Introduction

Autoimmune diseases are characterized by intrinsic immune alterations, which may lead to chronic inflammation in multiple organ systems. The diseases themselves and many of their treatments are associated with increased risk for severe infections [1,2,3,4]. Detailed research on the pathogenesis of these diseases has led to the development of targeted therapies, aimed mainly at cytokines and cells, which play key roles in the chronic inflammatory process.

During the COVID-19 pandemic, the risk and outcomes of patients with autoimmune diseases have become a cause of concern and require proper investigation. On the one hand, this population is, by virtue of the disease and its treatment, potentially vulnerable to suffer severely from COVID-19; on the other hand, many of these patients have been and continue to be exposed to immune modifying therapies which may have the potential to treat some features and improve outcomes of COVID-19 itself.

The initial link came from observations indicating that severe COVID-19 infection, manifested by acute respiratory distress syndrome (ARDS) and/or multi-organ failure, is associated in a proportion of patients with a “cytokine storm” [5,6,7,8]. It follows then that certain cytokine-blocking agents may be beneficial in this scenario [9, 10] and several of them are presently being tested in clinical trials internationally. In the meantime, data on the risks and outcomes of COVID-19 infection in patients with autoimmune diseases who are already receiving such agents could be informative but remain scarce [11,12,13].

In this narrative review, we outline the pathophysiologic mechanisms that underpin the potential utility of anti-rheumatic therapies in the context of severe COVID-19 infection and summarize the current evidence regarding the risks and outcomes of COVID-19 infection in patients with systemic autoimmune diseases.

Rational for the use of anti-rheumatic drugs in the treatment of severe COVID-19 infection

Driven by data indicating that an immune overreaction, a “cytokine storm”, may be a critical component of severe COVID-19 infection [5,6,7], it has been hypothesized that the use of drugs blocking specific cytokines could be beneficial for some infected patients [9, 10]. Previous observations in patients infected with different viruses such as the influenza virus [14, 15] and other coronaviruses like the SARS-CoV and MERS-CoV [16] suggested that an exaggerated host immune response could participate in tissue injury leading to serious outcomes.

Data regarding the exact nature of host immune responses’ against COVID-19 are now starting to emerge [17]. As is the case with other respiratory viruses, the initial innate antiviral immune responses in the respiratory tract are crucial for effective control of the infection. In the majority of cases (> 80%), these responses are adequate, leading to viral clearance and clinical improvement without significant associated tissue (mainly lung) injury. Data regarding immune responses in recovered patients following mild or moderate illness are limited so far.

On the other hand, in a proportion of patients (~ 20%), severe lung injury is observed which may lead in approximately 5% of patients to ARDS, respiratory and multi-organ failure or even death (~ 1%). In such cases, a dysregulated immune response has been described, usually during the 2nd week of infection, characterized more often by T cell lymphopenia (both CD4 and CD8 cells) [5], enhanced production and release of pro-inflammatory cytokines such as interleukin-6 (IL-6), tumor necrosis factor-a (TNF-a), IL-1 and G-CSF [5,6,7, 18], decreased production of antiviral type I and III interferons [17] and T cell exhaustion [17].This “hyperinflammatory state” which is being referred to as “cytokine storm” shares certain clinical (high fever), laboratory (increased C-reactive protein-CRP, ferritin, serum IL-6 and D-dimer levels) and immunological (macrophage and T cell activation) features with the macrophage activation syndrome or secondary hemophagocytic lymphohistiocytosis (HLH) seen in a subset of patients with infections (bacterial or viral), auto-immune or auto-inflammatory diseases (such as systemic onset juvenile idiopathic arthritis, adult Still’s disease and systemic lupus erythematosus) and after CART cell therapy in patients with leukemia or lymphoma [19,20,21].

Such observations underpinned the hypothesis that targeting pro-inflammatory cytokines with specific drugs used in autoimmune diseases may control the hyperinflammatory state in COVID-19 infection; however, this approach also raises concerns for a blunted response to viral invasion [22].

Targeted anti-cytokine therapies neutralize individual mediators of inflammation rather than causing a complete shutdown of the innate and adaptive immune responses. Hence, the risk from specific pathogens may differ between therapies. The most commonly used biologic Disease Modifying Anti-Rheumatic Drugs (bDMARDs), i.e. the anti-TNF agents (infliximab, etanercept, adalimumab, golimumab and certolizumab) increase the risk of serious bacterial infections [23] while there is some limited evidence that the may also increase the risk for influenza infection [24]. Humoral immunity may be affected more by some drugs (such as the anti-CD20 monoclonal antibody, rituximab) than other targeted anti-cytokine therapies, such as anti-IL-23,-IL-4/IL-13 or -IL-17A, which have not been associated with a significant increase in the risk of viral infections.

So far there have been few studies assessing the role of cytokine-targeted (mainly IL-1 and IL-6) therapies in patients with COVID-19 infection [25,26,27,28,29] but none of them were randomized controlled studies.

However, drugs with the potential to inhibit multiple cytokines may be riskier. These include the more recently developed targeted synthetic (ts)DMARDs, mainly the JAK inhibitors. These drugs target JAK1 and JAK3, affecting the function of several cytokines that are involved in antiviral responses (including type I interferons, IL-2, IL-15, IL-21 and IFNγ) and could theoretically inhibit viral clearance [30]. The flip side of this is that JAK2 inhibition may block SARS-CoV-2 entry and pro-inflammatory cytokines that depend on JAK signaling (including IL-17, IL-6 and GM-CSF) [31, 32].

The majority of patients with autoimmune diseases, however, still receive the older conventional synthetic (cs)DMARDs and/or variable doses of glucocorticoids that may also have significant effects in the context of COVID-19 infection. Hydroxychloroquine has been suggested as a potential therapeutic agent for prophylaxis or therapy of COVID-19 infection but the results so far have been conflicting [25, 33,34,35,36,37,38,39]. There is concern regarding the use of moderate- to high-dose glucocorticoids in COVID-19 infection [40, 41] and so far these are only recommended for mechanically ventilated patients with ARDS [42]. There are no data so far for other commonly used csDMARDs, such as methotrexate, azathioprine, cyclosporine and mycophenolate mofetil.

In summary, whereas on the one hand b- or ts-DMARDs may theoretically increase the risk and severity of viral infections, there are, on the other hand, mechanistic and clinical observations suggesting that they may be useful in controlling specific aspects of the immune overreaction and thus improve the outcomes of severe COVID-19 infection. Indicative answers for both questions may arise from well-conducted observational studies in patients with autoimmune diseases receiving DMARD therapy compared with the general population. A definitive answer for the latter question could only be provided as a result of appropriately conducted randomized controlled trials (RCTs) in patients with severe COVID-19 infection [43].

Current data on COVID-19 infection in patients with rheumatic diseases

A literature search in the electronic databases PubMed, Scopus and Web of Science was performed until May 29th, using the terms “COVID-19” and “rheumatic diseases”, “disease-modifying anti-rheumatic drugs”, “biologics”, “rheumatoid arthritis”, “systemic lupus erythematosus”, “spondyloarthropathies”, “psoriatic arthritis”, “ankylosing spondylitis”, “vasculitis”, “scleroderma”, “JAK inhibitors”, glucocorticoids” or “corticosteroids” in order to identify relevant publications of patients with rheumatic diseases and COVID-19 infection. Articles were initially selected by their title and abstract and then the full text was searched for the appropriate relevant content. References from the selected articles were also manually searched for relevant articles.

In Table 1, we summarize data from all published studies and case series which described COVID-19 infection in patients with autoimmune diseases. There is significant heterogeneity both in the general characteristics as well as the autoimmune disease characteristics and treatments of the patients included in these cohorts.

Table 1 Anti-inflammatory treatment and outcomes of patients with confirmed or suspected COVID-19 infection
Full size table

The largest series with detailed data so far is from New York [44]. It reports data on the outcome of 86 patients with autoimmune/inflammatory diseases who had either confirmed (59 patients) or suspected (27 patients) COVID-19 infection (Table 1) [44].The majority were females (57%) at a rather young age (mean 46 years). Most common diagnoses included spondyloarthritis (SpA) and/or psoriasis (PSO), inflammatory bowel diseases (IBD), rheumatoid arthritis (RA) or a combination of these diseases. Compared to other patient cohorts, comorbidities were rather uncommon (hypertension: 13%, COPD: 5%, diabetes: 6%). Most patients were on biologics or JAK inhibitors (72%) with few receiving glucocorticoids (9%). Hospitalization was required for 14 (16%), ICU admission or mechanical ventilation in 1 patient (7%) while there was only 1 death at arrival in the ER (7%). The rate for hospitalization was not different from that of the general population of New York (26%).

An observational study from France monitored the clinical course of COVID-19 infection in 17 patients with systemic lupus erythematosus (SLE) who were on long-term hydroxychloroquine therapy (median 7.5 years) (Table 1) [45]. Comorbidities were common in this group including obesity (59%) and chronic kidney disease (47%). All but one patient had clinically quiescent SLE, with a SLEDAI score equal to 0. Twelve (71%) patients were receiving glucocorticoids (usually at doses < 10 mg/day) and seven (41%) were receiving additional immunomodulatory drugs. Hydroxychloroquine and glucocorticoids were maintained at the same dose, while immunosuppressive drugs were discontinued or reduced. Fourteen patients required hospitalization (82%), half of them (n = 7) in the ICU and finally 2 patients (14%) died. Although this study is also limited by small numbers, the authors concluded that hydroxychloroquine does not appear to prevent severe COVID-19 infection. Regarding the severity of COVID-19 in SLE patients, no safe conclusion can be drawn, but the high incidence of other comorbidities may confound these observations.

In the relatively early phase of the outbreak in N. Italy, Monti et al. performed a survey in patients with chronic arthritis in their outpatient clinic to investigate potential infections with COVID-19 or high-risk contacts (Table 1) [46]. The authors gathered information from 320 patients (57% with RA, 43% with SpA, 52% treated with anti-TNFs, 40% with other bDMARDs and 8% with tsDMARDs). They identified 4 confirmed and 4 suspected COVID-19 infections while another 5 reported high-risk contacts but remained asymptomatic for the 2-week observation period. Three patients (one with confirmed and two with suspected COVID-19) were on hydroxychloroquine. All patients with symptoms of infection had their anti-rheumatic therapy temporarily withdrawn at the time of symptom onset. No significant relapses of the rheumatic disease occurred;' none of the patients with a confirmed or highly probable COVID-19 developed severe respiratory complications or died and only one patient with confirmed infection, aged 65, required admission to hospital and received low-flow oxygen supplementation for a few days. All patients with confirmed COVID-19 received at least one antibiotic course, and the hospitalized patient also received antiviral therapy and hydroxychloroquine. The authors also reported that among 700 patients admitted for severe COVID-19 during one month at their hospital, which was a referral center for COVID-19, none was receiving either bDMARDs or tsDMARDs [46].

An initiative driven by the Global Rheumatology Alliance aims to continuously collect data internationally for patients with rheumatic diseases infected with COVID-19. Initially data for 110 patients were reported [47] and more recently its updated form containing data for 600 patients (548 with confirmed and 52 with presumptive diagnosis of COVID-19) were published (Table 1) [48]. The most common diseases were RA (38%), SpA (20%), SLE (14%) and other diseases (33%, including vasculitis, Sjogren’s syndrome etc.). Medications included csDMARDs in 48%, bDMARDs in 29%, tsDMARDs in 4% and glucocorticoids in 27%. Common comorbidities included hypertension in 33%, lung disease in 21%, diabetes in 12%, cardiovascular disease in 11% and chronic renal insufficiency in 7%. Of the 600 infected patients, 46% were admitted to hospital and 9% died. By multivariable logistic regression analysis, age > 65 years (odds ratio—OR 2.56), certain comorbidities (hypertension or cardiovascular diseases, lung disease, diabetes and chronic renal disease, ORs 1.86–3.02) and high glucocorticoid dose (≥ 10 mg/day, OR 2.05) were associated with a higher risk for hospitalization, On the other hand, previous use of anti-TNF agents was associated with a lower hospitalization risk (OR 0.40) [48]. These results should be interpreted with caution since selection bias may have played a significant role here with inclusion of more severe cases.

Case reports of patients with autoimmune/rheumatic diseases who were infected with COVID-19 are also increasingly appearing in the literature. A brief report from Italy described the incidence of COVID-infection and outcome among 123 patients with autoimmune/inflammatory diseases [49]. Most patients, ~ 60% were receiving csDMARDs, 20% bDMARDs, while 64% were also receiving relatively low dose glucocorticoids (average 5.3 mg of prednisolone daily). Only one registered patient (0.81%) developed a confirmed COVID-19 infection; this incidence was similar to that of COVID-19 infection in the area (0.62% at that time).

An interesting case of a 52-year-old woman with granulomatosis with polyangiitis for 32 years, who developed severe COVID-19 infection while on therapy with rituximab and glucocorticoids has been reported [50]. The patient had ear, nose and throat, orbital, lung, joint and skin involvement, anti-proteinase 3 (PR3) antibodies and had been heavily pretreated (with cyclophosphamide—total dose of 41 g, anti-TNF, mycophenolate mofetil, methotrexate, leflunomide, rituximab and steroids). Few months before the pandemic she received four infusions of weekly rituximab (375 mg/m2) and a maintenance dose of rituximab (500 mg) was administered in March 2020, while still under prednisone 15 mg daily. She was diagnosed with COVID-19 having presented with bilateral interstitial pneumonia 4 days after her last rituximab dose. She received lopinavir/ritonavir for 3 days but required mechanical ventilation, for ARDS, 18 days after symptom onset. Hydroxychloroquine was started following intubation and the authors report rapid clinical improvement with no need for mechanical ventilation after 2 days and no need for oxygen supplementation after 7 days, while nasopharyngeal RT-PCR became negative and the patient was discharged. The authors hypothesized that glucocorticoids and rituximab may have limited the cytokine storm and delayed the worsening of clinical status of the patient.

A very brief discussion of a few cases of COVID-19 infection in patients with autoimmune/inflammatory diseases from the UK has also been published, without any details reported [51]. These cases include a 23-year-old male with Crohn’s disease on infliximab with mild disease who did not require hospitalization, a female with ulcerative colitis (of unspecified age) on tofacitinib and steroids who also did not require hospitalization and a 53-year-old female with Crohn’s disease on adalimumab who was hospitalized but did not require mechanical ventilation.

Conclusions

As of June 10th 2020 about 7.2 million individuals have tested positive for, and more than 410,000 have died from COVID-19 [52]. Despite these numbers, data regarding the risk of infection and the severity and outcome of the disease in patients with systemic autoimmune diseases remain extremely limited. The estimated prevalence of systemic autoimmune/inflammatory diseases in the general population ranges between 2 and 3% [53]. Assuming an equal risk for COVID-19 infection and death with the general population, one would have expected at least 5000 deaths of such patients so far. Were data on them to have been collected systematically through international collaborations, these are useful numbers for detailed analyses of risk with potentially powerful clinical messages. One should take into account though that the percentages of patients with systemic autoimmune disease who receive immunosuppressives may be lower within the elderly population which is more vulnerable to severe outcomes. Moreover, since males seem to have higher death rates than females [54], the female preponderance in systemic autoimmunity could be a protective factor.

Based on the available data at present, any conclusions would be unsafe and should be seen with great caution. Analysis of the published data presented in this review suggest the following: (A) there is thus far no convincing evidence that any of the DMARDs (conventional synthetic, biologic or targeted synthetic) including hydroxychloroquine, used in the management of systemic autoimmune diseases may protect against severe COVID-19 infection; (B) answers about the possible usefulness of DMARDs in the management of the cytokine storm and its sequelae during severe COVID-9 infection will only arise from the currently ongoing RCTs; (C) the risk of COVID-19 infection in patients with systemic autoimmune diseases does not seem to be much higher than in the general population with similar comorbidities [46, 49]; (D) the severity of COVID-19 infection (in terms of requirement for hospitalization, ICU admission or death) does not appear to be particularly dissimilar in this population compared with the general population, although data from large series of patients are not available yet; (E) it is impossible at this stage to draw any conclusions for differences in COVID-19 risk and outcome between different autoimmune diseases and between the various immunomodulatory therapies used for them.

More research in the field is obviously required, including as a minimum careful and systematic epidemiology, basic laboratory research and well-designed and appropriately controlled clinical trials.