Bacterial infections occurring during acute COVID-19 (i.e. detected within the first 48 hr) are relatively rare. Although early reports suggested a high rate of bacterial co-infection at presentation (Langford et al., 2020), most studies report rates between 3% and 9% (Petty et al., 2022; Vaughn et al., 2021; Cong et al., 2022; Cheng et al., 2020; Karaba et al., 2021; Moreno-García et al., 2022). Of these, about half occur in the respiratory tract and are caused by common, community-associated pathogens, including Staphylococcus aureus, Moraxella catarrhalis, Hemophilus influenzae, Streptococcus spp., and Escherichia coli (Petty et al., 2022; Cheng et al., 2020; Calcagno et al., 2021). It remains difficult to discriminate benign colonization, which may occur in more than half of individuals (Kolenda et al., 2020; Sharifipour et al., 2020), from symptomatic bacterial respiratory co-infections, which have been reported in less than 5% of the population (Vaughn et al., 2021; Russell et al., 2021; Karami et al., 2021). Bloodstream infections comprise approximately the other half of acute co-infections and are most commonly caused by S. aureus, coagulase-negative staphylococci, E. coli, Enterococcus faecalis, Acinetobacter spp., and Pseudomonas aeruginosa (Petty et al., 2022; Vaughn et al., 2021; Karaba et al., 2021; Sepulveda et al., 2020; Khatri et al., 2021). Differences in case definitions, isolation methods, geographic variability, and time periods make interpretation of the myriads of studies difficult, but the consensus in the field is that bacterial co-infections during acute COVID-19 are relatively rare.

In contrast, secondary bacterial infections during or following treatment of COVID-19 are more common, with estimates ranging between 10% and 40% depending on the clinical setting (Petty et al., 2022; Cong et al., 2022; Cheng et al., 2020; Ripa et al., 2022; Ripa et al., 2021; Grasselli et al., 2021b; Hedberg et al., 2022; Murgia et al., 2023; Budhiraja et al., 2022). Not surprisingly, rates of secondary pneumonia (frequently ventilator-associated) and bacteremia (frequently catheter-associated) are more common in critically ill patients (Petty et al., 2022; Cong et al., 2022; Cheng et al., 2020; Ripa et al., 2022; Ripa et al., 2021; Hedberg et al., 2022; Murgia et al., 2023; Budhiraja et al., 2022; Grasselli et al., 2021a). The bacterial species most commonly isolated in these patients include S. aureus, K. pneumoniae, Enterobacter spp., P. aeruginosa, Serratia spp., Acinetobacter baumannii, and E. coli (Calcagno et al., 2021; Hedberg et al., 2022; Murgia et al., 2023; Buehler et al., 2021; Ceballos et al., 2022; Chen et al., 2023; Bartoszewicz et al., 2023), reflecting the healthcare-associated nature of these infections. In some locations, secondary infections with multidrug-resistant gram-negative bacteria have also been reported (Ripa et al., 2021; Budhiraja et al., 2022; Bork et al., 2021). Moreover, bacterial bloodstream infections were more common in hospitalized COVID-19 patients (3–4%) (Petty et al., 2022; Khatri et al., 2021; Ripa et al., 2021; Grasselli et al., 2021b; Murgia et al., 2023) than during mild acute infection, likely due to a combination of prolonged hospitalization and immunosuppressive therapy to treat COVID-19 (Khatri et al., 2021; Rothe et al., 2021). There are conflicting reports of whether secondary bacterial infection is more common among COVID-19 patients compared with those infected with other respiratory pathogens (Hedberg et al., 2022; Thelen et al., 2021; Youngs et al., 2020). It remains unclear whether SARS-CoV-2 itself specifically increased the susceptibility to bacterial superinfection, or the high rates reflect instead the severity of infection, use of ventilators, and prolonged hospital course, which are also known to be associated with increased risk of secondary infection.

While initial studies suggested that co-infections do not exacerbate COVID-19 symptoms (Cheng et al., 2020; Grasselli et al., 2021b), most reports demonstrate higher mortality, longer hospital length of stay, and longer duration of ventilatory support in patients with bacterial co-infection (Petty et al., 2022; Ripa et al., 2021; Grasselli et al., 2021b; Murgia et al., 2023; Budhiraja et al., 2022; Buehler et al., 2021). This raises the possibility that bacterial co-infection may predispose to PASC. For example, one would anticipate that longer duration of ventilatory support might increase long-term pulmonary inflammation and subsequent fibrosis, a proposed mechanism for PASC. Indeed, pulmonary infection following recovery from acute COVID-19 is common in survivors (Jakubec and Genzor, 2022) and could reflect either heightened susceptibility to secondary consequences of COVID-19 or a causative role for late-onset infections in mediating PASC. While there is no direct evidence at this time that bacterial infections contribute to the development of PASC, these studies together highlight the need for additional investigations. Specific questions that should be addressed include: (1) Does bacterial co-infection, either at presentation or developing during COVID-19, contribute to the development of PASC? (2) Is secondary bacterial infection a marker for PASC? (3) Can undiagnosed co-infections be found in patients with relatively mild COVID-19 disease who developed PASC? (4) Are secondary infections merely a marker for severe COVID-19 or do they affect the severity of COVID-19? (5) How can we discriminate severe COVID-19 and the postinfection course from post-ICU syndrome sequelae?

M. tuberculosis

Given the global burden of tuberculosis (TB) (over 10 million cases and 1.6 million deaths worldwide in 2021; http://www.who.int/publications/i/item/9789240061729) and the large reservoir of people latently infected with M. tuberculosis (estimated as one-quarter of the world population; Cohen et al., 2019), the question has arisen of whether pre-existing M. tuberculosis infections increase the risk of PASC or other severe manifestations of COVID-19 and, vice versa, whether infection with SARS-CoV-2 may trigger TB reactivation in latently infected individuals. Some evidence exists, mostly based on meta-analyses, that TB (with or without concurrent HIV infection) increases COVID-19 severity and mortality (Boulle et al., 2021; Tamuzi et al., 2020). A recent, multi-site register-based cohort study further supports a deleterious interaction between COVID-19 and TB, as co-infection was found associated with increased mortality, particularly in high-risk subjects, such as elderly males (Group, 2022), who are also highly susceptible to COVID-19 or TB alone (Zhou et al., 2020; Nhamoyebonde and Leslie, 2014).

The observed harmful interactions between these two infections may have a biological basis. For example, immunological measurements have shown that (1) HIV-induced lymphopenia in HIV-TB patients was associated with reduced anti-viral antibody levels and frequencies of SARS-CoV-2-specific CD4+ T cells; (2) active TB was associated with reduced polyfunctionality of SARS-CoV-2-specific CD4+ T cell responses; and (3) acute SARS-CoV-2 infection diminished the pool of M. tuberculosis-specific memory T cell responses (Riou et al., 2021), possibly due to SARS-CoV-2-induced lymphopenia (Shariq et al., 2022). Thus, concurrent M. tuberculosis and SARS-CoV-2 infection can reciprocally reduce the magnitude and possibly the effectiveness of the CD4+ T cell response to the co-infecting pathogen. Other biological mechanisms have been invoked, including the M. tuberculosis-induced upregulation of ACE2 (Shariq et al., 2022), which is a critical SARS-CoV-2 receptor on host cells (Hoffmann et al., 2020). When interpreting epidemiological data, however, it is important to consider that potential bias may be introduced by unmeasured confounding factors (Boulle et al., 2021), particularly because the populations concurrently exposed to these infections may be highly vulnerable due to socioeconomic and geographic factors (Duarte et al., 2021). Moreover, it is widely recognized that the COVID-19 pandemic has had a devastating impact on essential tuberculosis control worldwide (Dheda et al., 2022), resulting in increased tuberculosis morbidity and mortality in 2021, following decades of decline (http://www.who.int/publications/i/item/9789240061729).

Given the clinical spectrum of M. tuberculosis infection (Barry et al., 2009) and the abovementioned enormous burden of asymptomatic infection (latent M. tuberculosis infection [LTBI]), concerns exist that SARS-CoV-2 infection in patients with LTBI may increase the risk of progression to active TB (Group, 2022; Shariq et al., 2022). These concerns are justified, since SARS-CoV-2 infection may induce lymphopenia (Shariq et al., 2022) and, in particular, as mentioned above , reduce the pool of M. tuberculosis-specific CD4+ T cells (Riou et al., 2021), which are central to immune control of M. tuberculosis infection (Flynn and Chan, 2022). Moreover, the use of corticosteroids in COVID-19 treatment may increase the risk of TB reactivation in latently infected patients (Friedman and DeGeorge, 2022). Assessing the association between history of SARS-CoV-2 infection and TB reactivation – at a large scale, and beyond occasional case reports (Khayat et al., 2021; Leonso et al., 2022) – will prove challenging due to the sources of bias mentioned above and the need for large-scale prospective studies.

Whether and how TB might contribute to PASC is even less understood. TB may give rise to chronic lung abnormalities in subjects who have completed therapy, referred to as post-TB lung disease (PTLD) (Allwood et al., 2020). PTLD is poorly recognized and insufficiently studied, despite its relatively high prevalence (lung impairment following pulmonary TB is found in 18–87% of cases, depending on the setting; Allwood et al., 2020) and its association with reduced quality of life. PTLD is highly heterogeneous, as it may include airflow obstruction, restrictive ventilatory defects, and/or impairment in gas exchange, which are in turn associated with various lung immunopathological processes, such as tissue necrosis, cavitation, and aberrant tissue repair leading to fibrosis (Ravimohan et al., 2018). These processes are mediated mostly by proinflammatory and fibrogenic cytokines and chemokines, and tissue-degrading enzymes such as matrix metalloproteinases (Ravimohan et al., 2018). While still speculative, it is quite possible that the TB-induced chronic lung damage may not only increase the risk of severe acute COVID-19 but also favor additional chronic lung damage following SARS-CoV-2 infection. Indeed, precedents exist for respiratory infections, such as aspergillosis, to become chronic due to underlying lung damage by preceding infectious (TB, nontuberculous mycobacterioses) or noninfectious diseases (sarcoidosis, chronic obstructive pulmonary disease) (Kosmidis and Denning, 2015).

Non-tuberculous mycobacteria

Non-tuberculous mycobacteria (NTM) comprise approximately 200 bacterial saprophytic species that are found in the environment, particularly soil and water, and can cause opportunistic infections in the presence of host risk factors, such as old age, immunosuppression, and underlying chronic lung disease (in particular cystic fibrosis or bronchiectasis) (Johansen et al., 2020). Little is known about the association between NTM and COVID-19. Isolated cases of co-infection have been reported (e.g. Rodriguez et al., 2021; Masoumi et al., 2021), and a single-center study in Korea including almost 4000 subjects reported an excess of COVID-19 incidence in subjects treated for NTM (Kim et al., 2022) relative to the general population. However, a National COVID Collaborative Cohort (N3C) study including >6 million subjects failed to observe increased COVID-19 morbidity and mortality associated with pulmonary NTM disease (Figueroa Castro and Hersh, 2021). Even less is known about NTM infection and increased risk of PASC. However, it is plausible that lung damage resulting from severe COVID-19 might increase the risk of subsequent NTM infection.