LAUSR

Severe COVID19 is associated with a poor prognosis among patients with systemic lupus erythematosus (SLE). Accordingly, patients with SLE receiving immunosuppressive drugs have been prioritised for vaccination in France. However, patients with autoimmune diseases—especially those receiving antiCD20— are now known to mount a suboptimal humoral response following COVID19 vaccination. Furthermore, vaccineinduced humoral protection against omicron, the current dominant variant of concern (VOC) worldwide, is not known in SLE. Patients with SLE were prospectively enrolled in the vaccine task force set up between March and May 2021 in our national centre for autoimmune diseases. Humoral vaccine responses against B (ancestral), alpha, delta and omicron variants were assessed using a specific multiplex ELISA assay (CoViDiag kit, Innobiochips, Loos, France). Humoral response was defined by a specific SARSCoV2 antispike IgG (antiS) level in serum >260 binding antibody units (BAU)/mL, according to French Health Authorities. Data were compared between groups using Fisher’s exact test for dichotomous variables and MannWhitney test for continuous variables. Fiftyfive patients with SLE were enrolled in the vaccine task force. Among them, 10 had prior COVID19 and 10 were under immunosuppressive drugs and received threedose primary series of BNT162b2 vaccine following national recommendation. Eventually, 35 COVID19naive patients with SLE (43.4 (36.0, 48.6) years; 88.6% female; table 1) and 9 healthy volunteers (HV) (59.0 (56.0, 62.0) years; 88.8% female) who received two doses of BNT162b2 vaccine 4 weeks apart were screened for humoral vaccine responses at baseline, at second dose, and 2 and 5 months after second dose. The twodose and threedose primary series of BNT162b2 vaccine were performed following an interval of 4 weeks between doses as recommended. Patients with SLE and HV had no detectable antiS and antinucleocapsid at baseline. Two months after the seconddose vaccine, 54.3% (n=19/35), 54.3%, 42.9% and 28.6% of patients with SLE had specific antiS titres>260 BAU/ml against B, alpha, delta and omicron variants, respectively. At the same time point, the percentage of subjects able to mount humoral response against VOC were lower in SLE as compared with HV (100% of HV for B, p=0.016; 100% for alpha; p=0.016; 88.9% for delta; p=0.023; 55.6% for omicron, p=0.235). In patients with SLE, when the humoral responses against VOC was obtained 2 months after the second dose, it was maintained at 5 months in only 10.5% (n=2/19), 10.5% (2/19), 13.3% (2/15) and 10% (1/10) for B, alpha, delta and omicron variants, respectively. Overall, 5 months after the seconddose vaccine, the percent of patients with SLE with humoral response against B (n=2/35, 5.7% vs n=7/9, 88.9% p<0.001), alpha (5.7% vs 77.8%, p<0.001) and delta (5.7% vs 55.6%, p=0.002) variants were dramatically low as compared with HV. Almost all vaccinated subjects (n=34/35, 99% patients with SLE and n=7/9, 77.8% HV, p=0.101) failed to mount longlasting humoral response against Omicron after two doses of BNT162b2 vaccine (figure 1). Of note, 50% (n=4/8), 50% (n=4/8), 37.5% (n=3/8) and 12.5% (n=1/8) of patients with SLE who were under immunosuppressive drugs (azathioprine, mycophenolate mofetil or antiCD20) and received threedose primary series of BNT162b2 had a preserved humoral response for B, alpha, delta and omicron variants, respectively, 5 months after the third dose. Moreover, all but one (83.4%, n=5/6) patients with SLE who had COVID19 6.2 (3.9–11.6) months before the first BNT162b2 dose had a sustained humoral response 5 months after the second dose against all VOC including omicron (online supplemental figures S2 and S3; online supplemental table S2). In this 6month prospective monocentric study, we show that more than 90% of COVID19naïve patients with SLE failed to reach vaccineinduced humoral response after two doses of BNT162b2. To our knowledge, this is the first evaluation of longlasting humoral responses against VOC including omicron variant in a cohort of patients with SLE. Neutralisation activity was not determined but the threshold of 260 BAU/mL for humoral response was consistent with previous published results and a strong correlation between antiS IgG BAU/mL, live viral neutralisation and pseudoneutralisation activity has been reported by our group. Since our patients with SLE received no immunosuppressive drugs, the poor immune response observed Letter