Risk of Bacterial Infection in Acquired Complement Deficiencies

Background: Acquired complement deficiency can occur in the setting of autoimmune syndromes, such as systemic lupus erythematosus (SLE), with very low or, occasionally, undetectable C3 levels. Based on data from patients with inherited complement defects, a perceived risk for serious bacterial infection exists amongst patients with transiently low complement, but the degree of risk related to C3 level is unknown. Methods: We performed a retrospective study of all pediatric patients with an undetectable total complement activity or absent individual complement components measured at our institution from 2002 to 2018. We assessed annual rate of serious bacterial infection (SBI) defined as requiring hospitalization and/or parenteral antibiotics. Among included SLE patients, we assessed the 30-day probability of SBI for given C3 measurements using a logistic regression model to determine risk. Results: Acquired complement deficiency secondary to SLE-related disease [n=44] was the most common underlying diagnosis associated with depressed complement levels. While controlling for immunosuppression level and lupus nephritis diagnosis, our logistic regression analysis of pediatric patients with SLE showed low C3 level was temporally associated with having an SBI event. Even in patients with equivalent immunosuppression, patients with an SBI were found to have lower C3 levels preceding the infection relative to patients without SBI. Conclusion: Pediatric patients with the diagnosis of SLE can develop very low C3 levels that are independently associated with risk of serious bacterial infection. Patients prone to complement consumption may particularly be at risk.

most common underlying diagnosis associated with depressed complement levels. While controlling for immunosuppression level and lupus nephritis diagnosis, our logistic regression analysis of pediatric patients with SLE showed low C3 level was temporally associated with having an SBI event. Even in patients with equivalent immunosuppression, patients with an SBI were found to have lower C3 levels preceding the infection relative to patients without SBI.
Conclusion: Pediatric patients with the diagnosis of SLE can develop very low C3 levels that are independently associated with risk of serious bacterial infection. Patients prone to complement consumption may particularly be at risk.

Background
Primary complement deficiencies are rare genetic conditions associated with clear susceptibility to encapsulated bacterial infections [1]. Secondary complement defects from an acquired disease process, such as systemic lupus erythematosus (SLE) or connective tissue disease (CTD), are usually partial and transient from immune complex formation and deposition. Temporary hypocomplementemia has been proposed to be a significant risk of infection among patients with SLE [2][3]. Although rarely do patients with SLE demonstrate undetectable complement activity and roughly half of patients with active lupus have normal complement levels [4]. With infections still ranking as one of the highest causes for mortality in patients with SLE, understanding the risk factors is critical [5][6][7].
In this study, we utilized a large catchment area as one of the sole centers with combined Pediatric Rheumatology and Immunology expertise in the greater Pacific Northwest. Here we describe the longitudinal infectious risk among patients with SLE/CTD-related disorders with a history of complement consumption including detailed clinical history to account for kidney disease, immune suppression and types of infection. We compared the risk of infectious complications in the secondary defect group to that of an established cohort of primary complement deficient patients.

Methods
Seattle Children's Hospital institutional review board waived consent due to the study not involving greater than minimal risk (IRB Approval STUDY00001067). A convenience cohort was identified through retrospective review of institutional electronic medical records.

Study Participants:
We performed an unbiased review of all complement levels measured at our hospital from 2002-2018 (n = 5876, from 1643 total patients). Patients clinically diagnosed with SLE or CTD were captured in our pediatric dataset as quarterly complement testing is usual standard practice at our center. A total of 70 subjects met 5 inclusion criteria, based on an undetectable complement measurement during the observation period. Both patients with primary complement deficiency (n = 18) and those with acquired complement deficiency (n = 52) were included in the cohort (Table 1).
Manual chart review identified total number of subjects with each diagnosis screened in our dataset. Primary complement deficiencies in this cohort included C1q deficiency (n = 3), C2 deficiency (n = 11), C6 deficiency (n = 3), and C8 deficiency (n = 1). Acquired  Table 1. We then utilized every C3 measurement for the SLE/CTD patients as a new observation (n = 1,150) and assessed the 30-day probability of having an SBI based upon level. Lupus nephritis and immunosuppression has been associated with SBI in SLE cohorts, thus these 7 were included as covariates. We excluded C3 levels drawn on the same day as the SBI diagnosis. We also categorically examined the risk of an infection in patients with low C3 (C3 < 83 mg/dL per reference range) compared to those with a normal C3 level. In our cohort of SLE/CTD patients with a history of at least one measurement with undetectable complement, we found a correlation between a low C3 level and the probability of being diagnosed with an SBI within 30 days (Fig. 1)

Discussion
Here, we demonstrate that severe C3 insufficiency is associated with a significant risk for serious bacterial infection within 30 days in pediatric SLE/CTD patients, even while adjusting for immunosuppression level and lupus nephritis diagnosis. The increased infection risk among SLE patients is likely multifactorial [9]. From these analyses, it is not possible to ascertain if the observed association between low C3 and infection in our SLE/CTD cohort is truly causative or a marker of other underlying mechanisms for infectious risk. Our findings are consistent with previous 8 studies that found suboptimal C3 levels at time of SLE diagnosis or decreased total complement activity (CH50 < 30 U/mL) to be associated with higher infection risk [2,10]. Although clinicians generally appreciate that patients with low complement levels predispose for bacterial infections, the degree of complement depletion and duration of deficiency remains hard to quantify the risk.
Pediatric SLE/CTD with a tendency to consume complement may benefit from hyperimmunization and rapid receipt of empiric antibiotics similar to primary complement deficient patients [11]. Thus far, physicians are failing to recommend vaccines for SLE patients [12].
A major strength of the study is the large catchment area and lack of other local centers with expertise for these patients, making it less likely that referral bias has affected our results. It is routine care to screen patients with SLE and CTD at our institution with quarterly C3 and C4 complement levels at our center, which lessens the ascertainment bias. The study also benefited from the long observation period to provide a window into longitudinal risk. Being at single referral center permitted normalization of the testing and serum handling methods known to be critical in complement assessment. A sub-analysis excluding complement levels drawn shortly after an SBI did not affect our results, which supports the association of hypocomplementemia and SBI rather than hypocomplementemia being secondary to SBI. Lastly, we made efforts to normalize for confounding variables including immunosuppressive agents, lupus nephritis diagnosis and vaccination status of our cohort.
Our study has several limitations. Our retrospective analysis was limited by missing data, as some of the recorded infections lacked corresponding C3 levels in the preceding 30 days. Due to the lack of a well-established consensus, the grading of immunosuppression level in our cohort was based on expert clinical experience alone. Further prospective studies may be needed to control for these limitations and better examine the risk of infection in patients with complement defects due to C3 consumption as SLE patients are high risk for infections [13].