2N/A)

  • - sampling was neither consecutive nor random (used secondary data);

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.
Cavanaugh et al.33/ PLOS ONE/ 2016/ Kenya Did not quoted. Characterize the value of various diagnostic tests, including the incremental yield of GeneXpert MTB/RIF and the culture on sputum smear microscopy. 778 people living with HIV. GeneXpert MTB/RIF, direct Ziehl-Neelsen smear microscopy, concentrated fluorescence microscopy, and MGIT 960 culture.

  • 88 (11.3%) cases of TB were confirmed. GeneXpert – detected 43 cases (incremental yield = 24.4%). In specific samples, it identified 48/84 cases (57.0%) while the morning samples identified 50/76 (66.0%) cases. Two GeneXpert tests detected 22/24 (92.0%) cases of TB with CD4 < 100 cells/μL and 30/45 (67.0%) cases with CD4 counts ≥ 100 cells/μl.

  • GeneXpert identified 45 (28.0%; 95%CI 16.6–39.3) patients with negative smears; identified six patients who were not identified by culture.

  • Microscopy – detected 25 (33.6%) cases in 76 morning sputum samples, 24 (31.5%; 95%CI 23.4–39.6) were identified by ZN microscopy and 29 (6.5% 95%CI 2.8–10.2) by FD.

  • Liquid culture – identified 10 (13.0%; 95%CI 6.4–19.7) additional cases that were not identified by microscopy or GeneXpert. In total, liquid culture and GeneXpert performed on a single morning sample identified 60 (79.0%; 95%CI 72.4–85) cases of pulmonary TB.

  • 7/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.
Mollel et al.34/ Tanzania Journal of Health Research/ 2017/ Tanzania Cross-sectional study. Evaluate the performance of GeneXpert MTB/RIF in the diagnosis of TB in TB/HIV co-infected patients. 69 individuals. GeneXpert MTB/RIF, microscopy and LJ method culture.

  • GeneXpert resulted in 09 positive results for TB and 60 negative results. Sensitivity – 100% (95%CI 66.4–100); specificity – 100% (95%CI 94–100); PPV 79.2% (95%CI 39–91.5); NPV 98.9% (95%CI 95–99.6).

  • Microscopy – 05 positives, 01 false positive, 59 negatives and 04 false negatives. Sensitivity – 55.6% (95%CI 21.4–86.3); specificity – 98.3% (95%CI 91–100); PPV 75.6% (95%CI 28.9–95.9). NPV 96.0% (95%CI 92.0–98.0).

  • 7/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.
Dorman et al.35/ The Lancet Infectious Diseases/ 2018/ South Africa, Uganda, Kenya, India, China, Georgia, Belarus, and Brazil Prospective study. Compare the diagnostic accuracy of the Xpert MTB/RIF Ultra test with that of the GeneXpert MTB/RIF for detection of negative smear tuberculosis and resistance to rifampicin. 1,753 adult individuals with symptoms of pulmonary tuberculosis. Xpert MTB/RIF Ultra test, smear microscopy using Ziehl-Neelsen and culture MGIT BACTEC 960 and LJ. The sensitivities of the Xpert Ultra and Xpert were 90% (95%CI 83–95) and 77% (95%CI 68–84), respectively, for the 115 HIV-positive participants with culture-positive sputum. The specificities of Xpert Ultra and Xpert for detection were respectively 96% and 98% overall and 93% and 98% for patients with a prior history of tuberculosis.

  • 6/8(2N/A)

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results;

  • - the study had 22.6% sample loss.
Auld et al.36/ Public Health Action/ 2014/ Cambodia Did not quoted. Describe the implementation and use of the GeneXpert MTB/RIF test in the diagnosis of tuberculosis among people living with HIV/AIDS. 497 individuals. GeneXpert MTB/RIF, Smear and Culture Microscopy. Among the 497 people with HIV, 357 (72%) were tested with sputum smear microscopy and 250 (50%) with Xpert; of which 25 (10%) were positive for TB and none resistant to rifampicin. Only 22/357 (4%) had sputum culture, being 04 (18%) positive for M. tuberculosis and 06 (27%) for non-tuberculous mycobacteria (NTM).

  • 4/8(2N/A)

  • - sampling was neither consecutive nor random (used secondary data);

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results;

  • - used sputum smear microscopy as a diagnostic method of comparison, which may not correctly classify a case of the disease;

  • - the reference standard test was not the same for all participants.
Balcells et al.37/ The International Journal of Tuberculosis and Lung Disease / 2012/ Chile Study cross-sectional. To assess the diagnostic accuracy of the GeneXpert MTB/RIF Assay in detecting pulmonary TB in HIV-infected patients in clinical care compared to traditional rapid smear microscopy and culture. 160 individuals. GeneXpert MTB/RIF, smear microscopy and cultures in LJ medium and MGIT BACTEC 960.

  • GeneXpert – sensitivity 91.7% (95%CI 64.6–98.5); specificity 99.3% (95%CI 96.3–99.9). PPV 91.7% (95%CI 64.6-98.5); NPV 99.3% (95%CI 96.3–99.9). Among the 12 culture-confirmed TB cases, two cases of resistance to rifampicin were detected by the Xpert MTB/RIF (16.6%).

  • Microscopy – sensitivity 66.7% (95%CI 39.1–86.2); specificity 98.6% (95%CI 95.2–99.6); PPV 80% (95%CI 49–94.3); NPV 97.3% (95%CI 93.3–99.3).

 8/8 (2N/A).
Lawn et al.38/ Clinical Infectious Diseases/ 2012/ South Africa Did not quoted. Report the characteristics of patients living with HIV/AIDS with a clinical diagnosis of TB and a negative result for GeneXpert MTB/RIF and report the Xpert status for subsequent pragmatic and clinical results. 602 sputum samples. GeneXpert MTB/RIF, Fluorescence Microscopy and Liquid Culture. M. tuberculosis was cultivated in 01 sample of 89 patients, which represented a prevalence of 17.0% (95%CI 13.9–20.5) of TB. Among them, 24 (27.0%) cases were positive for sputum smear microscopy and 52 (58.4%) for GeneXpert. When 02 samples were collected, the positivity for GeneXpert was 64 (71.9%). 8/8 (2N/A).
Cuervo et al.39/ Venezolan Archives of Pharmacology and Therapeutics/ 2018/Cuba Study retrospective. To reaffirm the impact-benefit of using GeneXpert MTB/RIF for the diagnosis of tuberculosis in HIV/AIDS patients with suspected pulmonary TB. 152 sputum samples. GeneXpert MTB/RIF, smear microscopy in Ziehl-Neelsen medium and culture by LJ. Of the 152 samples, GeneXpert detected 39 (25.65%) positive for Mycobacterium Tuberculosis, 36 (23.68%) being non-resistant to rifampicin, and 03 (2.00%) resistant to rifampicin. The percentage of positivity for sputum smear microscopy was 11.8% and for LJ culture was 21.7%. In comparison, researchers observed that the sensitivity, specificity, and agreement were above 90%, with no statistically significant differences between them.

  • 5/8 (2N/A)

  • - study is retrospective (possibly used secondary data);

  • - information about losses and exclusions of people from the sample was not revealed;

  • - it is not possible to know if the results of the reference test (sputum culture) were obtained without knowing GeneXpert results.

HIV: human immunodeficiency virus; 95%CI: 95% confidence interval; LJ: Löwenstein Jensen; MDR: multidrug-resistant; N/A: not applicable; WHO: World Health Organization; PLR: positive likelihood ratio; NLR: negative likelihood ratio; ART: antiretroviral therapy; TB: tuberculosis; PPV: positive predictive value; NPV: negative predictive value.