The Development or Exacerbation of an Autoimmune Disorder Associated with SARS-CoV-2 infection: a Systematic Review of Case Reports

Background: Infections have long been studied as environmental triggers of autoimmunity. Previous associations between coronavirus and autoimmune disorders make severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), a novel coronavirus identied in December 2019, a potential candidate for autoimmune disorder development. Indeed, an increasing number of these cases have been reported prompting a characterization and summarization of existing evidence. Objectives: This review aims to characterize and summarize case reports documenting the development or exacerbation of an autoimmune disorders associated with SARS-CoV-2 infection. Methods: A bibliographic search of Embase was performed from inception to July 2020. Studies included reported the development or exacerbation of an autoimmune disorder associated with SARS-CoV-2 infection. Quality was judge using Joanna Briggs Institute Critical Appraisal Checklist for Case Reports and piloted forms were used for data collection. Data were summarized using descriptive statistics and narrative synthesis. Results: From 304 entries, 85 different cases were included, 41%-99% with good appraisal across quality domains. Sixty-two (72.9%) patients had no previous autoimmune disorder and 70 (82.4%) developed it De novo, the most frequent being Guillain–Barré syndrome and multisystem inammatory syndrome in children. Eighteen (21.2%) cases of autoimmune disorders, mostly in remission, exacerbated. Organ-specic disease was found in 57 (67.1%) and systemic in 28. Disease affection was categorized into groups, the most frequent being neurological (36; 42.4%), vasculitis (19; 22.4%), blood (15; 17.6%), and connectivitis/systemic lupus erythematosus/antiphospholipid syndrome (6; 7.1%). The median coronavirus disease-2019 to De novo autoimmune disorder latency was 11 days (IQR = 5.75-16). Conclusions: Although, as a systematic review of case reports, this study cannot verify causality, it provides support for further studies on the relationship between SARS-CoV-2 and autoimmune disorders. Furthermore, hemolytic anemia (AIHA). A total of 22 different ADs developed, while 12 distinct AD exacerbations were found.

The most common laboratory nding was thrombocytopenia and anemia. Other virus screens were negative (Epstein-Barr virus, hepatitis virus, Human Immunode ciency Virus (HIV), Cytomegalovirus). Four patients had positive direct Coombs test, two with high cold agglutinin titers, one of these also with positive antinuclear antibodies (ANA). One case reported low ADAMTS-13 activity and positive ADAMTS-13 inhibitory antibodies, another antiplatelet antibodies, and another positive (albeit weak) anti-platelet factor 4/heparin antibodies. Two cases performed bone marrow aspirate, revealing normal cellularity with increased megakaryocyte count ( Table 2). Chest and brain CT revealed two pulmonary embolism and two intracerebral hemorrhages.
All patients received AD treatment, the majority with corticosteroids, IVIG, platelet transfusions and red blood cell transfusions. For COVID-19, four required no treatment and only three required ICU admission. Mean hospitalization time was 19.85 (11.16 SD) days and one death was reported ( Table 2).
Other organ-speci c ADs Three cases had an exacerbation of IBD: one ulcerative colitis (UC) and one Crohn's disease previously in remission, and one active Crohn's disease. The Crohn's disease patients also presented De novo IgA vasculitis and Kawasaki disease. Age ranged from 14 to 26 years old. The UC case was a pregnant woman with proctitis, Disease Activity Index score of 3, and active chronic in ammation on biopsies, treated with methylprednisolone and cyclosporine, and ultimately undergoing spontaneous abortion. Abdominal CT and MRI enterography found ileitis in both Crohn's disease cases. One (previously in remission) also had perianal abscess and stula requiring drainage, antibiotics, and in iximab, while the other received the regular adalimumab treatment (Table 1).
Concerning skin-related ADs, two exacerbations of psoriasis (one in remission) and one De novo livedo reticularis were reported. The psoriasis cases occurred in a 73-year-old male 14 days after COVID-19 onset, and in a 71-year-old female 4 days after, neither receiving AD treatment. The livedo reticularis case occurred in a 57-year-old male with high d-dimers and positive ANAs, 8 days after COVID-19 onset, and treated with low-molecular-weight heparin (Table 1).
For endocrinological-related ADs, one diabetes mellitus type 1 exacerbation in the form of diabetic ketoacidosis, and one autoimmune hypothyroidism exacerbation was reported. The diabetic ketoacidosis case occurred in a 60-year-old male 5 days after COVID-19 onset, complicating with secondary bacterial infection and pulmonary embolism. He was treated with insulin and required ICU admission. The autoimmune hypothyroidism exacerbation received no treatment and occurred concurrent with De novo ITP in a 65-year-old female with elevated thyroid peroxidase antibodies, 11 days after COVID-19 onset (Table 1).
An ophthalmological-related AD befell an 11-year-old male with left retinal vasculitis on ocular fundus examination, negative autoimmune panel, and chilblains (Table 1).
Seven (36.8%) cases reported high troponins and one (5.3%) elevated myoglobin. The IgA vasculitis case had high complement C4 and serum IgA, and one cutaneous small-vessel vasculitis case had anti-cardiolipin antibodies. The ANCA-associated vasculitis exacerbation case was cytoplasmic-ANCA positive, and the MIS-A case presented positive ANAs, anti-Ro/SSA and low complement C3 and C4. From 14 heart ultrasounds requested, seven reported depressed left ventricular systolic function, ve coronary artery dilatation, and three pericardial effusion. Abdominal CT exams showed bowel in ammation (ileitis and/or colitis) in four cases. Three patients underwent skin lesion biopsy, all wielding leukocytoclastic vasculitis as the major nding (Table 2).
Most patients received antibiotic prophylaxis, IVIG, corticosteroids, immunosuppressive therapy (tocilizumab, anakinra, or cyclophosphamide), and acetylsalicylic acid, while two received no AD treatment. Most patients required ICU admission with one Kawasaki disease, and seven MIS-C patients requiring rescue and supportive therapy for cardiogenic shock. Nevertheless, most required no COVID-19 treatment. Death was reported in two and mean hospitalization time was 14.92 (10.48 SD) days (Table 2).
Connectivitis/SLE-APS-related AD Median age was 67 (50-69.25 IQR) and three were female. Four cases had no previous AD and developed De novo APS, and two had APS/SLE (previously in remission) and APS exacerbations. Median COVID-19 to De novo AD latency was 25.5 (12-45 IQR) days ( Table 2).
The De novo APS cases had positive anti-beta-2-glycoprotein I and anti-cardiolipin antibodies, and high d-dimers, while three had high brinogen, prothrombin and partial thromboplastin time, and one high brin degradation products (Table 2). Patients had multiple cerebral (5; 83.3%) and bilateral cerebellar (3; 50.0%) infarctions on brain CT. On doppler ultrasound one patient had bilateral jugular venous thrombi. One De novo APS case complicated with index nger dry gangrene and another with two digits and lower limbs bilateral ischemia.
Three cases received AD treatment, one with acetylsalicylic acid, low-molecular-weight heparin and plasma exchange, and two with IVIG, one of which with eltrombopag, platelet transfusions and prednisone. Five required COVID-19 treatment, three with supportive therapy. Mean hospitalization time was 21.17 (18.76 SD) days with four ICU admissions and one reported death (APS exacerbation) ( Table 2).
Other systemic AD For Arthritis-related ADs, one ankylosing spondylitis exacerbation (previously in remission) and one De novo non-classi ed arthritis were reported. The ankylosing spondylitis case occurred in a 53-year-old female 22 days after COVID-19 onset with normal brain MRI and nerve conduction studies, and was treated with etanercept. The non-classi ed arthritis case occurred in a 57-year-old male 20 days after COVID-19 onset with negative ANAs, rheumatoid factor and anti-cyclic citrullinated peptide antibodies, and the synovial uid was crystal-free on polarized microscopic examination (Table 1).
One endocrinological-related AD case, a 32-year-old female with autoimmune polyglandular syndrome type 1, presented with hypoparathyroidism exacerbation with low serum calcium, and was treated with calcitriol and calcium supplementation. She required ICU admission for COVID-19 (Table 1).

Discussion
This review describes 22 different De novo ADs and 12 distinct AD exacerbations associated with SARS-CoV-2 infection.
As AD cases in children (<18 years) have been reported following SARS-CoV-2 infection, most notably MIS-C [14,15], one could argue the existence of more associations. In this regard, entry age was not restricted, and besides MIS-C and Kawasaki disease, AIHA and retinal vasculitis were also found. The reported higher prevalence of older patients, cardiovascular risk factor comorbidity, and cardiovascular disease could be explained by their added risk of severe COVID-19 and hospital admission [91,92], in line with this review's inpatient cohort. Contrary to the literature on AD gender differences [93], there were more males than females found (9:8).
COVID-19 was predominantly diagnosed through rRT-PCR, the present recommendation. Laboratory and radiologic ndings were similar to those found in the literature [7,91,94]. Treatment was heterogeneous, re ecting the immaturity of the disease knowledge. Most were treated with hydroxychloroquine and antibiotics (predominantly azithromycin), the recommendation at the time.
Additionally, similar to another review [94], a reduced number of patients received no treatment for COVID-19, and intubation, mechanical ventilator support, and mortality were higher than the general population, suggesting greater disease severity, which, again, may relate to the number of hospitalizations in this cohort. Furthermore, more patients than expected required ICU admission which may be explained by the association of AD complications.
Most cases had no previous AD and, ultimately, all developed one, which supports the causal relationship hypothesis. Furthermore, most patients with AD comorbidity exacerbations were previously in remission. In the nal major group, APS, four anti-beta-2-glycoprotein I and anti-cardiolipin antibody positive patients complicated with multiple thrombotic events. Though antiphospholipid antibodies can be present in acute infections, they usually present with low levels and are not associated with thrombotic events [110]. Infections are also suspected triggers as they preceded APS in many cases [111,112].

Limitations
Even though good appraisal in several quality domains was seen across studies, key elements of good reporting were lacking, and 21 cases did not report an outcome. Likewise, other reviews shared the same ndings [94]. A possible explanation might be the urgency to publish new information, as COVID-19 is a recently identi ed disease. This, however, may lead to a lack of important information that could further the knowledge of possible associations with ADs. Nevertheless, all cases were included as a way to nd the most possible number of associations.
Although numerous associations between infections and ADs are described, the acute state of infection and hyperin ammatory state which occurs with COVID-19 could mimic the immunological ndings of an AD. Since the reported autoimmune manifestations might encompass the immunological component of the infection, caution is warranted when interpreting the results. Nevertheless, a high suspicion rating is always necessary when considering an AD diagnosis.
Furthermore, milder AD cases might not be represented due to publication bias as more severe cases are more likely to be published and diagnosed.
Although a thorough search was done, using an expansive search query to be as inclusive as possible, methodological limitations are still present. First, only one database was used and, even though this review provides a look at the rst image of the COVID-19 pandemic, it may have missed studies that were published since the search date. Additionally, despite using piloted forms and training to minimize inaccuracies, only one reviewer performed data extraction.

Conclusions
This systematic review provides a characterization and summarization of AD cases associated with SARS-CoV-2 infection. There are some ndings that may suggest the existence of this relationship and although this study as a review of case reports cannot verify causality, it supports further studies and provides a resource for clinicians to be aware of possible AD complications of SARS-CoV-2 infected patients.

Availability of Supporting Data
The data set supporting the results of this article is included within the article (and its additional les).

Competing interests
None of the authors have any nancial or non-nancial competing interests to disclose.

Funding
None.
Author's contributions NG performed the screening, study selection, data collection, analysis, and interpretation, conceptualized and drafted the manuscript. EM performed screening and study selection, analyzed data, and contributed to the writing of the manuscript. CC aided in data analysis and interpretation and performed critical revision of the article. PMT contributed to conception and design of the work. All    Table 2. Total and by group summary of findings. General population and case characteristics, laboratory findings divided by areas, COVID-19 related radiologic findings and treatment, and AD-related treatment summarized with frequencies and descriptive statistics for the total and all major groups.