key: cord-1033933-cy0oxan7
authors: nan
title: Tuberculosis and COVID-19 co-infection: description of the global cohort
date: 2021-11-11
journal: Eur Respir J
DOI: 10.1183/13993003.02538-2021
sha: d4f856bdb8d37cc8c7e525fc0640a8017bbe3c2b
doc_id: 1033933
cord_uid: cy0oxan7

Information on tuberculosis (TB) and COVID-19 is still limited. The aim of this study is to describe the features of the TB/COVID-19 co-infected individuals from a prospective, anonymised, multi-country register-based cohort with special focus on the determinants of mortality and other outcomes. We enrolled all patients of any age with either active TB or previous TB and COVID-19. 172 centres from 34 countries provided individual data on 767 TB-COVID-19 co-infected patients, (>50% population-based). Of 767 patients, 553/747 (74.0%) had TB before COVID-19 (including 234/747 with previous TB), 71/747 (9.5%) had COVID-19 first and 123/747 (16.5%) had both diseases diagnosed within the same week (35, 4.6% on the same day). 85/767 patients died (11.08%) (41/289 (14.2%) in Europe and 44/478 (9.2%) outside Europe; (p=0.03)): 42 (49.4%) from COVID-19, 31 (36.5%) from COVID-19 and TB, 1/85 (1.2%) from TB and 11 from other causes. In the univariate analysis on mortality the following variables reached statistical significance: age, being male, having >1 comorbidity; diabetes mellitus, cardiovascular disease, chronic respiratory disease, chronic renal disease, presence of key symptoms, invasive ventilation and hospitalisation due to COVID-19. The final multivariable logistic regression model included age, male gender, and invasive ventilation as independent contributors to mortality. The data suggests TB and COVID-19 are a “cursed duet” and need immediate attention. TB should be considered a risk factor for severe COVID disease and patients with TB should be prioritised for COVID-19 preventative efforts, including vaccination.

Tuberculosis (TB), with its estimated 10 million cases and 1.3 million deaths annually, continues to be a global health priority [1] .The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) disease (COVID-19) pandemic has required concerted public health focus and action because of its rapid global spread, clinical severity, high mortality rate with 4 million deaths, and capacity to overwhelm healthcare systems [2] [3] [4] [5] .The impact of COVID-19 on TB services has been well described, with a reduction of the number TB cases diagnosed and managed in most countries as a combined result of reduced access, delayed diagnosis with more advanced forms and overstretched health services among other reasons [6] [7] [8] [9] [10] [11] . According to the World Health Organization (WHO) report, there was a 18% decrease of TB case notifications between 2019 and 2020 (from 7.1 to 5.8 million cases) [1] . Conservative models suggest a 20% increase in TB deaths in the next 5 years is likely as a result of the pandemic [12, 13] .

The clinical and immune-pathological interaction between the two diseases and the drivers of dual COVID-19/TB disease mortality are not yet fully understood [14] [15] [16] [17] . A first pilot study of the Global Tuberculosis Network (GTN) on 49 TB/COVID-19 co-infected patients from eight countries was published in 2020 [18] , suggesting that although signs and symptoms are largely the same, TB is frequently diagnosed concomitantly or after COVID-19 and that dual infection may be associated with increased case fatality rate. A second GTN study on 69 TB/COVID-19 patients [10] suggested an overall 12.6% case-fatality rate, higher than the 1-2% mortality rate reported for drug-susceptible TB [1] and for COVID-19 [4] , identifying age and co-morbidities as the main determinants for mortality. Subsequent studies from South Africa and the Philippines suggested that COVID-19 patients with TB have, respectively, a 2.7 [19] and 2.17 [20] higher risk of mortality compared with COVID-19 patients without TB [20] . No large multi-country cohort of TB and COVID-19 patients has been reported to date.

In 2020 the GTN, in collaboration with several organizations (GREPI, Groupe patients. The repository was shared with the WHO to inform the development of global guidance [1, 21] .The aim of this study is to describe the features of the TB and COVID-19 co-infected individuals using this repository, with special focus on the determinants of mortality and other short-term outcomes.

The study is based on a prospective, anonymised, multi-country register-based cohort (Annex 1).

We worked with WHO and the Global TB Network to identify respondents and send invitations to 175 centres in 37 countries [22] . The centres and countries providing data are listed in Annex 2, and Figure 1 ; we enrolled all patients (including children and adolescents) notified to these centres between March 2020 (first case reported on March 12th, 2020) and June 2021. The questionnaire and process was piloted and described previously [10, 18, 23] . We enrolled all patients of any age from these centres with either active TB or previous TB and COVID-19 [18] simultaneously.

The data were obtained via an electronic collection form using variables standardised and harmonised with WHO and piloted in our previous study [18, 21] , including anonymised patients' demographic data, laboratory, radiological, and clinical status at diagnosis of TB and COVID-19, and details on follow-up.

Case definitions follow WHO classification [1] . We define previous TB, as patients who had TB and completed anti-TB treatment at any time in the past before diagnosis of COVID-19. The TB/COVID-19 cases collected in our study were compared with country/regional surveillance systems to estimate coverage in agreement with investigators (Annex 2). All data were cleaned and harmonised throughout the dataset and investigators were contacted in at least two rounds of data cleaning to ensure quality of the dataset before final analysis. The cause of death was analysed as reported by each investigator.

A descriptive analysis was performed on all patients, presenting the details of TB and COVID-19 in the cohort. Considering the relevant proportion of patients from Europe and the number of European countries reporting (15/34) data were also stratified by geographical origin.

We summarised variables using frequencies and percentages and calculated mean and standard deviations (SD) for normally distributed and medians with interquartile ranges (IQR) for nonnormally distributed data. Unpaired t-tests were used to compare continuous variables with normal distributions and categorical variables were compared using chi-squared or Fisher exact test. We used non-parametric tests (e.g., Mann-Whitney U test) for data which could not be converted into a standard distribution.

We were interested in determinants of mortality of COVID-19 and evaluated the effect of prognostic factors on these endpoints by univariable and multivariable logistic regression models.

Covariates which were significant prognostic factors at single variable analysis (p<0.05) were tested for inclusion in the multivariable model in a forward fashion using likelihood ratio tests at each step and used Akaike's information criterion to decide on the final model. For all variables, two-sided pvalues ≤0.05 were considered statistically significant. All variables, when biologically plausible, were tested for interaction. Based on the results of the final multivariable model, we developed a nomogram for risk prediction (Annex 3). The nomogram displays the predicted and confounding probabilities for each variable and overall as points on a scale from 0 to 100 in a user-friendly graphical interface and the overall scale corresponds to the predicted overall probability of the outcome for a patient.

The Ethics Committee of the Maugeri Care and Research Institute, Tradate, Italy (The Coordinating Centre) approved the study on 26 May 2020 (CE 2020/May 26). Each participating centre or country signed a confidentiality and data-sharing agreement with the coordinating centre and obtained local ethics committee clearance or had a waiver indicating no requirement for ethical approval due to the local regulations [18, 23, 24] .

In total, 172 centres from 34 countries provided individual data on 767 TB-COVID-19 co-infected patients (Annex 2). Ascertainment of COVID-19/TB was very high and in most of countries (or regions/states or metropolitan areas, 18/34, 52.9%) more than 80% of these patients were notified to us.

The demographic, epidemiological and clinical characteristics of the 767 TB/COVID-19 patients are summarised in Table 1 .

Most patients were male (70.4%, 540/767), with a median (IQR) age of 44 (31-58) years. The majority were vaccinated with BCG (Bacillus Calmette-Guerin; 90.7%, 349/385). 11.1% (80/717) had a history of migration in the last 5 years and 11.5% (83/724) were HIV co-infected.

Of 767 patients, 553/747 (74.0 %) had TB before COVID-19 (including 234/747 with previous TB), 71/747 (9.5%) had COVID-19 first and 123/747 (16.5%) had both diseases diagnosed within the same week (35 of them (4.6%) on the same day).

As shown in Table 2 , the majority of patients had newly diagnosed TB (618/723, 85.5%) and

bacteriologically-confirmed disease (612/732, 83.6%) with pulmonary localisation (648/755, 85.8%); the majority (517/607, 85.2%) had pan-susceptible TB.

Overall, 248/633 (39.2%) patients presented unilateral or bilateral cavities. About one third of the patients (209/625, 33.4%) performed at least one lung function test, pulse oximetry being the most utilised.

The majority of patients with TB (388 of the 614 with information, 63.2%) were hospitalised during anti-TB treatment for a median (IQR) duration of 31 (14-90) days.

The SARS-CoV-2 laboratory confirmation was available for 723/763 patients (94.8%), the remaining patients diagnosis of COVID-19 was based on clinical and radiological criteria ( Azithromycin, hydroxychloroquine, antiretroviral drugs, corticosteroids and anticoagulants were the drugs most frequently prescribed during the first wave of the epidemic ( Table 3 ).

The number of co-morbidities in the patients who survived and died are summarised in Table 3 and Annex 3. Cardiovascular and endocrine co-morbidities were the most commonly observed, mostly hypertension and diabetes mellitus.

Out of 767 patients in the cohort 85 died (11.08%), 41/289 (14.2%) in Europe and 44/478 (9.2%) outside Europe (P=0.03) ( Table 4) .

Overall, the median (IQR) age of the patients in Europe was higher than outside Europe, 49 years More males were present among those who died vs those who survived (70/85, 82.4% vs 470/682, 68.9%, P=0.01) ( Table 4) .

The comorbidities per patients and geographical location, grouped into main categories, are summarised in Tables 3 and 4 , and Annex 3.

Patients with more than one comorbidity were more frequently observed among those who died In Table 5 the results of the logistic regression analysis to assess the relationship between demographic, epidemiological, clinical variables and mortality are summarised.

In the univariate analysis on mortality the following variables reached the statistical significance: age, being male, having more than one comorbidity, type 2 diabetes mellitus, cardiovascular disease, chronic respiratory disease, chronic renal disease, presence of key symptoms, invasive ventilation and hospitalisation due to COVID-19 (Table 5 ). The final multivariable logistic regression model included age (10 year increase), male gender, and need for invasive ventilation as independent contributors to mortality (Table 5) . Adding other covariates did not significantly increase the performance of the model. A nomogram for the estimation of the risk of death was generated on the basis of the final multivariable model. As depicted in Figure 2 each indicator is measured, and the corresponding points are assigned using the row "score". Thus, the sum is reported on the row "Total Score", and the corresponding probability of death is identified in the row "probability (%) of death".

In the overall cohort, the presence of previous TB was higher among the patients who died than in those who survived (34/85, 40.0% vs 200/682, 30.2%), the difference not being statistically significant; no difference was found between European vs non-European patients (Table 4) .

Patients with active TB had higher probability of death (OR: 1.5) compared with those with previous TB (Table 5) . 

Our study described, for the first time, the features of the TB and COVID-19 co-infected individuals in a large cohort of 767 patients from 172 centres in 34 countries with specific focus on the risk factors for mortality and other outcomes.

The main characteristics of the cohort confirmed our previously described findings from the pilot study [18] : the patients are young (median age: 44 years), the majority male, with drug-susceptible pulmonary TB. The commonest symptoms reported were fever, dry cough, dyspnoea, with about one out of 10 patients with typical symptoms for COVID-19 (olfactory and taste disorders). The majority of patients who underwent CT imaging presented typical or atypical ground glass opacities, confirming the relevance of this radiological sign for the diagnosis of COVID-19 [25] , which co-exist with the radiological features of TB (cavities and infiltrates).

Interestingly, 74% of the patients had TB diagnosed before COVID-19 (including 234 patients with previous TB, corresponding to 31.3% of the whole cohort), 16.5% were diagnosed within the same week (the presence of signs and symptoms suggested the clinicians to perform imaging, which revealed a potentially pre-existing TB on top of COVID-19) [18] and 9.5% had COVID-19 diagnosed first.

A key question from our preliminary study [18] was on the role of SARS-CoV-2 on the progression of TB infection to disease as observed in other viral diseases (e.g. HIV) [5, 18] . Whilst our study is not specifically designed to answer this question, we found 71 patients who had COVID-19 diagnosed before TB: of these 35 were diagnosed more than 30 days prior (with a sufficient time to develop TB disease) and 33 had pulmonary TB. Of 25 patients with complete radiological information, 12 (48%) had cavities, a condition which is likely to develop in more than 30 days. Therefore, this indirect evidence from our data suggests that COVID-19 may not have a major role in advancing TB infection to TB disease. Further longitudinal studies observing the patients with TB infection and COVID-19 over time and comparing the proportion of those who acquire TB disease with a control group without COVID-19 may offer better insight to an interaction.

The TB/COVID-19 patients with higher mortality are males, belong to older age-groups and underwent invasive ventilation, with more comorbidities than those with no need for (invasive) ventilation. These determinants of death are similar to those described for mono-disease COVID -19 or TB [4, 26] .

Another important question arising from previous studies [10, 14, 18, 27] relates to the resources required for managing patients with TB and COVID-19. The study results indicate that an important proportion of patients needed ventilation (18%. of whom 7.4% required intubation) and 32% supplemental oxygen, the vast majority during hospitalisation (61.7% of the patients required a median of 14 days of admission because of COVID-19, in addition to those needed for TB). The need for competent staff to manage TB/COVID patients with respiratory failure has been a problem in several countries, where clinicians working within the TB programme were re-deployed to work within the COVID-19 emergency [6] [7] [8] [9] 14] . Evidence is continuing to emerge on the negative impact of COVID-19 on TB services [9, 28] . A recent global study indicates a significant decline in TB and TB infections diagnosed, with an increase of tele-medicine use in 2020 in comparison with 2019 [9] . Reduction in the performance of global TB detection and care due to COVID-19 pandemic are expected to have devastating impact on TB mortality [29] .

An issue which recently gained increasing interest is that of Post-TB Lung Disease (PTLD), as 13-68% of new TB cases and 75-96% of patients with MDR-TB completing anti-TB treatment suffer from TB sequelae [30] [31] . This condition [30, 31, [33] [34] [35] [36] includes obstructive, restrictive or mixedpattern lung function abnormalities, reduced exercise capacity and impaired Quality of Life (QoL).

A summary of clinical standards to adequately manage PTLD, which includes post-treatment evaluation and identification of patients with sequelae likely to benefit from pulmonary rehabilitation has recently been published [36] .

Similarly, COVID-19 appears to commonly cause sequelae (the so-called "long-COVID" syndrome) [37] [38] [39] , characterised by fatigue, sleeping difficulties, low grade fever, depression, anxiety, impacting cardiac, pulmonary and renal functions and discussion is ongoing on the potential role of post-COVID-19 rehabilitation [14, [40] [41] [42] .

A combination of post-COVID-19 and PTLD sequelae and the need for assessment and potential follow-up and rehabilitation can pose additional stress on health services in terms of human and economic resources.

Our study has several strengths, including a large sample size and the inclusion of countries from all continents. Furthermore, several variables collected in our study are not routinely collected in the surveillance systems at country level, making the study important to better understand the TB/COVID-19 interactions and to design 'ad hoc' studies aimed at answering specific outstanding questions. Furthermore, about half of the countries/ territories (18/34) provided population-based data representative of their respective TB/COVID patients.

Among the main study limitations, Africa and Asia were under-represented, the number of paediatric patients was limited (6 patients, two of them below 1 year of age), some centres were unable to provide all the information requested on a few variables (particularly laboratory data) and about 10% of the patients had COVID-19 diagnosed based on the clinical and radiology findings, following the respective countries' policy during the emergency phases of the pandemic. The timing of our study also does not allow comment on the differential impact of emerging SARS-CoV-2 variants and TB, which will require ongoing monitoring and review.

In addition, as the cohort was composed of TB and COVID-19 patients, it was not possible undertake a comparative analysis against patients with TB or COVID-19 alone. It was also not possible to draw conclusions on the effect of the different drugs prescribed, and we note that our cohort was prescribed a range of therapies by treating clinicians, including some now demonstrated to have no impact on COVID outcomes. Future studies looking at the cohort will be able to examine the effect of steroids or monoclonal antibodies. Furthermore, it was not possible to perform the analysis of TB-specific outcomes as an important proportion of patients are still undergoing anti-TB treatment.

The study will continue to evaluate early and final anti-TB treatment outcomes through periodic updates, as to make the 'cohort' a 'living' one.

This first description of a large global cohort provides important information for clinical and public health management of patients co-infected by TB and COVID-19. The similarity of signs and symptoms for the two diseases has been confirmed alongside the importance of the radiological presence of ground glass opacities for the diagnosis of COVID-19. Preliminary information seems to suggest that COVID-19 is unlikely to represent a major determinant triggering TB infection to active TB.

The high (12%) mortality of co-infected patients may be explained by older age, male gender, with an important contribution also played by co-morbidities (particularly cardiovascular disease and diabetes mellitus). The reason why males died more than females may be explained by the potential higher prevalence of co-morbidities and risk factors.

Efforts to prevent SARS-CoV-2 infection in TB patients is warranted, including reinforcing of social distancing, mask wearing and other measures as appropriate to local epidemiology.

Encouraging vaccination against SARS-CoV-2 for people with a current or past diagnosis of TB will also be valuable in preventing morbidity and mortality related to COVID-19 disease.

The combination of COVID-19 and TB adds to the clinical complexity in patients' management (e.g. need for supplemental oxygen, invasive or non-invasive ventilation and specialized staff) significantly impacting health services. The impact of COVID-19 on long-term pulmonary sequelae in patients with TB and the need for pulmonary rehabilitation is yet to be determined.

As patients reported similar symptoms, it advisable for health services to screen patients for both diseases whenever possible, taking advantage of the possibility to rapidly obtain imaging, and stimulating adoption of rapid molecular testing for TB and COVID-19. Although our study does not provide specific data on this, it seems clinically advisable to treat both conditions as soon as possible following international recommendations.

Last but not least, the experience gained during the COVID-19 pandemic will allow us to make better use of telemedicine interventions, thus reducing the burden of physical access to health services and transmission. Unnecessary hospitalisation should be actively discouraged [7, 9, 27] .

Overall, the data suggests TB and COVID-19 are a 'cursed duet' and need immediate attention. 

The nomogram for the estimation of the risk of death was generated on the basis of the multivariable logistic regression analysis. As depicted in the figure, each indicator is measured, and the corresponding points are assigned using the row "Score". Thus, the sum is reported on the row "Total Score", and the corresponding probability of the outcome is identified in the row "Probability (%) of death". As an example on how to use this nomogram, an 80-year-old woman not requiring invasive ventilation, would have a probability of death < 20%. In contrast, 80-year-old woman requiring invasive ventilation during hospitalization would have a probability >80%. 

World Health Organization. Global tuberculosis report 2021. Geneva: World Health Organization

Date last accessed

Clinical Characteristics of Coronavirus Disease 2019 in China

Clinical findings in a group of patients infected with the 2019 novel coronavirus (SARS-Cov-2) outside of Wuhan, China: retrospective case series

Clinical course and risk factors for mortality of adult inpatients with COVID-19 in Wuhan, China: a retrospective cohort study

Epidemic and pandemic viral infections: impact on tuberculosis and the lung. A consensus by the World Association for Infectious Diseases and Immunological Disorders (WAidid), Global Tuberculosis Network (GTN) and members of ESCMID Study Group for Mycobacterial Infections (ESGMYC)

Worldwide Effects of Coronavirus Disease Pandemic on Tuberculosis Services

The impact of COVID-19 on TB: a review of the data

The potential impact of COVID-19-related disruption on tuberculosis burden

Gauging the impact of the COVID-19 pandemic on tuberculosis services: a global study

Tuberculosis, COVID-19 and migrants: Preliminary analysis of deaths occurring in 69 patients from two cohorts

Time to regain lost ground: Tuberculosis in the COVID-19 era

Potential impact of the COVID-19 pandemic on HIV, tuberculosis, and malaria in low-income and middle-income countries: a modelling study

The potential impact of the COVID-19 pandemic on the tuberculosis epidemic a modelling analysis

Tuberculosis and COVID-19 interaction: A review of biological, clinical and public health effects

Clinical characteristics of COVID-19 and active tuberculosis co-infection in an Italian reference hospital

Coinfection of tuberculosis and COVID-19 limits the ability to in vitro respond to SARS-CoV-2

Relationship of SARS-CoV-2-specific CD4 response to COVID-19 severity and impact of HIV-1 and tuberculosis coinfection

Active tuberculosis, sequelae and COVID-19 co-infection: first cohort of 49 cases

Risk factors for COVID-19 death in a population cohort study from the Western Cape Province, South Africa

Previous and active tuberculosis increases risk of death and prolongs recovery in patients with COVID-19

World Health Organization. The WHO Global Clinical Platform for COVID-19

The TB/COVID-19 Global Study Group. TB and COVID-19 co-infection: rationale and aims of a global study

Surveillance of adverse events in the treatment of drug-resistant tuberculosis: first global report

Effectiveness and safety of bedaquiline-containing regimens in the treatment of MDR-and XDR-TB: a multicentre study

Coronavirus Disease 2019: Initial Detection on Chest CT in a Retrospective Multicenter Study of 103 Chinese Patients

Different disease, same challenges: Social determinants of tuberculosis and COVID-19

Boom M; contributing members of the Global Tuberculosis Network. Tuberculosis, COVID-19 and hospital admission: Consensus on pros and cons based on a review of the evidence

COVID-19 related hospital re-organization and trends in tuberculosis diagnosis and admissions: reflections from Portugal

Predicted impact of the COVID-19 pandemic on global tuberculosis deaths in 2020

The need for pulmonary rehabilitation following tuberculosis treatment

Pulmonary rehabilitation is effective in patients with tuberculosis pulmonary sequelae

Persistent chronic respiratory symptoms despite TB cure is poorly correlated with lung function

Post-tuberculosis lung health: perspectives from the First International Symposium

Functional impact of sequelae in drug-susceptible and multidrug-resistant tuberculosis

Is there a rationale for pulmonary rehabilitation following successful chemotherapy for tuberculosis?

Clinical standards for the assessment, management, and rehabilitation of post-TB lung disease

Pulmonary Rehabilitation in Patients Recovering from COVID-19

Functional impairment during post-acute COVID-19 phase: Preliminary finding in 56 patients

The Role of Blood Gas Analysis in the Post-Acute Phase of COVID-19 Pneumonia

Time course of exercise capacity in patients recovering from COVID-19-associated pneumonia. Authors' reply

Post-COVID rehabilitation

Post-COVID sequelae

Extra-Pulmonary TB, n (%)

Median (IQR) sO2, % (n=214)

Median (IQR) FiO2, % (n=112)

Median (IQR) pO2, mmHg (n=40) 77.9 (65.7-93.8) Mean (SD) pCO2, mmHg (n=40)

Median (IQR) sO2, % (n=397)

Median (IQR) FiO2, %(n=269)

Median (IQR) pO2, mmHg (n=99)

Median (IQR) pCO2, mmHg (n=100)

Median (IQR) duration of hospitalization, days (n=395)

Concomitant hospitalization due to TB-COVID-19 co-infection

Median (IQR) duration of concomitant hospitalization, days

Median (IQR) from start treatment to PCR conversion, days (n=196) 14

Immunomodulators Glucocorticoids (methylprednisolone, betamethasone, ciclesonide , other glucocorticoids), n (%)

Chronic Renal Disease, n (%)

Alive: all patients in the cohort based on the latest information available (see text and Figure 3 for details)

The detailed causes of death are reported in the text and in Figure 3; TB: tuberculosis; COVID-19: acute respiratory syndrome coronavirus 2 (SARS-CoV-2)

*: including pts with PCR not done; ^: 2 patients with symptoms resolution remain hospitalized for TB **2 for multiple comorbidities, 1 for suspected cancer, 1 for sarcoidosis

2 for sepsis, 2 for multiple comorbidities, 1 for bilateral gram negative nosocomial pneumonia, M. morgany, 1 for pulmonary thrombo-embolism [with COVID-19 clinically diagnosed and PCR unknown]) PCR: polymerase chain reaction