Imaging and Circulating Biomarker-Defined Cardiac Pathology in Pulmonary Tuberculosis: A Systematic Review

Data sources, search strategy, and eligibility criteria

MEDLINE, Embase, Global Health, Web of Science, and Google Scholar were searched for studies evaluating cardiac pathology in patients with PTB using cardiac imaging techniques and/or biochemical markers of myocardial injury and remodelling.

The search strategy was performed using keywords for pulmonary tuberculosis (PTB), the imaging modalities, and cardiac biomarkers of injury and remodelling.

Our search terms included ‘pulmonary tuberculosis’ and related terms, ‘cardiac imaging’ including ‘echocardiography,’ ‘computed tomography coronary angiography,’ ‘cardiac magnetic resonance imaging,’ ‘positron-emission tomography,’ and ‘cardiac biomarkers’—see Appendix 1 – Search Strategy for full list of search criteria.

We included studies indexed from 1960 to October 31, 2023, and no restrictions were placed on language or study site.

Studies that used cardiac imaging and/or circulating biomarkers of cardiac injury and/or remodelling among patients with PTB were included. The primary exposure was bacteriological (microscopy, culture, or molecular diagnostics), radiological (chest x-ray or computed tomography), and/or clinical (according to internationally or locally accepted diagnostic algorithms) diagnosis of PTB.

Study selection

Two reviewers (MSS and KJH) screened titles and abstracts of retrieved studies; independently assessed their eligibility according to inclusion criteria; and resolved conflicts by consensus. A Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) chart was created (Figure 1). Data were extracted from studies by MSS and KJH and reviewed by ASVS.

Figure 1

Outcomes

The main outcome was the proportion of patients with PTB who had evidence of cardiac pathology identified through cardiac imaging and/or biochemical markers of cardiac injury and remodelling.

Cardiac pathology was defined as: presence of pericardial pathology (pericardial effusion >0.5 cm) or left ventricular systolic dysfunction on echocardiography; myocardial fibrosis based on late gadolinium enhancement using CMR; coronary stenosis on CTCA; vascular and/or myocardial inflammation on PET; and myocardial injury based on circulating levels of cTn and BNP.

Risk of bias

Risk of bias was independently assessed by two reviewers (MSS and KJH) using the National Heart, Lung and Blood Institute (NHLBI) quality assessment tools (19) for observational studies including cross-sectional, cohort, randomised controlled trial, and case series designs. These tools comprise 14 questions that assess the quality and risk of bias of studies and suggests inclusion, exclusion, or seeking further information prior to inclusion in the systematic review. Risk of bias was assessed primarily based on quality of study design.

Study characteristics

We identified 4,989 studies after deduplication, of which seven met specified inclusion criteria (Figure 1 and Table 1) (20, 21, 22, 23, 24, 25, 26). A total of four studies evaluated echocardiographic findings of cardiac pathology: three studies (20, 21, 22) evaluated 533 participants using echocardiography alone; the other study (23) evaluated cardiac Troponin T (cTnT), Creatine Kinase-MB (CK-MB), and NT-pro-BNP alongside echocardiography in 800 participants. Two studies evaluated PET findings of vascular inflammation: one study (25) evaluated 358 patients with both pulmonary and extrapulmonary TB (PTB and EPTB), of whom 100 had pulmonary TB; the other study (24) evaluated 96 patients established on anti-tuberculous therapy. One study involving 26 patients described findings of lung and cardiac inflammation on PET and cardiac function on CMR (26). No study evaluated CTCA or transoesophageal echocardiography.

Human immunodeficiency virus status was documented in six (20, 21, 22, 23, 25, 26) of seven studies: one study (22) exclusively recruited People living with HIV (PLHIV); three studies (20, 21, 26) recruited both PLHIV and HIV-negative participants; one study (25) exclusively recruited participants with a documented negative HIV test; and one study (23) excluded PLHIV but did not provide HIV testing data. In the remaining study (24), participants’ HIV status was not reported.

Of the seven studies, four originated from high-TB burden countries (21, 22, 23, 26) (one from South-East Asia, and three from sub-Saharan Africa), two from low-TB burden countries (20, 24), and one (25) was a multi-country study with both high- and low-TB burden countries—see Figure 2.

Figure 2

The study designs used were cross-sectional (5/7, 71.4%) and cohort studies (2/7, 28.6%). Date of publication ranged from 2000 to 2023, and sample size ranged from 26 to 800 participants.

Risk of bias assessment

Of the seven studies, three (42.9%) were classified as low (22, 25, 26); two medium (20, 21) (28.6%); and two (23, 24) (28.6%), high risk of bias. Of the included studies, three conducted statistical adjustment for confounders, none had adequate sample size justification, and one had a control population. Risk of bias assessment is described in Appendix 2.

Population characteristics

A total of 1,333 participants, with PTB with a mean age ranging from 33 to 39.7 years of whom 243 (14.1%) were PLHIV and 56.3% were male, were enrolled across the studies (see Table 2 – Population Characteristics). Time since PTB diagnosis was available for all patients who were enrolled at the time of or shortly after PTB diagnosis.

Pericardial pathology in PTB

Three studies reported echocardiographic evidence of pericardial effusion in patients with PTB, with prevalence ranging from 14.1 to 55.6%. Across these studies, we made several observations. First, microbiological confirmation of tubercular effusion was not attempted and diagnosis of PTB-associated pericardial effusion was inferred from bacteriological confirmation in sputum. Importantly, one study reported a significant association between pericardial effusion and subsequent culture confirmation of TB in sputum (21). Second, all three studies were limited by sample size and cross-sectional designs. Third, the majority of participants were PLHIV, with only 4/131 patients with pericardial effusion and PTB who had a negative HIV test. Our evidence synthesis may therefore not reflect the true prevalence of pericardial effusion in patients with PTB—particularly in those who are HIV-negative—and it remains unclear whether PTB is associated with higher frequency of dissemination to the pericardium independent of co-existent HIV infection. Prevalence of pericardial effusions in other studies is restricted to populations with known cardiac pathology and TB disease. In these populations, and especially in those populations studied prior to the introduction of HIV antiretroviral therapy (ART), the reported prevalence of pericardial pathologies is up to 85% (38, 39, 40, 41). Studies are now underway to systematically determine the prevalence and natural history of cardiac pathology, using cardiac ultrasound and biochemical markers, in patients with newly-diagnosed PTB, with and without HIV (42).

Tuberculosis is thought to contribute up to 50% of pericardial disease burden in LMICs (8) but systematic evaluation in the TB population remains sparse. Whereas bacterial pericarditis is almost uniformly fatal, TB pericarditis presents indolently with progressive accumulation of inflammatory serosal fluid and progressive calcification of inflamed pericardium (43). Despite its progressive symptomatology, TB pericarditis carries an up to 40% mortality at six months (43, 44). The Investigation of the Management of Pericarditis (IMPI) trial (45) randomised patients with tuberculous pericarditis to adjuvant steroid therapy. Prednisolone administration resulted in a 40% reduction in development of constrictive pericarditis and hospitalization but a three-fold increase in HIV-associated cancer. Importantly, two-thirds of the study population were PLHIV and only 14% established on ART therapy.

Myocardial pathology in PTB

Myocardial TB contributes <0.1% to incident TB cases, is associated with premature death from fatal arrhythmias, and diagnosed post-mortem in the few existing case studies (8, 46). However, our study highlights the significant paucity of data to make firm conclusions on the interaction between PTB and myocardial pathology with data originating from a single study investigating an Indian population (23). One in 20 patients had depressed systolic function which was associated with biochemical evidence of cardiac injury. Importantly, in 20% of patients, left ventricular systolic dysfunction either persisted or worsened over time (23). Mechanistic imaging data using advanced modalities also remains limited. We identified one study that evaluated 26 patients with PTB using CMR and late gadolinium enhancement but did not report on the prevalence of myocardial fibrosis (26). Two further studies evaluated inflammatory cellular infiltration with PET in PTB (24, 25) with a reported prevalence of perimyocardial inflammation at 0.6% and 21.8%. Myocardial TB may also be an unrecognised and important cause of significant cardiac electrical disturbance. A recent case series showed that of 13 patients presenting with unexplained ventricular tachycardia in India, 11 were found to have TB disease (47). Importantly, treatment with anti-TB medications and steroids resulted in improvement of left ventricular function and rhythm stability (47). Together, these findings further underline the need for systematic multi-modality imaging studies to delineate causal pathways in PTB-induced cardiac pathology at diagnosis, its potential impact on outcomes at TB treatment completion, and potential relationship with CVD burden in TB-endemic settings.

Clinical correlations

Our evidence synthesis demonstrates the paucity of systematic evidence guiding healthcare providers to screen, diagnose, and manage cardiac involvement in PTB. Current data and clinical guidelines are overwhelmingly focused on a narrow definition of cardiac tuberculosis, overlooking clinically relevant cardiovascular complications that occur in patients with PTB.

Our review highlights significant heterogeneity in the available evidence describing cardiac pathology in PTB. It is plausible that the burden of cardiovascular pathology is underappreciated in PTB, and significant research gaps remain in evaluating the burden of cardiovascular pathology in PTB. Our review provides critical evidence to guide the next phase of research to understand the burden and mechanisms of TB-associated CVD.