The study, published in Frontiers in Lupus by Hiroshi Oiwa and colleagues at Hiroshima City Hiroshima Citizens Hospital, tackles a practical problem that has quietly frustrated Japanese lupus specialists. Circulating immune complexes—clumps of autoantibody bound to self-antigen that drive much of lupus pathology—can be measured with a C1q-based solid-phase enzyme immunoassay. But that assay is no longer commercially available in Japan. Oiwa’s group had previously shown that detectable immune complexes, at levels of 1.5 micrograms per milliliter or higher, were the sole independent predictor of achieving clinical remission off corticosteroids, a milestone the field calls CR off C. Losing the assay meant losing a compass. The researchers’ response was to reverse-engineer it: if you cannot measure immune complexes directly, can you predict their presence from tests almost any clinic already runs?

To answer that question, the team retrospectively screened 124 patients treated at their tertiary center between April 2012 and December 2023, all of whom had active disease requiring initiation or escalation of glucocorticoid therapy. After exclusions, 74 patients formed the study population, which the investigators split into two overlapping cohorts. The evaluation cohort, 70 patients followed for at least one year, served to measure how often patients achieved steroid-free clinical remission. The derivation cohort, 61 patients who had circulating immune complexes actually measured, served to hunt for predictors of immune complex detection. Fifty-seven patients belonged to both groups—a design choice the authors openly flag as a limitation, since it means the cohorts were not independent.

The headline remission figure is sobering but hopeful. Over a median follow-up of 4.4 years, 24 of the 70 evaluation patients—34.3 percent—achieved clinical remission off corticosteroids, defined as zero clinical activity on the SLE Disease Activity Index 2000 while taking no corticosteroids at all. Serological activity, elevated anti-dsDNA antibodies, or hypocomplementemia were permitted under this definition, as were immunosuppressants and biologics. The median time to remission among achievers was 3.5 years. Notably, this cohort was treated aggressively with modern tools: 82.9 percent received hydroxychloroquine at some point, 78.6 percent received immunosuppressive agents such as mycophenolate mofetil, calcineurin inhibitors, or azathioprine, and 52.9 percent received biologics, overwhelmingly belimumab, with a small number treated with anifrolumab. One in three patients with active lupus reaching steroid-free remission in the biologic era is, by the standards of a decade ago, a remarkable shift.

The predictor hunt then zeroed in on the derivation cohort. The researchers first used receiver operating characteristic analysis with the Youden index to find optimal cutoff values for continuous variables tied to immune complex detection. Anti-dsDNA antibody measured by ELISA crossed the line at 57.5 international units per milliliter—nearly five times the assay’s upper reference limit of 12. Complement C3 fell below 68.5 milligrams per deciliter, against a reference interval of 73 to 138. Univariate logistic regression flagged 19 candidate variables, from male sex to rash to direct Coombs positivity. Because the sample was small—29 immune complex-positive and 32 immune complex-negative cases—the investigators capped the multivariate model at three variables, following the one-in-ten rule, and used Firth’s penalized likelihood logistic regression whenever complete separation occurred, as it did for male sex: all eight men in the derivation cohort had detectable immune complexes.

What survived the multivariate gauntlet was biologically coherent. Elevated anti-dsDNA antibody and low C3 remained significant and independent predictors in model 4, with odds ratios of 12.20 and 27.24 respectively. Male sex, tested in a Firth-adjusted model alongside the two serological factors, also held firm, with a β coefficient of 3.54 that was comparable in magnitude to the antibody (2.99) and complement (2.57) coefficients. Initial therapy with mycophenolate mofetil was also associated with immune complex detection, but the authors deliberately excluded it from the final score because treatment choices reflect disease severity and physician preference, inviting confounding. Anti-SSA antibodies, anti-cardiolipin antibodies, and direct Coombs positivity, which had looked promising in univariate analysis, were confounded by the anti-dsDNA and C3 signals and dropped out.

The resulting three-factor score—one point each for male sex, anti-dsDNA antibody at or above 57.5 IU/mL, and C3 below 68.5 mg/dL—produced a gradient the authors describe as almost stepwise. Patients scoring zero had detectable immune complexes 6.3 percent of the time. A score of one raised that to 11.8 percent. At two points, 91.7 percent of patients had detectable complexes, and at three points, every single patient did. The score’s discriminatory performance, measured as the area under the receiver operating characteristic curve, was 0.93 with a 95 percent confidence interval of 0.86 to 1.00, and bootstrap validation with 1,000 resamples yielded essentially the same result. In an exploratory Kaplan-Meier analysis, higher scores were also associated with achieving steroid-free clinical remission, with a log-rank p value of 0.04 and a crude hazard ratio of 1.71 per additional factor.

The biology behind the score is not novel, and the authors are careful to say so. Anti-dsDNA antibodies are the precursors of immune complexes; low complement reflects their consumption. The 57.5 IU/mL threshold, however, carries a mechanistic hint. C1q activation in the classical complement pathway requires two or more immunoglobulin Fc regions to sit close together within an immune complex. Large complexes—formed only when both double-stranded DNA antigen and anti-dsDNA antibody are abundant—may structurally favor that Fc clustering, promoting C1q binding. In other words, a relatively high antibody burden may be needed before immune complexes become detectable by this assay, which is why the cutoff sits so far above the routine reference range.

The male sex signal is the most counterintuitive element. Lupus is overwhelmingly a disease of women—87.1 percent of the evaluation cohort was female—and previous studies have shown that men with lupus develop nephritis more often, accumulate damage faster, and fare worse overall. Yet Oiwa’s earlier work paradoxically suggested men had more favorable outcomes with respect to steroid-free remission in the biologic era. The new data offer a possible explanation: men may be more likely to develop immune complex-mediated disease such as lupus nephritis, characterized by substantial immune complex deposition, and in this cohort five of nine men (56 percent) achieved steroid-free clinical remission compared with 19 of 61 women (31 percent). Male sex remained an independent predictor of immune complex detection even after adjustment for the two serological factors, which may partly explain why its inclusion strengthened the score’s association with remission.

The authors are candid about the study’s limits. It is retrospective, small, and single-center. Multiple statistical analyses risk overfitting. The assumption that immune complex detection predicts steroid-free remission has not been externally validated, and the substantial overlap between the derivation and evaluation cohorts means the remission association should be read as exploratory, not confirmatory. Still, the practical payoff is real: in any clinic where the C1q-based assay is unavailable, three routinely available variables—a sex, an anti-dsDNA antibody level, and a C3 level—may serve as a surrogate compass, identifying patients likely to carry detectable immune complexes and, by extension, those with a higher likelihood of one day living free of the steroids that have long been both lupus’s rescue and its slow poison. Confirmation in larger, independent cohorts is the necessary next step, but for now the score offers a hypothesis-generating bridge between what clinics can measure and what they cannot.

Subject of Research: Immunology

Article Title: Baseline factors defining immune complex detection in systemic lupus erythematosus: development of a three-factor score in a retrospective cohort study

Article References: Oiwa, H., Yahata, A., & Nakamoto, N. (2026). Baseline factors defining immune complex detection in systemic lupus erythematosus: development of a three-factor score in a retrospective cohort study. Frontiers in Lupus, 4, Article 1825240. https://doi.org/10.3389/flupu.2026.1825240

Image Credits: AI Generated

DOI: 10.3389/flupu.2026.1825240

Keywords: systemic lupus erythematosus, immune complexes, anti-dsDNA antibodies, complement C3, glucocorticoid-free remission, biologics, rheumatology, predictive score, Baseline, factors, defining, immune