The effect of sputum quality and volume on the yield of bacteriologically-confirmed TB by Xpert MTB/RIF and smear

Feature fusion method for pulmonary tuberculosis patient detection based on cough sound

Since the COVID-19, cough sounds have been widely used for screening purposes. Intelligent analysis techniques have proven to be effective in detecting respiratory diseases. In 2021, there were up to 10 million TB-infected patients worldwide, with an annual growth rate of 4.5%. Most of the patients were from economically underdeveloped regions and countries. The PPD test, a common screening method in the community, has a sensitivity of as low as 77%. Although IGRA and Xpert MTB/RIF offer high specificity and sensitivity, their cost makes them less accessible. In this study, we proposed a feature fusion model-based cough sound classification method for primary TB screening in communities. Data were collected from hospitals using smart phones, including 230 cough sounds from 70 patients with TB and 226 cough sounds from 74 healthy subjects. We employed Bi-LSTM and Bi-GRU recurrent neural networks to analyze five traditional feature sets including the Mel frequency cepstrum coefficient (MFCC), zero-crossing rate (ZCR), short-time energy, root mean square, and chroma_cens. The incorporation of features extracted from the speech spectrogram by 2D convolution training into the Bi-LSTM model enhanced the classification results. With traditional futures, the best TB patient detection result was achieved with the Bi-LSTM model, with 93.99% accuracy, 93.93% specificity, and 92.39% sensitivity. When combined with a speech spectrogram, the classification results showed 96.33% accuracy, 94.99% specificity, and 98.13% sensitivity. Our findings underscore that traditional features and deep features have good complementarity when fused using Bi LSTM modelling, which outperforms existing PPD detection methods in terms of both efficiency and accuracy.

Increased tuberculosis case detection in Tanzanian children and adults using African giant pouched rats

BACKGROUND

African giant pouched rats, trained by Anti-Persoonsmijnen Ontmijnende Product Ontwikkeling (APOPO), have demonstrated their ability to detect tuberculosis (TB) from sputum. We assessed rat-based case detection and compared the mycobacterium bacillary load (MTB-load) in children versus adults.

METHODS

From January-December 2022, samples were collected prospectively from 69 Directly Observed Therapy (DOT) facilities' presumed TB patients. Using an average of five rats, APOPO re-evaluated patients with bacteriologically negative (sputum-smear microscopy or Xpert MTB/RIF) results. Rat-positive samples were tested using concentrated smear light-emitting diode microscopy to confirm TB detection before treatment initiation. The rats' identification of pulmonary TB is based on smelling TB-specific volatile organic compounds (VOCs) in sputum. Using STATA, Chi-square for odds ratio and confidence interval was calculated and evaluated: (1) the yield of rat-based TB detection compared to that of the health facilities; (2) rat-based TB detection in children versus adults; and (3) rats' ability to detect TB across MTB-loads and between children and adults.

RESULTS

From 35,766 patients, 5.3% (1900/35,766) were smear-positive and 94.7% (33,866/35,766) were smear or Xpert-negatives at DOTS facility. Of those with negative results, 2029 TB cases were detected using rats, contributing to 52% (2029/3929 of total TB identified), which otherwise would have been missed. Compared to DOT facilities, rats were six-fold more likely to detect TB among Acid Fast Bacilli (AFB) 1+/scanty [90% (1829/2029) versus 60% (1139/1900), odds ratio, OR = 6.11, 95% confidence interval, CI: 5.14-7.26]; twice more likely to identify TB cases among children [71% (91/129) versus 51% (1795/3542), OR = 2.3, 95% CI: 1.59-3.42]; and twice more likely to identify TB cases among children with AFB 1+/scanty than adults with the same MTB-load [5% (86/1703) versus 3% (28/1067), OR = 2.0, 95% CI: 1.28-3.03].

CONCLUSIONS

Rats contributed over half of the TB cases identified in program settings, and children, especially those with a lower MTB-load, were more likely to be diagnosed with TB by rats. The chemical signatures, VOCs, were only available for adults, and further research describing the characteristics of VOCs in children versus adults may pave the way to enhance TB diagnosis in children.

Evaluation of the GeneXpert MTB/RIF assay performance in sputum samples with various characteristics from presumed pulmonary tuberculosis patients in Shiselweni region, Eswatini

BACKGROUND

Tuberculosis is a global health concern and has become more complex to diagnose due to mutations in the causative agent, Mycobacterium tuberculosis. In a setting with high TB prevalence, having a rapid and accurate diagnosis may reduce the rate of infections. The study aimed to evaluate the diagnostic performance of GeneXpert MTB/RIF assay in sputum samples from suspected tuberculosis patients.

METHODS

A total of 1 328 sputum samples were collected from patients, across 12 clinics in the Shiselweni region, Eswatini. One thousand one hundred and ten (1110; 84%) samples were simultaneously processed on GeneXpert MTB/RIF assay and MGIT culture.

RESULTS

Two hundred and ninety-seven (297) samples tested GeneXpert positive and 813 GeneXpert negatives, while 310 samples tested positive and 800 tested negative on the MGIT culture method. The positive predictive value on GeneXpert MTB/RIF assay was 83% while the negative predictive value was 97.80%. At p-value = .796, sputum quality did not affect the positivity of the GeneXpert MTB/RIF results. Sputum volume had a significant impact on the performance of the GeneXpert MTB/RIF with increased sensitivity in 4 ml and 6 ml samples.

CONCLUSIONS

Although detection of tuberculosis using the GeneXpert MTB/RIF assay in sputum samples is not limited to one specific characteristic, sputum volume assessment should be considered as an integral part of routine laboratory diagnosis of tuberculosis especially in high tuberculosis prevalent settings. However, the ability of the GeneXpert MTB/RIF to provide rapid TB diagnosis is not dependent on sputum quality.

Xpert Ultra versus Xpert MTB/RIF for pulmonary tuberculosis and rifampicin resistance in adults with presumptive pulmonary tuberculosis

BACKGROUND

Xpert MTB/RIF and Xpert MTB/RIF Ultra (Xpert Ultra) are World Health Organization (WHO)-recommended rapid tests that simultaneously detect tuberculosis and rifampicin resistance in people with signs and symptoms of tuberculosis. This review builds on our recent extensive Cochrane Review of Xpert MTB/RIF accuracy.

OBJECTIVES

To compare the diagnostic accuracy of Xpert Ultra and Xpert MTB/RIF for the detection of pulmonary tuberculosis and detection of rifampicin resistance in adults with presumptive pulmonary tuberculosis. For pulmonary tuberculosis and rifampicin resistance, we also investigated potential sources of heterogeneity. We also summarized the frequency of Xpert Ultra trace-positive results, and estimated the accuracy of Xpert Ultra after repeat testing in those with trace-positive results.

SEARCH METHODS

We searched the Cochrane Infectious Diseases Group Specialized Register, MEDLINE, Embase, Science Citation Index, Web of Science, LILACS, Scopus, the WHO ICTRP, the ISRCTN registry, and ProQuest to 28 January 2020 with no language restriction.

SELECTION CRITERIA

We included diagnostic accuracy studies using respiratory specimens in adults with presumptive pulmonary tuberculosis that directly compared the index tests. For pulmonary tuberculosis detection, the reference standards were culture and a composite reference standard. For rifampicin resistance, the reference standards were culture-based drug susceptibility testing and line probe assays.

DATA COLLECTION AND ANALYSIS

Two review authors independently extracted data using a standardized form, including data by smear and HIV status. We assessed risk of bias using QUADAS-2 and QUADAS-C. We performed meta-analyses comparing pooled sensitivities and specificities, separately for pulmonary tuberculosis detection and rifampicin resistance detection, and separately by reference standard. Most analyses used a bivariate random-effects model. For tuberculosis detection, we estimated accuracy in studies in participants who were not selected based on prior microscopy testing or history of tuberculosis. We performed subgroup analyses by smear status, HIV status, and history of tuberculosis. We summarized Xpert Ultra trace results.

MAIN RESULTS

We identified nine studies (3500 participants): seven had unselected participants (2834 participants). All compared Xpert Ultra and Xpert MTB/RIF for pulmonary tuberculosis detection; seven studies used a paired comparative accuracy design, and two studies used a randomized design. Five studies compared Xpert Ultra and Xpert MTB/RIF for rifampicin resistance detection; four studies used a paired design, and one study used a randomized design. Of the nine included studies, seven (78%) were mainly or exclusively in high tuberculosis burden countries. For pulmonary tuberculosis detection, most studies had low risk of bias in all domains. Pulmonary tuberculosis detection Xpert Ultra pooled sensitivity and specificity (95% credible interval) against culture were 90.9% (86.2 to 94.7) and 95.6% (93.0 to 97.4) (7 studies, 2834 participants; high-certainty evidence) versus Xpert MTB/RIF pooled sensitivity and specificity of 84.7% (78.6 to 89.9) and 98.4% (97.0 to 99.3) (7 studies, 2835 participants; high-certainty evidence). The difference in the accuracy of Xpert Ultra minus Xpert MTB/RIF was estimated at 6.3% (0.1 to 12.8) for sensitivity and -2.7% (-5.7 to -0.5) for specificity. If the point estimates for Xpert Ultra and Xpert MTB/RIF are applied to a hypothetical cohort of 1000 patients, where 10% of those presenting with symptoms have pulmonary tuberculosis, Xpert Ultra will miss 9 cases, and Xpert MTB/RIF will miss 15 cases. The number of people wrongly diagnosed with pulmonary tuberculosis would be 40 with Xpert Ultra and 14 with Xpert MTB/RIF. In smear-negative, culture-positive participants, pooled sensitivity was 77.5% (67.6 to 85.6) for Xpert Ultra versus 60.6% (48.4 to 71.7) for Xpert MTB/RIF; pooled specificity was 95.8% (92.9 to 97.7) for Xpert Ultra versus 98.8% (97.7 to 99.5) for Xpert MTB/RIF (6 studies). In people living with HIV, pooled sensitivity was 87.6% (75.4 to 94.1) for Xpert Ultra versus 74.9% (58.7 to 86.2) for Xpert MTB/RIF; pooled specificity was 92.8% (82.3 to 97.0) for Xpert Ultra versus 99.7% (98.6 to 100.0) for Xpert MTB/RIF (3 studies). In participants with a history of tuberculosis, pooled sensitivity was 84.2% (72.5 to 91.7) for Xpert Ultra versus 81.8% (68.7 to 90.0) for Xpert MTB/RIF; pooled specificity was 88.2% (70.5 to 96.6) for Xpert Ultra versus 97.4% (91.7 to 99.5) for Xpert MTB/RIF (4 studies). The proportion of Ultra trace-positive results ranged from 3.0% to 30.4%. Data were insufficient to estimate the accuracy of Xpert Ultra repeat testing in individuals with initial trace-positive results. Rifampicin resistance detection Pooled sensitivity and specificity were 94.9% (88.9 to 97.9) and 99.1% (97.7 to 99.8) (5 studies, 921 participants; high-certainty evidence) for Xpert Ultra versus 95.3% (90.0 to 98.1) and 98.8% (97.2 to 99.6) (5 studies, 930 participants; high-certainty evidence) for Xpert MTB/RIF. The difference in the accuracy of Xpert Ultra minus Xpert MTB/RIF was estimated at -0.3% (-6.9 to 5.7) for sensitivity and 0.3% (-1.2 to 2.0) for specificity. If the point estimates for Xpert Ultra and Xpert MTB/RIF are applied to a hypothetical cohort of 1000 patients, where 10% of those presenting with symptoms have rifampicin resistance, Xpert Ultra will miss 5 cases, and Xpert MTB/RIF will miss 5 cases. The number of people wrongly diagnosed with rifampicin resistance would be 8 with Xpert Ultra and 11 with Xpert MTB/RIF. We identified a higher number of rifampicin resistance indeterminate results with Xpert Ultra, pooled proportion 7.6% (2.4 to 21.0) compared to Xpert MTB/RIF pooled proportion 0.8% (0.2 to 2.4). The estimated difference in the pooled proportion of indeterminate rifampicin resistance results for Xpert Ultra versus Xpert MTB/RIF was 6.7% (1.4 to 20.1).

AUTHORS' CONCLUSIONS

Xpert Ultra has higher sensitivity and lower specificity than Xpert MTB/RIF for pulmonary tuberculosis, especially in smear-negative participants and people living with HIV. Xpert Ultra specificity was lower than that of Xpert MTB/RIF in participants with a history of tuberculosis. The sensitivity and specificity trade-off would be expected to vary by setting. For detection of rifampicin resistance, Xpert Ultra and Xpert MTB/RIF had similar sensitivity and specificity. Ultra trace-positive results were common. Xpert Ultra and Xpert MTB/RIF provide accurate results and can allow rapid initiation of treatment for rifampicin-resistant and multidrug-resistant tuberculosis.

Comparison of Xpert MTB/RIF (G4) and Xpert Ultra, including trace readouts, for the diagnosis of pulmonary tuberculosis in a TB and HIV endemic setting

BACKGROUND

There are limited data about Xpert-Ultra performance in different settings, in HIV-infected persons, in those with a history of previous TB, and with trace readouts.

METHODS

We evaluated the relative accuracy of Xpert-MTB/RIF and Xpert-Ultra in 272 selected but well-characterized archived sputum samples. Of these, 168 were culture-positive (64/168 smear-positive and 104/168 smear-negative), and 104 were culture-negative (102/104 from patients with previous TB and 2/104 from patients without a TB history). Assay-specific limit-of-detection (LOD) experiments were conducted using serial dilutions of Mycobacterium tuberculosis H37Rv.

RESULTS

Overall sensitivity (95%CI) in smear-negative culture-positive samples for Xpert-MTB/RIF and Xpert-Ultra were 71.2% (62.5-79.9) and 77% (68.9-85.1), respectively (and in HIV-infected persons: 63.5% (50-76.1) and 73.1% (61.1-85.2), respectively). The LOD for Xpert-Ultra was lower (9 versus 184 CFU/ml). There were a total of 9/272 (3.3%) Xpert Ultra trace readouts (6/104 [5.8%]) in smear-negative culture-positive persons, and 3/102 (3%) in culture-negative non-TB persons with a history of previous TB).

CONCLUSIONS

Xpert-Ultra had a lower LOD compared to Xpert-MTB/RIF. A small proportion of samples (<5%) from culture-negative patients but with a history of previous TB had a likely false-positive trace readout. These data inform the management of patients with suspected TB in endemic settings.

Improving sputum collection processes to increase tuberculosis case finding among HIV-positive persons in Botswana

SETTING

Twenty-two human immunodeficiency virus (HIV) clinics in Botswana.

OBJECTIVE

To compare sputum collection rates, sputum quality and volume, and tuberculosis (TB) diagnosis rates before and after field efforts to improve sputum collection among individuals newly diagnosed with HIV with TB symptoms.

DESIGN

Newly diagnosed individuals living with HIV attending 22 HIV clinics in Botswana were screened for TB from August 2012 to March 2014. Starting in May 2013, a field intervention composed of the introduction of a tracking log for presumed TB patients, and patient instructions and sputum induction to improve sputum collection rates was implemented.

RESULTS

Prior to the intervention, sputum collection rates were 44.1% (384/870). Subsequently, sputum collection increased to 58.3% (579/993) (P < 0.001). Sputum quality and volume also improved. Although rates of TB diagnosis increased from 9.7% (84/870) to 12.5% (120/993), this difference was not significant (P = 0.143).

CONCLUSION

Sputum collection rates among presumptive TB cases, as well as sputum quality and volume improved after implementation of the field intervention. To improve sputum collection rates, efforts at the program level should be ongoing.