Clinical characteristics and risk factors of infection in patients with systemic lupus erythematosus: A systematic review and meta-analysis of observational studies

Abstract

ObjectiveInfection is one of the major causes of morbidity and mortality in systemic lupus erythematosus (SLE) patients. We conducted a systematic review and meta-analysis to investigate the clinical characteristics and risk factors of infection in SLE by comparing demographic factors, laboratory data, clinical features, and therapeutic factors between infection and non-infection SLE patients. MethodsPubMed, Embase, and Cochrane databases were searched systematically without restricting the language or year (up to September 2019) by using MeSH terms and keywords pertaining to SLE and infection. Three independent reviewers selected all observational studies based on the established inclusion criteria. Odds ratio (OR) and standardized mean difference (SMD) along with 95% confidence intervals (CI) were used and the analyses were carried out by using a random/fixed-effects model. When necessary, different subgroup and sensitivity analyses were conducted. Study quality was assessed by the modified version (nine-star scoring system) of the Newcastle-Ottawa Scale (NOS) and publication bias was evaluated by funnel plots, and Egger's and Begg's tests. ResultsIn total, we included 39 studies (3709 infection SLE patients and 10526 non-infection SLE patients) based on the inclusion criteria.Compared with the SLE patients without infection, we found that infected SLE patients had a significantly higher incidence rate of the following: 1) lymphopenia (OR = 2.738 95%CI (1.017–7.376), P = 0.046, I2 = 81.4%), 2) thrombocytopenia (OR = 1.61 95%CI (1.4–1.85), P<0.001, I2 = 0%), 3) anemia (OR = 2.294 95%CI (1.402–3.755), P = 0.001, I2 = 83.0%), 4) hypoproteinemia (OR = 2.336 95%CI (1.408–3.876), P = 0.001, I2 = 84.2%), 5) C3 consumed (OR = 1.890 95%CI (1.190–3.002), P = 0.007, I2 = 77.4%), 6) diabetes mellitus (OR = 3.890 95%CI (2.450–6.160), P < 0.001, I2 = 0%), 7) elevated creatinine (OR = 1.954 95%CI (1.646–2.320), P < 0.001, I2 = 0.0%), 8) renal involvement (OR = 2.692 95%CI (2.000–3.623), P < 0.001, I2 = 76.0%), 9) serositis (OR = 3.877 95%CI (0.995–15.110), P = 0.051, I2 = 79.1%), and 10) use of steroid immunosuppressants (OR = 3.116 95%CI (1.959–4.957), P < 0.001, I2 = 77.9%). Furthermore, infected SLE patients had a significantly higher mean dose of prednisone (SMD = 2.088 95%CI (1.196–2.981), P < 0.001, I2 = 97.8%). In addition, SLE patients with infection showed a significantly lower incidence of antimalarial drug use (OR = 0.634 95%CI (0.451–0.892), P = 0.009, I2 = 56.0%).Infected SLE patients had a significantly higher level of 1) 24-h urinary protein (SMD = 0.560 95%CI (0.300–0.810), P < 0.001, I2 = 0%), 2) CRP (SMD = 0.437 95%CI (0.184–0.691), P = 0.001, I2 = 68.6%), and 3) SLE Collaborating Clinics damage index (SDI) (SMD = 0.451 95%CI (0.238–0.664), P < 0.001, I2 = 0.0%), along with a significantly lower level of albumin (SMD = -0.400 95%CI (-0.610–-0.200), P < 0.001, I2 = 0.0%). After adjustment for false discovery rate (FDR), lymphopenia and serositis were no longer associated with the occurrence of infection; however, the remaining factors were still associated with infection in SLE. According to the nine-star scoring system of NOS, 71.79% of the studies were considered as high methodological quality (low risk of bias). No significant publication bias, except for renal involvement, was detected from funnel plots or Egger's and Begg's test, while this publication bias of renal involvement did not impact the pooled estimates. ConclusionWe identified many factors including thrombocytopenia, anemia, hypoproteinemia, hypocomplementemia, hypoalbuminemia, higher level of CRP, higher SDI score, renal involvement and diabetes mellitus that were associated with infection in SLE patients. In addition, glucocorticoids (especially high-dose) and immunosuppressants (e.g. cyclophosphamide) rendered SLE patients more susceptible to infection, while antimalarial drug administration (hydroxychloroquine) was a protective factor against infection in SLE patients. SLE patients with the above clinical characteristics and risk factors might be at high risk from infection, which might contribute to the early identification of infection in SLE patients for better prognosis.