This retrospective study was carried out using laboratory and electronic records of our SLE patients attending outpatient and inpatient services of the Department of Clinical Immunology and Rheumatology between March 2013 and January 2015. Hospital data of patients fulfilling ACR 1990 or SLICC 2012 classification criteria for SLE who underwent anti-C1q antibody test during this period, were retrieved from laboratory register. Relevant clinical, laboratory and serological parameters corresponding to the time of anti-C1q assay were noted from hospital electronic medical record. Clinical parameters noted included presence of organ system involvement (e.g., arthritis, skin manifestations, serositis, and central nervous system involvement), thromboembolic events, major infections as well as demographic features like disease duration prior to presentation. Laboratory findings from hospital electronic medical records were also noted including ESR, haemoglobin, blood counts, complement C3 and C4, Urine Protein/Urine Creatinine ratio (UP/UC), presence of autoantibodies (like anti-dsDNA, anti nucleosome antibody and antiphospholipid antibodies) and biopsy results. Presence of lupus anticoagulant or anti cardiolipin antibody in our SLE patients was considered indicative of positive antiphospholipid antibody status. When other laboratory test results were not available at the precise date of anti-C1q antibody assays, test results within 15 days of the anti-C1q antibody measurement were accepted as concurrent test results [18]. Disease activity score was calculated by SLEDAI using the relevant data from the electronic record of the hospital at the time of anti-C1q assay. Accordingly, disease activity in all SLE patients was classified as mild, moderate, or severe, based on their SLEDAI scores (mild <8, moderate 8-18, and severe >18). Renal SLEDAI score was calculated to assess kidney disease activity as described earlier [19].

Blood samples for anti-C1q antibody assay were collected from patients during their hospital visits irrespective of disease duration or dose of immunosuppressants. Anti-C1q Antibody was assayed by commercially available ELISA kit (Demeditec Diagnostics GmbH, Germany). Results were expressed as unit/ml (U/ml) and serum level more than or equal to 10 U/ml (cut off value), was considered positive, as recommended by the manufacturer.

This study was approved by the Institutional Review Board (ethics and research committee) of Christian Medical College, Vellore, Tamil Nadu, India.

Continuous data are presented as mean (SD) or median (range) and categorical data are presented as frequency. Mann-Whitney test was performed to find out any significant difference of anti-C1q Ab titer between: a) Patients with and without LN; and b) between patients with active LN and Inactive LN. To identify any association of clinical and laboratory variables with anti-C1q antibody, univariate logistic regression analysis was done. Organ involvement, major infection and thromboembolic events were categorised as present or absent at the time of anti-C1q antibody assay. Laboratory variables included C3 and C4 levels, ESR and urine protein/urine creatinine ratio, and these were categorised as low or normal whereas anti- dsDNA antibody, anti nucleosome antibody and anti phospholipid antibodies were categorised as positive or negative. The relevant variables with p<0.05 by univariate logistic regression analysis were subjected to multivariable logistic regression analyses. A two tailed p-value of <0.05 was considered as significant. The statistical analysis of data was done using STATA 13.1 (StataCorp LP, Texas, USA) statistical package.

Baseline characteristics of 126 patients included in the study are shown [Table/Fig-1]. LN was present in 54.76% of the patients. Anti-C1q antibody was present in 42.85% of the patients. Anti-C1q levels were significantly higher in patients with LN, as compared to lupus patients without LN (p<0.05) [Table/Fig-2a]. Similarly, anti-C1q levels were significantly higher in patients with active LN, as compared to inactive LN (p<0.05) [Table/Fig-2b].

[Table/Fig-2]:

Scatter plot show anti-C1q Ab levels; (a). Between patients with LN (n-69) and without LN (n-57); (b). Between patients with active LN (n-32) and Inactive LN (n-37)

*Mann-Whitney U test was used.

Results of univariate analysis are shown [Table/Fig-3]. Variables found to be significant in univariate analysis were subjected to multivariate logistic regression with the exception of anti nucleosome antibodies and SLEDAI [Table/Fig-3]. Anti nucleosome antibody tests were available only for 91 patients, hence it was excluded from multivariable analysis. SLEDAI was also excluded from multivariate analysis, as it encompasses some of the other clinical and lab variables used in analysis. In multivariate analysis, low C4 (OR 3.11), mucocutaneous features (OR 4.72) and higher renal SLEDAI (OR 1.35) were found to be significantly associated with positive anti-C1q antibody. On removing renal SLEDAI from analysis, UP/UC {OR 1.77(0.63-4.95), (p<0.05)} also attained significance.

We also did a separate univariate analysis with laboratory markers and clinical features in LN patients (n-69) [Table/Fig-4]. Anti-nucleosome antibody, anti-dsDNA antibody, low C3 and low C4, high UP/UC and renal SLEDAI were found to be significantly associated with anti-C1q antibody. No significant association was found when these parameters underwent multivariate logistic regression.

Anti-C1q antibody is associated with SLE, in particular with LN. The result of present study on associations of anti-C1q antibody with LN in Asian Indian population is consistent with previous studies reported in western as well as Asian population. Multivariate analysis didn’t reveal any significant association of anti-C1q antibody in LN. In the whole SLE cohort, however, low C4, mucocutaneous and renal SLEDAI were independently associated with elevated serum anti-C1q antibody levels.

The frequency of anti-C1q antibodies in all SLE patients and LN subgroup were 42.3% and 50.7%, respectively in the present study as compared to 58.3% and 60%, respectively in a study from Western India [20]. The difference in frequencies could be accounted by larger sample size in our study. The prevalence of anti-C1q in another study consisting of multiethnic SLE patients was 28%, which is lower than the present study. In our study, anti-C1q antibody levels were significantly higher in patients with LN as compared to lupus patients without LN (p<0.05), which has not been reported previously [20–22]. Association of anti-C1q Ab with anti nucleosome Ab, anti-dsDNA Ab and urine protein/creatinine ratio in LN patients was not established in our multivariate analysis, possibly due to relatively lesser number of subjects with LN. Our finding of a pronounced association of anti-C1q antibody with anti nucleosome antibody and also a fairly strong association with anti dsDNA antibody may be explained by plausible biological basis. Nucleosome enhances binding of C1q in glomerular endothelial cells undergoing apoptosis and this interaction leads to potential binding of anti-C1q autoantibodies [23]. Anti-C1q antibody is considered to be predictor of active LN and in our study too, we found higher levels in active disease, as compared to inactive LN, similar to observation by others [24,25]. However, some reports also suggest that SLE patients negative for anti-C1q antibody are unlikely to get LN [10,26].

Association of renal SLEDAI with anti-C1q antibody observed in our study confirms utility of this antibody as a biomarker for renal involvement in SLE [15]. In the multivariate analysis, once renal SLEDAI was eliminated, UP/UC ratio {OR 1.77(0.63-4.95)} was also a significant association of anti-C1q antibody. This may imply that UP/UC levels are an important contributing factor towards renal SLEDAI in lupus nephritis patients with high levels of anti-C1q antibodies. This is again consistent with previous findings showing association of anti-C1q antibody with proteinuria and renal disease activity [8,27]. Further, combination of anti-C1q antibody and anti- dsDNA antibody predicts poorer renal outcome indicating active renal disease [28]. Results from this study also may suggest that anti-C1q antibody in combination with anti nucleosome antibody can be a better reflector of active LN.

In our SLE cohort, anti-C1q antibody is also highly associated with mucocutaneous involvement {OR 4.72 (1.38-16.05)}. Data on association of cutaneous features with anti-C1q antibody is scarce in SLE literature [29]. Bălănescu E et al., reported that SLE patients had increased titers of anti-C1q in serum [30]. Anti-C1q antibodies decrease complement proteins in circulation including C1q [31]. Congenital deficiency of C1q have been observed in early onset photosensitive SLE [32]. Although, low C1q in adult SLE patients is an acquired phenomenon caused by anti-C1q antibodies, could this have produced cutaneous manifestations as in patients with congenital C1q deficiency?

Nevertheless, association of early onset SLE with deficiency of C1q is well known [33]. In our study also, anti-C1q antibody levels were different between age groups with a cut off of 35 years of age; although this did not reach statistical significance, the trend was visible (p=0.055). Orbai AM et al., found that anti-C1q antibody level was higher in SLE patients with age below this age cut off [15]. These data support the notion that anti-C1q antibodies were more common in younger onset SLE patients.

Limitation of this study includes retrospective design and smaller sample size.

In conclusion, low complement protein C4, mucocutaneous features and higher renal SLEDAI were independently associated with high anti-C1q antibody titre. Anti-C1q antibody, therefore, may act as a surrogate marker of renal and cutaneous involvement in SLE.