BMJ Open Is Committed to Open Peer Review. As Part of This Commitment We Make the Peer Review History of Every Article We Publish Publicly Available

BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

BMJ Open is committed to open peer review. As part of this commitment we make the peer review history of every article we publish publicly available.

When an article is published we post the peer reviewers’ comments and the authors’ responses online. We also post the versions of the paper that were used during peer review. These are the versions that the peer review comments apply to.

The versions of the paper that follow are the versions that were submitted during the peer review process. They are not the versions of record or the final published versions. They should not be cited or distributed as the published version of this manuscript.

BMJ Open is an open access journal and the full, final, typeset and author-corrected version of record of the manuscript is available on our site with no access controls, subscription charges or pay-per-view fees (http://bmjopen.bmj.com).

If you have any questions on BMJ Open’s open peer review process please email [email protected] http://bmjopen.bmj.com/ on September 28, 2021 by guest. Protected copyright. BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

Relatively frequent rheumatic manifestations and autoantibodies in infective endocarditis: differentiating infective endocarditis from autoimmune disease is needed ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-031512

Article Type: Research

Date Submitted by the 18-May-2019 Author:

Complete List of Authors: Zhou, Zhuochao; Shanghai Jiao Tong University School of Medicine, ; Ye, Junna Teng, Jialin Liu, Honglei Cheng, Xiaobing Sun, Yue Su, Yutong Chi, Huihui Wang, Fan Yang, Chengde; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital Jin, Wei http://bmjopen.bmj.com/ Infective endocarditis, Rheumatic manifestation, ANCA, Antiphospholipid Keywords: antibody

on September 28, 2021 by guest. Protected copyright.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 6 7 8 9 I, the Submitting Author has the right to grant and does grant on behalf of all authors of the Work (as defined 10 in the below author licence), an exclusive licence and/or a non-exclusive licence for contributions from authors 11 who are: i) UK Crown employees; ii) where BMJ has agreed a CC-BY licence shall apply, and/or iii) in accordance 12 with the terms applicable for US Federal Government officers or employees acting as part of their official 13 duties; on a worldwide, perpetual, irrevocable, royalty-free basis to BMJ Publishing Group Ltd (“BMJ”) its 14 licensees and where the relevant Journal is co-owned by BMJ to the co-owners of the Journal, to publish the 15 Work in this journal and any other BMJ products and to exploit all rights, as set out in our licence. 16 17 The Submitting Author accepts and understands that any supply made under these terms is made by BMJ to 18 the Submitting Author Forunless you peer are acting as review an employee on behalf only of your employer or a postgraduate 19 student of an affiliated institution which is paying any applicable article publishing charge (“APC”) for Open 20 Access articles. Where the Submitting Author wishes to make the Work available on an Open Access basis (and 21 intends to pay the relevant APC), the terms of reuse of such Open Access shall be governed by a Creative 22 Commons licence – details of these licences and which Creative Commons licence will apply to this Work are set 23 out in our licence referred to above. 24 25 Other than as permitted in any relevant BMJ Author’s Self Archiving Policies, I confirm this Work has not been 26 accepted for publication elsewhere, is not being considered for publication elsewhere and does not duplicate 27 material already published. I confirm all authors consent to publication of this Work and authorise the granting 28 of this licence. 29 30 31 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 Relatively frequent rheumatic manifestations and 5 6 7 autoantibodies in infective endocarditis: 8 9 10 differentiating infective endocarditis from 11 12 autoimmune disease is needed 13 14 15 16 17 Zhuochao Zhou1,*, Junna Ye1,*, Jialin Teng1, Honglei Liu1, Xiaobing Cheng1, Yue 18 For peer review only 19 1 1 1 1 1,# 2,# 20 Sun , Yutong Su , Huihui Chi , Fan Wang , Chengde Yang , Wei Jin 21 22 23 24 25 1Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 26 27 University School of Medicine; 28 29 30 2Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School 31 32 33 of Medicine 34 35 36

37 http://bmjopen.bmj.com/ 38 *These authors contributed equally to this work. 39 40 41 42 43 #Correspondence: 44

45 on September 28, 2021 by guest. Protected copyright. 46 Wei Jin, MD, PhD 47 48 Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School of 49 50 51 Medicine, No. 197 Ruijin Second Road, Huangpu District, Shanghai 200025, China. 52 53 Tel.: +86 21 64370045; 54 55 56 Fax: +86 21 34186000; 57 58 59 E-mail: [email protected] 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 6 7 Chengde Yang, MD, PhD 8 9 Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 10 11 12 University School of Medicine, No. 197 Ruijin Second Road, Huangpu District, 13 14 Shanghai 200025, China. 15 16 17 Tel.: +86 21 64370045; 18 For peer review only 19 20 Fax: +86 21 34186000; 21 22 E-mail: [email protected] 23 24 25 26 27 28 29 30 31 32 33 Abstract 34 35 Objective: This study aimed to characterize rheumatic manifestations and 36

37 http://bmjopen.bmj.com/ 38 auto-antibodies of 432 patients diagnosed with infective endocarditis (IE). 39 40 Design, setting and participants: A retrospective study was conducted in Ruijin 41 42 43 Hospital from 1997 to 2017. The clinical and laboratory characteristics of a total of 44

45 on September 28, 2021 by guest. Protected copyright. 46 432 patients were analyzed. Besides, the differences between patients with positive 47 48 and negative antineutrophil cytoplasmic antibody (ANCA) and anti-phospholipid 49 50 51 antibody (aPL), as well as survival rate of these patients, were compared. 52 53 Results: 432 patients with 278 male patients and 154 female patients were included. 54 55 56 The mean age of patients was 45.61±16.12 years. At last, 346 patients (80.09%) had 57 58 59 cardiac surgery and 55 patients (12.73%) died in the hospital. Among IE patients, 104 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 of them were either tested ANCA or aPL, and were analyzed in different groups. 5 6 7 Interestingly, positive ANCA patients were all of proteinase 3(PR3)-ANCA. 8 9 Compared with ANCA-negative group, patients with positive ANCA had higher IgM 10 11 12 (P=0.048), lower hemoglobin (P=0.001), more presentation of arthritis (P=0.003), and 13 14 more serositis (P=0.06). However, IE patients with positive aPL had higher level of 15 16 17 erythrocyte sedimentation rate (ESR) compared to aPL-negative group (P=0.003). In 18 For peer review only 19 20 addition, survival rate of ANCA-positive IE patients was lower (P=0.032) than 21 22 ANCA-negative group, while there was no difference in patients with or without aPL 23 24 25 antibody (P=0.728). 26 27 Conclusion: The present study supports the claim that atypical rheumatic 28 29 30 manifestations and presentation of antibodies are frequent in patients with IE and lead 31 32 33 to early misdiagnosis. Physicians should pay more attention to the measurement of 34 35 autoantibodies in these patients. 36

37 http://bmjopen.bmj.com/ 38 39 40 Keywords: infective endocarditis, rheumatic manifestation, ANCA, aPL 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 Strengths and limitations of this study 53 54 55 1. The study analyzed the differences and survival rate between patients with and 56 57 without ANCA or aPL antibodies in IE patients, which were not analyzed together 58 59 60 before.

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 2. In patients who tested both ANCA and aPL, positive LAC and proteinase 5 6 7 3(PR3)-ANCA were the most common types. 8 9 3. Atypical rheumatic manifestations and presentation of antibodies are frequent in 10 11 12 patients with IE and lead to early misdiagnosis. 13 14 4. As it was a retrospective study, some patients did not detect autoantibodies when 15 16 17 diagnosed. 18 For peer review only 19 20 21 22 23 Data availability statement 24 25 26 27 All data relevant to the study are included in the article or uploaded as supplementary 28 29 30 information. 31 32 33 34 35 36

37 Declaration of conflicting interests http://bmjopen.bmj.com/ 38 39 40 The authors report no relationships that could be construed as a conflict of interest. 41 42 43 44

45 Funding on September 28, 2021 by guest. Protected copyright. 46 47 48 This work is supported by National Natural Science Foundation of China 49 50 (81871272, 81801592, 81370397, 81670266), Science and Technology Commission 51 52 53 of Shanghai Municipality (17140902500), Shanghai Sailing Program (18YF1414100), 54 55 Shanghai Jiao Tong University Interdisciplinary Research Project (YG2016QN60), 56 57 58 Excellent Youth B Project (GCQN-2017-B05) and Innovative research team of 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 high-level local universities in Shanghai. 5 6 7 8 9 10 11 12 13 14 15 Introduction 16 17 Infective endocarditis (IE) is a microbial infection of the endocardium and usually 18 For peer review only 19 20 influences heart valves. Despite being relatively rare worldwide, IE has been reported 21 22 by a steady incidence over the past three decades ranging from 2~6 per 100,000 23 24 25 individuals in the general population per year, with a considerable associated 26 27 mortality varying from 10% to 30%.1 Predisposing factors of IE include valvular heart 28 29 30 diseases such as congenital heart diseases, rheumatic valve disease, artificial valves, 31 32 2 33 intravenous drug use, dental procedure and hemodialysis. Diagnosis of IE is made 34 35 based on symptoms, blood cultures, cardiac ultrasound and pathological biopsy.3 36

37 http://bmjopen.bmj.com/ 38 However, superantigens induced by some bacteria like staphylococcus aureus can 39 40 stimulate the immune response, which could present related manifestations mimicking 41 42 43 rheumatic diseases. Rheumatic manifestations such as myalgia, arthralgia, and 44

45 on September 28, 2021 by guest. Protected copyright. 46 arthritis are prevalent in nearly 40% of patients with IE at presentation, weeks to 47 48 months before diagnosis of IE.4 When classic IE manifestations are less evident, it 49 50 51 could cause misdiagnosis. Thus, the distinction between IE and rheumatic diseases is 52 53 not crystal clear, and to understand the difference is of great importance. 54 55 56 57 58 59 Many special antibodies like antineutrophil cytoplasmic antibodies (ANCA) and 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 anti-phospholipid antibody (aPL), might be related with the pathophysiology of IE. 5 6 7 Due to positive ANCA tests, IE has been reported to mimic ANCA-associated 8 9 vasculitis. Patients with IE may present with multiple pulmonary nodules and 10 11 12 glomerulonephritis, mimicking granulomatosis with polyangiitis.5 Langlois V’s study 13 14 further underscored that ANCA might be associated with multiple valve 15 16 17 involvement.6 On the other hand, infection-associated elevated aPL levels in patients 18 For peer review only 19 20 with infective endocarditis are related to endothelial cell activation, thrombin 21 22 generation and impairment of fibrinolysis. This may contribute to the increased risk 23 24 25 for major embolic events in these patients.7 Therefore, ANCA and aPL could be 26 27 non-specific in IE patients and confusing. 28 29 30 31 32 33 In Turkey, it has been reported that IE occurs in relatively young patients. Prosthetic 34 35 and degenerative valve disease is becoming a predisposing factor.8 In Lebanon, 36

37 http://bmjopen.bmj.com/ 38 El-Chakhtoura N et al performed a retrospective study of 80 patients with IE from 39 40 2001 to 2014. The most commonly isolated organisms were streptococci (37%) and 41 42 43 staphylococcus aureus (11%).9 Although there was data reported in various countries 44

45 on September 28, 2021 by guest. Protected copyright. 46 about different aspects of IE, the analysis of the clinical and immunologic features 47 48 from Chinese IE patients was still limited and there is no report of patients both tested 49 50 51 with ANCA and aPL. We have already reported cases and literature review of IE 52 53 patients with ANCA before,10 the objective of this study is to make the profile of 54 55 56 clinical and laboratory manifestations of 432 IE patients in a tertiary hospital in China 57 58 59 from 1997 to 2017, and analyze the characteristics of IE patients tested with ANCA or 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 aPL. The characteristics of subgroups of positive or negative ANCA and aPL were 5 6 7 studied as well as survival rate of these IE patients. 8 9 10 11 12 Methods 13 14 Patients 15 16 17 We retrospectively analyzed the features of 432 patients hospitalized in Ruijin 18 For peer review only 19 20 Hospital affiliated to Shanghai Jiao Tong University School of Medicine, diagnosed 21 22 with IE from 1997 to 2017 according to the modified Duke criteria.3 We had ruled out 23 24 25 the patients with primary rheumatic disease or in immunosuppression condition. 26 27 Cardiac surgery and death were recorded during the period of hospitalization. 28 29 30 Relapses were defined as repeat episodes of IE caused by the same microorganism 31 32 3 33 less than 6 months after the initial episode as previously described. Three blood 34 35 samples were taken from a peripheral vein at 30 minutes intervals, and were incubated 36

37 http://bmjopen.bmj.com/ 38 in both aerobic and anaerobic bottles. The study followed the ethical standards for 39 40 human experimentation established in the Declaration of Helsinki and was approved 41 42 43 by the Institutional Research Ethics Committee of Ruijin Hospital (ID: 2016-62), 44

45 on September 28, 2021 by guest. Protected copyright. 46 Shanghai, China. 47 48 49 50 51 104 IE patients who had been tested for either ANCA or aPL were analyzed. Positive 52 53 ANCA and aPL were reported to be non-specific in patients with infection.5-7 In this 54 55 56 study, either positive myeloperoxidase(MPO)-ANCA or proteinase 3(PR3)-ANCA 57 58 59 patients were defined as ANCA-positive patients. Patients with any positive 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 antibodies of anticardiolipin antibodies (ACL), lupus anticoagulant (LAC) or anti-β2 5 6 7 glycoprotein I antibodies (anti-β2GPI) were defined as aPL-positive patients. Finally, 8 9 89 patients were classified as ‘ANCA-positive IE’ or ‘ANCA-negative IE’ and 53 10 11 12 patients as ‘aPL-positive IE’ or ‘aPL-negative IE’ according to their antibodies test 13 14 (Supplemental Figure 1). Then, the survival rate of sub groups of positive or negative 15 16 17 ANCA and aPL were analyzed as well. 38 patients were tested with both ANCA and 18 For peer review only 19 20 aPL, and the characteristics were analyzed as well. 21 22 23 24 25 Laboratory features and echocardiography 26 27 Levels of anti-PR3, anti-MPO and aPL antibodies in serum were measured by enzyme 28 29 30 linked immunosorbent assay (ELISA). The following laboratory data were recorded: 31 32 33 white blood cell counts in blood (WBC), hemoglobin (Hb), platelet (PLT), C-reactive 34 35 protein (CRP), erythrocyte sedimentation rate (ESR), serum lactate dehydrogenase 36

37 http://bmjopen.bmj.com/ 38 (LDH), hematuria, proteinuria, rheumatoid factor (RF), immunoglobulin G (IgG), 39 40 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), 41 42 43 anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 44

45 on September 28, 2021 by guest. Protected copyright. 46 antibodies (anti-β2GPI), anti-PR3 antibody (PR3-ANCA), anti-MPO antibody 47 48 (MPO-ANCA), anti-double stranded DNA antibody (anti-dsDNA). All IE patients 49 50 51 underwent transthoracic or transesophageal echocardiography. Vegetation or new 52 53 valvular regurgitation and other abnormalities of a valve, were classified according to 54 55 56 modified Duke criterion for IE.3 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 Statistical analysis 5 6 7 Data was analyzed using the Statistical Package for the Social Sciences for Windows 8 9 (version 23.0; SPSS, IBM, Chicago, IL, USA). Statistical analysis was performed by t 10 11 12 test or Chi-square test according to the type of continuous data or categorical data. 13 14 Survival curves were obtained using Kaplan-Meier’s method. The significance was 15 16 17 obtained by a log-rank test. P value of less than 0.05 was considered statistically 18 For peer review only 19 20 significant. 21 22 23 24 25 Patient and Public Involvement 26 27 The clinical and laboratory data of the patients were mainly obtained from the 28 29 30 medical records system of Ruijin hospital and follow-up phone calls from 1997-2017. 31 32 33 The study was a retrospective study, we continuously involved patients hospitalized in 34 35 Ruijin Hospital affiliated to Shanghai Jiao Tong University School of Medicine, 36

37 http://bmjopen.bmj.com/ 38 diagnosed with IE from 1997 to 2017 according to the modified Duke criteria. The 39 40 clinical and laboratory data was collected from the system of the hospital. We asked 41 42 43 the research questions and outcome measures of these patients by telephone. Patients 44

45 on September 28, 2021 by guest. Protected copyright. 46 were not involved in the design, recruitment to and conduct of the study. If this paper 47 48 is published, the paper version will disseminated to study participants by mail. 49 50 51 52 53 54 55 56 Results 57 58 59 Patient characteristics 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 As shown in Table 1, a total of 432 patients consisted of 278 (64.35%) male patients 5 6 7 and 154 (35.65%) female patients (male:female ratio=1.8:1) who met the modified 8 9 Duke criteria were analyzed. The mean age ± standard deviation (SD) of patients was 10 11 12 45.61±16.12 years. The mean duration of disease was 4.34±12.12 months. The mitral 13 14 valve (54.17%) was the mostly involved valve followed by the aortic (42.82%) valve. 15 16 17 The most common complications of IE patients were pulmonary arterial hypertension 18 For peer review only 19 20 (PAH) (32.64%), hypertension (18.98%) and congenital heart disease (15.28%). As 21 22 for thrombosis discovered or presented in hospital, brain (17.36%) and spleen (6.02%) 23 24 25 infarction were most common. For predisposing factors, dialysis accounted for 1.39%, 26 27 0.93% with dental procedure, and 0.46% with intravenous drug use (IVDU). In the 28 29 30 end, 346 patients (80.09%) had cardiac surgery and 55 patients (12.73%) died in 31 32 33 Ruijin hospital. 34 35 36

37 http://bmjopen.bmj.com/ 38 Clinical manifestations and laboratory features 39 40 Fever (84.26%) and new heart murmur (79.4%) were presented in most patients of IE 41 42 43 (Supplemental Table 1). The most common rheumatic manifestations were serositis 44

45 on September 28, 2021 by guest. Protected copyright. 46 (39.35%), arthritis (15.05%), and myalgia (6.48%), which might cause misdiagnosis 47 48 into rheumatic diseases. In addition, a few patients had non-specific manifestations 49 50 51 such as heptomegaly (4.4%), splenomegaly (22.92%) and fatigue (19.91%). Eighteen 52 53 patients had Janeway lesions (4.17%), 11 patients with Osler nodes (2.55%) and 5 54 55 56 patients with Roth spot (1.16%). 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 Blood cultures were positive in 50% of IE patients. As for immunologic features, 89 5 6 7 patients were tested for ANCA with 21 (23.6%) ANCA-positive. It is interesting that 8 9 all these 21 ANCA-positive patients were PR3-ANCA positive. Fifty-three patients 10 11 12 were tested for aPL and 21 was positive (45.28%). One patients had positive IgG aPL 13 14 (1.89%), 4 patients with IgM aPL (7.55%), 1 patients with IgA aPL (1.89%), 2 15 16 17 patients with anti-β2GPI (3.77%) and 17 patients with LAC (32.08%). In total, 17 18 For peer review only 19 20 patients were single aPL positive, 2 patients were double aPL positive and none was 21 22 tripple aPL positive. Eight (12.70%) patients had positive ANA, one had positive 23 24 25 ENA (1.59%). For urine tests, proteinuria (25.23%) and hematuria (41.90%) were 26 27 most commonly seen. Of 38 IE patients tested with both aPL and ANCA, 15 were 28 29 30 only aPL positive, 9 were only ANCA positive and 3 were both positive (Table 4). 31 32 33 34 35 Comparison of patients with positive and negative ANCA or APL 36

37 http://bmjopen.bmj.com/ 38 Of 432 IE patients, 104 patients who were either tested ANCA or aPL，were divided 39 40 into different groups by the presentation of antibody and analyzed. Of 89 patients who 41 42 43 was tested for ANCA, 21 (23.6%) were ANCA-positive and 68 (76.4%) were 44

45 on September 28, 2021 by guest. Protected copyright. 46 ANCA-negative. Clinical manifestations such as fever, fatigue, thrombosis, myalgia, 47 48 splenomegaly, heart murmur, and weight loss were not significantly different between 49 50 51 the two groups (P˃0.05). But arthritis (P=0.003) was more frequent, and serositis 52 53 (P=0.06) seems to be more common in ANCA-positive patients than in 54 55 56 ANCA-negative group (Table 2). Compared to ANCA-negative patients, IgM 57 58 59 (P=0.048) was higher, Hb (P=0.001) was significantly lower and elevated RF 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 (P=0.053) seems to be more common in ANCA-positive patients (Table 3). Of 53 5 6 7 patients who were tested for aPL, 21 (39.62%) were aPL-positive, the remaining 32 8 9 (60.38%) were aPL-negative. IE patients didn’t differ between the items listed as 10 11 12 clinical manifestations. However, it was interesting to find that ESR was significantly 13 14 higher in patients who were positive for aPL (P=0.003) (Table 2-3). 15 16 17 18 For peer review only 19 20 Microbiology and treatment 21 22 Streptococci (25.46%) accounted for the majority of pathogen followed by 23 24 25 staphylococcus aureus (14.81%) and staphylococcus epidermidis (6.94%). Other 26 27 common types included enterococci (6.02%), enterobacter (4.63%) and other bacillus 28 29 30 (4.17%). Furthermore, commonly used antibiotics were vancomycin (52.77%), 31 32 33 cephalosporin (37.73%), carbopenems (25.69%), penicillin (25.23%) and Quinolones 34 35 (21.99%). 36

37 http://bmjopen.bmj.com/ 38 39 40 Outcomes 41 42 43 Overall, 9 patients tested for aPL or ANCA antibodies died in hospital. 4 patients 44

45 on September 28, 2021 by guest. Protected copyright. 46 were ANCA positive and 2 patients were aPL positive. One patient died of renal 47 48 failure, four patients of acute heart failure, two of septic shock and two of stroke. 49 50 51 During the 3-month observation period from the day of diagnosis, no patient died 52 53 after discharged from hospital. The survival rate was significantly lower in 54 55 56 ANCA-positive IE (P =0.032), and there was no significant difference between the 57 58 59 aPL groups (P = 0.728) (Fig. 1). 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 6 7 Discussion 8 9 The clinical presentation and laboratory findings of IE have overlaps with rheumatic 10 11 12 diseases, making it difficult to diagnose. Thus, it is pivotal to analyse rheumatic 13 14 manifestations of IE so as to help physicians differentiate it when encountering these 15 16 17 clinical scenarios. In this study, we analyzed clinical and laboratory features of 432 IE 18 For peer review only 19 20 patients in Ruijin Hospital and compared the characteristics of 104 patients tested for 21 22 ANCA or aPL during the past two decades. Our study showed that ANCA-positive IE 23 24 25 patients had higher IgM, lower Hb and more presentation of arthritis, seemed to have 26 27 more serositis than ANCA-negative IE patients. Moreover, aPL-positive IE patients 28 29 30 had significantly higher ESR than aPL-negative group. 31 32 33 34 35 Talking about blood culture, in a university hospital in northern Italy, staphylococcus 36

37 http://bmjopen.bmj.com/ 38 aureus has become the leading cause of IE.11 However, data from China was different. 39 40 In one study, streptococcus was found in 61.9% of patients and glycopeptide and 41 42 43 cephalosporins were the most frequently used antibiotics.12 In our study, the most 44

45 on September 28, 2021 by guest. Protected copyright. 46 common microbe of IE was streptococcus accounting for 25.46%, and the second 47 48 common one was staphylococcus aureus accounting for 14.81%. In terms of used 49 50 51 antibiotics, 52.77% was vancomycin and 37.73% was cephalosporin, being the two 52 53 leading treatments. The use of antibiotics in our hospital was decided according to 54 55 56 drug sensitivity tests. 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 The phenomenon of IE patients with positive ANCA has long been reported. Alfred 5 6 7 Mahr et al reported that 8% IE patients had PR3-ANCA or MPO-ANCA. It was 8 9 associated with young age (P=0.022), echocardiographic vegetations (P=0.043), and 10 11 12 elevated IgG levels (P=0.017).13 It has also been reported that most gram-positive 13 14 endocarditis could cause an anti-PR3 specificity. Possibly during delayed 15 16 17 polymorphonuclear leucocyte apoptosis, PR3 contact with the immune system and 18 For peer review only 19 14 20 caused increased production of anti-PR3 ANCA. It was also indicated that PR3 was 21 22 a common neutrophil serine protease. In azurophillic granules, it has a mean 23 24 25 concentration of 13.4 mM.15 The function of PR3 include degradation of extracellular 26 27 matrix, platelet activation, cleavage of inflammatory mediators and regulation of 28 29 30 granulopoiesis.16 It was interesting to find that in our IE patients, all of them were 31 32 33 PR3 ANCA positive. One important presentation of ANCA related vasculitis is 34 35 glomerular nephritis. But IE-associated glomerular nephritis often presents with acute 36

37 http://bmjopen.bmj.com/ 38 kidney injury. The most common biopsy finding was necrotizing and crescentic 39 40 glomerular nephritis (53%), followed by endocapillary proliferative glomerular 41 42 43 nephritis (37%).17 In our study, IE patients who had proteinuria and hematuria 44

45 on September 28, 2021 by guest. Protected copyright. 46 accounted for 25.2% and 41.9%, respectively. However, kidney biopsy was not 47 48 routinely done in these patients. 49 50 51 52 53 Apart from ANCA, aPL was also highly paid attention to. In terms of the relationship 54 55 56 between aPL and IE, it has been indicated that IgG aPL might be associated with Q 57 58 59 Fever Endocarditis and antiphospholipid antibodies were associated with thrombosis 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 18-19 4 during acute Q fever. Furthermore, the presence of IgM aPL and anti-β2GPI was 5 6 7 associated with embolic events, particularly cerebral ones, which could contribute to 8 9 assess the embolic risk of IE.20 In our study, APL-positive IE patients had higher 10 11 12 ESR, and LAC was the most commonly positive type of aPL in IE patients. This 13 14 meant that aPL was related to an inflammatory status in vivo. 15 16 17 18 For peer review only 19 20 As for the outcomes, in our study, the survival rate was significantly lower in 21 22 ANCA-positive IE than ANCA-negative group and there was no difference between 23 24 25 the aPL groups, indicating that ANCA positive might be related to mortality. 26 27 28 29 30 There were several limitations of this study. First, it was retrospective. The 31 32 33 measurement of autoantibodies were not conducted in all IE patients when they were 34 35 diagnosed, with only 89 with ANCA and 53 with aPL respectively, and only 38 36

37 http://bmjopen.bmj.com/ 38 patients were both tested with ANCA and aPL. Second, the time of study spanned 39 40 quite long and some patients lost follow ups and the measurement of autoantibodies 41 42 43 could not be done again. 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 Conclusions 49 50 51 The present study supports the hypothesis that rheumatic complications are frequent 52 53 in patients with IE. For rheumatologists, it’s indispensable to routinely exclude IE 54 55 56 from patients with atypical rheumatic features such as arthritis, serositis, 57 58 59 splenomegaly, myalgia, proteinuria and hematuria. For cardiologists, more attention 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 should be paid to the measurement of autoantibodies in IE patients, which might 5 6 7 contribute to the evaluation of prognosis. 8 9 10 11 12 13 14 15 16 17 References 18 For peer review only 19 20 1. Fedeli U, Schievano E, Buonfrate D, et al. Increasing incidence and mortality of infective 21 22 endocarditis: a population-based study through a record-linkage system. BMC Infect Dis 23 24 25 2011;11:48. 26 27 2. Monteiro TS, Correia MG, Golebiovski WF, et al. Asymptomatic and symptomatic embolic 28 29 30 events in infective endocarditis: associated factors and clinical impact. Braz J Infect Dis 31 32 33 2017;21:240-247. 34 35 3. Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC Guidelines for the management of 36

37 http://bmjopen.bmj.com/ 38 infective endocarditis: The Task Force for the Management of Infective Endocarditis of the 39 40 European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic 41 42 43 Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J 44

45 on September 28, 2021 by guest. Protected copyright. 46 2015;36:3075-3128. 47 48 4. Misra DP, Chowdury AC, Edavalath S, et al. Endocarditis: the great mimic of rheumatic 49 50 51 diseases. Trop Doct 2016;46:180-186. 52 53 5. Peng H, Chen WF, Wu C, et al. Culture-negative subacute bacterial endocarditis masquerades 54 55 56 as granulomatosis with polyangiitis (Wegener's granulomatosis) involving both the kidney and 57 58 59 lung. BMC Nephrol 2012;13:174. 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 6. Langlois V, Lesourd A, Girszyn N, et al. Antineutrophil Cytoplasmic Antibodies Associated 5 6 7 With Infective Endocarditis. Medicine (Baltimore) 2016;95:e2564. 8 9 7. Kupferwasser LI, Hafner G, Mohr-Kahaly S, et al. The presence of infection-related 10 11 12 antiphospholipid antibodies in infective endocarditis determines a major risk factor for embolic 13 14 events. J Am Coll Cardiol 1999;33:1365-71. 15 16 17 8. Demircan F, Akbulut S, Yavuz R, et al. The effect of laser epilation on recurrence and 18 For peer review only 19 20 satisfaction in patients with sacrococcygeal pilonidal disease: a prospective randomized controlled 21 22 trial. Int J Clin Exp Med 2015;8:2929-33. 23 24 25 9. El-Chakhtoura N, Yasmin M, Kanj SS, et al. A 27-year experience with infective endocarditis 26 27 in Lebanon. J Infect Public Health 2017;10:734-739. 28 29 30 10. Ying CM, Yao DT, Ding HH, et al. Infective endocarditis with antineutrophil cytoplasmic 31 32 33 antibody: report of 13 cases and literature review. PLoS One 2014;9:e89777. 34 35 11. Ferraris L, Milazzo L, Ricaboni D, et al. Profile of infective endocarditis observed from 2003 - 36

37 http://bmjopen.bmj.com/ 38 2010 in a single center in Italy. BMC Infect Dis 2013;13:545. 39 40 12. Xu H, Cai S, Dai H. Characteristics of Infective Endocarditis in a Tertiary Hospital in East 41 42 43 China. PLoS One 2016;11:e0166764. 44

45 on September 28, 2021 by guest. Protected copyright. 46 13. Mahr A, Batteux F, Tubiana S, et al. Brief report: prevalence of antineutrophil cytoplasmic 47 48 antibodies in infective endocarditis. Arthritis Rheumatol 2014;66:1672-7. 49 50 51 14. Aslangul E, Goulvestre C, Mallat Z, et al. Human bartonella infective endocarditis is 52 53 associated with high frequency of antiproteinase 3 antibodies. J Rheumatol 2014;41:408-10. 54 55 56 15. Campbell EJ1, Campbell MA, Owen CA. Bioactive proteinase 3 on the cell surface of human 57 58 59 neutrophils: quantification, catalytic activity, and susceptibility to inhibition. J Immunol 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 23 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 2000;165:3366-74. 5 6 7 16. Yang TY, Zhou WJ, Du Y, et al. Role of saliva proteinase 3 in dental caries. Int J Oral Sci 8 9 2015;7:174-8. 10 11 12 17. Boils CL, Nasr SH, Walker PD, et al. Update on endocarditis-associated glomerulonephritis. 13 14 Kidney Int 2015;87:1241-9. 15 16 17 18. Million M, Walter G, Bardin N, et al. Immunoglobulin G anticardiolipin antibodies and 18 For peer review only 19 20 progression to Q fever endocarditis. Clin Infect Dis 2013;57:57-64. 21 22 19. Million M, Bardin N, Bessis S, et al. Thrombosis and antiphospholipid antibody syndrome 23 24 25 during acute Q fever: A cross-sectional study. Medicine (Baltimore) 2017;96:e7578. 26 27 20. Selton-Suty C, Maigrat CH, Devignes J, et al. Possible relationship between antiphospholipid 28 29 30 antibodies and embolic events in infective endocarditis. Heart 2018;104:509-516. 31 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 Contribution 41 42 43 For authors, Jialin Teng, Honglei Liu, Xiaobing Cheng, Huihui Chi and Fan Wang 44

45 on September 28, 2021 by guest. Protected copyright. 46 collected the clinical data. Yutong Su and Yue Sun analyzed the data. Zhuochao Zhou 47 48 and Junna Ye wrote the manuscript. Wei Jin and Chengde Yang designed the study 49 50 51 and critically read the manuscript. We would like to thank the patients for their 52 53 participation in this study. 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 6 Table 1 7 General characteristics of infective endocarditis (IE) patients. 8 9 Item N (%) 10 N 432 11 Female/male 154(35.65)/278(64.35) 12 13 Age (years, mean±SD) 45.61±16.12 14 Duration (months, mean±SD) 4.34±12.12 15 Predisposing factors 16 Dialysis 6(1.39) 17 18 DentalFor procedure peer review 4(0.93)only 19 Intravenous drug use 2(0.46) 20 Co-morbidities 21 22 Pulmonary arterial hypertension 141(32.64) 23 Hypertension 82(18.98) 24 Congenital heart disease 66(15.28) 25 26 Diabetes 27(6.25) 27 Cancer 11(2.55) 28 Thrombosis 29 Brain 75(17.36) 30 31 Spleen 26(6.02) 32 Kidney 8(1.85) 33 extremity 8(1.85) 34 35 Coronary 4(0.92) 36 Lung 4(0.92)

37 Infected valve http://bmjopen.bmj.com/ 38 39 Prosthetic valve 42(9.72) 40 Native valve 41 Mitral valve 234(54.17) 42 Aortic valve 185(42.82) 43 44 Tricuspid valve 17(3.94)

45 Pulmonary valve 10(2.31) on September 28, 2021 by guest. Protected copyright. 46 Outcome 47 48 Cardiac surgery 346(80.09) 49 Relapse 13(3.01) 50 In-hospital Death 55(12.73) 51 52 Note: standard deviation (SD). 53 54 Table 2 55 56 Clinical features of patients positive or negative for ANCA and aPL. 57 ANCA aPL 58 ANCA(+) ANCA(-) p valve aPL(+) aPL(-) p valve 59 60 N 21 68 / 21 32 /

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 23 BMJ Open

1 2

3 Positive Blood 13(61.9) 37(57.8) 0.741 10(50.0) 17(53.1) 0.826 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 culture 6 Fever 19(90.5) 57(83.8) 0.688 18(85.7) 23(71.9) 0.400 7 Arthritis 11(52.4) 13(19.1) 0.003 5(23.8) 8(25.0) 0.922 8 Fatigue 3(14.3) 16(23.5) 0.549 7(33.3) 9(28.1) 0.686 9 10 Serositis 8(80.0) 23(41.8) 0.060 13(61.9) 15(46.9) 0.284 11 Thrombosis 3(14.3) 17(25.0) 0.466 8(38.1) 10(31.3) 0.607 12 Myalgia 0 6(10.9) 0.672 2(9.5) 2(6.5) 1.000 13 14 Splenomegaly 6(28.6) 20(29.4) 0.941 5(23.8) 6(18.8) 0.922 15 Heart murmur 10(100.0) 43(78.2) 0.233 18(85.7) 28(87.5) 1.000 16 Weight loss 3(30.0) 9(16.4) 0.562 3(14.3) 7(21.9) 0.740 17 18 Cardiac surgery For5(50.0) peer42(72.4) review0.295 only16(76.2) 19(59.4) 0.206 19 In-hospital Death 4(19.0) 3(4.4) 0.087 2(9.5) 4(12.5) 1.000 20 Note: antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 21 22 23 Table 3 24 Laboratory features of patients positive or negative for ANCA and aPL. 25 26 ANCA aPL 27 ANCA(+) ANCA(-) p value aPL(+) aPL(-) p value 28 N 21 68 / 21 32 / 29 30 WBC (*109/L) 8.39±2.53 8.56±4.19 0.896 7.66±2.54 8.58±4.52 0.403 31 Hb (g/L) 87.57±26.83 105.51±18.76 0.001 99.57±15.28 102.34±26.22 0.664 32 PLT (*109/L) 194.70±110.91 201.25±127.06 0.879 184.95±83.86 193.47±151.31 0.819 33 34 LDH (IU/L) 206.125±53.45 252.00±96.73 0.200 278.87±296.11 247.72±103.72 0.632 35 ESR (mm/h) 64.76±36.87 51.85±33.47 0.135 69.14±34.82 39.09±33.48 0.003 36 IgG (mg/dl) 2254.29±1001.40 1660.05±528.67 0.172 1959.69±747.29 1751.67±707.30 0.379

37 http://bmjopen.bmj.com/ 38 IgA (mg/dl) 232.29±89.77 301.50±131.20 0.188 276.63±94.03 283.38±130.21 0.859 39 IgM (mg/dl) 312.71±183.51 140.70±64.03 0.048 193.07±119.30 153.79±119.39 0.324 40 IgE (mg/dl) 37.00±22.39 354.85±569.90 0.286 357.31±437.81 351.73±635.24 0.983 41 ACL positive 2(25.0) 2(5.3) 0.436 5(23.8) 0 0.004 42 43 aβ2GPI positive 0 1(4.8) 1.000 2(12.5) 0 0.057 44 LAC positive 2(28.6) 11(37.9) 0.981 17(81.0) 0 0.000

45 ANA positive 2(15.4) 6(11.8) 1.000 3(15.8) 3(10.7) 0.618 on September 28, 2021 by guest. Protected copyright. 46 47 Elevated CRP 18(90.0) 51(85.0) 0.851 14(73.7) 21(75.0) 0.921 48 Elevated RF 10(58.8) 14(31.8) 0.053 6(30.8) 6(25.0) 0.336 49 Note: White blood cell counts (WBC), hemoglobin (Hb), platelet (PLT), erythrocyte 50 51 sedimentation rate (ESR), serum lactate dehydrogenase (LDH), immunoglobulin G (IgG), 52 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), anticardiolipin 53 antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I antibodies (anti-β2GPI), 54 55 antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL).. 56 57 Table 4 58 Antibodies of patients tested with both aPL and ANCA. 59 60 N

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from Only aPL positive 15 4 5 ACL positive 1(IgM ACL) 6 aβ2GP1 positive 1 7 LAC positive 10 8 9 ACL+LAC 3(2 IgM ACL +1 IgA ACL) 10 Only ANCA positive 9 11 Both aPL and ANCA positive 3(LAC, LAC+IgM ACL, IgM ACL) 12 13 Note: anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 14 antibodies (anti-β2GPI), antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody 15 (aPL). 16 17 18 For peer review only 19 Figure 1 20 Kaplan-Meier’s survival curves. 21 22 a. Kaplan-Meier’s survival curves of aPL tested patients; b. Kaplan-Meier’s survival curves of 23 ANCA tested patients. 24 25 26 27 28 29 30 31 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 23 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40 90x90mm (300 x 300 DPI) 41 on September 28, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 23

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from Supplemental Table 1 4 5 Clinical manifestations of infective endocarditis patients. 6 Item N (%) 7 Heart murmur 343(79.40) 8 9 Fever 364(84.26) 10 Arthritis 65(15.05) 11 Fatigue 86(19.91) 12 13 Serositis 170(39.35) 14 Myalgia 28(6.48) 15 Hepatomegaly 19(4.40) 16 17 Splenomegaly 99(22.92) 18 PeripheralFor polyneuropathy peer review 4(0.93)only 19 Janeway lesion 18(4.17) 20 Osler node 11(2.55) 21 22 Roth spot 5(1.16) 23 24

25 26 27 Supplemental Figure 1 28 Distribution of patients diagnosed with infectious endocarditis. 29 30 31 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 * Patients who were either tested with aPL or ANCA were included and compared in different 48 groups. 49 Note: infectious endocarditis (IE), antineutrophil cytoplasmic antibody (ANCA), 50 anti-phospholipid antibody (aPL). 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

Relatively frequent rheumatic manifestations and autoantibodies in infective endocarditis: a 20-year retrospective study in China ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-031512.R1

Article Type: Original research

Date Submitted by the 29-Aug-2019 Author:

Hospital, Rheumatology and Immunology http://bmjopen.bmj.com/ Su, Yutong; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Chi, Huihui; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Wang, Fan; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Yang, Chengde; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Jin, Wei; Shanghai Jiao Tong University Medical School Affiliated Ruijin on September 28, 2021 by guest. Protected copyright. Hospital, Cardiology

Primary Subject Rheumatology Heading:

Secondary Subject Heading: Rheumatology

Infective endocarditis, Rheumatic manifestation, ANCA, Antiphospholipid Keywords: antibody

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 25 BMJ Open

1 2

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 1 Relatively frequent rheumatic manifestations and 5 6 7 2 autoantibodies in infective endocarditis: a 20-year 8 9 3 10 retrospective study in China 11 12 4 13 14 5 Zhuochao Zhou1,*, Junna Ye1,*, Jialin Teng1, Honglei Liu1, Xiaobing Cheng1, Yue 15 16 17 6 Sun1, Yutong Su1, Huihui Chi1, Fan Wang1, Chengde Yang1,#, Wei Jin2,# 18 For peer review only 19 20 7 21 22 8 1Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 23 24 25 9 University School of Medicine; 26 27 10 2Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School 28 29 30 11 of Medicine 31 32 33 12 34 35 13 *These authors contributed equally to this work. 36

37 http://bmjopen.bmj.com/ 38 14 39 40 15 #Correspondence: 41 42 43 16 Wei Jin, MD, PhD 44

45 on September 28, 2021 by guest. Protected copyright. 46 17 Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School of 47 48 18 Medicine, No. 197 Ruijin Second Road, Huangpu District, Shanghai 200025, China. 49 50 51 19 Tel.: +86 21 64370045; 52 53 20 Fax: +86 21 34186000; 54 55 56 21 E-mail: [email protected] 57 58 59 22 60 1

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 23 Chengde Yang, MD, PhD 5 6 7 24 Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 8 9 25 University School of Medicine, No. 197 Ruijin Second Road, Huangpu District, 10 11 12 26 Shanghai 200025, China. 13 14 27 Tel.: +86 21 64370045; 15 16 17 28 Fax: +86 21 34186000; 18 For peer review only 19 20 29 E-mail: [email protected] 21 22 30 23 24 25 31 26 27 32 28 29 30 33 Abstract 31 32 33 34 Objective: This study aimed to characterize rheumatic manifestations and 34 35 35 autoantibodies of 432 patients diagnosed with infective endocarditis (IE). 36

37 http://bmjopen.bmj.com/ 38 36 Design, setting and participants: A retrospective study was conducted in Ruijin 39 40 37 Hospital from 1997 to 2017. The clinical and laboratory characteristics of a total of 41 42 43 38 432 patients were analyzed. Besides, the differences between patients with positive 44

45 on September 28, 2021 by guest. Protected copyright. 46 39 and negative antineutrophil cytoplasmic antibody (ANCA) and anti-phospholipid 47 48 40 antibody (aPL), as well as survival rate of these patients, were compared. 49 50 51 41 Results: 432 patients with 278 male patients and 154 female patients were included. 52 53 42 The mean age of patients was 45.61±16.12 years. Until 2017, the end of the study, 54 55 56 43 346 patients (80%) had cardiac surgery and 55 patients (13%) died in the hospital. 57 58 59 44 The most common rheumatic manifestations were serositis (39%), arthritis (15%), and 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 45 myalgia. Among the IE patients, 104 of them were either tested for ANCA or aPL, 5 6 7 46 and were analyzed in different groups. 21 (24%) positive ANCA patients were all of 8 9 47 proteinase 3(PR3)-ANCA. Compared with the ANCA-negative group, patients with 10 11 12 48 positive ANCA had higher IgM (P=0.048), lower hemoglobin (P=0.001), and more 13 14 49 presentation of arthritis (P=0.003). 21 (40%) aPL-positive patients had a higher level 15 16 17 50 of erythrocyte sedimentation rate (ESR) compared to aPL-negative group (P=0.003). 18 For peer review only 19 20 51 In addition, survival rate of ANCA-positive IE patients was lower (P=0.03) than 21 22 52 ANCA-negative group, while there was no difference in patients with or without aPL 23 24 25 53 antibody (P=0.728). 26 27 54 Conclusion: The present study supports the claim that rheumatic manifestations and 28 29 30 55 presentation of antibodies are frequent in patients with IE and lead to early 31 32 33 56 misdiagnosis. Physicians should pay more attention to the measurement of 34 35 57 autoantibodies in these patients. 36

37 http://bmjopen.bmj.com/ 38 58 39 40 59 Keywords: infective endocarditis, rheumatic manifestation, ANCA, aPL 41 42 43 60 44

45 on September 28, 2021 by guest. Protected copyright. 46 61 47 48 62 49 50 51 52 63 Strengths and limitations of this study 53 54 55 64 1. The study analyzed the differences and survival rate between patients with and 56 57 65 without ANCA or aPL antibodies in IE patients. 58 59 60 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 66 2. In patients who were tested for both ANCA and aPL, positive lupus anticoagulant 5 6 7 67 (LAC) and proteinase 3(PR3)-ANCA were the most common types. 8 9 68 3. Rheumatic manifestations and presentation of antibodies are frequent in patients 10 11 12 69 with IE and lead to early misdiagnosis. 13 14 70 4. As it was a retrospective study, not all patients detected ANCA or aPL 15 16 17 71 autoantibodies in hospital. 18 For peer review only 19 20 72 21 22 23 73 Data availability statement 24 25 26 27 74 All data relevant to the study are included in the article or uploaded as supplementary 28 29 30 75 information. 31 32 33 34 76 35 36

37 77 Declaration of conflicting interests http://bmjopen.bmj.com/ 38 39 40 78 The authors report no relationships that could be construed as a conflict of interest. 41 42 43 79 44

45 80 Funding on September 28, 2021 by guest. Protected copyright. 46 47 48 81 This work is supported by National Natural Science Foundation of China 49 50 82 (81871272, 81801592, 81370397, 81670266), Science and Technology Commission 51 52 53 83 of Shanghai Municipality (17140902500), Shanghai Sailing Program (18YF1414100), 54 55 Shanghai Jiao Tong University Interdisciplinary Research Project (YG2016QN60), 56 84 57 58 85 Excellent Youth B Project (GCQN-2017-B05) and Innovative research team of 59 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 86 high-level local universities in Shanghai. 5 6 7 87 8 9 88 10 11 12 89 13 14 15 90 Introduction 16 17 91 Infective endocarditis (IE) is a microbial infection of the endocardium and usually 18 For peer review only 19 20 92 influences the heart valves. Despite being relatively rare worldwide, IE has been 21 22 93 reported by a steady incidence over the past three decades ranging from 2~6 per 23 24 25 94 100,000 individuals in the general population per year, with a considerable 26 27 95 associated mortality varying from 10% to 30%.1 Predisposing factors of IE include 28 29 30 96 predisposing conditions such as congenital heart diseases, rheumatic valve disease, 31 32 33 97 artificial valves, and predisposing procedures like intravenous drug use, dental 34 35 98 procedure and hemodialysis.2 Diagnosis of IE is made based on symptoms, blood 36

37 http://bmjopen.bmj.com/ 38 99 cultures, echocardiography and pathological biopsy from valve surgery.3 39 40 100 However, superantigens induced by some bacteria like staphylococcus aureus can 41 42 43 101 stimulate the immune response, which could be interfering with antibody 44

45 4 on September 28, 2021 by guest. Protected copyright. 46 102 production. Rheumatic manifestations such as myalgia, arthralgia, and arthritis 47 48 103 are prevalent in nearly 40% of patients with IE at presentation, weeks to months 49 50 51 104 before diagnosis of IE.5 When classic IE manifestations are less evident, it could 52 53 105 cause misdiagnosis, which might lead to delayed initiation of antibiotic treatment. 54 55 56 106 Thus, the distinction between IE and rheumatic diseases is not crystal clear, and to 57 58 59 107 understand the difference is of great importance. 60 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 108 5 6 7 109 Many specific antibodies like antineutrophil cytoplasmic antibodies (ANCA) and 8 9 110 anti-phospholipid antibody (aPL), might be related with the pathophysiology of IE. 6-7 10 11 12 111 Due to positive ANCA tests, IE has been reported to mimic ANCA-associated 13 14 112 vasculitis. Patients with IE may present with multiple pulmonary nodules and 15 16 17 113 glomerulonephritis, mimicking granulomatosis with polyangiitis.8 A study further 18 For peer review only 19 6 20 114 underscored that ANCA might be associated with multiple valve involvement. 21 22 115 Additionally, infection-associated elevated aPL levels in patients with infective 23 24 25 116 endocarditis are related to endothelial cell activation, thrombin generation and 26 27 117 impairment of fibrinolysis, which may contribute to the increased risk for major 28 29 30 118 embolic events in these patients.7 Therefore, ANCA and aPL are non specific in IE 31 32 33 119 patients. 34 35 120 36

37 http://bmjopen.bmj.com/ 38 121 The objective of this study is to make the profile of clinical and laboratory 39 40 122 manifestations of 432 IE patients in a tertiary hospital in China from 1997 to 2017, 41 42 43 123 and analyze the characteristics of IE patients tested with ANCA or aPL. The 44

45 on September 28, 2021 by guest. Protected copyright. 46 124 characteristics of subgroups of positive or negative ANCA and aPL were studied as 47 48 125 well as survival rate of these IE patients. 49 50 51 126 52 53

54 127 Methods 55 56 128 Patient and Public Involvement 57 58 59 129 The study was a retrospective study, we continuously involved 432 patients 60 6

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 130 hospitalized in Ruijin Hospital affiliated to Shanghai Jiao Tong University School of 5 6 7 131 Medicine, diagnosed with IE from 1997 to 2017. The clinical and laboratory data of 8 9 132 the patients were mainly obtained from the medical records system of Ruijin hospital 10 11 12 133 and follow-up phone calls. Patients were not involved in the design, recruitment to 13 14 134 and conduct of the study. 15 16 17 135 18 For peer review only 19 20 136 Only patients with definite infective endocarditis, either by the Duke criteria (up to 21 22 137 2000) or the modified Duke criteria (from 2000 onwards) were included.3,9 We had 23 24 25 138 ruled out 33 patients with primary rheumatic disease or in immunosuppression 26 27 139 condition. Duration of disease means the period from the time patients presented first 28 29 30 140 clinical feature to definite diagnosis. Cardiac surgery and death were recorded during 31 32 33 141 the period of hospitalization. Three blood samples were taken from a peripheral vein 34 35 142 at 30 minutes intervals once the patient had fever of over 38.5℃ at the same day for 36

37 http://bmjopen.bmj.com/ 38 143 higher microorganism positivity, and were incubated in both aerobic and anaerobic 39 40 144 bottles. The study followed the ethical standards for human experimentation 41 42 43 145 established in the Declaration of Helsinki and was approved by the Institutional 44

45 on September 28, 2021 by guest. Protected copyright. 46 146 Research Ethics Committee of Ruijin Hospital (ID: 2016-62), Shanghai, China. 47 48 147 49 50 51 148 In this study, either positive myeloperoxidase(MPO)-ANCA or proteinase 52 53 149 3(PR3)-ANCA patients were defined as ANCA-positive patients. Patients with any 54 55 56 150 positive antibodies of anticardiolipin antibodies (ACL), lupus anticoagulant (LAC) or 57 58 59 151 anti-β2 glycoprotein I antibodies (anti-β2GPI) were defined as aPL-positive patients. 60 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 152 In-hospital mortality means diagnosed IE patients in our study who were dead during 5 6 7 153 hospitalization. There was a 3-month observation period for patient mortality from the 8 9 154 day of diagnosis. The survival rate of subgroups of positive or negative ANCA and 10 11 12 155 aPL were analyzed. 13 14 156 15 16 17 157 Laboratory features and echocardiography 18 For peer review only 19 20 158 Levels of anti-PR3, anti-MPO and aPL antibodies in serum were measured by enzyme 21 22 159 linked immunosorbent assay (ELISA). The following laboratory data were recorded: 23 24 25 160 white blood cell counts in blood (WBC), hemoglobin (Hb), platelet (PLT), C-reactive 26 27 161 protein (CRP), erythrocyte sedimentation rate (ESR), serum lactate dehydrogenase 28 29 30 162 (LDH), rheumatoid factor (RF), immunoglobulin G (IgG), immunoglobulin A (IgA), 31 32 33 163 immunoglobulin M (IgM), antinuclear antibody (ANA), anticardiolipin antibodies 34 35 164 (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I antibodies (anti-β2GPI), 36

37 http://bmjopen.bmj.com/ 38 165 anti-PR3 antibody (PR3-ANCA), anti-MPO antibody (MPO-ANCA), hematuria, 39 40 166 proteinuria. All IE patients underwent transthoracic or transesophageal 41 42 43 167 echocardiography. 44

45 on September 28, 2021 by guest. Protected copyright. 46 168 47 48 169 Statistical analysis 49 50 51 170 Data was analyzed using the Statistical Package for the Social Sciences for Windows 52 53 171 (version 23.0; SPSS, IBM, Chicago, IL, USA). Statistical analysis was performed by t 54 55 56 172 test or Chi-square test according to the type of continuous data or categorical data. 57 58 59 173 Survival curves were obtained using Kaplan-Meier’s method. The significance was 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 174 obtained by a log-rank test. P value of less than 0.05 was considered statistically 5 6 7 175 significant. 8 9 176 10 11 12 177 13 14 15 178 Results 16 17 179 Patient characteristics 18 For peer review only 19 20 180 As shown in Table 1, a total of 432 patients consisted of 278 (64%) male patients and 21 22 181 154 (36%) female patients (male:female ratio=1.8:1) were analyzed. The mean age ± 23 24 25 182 standard deviation (SD) of patients was 45.61±16.12 years. The median duration of 26 27 183 disease was 1.56(0.75, 3) months. The mitral valve (54%) was the mostly involved 28 29 30 184 valve followed by the aortic (43%) valve. The most common complication of IE 31 32 33 185 patients was pulmonary arterial hypertension (PAH) (33%). As for embolic events 34 35 186 discovered or presented in hospital, brain (14%) and spleen (6%) infarction were most 36

37 http://bmjopen.bmj.com/ 38 187 common. For predisposing factors, dialysis accounted for 1%, dental procedure 39 40 188 accounted for 1%, and intravenous drug use (IVDU) accounted for 0.5%. In the end, 41 42 43 189 346 patients (80%) had cardiac surgery and 55 patients (13%) died in Ruijin hospital. 44

45 on September 28, 2021 by guest. Protected copyright. 46 190 The pathogen and treatment of infective endocarditis patients were listed in 47 48 191 supplemental table 1. 49 50 51 192 52 53 193 Clinical manifestations and laboratory features 54 55 56 194 Fever (84%) and new heart murmur (79%) were presented in most patients with IE 57 58 59 195 (Supplemental Table 2). The most common rheumatic manifestations were serositis 60 9

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 196 (39%), arthritis (15%), and myalgia (6%). In addition, a few patients had non-specific 5 6 7 197 manifestations such as hepatomegaly (4%), splenomegaly (23%) and fatigue (20%). 8 9 198 Eighteen patients had Janeway lesions (4%), 11 patients with Osler nodes (3%) and 5 10 11 12 199 patients with Roth spot (1%). 13 14 200 15 16 17 201 Blood cultures were positive in 50% of IE patients. Echocardiographic data showed 18 For peer review only 19 20 202 that 266/432(62%) patients had IE specific characteristic like vegetation, abscess, 21 22 203 pseudoaneurysm, intracardiac valvular perforation and abnormal activity around the 23 24 25 204 site of prosthetic valve. As for immunologic features, 89 patients were tested for 26 27 205 ANCA with 21 (24%) ANCA-positive, all of which were PR3-ANCA positive. 28 29 30 206 Fifty-three patients were tested for aPL and 21 were positive (45%). One patient had 31 32 33 207 positive IgG aPL (2%), 4 patients with IgM aPL (8%), 1 patient with IgA aPL (2%), 2 34 35 208 patients with anti-β2GPI (4%) and 17 patients with LAC (32%). In total, 17 patients 36

37 http://bmjopen.bmj.com/ 38 209 were single aPL positive, 2 patients were double aPL positive and none was triple aPL 39 40 210 positive. Eight (13%) patients had positive ANA, one had positive extractable nuclear 41 42 43 211 antigen (ENA) (2%). Proteinuria (25%) and hematuria (42%) were most commonly 44

45 on September 28, 2021 by guest. Protected copyright. 46 212 seen. Of 38 IE patients tested with both aPL and ANCA, 15 were only aPL positive, 9 47 48 213 were only ANCA positive and 3 were positive for both antibodies (Table 2). Clinical 49 50 51 214 features and laboratory features of patients tested for antibodies or not were presented 52 53 215 in Supplemental Table 3. Arthritis was more common in patients tested for antibodies 54 55 56 216 than those were not, indicating the relevance between autoantibodies and rheumatic 57 58 59 217 features. Besides, patients that were not tested for antibodies had higher rate of 60 10

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 218 cardiac surgery, which might be related to the complexity of patients tested for 5 6 7 219 antibodies. 8 9 220 10 11 12 221 Comparison of patients with positive and negative ANCA or APL 13 14 222 Of 432 IE patients, 104 IE patients who had been tested for either ANCA or aPL were 15 16 17 223 analyzed. Finally, 89 patients were classified as ‘ANCA-positive IE’ or 18 For peer review only 19 20 224 ‘ANCA-negative IE’ and 53 patients as ‘aPL-positive IE’ or ‘aPL-negative IE’, 21 22 225 according to their antibodies test (Supplemental Figure 1). 38 patients were tested 23 24 25 226 with both ANCA and aPL, and the characteristics were analyzed. Of 89 patients who 26 27 227 were tested for ANCA, 21 (24%) were ANCA-positive and 68 (76%) were 28 29 30 228 ANCA-negative. Clinical manifestations such as fever, fatigue, thrombosis, myalgia, 31 32 33 229 splenomegaly, heart murmur, and weight loss were not significantly different between 34 35 230 the two groups (P˃0.05). But arthritis (P=0.003) was more frequent, and there was a 36

37 http://bmjopen.bmj.com/ 38 231 tendency against serositis (P=0.06) in ANCA-positive patients than in 39 40 232 ANCA-negative group (Table 3). Compared to ANCA-negative patients, IgM 41 42 43 233 (P=0.048) was higher, Hb (P=0.001) was significantly lower and elevated RF 44

45 on September 28, 2021 by guest. Protected copyright. 46 234 (P=0.053) seems to be more common in ANCA-positive patients (Table 4). Of 53 47 48 235 patients who were tested for aPL, 21 (40%) were aPL-positive, the remaining 32 49 50 51 236 (60%) were aPL-negative. ESR was significantly higher in patients who were positive 52 53 237 for aPL (P=0.003) (Table 4). 54 55 56 238 57 58 59 239 Outcomes 60 11

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 240 Overall, 9 patients tested for aPL or ANCA antibodies died in hospital. Among these 5 6 7 241 patients, 4 patients were ANCA positive and 2 patients were aPL positive. One patient 8 9 242 died of renal failure, four patients of acute heart failure, two of septic shock and two 10 11 12 243 of stroke. During the 3-month observation period from the day of diagnosis, no 13 14 244 patient died after discharged from hospital. The survival rate was significantly lower 15 16 17 245 in ANCA-positive IE (P =0.032), and there was no significant difference between the 18 For peer review only 19 20 246 aPL groups (P = 0.728) (Fig. 1). 21 22 247 23 24 25 248 Discussion 26 27 249 Although there were data reported in various countries about different aspects of IE, 28 29 30 250 10-12 the analysis of the clinical and immunologic features from Chinese IE patients 31 32 33 251 was still limited. The clinical presentation and laboratory findings of IE are similar to 34 35 252 many rheumatic diseases, making it difficult to diagnose. Thus, it is pivotal to analyze 36

37 http://bmjopen.bmj.com/ 38 253 rheumatic manifestations of IE, so as to help physicians differentiate it when 39 40 254 encountering these clinical scenarios. We have already reported cases and literature 41 42 43 255 review of IE patients with ANCA before.12 In this study, we analyzed clinical and 44

45 on September 28, 2021 by guest. Protected copyright. 46 256 laboratory features of 432 IE patients in Ruijin Hospital and compared the 47 48 257 characteristics of 104 patients tested for ANCA or aPL during the past two decades. 49 50 51 258 The mean age of patients was 46 years, which was much younger than other studies. 52 53 259 Although in many researches, the mean age of general population was elder, there 54 55 56 260 were others not. In a research of Brazil, the mean age of IE patients was 45.26 years.2 57 58 59 261 In Turkey, it has been reported that IE occurs in relatively young patients (46.94 ± 60 12

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 262 17.11 years).13 While in India, the mean age of 44 IE patients was 31 years.14 In 5 6 7 263 addition, elder people were more likely to have valve replacement surgery because of 8 9 264 degenerative valve disease. Prosthetic valvec endocarditis accounts for more 10 11 12 265 percentage of all cased of IE than that in our study3. As reported in the reference of 13 14 266 2015 guideline, blood culture–negative IE (BCNIE) can occur in up to 31% of all 15 16 17 267 cases of IE and often poses considerable diagnostic and therapeutic dilemmas. BCNIE 18 For peer review only 19 20 268 most commonly arises as a consequence of previous antibiotic administration, and can 21 22 269 be caused by fungi or fastidious bacteria, notably obligatory intracellular bacteria. As 23 24 25 270 a tertiary hospital, before hospitalized in Ruijin hospital, a lot of patients had been 26 27 271 treated with antibiotics empirically in other hospitals with uncertain diagnosis. As a 28 29 30 272 result, there were only 50% patients with culture positive. Patients with culture 31 32 33 273 negative were diagnosed by synthesis of other criteria like echocardiography and 34 35 274 histological examination of a vegetation after operation. We had checked the 36

37 http://bmjopen.bmj.com/ 38 275 diagnosis for certainty according to the guideline. Our study showed that 39 40 276 ANCA-positive IE patients had higher IgM, lower Hb and more presentation of 41 42 43 277 arthritis, and seemed to have more serositis than ANCA-negative IE patients. 44

45 on September 28, 2021 by guest. Protected copyright. 46 278 Moreover, aPL-positive IE patients had significantly higher ESR than aPL-negative 47 48 279 group. Positive ANCA and aPL were reported to be non-specific in patients with 49 50 51 280 infection.6-8 52 53 281 54 55 56 282 The phenomenon of IE patients with positive ANCA has long been reported. A 57 58 59 283 research from France reported that 8% IE patients had PR3-ANCA or MPO-ANCA. It 60 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 284 was associated with young age (P=0.022), echocardiographic vegetations (P=0.043), 5 6 15 7 285 and elevated IgG levels (P=0.017). It has also been reported that most 8 9 286 Gram-positive endocarditis could cause an anti-PR3 specificity. Possibly during 10 11 12 287 delayed polymorphonuclear leucocyte apoptosis, PR3 contact with the immune 13 14 288 system and caused increased production of anti-PR3 ANCA.16 It was interesting to 15 16 17 289 find that in our IE patients, all of them were PR3 ANCA positive. One important 18 For peer review only 19 20 290 presentation of ANCA related vasculitis is glomerular nephritis. But IE-associated 21 22 291 glomerular nephritis often presents with acute kidney injury.17 The most common 23 24 25 292 biopsy finding was necrotizing and crescentic glomerular nephritis (53%), followed 26 27 293 by endocapillary proliferative glomerular nephritis (37%).17 In our study, IE patients 28 29 30 294 who had proteinuria and hematuria accounted for 25% and 42%, respectively. 31 32 33 295 However, kidney biopsy was not routinely done in these patients. 34 35 296 36

37 http://bmjopen.bmj.com/ 38 297 Apart from ANCA, aPL was also paid attention to. In terms of the relationship 39 40 298 between aPL and IE, it has been indicated that IgG aPL might be associated with Q 41 42 43 299 Fever Endocarditis and antiphospholipid antibodies, and has previously been shown 44

45 18-19 on September 28, 2021 by guest. Protected copyright. 46 300 to be associated with thrombosis during acute Q fever. There was a research in 47 48 301 Brazil on embolic events in infective endocarditis.2 Furthermore, the presence of IgM 49 50 51 302 aPL and anti-β2GPI was associated with embolic events, particularly cerebral ones, 52 53 303 which could contribute to assess the embolic risk of IE.20 In our study, there was no 54 55 56 304 difference of thrombosis between APL-positive patients and APL-negative group. 57 58 59 305 APL-positive IE patients had higher ESR, and LAC was the most commonly positive 60 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 306 type of aPL in IE patients. This might indicated that aPL was related to an 5 6 7 307 inflammatory status. 8 9 308 10 11 12 309 As for the outcomes, in our study, the survival rate was significantly lower in 13 14 310 ANCA-positive IE than ANCA-negative group. The ANCA positive patients had 15 16 17 311 lower Hgb than the ANCA negative, which might be the reason for the higher 18 For peer review only 19 20 312 mortality. Previously, we have reported that the survival rate was significantly lower 21 22 313 in ANCA-positive IE than ANCA-positive IE.12 There was no difference of survival 23 24 25 314 rate between the aPL groups in our study. Thus, ANCA positive patients probably 26 27 315 were those with a delay in diagnosis and with more heart failure. So that the 28 29 30 316 antibodies may be only a marker of poor prognosis, which might be the reason for the 31 32 33 317 higher mortality. 34 35 318 36

37 http://bmjopen.bmj.com/ 38 319 There were several limitations of this study. First, it was retrospective. The 39 40 320 measurement of autoantibodies was not conducted in all IE patients when they were 41 42 43 321 diagnosed, with only 89 with ANCA and 53 with aPL respectively, and only 38 44

45 on September 28, 2021 by guest. Protected copyright. 46 322 patients were both tested with ANCA and aPL. Second, the time of study spanned 47 48 323 quite long and some patients lost follow ups. 49 50 51 324 52 53 54 325 Conclusions 55 56 326 The present study supports the hypothesis that rheumatic complications are frequent 57 58 59 327 in patients with IE. For rheumatologists, it’s indispensable to routinely exclude IE 60 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 328 from patients with rheumatic features such as arthritis, serositis, splenomegaly, 5 6 7 329 myalgia, proteinuria and hematuria. For cardiologists, more attention should be paid 8 9 330 to the measurement of autoantibodies in IE patients, which might contribute to the 10 11 12 331 evaluation of prognosis. 13 14 332 15 16 17 333 18 For peer review only 19 20 334 21 22 335 References 23 24 25 336 1. Fedeli U, Schievano E, Buonfrate D, et al. Increasing incidence and mortality of infective 26 27 337 endocarditis: a population-based study through a record-linkage system. BMC Infect Dis 28 29 30 338 2011;11:48. 31 32 33 339 2. Monteiro TS, Correia MG, Golebiovski WF, et al. Asymptomatic and symptomatic embolic 34 35 340 events in infective endocarditis: associated factors and clinical impact. Braz J Infect Dis 36

37 http://bmjopen.bmj.com/ 38 341 2017;21:240-247. 39 40 342 3. Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC Guidelines for the management of 41 42 43 343 infective endocarditis: The Task Force for the Management of Infective Endocarditis of the 44

45 on September 28, 2021 by guest. Protected copyright. 46 344 European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic 47 48 345 Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J 49 50 51 346 2015;36:3075-3128. 52 53 347 4. Spaulding AR, Salgado-Pabon W, Kohler PL, et al. Staphylococcal and Streptococcal 54 55 56 348 Superantigen Exotoxins. Clin Microbiol Rev 2013;26:422-447. 57 58 59 349 5. Misra DP, Chowdury AC, Edavalath S, et al. Endocarditis: the great mimic of rheumatic 60 16

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 350 diseases. Trop Doct 2016;46:180-186. 5 6 7 351 6. Langlois V, Lesourd A, Girszyn N, et al. Antineutrophil Cytoplasmic Antibodies Associated 8 9 352 With Infective Endocarditis. Medicine (Baltimore) 2016;95:e2564. 10 11 12 353 7. Kupferwasser LI, Hafner G, Mohr-Kahaly S, et al. The presence of infection-related 13 14 354 antiphospholipid antibodies in infective endocarditis determines a major risk factor for embolic 15 16 17 355 events. J Am Coll Cardiol 1999;33:1365-71. 18 For peer review only 19 20 356 8. Peng H, Chen WF, Wu C, et al. Culture-negative subacute bacterial endocarditis masquerades 21 22 357 as granulomatosis with polyangiitis (Wegener's granulomatosis) involving both the kidney and 23 24 25 358 lung. BMC Nephrol 2012;13:174. 26 27 359 9. Durack DT, Lukes AS, Bright DK, Duke Endocarditis Service. New criteria for diagnosis of 28 29 30 360 infective endocarditis: utilization of specific echocardiographic findings. Am J Med 1994; 96:200– 31 32 33 361 9.) 34 35 362 10. El-Chakhtoura N, Yasmin M, Kanj SS, et al. A 27-year experience with infective endocarditis 36

37 http://bmjopen.bmj.com/ 38 363 in Lebanon. J Infect Public Health 2017;10:734-739. 39 40 364 11. Ferraris L, Milazzo L, Ricaboni D, et al. Profile of infective endocarditis observed from 2003 - 41 42 43 365 2010 in a single center in Italy. BMC Infect Dis 2013;13:545. 44

45 on September 28, 2021 by guest. Protected copyright. 46 366 12. Ying CM, Yao DT, Ding HH, et al. Infective endocarditis with antineutrophil cytoplasmic 47 48 367 antibody: report of 13 cases and literature review. PLoS One 2014;9:e89777. 49 50 51 368 13. Şimşek-Yavuz S, Şensoy A, Kaşıkçıoğlu H, et al. Infective endocarditis in Turkey: aetiology, 52 53 369 clinical features, and analysis of risk factors for mortality in 325 cases. International Journal of 54 55 56 370 Infectious Diseases 2015, 30:106-114. 57 58 59 371 14. Ghosh S, Sahoo R, Nath RK, et al. A Study of Clinical, Microbiological, and 60 17

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 25 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 372 Echocardiographic Profile of Patients of Infective Endocarditis. Int Sch Res Notices 5 6 7 373 2014;2014:1-9. 8 9 374 15. Mahr A, Batteux F, Tubiana S, et al. Brief report: prevalence of antineutrophil cytoplasmic 10 11 12 375 antibodies in infective endocarditis. Arthritis Rheumatol 2014;66:1672-7. 13 14 376 16. Aslangul E, Goulvestre C, Mallat Z, et al. Human bartonella infective endocarditis is 15 16 17 377 associated with high frequency of antiproteinase 3 antibodies. J Rheumatol 2014;41:408-10. 18 For peer review only 19 20 378 17. Boils CL, Nasr SH, Walker PD, et al. Update on endocarditis-associated glomerulonephritis. 21 22 379 Kidney Int 2015;87:1241-9. 23 24 25 380 18. Million M, Walter G, Bardin N, et al. Immunoglobulin G anticardiolipin antibodies and 26 27 381 progression to Q fever endocarditis. Clin Infect Dis 2013;57:57-64. 28 29 30 382 19. Million M, Bardin N, Bessis S, et al. Thrombosis and antiphospholipid antibody syndrome 31 32 33 383 during acute Q fever: A cross-sectional study. Medicine (Baltimore) 2017;96:e7578. 34 35 384 20. Selton-Suty C, Maigrat CH, Devignes J, et al. Possible relationship between antiphospholipid 36

37 http://bmjopen.bmj.com/ 38 385 antibodies and embolic events in infective endocarditis. Heart 2018;104:509-516. 39 40 386 41 42 43 387 44

45 on September 28, 2021 by guest. Protected copyright. 46 388 47 48 389 Contribution 49 50 51 390 For authors, Jialin Teng, Honglei Liu, Xiaobing Cheng, Huihui Chi and Fan Wang 52 53 391 collected the clinical data. Yutong Su and Yue Sun analyzed the data. Zhuochao Zhou 54 55 56 392 and Junna Ye wrote the manuscript. Wei Jin and Chengde Yang designed the study 57 58 59 393 and critically read the manuscript. We would like to thank the patients for their 60 18

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 394 participation in this study. 5 6 7 395 8 9 396 10 11 12 397 13 14 398 Table 1 15 399 General characteristics of infective endocarditis (IE) patients. 16 Item N (%) 17 18 N For peer review 432only 19 Female/male 154(36)/278(64) 20 Age (years, mean±SD) 45.61±16.12 21 22 Duration (months, median, interquartile 1.56(0.75,3) 23 range) 24 Predisposing factors 25 26 Dialysis 6(1) 27 Dental procedure 4(1) 28 Intravenous drug use 2(0.5) 29 Congenital heart disease 66(15) 30 31 Co-morbidities 32 Pulmonary arterial hypertension 141(33) 33 Hypertension 82(19) 34 35 Diabetes 27(6) 36 Cancer 11(3)

37 Thrombosis http://bmjopen.bmj.com/ 38 39 Brain 67(14) 40 Spleen 26(6) 41 Kidney 8(2) 42 extremity 8(2) 43 44 Coronary 4(1)

45 Lung 4(1) on September 28, 2021 by guest. Protected copyright. 46 Infected valve 47 48 Prosthetic valve 42(10) 49 Native valve 50 Mitral valve 234(54) 51 52 Aortic valve 185(43) 53 Tricuspid valve 17(4) 54 Pulmonary valve 10(2) 55 Outcome 56 57 Cardiac surgery 346(80) 58 Relapse 13(3) 59 In-hospital mortality 55(13) 60 19

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 25 BMJ Open

1 2

3 400 Note: standard deviation (SD). BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 401 6 402 Table 2 7 403 Antibodies of patients tested with both aPL and ANCA. 8 9 N 10 Only aPL positive 15 11 ACL positive 1(IgM ACL) 12 13 aβ2GP1 positive 1 14 LAC positive 10 15 ACL+LAC 3(2 IgM ACL +1 IgA ACL) 16 17 Only ANCA positive 9 18 Both aPL and ANCAFor positive peer review3(LAC, LAC+IgM only ACL, IgM ACL) 19 404 Note: anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 20 21 405 antibodies (anti-β2GPI), antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody 22 406 (aPL). 23 407 24 Table 3 25 408 26 409 Clinical features of patients positive or negative for ANCA and aPL. 27 ANCA aPL 28 29 ANCA(+) ANCA(-) p valve aPL(+) aPL(-) p valve 30 N 21 68 / 21 32 / 31 Positive Blood 13(61.9) 37(57.8) 0.741 10(50.0) 17(53.1) 0.826 32 33 culture 34 Fever 19(90.5) 57(83.8) 0.688 18(85.7) 23(71.9) 0.400 35 Arthritis 11(52.4) 13(19.1) 0.003 5(23.8) 8(25.0) 0.922 36 Fatigue 3(14.3) 16(23.5) 0.549 7(33.3) 9(28.1) 0.686 37 http://bmjopen.bmj.com/ 38 Serositis 8(80.0) 23(41.8) 0.060 13(61.9) 15(46.9) 0.284 39 Thrombosis 3(14.3) 17(25.0) 0.466 8(38.1) 10(31.3) 0.607 40 Myalgia 0 6(10.9) 0.672 2(9.5) 2(6.5) 1.000 41 42 Splenomegaly 6(28.6) 20(29.4) 0.941 5(23.8) 6(18.8) 0.922 43 Heart murmur 10(100.0) 43(78.2) 0.233 18(85.7) 28(87.5) 1.000 44 Weight loss 3(30.0) 9(16.4) 0.562 3(14.3) 7(21.9) 0.740

45 on September 28, 2021 by guest. Protected copyright. 46 Cardiac surgery 5(50.0) 42(72.4) 0.295 16(76.2) 19(59.4) 0.206 47 In-hospital 4(19.0) 3(4.4) 0.087 2(9.5) 4(12.5) 1.000 48 mortality 49 50 410 Note: antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 51 411 52 412 Table 4 53 54 413 Laboratory features of patients positive or negative for ANCA and aPL. 55 ANCA aPL 56 ANCA(+) ANCA(-) p value aPL(+) aPL(-) p value 57 58 N 21 68 / 21 32 / 59 WBC (*109/L) 8.39±2.53 8.56±4.19 0.896 7.66±2.54 8.58±4.52 0.403 60 20

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 25

1 2

3 Hb (g/L) 87.57±26.83 105.51±18.76 0.001 99.57±15.28 102.34±26.22 0.664 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 9 5 PLT (*10 /L) 194.70±110.91 201.25±127.06 0.879 184.95±83.86 193.47±151.31 0.819 6 LDH (IU/L) 206.125±53.45 252.00±96.73 0.200 278.87±296.11 247.72±103.72 0.632 7 ESR (mm/h) 64.76±36.87 51.85±33.47 0.135 69.14±34.82 39.09±33.48 0.003 8 IgG (mg/dl) 2254.29±1001.40 1660.05±528.67 0.172 1959.69±747.29 1751.67±707.30 0.379 9 10 IgA (mg/dl) 232.29±89.77 301.50±131.20 0.188 276.63±94.03 283.38±130.21 0.859 11 IgM (mg/dl) 312.71±183.51 140.70±64.03 0.048 193.07±119.30 153.79±119.39 0.324 12 IgE (mg/dl) 37.00±22.39 354.85±569.90 0.286 357.31±437.81 351.73±635.24 0.983 13 14 ACL positive 2(25.0) 2(5.3) 0.436 5(23.8) 0 0.004 15 aβ2GPI positive 0 1(4.8) 1.000 2(12.5) 0 0.057 16 LAC positive 2(28.6) 11(37.9) 0.981 17(81.0) 0 0.000 17 18 ANA positive 2(15.4) For peer6(11.8) review1.000 3(15.8) only 3(10.7) 0.618 19 Elevated CRP 18(90.0) 51(85.0) 0.851 14(73.7) 21(75.0) 0.921 20 Elevated RF 10(58.8) 14(31.8) 0.053 6(30.8) 6(25.0) 0.336 21 22 414 Note: White blood cell counts (WBC), hemoglobin (Hb), platelet (PLT), erythrocyte 23 415 sedimentation rate (ESR), serum lactate dehydrogenase (LDH), immunoglobulin G (IgG), 24 416 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), anticardiolipin 25 26 417 antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I antibodies (anti-β2GPI), 27 418 antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 28 29 419 30 31 420 Figure 1 32 421 Kaplan-Meier’s survival curves. 33 422 a. Kaplan-Meier’s survival curves of aPL tested patients; b. Kaplan-Meier’s survival curves of 34 35 423 ANCA tested patients. 36 424 37 http://bmjopen.bmj.com/ 38 39 425 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 25 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

1 2

3 Supplemental Table 1 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Pathogen and treatment of infective endocarditis patients. 6 Item N (%) 7 Pathogen 8 9 Streptococcus 110(25.46) 10 Staphylococcus aureus 64(14.81) 11 Staphylococcus epidermidis 30(6.94) 12 13 Enterococci 26(6.02) 14 Enterobacter 20(4.63) 15 Other Gram-negative Bacillus 18(4.17) 16 Candida albicans 9(2.08) 17 18 LactococcusFor lactis peer review 4(0.92)only 19 Gemella mobillorum 4(0.92) 20 Kocuria sp 4(0.92) 21 22 Aerococcus urinae 3(0.69) 23 micrococcus 2(0.46) 24 Aerococcus viridans 2(0.46) 25 26 pseudomonas maltophilia 2(0.46) 27 Treatment 28 Vancomycin 228(52.77) 29 Cephalosporin 163(37.73) 30 31 carbapenems 111(25.69) 32 Penicillin 109(25.23) 33 Quinolones 95(21.99) 34 35 Linezolid 74(17.13) 36 Aminoglycoside 41(9.49)

37 Teicoplanin 21(4.86) http://bmjopen.bmj.com/ 38 39 Antifungal 21(4.86) 40 Phosphonomycin 20(4.63) 41 Metronidazole 7(1.62) 42 Macrolides 5(1.16) 43 44 Daptomycin 4(0.92)

45 on September 28, 2021 by guest. Protected copyright. 46 47 Supplemental Table 2 48 Clinical manifestations of infective endocarditis patients at time of IE diagnosis. 49 Item N (%) 50 51 Heart murmur 343(79.40) 52 Fever 364(84.26) 53 Arthritis 65(15.05) 54 Fatigue 86(19.91) 55 56 Serositis 170(39.35) 57 Myalgia 28(6.48) 58 Hepatomegaly 19(4.40) 59 60 Splenomegaly 99(22.92)

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 25 BMJ Open

1 2

3 Peripheral polyneuropathy 4(0.93) BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Janeway lesion 18(4.17) 6 Osler node 11(2.55) 7 Roth spot 5(1.16) 8 9 10 Supplemental Table 3 11 Clinical features and laboratory features of patients tested with antibodies or not. 12 13 Patients tested with antibodies Patients not tested with antibodies 14 N 104 328 / 15 Female/male 40/64 114/214 0.492 16 17 Age (years, mean±SD) 45.65±14.22 45.59±16.97 0.973 18 Positive Blood cultureFor 56 (5peer6.6) review154 (only47.0) 0.108 19 Fever 89(85.6) 275(83.8) 0.758 20 21 Arthritis 27(26.0) 41(12.5) 0.001 22 Fatigue 26(25.0) 67(20.4) 0.323 23 Serositis 35(43.8) 139(42.4) 0.824 24 Thrombosis 25(24.0) 116(35.4) 0.032 25 26 Myalgia 7(8.9) 23(7.0) 0.572 27 Splenomegaly 29(27.9) 70(21.3) 0.167 28 Heart murmur 67(83.8) 288(87.8) 0.333 29 30 Weight loss 14(17.5) 47(14.3) 0.476 31 Cardiac surgery 54(65.1) 290(88.4) 0.000 32 In-hospital mortality 9(8.7) 46(14.0) 0.178 33 9 34 WBC (*10 /L) 8.39±2.53 8.39±2.53 0.614 35 Hb (g/L) 87.57±26.83 87.57±26.83 0.105 36 PLT (*109/L) 194.70±110.91 194.70±110.91 0.659

37 http://bmjopen.bmj.com/ ESR (mm/h) 64.76±36.87 64.76±36.87 0.116 38 39 Proteinuria 27(27.3) 85(25.9) 0.788 40 hematuria 50(50.5) 133(40.5) 0.079 41 42 Note: White blood cell counts (WBC), hemoglobin (Hb), platelet (PLT), erythrocyte 43 sedimentation rate (ESR). 44

45 on September 28, 2021 by guest. Protected copyright. Supplemental Figure 1 46 47 Distribution of patients diagnosed with infectious endocarditis. 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 25

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 * Patients who were either tested with aPL or ANCA were included and compared in different 32 groups. 33 Note: infectious endocarditis (IE), antineutrophil cytoplasmic antibody (ANCA), 34 35 anti-phospholipid antibody (aPL). 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

Clinical Characteristics of Infective Endocarditis in Patients with Antineutrophil Cytoplasmic Antibody or Antiphospholipid Antibody in Shanghai ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-031512.R2

Article Type: Original research

Date Submitted by the 29-Oct-2019 Author:

Primary Subject Rheumatology Heading:

Secondary Subject Heading: Rheumatology

Infective endocarditis, Rheumatic manifestation, ANCA, Antiphospholipid Keywords: antibody

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 27 BMJ Open

1 2

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 1 Clinical Characteristics of Infective Endocarditis in Patients with Antineutrophil 5 6 7 2 Cytoplasmic Antibody or Antiphospholipid Antibody in Shanghai 8 9 3 10 11 12 4 Zhuochao Zhou1,*, Junna Ye1,*, Jialin Teng1, Honglei Liu1, Xiaobing Cheng1, Yue 13 14 5 Sun1, Yutong Su1, Huihui Chi1, Fan Wang1, Chengde Yang1,#, Wei Jin2,# 15 16 17 6 18 For peer review only 19 1 20 7 Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 21 22 8 University School of Medicine; 23 24 25 9 2Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School 26 27 10 of Medicine 28 29 30 11 31 32 * 33 12 These authors contributed equally to this work. 34 35 13 36

37 http://bmjopen.bmj.com/ 38 14 #Correspondence: 39 40 15 Wei Jin, MD, PhD 41 42 43 16 Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School of 44

45 on September 28, 2021 by guest. Protected copyright. 46 17 Medicine, No. 197 Ruijin Second Road, Huangpu District, Shanghai 200025, China. 47 48 18 Tel.: +86 21 64370045; 49 50 51 19 Fax: +86 21 34186000; 52 53 20 E-mail: [email protected] 54 55 56 21 57 58 59 22 Chengde Yang, MD, PhD 60 1

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 23 Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 5 6 7 24 University School of Medicine, No. 197 Ruijin Second Road, Huangpu District, 8 9 25 Shanghai 200025, China. 10 11 12 26 Tel.: +86 21 64370045; 13 14 27 Fax: +86 21 34186000; 15 16 17 28 E-mail: [email protected] 18 For peer review only 19 20 29 21 22 30 23 24 25 31 Abstract 26 27 32 Objective: This study aimed to characterize rheumatic manifestations and 28 29 30 33 autoantibodies in 432 patients diagnosed with infective endocarditis (IE) in Shanghai. 31 32 33 34 Design, setting and participants: A retrospective study was conducted in Ruijin 34 35 35 Hospital from 1997 to 2017. The clinical and laboratory characteristics of a total of 36

37 http://bmjopen.bmj.com/ 38 36 432 patients were analysed. In addition, the differences between patients with positive 39 40 37 and negative antineutrophil cytoplasmic antibodies (ANCA) and antiphospholipid 41 42 43 38 antibodies (aPL) as well as the survival rates of these patients were compared. 44

45 on September 28, 2021 by guest. Protected copyright. 46 39 Results: A total of 432 patients, including 278 male patients and 154 female patients, 47 48 40 were included. The mean age of the patients was 45.61±16.12 years. A total of 346 49 50 51 41 patients (80%) had cardiac surgery, and 55 patients (13%) died in the hospital. 52 53 42 Among the IE patients, 104 were tested for either ANCA or aPL and were analysed in 54 55 56 43 different groups. Twenty-one (24%) positive ANCA patients were proteinase 57 58 59 44 3(PR3)-ANCA. Compared with the ANCA-negative group, patients with positive 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 45 ANCA had higher IgM (P=0.048), lower haemoglobin (P=0.001), and a higher 5 6 7 46 likelihood of arthritis (P=0.003). Twenty-one (40%) aPL-positive patients had a 8 9 47 higher erythrocyte sedimentation rate (ESR) than was found in the aPL-negative 10 11 12 48 group (P=0.003). In addition, the survival rate of the ANCA-positive IE patients was 13 14 49 lower (P=0.03) than that of the ANCA-negative group, while there was no difference 15 16 17 50 between patients with or without aPL antibodies (P=0.728). 18 For peer review only 19 20 51 Conclusion: The present study supports the claim that rheumatic manifestations and 21 22 52 autoantibodies are frequently present in patients with IE and might lead to early 23 24 25 53 misdiagnosis. Physicians should pay more attention to the measurement of 26 27 54 autoantibodies in these patients. 28 29 30 55 31 32 33 56 Keywords: infective endocarditis, rheumatic manifestation, ANCA, aPL 34 35 57 36

37 http://bmjopen.bmj.com/ 38 58 39 40 41 42 59 Strengths and limitations of this study 43 44 60 1. The study analysed the clinical characteristics and survival rates of IE patients with

45 on September 28, 2021 by guest. Protected copyright. 46 47 61 and without ANCA or aPL antibodies. 48 49 62 2. As it was a retrospective study, not all patients underwent detection of ANCA and 50 51 52 63 aPL autoantibodies. 53 54 55 64 3.Although there was a 4-month observation period for patients’ mortality from the 56 57 65 day of diagnosis during hospitalization, we did not register deaths occurring in other 58 59 60 66 hospitals or patients who died at home. 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 67 5 6 7 68 8 9 69 Introduction 10 11 12 70 Infective endocarditis (IE) is a microbial infection of the endocardium that usually 13 14 71 involves the heart valves. Despite being relatively rare worldwide, IE has been 15 16 17 72 reported in a steady incidence over the past three decades ranging from 2-6 per 18 For peer review only 19 20 73 100,000 individuals in the general population per year and has a considerable 21 22 74 associated mortality rate that varies from 10% to 30%.1 Predisposing factors for IE 23 24 25 75 include conditions such as congenital heart diseases, rheumatic valve disease, 26 27 76 artificial valves, and other factors, such as intravenous drug use, dental procedures 28 29 30 77 and haemodialysis.2 The diagnosis of IE is made based on symptoms, blood 31 32 3 33 78 cultures, echocardiography and pathological biopsy from valve surgery. 34 35 79 However, superantigens induced by some bacteria, such as Staphylococcus 36

37 http://bmjopen.bmj.com/ 38 80 aureus, can stimulate an immune response, which could interfere with antibody 39 40 81 production.4 Rheumatic manifestations, such as myalgia, arthralgia, and arthritis, 41 42 43 82 are prevalent, occurring in nearly 40% of patients with IE at presentation or weeks 44

45 5 on September 28, 2021 by guest. Protected copyright. 46 83 to months before the diagnosis of IE. When classic IE manifestations are less 47 48 84 evident, it can cause misdiagnosis, which might lead to the delayed initiation of 49 50 51 85 antibiotic treatment. Thus, the distinction between IE and rheumatic diseases is not 52 53 86 crystal clear, and improving understanding this difference is of great importance. 54 55 56 87 57 58 59 88 Many specific antibodies, such as antineutrophil cytoplasmic antibodies (ANCA) and 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 89 anti-phospholipid antibody (aPL), might be related to the pathophysiology of IE. 6-7 5 6 7 90 Due to positive ANCA tests, IE has been reported to mimic ANCA-associated 8 9 91 vasculitis (AAV). Patients with IE may present with multiple pulmonary nodules and 10 11 12 92 glomerulonephritis, which mimics granulomatosis with polyangiitis.8 One study 13 14 93 further underscored that ANCA might be associated with multiple valve 15 16 17 94 involvement.6 Additionally, infection-associated elevated aPL levels in patients with 18 For peer review only 19 20 95 infective endocarditis are related to endothelial cell activation, thrombin generation 21 22 96 and impairment of fibrinolysis, which may contribute to the increased risk of major 23 24 25 97 embolic events in these patients.7 Therefore, ANCA and aPL are not specific to IE 26 27 98 patients. 28 29 30 99 31 32 33 100 The objective of this study was to compare the clinical characteristics and survival 34 35 101 rates of IE patients with and without ANCA or aPL autoantibodies in a tertiary 36

37 http://bmjopen.bmj.com/ 38 102 hospital in Shanghai, China from 1997 to 2017. 39 40 103 41 42 43 104 Methods 44

45 on September 28, 2021 by guest. Protected copyright. 46 105 Patients 47 48 106 This study was a retrospective study. We continuously included 432 patients 49 50 51 107 hospitalized in Ruijin Hospital affiliated to Shanghai Jiao Tong University School of 52 53 108 Medicine who were diagnosed with IE from 1997 to 2017 after ruling out 33 patients 54 55 56 109 with primary rheumatic disease or in immunosuppression condition. The clinical and 57 58 59 110 laboratory data of the patients were mainly obtained from the medical records system 60 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 111 of Ruijin hospital. Patients were not involved in the design, recruitment or conduct of 5 6 7 112 the study. 8 9 113 10 11 12 114 Only patients with definite IE determined by either the Duke criteria (up to 2000) or 13 14 115 the modified Duke criteria (from 2000 onwards) were included.3,9 Duration of disease 15 16 17 116 means the period from the time when the patients presented with the first clinical 18 For peer review only 19 20 117 feature to a definite IE diagnosis. In patients with long-term fever, the autoantibodies 21 22 118 were tested to rule out rheumatic diseases. Cardiac surgery and death were recorded 23 24 25 119 during the period of hospitalization. The study followed the ethical standards for 26 27 120 human experimentation established in the Declaration of Helsinki and was approved 28 29 30 121 by the Institutional Research Ethics Committee of Ruijin Hospital (ID: 2016-62), 31 32 33 122 Shanghai, China. 34 35 123 36

37 http://bmjopen.bmj.com/ 38 124 In this study, patients positive for either myeloperoxidase (MPO)-ANCA or 39 40 125 proteinase 3(PR3)-ANCA were defined as ANCA-positive patients. Patients with any 41 42 43 126 positive test for anticardiolipin antibodies (ACL), lupus anticoagulant (LAC) or 44

45 on September 28, 2021 by guest. Protected copyright. 46 127 anti-β2 glycoprotein I antibodies (aβ2GPI) were defined as aPL-positive patients. 47 48 128 Congestive heart failure was defined according to the New York Heart Association 49 50 51 129 classification system10. In-hospital mortality means IE patients who died during 52 53 130 hospitalization. There was a 4-month observation period for patient mortality from the 54 55 56 131 day of diagnosis. The survival rates of subgroups positive or negative for ANCA and 57 58 59 132 aPL were then analysed. 60 6

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 133 5 6 7 134 Laboratory features and echocardiography 8 9 135 Levels of anti-PR3, anti-MPO and aPL antibodies in serum were measured by enzyme 10 11 12 136 linked immunosorbent assay (ELISA). The following laboratory data were recorded: 13 14 137 white blood cell counts in blood (WBC), haemoglobin (Hb), platelets (PLT), 15 16 17 138 C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), lactate 18 For peer review only 19 20 139 dehydrogenase (LDH), rheumatoid factor (RF), immunoglobulin G (IgG), 21 22 140 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), 23 24 25 141 anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 26 27 142 antibodies (aβ2GPI), anti-PR3 antibodies (PR3-ANCA), anti-MPO antibodies 28 29 30 143 (MPO-ANCA), haematuria, and proteinuria. All IE patients underwent transthoracic 31 32 33 144 or transoesophageal echocardiography. Furthermore, each patient underwent 34 35 145 abdominal ultrasounds in the hospital. 36

37 http://bmjopen.bmj.com/ 38 146 39 40 147 Statistical analysis 41 42 43 148 Data were analysed using the Statistical Package for the Social Sciences for Windows 44

45 on September 28, 2021 by guest. Protected copyright. 46 149 (version 23.0; SPSS, IBM, Chicago, IL, USA). Statistical analyses were performed by 47 48 150 t test or Chi-square test according to the type of data (continuous or categorical, 49 50 51 151 respectively). Survival curves were obtained using Kaplan-Meier’s method. 52 53 152 Significance was obtained by a log-rank test. A p value of less than 0.05 was 54 55 56 153 considered statistically significant. 57 58 59 154 60 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 155 Patient and public involvement 5 6 7 156 Patients were not informed by the development of the research question and outcome 8 9 157 measures. Patients were not involved in the design, the recruitment to and conduct of 10 11 12 158 the study. The results were not disseminated to study participants. 13 14 159 15 16 17 160 Results 18 For peer review only 19 20 161 Characteristics of Patients 21 22 162 As shown in Table 1, a total of 432 patients, including 278 (64%) male patients and 23 24 25 163 154 (36%) female patients (male:female ratio=1.8:1), were analysed. The mean age ± 26 27 164 standard deviation (SD) of patients was 45.61±16.12 years. The median duration of 28 29 30 165 disease was 1.56 (0.75, 3) months. The mitral valve (54%) was the most frequently 31 32 33 166 involved valve followed by the aortic (43%) valve. The most common complication 34 35 167 of IE patients was congestive heart failure in 141 patients (33%). For embolic events 36

37 http://bmjopen.bmj.com/ 38 168 discovered or presented in the hospital, brain (14%) and spleen (6%) infarction were 39 40 169 most commonly seen. For predisposing factors, dialysis accounted for 1%, dental 41 42 43 170 procedures accounted for 1%, and intravenous drug use (IVDU) accounted for 0.5%. 44

45 on September 28, 2021 by guest. Protected copyright. 46 171 Additionally, 58 (13%) patients had rheumatic valvulopathy. In the end, 346 patients 47 48 172 (80%) had cardiac surgery, and 55 patients (13%) died in Ruijin Hospital. The 49 50 51 173 pathogens associated with IE patients are listed in supplemental table 1. 52 53 174 54 55 56 175 Clinical manifestations and laboratory features 57 58 59 176 Fever (84%) and new heart murmur (79%) were present in most patients with IE 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 177 (Supplemental Table 2). The most common rheumatic manifestations were arthritis 5 6 7 178 (15%), and myalgia (6%). In addition, a few patients had non-specific manifestations, 8 9 179 such as hepatomegaly (4%), splenomegaly (23%) and fatigue (20%). Eighteen 10 11 12 180 patients had Janeway lesions (4%), 11 patients had Osler nodes (3%) and 5 patients 13 14 181 had Roth spot (1%). 15 16 17 182 18 For peer review only 19 20 183 Blood cultures were positive in 50% of IE patients. Echocardiographic data showed 21 22 184 that 266/432 (62%) patients had IE-specific characteristics, such as vegetation, 23 24 25 185 abscess, pseudoaneurysm, intracardiac valvular perforation and abnormal activity 26 27 186 around the site of the prosthetic valve. Regarding immunologic features, 89 patients 28 29 30 187 were tested for ANCA, 21 (24%) of these were ANCA-positive, and all of these were 31 32 33 188 PR3-ANCA positive. Fifty-three patients were tested for aPL, and 21 were positive 34 35 189 (45%). One patient had positive IgG aPL (2%), 4 patients had IgM aPL (8%), 1 36

37 http://bmjopen.bmj.com/ 38 190 patient had IgA aPL (2%), 2 patients had aβ2GPI (4%) and 17 patients had LAC 39 40 191 (32%). Eight (13%) patients were positive for ANA, and one was positive for 41 42 43 192 extractable nuclear antigen (ENA) (2%). Proteinuria (109/432, 25%) and haematuria 44

45 on September 28, 2021 by guest. Protected copyright. 46 193 (181/432, 42%) were most commonly observed. Of 38 IE patients tested for both aPL 47 48 194 and ANCA, 15 were only aPL-positive, 9 were only ANCA-positive and 3 were 49 50 51 195 positive for both antibodies (Table 2). The clinical features and laboratory features of 52 53 196 patients tested for antibodies are presented in Supplemental Table 3. Arthritis was 54 55 56 197 more common in patients tested for antibodies than in those who were not, indicating 57 58 59 198 the relevance between autoantibodies and rheumatic features. 60 9

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 199 5 6 7 200 Comparison of patients with positive and negative ANCA or aPL 8 9 201 Of 432 IE patients, the 104 who were tested for either ANCA or aPL were analysed. 10 11 12 202 Finally, 89 patients were classified as ‘ANCA-positive IE’ or ‘ANCA-negative IE’, 13 14 203 and 53 patients were classified as ‘aPL-positive IE’ or ‘aPL-negative IE’ according to 15 16 17 204 their antibodies test (Supplemental Figure 1). Thirty-eight patients were tested for 18 For peer review only 19 20 205 both ANCA and aPL antibodies, and their characteristics were analysed. Of 89 21 22 206 patients who were tested for ANCA, 21 (24%) were ANCA-positive, and 68 (76%) 23 24 25 207 were ANCA-negative. Clinical manifestations, such as fever, fatigue, embolic events, 26 27 208 myalgia, splenomegaly, heart murmur, and weight loss, were not significantly 28 29 30 209 different between the two groups (P˃0.05). However, arthritis (P=0.003) was more 31 32 33 210 frequent in ANCA-positive patients than in ANCA-negative group (Table 3). 34 35 211 Compared to ANCA-negative patients, in ANCA-negative patients, IgM (P=0.048) 36

37 http://bmjopen.bmj.com/ 38 212 was higher, Hb (P=0.001) was significantly lower, and elevated RF (P=0.053) seems 39 40 213 to be more common (Table 4). Of 53 patients who were tested for aPL, 21 (40%) were 41 42 43 214 aPL-positive, and the remaining 32 (60%) were aPL-negative. ESR was significantly 44

45 on September 28, 2021 by guest. Protected copyright. 46 215 higher in patients who were positive for aPL and in those who were negative 47 48 216 (P=0.003) (Table 4). 49 50 51 217 52 53 218 Survival Rate 54 55 56 219 Overall, 9 patients tested for aPL or ANCA antibodies died in the hospital. Among 57 58 59 220 these patients, 4 were ANCA-positive, and 2 patients were aPL-positive. One patient 60 10

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 221 died of renal failure, four of acute heart failure, two of septic shock and two of stroke. 5 6 7 222 The survival rate was significantly lower in ANCA-positive IE patients (P =0.032), 8 9 223 but there was no significant difference between the aPL groups (P = 0.728) (Fig. 1). 10 11 12 224 13 14 225 15 16 17 226 Discussion 18 For peer review only 19 11-13 20 227 Although it has been reported in various countries about different aspects of IE, 21 22 228 the analysis of the clinical and immunologic features of Chinese IE patients is still 23 24 25 229 limited. The clinical presentations and laboratory findings of IE are similar to those of 26 27 230 many rheumatic diseases, making it difficult to diagnose. Thus, it is pivotal to analyse 28 29 30 231 the rheumatic manifestations of IE to help physicians differentiate it when 31 32 33 232 encountering these clinical scenarios. In this study, we analysed the clinical and 34 35 233 laboratory features of 432 IE patients in Ruijin Hospital and compared the 36

37 http://bmjopen.bmj.com/ 38 234 characteristics of 104 patients tested for ANCA or aPL during the past two decades. 39 40 235 The mean age of the IE patients was 46 years. In many studies of IE patients, the 41 42 43 236 mean age was higher than that in our study, while other IE patients were similar to 44

45 2,14-15 on September 28, 2021 by guest. Protected copyright. 46 237 ours . In addition, elderly people were more likely to have valve replacement 47 48 238 surgery because of degenerative valve disease. With regard for blood culture, it has 49 50 51 239 been reported that blood culture-negative IE (BCNIE) may occur in up to 31% of all 52 53 240 cases of IE and often poses considerable diagnostic and therapeutic dilemmas3. 54 55 56 241 BCNIE most commonly arises as a consequence of previous antibiotic administration. 57 58 59 242 As a tertiary hospital, before hospitalization in Ruijin Hospital, our patients had been 60 11

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 243 treated with antibiotics empirically in other hospitals with uncertain diagnoses. As a 5 6 7 244 result, only approximately 50% of the patients had a positive blood culture. Patients 8 9 245 who were culture-negative were diagnosed based on the synthesis of other criteria, 10 11 12 246 such as echocardiography findings and a postoperative histological examination of 13 14 247 vegetation. In addition, our study showed that ANCA-positive IE patients had higher 15 16 17 248 IgM and lower Hb levels, were more likely to present with arthritis than was found in 18 For peer review only 19 20 249 ANCA-negative IE patients. Moreover, aPL-positive IE patients had significantly 21 22 250 higher ESR than was found in aPL-negative patients. 23 24 25 251 26 27 252 Positive ANCA and aPL were previously reported to be non-specific in patients with 28 29 30 253 infection.6-8 The phenomenon of IE patients positive for ANCA has long been 31 32 33 254 reported. A study from France reported that 8% of IE patients had PR3-ANCA or 34 35 255 MPO-ANCA, which are associated with young age (P=0.022), echocardiographic 36

37 http://bmjopen.bmj.com/ 38 256 vegetation (P=0.043), and elevated IgG levels (P=0.017).16 It has also been reported 39 40 257 that most Gram-positive endocarditis can cause an anti-PR3 specificity. During 41 42 43 258 delayed polymorphonuclear leucocyte apoptosis, PR3 contact with the immune 44

45 17 on September 28, 2021 by guest. Protected copyright. 46 259 system could possibly cause an increase in the production of anti-PR3 ANCA. It 47 48 260 was interesting to find that among our IE patients, all of them were 49 50 51 261 PR3-ANCA-positive. One important presentation of AAV is glomerulonephritis. 52 53 262 However, IE-associated glomerular nephritis often presents with acute kidney 54 55 56 263 injury.18 The most common biopsy finding was necrotizing and crescentic glomerular 57 58 18 59 264 nephritis (53%), followed by endocapillary proliferative glomerular nephritis (37%). 60 12

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 265 In our study, among IE patients, those with proteinuria and haematuria accounted for 5 6 7 266 25% and 42%, respectively. However, the kidney biopsy was not routinely performed 8 9 267 in these patients. 10 11 12 268 13 14 269 Apart from ANCA, aPL was also analysed in our study. In terms of the relationship 15 16 17 270 between aPL and IE, it has been indicated that IgG aPL might be associated with Q 18 For peer review only 19 20 271 fever endocarditis and aPL antibodies and has previously been shown to be associated 21 22 272 with thrombosis during acute Q fever.19-20 In our study, aPL-positive patients did not 23 24 25 273 have positive serology of Coxiella. Furthermore, the presence of IgM aPL and aβ2GPI 26 27 274 was associated with embolic events, especially cerebral events, which could 28 29 30 275 contribute to assessing the embolic risk of IE.21 In our study, there was no difference 31 32 33 276 in thrombosis between aPL-positive patients and aPL-negative patients. APL-positive 34 35 277 IE patients had higher ESR, and LAC was the most commonly positive type of aPL in 36

37 http://bmjopen.bmj.com/ 38 278 IE patients. This might indicate that aPL was related to an inflammatory status. 39 40 279 41 42 43 280 In terms of the survival rate found in our study, it was significantly lower in the 44

45 on September 28, 2021 by guest. Protected copyright. 46 281 ANCA-positive IE group than in the ANCA-negative group. ANCA-positive patients 47 48 282 were more likely to be confounded with AAV, leading to a delay in diagnosis. Thus, 49 50 51 283 these antibodies may be associated with a poor prognosis, which might be the reason 52 53 284 for a higher mortality. Previously, we reported that the survival rate was significantly 54 55 56 285 lower in ANCA-positive IE patients than in ANCA-negative group.13 However, there 57 58 59 286 was no difference in the survival rate between the aPL groups in our study. 60 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 287 5 6 7 288 There were several limitations of this study. First, it was retrospective. The 8 9 289 measurement of autoantibodies was not conducted in all IE patients when they were 10 11 12 290 diagnosed, with only 89 undergoing ANCA and 53 aPL testing, and only 38 patients 13 14 291 were tested for both ANCA and aPL antibodies. Second, the duration of the study 15 16 17 292 spanned 20 years, and some patients were lost to follow-up. Third, although there was 18 For peer review only 19 20 293 a 4-month observation period for patients’ mortality from the day of diagnosis during 21 22 294 hospitalization, we did not register deaths occurring in other hospitals or patients who 23 24 25 295 died at home. 26 27 296 28 29 30 297 Conclusions 31 32 33 298 The present study supports the hypothesis that rheumatic complications are frequent 34 35 299 in patients with IE. For rheumatologists, it is indispensable to routinely exclude IE in 36

37 http://bmjopen.bmj.com/ 38 300 patients with rheumatic features, such as arthritis, splenomegaly, myalgia, proteinuria 39 40 301 and haematuria. For cardiologists, more attention should be paid to the measurement 41 42 43 302 of autoantibodies in IE patients as this might contribute to the evaluation of prognosis. 44

45 on September 28, 2021 by guest. Protected copyright. 46 303 47 48 304 49 50 51 305 52 53 306 References 54 55 56 307 1. Fedeli U, Schievano E, Buonfrate D, et al. Increasing incidence and mortality of infective 57 58 59 308 endocarditis: a population-based study through a record-linkage system. BMC Infect Dis 60 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 309 2011;11:48. 5 6 7 310 2. Monteiro TS, Correia MG, Golebiovski WF, et al. Asymptomatic and symptomatic embolic 8 9 311 events in infective endocarditis: associated factors and clinical impact. Braz J Infect Dis 10 11 12 312 2017;21:240-247. 13 14 313 3. Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC Guidelines for the management of 15 16 17 314 infective endocarditis: The Task Force for the Management of Infective Endocarditis of the 18 For peer review only 19 20 315 European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic 21 22 316 Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J 23 24 25 317 2015;36:3075-3128. 26 27 318 4. Spaulding AR, Salgado-Pabon W, Kohler PL, et al. Staphylococcal and Streptococcal 28 29 30 319 Superantigen Exotoxins. Clin Microbiol Rev 2013;26:422-447. 31 32 33 320 5. Misra DP, Chowdury AC, Edavalath S, et al. Endocarditis: the great mimic of rheumatic 34 35 321 diseases. Trop Doct 2016;46:180-186. 36

37 http://bmjopen.bmj.com/ 38 322 6. Langlois V, Lesourd A, Girszyn N, et al. Antineutrophil Cytoplasmic Antibodies Associated 39 40 323 With Infective Endocarditis. Medicine (Baltimore) 2016;95:e2564. 41 42 43 324 7. Kupferwasser LI, Hafner G, Mohr-Kahaly S, et al. The presence of infection-related 44

45 on September 28, 2021 by guest. Protected copyright. 46 325 antiphospholipid antibodies in infective endocarditis determines a major risk factor for embolic 47 48 326 events. J Am Coll Cardiol 1999;33:1365-71. 49 50 51 327 8. Peng H, Chen WF, Wu C, et al. Culture-negative subacute bacterial endocarditis masquerades 52 53 328 as granulomatosis with polyangiitis (Wegener's granulomatosis) involving both the kidney and 54 55 56 329 lung. BMC Nephrol 2012;13:174. 57 58 59 330 9. Durack DT, Lukes AS, Bright DK, Duke Endocarditis Service. New criteria for diagnosis of 60 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 331 infective endocarditis: utilization of specific echocardiographic findings. Am J Med 1994; 96:200– 5 6 7 332 9.) 8 9 333 10. Yancy CW, Jessup M, Bozkurt B, et al., American College of Cardiology Foundation; 10 11 12 334 American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for 13 14 335 the management of heart failure: a report of the American College of Cardiology 15 16 17 336 Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 18 For peer review only 19 20 337 2013;62:e147–e239. 21 22 338 11. El-Chakhtoura N, Yasmin M, Kanj SS, et al. A 27-year experience with infective endocarditis 23 24 25 339 in Lebanon. J Infect Public Health 2017;10:734-739. 26 27 340 12. Ferraris L, Milazzo L, Ricaboni D, et al. Profile of infective endocarditis observed from 2003 - 28 29 30 341 2010 in a single center in Italy. BMC Infect Dis 2013;13:545. 31 32 33 342 13. Ying CM, Yao DT, Ding HH, et al. Infective endocarditis with antineutrophil cytoplasmic 34 35 343 antibody: report of 13 cases and literature review. PLoS One 2014;9:e89777. 36

37 http://bmjopen.bmj.com/ 38 344 14. Şimşek-Yavuz S, Şensoy A, Kaşıkçıoğlu H, et al. Infective endocarditis in Turkey: aetiology, 39 40 345 clinical features, and analysis of risk factors for mortality in 325 cases. International Journal of 41 42 43 346 Infectious Diseases 2015, 30:106-114. 44

45 on September 28, 2021 by guest. Protected copyright. 46 347 15. Ghosh S, Sahoo R, Nath RK, et al. A Study of Clinical, Microbiological, and 47 48 348 Echocardiographic Profile of Patients of Infective Endocarditis. Int Sch Res Notices 49 50 51 349 2014;2014:1-9. 52 53 350 16. Mahr A, Batteux F, Tubiana S, et al. Brief report: prevalence of antineutrophil cytoplasmic 54 55 56 351 antibodies in infective endocarditis. Arthritis Rheumatol 2014;66:1672-7. 57 58 59 352 17. Aslangul E, Goulvestre C, Mallat Z, et al. Human bartonella infective endocarditis is 60 16

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 353 associated with high frequency of antiproteinase 3 antibodies. J Rheumatol 2014;41:408-10. 5 6 7 354 18. Boils CL, Nasr SH, Walker PD, et al. Update on endocarditis-associated glomerulonephritis. 8 9 355 Kidney Int 2015;87:1241-9. 10 11 12 356 19. Million M, Walter G, Bardin N, et al. Immunoglobulin G anticardiolipin antibodies and 13 14 357 progression to Q fever endocarditis. Clin Infect Dis 2013;57:57-64. 15 16 17 358 20. Million M, Bardin N, Bessis S, et al. Thrombosis and antiphospholipid antibody syndrome 18 For peer review only 19 20 359 during acute Q fever: A cross-sectional study. Medicine (Baltimore) 2017;96:e7578. 21 22 360 21. Selton-Suty C, Maigrat CH, Devignes J, et al. Possible relationship between antiphospholipid 23 24 25 361 antibodies and embolic events in infective endocarditis. Heart 2018;104:509-516. 26 27 362 28 29 30 363 Contribution 31 32 33 364 For authors, Jialin Teng, Honglei Liu, Xiaobing Cheng, Huihui Chi and Fan Wang 34 35 365 collected the clinical data. Yutong Su and Yue Sun analyzed the data. Zhuochao Zhou 36

37 http://bmjopen.bmj.com/ 38 366 and Junna Ye wrote the manuscript. Wei Jin and Chengde Yang designed the study 39 40 367 and critically read the manuscript. We would like to thank the patients for their 41 42 43 368 participation in this study. 44

45 on September 28, 2021 by guest. Protected copyright. 46 369 47 48 370 Data availability statement 49 50 51 52 371 All data relevant to the study are included in the article or uploaded as supplementary 53 54 55 372 information. 56 57 58 373 Declaration of conflicting interests 59 60 17

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 374 The authors report no relationships that could be construed as a conflict of interest. 5 6 7 375 8 9 376 Funding 10 11 12 377 This work is supported by National Natural Science Foundation of China 13 14 378 (81871272, 81801592, 81370397, 81670266), Science and Technology Commission 15 16 17 379 of Shanghai Municipality (17140902500), Shanghai Sailing Program (18YF1414100), 18 For peer review only 19 20 380 Shanghai Jiao Tong University Interdisciplinary Research Project (YG2016QN60), 21 22 381 Excellent Youth B Project (GCQN-2017-B05) and Innovative research team of 23 24 25 382 high-level local universities in Shanghai. 26 27 383 28 29 30 384 Acknowledgement 31 32 33 385 We would like to thank the patients and patient advisers for their participation in this 34 35 386 study. We also thank Dr Shuzhen Xiao for her critical reading of this manuscript. 36

37 http://bmjopen.bmj.com/ 38 387 39 40 388 41 42 43 389 44

45 390 Table 1 on September 28, 2021 by guest. Protected copyright. 46 391 General characteristics of infective endocarditis (IE) patients. 47 48 Item N (%) 49 N 432 50 Female/male 154(36)/278(64) 51 52 Age (years, mean±SD) 45.61±16.12 53 Duration (months, median, interquartile 1.56(0.75,3) 54 range) 55 Predisposing factors 56 57 Dialysis 6(1) 58 Dental procedure 4(1) 59 Intravenous drug use 2(0.5) 60 18

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 27

1 2

3 Congenital heart disease 66(15) BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Rheumatic valvulopathy 58(13) 6 Complications 7 Congestive heart failure 141(33) 8 Hypertension 82(19) 9 10 Diabetes 27(6) 11 Cancer 11(3) 12 Embolic events 13 14 Brain 67(14) 15 Spleen 26(6) 16 Kidney 8(2) 17 18 ExtremityFor peer review 8(2)only 19 Coronary 4(1) 20 Lung 4(1) 21 Infected valve 22 23 Prosthetic valve 42(10) 24 Native valve 25 Mitral valve 234(54) 26 27 Aortic valve 185(43) 28 Tricuspid valve 17(4) 29 Pulmonary valve 10(2) 30 31 Outcome 32 Cardiac surgery 346(80) 33 Relapse 13(3) 34 In-hospital mortality 55(13) 35 36 392 Note: standard deviation (SD).

37 393 http://bmjopen.bmj.com/ 38 Table 2 39 394 40 395 Antibodies in 38 patients tested for both ANCA and aPL. 41 N 42 43 Only aPL-positive 15 44 ACL-positive 1(IgM ACL)

45 aβ2GP1-positive 1 on September 28, 2021 by guest. Protected copyright. 46 LAC-positive 10 47 48 ACL+LAC 3(2 IgM ACL +1 IgA ACL) 49 Only ANCA-positive 9 50 Both aPL- and ANCA-positive 3(LAC, LAC+IgM ACL, IgM ACL) 51 52 396 Note: anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 53 397 antibodies (anti-β2GPI), antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody 54 398 (aPL). 55 56 399 57 400 Table 3 58 401 Clinical features of patients with positive or negative ANCA and aPL. 59 60 ANCA aPL 19

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 27 BMJ Open

1 2

3 ANCA(+) ANCA(-) p valve aPL(+) aPL(-) p valve BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 N 21 68 / 21 32 / 6 Positive Blood culture 13(62) 37(58) 0.741 10(50) 17(53) 0.826 7 Fever 19(91) 57(84) 0.688 18(86) 23(72) 0.400 8 9 Arthritis 11(52) 13(19) 0.003 5(24) 8(25) 0.922 10 Fatigue 3(14) 16(24) 0.549 7(33) 9(28) 0.686 11 Pleural effusion 3(14) 15(22) 0.642 6(29) 9(28) 0.972 12 13 Pericardial effusion 4(19) 12(18) 0.884 6(29) 8(25) 0.773 14 Thrombosis 3(14) 17(25) 0.466 8(38) 10(31) 0.607 15 Myalgia 0 6(11) 0.672 2(10) 2(7) 1.000 16 Splenomegaly 6(29) 20(29) 0.941 5(24) 6(19) 0.922 17 18 Heart murmur For10(100) peer43(78) review0.233 only18(86) 28(88) 1.000 19 Weight loss 3(30) 9(16) 0.562 3(14) 7(22) 0.740 20 Cardiac surgery 5(50) 42(72) 0.295 16(76) 19(59) 0.206 21 22 In-hospital mortality 4(19) 3(4) 0.087 2(10) 4(13) 1.000 23 402 Note: antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 24 403 25 26 404 Table 4 27 405 Laboratory features of patients positive or negative for ANCA and aPL. 28 ANCA aPL 29 30 ANCA(+) ANCA(-) p value aPL(+) aPL(-) p value 31 N 21 68 / 21 32 / 32 WBC (*109/L) 8.39±2.53 8.56±4.19 0.896 7.66±2.54 8.58±4.52 0.403 33 34 Hb (g/L) 87.57±26.83 105.51±18.76 0.001 99.57±15.28 102.34±26.22 0.664 35 PLT (*109/L) 194.70±110.91 201.25±127.06 0.879 184.95±83.86 193.47±151.31 0.819 36 LDH (IU/L) 206.125±53.45 252.00±96.73 0.200 278.87±296.11 247.72±103.72 0.632

37 http://bmjopen.bmj.com/ 38 ESR (mm/h) 64.76±36.87 51.85±33.47 0.135 69.14±34.82 39.09±33.48 0.003 39 IgG (mg/dl) 2254.29±1001.40 1660.05±528.67 0.172 1959.69±747.29 1751.67±707.30 0.379 40 IgA (mg/dl) 232.29±89.77 301.50±131.20 0.188 276.63±94.03 283.38±130.21 0.859 41 42 IgM (mg/dl) 312.71±183.51 140.70±64.03 0.048 193.07±119.30 153.79±119.39 0.324 43 IgE (mg/dl) 37.00±22.39 354.85±569.90 0.286 357.31±437.81 351.73±635.24 0.983 44 ACL-positive 2(25) 2(5) 0.436 5(24) 0 0.004

45 on September 28, 2021 by guest. Protected copyright. aβ2GPI-positive 0 1(5) 1.000 2(13) 0 0.057 46 47 LAC-positive 2(29) 11(38) 0.981 17(81) 0 0.000 48 ANA-positive 2(15) 6(12) 1.000 3(16) 3(11) 0.618 49 Elevated CRP 18(90) 51(86) 0.851 14(74) 21(75) 0.921 50 51 Elevated RF 10(59) 14(32) 0.053 6(31) 6(25) 0.336 52 406 Note: White blood cell counts (WBC), haemoglobin (Hb), platelets (PLT), erythrocyte 53 407 sedimentation rate (ESR), serum lactate dehydrogenase (LDH), immunoglobulin G (IgG), 54 55 408 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), anticardiolipin 56 409 antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I antibodies (aβ2GPI), 57 410 antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 58 59 411 60 20

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 27

1 2

3 412 Figure 1 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 413 Kaplan-Meier survival curves. 6 414 a. Kaplan-Meier survival curves of patients who were tested for aPL; b. Kaplan-Meier survival 7 415 curves of patients who were tested for ANCA. 8 9 416 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 27 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

1 2

3 Supplemental Table 1 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Pathogen and treatment of infective endocarditis patients. 6 Item N (%) 7 Pathogen* 8 9 Culture negative 222(51) 10 Gram-positive 11 Streptococcus viridans 25(6) 12 13 Streptococcus sanguis 24(6) 14 Streptococcus mitis 23(5) 15 β-hemolytic streptococci 6(1) 16 17 Other Streptococcus 32(7) 18 StaphylococcusFor aureuspeer review 64(1only5) 19 Staphylococcus epidermidis 30(7) 20 Enterococci 26(6) 21 22 Lactococcus lactis 4(1) 23 Gemella mobillorum 4(1) 24 Aerococcus urinae 3(1) 25 26 Micrococcus 2(0.5) 27 Aerococcus viridans 2(0.5) 28 Kocuria sp 4(1) 29 30 Gram-negative 31 Enterobacter 20(5) 32 Stenotrophomonas maltophilia 2(0.5) 33 Other Gram-negative Bacillus 18(4) 34 35 Fungi 36 Candida albicans 9(2) 37 *Several patients had more than one microorganism. http://bmjopen.bmj.com/ 38 39 40 Supplemental Table 2 41 Clinical manifestations of infective endocarditis patients at time of IE diagnosis. 42 43 Item N (%) 44 Heart murmur 343(79)

45 on September 28, 2021 by guest. Protected copyright. Fever 364(84) 46 47 Arthritis 65(15) 48 Fatigue 86(20) 49 Pleural effusion 119 (28) 50 51 Pericardial effusion 103(24) 52 Myalgia 28(7) 53 Hepatomegaly 19(4) 54 55 Splenomegaly 99(23) 56 Peripheral polyneuropathy 4(1) 57 Janeway lesion 18(4) 58 Osler node 11(3) 59 60 Roth spot 5(1)

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 25 of 27 BMJ Open

1 2

3 Supplemental Table 3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Clinical features and laboratory features of patients tested with antibodies or not. 6 Patients tested with antibodies Patients not tested with antibodies p valve 7 N 104 328 / 8 9 Female/male 40/64 114/214 0.492 10 Age (years, mean±SD) 45.65±14.22 45.59±16.97 0.973 11 Positive Blood culture 56(57) 154(47) 0.108 12 13 Fever 89(86) 275(84) 0.758 14 Arthritis 27(26) 41(13) 0.001 15 Fatigue 26(25) 67(20) 0.323 16 17 Pleural effusion 21(20) 98(30) 0.054 18 Pericardial effusion For19 (18)peer review84(26) only 0.126 19 Thrombosis 25(24) 116(35) 0.032 20 Myalgia 7(9) 23(7) 0.572 21 22 Splenomegaly 29(28) 70(21) 0.167 23 Heart murmur 67(84) 288(88) 0.333 24 Weight loss 14(18) 47(14) 0.476 25 26 Cardiac surgery 54(65) 290(88) 0.000 27 In-hospital mortality 9(9) 46(14) 0.178 28 WBC (*109/L) 8.39±2.53 8.39±2.53 0.614 29 30 Hb (g/L) 87.57±26.83 87.57±26.83 0.105 9 31 PLT (*10 /L) 194.70±110.91 194.70±110.91 0.659 32 ESR (mm/h) 64.76±36.87 64.76±36.87 0.116 33 Proteinuria 27(27) 85(26) 0.788 34 35 hematuria 50(51) 133(41) 0.079 36 Note: White blood cell counts (WBC), hemoglobin (Hb), platelet (PLT), erythrocyte

37 http://bmjopen.bmj.com/ sedimentation rate (ESR). 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50

51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 27

1 2

3 Supplemental Figure 1 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Distribution of patients diagnosed with infectious endocarditis. 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 * Patients who were either tested with aPL or ANCA were included and compared in different 34 35 groups. 36 Note: infectious endocarditis (IE), antineutrophil cytoplasmic antibody (ANCA),

37 anti-phospholipid antibody (aPL). http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 27 BMJ Open

1 2 STROBE Statement—checklist of items that should be included in reports of observational studies

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Item Page 6 No Recommendation 7 Title and abstract 1 (a) Indicate the study’s design with a commonly used term in the title or 1-2 8 the abstract 9 10 (b) Provide in the abstract an informative and balanced summary of what 2-3 11 was done and what was found 12 Introduction 13 14 Background/rationale 2 Explain the scientific background and rationale for the investigation 4-5 15 being reported 16 Objectives 3 State specific objectives, including any prespecified hypotheses 5 17 18 Methods For peer review only 19 Study design 4 Present key elements of study design early in the paper 5-6 20 Setting 5 Describe the setting, locations, and relevant dates, including periods of 5-6 21 recruitment, exposure, follow-up, and data collection 22 23 Participants 6 (a) Cohort study—Give the eligibility criteria, and the sources and 5-6 24 methods of selection of participants. Describe methods of follow-up 25 Case-control study—Give the eligibility criteria, and the sources and 26 27 methods of case ascertainment and control selection. Give the rationale 28 for the choice of cases and controls 29 Cross-sectional study—Give the eligibility criteria, and the sources and 30 methods of selection of participants 31 32 (b) Cohort study—For matched studies, give matching criteria and / 33 number of exposed and unexposed 34 Case-control study—For matched studies, give matching criteria and the 35 number of controls per case 36 Variables 7 Clearly define all outcomes, exposures, predictors, potential confounders, 6 37 http://bmjopen.bmj.com/ 38 and effect modifiers. Give diagnostic criteria, if applicable 39 Data sources/ 8* For each variable of interest, give sources of data and details of methods 7 40 measurement of assessment (measurement). Describe comparability of assessment 41 42 methods if there is more than one group 43 Bias 9 Describe any efforts to address potential sources of bias 7 44 Study size 10 Explain how the study size was arrived at 6

45 on September 28, 2021 by guest. Protected copyright. 46 Quantitative variables 11 Explain how quantitative variables were handled in the analyses. If 7 47 applicable, describe which groupings were chosen and why 48 Statistical methods 12 (a) Describe all statistical methods, including those used to control for 7 49 confounding 50 51 (b) Describe any methods used to examine subgroups and interactions 7 52 (c) Explain how missing data were addressed 7 53 (d) Cohort study—If applicable, explain how loss to follow-up was 7 54 addressed 55 56 Case-control study—If applicable, explain how matching of cases and 57 controls was addressed 58 Cross-sectional study—If applicable, describe analytical methods taking 59 account of sampling strategy 60 (e) Describe any sensitivity analyses 7

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 27

1 2 Results

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 Participants 13* (a) Report numbers of individuals at each stage of study—eg numbers potentially 8-9 5 eligible, examined for eligibility, confirmed eligible, included in the study, 6 completing follow-up, and analysed 7 (b) Give reasons for non-participation at each stage / 8 9 (c) Consider use of a flow diagram 10 suppl 10 fig1 11 Descriptive data 14* (a) Give characteristics of study participants (eg demographic, clinical, social) and 8-10 12 information on exposures and potential confounders 13 14 (b) Indicate number of participants with missing data for each variable of interest 8-10 15 (c) Cohort study—Summarise follow-up time (eg, average and total amount) 7 16 Outcome data 15* Cohort study—Report numbers of outcome events or summary measures over time 7 17 18 Case-controlFor peer study—Report review numbers in each exposure only category, or summary / 19 measures of exposure 20 Cross-sectional study—Report numbers of outcome events or summary measures / 21 Main results 16 (a) Give unadjusted estimates and, if applicable, confounder-adjusted estimates 7 22 23 and their precision (eg, 95% confidence interval). Make clear which confounders 24 were adjusted for and why they were included 25 (b) Report category boundaries when continuous variables were categorized / 26 27 (c) If relevant, consider translating estimates of relative risk into absolute risk for a / 28 meaningful time period 29 Other analyses 17 Report other analyses done—eg analyses of subgroups and interactions, and 7-10 30 sensitivity analyses 31 32 Discussion 33 Key results 18 Summarise key results with reference to study objectives 7-10 34 Limitations 19 Discuss limitations of the study, taking into account sources of potential bias or 13-14 35 36 imprecision. Discuss both direction and magnitude of any potential bias

37 Interpretation 20 Give a cautious overall interpretation of results considering objectives, limitations, 11-14 http://bmjopen.bmj.com/ 38 multiplicity of analyses, results from similar studies, and other relevant evidence 39 Generalisability 21 Discuss the generalisability (external validity) of the study results / 40 41 Other information 42 Funding 22 Give the source of funding and the role of the funders for the present study and, if 17 43 44 applicable, for the original study on which the present article is based

45 on September 28, 2021 by guest. Protected copyright. 46 *Give information separately for cases and controls in case-control studies and, if applicable, for exposed and 47 unexposed groups in cohort and cross-sectional studies. 48 49 50 Note: An Explanation and Elaboration article discusses each checklist item and gives methodological background and 51 published examples of transparent reporting. The STROBE checklist is best used in conjunction with this article (freely 52 available on the Web sites of PLoS Medicine at http://www.plosmedicine.org/, Annals of Internal Medicine at 53 54 http://www.annals.org/, and Epidemiology at http://www.epidem.com/). Information on the STROBE Initiative is 55 available at www.strobe-statement.org. 56 57 58 59 60

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

Clinical Characteristics of Infective Endocarditis in Patients with Antineutrophil Cytoplasmic Antibody or Antiphospholipid Antibody: A Retrospective Study in Shanghai

ForJournal: peerBMJ Open review only Manuscript ID bmjopen-2019-031512.R3

Article Type: Original research

Date Submitted by the 16-Dec-2019 Author:

Complete List of Authors: Zhou, Zhuochao; Shanghai Jiao Tong University School of Medicine; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Ye, Junna; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Teng, Jialin; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Liu, Honglei; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Cheng, Xiaobing; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Sun, Yue; Shanghai Jiao Tong University Medical School Affiliated Ruijin http://bmjopen.bmj.com/ Hospital, Rheumatology and Immunology Su, Yutong; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Chi, Huihui; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Wang, Fan; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Rheumatology and Immunology Yang, Chengde; Shanghai Jiao Tong University Medical School Affiliated

Ruijin Hospital, Rheumatology and Immunology on September 28, 2021 by guest. Protected copyright. Jin, Wei; Shanghai Jiao Tong University Medical School Affiliated Ruijin Hospital, Cardiology

Primary Subject Rheumatology Heading:

Secondary Subject Heading: Rheumatology

Infective endocarditis, Rheumatic manifestation, ANCA, Antiphospholipid Keywords: antibody

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 27 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

33 http://bmjopen.bmj.com/ 34 35 36 37 38 39 40

41 on September 28, 2021 by guest. Protected copyright. 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 27

1 2

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 1 Clinical Characteristics of Infective Endocarditis in Patients with Antineutrophil 5 6 7 2 Cytoplasmic Antibody or Antiphospholipid Antibody: A Retrospective Study in 8 9 3 Shanghai 10 11 12 4 13 14 5 Zhuochao Zhou1,*, Junna Ye1,*, Jialin Teng1, Honglei Liu1, Xiaobing Cheng1, Yue 15 16 17 6 Sun1, Yutong Su1, Huihui Chi1, Fan Wang1, Chengde Yang1,#, Wei Jin2,# 18 For peer review only 19 20 7 21 22 8 1Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 23 24 25 9 University School of Medicine; 26 27 10 2Department of Cardiology, Ruijin Hospital, Shanghai Jiao Tong University School 28 29 30 11 of Medicine 31 32 33 12 34 35 13 *These authors contributed equally to this work. 36

37 http://bmjopen.bmj.com/ 38 14 39 40 15 #Correspondence: 41 42 43 16 Wei Jin, MD, PhD 44

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 23 Chengde Yang, MD, PhD 5 6 7 24 Department of Rheumatology and Immunology, Ruijin Hospital, Shanghai Jiao Tong 8 9 25 University School of Medicine, No. 197 Ruijin Second Road, Huangpu District, 10 11 12 26 Shanghai 200025, China. 13 14 27 Tel.: +86 21 64370045; 15 16 17 28 Fax: +86 21 34186000; 18 For peer review only 19 20 29 E-mail: [email protected] 21 22 30 23 24 25 31 26 27 32 Abstract 28 29 30 33 Objective: This study aimed to characterize rheumatic manifestations and 31 32 33 34 autoantibodies in 432 patients diagnosed with infective endocarditis (IE) in Shanghai. 34 35 35 Design, setting and participants: A retrospective study was conducted in Ruijin 36

37 http://bmjopen.bmj.com/ 38 36 Hospital from 1997 to 2017. The clinical and laboratory characteristics of a total of 39 40 37 432 patients were analysed. In addition, the differences between patients with positive 41 42 43 38 and negative antineutrophil cytoplasmic antibodies (ANCA) and antiphospholipid 44

45 on September 28, 2021 by guest. Protected copyright. 46 39 antibodies (aPL) as well as the survival rates of these patients were compared. 47 48 40 Results: A total of 432 patients, including 278 male patients and 154 female patients, 49 50 51 41 were included. The mean age of the patients was 46±16 years. A total of 346 patients 52 53 42 (80%) had cardiac surgery, and 55 patients (13%) died in the hospital. Among the IE 54 55 56 43 patients, 104 were tested for either ANCA or aPL and were analysed in different 57 58 59 44 groups. Twenty-one (24%) positive ANCA patients were proteinase 3(PR3)-ANCA 60 2

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 45 positive. Compared with the ANCA-negative group, patients with positive ANCA had 5 6 7 46 higher IgM (P=0.048), lower haemoglobin (P=0.001), and a higher likelihood of 8 9 47 arthritis (P=0.003). Twenty-one (40%) aPL-positive patients had a higher erythrocyte 10 11 12 48 sedimentation rate (ESR) than was found in the aPL-negative group (P=0.003). In 13 14 49 addition, the survival rate of the ANCA-positive IE patients was lower (P=0.032) than 15 16 17 50 that of the ANCA-negative group, while there was no difference between patients 18 For peer review only 19 20 51 with or without aPL antibodies (P=0.728). 21 22 52 Conclusion: The present study supports the claim that rheumatic manifestations and 23 24 25 53 autoantibodies are frequently present in patients with IE and might lead to early 26 27 54 misdiagnosis. Physicians should pay more attention to the measurement of 28 29 30 55 autoantibodies in these patients. 31 32 33 56 34 35 57 Keywords: infective endocarditis, rheumatic manifestation, ANCA, aPL 36

37 http://bmjopen.bmj.com/ 38 58 39 40 59 41 42 43 44 60 Strengths and limitations of this study

45 on September 28, 2021 by guest. Protected copyright. 46 47 61 1. It was a retrospective study in a tertiary clinical center spanned over twenty years. 48 49 62 2. This study filled the gap in the study of the IE population related to ANCA or aPL 50 51 52 63 antibodies in China. 53 54 55 64 3. The study analysed the clinical characteristics and survival rates of IE patients with 56 57 65 and without ANCA or aPL antibodies. 58 59 60 66 4. As it was a retrospective study, not all patients underwent analysis for ANCA and 3

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 67 aPL autoantibodies. 5 6 7 68 5. Deaths occurring at home or in other hospitals were not registered. 8 9 69 10 11 12 70 13 14 15 71 Introduction 16 17 72 Infective endocarditis (IE) is a microbial infection of the endocardium that usually 18 For peer review only 19 20 73 involves the heart valves. Despite being relatively rare worldwide, IE has been 21 22 74 reported in a steady incidence over the past three decades ranging from 2-6 per 23 24 25 75 100,000 individuals in the general population per year and has a considerable 26 27 76 associated mortality rate that varies from 10% to 30%.1 Predisposing factors for IE 28 29 30 77 include conditions such as congenital heart diseases, rheumatic valve disease, 31 32 33 78 artificial valves, and other factors, such as intravenous drug use, dental procedures 34 35 79 and haemodialysis.2 The diagnosis of IE is made based on symptoms, blood 36

37 http://bmjopen.bmj.com/ 38 80 cultures, echocardiography and pathological biopsy from valve surgery.3 39 40 81 However, superantigens induced by some bacteria, such as Staphylococcus 41 42 43 82 aureus, can stimulate an immune response, which could interfere with antibody 44

45 4 on September 28, 2021 by guest. Protected copyright. 46 83 production. Rheumatic manifestations, such as myalgia, arthralgia, and arthritis, 47 48 84 are prevalent, occurring in nearly 40% of patients with IE at presentation of weeks 49 50 51 85 to months before the diagnosis of IE.5 When classic IE manifestations are less 52 53 86 evident, patients can be misdiagnosed as having a rheumatological disease, which 54 55 56 87 might lead to delayed initiation of antibiotic treatment. Thus, the distinction 57 58 59 88 between IE and rheumatic diseases is not always easy, and to improve the 60 4

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 89 understanding concerning differences between these conditions is of great 5 6 7 90 importance. 8 9 91 10 11 12 92 Many specific antibodies, such as antineutrophil cytoplasmic antibodies (ANCA) and 13 14 93 anti-phospholipid antibody (aPL), might be related to the pathophysiology of IE. 6-7 15 16 17 94 Due to positive ANCA tests, IE has been reported to mimic ANCA-associated 18 For peer review only 19 20 95 vasculitis (AAV). Patients with IE may present with multiple pulmonary nodules and 21 22 96 glomerulonephritis, which mimics granulomatosis with polyangiitis.8 One study 23 24 25 97 further underscored that ANCA might be associated with multiple valve 26 27 98 involvement.6 Additionally, infection-associated elevated aPL levels in patients with 28 29 30 99 infective endocarditis are related to endothelial cell activation, thrombin generation 31 32 33 100 and impairment of fibrinolysis, which may contribute to the increased risk of major 34 35 101 embolic events in these patients.7 Therefore, ANCA and aPL are not specific to IE 36

37 http://bmjopen.bmj.com/ 38 102 patients. 39 40 103 41 42 43 104 The objective of this study was to compare the clinical characteristics and survival 44

45 on September 28, 2021 by guest. Protected copyright. 46 105 rates of IE patients with and without ANCA or aPL autoantibodies in a tertiary 47 48 106 hospital in Shanghai, China from 1997 to 2017. 49 50 51 107 52 53 54 108 Methods 55 56 109 Patients 57 58 59 110 This study was a retrospective study. We continuously included 432 patients 60 5

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 111 hospitalized in Ruijin Hospital affiliated to Shanghai Jiao Tong University School of 5 6 7 112 Medicine who were diagnosed with IE from 1997 to 2017 after ruling out 33 patients 8 9 113 with primary rheumatic disease or in immunosuppression condition. The clinical and 10 11 12 114 laboratory data of the patients were mainly obtained from the medical records system 13 14 115 of Ruijin hospital. 15 16 17 116 18 For peer review only 19 20 117 Only patients with definite IE determined by either the Duke criteria (up to 2000) or 21 22 118 the modified Duke criteria (from 2000 onwards) were included.3,9 Duration of disease 23 24 25 119 was defined as the period from the time when the patients presented with the first 26 27 120 clinical feature to a definite IE diagnosis. In patients with long-term fever, the 28 29 30 121 autoantibodies were tested to rule out rheumatic diseases. Cardiac surgery and death 31 32 33 122 were recorded during the period of hospitalization. The study followed the ethical 34 35 123 standards for human experimentation established in the Declaration of Helsinki and 36

37 http://bmjopen.bmj.com/ 38 124 was approved by the Institutional Research Ethics Committee of Ruijin Hospital (ID: 39 40 125 2016-62), Shanghai, China. 41 42 43 126 44

45 on September 28, 2021 by guest. Protected copyright. 46 127 In this study, patients positive for either myeloperoxidase (MPO)-ANCA or 47 48 128 proteinase 3(PR3)-ANCA were defined as ANCA-positive patients. Patients with any 49 50 51 129 positive test for anticardiolipin antibodies (ACL), lupus anticoagulant (LAC) or 52 53 130 anti-β2 glycoprotein I antibodies (aβ2GPI) were defined as aPL-positive patients. 54 55 56 131 Congestive heart failure was defined according to the New York Heart Association 57 58 10 59 132 classification system . In-hospital mortality was defined as IE patients who died 60 6

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 133 during hospitalization. There was a 4-month observation period for patient mortality 5 6 7 134 from the day of diagnosis. The survival rates of subgroups positive or negative for 8 9 135 ANCA and aPL were then analysed. 10 11 12 136 13 14 137 Laboratory features and echocardiography 15 16 17 138 Levels of anti-PR3, anti-MPO and aPL antibodies in serum were measured by enzyme 18 For peer review only 19 20 139 linked immunosorbent assay (ELISA). The following laboratory data were recorded: 21 22 140 white blood cell counts in blood (WBC), haemoglobin (Hb), platelets (PLT), 23 24 25 141 C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), lactate 26 27 142 dehydrogenase (LDH), rheumatoid factor (RF), immunoglobulin G (IgG), 28 29 30 143 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), 31 32 33 144 anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 34 35 145 antibodies (aβ2GPI), anti-PR3 antibodies (PR3-ANCA), anti-MPO antibodies 36

37 http://bmjopen.bmj.com/ 38 146 (MPO-ANCA), haematuria, and proteinuria. All IE patients underwent transthoracic 39 40 147 or transoesophageal echocardiography. Furthermore, each patient underwent 41 42 43 148 abdominal ultrasounds. 44

45 on September 28, 2021 by guest. Protected copyright. 46 149 47 48 150 Statistical analysis 49 50 51 151 Data were analysed using the Statistical Package for the Social Sciences for Windows 52 53 152 (version 23.0; SPSS, IBM, Chicago, IL, USA). Statistical analyses were performed by 54 55 56 153 t test or Chi-square test according to the type of data (continuous or categorical, 57 58 59 154 respectively). Survival curves were obtained using Kaplan-Meier’s method. 60 7

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 155 Significance was obtained by a log-rank test. A p value of less than 0.05 was 5 6 7 156 considered statistically significant. 8 9 157 10 11 12 158 Patient and public involvement 13 14 159 Patients were not informed by the development of the research question and outcome 15 16 17 160 measures. Patients were not involved in the design, the recruitment to and conduct of 18 For peer review only 19 20 161 the study as the study was retrospective. The results were not shared with study 21 22 162 participants. 23 24 25 163 26 27 28 164 Results 29 30 165 Characteristics of Patients 31 32 33 166 As shown in Table 1, a total of 432 patients, including 278 (64%) male patients and 34 35 167 154 (36%) female patients (male:female ratio=1.8:1), were analysed. The mean age ± 36

37 http://bmjopen.bmj.com/ 38 168 standard deviation (SD) of patients was 46±16 years. The median duration of disease 39 40 169 was 1.56 (0.75, 3) months. The mitral valve (54%) was the most frequently involved 41 42 43 170 valve followed by the aortic (43%) valve. The most common complication of IE 44

45 on September 28, 2021 by guest. Protected copyright. 46 171 patients was congestive heart failure in 141 patients (33%). For embolic events 47 48 172 discovered or presented in the hospital, brain (14%) and spleen (6%) infarction were 49 50 51 173 most commonly seen. For predisposing factors, dialysis accounted for 1%, dental 52 53 174 procedures accounted for 1%, and intravenous drug use (IVDU) accounted for 0.5%. 54 55 56 175 Additionally, 58 (13%) patients had rheumatic valvulopathy. During hospitalization, 57 58 59 176 346 patients (80%) had cardiac surgery, and 55 patients (13%) died. The pathogens 60 8

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 177 associated with IE patients are listed in supplemental table 1. 5 6 7 178 8 9 179 Clinical manifestations and laboratory features 10 11 12 180 Fever (84%) and new heart murmur (79%) were present in most patients with IE 13 14 181 (Supplemental Table 2). The most common rheumatic manifestations were arthritis 15 16 17 182 (15%), and myalgia (6%). In addition, a few patients had non-specific manifestations, 18 For peer review only 19 20 183 such as hepatomegaly (4%), splenomegaly (23%) and fatigue (20%). Eighteen 21 22 184 patients had Janeway lesions (4%), 11 patients had Osler nodes (3%) and 5 patients 23 24 25 185 had Roth spot (1%). 26 27 186 28 29 30 187 Blood cultures were positive in 50% of IE patients. Echocardiographic data showed 31 32 33 188 that 266/432 (62%) patients had IE-specific characteristics, such as vegetation, 34 35 189 abscess, pseudoaneurysm, intracardiac valvular perforation and abnormal activity 36

37 http://bmjopen.bmj.com/ 38 190 around the site of the prosthetic valve. Regarding immunologic features, 89 patients 39 40 191 were tested for ANCA, 21 (24%) of these were ANCA-positive, and all of these were 41 42 43 192 PR3-ANCA positive. Fifty-three patients were tested for aPL, and 21 were positive 44

45 on September 28, 2021 by guest. Protected copyright. 46 193 (45%). One patient had positive IgG aPL (2%), 4 patients had IgM aPL (8%), 1 47 48 194 patient had IgA aPL (2%), 2 patients had aβ2GPI (4%) and 17 patients had LAC 49 50 51 195 (32%). Eight (13%) patients were positive for ANA, and one was positive for 52 53 196 extractable nuclear antigen (ENA) (2%). Proteinuria (109/432, 25%) and haematuria 54 55 56 197 (181/432, 42%) were most commonly observed. Of 38 IE patients tested for both aPL 57 58 59 198 and ANCA, 15 were aPL-positive, 9 were ANCA-positive and 3 were positive for 60 9

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 199 both antibodies (Table 2). The clinical features and laboratory features of patients 5 6 7 200 tested for antibodies or not are presented in Supplemental Table 3. Arthritis was more 8 9 201 common in patients tested for antibodies than in those who were not, indicating the 10 11 12 202 relevance between autoantibodies and rheumatic features. 13 14 203 15 16 17 204 Comparison of patients with positive and negative ANCA or aPL 18 For peer review only 19 20 205 Of 432 IE patients, the 104 who were tested for either ANCA or aPL were analysed. 21 22 206 Finally, 89 patients were classified as ‘ANCA-positive IE’ or ‘ANCA-negative IE’, 23 24 25 207 and 53 patients were classified as ‘aPL-positive IE’ or ‘aPL-negative IE’ according to 26 27 208 their antibodies test (Supplemental Figure 1). Thirty-eight patients were tested for 28 29 30 209 both ANCA and aPL antibodies, and their characteristics were analysed. Of 89 31 32 33 210 patients who were tested for ANCA, 21 (24%) were ANCA-positive, and 68 (76%) 34 35 211 were ANCA-negative. Arthritis (P=0.003) was more frequent in ANCA-positive 36

37 http://bmjopen.bmj.com/ 38 212 patients than in ANCA-negative patients, while more cardiac surgery was done in 39 40 213 ANCA-negative patients (P=0.002) (Table 3). Other clinical manifestations, such as 41 42 43 214 fever, fatigue, embolic events, myalgia, splenomegaly, heart murmur, and weight loss, 44

45 on September 28, 2021 by guest. Protected copyright. 46 215 were not significantly different between the two groups (P˃0.05). Compared to 47 48 216 ANCA-negative patients, in ANCA-positive patients, IgM (P=0.048) was higher, Hb 49 50 51 217 (P=0.001) was significantly lower, and elevated RF (P=0.053) seems to be more 52 53 218 common (Table 4). Of 53 patients who were tested for aPL, 21 (40%) were 54 55 56 219 aPL-positive, and the remaining 32 (60%) were aPL-negative. ESR was significantly 57 58 59 220 higher in patients who were positive for aPL than in those who were negative 60 10

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 221 (P=0.003) (Table 4). 5 6 7 222 8 9 223 Survival Rate 10 11 12 224 Overall, 9 patients tested for aPL or ANCA antibodies died in the hospital. Among 13 14 225 these patients, 4 were ANCA-positive, and 2 patients were aPL-positive. One patient 15 16 17 226 died of renal failure, four of acute heart failure, two of septic shock and two of stroke. 18 For peer review only 19 20 227 The survival rate was significantly lower in ANCA-positive IE patients (P =0.032), 21 22 228 but there was no significant difference between the aPL groups (P = 0.728) (Fig. 1). 23 24 25 229 26 27 230 28 29 30 231 Discussion 31 32 11-13 33 232 Although it has been reported about different aspects of IE, the analysis of the 34 35 233 clinical and immunologic features of Chinese IE patients is still limited. The clinical 36

37 http://bmjopen.bmj.com/ 38 234 presentations and laboratory findings of IE are similar to those of many rheumatic 39 40 235 diseases, making it difficult to diagnose. Thus, it is pivotal to analyse the rheumatic 41 42 43 236 manifestations of IE to help physicians differentiate it when encountering these 44

45 on September 28, 2021 by guest. Protected copyright. 46 237 clinical scenarios. In this study, we analysed the clinical and laboratory features of 47 48 238 432 IE patients in Ruijin Hospital and compared the characteristics of 104 patients 49 50 51 239 tested for ANCA or aPL during the past two decades. The mean age of the IE patients 52 53 240 was 46 years. In many studies of IE patients, the mean age was higher than that in our 54 55 56 241 study, while other IE patients were similar to ours.2,14-15 In addition, elderly people 57 58 59 242 were more likely to have valve replacement surgery because of degenerative valve 60 11

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 243 disease. With regard for blood culture, it has been reported that blood culture-negative 5 6 7 244 IE (BCNIE) may occur in up to 31% of all cases of IE and often poses considerable 8 9 245 diagnostic and therapeutic dilemmas.3 BCNIE most commonly arises as a 10 11 12 246 consequence of previous antibiotic administration. As a tertiary hospital, before 13 14 247 hospitalization in Ruijin Hospital, our patients had been treated with antibiotics 15 16 17 248 empirically in other hospitals with uncertain diagnoses. As a result, only 18 For peer review only 19 20 249 approximately 50% of the patients had a positive blood culture. Patients who were 21 22 250 culture-negative were diagnosed based on other criteria, such as echocardiography 23 24 25 251 findings and a postoperative histological examination of vegetation. In addition, our 26 27 252 study showed that ANCA-positive IE patients had higher IgM and lower Hb levels, 28 29 30 253 were more likely to present with arthritis than was found in ANCA-negative IE 31 32 33 254 patients. Moreover, aPL-positive IE patients had significantly higher ESR than was 34 35 255 found in aPL-negative patients. 36

37 http://bmjopen.bmj.com/ 38 256 39 40 257 Positive ANCA and aPL were previously reported to be non-specific in patients with 41 42 43 258 infection.6-8 The phenomenon of IE patients positive for ANCA has long been 44

45 on September 28, 2021 by guest. Protected copyright. 46 259 reported. A study from France reported that 8% of IE patients had PR3-ANCA or 47 48 260 MPO-ANCA, which are associated with young age (P=0.022), echocardiographic 49 50 51 261 vegetation (P=0.043), and elevated IgG levels (P=0.017).16 It has also been reported 52 53 262 that most Gram-positive endocarditis can cause an anti-PR3 specificity. During 54 55 56 263 delayed polymorphonuclear leucocyte apoptosis, PR3 contact with the immune 57 58 17 59 264 system could possibly cause an increase in the production of anti-PR3 ANCA. It 60 12

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 265 was interesting to find that among our IE patients tested for ANCA, all of them were 5 6 7 266 PR3-ANCA-positive. One important presentation of AAV is glomerulonephritis. 8 9 267 However, IE-associated glomerular nephritis often presents with acute kidney 10 11 12 268 injury.18 The most common biopsy finding was necrotizing and crescentic glomerular 13 14 269 nephritis (53%), followed by endocapillary proliferative glomerular nephritis (37%).18 15 16 17 270 In our study, among IE patients, those with proteinuria and haematuria accounted for 18 For peer review only 19 20 271 25% and 42%, respectively. However, the kidney biopsy was not routinely performed 21 22 272 in these patients. 23 24 25 273 26 27 274 Apart from ANCA, aPL was also analysed in our study. In terms of the relationship 28 29 30 275 between aPL and IE, it has been indicated that IgG aPL might be associated with Q 31 32 33 276 fever endocarditis and aPL antibodies has previously been shown to be associated 34 35 277 with thrombosis during acute Q fever.19-20 In our study, aPL-positive patients did not 36

37 http://bmjopen.bmj.com/ 38 278 have positive serology of Coxiella. Furthermore, the presence of IgM aPL and aβ2GPI 39 40 279 was associated with embolic events, especially cerebral events, which could 41 42 43 280 contribute to assessing the embolic risk of IE.21 In our study, there was no difference 44

45 on September 28, 2021 by guest. Protected copyright. 46 281 in thrombosis between aPL-positive patients and aPL-negative patients. APL-positive 47 48 282 IE patients had higher ESR, and LAC was the most commonly positive type of aPL in 49 50 51 283 IE patients. This might indicate that aPL was related to an inflammatory status. 52 53 284 54 55 56 285 In terms of the survival rate found in our study, it was significantly lower in the 57 58 59 286 ANCA-positive IE group than in the ANCA-negative group. ANCA-positive patients 60 13

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 287 were more likely to be mistaken for AAV, leading to a delay in diagnosis. Thus, these 5 6 7 288 antibodies may be associated with a poor prognosis, which might be the reason for a 8 9 289 higher mortality. Previously, we reported that the survival rate was significantly lower 10 11 12 290 in ANCA-positive IE patients than in ANCA-negative group.13 However, there was 13 14 291 no difference in the survival rate between the aPL groups in our study. 15 16 17 292 18 For peer review only 19 20 293 There were several limitations of this study. First, it was retrospective. The 21 22 294 measurement of autoantibodies was not conducted in all IE patients when they were 23 24 25 295 diagnosed, with only 89 undergoing ANCA and 53 aPL testing, and only 38 patients 26 27 296 were tested for both ANCA and aPL antibodies. Second, the duration of the study 28 29 30 297 spanned over 20 years, and some patients were lost to follow-up. Third, although 31 32 33 298 there was a 4-month observation period for patients’ mortality from the day of 34 35 299 diagnosis during hospitalization, we did not register deaths occurring in other 36

37 http://bmjopen.bmj.com/ 38 300 hospitals or patients who died at home. 39 40 301 41 42 43 302 Conclusions 44

45 on September 28, 2021 by guest. Protected copyright. 46 303 The present study supports the hypothesis that rheumatic complications are frequent 47 48 304 in patients with IE. For rheumatologists, it is indispensable to routinely exclude IE in 49 50 51 305 patients with rheumatic features, such as arthritis, splenomegaly, myalgia, proteinuria 52 53 306 and haematuria. For cardiologists, more attention should be paid to the measurement 54 55 56 307 of autoantibodies in IE patients as this might contribute to the evaluation of prognosis. 57 58 59 308 60 14

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 309 5 6 7 310 8 9 311 References 10 11 12 312 1. Fedeli U, Schievano E, Buonfrate D, et al. Increasing incidence and mortality of infective 13 14 313 endocarditis: a population-based study through a record-linkage system. BMC Infect Dis 15 16 17 314 2011;11:48. 18 For peer review only 19 20 315 2. Monteiro TS, Correia MG, Golebiovski WF, et al. Asymptomatic and symptomatic embolic 21 22 316 events in infective endocarditis: associated factors and clinical impact. Braz J Infect Dis 23 24 25 317 2017;21:240-247. 26 27 318 3. Habib G, Lancellotti P, Antunes MJ, et al. 2015 ESC Guidelines for the management of 28 29 30 319 infective endocarditis: The Task Force for the Management of Infective Endocarditis of the 31 32 33 320 European Society of Cardiology (ESC). Endorsed by: European Association for Cardio-Thoracic 34 35 321 Surgery (EACTS), the European Association of Nuclear Medicine (EANM). Eur Heart J 36

37 http://bmjopen.bmj.com/ 38 322 2015;36:3075-3128. 39 40 323 4. Spaulding AR, Salgado-Pabon W, Kohler PL, et al. Staphylococcal and Streptococcal 41 42 43 324 Superantigen Exotoxins. Clin Microbiol Rev 2013;26:422-447. 44

45 on September 28, 2021 by guest. Protected copyright. 46 325 5. Misra DP, Chowdury AC, Edavalath S, et al. Endocarditis: the great mimic of rheumatic 47 48 326 diseases. Trop Doct 2016;46:180-186. 49 50 51 327 6. Langlois V, Lesourd A, Girszyn N, et al. Antineutrophil Cytoplasmic Antibodies Associated 52 53 328 With Infective Endocarditis. Medicine (Baltimore) 2016;95:e2564. 54 55 56 329 7. Kupferwasser LI, Hafner G, Mohr-Kahaly S, et al. The presence of infection-related 57 58 59 330 antiphospholipid antibodies in infective endocarditis determines a major risk factor for embolic 60 15

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 331 events. J Am Coll Cardiol 1999;33:1365-71. 5 6 7 332 8. Peng H, Chen WF, Wu C, et al. Culture-negative subacute bacterial endocarditis masquerades 8 9 333 as granulomatosis with polyangiitis (Wegener's granulomatosis) involving both the kidney and 10 11 12 334 lung. BMC Nephrol 2012;13:174. 13 14 335 9. Durack DT, Lukes AS, Bright DK, Duke Endocarditis Service. New criteria for diagnosis of 15 16 17 336 infective endocarditis: utilization of specific echocardiographic findings. Am J Med 1994; 96:200– 18 For peer review only 19 20 337 9.) 21 22 338 10. Yancy CW, Jessup M, Bozkurt B, et al., American College of Cardiology Foundation; 23 24 25 339 American Heart Association Task Force on Practice Guidelines. 2013 ACCF/AHA guideline for 26 27 340 the management of heart failure: a report of the American College of Cardiology 28 29 30 341 Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 31 32 33 342 2013;62:e147–e239. 34 35 343 11. El-Chakhtoura N, Yasmin M, Kanj SS, et al. A 27-year experience with infective endocarditis 36

37 http://bmjopen.bmj.com/ 38 344 in Lebanon. J Infect Public Health 2017;10:734-739. 39 40 345 12. Ferraris L, Milazzo L, Ricaboni D, et al. Profile of infective endocarditis observed from 2003 - 41 42 43 346 2010 in a single center in Italy. BMC Infect Dis 2013;13:545. 44

45 on September 28, 2021 by guest. Protected copyright. 46 347 13. Ying CM, Yao DT, Ding HH, et al. Infective endocarditis with antineutrophil cytoplasmic 47 48 348 antibody: report of 13 cases and literature review. PLoS One 2014;9:e89777. 49 50 51 349 14. Şimşek-Yavuz S, Şensoy A, Kaşıkçıoğlu H, et al. Infective endocarditis in Turkey: aetiology, 52 53 350 clinical features, and analysis of risk factors for mortality in 325 cases. International Journal of 54 55 56 351 Infectious Diseases 2015, 30:106-114. 57 58 59 352 15. Ghosh S, Sahoo R, Nath RK, et al. A Study of Clinical, Microbiological, and 60 16

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 353 Echocardiographic Profile of Patients of Infective Endocarditis. Int Sch Res Notices 5 6 7 354 2014;2014:1-9. 8 9 355 16. Mahr A, Batteux F, Tubiana S, et al. Brief report: prevalence of antineutrophil cytoplasmic 10 11 12 356 antibodies in infective endocarditis. Arthritis Rheumatol 2014;66:1672-7. 13 14 357 17. Aslangul E, Goulvestre C, Mallat Z, et al. Human bartonella infective endocarditis is 15 16 17 358 associated with high frequency of antiproteinase 3 antibodies. J Rheumatol 2014;41:408-10. 18 For peer review only 19 20 359 18. Boils CL, Nasr SH, Walker PD, et al. Update on endocarditis-associated glomerulonephritis. 21 22 360 Kidney Int 2015;87:1241-9. 23 24 25 361 19. Million M, Walter G, Bardin N, et al. Immunoglobulin G anticardiolipin antibodies and 26 27 362 progression to Q fever endocarditis. Clin Infect Dis 2013;57:57-64. 28 29 30 363 20. Million M, Bardin N, Bessis S, et al. Thrombosis and antiphospholipid antibody syndrome 31 32 33 364 during acute Q fever: A cross-sectional study. Medicine (Baltimore) 2017;96:e7578. 34 35 365 21. Selton-Suty C, Maigrat CH, Devignes J, et al. Possible relationship between antiphospholipid 36

37 http://bmjopen.bmj.com/ 38 366 antibodies and embolic events in infective endocarditis. Heart 2018;104:509-516. 39 40 367 41 42 43 368 Contribution 44

45 on September 28, 2021 by guest. Protected copyright. 46 369 For authors, Jialin Teng, Honglei Liu, Xiaobing Cheng, Huihui Chi and Fan Wang 47 48 370 collected the clinical data. Yutong Su and Yue Sun analyzed the data. Zhuochao Zhou 49 50 51 371 and Junna Ye wrote the manuscript. Wei Jin and Chengde Yang designed the study 52 53 372 and critically read the manuscript. We would like to thank the patients for their 54 55 56 373 participation in this study. 57 58 59 374 60 17

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 27

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 375 Data availability statement 5 6 7 8 376 All data relevant to the study are included in the article or uploaded as supplementary 9 10 377 information. 11 12 13 14 378 Declaration of conflicting interests 15 16 17 379 The authors report no relationships that could be construed as a conflict of interest. 18 For peer review only 19 380 20 21 22 381 Funding 23 24 25 382 This work is supported by National Natural Science Foundation of China 26 27 383 (81871272, 81801592, 81370397, 81670266), Science and Technology Commission 28 29 30 384 of Shanghai Municipality (17140902500), Shanghai Sailing Program (18YF1414100), 31 32 385 Shanghai Jiao Tong University Interdisciplinary Research Project (YG2016QN60), 33 34 35 386 Excellent Youth B Project (GCQN-2017-B05) and Innovative research team of 36

37 http://bmjopen.bmj.com/ 38 387 high-level local universities in Shanghai. 39 40 388 41 42 43 389 Acknowledgement 44

45 390 We would like to thank the patients and patient advisers for their participation in this on September 28, 2021 by guest. Protected copyright. 46 47 48 391 study. We also thank Dr Shuzhen Xiao for her critical reading of this manuscript. 49 50 51 392 52 53 393 54 55 56 394 57 58 395 Table 1 59 396 General characteristics of infective endocarditis (IE) patients. 60 18

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 27 BMJ Open

1 2

3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from Item N 4 5 N 432 6 Female/male 154(36)/278(64) 7 Age (years, mean±SD) 46±16 8 9 Duration (months, median, interquartile 1.56(0.75,3) 10 range) 11 Predisposing factors 12 13 Dialysis 6(1) 14 Dental procedure 4(1) 15 Intravenous drug use 2(0.5) 16 Congenital heart disease 66(15) 17 18 RheumaticFor valvulopathy peer review 58(13)only 19 Complications 20 Congestive heart failure 141(33) 21 22 Hypertension 82(19) 23 Diabetes 27(6) 24 Cancer 11(3) 25 26 Embolic events 27 Brain 67(14) 28 Spleen 26(6) 29 Kidney 8(2) 30 31 Extremity 8(2) 32 Coronary 4(1) 33 Lung 4(1) 34 35 Infected valve 36 Prosthetic valve 42(10)

37 Native valve http://bmjopen.bmj.com/ 38 39 Mitral valve 234(54) 40 Aortic valve 185(43) 41 Tricuspid valve 17(4) 42 Pulmonary valve 10(2) 43 44 Outcome

45 Cardiac surgery 346(80) on September 28, 2021 by guest. Protected copyright. 46 Relapse 13(3) 47 48 In-hospital mortality 55(13) 49 397 Note: standard deviation (SD). 50 398 51 52 399 Table 2 53 400 Antibodies in 38 patients tested for both ANCA and aPL. 54 N 55 56 aPL-positive 15 57 ACL-positive 1(IgM ACL) 58 aβ2GP1-positive 1 59 60 LAC-positive 10 19

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 27

1 2

3 ACL+LAC 3(2 IgM ACL +1 IgA ACL) BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 ANCA-positive 9 6 Both aPL- and ANCA-positive 3(LAC, LAC+IgM ACL, IgM ACL) 7 401 Note: anticardiolipin antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I 8 9 402 antibodies (anti-β2GPI), antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody 10 403 (aPL). 11 404 12 13 405 Table 3 14 406 Clinical features of patients with positive or negative ANCA and aPL. 15 ANCA aPL 16 17 ANCA(+), ANCA(-), p value aPL(+), aPL(-), p value 18 ForN(%) peerN(%) review onlyN(%) N(%) 19 Item 21(24) 68(76) / 21(40) 32(60) / 20 21 Positive Blood culture 13(62) 37(58) 0.741 10(50) 17(53) 0.826 22 Fever 19(90) 57(84) 0.688 18(86) 23(72) 0.400 23 Arthritis 11(52) 13(19) 0.003 5(24) 8(25) 0.922 24 25 Fatigue 3(14) 16(24) 0.549 7(33) 9(28) 0.686 26 Pleural effusion 3(14) 15(22) 0.642 6(29) 9(28) 0.972 27 Pericardial effusion 4(19) 12(18) 0.884 6(29) 8(25) 0.773 28 Thrombosis 3(14) 17(25) 0.466 8(38) 10(31) 0.607 29 30 Myalgia 0 6(9) 0.362 2(10) 2(6) 1.000 31 Splenomegaly 6(29) 20(29) 0.941 5(24) 6(19) 0.922 32 Heart murmur 10(48) 43(63) 0.202 18(86) 28(88) 1.000 33 34 Weight loss 3(14) 9(13) 1.000 3(14) 7(22) 0.740 35 Cardiac surgery 5(24) 42(62) 0.002 16(76) 19(59) 0.206 36 In-hospital mortality 4(19) 3(4) 0.087 2(10) 4(13) 1.000

37 http://bmjopen.bmj.com/ 38 407 Note: antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 39 408 40 409 Table 4 41 42 410 Laboratory features of patients positive or negative for ANCA and aPL. 43 ANCA aPL 44 ANCA(+) ANCA(-) p value aPL(+) aPL(-) p value

45 on September 28, 2021 by guest. Protected copyright. 46 Item 21(24) 68(76) / 21(40) 32(60) / 47 WBC (*109/L) 8.39±2.53 8.56±4.19 0.896 7.66±2.54 8.58±4.52 0.403 48 Hb (g/L) 87.57±26.83 105.51±18.76 0.001 99.57±15.28 102.34±26.22 0.664 49 9 50 PLT (*10 /L) 194.70±110.91 201.25±127.06 0.879 184.95±83.86 193.47±151.31 0.819 51 LDH (IU/L) 206.125±53.45 252.00±96.73 0.200 278.87±296.11 247.72±103.72 0.632 52 ESR (mm/h) 64.76±36.87 51.85±33.47 0.135 69.14±34.82 39.09±33.48 0.003 53 54 IgG (mg/dl) 2254.29±1001.40 1660.05±528.67 0.172 1959.69±747.29 1751.67±707.30 0.379 55 IgA (mg/dl) 232.29±89.77 301.50±131.20 0.188 276.63±94.03 283.38±130.21 0.859 56 IgM (mg/dl) 312.71±183.51 140.70±64.03 0.048 193.07±119.30 153.79±119.39 0.324 57 IgE (mg/dl) 37.00±22.39 354.85±569.90 0.286 357.31±437.81 351.73±635.24 0.983 58 59 ACL-positive 2(25) 2(5) 0.436 5(24) 0 0.004 60 aβ2GPI-positive 0 1(5) 1.000 2(13) 0 0.057 20

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 23 of 27 BMJ Open

1 2

3 LAC-positive 2(29) 11(38) 0.981 17(81) 0 0.000 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 ANA-positive 2(15) 6(12) 1.000 3(16) 3(11) 0.618 6 Elevated CRP 18(90) 51(86) 0.851 14(74) 21(75) 0.921 7 Elevated RF 10(59) 14(32) 0.053 6(31) 6(25) 0.336 8 9 411 Note: White blood cell counts (WBC), haemoglobin (Hb), platelets (PLT), erythrocyte 10 412 sedimentation rate (ESR), serum lactate dehydrogenase (LDH), immunoglobulin G (IgG), 11 413 immunoglobulin A (IgA), immunoglobulin M (IgM), antinuclear antibody (ANA), anticardiolipin 12 13 414 antibodies (ACL), lupus anticoagulant (LAC), anti-β2 glycoprotein I antibodies (aβ2GPI), 14 415 antineutrophil cytoplasmic antibody (ANCA), anti-phospholipid antibody (aPL). 15 16 416 17 18 417 Figure 1 For peer review only 19 418 Kaplan-Meier survival curves. 20 419 a. Kaplan-Meier survival curves of patients who were tested for aPL; b. Kaplan-Meier survival 21 22 420 curves of patients who were tested for ANCA. 23 24 421 25 26 27 28 29 30 31 32 33 34 35 36

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 24 of 27 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 For peer review only 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

1 2

3 Supplemental Table 1 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Pathogens of infective endocarditis patients. 6 Item N (%) 7 Pathogen* 8 9 Culture negative 222(51) 10 Gram-positive 11 Streptococcus viridans 25(6) 12 13 Streptococcus sanguis 24(6) 14 Streptococcus mitis 23(5) 15 β-hemolytic streptococci 6(1) 16 17 Other Streptococcus 32(7) 18 StaphylococcusFor aureuspeer review 64(1only5) 19 Staphylococcus epidermidis 30(7) 20 Enterococci 26(6) 21 22 Lactococcus lactis 4(1) 23 Gemella mobillorum 4(1) 24 Aerococcus urinae 3(1) 25 26 Micrococcus 2(0.5) 27 Aerococcus viridans 2(0.5) 28 Kocuria sp 4(1) 29 30 Gram-negative 31 Enterobacter 20(5) 32 Stenotrophomonas maltophilia 2(0.5) 33 Other Gram-negative Bacillus 18(4) 34 35 Fungi 36 Candida albicans 9(2) 37 *Several patients had more than one microorganism. http://bmjopen.bmj.com/ 38 39 40 Supplemental Table 2 41 Clinical manifestations of infective endocarditis patients at time of IE diagnosis. 42 43 Item N (%) 44 Heart murmur 343(79)

45 on September 28, 2021 by guest. Protected copyright. Fever 364(84) 46 47 Arthritis 65(15) 48 Fatigue 86(20) 49 Pleural effusion 119 (28) 50 51 Pericardial effusion 103(24) 52 Myalgia 28(6) 53 Hepatomegaly 19(4) 54 55 Splenomegaly 99(23) 56 Peripheral polyneuropathy 4(1) 57 Janeway lesion 18(4) 58 Osler node 11(3) 59 60 Roth spot 5(1)

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 26 of 27

1 2

3 Supplemental Table 3 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Clinical features and laboratory features of patients tested for antibodies or not. 6 Patients tested for antibodies Patients not tested for antibodies p value 7 Item 104 328 / 8 9 Female/male 40/64 114/214 0.492 10 Age (years, mean±SD) 45.65±14.22 45.59±16.97 0.973 11 Positive Blood culture 56(57) 154(47) 0.108 12 13 Fever 89(86) 275(84) 0.758 14 Arthritis 27(26) 41(13) 0.001 15 Fatigue 26(25) 67(20) 0.323 16 17 Pleural effusion 21(20) 98(30) 0.054 18 Pericardial effusion For19 (18)peer review84(26) only 0.126 19 Thrombosis 25(24) 116(35) 0.032 20 Myalgia 7(9) 23(7) 0.572 21 22 Splenomegaly 29(28) 70(21) 0.167 23 Heart murmur 67(84) 288(88) 0.333 24 Weight loss 14(18) 47(14) 0.476 25 26 Cardiac surgery 54(65) 290(88) 0.000 27 In-hospital mortality 9(9) 46(14) 0.178 28 WBC (*109/L) 8.39±2.53 8.39±2.53 0.614 29 30 Hb (g/L) 87.57±26.83 87.57±26.83 0.105 9 31 PLT (*10 /L) 194.70±110.91 194.70±110.91 0.659 32 ESR (mm/h) 64.76±36.87 64.76±36.87 0.116 33 Proteinuria 27(27) 85(26) 0.788 34 35 Hematuria 50(51) 133(41) 0.079 36 Note: White blood cell counts (WBC), hemoglobin (Hb), platelet (PLT), erythrocyte

37 http://bmjopen.bmj.com/ sedimentation rate (ESR). 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50

51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 27 of 27 BMJ Open

1 2

3 Supplemental Figure 1 BMJ Open: first published as 10.1136/bmjopen-2019-031512 on 12 February 2020. Downloaded from 4 5 Distribution of patients diagnosed with infectious endocarditis. 6 7 8 9 10 11 12 13 14 15 16 17 18 For peer review only 19 20 21 22 23 24 25 26 27 28 29 30 31 32 * Patients who were either tested with aPL or ANCA were included and compared in different 33 groups. 34 35 Note: infectious endocarditis (IE), antineutrophil cytoplasmic antibody (ANCA), 36 anti-phospholipid antibody (aPL).

37 http://bmjopen.bmj.com/ 38 39 40 41 42 43 44

45 on September 28, 2021 by guest. Protected copyright. 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 28 of 27

1 2 STROBE Statement—checklist of items that should be included in reports of observational studies

1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 29 of 27 BMJ Open

1 2 Results

2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml