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

To the Editor:

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) disease (COVID-19) pandemic has attracted interest because of its global rapid spread, clinical severity, high mortality rate and capacity to overwhelm healthcare systems [1, 2]. SARS-CoV-2 transmission occurs mainly through droplets, although surface contamination contributes and debate continues on aerosol transmission [3–5].

The disease is usually characterised by initial signs and symptoms [4–9] similar to those of related viral infections (e.g. influenza, SARS, Middle East respiratory syndrome) and tuberculosis (TB), although prognosis and complications sometimes differ. Experience with concomitant TB and COVID-19 is extremely limited.

One case–control study of COVID-19 patients with interferon-γ release assay-confirmed TB infection [10] and a single case of TB with COVID-19 have been submitted to, but not yet published in, peer-reviewed journals [11]. In a recent analysis of 1217 consecutive respiratory specimens collected from COVID-19 patients (Mycobacterium tuberculosis was not tested), the authors concluded that higher rates of co-infection between SARS-CoV-2 and other respiratory pathogens can be expected [12].

The present study describes the first-ever global cohort of current or former TB patients (post-TB treatment sequelae) with COVID-19, recruited by the Global Tuberculosis Network (GTN) in eight countries and three continents. No analysis for determinants of outcome was attempted.

The study is nested within the GTN project monitoring adverse drug reactions [13, 14] for which the coordinating centre has an ethics committee approval, alongside ethics clearance from participating centres according to respective national regulation [13, 14]. A specific nested database was created in collaboration with the eight countries reporting patients with TB and COVID-19; the remaining countries had not yet observed COVID-19 in their patients at the time this manuscript was written.

Continuous variables, if not otherwise specified, are presented as medians with interquartile ranges.

Overall, 49 consecutive patients with current or former TB and COVID-19 from 26 centres in Belgium (n=1), Brazil (Porto Alegre, Rio Grande do Sul State; n=1), France (n=12), Italy (n=17), Russia (Moscow Region; n=6), Singapore (n=1), Spain (n=10) and Switzerland (Vaud Canton; n=1) were recruited (dataset updated as of 25 April, 2020) (table 1).

The first onset of COVID-19 in the cohort was observed in an Italian patient with TB sequelae on 12 March, 2020 (symptoms from 6 March).

Of 49 patients, 26 (53.0%) had TB before COVID-19, 14 (28.5%) had COVID-19 first and nine (18.3%) had both diseases diagnosed within the same week (four on the same day).

42 (85.7%) patients had active TB (median age 45.5 (28.0–63.0) years) and seven (14.3%) had post-TB treatment sequelae (median age 69.0 (66.0–70.0) years; p=0.01); the patients with TB sequelae (from five centres in Italy, Singapore, Spain and Switzerland) were cured 8.2 (2.7–44.3) years earlier.

Overall, 26/49 (53.1%) patients were migrants, 15/48 (31.3%) unemployed, and 2/48 (4.1%) healthcare workers (medical doctor and radiology technician).

46 (93.9%) patients had confirmed SARS-CoV-2 infection and three other patients (6.1%) had chest high-resolution computed tomography (HRCT) highly suggestive of COVID-19 related pneumonia (bilateral ground glass opacities) [15].

48 patients had pulmonary TB (one caused by Mycobacterium bovis).

37 patients had drug-susceptible (or were treated with the standard first-line regimen for new cases) and eight had drug-resistant TB (and were treated with second-line drugs).

Of the 14 non-clustered patients with COVID-19 diagnosis preceding TB, a child of Gambian origin (3 months old) had SARS-CoV-2 identified 3 days before TB diagnosis although TB was probably pre-existing (pulmonary and extra-pulmonary TB, meningitis). The child is continuing anti-TB treatment and has now recovered from COVID-19. Altogether the diagnosis of COVID-19 preceded that of TB by a median (range) time of 4 (2–10) days.

Those 14 patients, managed in nine centres (in France, Italy, Russia, Spain) were young (median age 33 (26.0–46.0) years); 11/14 (78.5%) were migrants.

Of the 19 patients undergoing anti-TB treatment, the diagnosis of COVID-19 was made during month 1–2 for 10 (52.6%), month 3–4 for three (15.7%), month 4–6 for three (15.7%) and after 6 months for three (15.7%) patients.

Signs and symptoms attributed to COVID-19 included (in different combinations) fever (32/48, 81.2%), dry cough (27/48, 56.2%) and dyspnoea (17/48, 35.4%).

Radiological information was available for 48/49 (98.0%) patients: 23 (47.9%) presented cavities (table 1).

21 patients manifested a typical HRCT COVID-19 pattern (bilateral ground glass opacities), whereas 23 had different patterns at HRCT or chest radiography essentially reporting TB-related lesions (infiltrates, consolidations, cavities, etc.) and five were not studied during the course of COVID-19 disease.

43 (87.8%) patients were hospitalised and, provisionally, the overall median (interquartile range) number of hospital admission days was 15 (8–27). Six patients needed noninvasive ventilation and 14 oxygen supply.

Medication for COVID-19 was reported for 28 patients in all countries except Belgium and Brazil (which had a single case each): 22 (78.6%) received, in different combinations, hydroxychloroquine, 12 (42.9%) received an anti-HIV protease inhibitor (i.e. lopinavir/ritonavir, darunavir/cobicistat), seven (25.0%) patients received azithromycin, and one (3.6%) patient received other drugs (i.e. enoxaparin and N-acetyl-cysteine (NAC)). 17 (60.8%) patients received monotherapy (i.e. 11 hydroxychloroquine, five lopinavir/ritonavir and one azithromycin), nine (32.1%) received a combination of two drugs (i.e. five hydroxychloroquine and protease inhibitor; four hydroxychloroquine plus azithromycin), two (7.1%) received three or more drugs (i.e. one hydroxychloroquine, lopinavir/ritonavir and azithromycin; one hydroxychloroquine, lopinavir/ritonavir and azithromycin, enoxaparine and NAC).

The case fatality rate was high (6/49, 12.3%); five of these six were >60 years old and all of them had at least one comorbidity. Given the small number of deaths, larger studies are necessary.

This preliminary analysis suggests that:

• 1) In 19/49 (38.8%) patients, COVID-19 appeared during anti-TB treatment and limited or no protection against COVID-19 might have favoured SARS-CoV-2 infection (which affected two healthcare workers).

• 2) The diagnosis of TB and COVID-19 was done simultaneously or within 7 days in nine patients, posing differential diagnosis challenges, suggesting that clinical assessments to investigate COVID-19 (e.g. clinical picture, HRCT) facilitated the identification of (a probably pre-existing) TB. Any contribution of COVID-19 to TB pathogenesis cannot be excluded or confirmed.

• 3) Although the diagnosis of COVID-19 preceded that of TB in 14 patients, larger studies are needed to understand any role played by SARS-CoV-2 in the progression of TB infection to disease. Given that up to a quarter of the population in some of the regions studied are latently infected, SARS-CoV-2 infection might boost the development of active TB in the coming months [10]. As we do not include individuals with latent TB infection followed up over time, we cannot report on the potential contribution of COVID-19 towards development of active disease. Probably, an overlap of signs/symptoms of COVID-19 and TB occurred, and COVID-19 was diagnosed earlier because of a higher index of suspicion, while TB may have been previously present. Or, differently, COVID brought to clinical evaluation/diagnostic assessment TB patients at an earlier stage of disease, before the occurrence of TB-related symptoms.

• 4) In seven cases, COVID-19 occurred in patients with TB sequelae. They were older than patients under anti-TB treatment and presented higher (although not statistically significant) mortality. All but one had comorbidities (four COPD; one HIV co-infection plus liver and kidney diseases, hypertension and cancer present in different combinations). They presented unilateral or bilateral radiological sequelae of previous infiltrates (four patients) or cavities (three patients). Larger numbers are necessary to further understand the role played by TB sequelae.

• 5) The impact on the healthcare system (days of admission, intensive care unit beds, etc.) was relevant, and will deserve further evaluation.

• 6) The information on BCG (bacillus Calmette–Guérin) vaccination is modest (30 patients with information, 19 previously vaccinated in all eight countries) and no significant elements can be provided to the ongoing debate on its protective role.

• 7) We presently have no data on drug–drug interactions.

This is, to our knowledge, the first published cohort of patients with active TB and COVID-19. Our study represents a “snapshot” of a cohort of patients at different stages of disease. No attempt was made to obtain representation of the larger universe of patients with both diseases and the small sample size precludes an analysis of risk factors. We cannot exclude that some findings have a casual origin.

The information available on the patients recently admitted was accurate, but some details on previous TB were incomplete, and some examinations were not performed either because the patients refused or the patients’ condition was too severe. Although case reporting is comprehensive in the countries/regions covered by GTN, the study reflects the initial stages of the COVID-19 epidemic only, and representative longitudinal observations will be necessary to evaluate the interactions between COVID-19 and TB.

We believe that this descriptive research can motivate larger studies to enable analyses of interactions and determinants of outcomes in patients with both diseases. The study will continue to follow-up the patients and accrue more records. We therefore invite interested clinicians and programmes to contact the corresponding author and help improve the understanding of how to optimise care for these patients.

Shareable PDF