Objective To determine the incidence and baseline factors associated with breakthrough coronavirus disease 2019 (COVID-19) after preexposure prophylaxis (PrEP) with tixagevimab/cilgavimab among patients with systemic autoimmune rheumatic diseases (SARDs).

Methods We performed a retrospective cohort study among patients with SARDs who received tixagevimab/cilgavimab between January 2, 2022, and November 16, 2022. The primary outcome was breakthrough COVID-19 after tixagevimab/cilgavimab. We performed multivariable Cox regression models adjusted for baseline factors to identify risk factors for breakthrough COVID-19.

Results We identified 444 patients with SARDs who received tixagevimab/cilgavimab (mean age 62.0 years, 78.2% female). There were 83 (18.7%) breakthrough COVID-19 cases (incidence rate 31.5/1000 person-months, 95% CI 24.70-38.24), 7 (1.6%) hospitalizations, and 1 (0.2%) death. Older age was inversely associated with breakthrough COVID-19 (adjusted hazard ratio [aHR] 0.86/10 years, 95% CI 0.75-0.99). Higher baseline spike antibody levels were associated with lower risk of breakthrough COVID-19 (aHR 0.42, 95% CI 0.18-0.99 for spike antibody levels > 200 vs < 0.4 units). CD20 inhibitor users had a similar risk of breakthrough COVID-19 (aHR 1.05, 95% CI 0.44-2.49) compared to conventional synthetic disease-modifying antirheumatic drug (DMARD) users.

Conclusion We found that patients with SARDs had frequent breakthrough COVID-19, but the proportion experiencing severe COVID-19 was low. DMARD type, including CD20 inhibitors, did not significantly affect risk of breakthrough COVID-19. Evidence of prior humoral immunity was protective against breakthrough infection, highlighting the continued need for a multimodal approach to prevent severe COVID-19 as novel PrEP therapies are being developed.

Vaccines and therapeutics aimed at preventing and treating coronavirus disease 2019 (COVID-19) have been developed and authorized at a remarkable pace since the beginning of the COVID-19 pandemic.1-3 Unique among these options is tixagevimab/cilgavimab, a combination of monoclonal antibodies against the receptor binding domain of the SARS-CoV-2 spike protein, which received emergency use authorization (EUA) from the US Food and Drug Administration in December 2021 for use as preexposure prophylaxis (PrEP) among patients who are moderately to severely immunocompromised. The original clinical trial demonstrating the efficacy of tixagevimab/cilgavimab enrolled participants with various risk factors for severe COVID-19, but only 3.3% of the participants were eligible due to immunosuppressive medications.4 Nevertheless, tixagevimab/cilgavimab was approved for use in immunosuppressed patients, including those with systemic autoimmune rheumatic diseases (SARDs). Given its limited availability particularly early on, it was often reserved for those receiving B cell depleting therapy because of concerns about their risk of breakthrough SARS-CoV-2 infections, severe outcomes, and poor humoral response to vaccinations.5-8 However, little is known regarding the real-world experience with tixagevimab/cilgavimab in patients with SARDs. Therefore, we aimed to determine the incidence of breakthrough COVID-19 after tixagevimab/cilgavimab among immunocompromised patients with SARDs and identify potential risk factors for breakthrough COVID-19 in this population.

Study design and population. We performed a retrospective cohort study within the Mass General Brigham (MGB) Healthcare system in Massachusetts, USA. MGB encompasses a total of 14 healthcare facilities including tertiary care hospitals, community hospitals, and outpatient centers. We identified all patients who received tixagevimab/cilgavimab as PrEP between January 2, 2022, when tixagevimab/cilgavimab first became available within our system, and November 16, 2022, when the predominant viral strain circulating in Massachusetts became resistant to tixagevimab/cilgavimab.9 The study was approved by the MGB Institutional Review Board (2020P000833).

Identification of patients with SARDs and index date. We identified all patients ≥ 18 years of age within the MGB system who received tixagevimab/cilgavimab between January 2, 2022, and November 16, 2022, using an electronic query of the MGB data warehouse. We then screened for patients with SARDs by filtering for at least 1 International Classification of Diseases code for a SARD diagnosis and use of immunomodulatory medications or glucocorticoids within 1 year prior to the first tixagevimab/cilgavimab receipt, as described in prior studies.10 We did not include patients who were being treated for osteoarthritis, crystalline arthritis, mechanical back pain, or fibromyalgia, because these patients are often cared for by nonrheumatologists. We confirmed the SARD diagnosis by manual medical record review.

The index date was the date of administration of the first dose of tixagevimab/cilgavimab. For patients who initially received the 150 mg/150 mg dose of tixagevimab/cilgavimab (prior to the change in dosage to 300 mg/300 mg in February 2022),11 the first date of tixagevimab/cilgavimab administration was considered the index date, but we quantified the number of patients who received a supplementary dose.

Patient characteristics. We collected demographic information, comorbidities, smoking history, vaccination status, disease-modifying antirheumatic drug (DMARD) use, prior SARS-CoV-2 infection status, and SARS-CoV-2 spike antibody levels from electronic query. Comorbidities included hypertension, diabetes, coronary artery disease, heart failure, asthma, chronic obstructive pulmonary disease, obstructive sleep apnea, chronic kidney disease, and interstitial lung disease. Vaccination status was defined as follows: (1) unvaccinated, (2) partially vaccinated, (3) 2 doses of mRNA vaccines or 1 dose of viral vector vaccine, (4) 3 doses of mRNA vaccines or 1 dose of viral vector vaccine followed by an mRNA vaccine, and (5) any additional vaccine doses, as described previously.12 Specific SARD type and presence of other indications for tixagevimab/cilgavimab (ie, active cancer, organ transplant, other immune-mediated inflammatory disease [eg, multiple sclerosis], primary immunodeficiency) were collected by manual medical record review. For patients on CD20 inhibitor treatment for their SARD, we also collected clinical data on time from the last CD20 inhibitor infusion to the index date and peripheral circulating B cell levels where available.

Outcome: breakthrough COVID-19. The primary outcome was breakthrough COVID-19 after the index date as identified by electronic query and confirmed by manual medical record review. We included both PCR tests and home rapid antigen test results if the latter was reported to a physician. We performed medical record review on all patients to identify COVID-19 outcomes that included patient messages to physicians about positive test results up to the end of study. We also evaluated for severe COVID-19, which we defined as hospitalization or death within 30 days from the positive SARS-CoV-2 test result. We chose SARS-CoV-2 infection as the primary outcome to match the primary efficacy endpoint in the original study for tixagevimab/cilgavimab.4 We report on severe COVID-19 requiring hospitalization but did not expect many severe COVID-19 events given the improvement in outcomes during the Omicron variant era, as we previously reported.12

Covariates. We considered age as a continuous variable per 10 years. For patients receiving multiple immunosuppressive medications, we stratified the medications based on a hierarchical system as in prior studies: CD20 inhibitors > biologic or targeted synthetic DMARDs > conventional synthetic DMARDs (csDMARDs).10 For instance, a patient using methotrexate and a tumor necrosis factor inhibitor would be categorized as a biologic DMARD user. We also considered calendar time as a covariate (categorized into 3 periods in 2022: before April 6, between April 6 and June 30, and after June 30), given the shift in local prescribing patterns and supply of tixagevimab/cilgavimab over time. Due to the initial limited supply at our institution, tixagevimab/cilgavimab was rolled out in phases. CD20 inhibitor users, organ transplant recipients, and those with primary immunodeficiency were prioritized in the first phase before April 6, 2022; subsequently, all patients with low spike antibody levels became eligible; finally, after June 30, 2022, any patient with risk factors for severe illness with COVID-19 became eligible.

We also evaluated surrogates of immunity including baseline spike antibody levels and prior COVID-19 before the index date. We categorized the spike antibodies as < 0.4 units/mL (undetectable), 0.4-200 units/mL, and > 200 units/mL based on the negative, suboptimal, and low immune response categories, respectively, based on the ongoing ACV01 clinical trial for COVID-19 booster vaccines in patients with SARDs.13 For patients with missing spike antibody levels, those who were concomitantly receiving CD20 inhibitors were imputed to have undetectable (< 0.4 units) spike antibody levels.

Statistical analysis. We reported the following descriptive statistics at baseline of tixagevimab/cilgavimab receipt: frequencies and proportions of categorical variables and mean (SD) and median (IQR) of continuous variables, as appropriate. Follow-up time for each patient was measured from the index date to the earliest of breakthrough COVID-19, death, or end of study (November 16, 2022). The incidence rate (95% CI) of breakthrough COVID-19 after tixagevimab/cilgavimab receipt was calculated. We also performed a sensitivity analysis to censor patients upon receipt of the earliest of a second full dose of tixagevimab/cilgavimab or 6 months from the index date, in addition to the other censoring events in the primary analysis.

We then assessed for associations of baseline factors that may predispose to breakthrough COVID-19 after tixagevimab/cilgavimab, using unadjusted and adjusted multivariable Cox regression models to estimate hazard ratios (HRs) and 95% CIs for breakthrough COVID-19. In the multivariable models, we adjusted for age, sex, calendar date, spike antibody level, DMARD use, and previous COVID-19 based on findings from univariate analyses and known risk factors for breakthrough infection.

Two-sided P values of < 0.05 were considered statistically significant. All analyses were completed using SAS statistical software version 9.4 (SAS Institute).

Study cohort characteristics. We identified 501 patients who received tixagevimab/cilgavimab and met our screening for SARD diagnosis; after manual medical record review, we identified 444 patients with confirmed SARD who received tixagevimab/cilgavimab between January 2, 2022, and November 16, 2022 (Supplementary Figure S1, available with the online version of this article). The majority were female (78.2%), and the mean age was 62.0 (SD 14.0) years. Most patients were White (79.3%), followed by Black (8.3%), other or multiple (5.4%), and Asian (5.2%). Comorbidities were common (median Charlson Comorbidity Index of 2, IQR 1-5) and the most frequent comorbidities were hypertension (45.7%), interstitial lung disease (23%), and chronic kidney disease (21%). A total of 88 (19.8%) patients had other indications for receiving tixagevimab/cilgavimab in addition to their SARD diagnosis, including history of solid organ or bone marrow transplant (10.4%), active cancer treatment (6.1%), other immune-mediated inflammatory disease (3.8%), and primary immunodeficiency (0.7%; Table 1).

Baseline demographic and clinical characteristics of patients with SARDs at time of initial tixagevimab/cilgavimab receipt.

The most common SARD diagnoses were rheumatoid arthritis (RA; 43.7%), followed by systemic lupus erythematosus (14.9%), and antineutrophil cytoplasmic antibodies–associated vasculitis (11.7%; Table 2). CD20 inhibitors were used by 48.7% of the cohort. Specific details on the patients receiving CD20 inhibitors including their SARD diagnoses, time from last infusion to the index date, and number of infusions prior to the index date, as well as measurement of circulating B cell levels, where available, are shown in Supplementary Table S1 (available with the online version of this article). Since the recommended dose of tixagevimab/cilgavimab was revised from 150 mg/150 mg to 300 mg/300 mg on February 25, 2022, we report postbaseline variables including additional tixagevimab/cilgavimab doses as well as any vaccinations after the index date (Supplementary Table S2).

Baseline systemic autoimmune rheumatic disease characteristics at time of initial tixagevimab/cilgavimab receipt.

COVID-19 outcomes. We identified 83 (18.7%) patients with breakthrough COVID-19 during 2637.6 person-months of follow-up (Table 3). The incidence rate of breakthrough COVID-19 was 31.5/1000 person-months (95% CI 24.70-38.24). The median time from the index date to symptomatic COVID-19 among cases was 146 (IQR 98-209) days. The distribution of breakthrough infections after the index date is shown in Supplementary Figure S2. Of these, 7 (1.6% of total cohort, incidence rate 2.7, 95% CI 0.69-4.62) had severe COVID-19 (hospitalization or death within 30 days of COVID-19). One death occurred in a patient with RA who developed acute limb ischemia in the setting of COVID-19. Sixty-eight of the 83 patients (81.9%) with breakthrough infection received COVID-19–specific treatments, with nirmatrelvir/ritonavir being the most common treatment (43/83, 51.8%).

Breakthrough COVID-19 outcomes after tixagevimab/cilgavimab among patients with SARDs.

Baseline variables associated with breakthrough COVID-19. Older age was associated with a lower risk of breakthrough infection (adjusted HR [aHR] 0.86/10 years, 95% CI 0.75-0.99). Higher baseline spike antibody levels were also associated with lower risk of breakthrough COVID-19 (aHR 0.42, 95% CI 0.18-0.99 for spike antibody levels > 200 units vs < 0.4 units). The cumulative incidence curve stratified by spike antibody levels is shown in the Figure. Prior COVID-19 infection was not associated with risk for breakthrough COVID-19. In the unadjusted and multivariable models, different DMARD classes had no significant association with the risk of breakthrough infection. Notably, CD20 inhibitor users did not have higher risk of breakthrough COVID-19 (aHR 1.05, 95% CI 0.44-2.49) compared to patients on csDMARDs (Table 4).

Cumulative incidence curve of breakthrough COVID-19 stratified by spike antibody level. aHR 0.42 (95% CI 0.18-0.99) for breakthrough COVID-19 for spike antibody level > 200 units vs < 0.4 units. aHR: adjusted hazard ratio; COVID-19: coronavirus disease 2019.

Associations of baseline demographic- and treatment-related factors with COVID-19 breakthrough after tixagevimab/cilgavimab receipt.

Sensitivity analyses. We performed a sensitivity analysis censoring patients at 6 months after their index date (when we presume the efficacy of monoclonal antibodies would wane), upon receipt of the second full dose of tixagevimab/cilgavimab, death, COVID-19, or end of study (Supplementary Table S3, available with the online version of this article). In this analysis, the observed incidence rate (27.0/1000 person-months, 95% CI 19.59-34.40) of COVID-19 was similar to the primary analysis. In addition, initiation of tixagevimab/cilgavimab later in the pandemic (index date on or after July 1, 2022) was associated with a higher risk of breakthrough infection (aHR 2.48, 95% CI 1.13-5.49).

In this large retrospective cohort study among patients with SARDs who received tixagevimab/cilgavimab during its period of viral neutralizing ability, we found that breakthrough COVID-19 was frequent, but severe infections were uncommon. We did not find disease- or treatment-specific factors associated with breakthrough COVID-19. Importantly, though patients using CD20 inhibitors have been reported to be at an increased risk of breakthrough infection and severe COVID-19,2,5,14 those who received tixagevimab/cilgavimab had a similar risk of breakthrough COVID-19 as those using csDMARDs alone.

We found a lower risk for breakthrough infection among older patients, which may in part be a result of behavioral differences such as more shielding practices, smaller social contacts, or differences in COVID-19 testing practices. In addition, higher baseline spike antibodies were protective against breakthrough COVID-19, highlighting the additional protection afforded by prior immunity from vaccination. Indeed, temporarily interrupting immunosuppression (eg, holding methotrexate or mycophenolate mofetil) or delaying CD20 inhibitor treatment to administer vaccines,15 even if not as protective as in the general population, may be an important complementary strategy in patients using these medications. Alternatively, the spike antibody may be a surrogate for overall immune system function, including cellular immunity that may have prevented symptomatic COVID-19.

Several other studies have described early experiences with tixagevimab/cilgavimab in patients with SARDs and observed a low proportion of patients with breakthrough COVID-19 (2-14%) as well as overall mild infections.16-18 We observed a slightly higher proportion (18.7%) with breakthrough COVID-19 in our cohort, which may be a result of longer follow-up time (median time from index date to COVID-19 of 146 days) as well as inclusion of patients with SARDs with substantial comorbidities such as organ transplant or active cancer. A recent study of patients with SARDs receiving tixagevimab/cilgavimab in France reported similar proportions of breakthrough COVID-19 (20%) as in our study.19 The incidence rate of breakthrough COVID-19 was higher in our current study (31.5/1000 person-months, 95% CI 24.70-38.24) compared to our prior study (2.6/1000 person-months, 95% CI 2.27-2.91),10 likely owing to the difference in the study period. Although there are few studies reporting on rates of SARS-CoV-2 infections among patients with SARDs during the Omicron era, studies from other countries indicate high rates of breakthrough infection (17-23%) among patients with SARDs after January 2022.20,21

In our study, approximately 10% of the patients with symptomatic COVID-19 required hospitalization; although this is lower than in prior studies in patients with SARDs,22-24 it is still substantially higher than the values reported for the general population during a similar period,25 thus highlighting the ongoing need for additional strategies to protect this group. The similar risk of infection observed in CD20 inhibitor users and those using csDMARDs highlights the importance of the PrEP strategy, especially for our most vulnerable patients with SARDs.

At the end of 2022, it was recognized that tixagevimab/cilgavimab did not neutralize the circulating SARS-CoV-2 variants, leading to removal of its EUA status. Thus, ours is the first study to our knowledge that investigated the real-world experience of tixagevimab/cilgavimab among SARDs throughout its entire period of viral neutralizing activity. This revocation has left many of the most vulnerable patients with SARDs at high risk again for severe COVID-19. Our findings highlight the importance of identifying other forms of PrEP, either antivirals26 or monoclonal antibodies, to protect patients with SARDs and inform future PrEP strategies. Additional studies are warranted that investigate these strategies in the highest risk patients. A clinical trial investigating a newer-generation PrEP is ongoing.27

Strengths of our study include the systematic identification of large numbers of patients with SARDs receiving tixagevimab/cilgavimab, as well as confirmation of SARD diagnoses and other clinically relevant risk factors by medical record review. We were also able to ascertain COVID-19 outcomes in the electronic medical record. Finally, we were able to include many patients on CD20 inhibitors, a population considered particularly vulnerable to COVID-19 because of the limited humoral response despite vaccination.7,28 Also, we were able to examine the real-world experience of the entire clinical experience of tixagevimab/cilgavimab at our center during its active EUA period.

Our study has several limitations. First, we did not include a comparator group of patients with SARDs who did not receive tixagevimab/cilgavimab. Thus, we are unable to estimate the comparative effectiveness of use or nonuse of tixagevimab/cilgavimab. It would be challenging to identify an appropriate comparator group that was at high risk for COVID-19 but did not receive tixagevimab/cilgavimab, as the clinical recommendation for much of 2022 encouraged the use of PrEP in immunocompromised patients.9 Also, patients with SARDs who sought PrEP may have important unmeasured differences related to health behaviors and access to care. Despite this, we made important observations regarding the incidence and risk factors of breakthrough COVID-19 after tixagevimab/cilgavimab. Second, since this was a retrospective study, circulating B cells or immunoglobulin levels were not clinically tested in many patients, limiting the conclusions we can draw regarding the association of these factors with the risk of infection. Third, we identified patients with COVID-19 diagnosed by PCR or self-reported home antigen test. As such, we likely did not capture all infections, especially asymptomatic or minimally symptomatic COVID-19. Thus, the rate of infection in our study, especially for nonsevere disease, may be an underestimate. However, given that the patients in the cohort are among the most immunosuppressed patients with SARDs, they are likely to reach out to their providers for guidance in the setting of a symptomatic infection. Finally, the majority of patients also received outpatient treatment with COVID-19 specific treatments such as nirmatrelvir/ritonavir. Thus, the relatively mild infections observed in our cohort could be attributed to the outpatient treatments20 in addition to the effects of PrEP with tixagevimab/cilgavimab.

In summary, we found that breakthrough COVID-19 following tixagevimab/cilgavimab was common but that the vast majority of infections in this high-risk population were nonsevere. Further, we found no associations between DMARD type, including CD20 inhibitors, with the risk of breakthrough COVID-19 among patients with SARDs who received tixagevimab/cilgavimab. As of July 2023, there is no PrEP against SARS-CoV-2 for severely immunocompromised patients such as patients with SARDs using CD20 inhibitors or other potent immunosuppressive agents. Our findings highlight the high clinical need for PrEP to prevent COVID-19 among particularly vulnerable patients.

YK is supported by the National Institutes of Health (NIH) Ruth L. Kirschstein Institutional National Research Service Award (T32 AR007530). ZSW is funded by the NIH/National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS; R01 AR080659, K23 AR073334, and R03 AR078938). JAS is funded by NIH/NIAMS (grant numbers R01 AR080659, R01 AR077607, P30 AR070253, and P30 AR072577), the R. Bruce and Joan M. Mickey Research Scholar Fund, and the Llura Gund Award funded by the Gordon and Llura Gund Foundation.

NJP reports consulting fees from FVC Health, Chronius, and Arrivo Bio unrelated to this work. ZSW reports research support from BMS and Principia/Sanofi and consulting/advisory board fees from Viela Bio, Zenas BioPharma, PPD, Visterra, Novartis, BioCryst, Sanofi, Horizon, and MedPace. JAS reports research support from BMS and consultancy fees from AbbVie, Amgen, Boehringer Ingelheim, BMS, Gilead, Inova, Janssen, Optum, Pfizer, and ReCor unrelated to this work. The remaining authors declare no conflicts of interest relevant to this article.

Z.S. Wallace and J.A. Sparks are co-senior authors.

This is an Open Access article, which permits use, distribution, and reproduction, without modification, provided the original article is correctly cited and is not used for commercial purposes.