Antibody durability 6 months after two doses of SARS-CoV-2 mRNA vaccines in patients with rheumatic and musculoskeletal disease

Antibody response following SARS-CoV-2 vaccination has been shown for up to 6 months in immunocompetent populations.¹ Although we previously described overall stability of SARS-CoV-2 antibody titres at 3 months in patients with rheumatic and musculoskeletal disease,2, 3 few longitudinal data exist beyond 3 months in immunosuppressed patients with rheumatic and musculoskeletal disease. Additionally, with the US Food and Drug Administration authorisation of third vaccine doses for immunosuppressed individuals,4, 5 the optimal timing for such doses in this patient group remains unknown. In this Comment, we describe antibody responses in a cohort of patients with rheumatic and musculoskeletal diseases on immunosuppression 6 months after completing the two-dose SARS-CoV-2 mRNA vaccine series.

We studied adults with rheumatic and musculoskeletal disease on immunosuppression who had received two doses of mRNA vaccine (BNT162b2 or mRNA-1273) between Jan 5 and April 28, 2021. Participants gave informed consent to participate before taking part in this study. This study was approved by the Johns Hopkins Institutional Review Board (IRB00248540). As previously described, demographics and clinical characteristics were collected via electronic questionnaire.⁵ Serial antibody testing was undertaken at months 1, 3, and 6 after the second vaccine dose with an anti-SARS-CoV-2 S enzyme immunoassay (Roche Elecsys, Rotkreuz, Switzerland), which tests for antibodies against the receptor binding domain (RBD) of the spike protein and represents a consistent correlate of immunity.⁶ Low positive antibody response was defined as anti-RBD pan immunoglobulin (Ig) 0·8–50·0 units per mL and high positive antibody response was defined as anti-RBD pan Ig of at least 50·0 units per mL on the basis of plasma-neutralising capacity in patients with convalescent COVID-19 and a modelling study across SARS-CoV-2 vaccine trials.7, 8 Participants with previous SARS-CoV-2 infection and recipients of additional vaccine doses before antibody testing at 6 months were excluded. We adapted survival methods to address right truncation of titres, in which the exact titre value was known for some participants and was only known to have exceeded a particular value (>250·0 units per mL or >2500·0 units per mL [ie, censorship]) for others. This methodology was used to calculate the medians (IQR).

Of 326 participants, 302 (93%) were female and median age was 49 years (IQR 39–59; appendix p 1). Inflammatory arthritis was the most common disorder, diagnosed in 139 (43%) patients. Hydroxychloroquine was the most frequently reported conventional synthetic disease modifying antirheumatic drug (135 patients [41%]), whereas TNF inhibitors were the most common biological therapy (84 patients [26%]). The majority of participants (207 [63%]) were on combination therapy, which was defined as two or more concomitant immunosuppressive medications.

At a median of 29 days (IQR 28–32) after the second dose, 312 (96%) of 326 patients had positive anti-spike titres with a median titre of 1175 units per mL (447·0 to >2500·0; table ). Among participants with a positive titre at 1 month, only two (<1%) had a titre below the threshold of positivity at 6 months. At a median of 91 days (87–93) after the second dose, 311 patients (95%) had positive anti-spike titres with a median titre of 647·5 units per mL (165·2–1373·0). At a median of 181 days (176–187) after the second dose, 313 participants (96%) had positive anti-spike titres with a median titre of 419 units per mL (91·9–861·0).

Table.

Anti-spike antibody response 6 months after two doses of SARS-CoV-2 mRNA vaccine in patients with rheumatic and musculoskeletal disease stratified by antibody response at 1 month

	Negative response	Low positive response	High positive response	Total at 1 month
Negative response at 1 month	11/14 (79%)	2/14 (14%)	1/14 (7%)	14/326 (4%)
Low positive response at 1 month	2/25 (8%)	14/25 (56%)	9/25 (36%)	25/326 (8%)
High positive response at 1 month	0/287 (0%)	37/287 (13%)	250/287 (87%)	287/326 (88%)
Total at 6 months	13/326 (4%)	53/326 (16%)	260/326 (80%)	..

Data are n/N (%). Anti-spike antibody response was defined per manufacturer data for the anti-SARS-CoV-2 S enzyme immunoassay (Roche Elecsys, Rotkreuz, Switzerland): negative was an anti-RBD pan Ig of <0·8 units per mL, low positive was an anti-RBD pan Ig of 0·8–50·0 units per mL, and high positive was an anti-RBD pan Ig of ≥50 units per mL. The initial ceiling of 250·0 units per mL increased to 2500·0 units per mL (positive ≥0·8 units per mL) in April, 2021. RBD=receptor binding domain. Ig=immunoglobulin.

Overall, 313 (96%) of 326 patients had positive antibody titres 6 months after two doses of the mRNA vaccines. Of 287 participants with high positive titre values at 1 month, 250 (87%) had a high positive titre value at 6 months, compared with 37 (13%) with a low positive value. A greater proportion of participants who had received the mRNA-1273 vaccine had high positive antibody titres than did those who had received the BNT162b2 vaccine (89% vs 72%; p<0·001; appendix p 1), with median titre of 612·4 units per mL (158·6–1042·0) in mRNA-1273 recipients and of 194·8 units/mL (40·9–618·0) in BNT162b2 recipients at 6 months. Additionally, a greater proportion of participants on monotherapy had high positive antibody titres than did those on combination therapy (94% vs 71%; p<0·001). Among the 12 participants with an augmented antibody response over 6 months, all individuals reported use of antimetabolites (ten with mycophenolate and two with methotrexate), whereas four participants were noted to have a decay in titre from 3 months to 6 months. All 13 patients (4%) with negative response at 6 months were on lymphodepleting therapy consisting of rituximab (ten patients), abatacept (two), or belimumab (one) in combination with another immunosuppressive therapy; nine (69%) were on concomitant glucocorticoid therapy.

Limitations of this study included an absence of an immunocompetent control group and an absence of data on both memory B cells and cellular response. We included patients without known COVID-19; however, given that we did not complete anti-nucleocapsid testing, asymptomatic infection could not be excluded.

Antibody response against SARS-CoV-2 decreased by 2·8 times from 1 month to 6 months, but remained above the threshold of predicted neutralising capacity⁸ in the majority of patients 6 months after two doses of mRNA vaccine. Participants who had received the mRNA-1273 vaccine or reported use of immunosuppressive monotherapy were more likely to have a high positive antibody response at 6 months than were participants who had received the BNT162b2 vaccine or those who reported use of combination immunosuppression. Consistent with previous findings, patients on lymphodepleting or combination immunosuppression more commonly had an undetectable antibody response than did patients on other immunosuppressive therapies.5, 9 Notably, we observed increased titres in a small proportion of our cohort. All of these participants reported antimetabolite use, which might suggest a delayed antibody response in patients on some lymphodepleting therapies and is consistent with previous findings.3, 10

Longitudinal studies evaluating the effects of differential immunosuppression and vaccine platform, as well as clinical correlates of protection, are required to inform the optimal vaccination schedule to ensure durable protective immunity in this susceptible patient population.

DLS reports consulting from CSL Behring, Novartis, Jazz Pharmaceuticals, Veloxis, Mallinckrodt, and Thermo Fisher Scientific; and speaking honoraria from Sanofi and Novartis. LC-S reports consultant fees from Janssen, Boehringer-Ingelheim, Mallinckrodt, Serono, ArgenX, Allogene, Roivant, and Octapharm; royalties for intellectual property related to anti-HMGCR assay from Inova Diagnostics; grants from Pfizer, Corbus, and Kezar; payment for expert testimony from Bendin Sumrall and Ladner, Feldman, Kleidman Coffey & Sappe, Downs Ward Bender Hauptmann & Herzog, and Sulloway and Hollis. All other authors declare no competing interests. SF and TP-YC contributed equally. DLS and JJP were co-senior authors. This work was made possible by the generous support of the Ben Dov family. This work was supported by grant number F32DK124941 to BJB, T32DK007713 from the National Institute of Diabetes and Digestive and Kidney Diseases to JLA, K24AI144954 to DLS, U01AI138897 and K23AI157893 from National Institute of Allergy and Infectious Diseases to WAW, and K23AR073927 from National Institute of Arthritis and Musculoskeletal and Skin Diseases to JJP. The analyses described in this Comment are the responsibility of the authors alone and do not necessarily reflect the views or policies of the US Department of Health and Human Services, nor does mention of trade names, commercial products, or organisations imply endorsement by the US Government.