Introduction

Systemic lupus erythematosus (SLE) is a systemic autoimmune inflammatory disease with variable clinical manifestations. SLE mainly affects females. In the past decades, the survival rate of SLE patients has been greatly improved due to earlier diagnoses and administration of high-dose glucocorticoids and other immunosuppressive agents. With these intensive treatments for active SLE manifestations, the most frequent cause of death in patients with SLE is now infection [1, 2]. These infections can be viral, bacterial, fungal, or parasitic involving multiple organs [3]. The most common infection sites are the lung, blood stream, urinary tract, and gastrointestinal system. Joint infections in patients with SLE, which represent a destructive form of acute arthritis, have rarely been described.

The annual incidence of joint infection in the general population varies from 2 to 10 per 100,000 patients [4]. The incidence increases to 28–38 cases per 100,000 in patients with pre-existing inflammatory arthritis [5]. Despite the wide use of antibiotics, the mortality rate for in-hospital joint infections ranges from 7 to 15% [6,7,8,9,10] (most of these deaths are directly attributable to sepsis). Currently, there is still no single confirmatory test available for diagnosing joint infections. Joint infections are difficult to diagnose at an early stage, especially in SLE patients, more than half of whom present with arthritis [11, 12]. If infections are undetected, they can lead to rapid joint destruction. Therefore, identifying an infection and pathogen early is essential for successful treatment. In this study, we investigated the clinical characteristics and risk factors of joint infections in patients with SLE in China.

Patients and methods

Patients

We retrospectively reviewed the medical charts of SLE inpatients admitted to Peking Union Medical College Hospital (PUMCH) from January 1990 to December 2016 in the medical record system. Demographic data, clinical features, laboratory findings, treatments, and outcomes were recorded. Of the 11,734 SLE patients, 20 had joint infections. Subjects without joint infections (designated as control patients) were selected from the pool of SLE patients using a 1:4 systematic sampling method (selected during the same time period as the SLE patients). All patients fulfilled the 2009 American College of Rheumatology revised classification criteria for SLE. Active SLE was defined as an SLE Disease Activity Index (SLEDAI) score ≥ 5. A definitive diagnosis of a joint infection was based on clinical manifestations and one of the following criteria: a positive result for any microorganism in a tissue specimen, joint fluid, blood culture, Gram stain, or acid-fast stain; typical imaging suggestive of joint infections; or effectiveness of an empirical treatment. The institutional review board of PUMCH approved this study. The requirement for written informed consent was waived because this study was retrospective and only involved the review of records.

Statistical analyses

The software package SPSS 18.0 (IBM, Armonk, NY, USA) was used to perform statistical analyses. Means ± standard deviations (SDs) were used for descriptive analyses. A chi-square test and Fisher exact test were used to compare categorical data. An independent sample t test was used to compare quantitative data between groups. Statistical significance was set at P < 0.05.

Results

Patient demographics and clinical features

The overall prevalence of joint infections in the SLE patients was 0.2% (20 of 11,734). The male-to-female ratio was 1:9. The mean age at onset of a joint infection was 32.2 ± 10.8(range 17–51) years old. Two patients had undergone joint needle aspiration prior to the infection, and only 1 patient had experienced trauma before the joint infection. The duration from the onset of SLE to joint infection varied from 4 months to 13 years (median duration 23 months). The clinical features at presentation are shown in Table 1. Fourteen patients presented with a single joint infection, and 6 patients had infections in 2 joints. The knee joint was infected in 15 patients, followed by the hip (6 cases), the ankle (1 case), and the sacroiliac joint (1 case). Joint pain was the most frequently observed symptom (20 cases, 100%). Joint swelling was noted in 14 patients, except for 5 patients with infection of the hip infections or the sacroiliac joint (swelling of this joint is usually undetectable). Twelve patients experienced joint warmth, while 15 patients experienced fever.

Table 1 Characteristics of SLE patients with joint infections
Full size table

All patients were treated with steroids at the time of SLE diagnosis. The median duration of glucocorticoid treatment was 11 months (range 1–156 months). Eighteen patients received a 1 mg/kg/day equivalent dose of prednisone, and 6 patients underwent methylprednisolone pulse therapy. Following the methylprednisolone pulse therapy, 3 patients were experiencing joint infections 3 months after therapy, and the other 3 patients were experiencing joint infections more than 3 years after therapy. When joint infections were diagnosed, seven patients were managed with a daily dose of prednisone ≥1 mg/kg/day. Four patients received ≥0.5 and <1 mg/kg/day prednisone. Nine patients received <0.5 mg/kg/day prednisone (≥15 mg in 3 cases, <15 mg in 6 cases). Seventeen patients received additional immunosuppressants, and 15 patients received a monotherapy: cyclophosphamide (CTX) (9 cases), leflunomide (2 cases), cyclosporine A (1 case), or mycophenolatemofetil (MMF) (3 cases). Two patients were treated with CTX and MMF (1 case) or CTX and Tripterygiumwilfordii (1 case). It is worth mentioning that the treatment of all patients was adjusted after the infection was confirmed.

Pathogens

Seven patients had Salmonella infections (5 cases of Salmonella typhi, 2 cases of Salmonella D). Five patients had Staphylococcus aureus infections (2 cases of methicillin-resistant S. aureus, 3 cases of methicillin-sensitive S. aureus). One patient was infected with β-hemolytic Streptococcus. Seven patients had suspected Mycobacterium infections. The pathogen was confirmed in only 6 of these cases, and 1 patient in which Mycobacterium was not found (She had acute unilateral knee pain and swelling. MRI revealed cartilage destruction and purulent effusion. Purulent synovial fluid was aspirated from the right knee. The patient was treated with anti-tuberculosis drugs, and the symptoms improved.)

Six patients had simultaneous periarticular abscesses caused by the same pathogen, which was the most common coexisting infection. Bacteremia was found in 3 patients. Other organ infections occurred in 3 patients (pulmonary tuberculosis, intestinal infection, and vertebral infection). Osteomyelitis was noted in 3 patients.

Laboratory characteristics

Leukopenia was identified in three patients (15.0%), and lymphocytopenia was identified in 13 patients (65.0%, mean 0.9 ± 0.5 × 109/L). A lymphocyte subset analysis was performed in seven patients. Depressed B cell function was observed in 7 patients (77.8%), and the CD4+ T cell count was decreased in all patients (100%) and was <200 cells/mm3 in 6 patients (66.7%). Hypoalbuminemia was observed in eight patients (40.0%, mean 31.1 ± 5.4 g/L). Elevated erythrocyte sedimentation rates (ESR) were observed in all patients with joint infections. Elevated C-reactive protein (CRP) levels were observed in 16 patients. The CRP levels were more than 50 mg/L in 10 patients (50.0%).

Imaging examinations of the involved joints were performed for all patients. The examinations revealed evidence of involved joint swelling, soft tissue thickening, and joint space narrowing (Fig. 1). Synovial fluid aspiration was performed for 16 patients. The fluid had a hazy, opaque, or bloody appearance and was characterized by an increased number of leukocytes. Seven patients underwent arthroscopy, and we observed narrowing of the joint space, joint cavities filled with necrotic tissue, synovial hyperplasia, and cartilage destruction. Caseating granulomas were visualized in patients infected with Mycobacterium. Infectious organisms were isolated from the synovial fluid cultures of 17 patients (obtained by synovial fluid aspiration or arthroscopy). Infectious organisms were isolated from two patients using synovial tissue pathology.

Fig. 1
figure 1

MRI from an SLE patient with a right hip joint infection. The right side of the articular surface was not flat. The femoral head had long T1 and T2 signal intensities with fluid collection (long T1 and T2 signals)

Full size image

Comparison of clinical characteristics between SLE patients with and without joint infections

Table 2 shows a comparison of the SLE patients with and without joint infections. The proportion of patients with pre-existing arthritis was significantly higher among the patients with joints infections. However, these patients had less gastrointestinal involvement and cardiac damage. Additionally, the joint infection group had significantly lower SLEDAI scores than the control group. Patients with joint infections had higher incidence of elevated CRP levels (80.0 vs 22.5%, P < 0.01) than the control group.

Table 2 Comparison of clinical characteristics of SLE patients with and without joint infections
Full size table

Discussion

Joint infection is a serious condition that can lead to joint and cartilage destruction. Indeed, delayed or inadequate treatment of joint infection can lead to irreversible joint destruction, with subsequent disability. This study is the first serial investigation of joint infections in SLE patients in China. Distinguishing infection from underlying arthritis in SLE patients is a diagnostic challenge of great importance and urgency. Invasion of a pathogen into the synovial space generally occurs by 2 routes: hematogenous spread (most common) or direct invasion [13]. The risk factors for joint infections are penetrating trauma, therapeutic procedures, joint prostheses, periarticular abscesses, pre-existing joint disease, and aging [9, 14]. We observed 2 patients who had joint needle aspiration and 1 patient with trauma prior to joint infection, and no patients had a history of joint prosthesis. Due to their immune dysfunction, the risk factors for joint infections among SLE patients are higher than for normal people, and joint infection among SLE patients occurs at a younger age compared to the general public (32.2 vs 57.5 years old) [14]. Moreover, lymphopenia and depressed B and T cell function are observed in the majority of SLE joint infection cases. Humoral and cellular immunological dysfunction induced by SLE and treatment for SLE is assumed to be responsible for the development of infections, especially at uncommon sites [3, 15, 16]. Findings indicate that lymphocytes should be monitored to avoid excessive inhibition of immunity during treatment. Moreover, the occurrence of pre-existing arthritis was also higher among the SLE patients with joint infections. These symptoms may be related to SLE arthritis which can cause joint congestion and swelling. These symptoms further increase the challenge of differentiating an infection from SLE-associated arthritis at an early stage.

The classic presentation of a joint infection is warmth, swelling, pain, and restricted movement. These symptoms are sometimes accompanied by fever and chill. However, a few patients may not present with classical symptoms. In our study, 25.0% of the SLE patients (5 cases) did not have fever when the joint infections occurred. In addition, disease activity was lower among the patients with joint infections. Of note, 18 patients (90.0%) received a 1 mg/kg/day equivalent dose of prednisone at the time of SLE diagnosis, and 17(85.0%) received additional immunosuppressants.

In this study, the joint infection group had significantly lower SLEDAI scores than the control group. Overall, infections often attributed to the use of corticosteroid and immunosuppressant medications [17]. Most infections occurring in SLE patients tend to be stabilized after progressive therapy. The risk of infection increases after the immune system is suppressed. However, most SLE-associated arthritis is relatively sensitive to glucocorticoids. These features may help us distinguish between SLE-associated arthritis and joint infections. All patients had oligo-arthritis in our series. The most frequently involved joints were large joints, such as hips and knees, which contrasts the multiple, symmetrical arthritis of hand joints among SLE patients. Therefore, patients who present with a recent emergence or exacerbation of joint symptoms while taking medium-high steroid doses should be evaluated for a joint infection.

ESR and CRP are acute-phase reactants that become elevated during infections and many autoimmune states. According to our data, SLE patients had a higher incidence of CRP elevation (P < 0.01), and the CRP levels were greater 50 mg/L in 10 patients. Moreover, all of the patients with joint infections had elevated ESR, but there was no significant difference between the joint infection group and the control group. Many conditions can influence ESR levels. However, CRP levels increase more quickly compared to ESR. Elevated CRP levels are more common among patients with infections compared to SLE patients with active disease (and does not reflect SLE disease activity) [3]. Therefore, CRP levels can distinguish infections from lupus flares, especially when CRP levels are >50.0 mg/L [18]. In other standard laboratory settings, serum procalcitonin showed potential for diagnostic prediction of acute bacterial joint infections and was modestly correlated with fever, CRP, serum, and synovial fluid WBC [19].

Salmonella and S. aureus are the predominant pathogens of joint infections [20,21,22], which is consistent with our results. Huang et al. [21] previously reported that SLE patients with Salmonella joint infections were relatively young. Changes in microorganisms involved in joint infections may be related to greater therapeutic aggressiveness (such as antibiotics and biological agents). S. aureus was the cause of 70% of joint infections before 2001 and only 35.8% after 2001 [20]. More unusual organisms, such as Listeria monocytogenes and Mycobacterium tuberculosis [23], caused some infections after 2001, especially in patients treated with TNF inhibitors. Because China has high tuberculosis rates, the prevalence of Mycobacterium in SLE patients in China (9.3%) is noteworthy [24]. Approximately 10% of extra-pulmonary tuberculosis cases are bone and joint tuberculosis [25]. In our study, we observed six confirmed cases of Mycobacterium, which is much higher than other reports on SLE joint infections (30.0 vs 0.0%) [18, 21]. Tuberculosis may be an occult infection and difficult to prove in some cases. However, eradicating the organism usually results in symptom relief.

Infections that cause joint inflammation must be diagnosed quickly. Because aspirated fluid is easy to check, especially the knee and wrist, joint aspiration is an urgent and vital step for assessing any acutely, hot, swollen joint when a joint infection is possible. For hip and shoulder infections, arthroscopy is often preferred because it allows for easier irrigation and provides better visualization of the joint to identify the pathogen.

Synovial fluid and tissue should be sent for culture, Gram stain, pathology, and polymerase chain reaction-restriction fragment length polymorphism analysis. In our study, pathogens were found by joint aspiration in 15 patients (75.0%), and coexisting infections were also higher among SLE patients with joint infections. Simultaneous examination for potential sites, such as blood culture, may help identify the pathogen. Additionally, if joint infection is suspected, empiric antibiotic therapy should be initiated immediately. The goal of treatment is to rapidly eradicate any infection and protect the joint. Thus, joint drainage or open surgical drainage is crucial for medical treatment, especially during osteomyelitis or hip infections. Early exercise may reduce the destruction of cartilage. The prognosis of joint infections depends mainly on the virulence of the pathogen and timing of effective treatments.

In summary, when an SLE patient with pre-existing arthritis suddenly develops asymmetric oligo- or large-joint involvement, especially when associated with elevated CRP levels and low disease activity, joint infections should be considered. A synovial fluid Gram stain, culture, blood culture, and arthroscopy should be promptly performed, and appropriate antibiotics (or empirical antibiotics) and joint drainage should be administered. Early diagnoses and timely treatments are critical for improving prognosis.