Skip to content
ALL Metrics
-
Views
Get PDF
Get XML
Cite
Export
Track
Systematic Review

Systematic review of diagnostic and prognostic host blood transcriptomic signatures of tuberculosis disease in people living with HIV

[version 1; peer review: 1 approved]
PUBLISHED 03 Feb 2023
Author details Author details
OPEN PEER REVIEW
REVIEWER STATUS

Abstract

Background
HIV-associated tuberculosis (TB) has high mortality; however, current triage and prognostic tools offer poor sensitivity and specificity, respectively. We conducted a systematic review of diagnostic and prognostic host-blood transcriptomic signatures of TB in people living with HIV (PLHIV).

Methods
We systematically searched online databases for studies published in English between 1990-2020. Eligible studies included PLHIV of any age in test or validation cohorts, and used microbiological or composite reference standards for TB diagnosis. Inclusion was not restricted by setting or participant age. Study selection, quality appraisal using the QUADAS-2 tool, and data extraction were conducted independently by two reviewers. Thereafter, narrative synthesis of included studies, and comparison of signatures performance, was performed.

Results
We screened 1,580 records and included 12 studies evaluating 31 host-blood transcriptomic signatures in 10 test or validation cohorts of PLHIV that differentiated individuals with TB from those with HIV alone, latent Mycobacterium tuberculosis infection, or other diseases (OD). Two (2/10; 20%) cohorts were prospective (29 TB cases; 51 OD) and 8 (80%) case-control (353 TB cases; 606 controls) design. All cohorts (10/10) were recruited in Sub-Saharan Africa and 9/10 (90%) had a high risk of bias. Ten signatures (10/31; 32%) met minimum WHO Target Product Profile (TPP) criteria for TB triage tests. Only one study (1/12; 8%) evaluated prognostic performance of a transcriptomic signature for progression to TB in PLHIV, which did not meet the minimum WHO prognostic TPP.

Conclusions
Generalisability of reported findings is limited by few studies enrolling PLHIV, limited geographical diversity, and predominantly case-control design, which also introduces spectrum bias. New prospective cohort studies are needed that include PLHIV and are conducted in diverse settings. Further research exploring the effect of HIV clinical, virological, and immunological factors on diagnostic performance is necessary for development and implementation of TB transcriptomic signatures in PLHIV.

Keywords

Host, blood, diagnostic, prognostic, gene, signatures, tuberculosis, HIV

Introduction

There were an estimated 703,000 HIV-associated incident tuberculosis (TB) cases in 2021 however only 368,600 (52%) were notified, with a resultant case fatality rate of 27%1. Earlier diagnosis and initiation of treatment, or disease prevention through targeted short-course TB preventive therapy (TPT), may reduce this burden. However, we lack adequate TB mass screening tools to direct confirmatory testing or prognostic tools to guide preventive therapy in the outpatient or community setting. Symptom screening, the most widely used TB triage tool, has low specificity in antiretroviral therapy (ART-) naïve and low sensitivity in ART-experienced people living with HIV (PLHIV)2. The addition of chest radiography improves sensitivity, at a cost of reduced specificity2. With almost three-quarters of the 38 million PLHIV globally now receiving ART3, new tools should be efficacious in this group.

The WHO currently recommends that PLHIV with a positive or unknown tuberculin skin test (TST) result should receive TPT, if active TB has been excluded4. There is strong evidence to support such an approach5. However, TST and interferon-γ release assay (IGRA) reflect a memory T-cell response following Mycobacterium tuberculosis (Mtb) exposure (sensitisation) and not necessarily ongoing infection. In TB-endemic countries with high rates of Mtb transmission and exposure, these tests have limited utility for guiding TPT6,7. In addition, loss or dysfunction of Mtb-specific memory T-cells among immunocompromised PLHIV with low CD4 cell counts results in lower IGRA positivity and may reduce sensitivity for predicting progression to disease8,9. There is also limited evidence regarding repeat courses of TPT among PLHIV; a recent study demonstrated that universal retreatment after one year did not provide additional benefit10.

Biomarker-guided treatment has been proposed to target therapy to those that need it most, reducing unnecessary pill burden, drug interactions, and side effects in individuals, and increasing efficacy and cost-effectiveness of mass screening11,12. Host-response blood transcriptomic signatures can identify those with active TB and those who are progressing to disease1315. Performance of most signatures in adults without HIV meet at least one of the minimum World Health Organization (WHO) Target Product Profile (TPP) TB triage test performance criteria (sensitivity 90% and specificity 70%) for diagnosing prevalent TB16,17. Several signatures have been shown to meet minimum prognostic benchmarks (sensitivity 75% and specificity 75%)18 for short-term prediction of progression to TB disease within six months of testing1921. However, only a couple of signatures, Roe1 and Roe322,23, meet these criteria through 12 months of follow-up for progression21. Transcriptomic biomarkers selected for advancement through the diagnostics pipeline for development as point-of-care assays should also perform well in PLHIV. We systematically reviewed the published literature on host-response blood transcriptomic biomarkers for diagnosing prevalent and predicting progression to incident TB disease in PLHIV, and compared performance to the WHO TPP criteria.

Methods

Protocol and registration

This review is reported in line with the Preferred Reporting Items for Systematic reviews and Meta-Analysis of Diagnostic Test Accuracy Studies (PRISMA-DTA)24 recommendations (Table 1). The systematic review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) on 02 January 2021 with registration number CRD42021224155 and published in the BMJ Open25.

Table 1. PRISMA for Diagnostic Test Accuracy 2018 checklist.

This checklist has been adapted from McInnes MDF, Moher D, Thombs BD, McGrath TA, Bossuyt PM, The PRISMA-DTA Group (2018). Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies: The PRISMA-DTA Statement. JAMA. 2018 Jan 23;319(4):388-396. doi: 10.1001/jama.2017.19163.

Section/topic#PRISMA-DTA Checklist ItemReported on
page #
TITLE / ABSTRACT
Title 1Identify the report as a systematic review (+/- meta-analysis) of diagnostic test accuracy
(DTA) studies.
Page 1
Abstract2Abstract: See PRISMA-DTA for abstracts.Page 1
INTRODUCTION
Rationale 3Describe the rationale for the review in the context of what is already known. Page 3
Clinical role of index
test
D1State the scientific and clinical background, including the intended use and clinical role
of the index test, and if applicable, the rationale for minimally acceptable test accuracy
(or minimum difference in accuracy for comparative design).
Page 3
Objectives 4Provide an explicit statement of question(s) being addressed in terms of participants,
index test(s), and target condition(s).
Page 3
METHODS
Protocol and
registration
5Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address),
and, if available, provide registration information including registration number.
Page 3
Eligibility criteria 6Specify study characteristics (participants, setting, index test(s), reference standard(s),
target condition(s), and study design) and report characteristics (e.g., years considered,
language, publication status) used as criteria for eligibility, giving rationale.
Page 3
Information sources 7Describe all information sources (e.g., databases with dates of coverage, contact with
study authors to identify additional studies) in the search and date last searched.
Page 3
Search 8Present full search strategies for all electronic databases and other sources searched,
including any limits used, such that they could be repeated.
Page 3
Study selection 9State process for selecting studies (i.e. screening, eligibility, included in systematic
review, and, if applicable, included in meta-analysis).
Page 4
Data collection
process
10Describe method of data extraction from reports (e.g., piloted forms, independently, in
duplicate) and any processes for obtaining and confirming data from investigators.
Page 4
Definitions for data
extraction
11Provide definitions used in data extraction and classifications of target condition(s),
index test(s), reference standard(s) and other characteristics (e.g. study design, clinical
setting).
Page 3
Risk of bias &
applicability
12Describe methods used for assessing risk of bias in individual studies and concerns
regarding the applicability to the review question.
Page 6
Diagnostic accuracy
measures
13State the principal diagnostic accuracy measure(s) reported (e.g. sensitivity, specificity)
and state the unit of assessment (e.g. per-patient, per-lesion).
Page 5
Synthesis of results 14Describe methods of handling data, combining results of studies and describing
variability between studies. This could include, but is not limited to: a) handling of
multiple definitions of target condition. b) handling of multiple thresholds of test
positivity, c) handling multiple index test readers, d) handling of indeterminate test
results, e) grouping and comparing tests, f) handling of different reference standards
Page 5
Meta-analysisD2Report the statistical methods used for meta-analyses, if performed.NA
Additional analyses 16Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-
regression), if done, indicating if pre-specified.
NA
RESULTS
Study selection 17Provide numbers of studies screened, assessed for eligibility, included in the review (and
included in meta-analysis, if applicable) with reasons for exclusions at each stage, ideally
with a flow diagram.
Page 6
Figure 1
Study characteristics 18For each included study provide citations and present key characteristics including:
a) participant characteristics (presentation, prior testing), b) clinical setting, c) study
design, d) target condition definition, e) index test, f) reference standard, g) sample size,
h) funding sources
Table 2, Table 3
Risk of bias and
applicability
19Present evaluation of risk of bias and concerns regarding applicability for each study.Page 14, Figure 2
Results of individual
studies
20For each analysis in each study (e.g. unique combination of index test, reference
standard, and positivity threshold) report 2x2 data (TP, FP, FN, TN) with estimates of
diagnostic accuracy and confidence intervals, ideally with a forest or receiver operator
characteristic (ROC) plot.
Figure 3Figure 6
Synthesis of results 21Describe test accuracy, including variability; if meta-analysis was done, include results
and confidence intervals.
Pages 14–17
Additional analysis 23Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses,
meta-regression; analysis of index test: failure rates, proportion of inconclusive results,
adverse events).
NA
DISCUSSION
Summary of evidence 24Summarize the main findings including the strength of evidence.Pages 17–19
Limitations 25Discuss limitations from included studies (e.g. risk of bias and concerns regarding
applicability) and from the review process (e.g. incomplete retrieval of identified
research).
Pages 17–19
Conclusions 26Provide a general interpretation of the results in the context of other evidence. Discuss
implications for future research and clinical practice (e.g. the intended use and clinical
role of the index test).
Pages 17–19
FUNDING
Funding 27For the systematic review, describe the sources of funding and other support and the
role of the funders.
Page 2

Eligibility criteria

We considered cross-sectional and case-control studies, prospective and retrospective cohort studies, and randomised control trials evaluating diagnostic and/or prognostic performance of human host-blood transcriptomic signatures of TB (index tests). Eligible studies included PLHIV in the signature test and/or validation cohorts. Studies that only reported signature discovery cohort performance, or treatment response and failure monitoring cohorts, were not considered. PLHIV of all ages, ethnicities, and in all settings were considered. Studies which did not report any measures of signature performance (sensitivity and specificity, or reported data which enable the reconstruction of a two-by-two table for test accuracy calculation for PLHIV), did not clearly state the case definition of TB disease, did not report primary data, or did not independently report signature performance in PLHIV, were excluded.

Endpoint definitions

The primary TB disease endpoint (target condition) was defined by a positive microbiological test, such as mycobacterial culture or the Xpert MTB/RIF assay (reference standards), in sputum or other bodily fluid sample. Microbiologically-confirmed extra-pulmonary TB disease was also considered. The secondary TB disease endpoint was defined by non-microbiologically-confirmed, presumptive TB diagnosed via composite clinical features. TB disease diagnosed within one month of the index test was presumed to be prevalent disease (diagnostic studies). Prognostic studies were defined as prospective studies in which participants were followed up for progression to incident TB disease with measurement of a transcriptomic signature from blood samples collected at enrolment. Eligible studies included healthy individuals, latent Mtb-infected individuals, or individuals with other respiratory or systemic diseases as a control group. Latent Mtb infection was defined by a positive TST or IGRA.

Search strategy and information sources

We systematically searched PubMed (MEDLINE), WOS Core Collection, Biological Abstracts, and SciELO Citation Index (via Web of Science), Africa-Wide Information and General Science Abstracts (via EBSCOhost), Scopus, and Cochrane Central Register of Controlled Trials databases for full-text articles published in English between 1 January 1990 and 31 December 2020 using Medical Subject Headings (MeSH) and keyword search terms for "Diagnosis", "Messenger RNA", "Biomarkers/blood”, “Tuberculosis”, and “HIV”. The search strategy, including publication date range, were prespecified and published in a systematic review protocol25. We reviewed reference lists of eligible articles and performed forward citation tracking using Science Citation Index (via Web of Science) to identify further articles and reports missed by the electronic database search26.

Study selection and data collection

Two reviewers (SCM and SV) independently conducted the literature search and screened the search outputs for potential inclusion using EndNote bibliographic software to manage references, as previously described27. After removal of duplicates, the selection process included an initial screening of titles and abstracts for relevance, followed by full text review for eligibility. The two reviewers resolved any disagreements or uncertainties by discussion. Data elements of included studies were then independently extracted by the two reviewers. Corresponding authors of potentially eligible studies were contacted to provide deidentified participant-level data to reconstruct two-by-two tables or summary performance data for the PLHIV subgroup. Studies without summary or participant-level data available for the PLHIV subgroup were excluded.

Data analysis

We performed a narrative synthesis of the eligible study cohorts and signatures, including study design, cohort and signature characteristics, and diagnostic and prognostic performance of signatures stratified by study control groups (healthy, latent-Mtb infected, or other disease), and diagnostic reference standards (microbiological or composite clinical). Studies and cohorts were designated by the first author name and year of publication (e.g. Author2019a) and signatures by first author and number of transcripts (e.g. Author11). Signature area under the curve (AUC), sensitivity, and specificity were summarised in forest plots (R forestplot package28). For studies with available participant-level data for the PLHIV subgroup, we were able to recalculate AUC (R pROC package29), and benchmark sensitivity and specificity against the WHO TPP minimum performance criteria for a triage (70% specificity and 90% sensitivity)16 or prognostic (75% specificity and 75% sensitivity)18 test. 95% Confidence intervals for AUCs and sensitivity and specificity, were calculated using the DeLong30 and Wilson binomial31 methods, respectively. For studies in which participant-level data were not available, we report summary AUC, sensitivity, and specificity estimates, and 95% confidence intervals, for the PLHIV subgroup as published in the original papers. Most of these estimates were not specifically benchmarked against the WHO TPP minimum performance criteria for a triage test.

Risk of bias, applicability, and quality of evidence

The methodological quality and applicability concerns of included studies was assessed by the two reviewers using the Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool32 and graphically represented using traffic-light plots (R robvis package33). Risk of bias and applicability concerns for individual study was evaluated in four domains relating to (1) patient selection, (2) measurement of the index test, (3) measurement of the reference standard, and (4) study flow and timing of investigational and diagnostic procedures. Risks of bias are reported as low risk, some concerns, or high risk. We assessed the cumulative quality of evidence synthesised by the systematic review using the “Grading of Recommendations Assessment, Development and Evaluation” (GRADE) approach34 with classification based on study design and limitations, indirectness, inconsistency, imprecision, and publication bias35,36.

Results

Search results

We performed the literature search in January 2021, identifying 1,580 unique records published between 1 January 1990 and 31 December 2020, of which 98 full-text articles were assessed for eligibility, and 12 studies13,22,3747 met all criteria for inclusion (Figure 1). The main reasons for study exclusion were absence of PLHIV in study cohorts (n=29), absence of an independent test or external validation cohort which included PLHIV (n=20), and inappropriate index test (n=16) or study design (n=10). Nine of 10 studies excluded for inappropriate design were commentaries or reviews. In addition, deidentified participant-level data to reconstruct two-by-two tables, or summary performance data for PLHIV, were not available and no data were received from corresponding authors for 9 records19,4855.

de232003-6657-44ea-b313-10efa7fbfb91_figure1.gif

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram.

Study cohorts included in quantitative synthesis

The 12 eligible studies included 10 independent test or validation cohorts featuring PLHIV (Table 2), cumulatively evaluating diagnostic performance of 31 transcriptomic signatures (Table 3) incorporating over 700 unique transcripts. All independent test and validation cohorts enrolled PLHIV from outpatient clinics or hospital inpatients, most with suspicion (symptom-positive) or high risk (initiating antiretroviral therapy) of prevalent TB (Table 2). No studies prospectively enrolled participants without significant risk factors from community settings with a high burden of subclinical TB. Cohort sizes were generally small, with a maximum of 97 TB cases and 176 controls with HIV in the Kaforou201337 study, and 67 TB cases and 134 controls with HIV in the Södersten202046 study. All other cohorts included less than 50 TB cases with HIV. Only one of the 12 studies (Darboe2019)42 reported transcriptomic signature prognostic performance for incident TB disease (disease recurrence after completion of TB treatment). The remainder reported diagnostic performance for differentiating patients with pulmonary TB disease from healthy controls, from controls with latent TB infection, or from other diseases. Most studies were conducted in Africa, with 8 out of 10 cohorts enrolling participants in South Africa. Only one cohort (Södersten2020)46 included participants from another continent (Peru, South America). Many studies used the same microarray datasets, GSE3725037 and GSE3994138, from the Kaforou2013 (Malawi and South Africa) and Anderson2014 (Kenya, Malawi and South Africa) studies respectively, for external validation of signatures. Three cohorts measured signature scores using near-point-of-care RT-qPCR platforms (Darboe2019 and Penn-Nicholson2019a,b)42,45, while only one study (Södersten2020)46 used a point-of-care device, the Cepheid Xpert-MTB Host-Response (HR) Prototype, to measure the Sweeney3 signature19,49 in biobanked whole blood RNA samples. No included studies validated signature performance for diagnosing extrapulmonary TB. However, one cohort (Anderson2014b)38 included 17 HIV-infected children with non-microbiologically-confirmed, presumptive clinical TB with clinical and radiologic features that prompted empirical treatment. All other study cohorts, with the exception of Anderson2014, enrolled adults and used a microbiological reference standard (culture or Xpert MTB/RIF).

Table 2. Study cohorts included in quantitative synthesis.

Cohort ID*Cohort name
or accession
number
Study designStudy settingCountryAge
group
Cohort
type
CasesControlsSample
type
Signature
measurement
method
HIV+
TB
cases
HIV+
controls
Anderson2014a38GSE39941Case-control
nested within
prospective cohort
Children with
suspected TB
(Hospital setting)
Malawi,
South Africa
Children DiagnosticSputum culture+ ODWhole
blood
Microarray 4065
Anderson2014b38GSE39941Case-control
nested within
prospective cohort
Children with
suspected TB
(Hospital setting)
Kenya Children DiagnosticSputum culture+
(n=10) or composite
reference standard
(n=17)
ODWhole
blood
Microarray2727
Darboe201942TRuTH CohortCase-control
nested within
prospective cohort
Outpatients
who recently
completed TB
treatment
South Africa Adults PrognosticSputum culture+
recurrent TB within
3 years of treatment
completion
No
recurrent
TB
PBMCRT-qPCR3884
Kaforou2013a37GSE37250Case-control studyHospital
inpatient,
community, clinic
Malawi,
South Africa
AdultsDiagnosticSputum culture+ LTBIWhole
blood
Microarray9784
Kaforou2013b37GSE37250Case-control studyHospital
inpatient,
community, clinic
Malawi,
South Africa
AdultsDiagnosticSputum culture+ ODWhole
blood
Microarray92
Penn-
Nicholson2019a45
CTBC CohortCase-control studyClinic outpatient South Africa Adults DiagnosticSputum culture+ or
Xpert MTB/RIF+
LTBI+HCWhole
blood
RT-qPCR4440
Penn-
Nicholson2019b45
ScreenTB and
AE-TBC Cohorts
Prospective cohortSymptomatic
clinic attendees
South Africa Adults DiagnosticSputum culture+ or
Xpert MTB/RIF+
ORDWhole
blood
RT-qPCR1224
Rajan201844,47Data not
publicly
available
Case-control
nested within
prospective cohort
Outpatients
initiating ART
UgandaAdultsDiagnosticSputum culture+ LTBI+HCWhole
blood
Microarray4080
Södersten202046Biobanked
PAXgene
samples
Nested case-
control study
Symptomatic
inpatients
Peru, South
Africa
AdultsDiagnosticSputum/blood culture+
or sputum/urine Xpert
MTB/RIF+
ODWhole
blood
Cepheid
Xpert-MTB-HR-
Prototype
67134
Turner202013Array Express:
E-MTAB-8290
Prospective cohortSymptomatic
clinic attendees
South Africa Adults DiagnosticSputum culture+
or Xpert MTB/RIF+
ORDWhole
blood
RNA sequencing1727

* A study may evaluate multiple signatures using several validation cohorts. Studies and cohorts are designated by the first author name and year of publication (e.g. Author2019a) and signatures by first author and number of transcripts (e.g. Author11).

GSE, Genomic Spatial Event (database). TruTH, TB Recurrence upon Treatment with HAART. CTBC, Cross-sectional TB Cohort. AE-TBC, African-European Tuberculosis Consortium. LTBI, latent tuberculosis infection. HC, healthy controls. OD, other diseases. ORD, other respiratory diseases. PBMC, peripheral blood mononuclear cell. RT-qPCR, real-time quantitative polymerase chain reaction.

Transcriptomic signatures included in quantitative synthesis

All 31 signatures evaluated were discovered in case-control cohorts, eight of which were nested within prospective cohorts (Table 3). Only one of the 31 signatures (Rajan5)44 was discovered in a cohort exclusively consisting of PLHIV. Seven signatures (Duffy1043, Gliddon355, Gliddon455, Kaforou2737, Kaforou4437, Kaforou5337, and Sweeney319,49) were discovered using the Kaforou2013 cohort37 and two signatures (Anderson42 and Anderson51)38 using the Anderson2014 cohort38, which included participants both with and without HIV. The 31 signatures were predominantly discovered in adult cohorts, with only 6 out 31 derived in paediatric (Gjøen751, Anderson42, and Anderson51) or adolescent (Darboe1156, Penn-Nicholson645, and Zak1640) cohorts. Signature discovery cohorts were geographically diverse, with recruitment in Asia (3 signatures), Europe (10), North America (4), and South America (1); however there was a predominance in Africa (16), particularly South Africa (15).

Table 3. Transcriptomic signatures included in quantitative synthesis.

Signature ID*Year
published
Intended useSignature
model
Discovery cohort
study design
Discovery
cohort
country
Discovery
cohort age
group
Discovery
cohort
HIV
status
Discovery dataDiscovery approach
Anderson42 (39.
LTBI)38
2014Diagnostic (TB vs
LTBI)
Disease risk
score
Case-control
nested in
prospective cohort
Malawi, South
Africa
ChildrenMixedGenome-wide RNA
Microarray
Elastic net
Anderson51 (39.
OD)38
2014Diagnostic (TB vs
OD)
Disease risk
score
Case-control
nested in
prospective cohort
Malawi, South
Africa
ChildrenMixedGenome-wide RNA
Microarray
Elastic net
Berry86582010Diagnostic (TB vs
OD)
k-nearest
neighbours
algorithm
Case-controlUnited KingdomAdultsNegativeGenome-wide RNA
Microarray
Differential expression
Berry393582010Diagnostic (TB vs
LTBI+HC)
k-nearest
neighbours
algorithm
Case-controlSouth Africa,
United
Kingdom
AdultsNegativeGenome-wide RNA
Microarray
Differential expression
Darboe11562018Prognostic (Incipient
TB vs HC)
Pair-wise SVM
ensemble
Case-control
nested in
prognostic cohort
South AfricaAdolescentsNegativeRT-qPCR using
selected target gene
set (Zak16)
Refinement of Zak16:
most reproducible and
least redundant genes
de Araujo1
(NPC2)59
2016Diagnostic (TB vs HC
+ LTBI)
Standardised
expression
Case-controlBrazilAdultsNot statedGenome-wide RNA
Sequencing
Differential expression
Duffy10432019Diagnostic (TB vs
LTBI + OD)
Six-class
multinomial
random forest
Case-controlMalawi, South
Africa
AdultsMixedGenome-wide RNA
Microarray
Multinomial random
forest
Gjøen8 (7)512017Diagnostic (TB vs
HC)
LASSO
regression
Case-controlIndiaChildrenNegative198 selected target
genes measured by
dc-RT MLPA
LASSO
Gliddon3552020Diagnostic (TB vs
LTBI)
Disease risk
score
Case-controlMalawi, South
Africa
AdultsMixedGenome-wide RNA
Microarray
FS-PLS
Gliddon4552020Diagnostic (TB vs
OD)
Disease risk
score
Case-controlMalawi, South
Africa
AdultsMixedGenome-wide RNA
Microarray
FS-PLS
Huang13 (11)602015Diagnostic (TB vs HC
+ OD)
SVMCase-controlUnited
Kingdom
AdultsNegativeGenome-wide RNA
Microarray
LASSO, L1/2, and elastic
net
Kaforou27 (25)372013Diagnostic (TB vs
LTBI)
Disease risk
score
Case-controlMalawi, South
Africa
AdultsMixedGenome-wide RNA
Microarray
Elastic net
Kaforou44 (39)372013Diagnostic (TB vs
OD)
Disease risk
score
Case-controlMalawi, South
Africa
AdultsMixedGenome-wide RNA
Microarray
Elastic net
Kaforou53 (45)372013Diagnostic (TB vs
OD+LTBI)
Disease risk
score
Case-controlMalawi, South
Africa
AdultsMixedGenome-wide RNA
Microarray
Elastic net
Maertzdorf4482015Diagnostic (TB vs HC
+ LTBI)
Random forestCase-controlIndiaAdultsNegative360 selected target
genes measured by
RT-qPCR
Random forest
Penn-
Nicholson645
2020Prognostic (Incipient
TB vs HC)
Pair-wise
ensemble
structure
Case-control
nested in
prognostic cohort
South AfricaAdolescentsNegativeGenome-wide RNA
Sequencing and
targeted RT-qPCR
SVM-based gene pair
ratios approach
Qian17502016Diagnostic (TB vs HC
+ OD)
Sum of
standardised
expression
Case-controlUnited
Kingdom
AdultsNegative862 Nrf2-mediated
genes measured by
RNA Microarray
Differential expression
Rajan5442018Diagnostic (HIV+TB
vs HIV+HC)
Unsigned
sums
Case-control
nested in
prospective cohort
UgandaAdultsPositiveGenome-wide RNA
Microarray
Differential expression
Roe1 (BATF2)222016Diagnostic (TB vs HC
and TB patients 2-4
years post recovery)
Standardised
expression
Case-controlUnited KingdomAdultsNegativeGenome-wide RNA
Microarray
SVM (linear kernel)
Roe3232020Prognostic (Incipient
TB vs HC)
SVM (linear
kernel)
Case-controlUnited
Kingdom
AdultsNegativeGenome-wide RNA
Sequencing
Stability selection and
SVM
Roe4222016Diagnostic (TB vs
OD)
SVM (linear
kernel)
Case-controlUnited
Kingdom
AdultsNegativeGenome-wide RNA
Microarray
SVM (linear kernel)
Roe5222016Diagnostic (TB vs HC
+ OD)
SVM (linear
kernel)
Case-controlUnited
Kingdom
AdultsNegativeGenome-wide RNA
Microarray
SVM (linear kernel)
Sambarey10412017Diagnostic (TB vs HC
+ LTBI)
Linear
discriminant
analysis
Case-controlIndiaAdultsNegativeGenome-wide RNA
Sequencing
Human Protein–
Protein Interaction
Network, Condition-
specific Weighted
Networks, Differential
expression
Singhania20532018Diagnostic (TB vs
LTBI)
Disease
risk score
(modified)
Case-controlUnited
Kingdom
AdultsNegativeGenome-wide RNA
Sequencing
Random forest using
modular approach
Suliman2612018Prognostic (Incipient
TB vs HC)
Pair-wise
ensemble
structure
Case-control
nested in
prognostic cohort
The Gambia,
South Africa
AdultsNegativeGenome-wide RNA
Sequencing
SVM-based gene pair
ratios approach
Suliman4612018Prognostic (Incipient
TB vs HC)
Pair-wise
ensemble
structure
Case-control
nested in
prognostic cohort
The Gambia,
Ethiopia, South
Africa
AdultsNegativeGenome-wide RNA
Sequencing
SVM-based gene pair
ratios approach
Sweeney319,492016Diagnostic (TB vs HC
+ LTBI + OD)
[(GBP5 +
DUSP3) / 2]
– KLF2
Multicohort
analysis (Case-
control)
France,
Malawi, South
Africa, United
Kingdom,
USA
AdultsMixedGenome-wide RNA
Microarray
Significance
thresholding and
forward search
Walter47 (32)392016Diagnostic (TB vs
OD)
SVMCase-controlUSAAdultsNegativeGenome-wide RNA
Microarray
SVM with recursive
feature elimination
Walter51 (46)392016Diagnostic (TB vs
LTBI)
SVMCase-controlUSAAdultsNegativeGenome-wide RNA
Microarray
SVM with recursive
feature elimination
Walter119 (101)392016Diagnostic (TB vs
LTBI + OD)
SVMCase-controlUSAAdultsNegativeGenome-wide RNA
Microarray
SVM with recursive
feature elimination
Zak16402016Prognostic (Incipient
TB vs LTBI)
Pair-wise SVM
ensemble
Case-control
nested in
prognostic cohort
South AfricaAdolescentsNegativeGenome-wide RNA
Sequencing
SVM-based gene pair
approach

* A study may evaluate multiple signatures using several validation cohorts. Studies and cohorts are designated by the first author name and year of publication (e.g. Author2019a) and signatures by first author and number of transcripts (e.g. Author11). Numbers in brackets indicate the subsequently reduced number of transcripts in a given signature due to duplicate transcript symbols (IDs) or where transcript sequences from an original discovery cohorts could not be mapped to a more recent reference transcriptome.

LTBI, latent tuberculosis infection. OD, other diseases. HC, healthy controls. SVM, support vector machines. LASSO, Least Absolute Shrinkage and Selection Operator. FS-PLS, Forward Selection-Partial Least Squares. RT-qPCR, real-time quantitative polymerase chain reaction. Dc-RT MLPA, dual-color-Reverse-Transcriptase-Multiplex-Ligation-dependent-Probe-Amplification.

The 40 most frequent transcripts, all included in 3 or more signatures, are listed in Table 4. We used the INTERFEROME database57 to classify interferon-stimulated genes (ISGs): We defined ISGs as genes significantly up or down regulated in expression (>1.5 fold change) in any human samples treated with Type-I IFN, relative to control samples. Almost all (38/39) of the most common transcripts with available gene annotations were classified as ISGs. The six transcripts most frequently included in signatures were Guanylate Binding Protein (GBP) 5 (11 signatures), GBP6 (10 signatures), Complement C1q B Chain (C1QB) and Fc fragment of IgG receptor Ia (FCGR1A) (7 signatures each), and Basic Leucine Zipper ATF-Like Transcription Factor 2 (BATF2) and GBP2 (6 signatures each).

Table 4. Forty most frequent transcripts included in 3 or more transcriptomic signatures, sorted by transcript use frequency.

Gene IDTranscript use
frequency
*Interferon-
stimulated gene
Anderson42 (39)Anderson51 (39)Berry393Berry86Darboe11de Araujo1Duffy10Gjøen8Gliddon3Gliddon4Huang13 (11)Kaforou27 (25)Kaforou44 (39)Kaforou53 (45)Maertzdorf4
GBP511 Yes111111
GBP610 Yes11111111
C1QB7 Yes111111
FCGR1A7 Yes1111
BATF26 Yes11
GBP26 Yes11
ANKRD225 Yes111
DUSP35 Yes1111
FCGR1B5 Yes1111
GBP15 Yes111
PRDM15 Yes111
SCARF15 Yes1
SEPT45 Yes111
SERPING15 Yes111
ALDH1A14 No1111
CYB5614 Yes111
DHRS94 Yes11
FCGR1C4 Yes111
GAS64 Yes111
IFITM34 Yes111
LHFPL24 Yes1111
SMARCD34 Yes111
STAT14 Yes11
VAMP54 Yes111
ACTA23 Yes11
BLK3 Yes11
C20ORF103 (LAMP5)3 Yes1
CD743 Yes111
CREB53 Yes111
DEFA13 Yes111
ETV73 Yes11
FER1L3 (MYOF)3 Yes111
ID33 Yes111
KREMEN13 Yes11
LOC3893863 NA**111
OSBPL103 Yes11
POLB3 Yes11
TAP13 Yes11
TMCC13 Yes111
TRAFD13 Yes11

Gene IDTranscript use
frequency
*Interferon-
stimulated gene
Penn-
Nicholson6
Qian17Rajan5Roe1Roe3Roe4Roe5Sambarey10Singhania20Suliman2Suliman4Sweeney3Walter47 (32)Walter51 (46)Walter119 (101)Zak16
GBP511 Yes11111
GBP610 Yes11
C1QB7 Yes1
FCGR1A7 Yes111
BATF26 Yes1111
GBP26 Yes1111
ANKRD225 Yes11
DUSP35 Yes1
FCGR1B5 Yes1
GBP15 Yes11
PRDM15 Yes11
SCARF15 Yes1111
SEPT45 Yes11
SERPING15 Yes11
ALDH1A14 No
CYB5614 Yes1
DHRS94 Yes11
FCGR1C4 Yes1
GAS64 Yes1
IFITM34 Yes1
LHFPL24 Yes
SMARCD34 Yes1
STAT14 Yes11
VAMP54 Yes1
ACTA23 Yes1
BLK3 Yes1
C20ORF103 (LAMP5)3 Yes11
CD743 Yes
CREB53 Yes
DEFA13 Yes
ETV73 Yes1
FER1L3 (MYOF)3 Yes
ID33 Yes
KREMEN13 Yes1
LOC3893863 NA**
OSBPL103 Yes1
POLB3 Yes1
TAP13 Yes1
TMCC13 Yes
TRAFD13 Yes1

* Gene significantly up or down regulated in expression (>1.5 fold change) in samples treated with Type-I interferon, relative to control samples, were defined as interferon-stimulated genes (ISGs). Data from INTERFEROME database (Samarajiwa SA, Forster S, Auchettl K, Hertzog PJ. INTERFEROME: the database of interferon regulated genes. Nucleic Acids Res. 2009;37:D852-7).

** No updated gene annotation available. NA, not available.

No experimental data demonstrating >1.5 fold-change in expression in response to Type-I interferon, however, there was >1.5 fold-change in expression response to Type-II interferon.

Quality appraisal of eligible studies

In the patient selection domain, 5 of the 10 independent test and external validation cohorts consecutively or randomly enrolled participants, with the remaining cohorts not reporting sampling method (Figure 2). Eight of the 10 cohorts utilised a case-control design (Table 2), with exclusion of participants with uncertain diagnosis introducing a high risk of spectrum bias and potentially inflating diagnostic accuracy. The Penn-Nicholson2019b cohort45 used a prospective design in recruiting symptomatic clinic attendees, but excluded probable and uncertain TB cases from analysis. Only one prospective diagnostic accuracy study, Turner202013, measured signature scores and tested performance in all enrolled participants with clinically suspected tuberculosis, including those with uncertain diagnosis, representative of the target population of symptomatic clinic attendees.

de232003-6657-44ea-b313-10efa7fbfb91_figure2.gif

Figure 2. Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2).

(a) Assessment of bias in individual cohorts included in the systematic review and (b) summary of results of the QUADAS-2 assessment in the four domains: patient selection, index test, reference standard, and flow and timing.

In the index test measurement domain, transcriptomic signature scores were interpreted without knowledge of the reference standard (i.e. blinded) in the Turner202013 and Penn-Nicholson2019b45 cohorts, with unclear reporting for the other studies. Due to the early stage of biomarker development and diverse signature measurement and score calculation methodologies, no studies used pre-specified signature score thresholds. The risk of bias in the reference standard domain was deemed to be low in all cohorts with use of appropriate and standardised microbiological confirmatory TB testing (Mtb culture and Xpert MTB/RIF) likely to correctly classify the target condition. Reference standard results were interpreted without knowledge of the results of the transcriptomic signature scores (i.e. blinded) for all included studies. In the study flow and timing domain, all studies used an appropriate interval between index test and reference standard sample collection, and all participants received the same reference standard tests. However, only the Turner202013 study included all participants in analysis. In terms of applicability concerns, selection of participants, and measurement and interpretation of index tests and reference standard matched the review question for all included studies. Overall 9 of the 10 cohorts had a high risk of bias, and one study (Turner202013) had some concerns (Figure 2).

Transcriptomic signature diagnostic performance

Nine independent test or external validation diagnostic cohorts with PLHIV subgroups were included in the systematic review (Table 2). Three cohorts evaluated diagnostic performance of 12 signatures for discriminating HIV-infected adults with prevalent TB disease from latent-Mtb infected individuals or healthy controls with HIV (Figure 3), and 6 cohorts evaluated diagnostic performance of 29 signatures for discriminating HIV-infected adults or children with prevalent TB disease from those with HIV and other respiratory or systemic diseases (Figure 4).

de232003-6657-44ea-b313-10efa7fbfb91_figure3.gif

Figure 3. Transcriptomic signature diagnostic performance for differentiation of patients with microbiologically confirmed TB disease from latent-Mtb infected individuals or healthy controls in independent test or external validation cohorts.

(a) Signature diagnostic performance (area under the curve, AUC) recalculated using participant-level data retrieved from supplementary data or supplied by study authors, with sensitivity and specificity benchmarked against the WHO TPP minimum performance criteria for a triage test (70% specificity and 90% sensitivity, red vertical lines)16. 95% confidence intervals for AUC and proportions (sensitivity and specificity) were calculated using the DeLong30 and binomial proportion (Wilson)31 methods, respectively.

(b) Signature diagnostic performance and 95% confidence intervals as originally reported in studies where participant-level data were not available. *Data for other signatures measured in the full Rajan2018 cohort were extracted from an ATS Conference abstract (Rajan, AJRCCM 2017)47. The number of participants included in the Rajan2018 test set was not specified (ns). #100 TB cases were reported in the paper by Walter et al. (J Clin Microbiol, 2016)39, however only 97 TB cases were included in the original microarray analysis37.

de232003-6657-44ea-b313-10efa7fbfb91_figure4.gif

Figure 4. Transcriptomic signature diagnostic performance for differentiation of patients with microbiologically confirmed TB disease from patients with other diseases in independent test or external validation cohorts.

(a) Signature diagnostic performance (area under the curve, AUC) recalculated using participant-level data retrieved from supplementary data or supplied by study authors, with sensitivity and specificity benchmarked against the WHO TPP minimum performance criteria for a triage test (70% specificity and 90% sensitivity, red vertical lines)16. 95% confidence intervals for AUC and proportions (sensitivity and specificity) were calculated using the DeLong30 and binomial proportion (Wilson)31 methods, respectively.

(b) Signature diagnostic performance and 95% confidence intervals as originally reported in studies where participant-level data were not available. #102 TB cases were reported in the paper by Walter et al. (J Clin Microbiol, 2016)39. however only 97 TB cases were included in the original microarray analysis37. ns, not specified.

Only the Kaforou27 (Kaforou2013a and Rajan2018 cohorts), Penn-Nicholson6 (Penn-Nicholson2020a cohort), and Rajan5 (Rajan2018 cohort) signatures met the WHO TPP minimum performance criteria for a triage test for differentiating adults with prevalent TB disease from latent-Mtb infected individuals or healthy controls with HIV in these case-control studies (Figure 3). However, the upper bounds of the 95% confidence intervals for 4 other signatures met the TPP benchmarks: Darboe11 (Penn-Nicholson2020a cohort), Duffy10 (Kaforou2013a cohort), Kaforou53 (Rajan2018 cohort), and Sweeney3 (Kaforou2013a and Rajan2018 cohorts).

Signatures which are able to distinguish PLHIV with prevalent TB disease from those with other respiratory or systemic diseases are more clinically useful in the inpatient or outpatient setting, as compared to the community. Seven out of 29 signatures met the WHO TPP minimum performance criteria for a triage test in this context (Figure 4): Anderson51 in the Anderson2014b paediatric cohort; Kaforou44 in the Kaforou2013b adult cohort; and Maertzdorf4, Qian17, Roe1, Roe3, and Sweeney3 in the Turner2020 adult cohort. A further 8 signatures (Anderson39.OD, de Araujo1, Kaforou25, Kaforou39, Penn-Nicholson6, Rajan5, Suliman2, and Zak16) had upper 95% confidence interval bounds for sensitivity which met the 90% TPP benchmark (with specificity set at 70%) in the Turner2020 prospective diagnostic accuracy cohort. However, the PLHIV subgroup of this cohort only included 17 TB cases and 27 controls with other respiratory diseases. In other cohorts, the Anderson51 (Anderson2014a cohort), Penn-Nicholson6 (Penn-Nicholson2020b cohort), and Sweeney3 (Kaforou2013b cohort) signatures also approached the WHO minimum TPP triage test criteria. Only one signature (Sweeney3) was tested using a point-of-care device, the Cepheid Xpert-MTB-HR Prototype in the Södersten2020 cohort on biobanked whole blood RNA samples from symptomatic inpatients with HIV. Sweeney3 achieved sensitivity of 91.0% (95%CI 81.8–95.8) and specificity of 55.8% (95%CI 47.2–64.1), falling short of the WHO TPP benchmark. No signatures met the WHO TPP benchmark criteria for a non-sputum confirmatory diagnostic test in PLHIV (98% specificity and 80% sensitivity)16 in any cohorts.

Participant-level data were available for only 3 signatures in the Anderson2014b cohort, which also included 17 children with non-microbiologically confirmed, presumptive clinically-diagnosed TB disease (Table 2). All signatures performed poorly in differentiating TB from other diseases in this subset (Figure 5).

de232003-6657-44ea-b313-10efa7fbfb91_figure5.gif

Figure 5. Transcriptomic signature diagnostic performance for non-microbiologically confirmed presumptive TB disease.

Transcriptomic signature diagnostic performance for differentiation of children with non-microbiologically confirmed (i.e. presumptive clinically-diagnosed) TB disease from children with other diseases in an external validation cohort. Signature diagnostic performance (area under the curve, AUC) recalculated using participant-level data supplied by study authors, with sensitivity and specificity benchmarked against the WHO TPP minimum performance criteria for a triage test (70% specificity and 90% sensitivity, red vertical lines)16. 95% confidence intervals for AUC and proportions (sensitivity and specificity) were calculated using the DeLong30 and binomial proportion (Wilson)31 methods, respectively.

Only the Södersten2020 study reported signature performance stratified by CD4 cell count, with lower Sweeney3 specificity in the 56 inpatients with CD4 cell count less than 200 (66.7%, 95%CI 46.7–82.0) as compared to 129 inpatients with CD4 greater than or equal to 200 (94.7%, 95%CI 88.9–97.5)46. Conversely, sensitivity was higher in the inpatients with CD4 cell count less than 200 (93.8%, 95%CI 79.9–98.3) as compared to the inpatients with CD4 greater than or equal to 200 (75.0%, 95%CI 50.5–89.9). No participant-level data were available to perform further subgroup analyses by CD4 cell count, HIV plasma viral load, TPT or ART status, or other variables in any cohorts.

There was considerable clinical and methodological heterogeneity between cohorts, with participants recruited in diverse settings, with dissimilar eligibility criteria, and distinct composition of the control groups. Studies also used different signature measurement methods (microarray, RNA sequencing, and RT-qPCR), different methods of signature score calculation, and there were no standardised signature score thresholds. Due to the significant heterogeneity in study design and index test measurement, and limited participant data, a meta-analysis was not deemed appropriate.

Transcriptomic signature prognostic performance

Only one study (Darboe2019)42 evaluated transcriptomic signature prognostic performance for recurrent TB disease in adults with HIV who had recently completed TB therapy. There were no eligible studies evaluating performance for prediction of progression to incident TB disease in PLHIV without prior TB disease. We stratified the prognostic performance of the Darboe11 signature by time from measurement until TB disease recurrence (Figure 6). Signature prognostic performance was best within 90 days of signature measurement, and waned thereafter. Sensitivity and specificity of the Darboe11 signature did not meet the minimum WHO TPP performance criteria for a prognostic test (75% sensitivity and 75% specificity)18 in any time window. Sample size was not sufficient to perform subgroup prognostic performance analyses, however Darboe11 signature scores were higher in individuals with detectable plasma HIV viral load (>400 copies/mL) as compared to those with an undetectable plasma viral load (<400 copies/mL; p<0.0001)42.

de232003-6657-44ea-b313-10efa7fbfb91_figure6.gif

Figure 6. Transcriptomic signature prognostic performance for microbiologically confirmed TB disease.

Darboe11 transcriptomic signature prognostic performance for microbiologically confirmed recurrent TB disease in the TB Recurrence upon Treatment with HAART (TRuTH) cohort (Darboe2019)42. Signature prognostic performance (area under the curve, AUC) is stratified by time window following completion of TB treatment: 0–90, 91–180, 181–360, and >360 days. Sensitivity and specificity benchmarked against the WHO TPP minimum performance criteria for a prognostic test (75% sensitivity and 75% specificity, red vertical lines)18.

GRADE evidence summary

A total of 10 cohorts, 2 prospective cohorts (29 TB cases; 51 other respiratory diseases) and 8 case-control studies (353 TB cases; 606 controls), were included in this systematic review of the diagnostic and prognostic accuracy of host blood transcriptomic signatures in PLHIV. All studies used reliable reference standards for definitive TB diagnosis. However, we adjudged that there was a very serious risk of bias due to exclusion of participants with indeterminate (non-microbiologically confirmed) TB in numerous studies, removing diagnostic uncertainty, and resulting in reduced diversity of clinical TB disease. Several case-control studies also included healthy asymptomatic controls, healthy Mtb-sensitised (latent Mtb infected) individuals (IGRA or TST positive), or individuals with other uncommon diseases, not reflective of the target population or setting, further exacerbating the spectrum bias. The inclusion of severe TB cases and healthy controls (or controls with inappropriate other diseases) may have resulted in misleadingly high diagnostic accuracy in some of these studies. Other limitations included uncertainty regarding consecutive recruitment, blinding status not clearly stated, and lack of a priori score thresholds.

Indirectness is synonymous with applicability, generalisability, translatability, and external validity of the evidence36. Included studies evaluated diagnostic performance among adults or children, within clinical outpatient and hospital inpatient settings, prospectively among symptomatic clinic attendees or within matched case-control cohorts. Some of these settings and populations are not appropriate or relevant to clinical practice, and results are unlikely to be generalisable. The lack of diagnostic uncertainty, spectrum bias, and inappropriate control selection is a concern for external validity of these results. Point-of-care device translatability has only been tested for one signature (Sweeney3) among PLHIV, with unsatisfactory diagnostic accuracy46. Technical variability and operator reliability have not been tested on point-of-care platforms for any tests.

With regards to downstream effects, false negative signature results (patients incorrectly classified as not having TB) may have serious consequences, with delayed TB diagnosis resulting in Mtb transmission to close contacts, and increased risk of morbidity and mortality. While consequences are less serious, false positive results (individuals incorrectly classified as having TB) may result in costly further investigations, or 6 months of curative therapy with potential adverse effects and without apparent benefit. Incorrect diagnosis of TB may result in missed or delayed alternate diagnosis and treatment, with potential downstream consequences. Misdiagnosis of TB may also result in stigmatisation from family and community, and psychological distress. There is no uncertainty regarding true positive and true negative results.

We found very serious risks of inconsistency, with significant unexplained heterogeneity of diagnostic sensitivity and specificity estimates for signatures in different validation cohorts and settings. There was also a very serious risk of imprecision, with small sample sizes and wide confidence intervals for estimates of test accuracy among PLHIV, with no pooling of data. Publication bias was also likely, with data only available for a few small studies supporting host blood transcriptomic signatures diagnostics in PLHIV. Data was not available or accessible for the PLHIV subgroup in numerous cohorts. In summary, the data included in this review provides very low quality evidence and we would not recommend any changes to clinical practice based on these results.

Discussion

TB transcriptomic biomarkers selected for advancement through the diagnostics pipeline for further development as point-of-care tests should ideally perform well in PLHIV. We systematically searched online databases for studies which evaluated the performance of host-blood transcriptomic signatures for diagnosing prevalent TB and identifying those who are progressing to incident TB in PLHIV, and compared performance to the WHO TPP criteria. We found 12 studies published prior to 2021 which included 10 independent test or validation cohorts featuring PLHIV, evaluating 31 transcriptomic signatures. Several of the signatures approached or met the WHO TPP minimum performance criteria for a triage test for differentiating people with prevalent TB disease from latent-Mtb infected individuals, healthy controls, or individuals with other respiratory or systemic diseases16. However, no transcriptomic signatures met the TPP benchmark criteria for a non-sputum confirmatory diagnostic test among PLHIV. The signatures also performed poorly for diagnosing non-microbiologically confirmed, presumptive TB disease.

Only one cohort evaluated a signature for predicting TB disease recurrence in individuals who recently completed TB treatment and initiated ART. Prognostic performance appeared to be superior proximally to incident TB disease, with highest AUC and specificity in the 3 months preceding TB recurrence42. The Darboe11 signature did not meet the WHO TPP for a prognostic test in any time window in this population18.

Among the 31 signatures evaluated, we found that the genes most frequently incorporated in TB transcriptomic signatures were ISGs, which may also be upregulated by chronic HIV viraemia62,63. We hypothesised that these signatures would be less discriminatory for TB in PLHIV, particularly among viraemic ART-naïve individuals, due to an increased abundance of circulating type-I IFNs. While no studies performed subgroup analyses of diagnostic accuracy by HIV plasma viral load, Darboe and colleagues42 reported higher signature scores in individuals with a viral load greater than 400 copies per mL, as compared to those with an undetectable viral load (<400 copies/mL). Södersten and colleagues46 demonstrated decreased Sweeney3 specificity in adults with a CD4 cell count less than 200, as compared to those with CD4 cell count greater than 200. Low CD4 cell count is a proxy for ART-naivety and high HIV plasma viral load. The lower specificity is possibly due to higher signature scores in the control group, either due to HIV viraemia, undiagnosed early or minimal TB, or other opportunistic infections. By specifically excluding ISGs, Esmail and colleagues62 have demonstrated that classical complement pathway and Fc-γ receptor 1 (FCGR1) genes are also differentially expressed in individuals with subclinical HIV-associated TB. While traditional discriminant analysis yields an overabundance of ISGs and an underabundance of B- and T-cell genes in active TB patients versus latently Mtb-infected controls, Singhania and colleagues53 have shown that a modular approach (i.e. pre-filtering genes by functional modules) results in a more diverse gene set.

The synthesised systematic review results represent an overall low quality of evidence, with lack of generalisability and external validity, inconsistency in results between studies, and imprecision in estimates. Most of the discovery and validation studies were conducted in Africa, particularly South Africa, limiting geographic diversity and generalisability of results. There were also few training and test datasets including PLHIV, with a notable overreliance on the Kaforou2013 and Anderson2014 datasets for signature discovery and validation, further limiting generalisability. Also, all eligible validation cohorts were from outpatient or inpatient settings. It is notable that signatures generally performed best in small test sets derived from the same population as the signature training cohort (e.g. Rajan5 in Rajan2018 test cohort, Kaforou27 in Kaforou2013a test cohort, Kaforou44 in Kaforou2013b, Anderson51 in Anderson2014b, and Penn-Nicholson6 in Penn-Nicholson2020a cohort), and performance waned in subsequent external validation. Multicohort gene meta-analytical methods, similar to those employed by Sweeney and colleagues49, may help to overcome such limitations.

There is insufficient evidence in the literature to suggest that signatures discovered in cohorts of PLHIV only (e.g. Rajan5) have greater diagnostic accuracy in populations with HIV than signatures discovered among HIV-uninfected (e.g. Darboe11, Penn-Nicholson6, Roe1, and Roe3) or mixed (e.g. Duffy10, Kaforou27, and Sweeney3) cohorts. However, there is a need for larger prospective cohorts, which include PLHIV, for signature discovery and validation, to confirm the utility of transcriptomic signatures among PLHIV. Such cohorts should enrol clearly defined, clinically relevant populations, such as symptomatic clinic attendees, HIV-infected outpatients and other high risk groups; or healthy, asymptomatic individuals from communities in diverse high incidence settings. To reduce spectrum bias, all participants should be enrolled and included in analysis, irrespective of diagnostic uncertainty. Prolonged follow-up over 3 to 6 months may aid diagnosis in unconfirmed cases. Most studies used microarray and RNA sequencing to measure signature scores. Only a few signatures were tested using near-point-of-care benchtop PCR instruments, and one with a point-of-care RT-qPCR device46. Future field evaluation studies should move towards technologies implementable at the point-of-care, such as the Cepheid Xpert-MTB-HR Prototype46,64,65.

Strengths of this systematic review include the comprehensive search strategy, with rigorous eligibility criteria, and publication of a peer-reviewed study protocol. The review also had several limitations. The pre-specified literature search strategy only included studies published prior to 2021; there were few studies with data available for an PLHIV subgroup in test or validation sets in this period, and most included cohorts were small and underpowered. The use of published summary performance data is also problematic due to different statistical methods used in original papers. We were also unable to obtain summary diagnostic performance estimates, or participant-level data to reconstruct two-by-two tables, for PLHIV subgroups from 9 studies, despite contacting corresponding authors. Most papers did not conform to the standard reporting guidelines for diagnostic accuracy studies (STARD)66, with missing information particularly relating to study design, participant recruitment, and blinding to reference standard result. Requisite anonymised participant data with signature scores were only available for a handful of studies. The lack of participant-level metadata precluded subgroup analyses. We were thus unable to systematically determine the effect of HIV viral load, CD4 cell count, TPT, and ART on transcriptomic signature diagnostic performance. Metanalysis was also deemed inappropriate due to the clinical and methodological heterogeneity and high risk of bias in cohort design and signature measurement method, with no predetermined score cut-offs or methods of standardising scores across platforms.

In the two years subsequent to the completion of this systematic literature review, diagnostic and prognostic performance of several transcriptomic signatures measured by qPCR were prospectively tested for mass screening in a South African community setting amongst predominantly asymptomatic PLHIV who were not seeking care21,67. Most transcriptomic signatures measured in this cohort met WHO triage test TPP benchmark criteria among symptomatic participants, however the signatures were upregulated by respiratory viral infection and HIV viraemia, and offered poor specificity for diagnosing sub-clinical TB. The Roe1 signature met the WHO prognostic test TPP benchmarks through 15 months of follow-up among PLHIV, and Darboe11 and Roe3 approached this threshold. These signatures could enable the screening of symptomatic adults seeking care and predict risk of progression to TB disease, thus enabling targeting of preventive therapy68.

Sputum-free TB diagnostics remain an appealing prospect in populations who are unable to produce sputum, such as children and non-ambulant hospitalised patients, and among individuals with paucibacillary or negative sputum samples, such as disseminated TB disease cases or those with advanced HIV6971. Recent studies have evaluated transcriptomic signature diagnostic performance for extrapulmonary TB and TB immune reconstitution inflammatory syndrome (IRIS), both of which are difficult to diagnose, and have high morbidity and mortality among PLHIV. The Sweeney3 and the Penn-Nicholson6 signatures were shown to predict TB immune reconstitution inflammatory syndrome (IRIS) prior to ART initiation and at IRIS diagnosis in children, and paradoxical worsening of TB within half a week of starting ART in adults with TB disease72. The signatures were also able to distinguish IRIS in adults with TB meningitis from HIV-infected patients who did not develop IRIS after starting ART. In this context, transcriptomic signatures could potentially be used to delay ART initiation to prevent paradoxical worsening of TB. While performance is promising, these studies are generally underpowered with clinically inappropriate control groups, suffering from similar flaws in study design as discussed previously. Larger, prospective studies are required to validate transcriptomic signatures for diagnosis of disseminated TB and TB IRIS.

Additionally, most of the cohorts evaluated in this systematic review included adults, with only the Anderson2014 paediatric cohort eligible for inclusion. Paediatric TB is particularly difficult to diagnose due to its paucibacillary nature and difficulty in obtaining sputum samples from small children7376. An accurate non-sputum diagnostic, such as host response transcriptomic signatures, would transform the diagnosis of TB in children77. However, young children are frequent vectors for respiratory and gastrointestinal viruses, and specificity of transcriptomic signatures may be low due to induction of ISG signalling from viral infections.

Transcriptomic signatures show promise for screening for prevalent TB to guide further investigations and predicting progression to incident TB for targeted TB preventive therapy17. However, evidence among PLHIV is limited and mostly from small case-control studies with high risk of spectrum bias. This review emphasises the need for larger heterogenous prospective discovery and validation cohorts exclusively consisting of PLHIV, or mixed cohorts which include PLHIV. Such cohorts should ideally comparatively test biomarker performance side-by-side to determine which signature should be advanced through the developmental pipeline. Further research exploring the effect of HIV clinical, virological, and immunological status is necessary for the design and implementation of TB transcriptomic signatures in this population who are at heightened risk of TB and its sequelae.

Comments on this article Comments (0)

Version 1
VERSION 1 PUBLISHED 03 Feb 2023
Comment
Author details Author details
Competing interests
Grant information
Copyright
Download
 
Export To
metrics
Views Downloads
Gates Open Research - -
PubMed Central
Data from PMC are received and updated monthly.
- -
Citations
CITE
how to cite this article
Mendelsohn SC, Verhage S, Mulenga H et al. Systematic review of diagnostic and prognostic host blood transcriptomic signatures of tuberculosis disease in people living with HIV [version 1; peer review: 1 approved] Gates Open Res 2023, 7:27 (https://doi.org/10.12688/gatesopenres.14327.1)
NOTE: it is important to ensure the information in square brackets after the title is included in all citations of this article.
track
receive updates on this article
Track an article to receive email alerts on any updates to this article.

Open Peer Review

Current Reviewer Status: ?
Key to Reviewer Statuses VIEW
ApprovedThe paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approvedFundamental flaws in the paper seriously undermine the findings and conclusions
Version 1
VERSION 1
PUBLISHED 03 Feb 2023
Views
0
Cite
Reviewer Report 02 Mar 2023
Theresa Ryckman, Bloomberg School of Public Health, Department of Epidemiology, Johns Hopkins University, Baltimore, Maryland, USA 
Approved
VIEWS 0
This study systematically reviewed publications that assessed the performance of host blood transcriptomic signatures in diagnosing active TB and/or predicting risk of progression to active TB among PLHIV. The study found that most publications were at high risk of bias, ... Continue reading
CITE
CITE
HOW TO CITE THIS REPORT
Ryckman T. Reviewer Report For: Systematic review of diagnostic and prognostic host blood transcriptomic signatures of tuberculosis disease in people living with HIV [version 1; peer review: 1 approved]. Gates Open Res 2023, 7:27 (https://doi.org/10.21956/gatesopenres.15638.r32915)
NOTE: it is important to ensure the information in square brackets after the title is included in all citations of this article.

Comments on this article Comments (0)

Version 1
VERSION 1 PUBLISHED 03 Feb 2023
Comment
Alongside their report, reviewers assign a status to the article:
Approved - the paper is scientifically sound in its current form and only minor, if any, improvements are suggested
Approved with reservations - A number of small changes, sometimes more significant revisions are required to address specific details and improve the papers academic merit.
Not approved - fundamental flaws in the paper seriously undermine the findings and conclusions

Are you a Gates-funded researcher?

If you are a previous or current Gates grant holder, sign up for information about developments, publishing and publications from Gates Open Research.

You must provide your first name
You must provide your last name
You must provide a valid email address
You must provide an institution.

Thank you!

We'll keep you updated on any major new updates to Gates Open Research

Sign In
If you've forgotten your password, please enter your email address below and we'll send you instructions on how to reset your password.

The email address should be the one you originally registered with F1000.

Email address not valid, please try again

You registered with F1000 via Google, so we cannot reset your password.

To sign in, please click here.

If you still need help with your Google account password, please click here.

You registered with F1000 via Facebook, so we cannot reset your password.

To sign in, please click here.

If you still need help with your Facebook account password, please click here.

Code not correct, please try again
Email us for further assistance.
Server error, please try again.