Background Reduced response to pandemic (2009) H1N1 (pH1N1) vaccine in patients with rheumatoid arthritis (RA) was recently reported.
Objectives To evaluate the contribution of age, disease activity, medication and previous antibody levels to this reduced response.
Methods 340 adult RA patients and 234 healthy controls were assessed before and 21 days after adjuvant-free influenza A/California/7/2009 (pH1N1) vaccine. Disease activity (DAS28), current treatment and pH1N1 antibody titres were collected. Seroprotection, seroconversion and factor increase in geometric mean titre (GMT) were calculated and adverse events registered.
Results RA and controls showed similar (p>0.05) prevaccination GMT (8.0 vs 9.3) and seroprotection (10.8% vs 11.5%). After vaccination a significant reduction (p<0.001) was observed in all endpoints: GMT and factor increase in GMT, seroprotection and seroconversion rates. Disease activity did not preclude seroconversion or seroprotection and remained unchanged in 97.4% of patients. Methotrexate was the only disease-modifying antirheumatic drug associated with reduced responses (p=0.001). Vaccination was well tolerated.
Conclusions The data confirmed both short-term anti-pH1N1 vaccine safety and, different from most studies with seasonal influenza, reduced seroprotection in RA patients, unrelated to disease activity and to most medications (except methotrexate). Extrapolation of immune responses from one vaccine to another may therefore not be possible and specific immunisation strategies (possibly booster) may be needed.
Clinicaltrials.gov no NCT01151644.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
The benefits of vaccination in avoiding vaccine-preventable diseases in patients with immune-inflammatory diseases such as rheumatoid arthritis (RA) are accepted.1,–,3 Nevertheless, there is concern regarding vaccine safety and efficacy.2 This is particularly important for emergent pandemic diseases such as pandemic (2009) H1N1 (pH1N1) influenza.
After the pH1N1A/California/7/2009 pandemic,4 inactivated monovalent vaccines became available through national programmes.5 Vaccination dosing schedules were developed in healthy individuals.6 We recently reported, in a large cohort of autoimmune rheumatic diseases including 343 RA patients, the overall short-term safety but reduced immunogenicity of a new adjuvant-free influenza A p/H1N1 vaccine,7 without addressing correlation with RA activity and medication. Our study assessed the relationship between reduced response to adjuvant-free pH1N1 vaccine and both RA disease activity and treatment with classic and biological disease-modifying antirheumatic drugs (DMARD).
Study design and participants
This prospective single-centre study was conducted during the public national health influenza A pH1N1(2009) vaccination campaign in Brazil. It was approved by the Institutional Review Board (clinicaltrials.gov no NCT01151644) and included two stages, vaccination (from 22 March to 2 April) and a 21-day follow-up.
RA (1987 American College of Rheumatology criteria)8 patients in regular follow-up at the outpatient clinics were recruited through invitation letters. Out of 562 invited patients, 391 (69.6%) received the vaccine and 349 (89.3%) completed the two phases; 340 (87.0%) patients were included in this study (complete serology, clinical and therapeutic data). Healthy control individuals were recruited from the hospital's immunisation centre. Among 326 vaccinated controls, 234 (71.8%) completed the study.
Inclusion and exclusion criteria
Inclusion and exclusion criteria are available in supplementary table S1 (available online only) and were similar to those previously described.7
The vaccine, Sanofi Pasteur Influenza A (H1N1) 2009 was a novel monovalent adjuvant-free vaccine (A/California/7/2009/Butantan Institute/Sanofi Pasteur, São Paulo, Brazil). The active component was a split inactivated influenza virus containing 15 μg haemagglutinin of influenza A/California/07/2009 (pH1N1) virus-like strain (NYMCx-179A) per 0.5-ml dose.7 It was available as 5-ml multidose vials with thimerosal (45 μg/0.5-ml dose) added as preservative.
All subjects were vaccinated with a single intramuscular dose of pH1N1 vaccine. Disease activity (disease activity score in 28 joints (DAS28)–erythrocyte sedimentation rate (ESR))9 evaluation and blood sample collection were performed before and 21 days after the vaccination.
A 21-day symptom diary card for prospective completion was given to each participant following vaccination and was returned 21 days later. All new symptoms, recorded or not in the diary, were reviewed by the investigators and causal relation with the vaccine was assessed. All RA patients answered one specific (yes or no) question about their perception of vaccine interference on disease activity responses.
The immunogenicity of pH1N1 A/California/7/2009-like virus vaccine was evaluated (haemagglutination inhibition assay) at Adolfo Lutz Institute.7 10 Antibody titres were assessed at baseline and 21 days post-immunisation. Geometric mean titres (GMT) were calculated. Serological endpoints were evaluated: seroprotection rate defined as the percentage of patients with titre 1:40 and seroconversion rate as the percentage of patients with a fourfold or greater increase in vaccination titre if prevaccination titre was 1:10 or greater, or postvaccination titre was 1:40 or greater if prevaccination titre was less than 1/10. The factor increase in GMT was also calculated. ESR and CRP were assessed.
Two-sided 95% CI were calculated assuming binomial distributions for dichotomous variables (Clopper–Pearson method) and log normal distribution for haemagglutination inhibition titres. Categorical variables were compared by Fisher's exact test or the Fisher–Freeman–Halton exact test; normally or non-normally distributed variables were compared using the t test or Wilcoxon rank-sum test. A multiple logistic regression model was applied to analyse interaction between demographic characteristics, medications and seroconversion. All tests were two-sided, with a 0.05 significance level.
Age (55.8±11.5 vs 36.63±12.5 years, p<0.0001) and female predominance (86.7% vs 66.8%, p<0.0001) were significantly higher in RA patients compared with controls. Clinical data are depicted in table 1.
Disease activity modifications due to immunisation
Nearly all (97.4%) RA patients answering one specific question about changes in disease activity after immunisation reported no alterations. All nine (2.6%) patients describing some change related worsening of symptoms. In fact, their DAS28 increased due to a significant rise in all components, with no baseline variable able to single out this small patient group (see supplementary table S2, available online only).
Clinical and laboratory parameters of disease activity remained unchanged in the group as a whole, except for slight non-clinically significant decreases in DAS28 and swollen joint number (table 1).
Pre-existent antibodies to pH1N1 (2009) and response to immunisation
Prevaccination seroprotection and GMT were similar in both groups (p>0.05; table 2). Conversely, following vaccination, RA patients showed reduced seroprotection and seroconversion rates compared with controls. Despite a significant increase in postvaccination GMT in both groups, GMT and factor increase in GMT were significantly lower in RA patients compared with controls (table 2).
Due to age differences, subanalysis was performed comparing 88 RA patients and 184 age-matched controls (mean ages 40.8 years (39.4 to 42.1) vs 40.8 years (39.5 to 42.1); p>0.05). There were comparable (p>0.05) prevaccination seroprotection rates (8.0% (2.3 to 13.6); 9.8% (5.5 to 14.1)) and GMT (8.2 (6.9 to 9.8); 9.0 (7.9 to 10.2)) and lower postvaccination seroprotection rates (67.0% (57.2 to 76.9); 83.2% (77.8 to 88.6); p=0.04) and seroconversion rates (63.6% (53.5 to 73.7); 76.1% (69.9 to 82.3); p=0.005) in RA patients. Postvaccination GMT in this RA subgroup (83.9 (60.3 to 116.7); 116.6 (96.2 to 141.3)) and factor increase in GMT (10.2 (7.5 to 13.9); 13.0 (10.7 to 15.8)) values were non-statistically significantly lower. Data on age subsets are shown in supplementary table S3 and figure S1, available online only.
Influence of disease characteristics and medication
High disease activity levels (14.5% of patients) did not preclude immune response (figure 1).
Patients were in regular follow-up receiving traditional and biological DMARD, usually in combination. Pre and postvaccination GMT and seroprotection rates, factor increase in GMT and seroconversion rates in patients with or without medications are depicted in table 2 and compared with controls.
Anti-tumour necrosis factor (TNF) agents were grouped together due to small patient numbers: 20, 16 and 11 on infliximab, adalimumab and etanercept, respectively. Only 11 (3.2%) patients were not medicated and 32 (9.4%) were on monotherapy. Patients on methotrexate had the lowest responses (table 2).
After multivariate analyses, only age, RA and methotrexate were associated with impaired seroconversion (p<0.05). The probability of methotrexate-related reduction in seroconversion was 49% (OR 0.51; CI 95% 0.32 to 0.82; p=0.005).
The vaccine was well tolerated; no severe side effects were reported during follow-up. No patient sought extra consultations, although reporting significantly more adverse events (140 events/100 patients vs 87/100 controls, p<0.005). Minor local reactions were more frequent in controls (p=0.032) and mild systemic reactions (p=0.01) in patients (see supplementary table S4, available online only). Local symptoms prevailed during the first days (88%), while systemic symptoms clustered around the first 15 days (73%) postvaccination.
To the best of our knowledge, this is the largest study addressing adjuvant-free pH1N1 (2009) vaccine in active RA patients undergoing immunosuppressive treatment under routine care. Although our patients were predominantly women (seven women and one man), as usual in Brazilian RA series,11 gender involvement is unlikely, because women reportedly have higher antibody titres to viral and bacterial vaccines.12
Age is a recognised factor in vaccine response. However, immunogenicity was similarly impaired in different age subsets and after age-matched subanalysis, as also reported for heterogeneous autoimmune rheumatic diseases.7 13 Furthermore, after multivariate analysis, RA was an independent contributor to immune response impairment.
This reduced response could be related to the intrinsic RA immune state, disease activity, medication or lower vaccine immunogenicity. Disease activity neither precluded immune response nor was related to adverse events, a fact relevant to future vaccination guidelines.1 As most patients were on combination therapy, neither RA itself nor medications could be independently assessed. Nevertheless, average/high effective doses of methotrexate were associated with reduced response at all endpoints. A recent study of adjuvanted pH1N1 vaccine also linked methotrexate to a lower antibody vaccine response,14 while studies with seasonal trivalent vaccine showed no similar interference.1 2 14,–,16
We noted no effect of corticosteroids, traditional DMARD and TNF-blockers on vaccine immune response. Conversely, leflunomide and other immunosuppressors were associated with lower postvaccination GMT in an adjuvanted pandemic H1N1 vaccine trial.14 In line with our observations, no interference of TNF-blockers was verified. Former trials with trivalent seasonal influenza vaccine reported no consistent influence of classic DMARD or corticosteroids.1 2 14 16 17 Overall, TNF inhibitors lead to adequate or slightly reduced seroprotection rates.1,–,3 14,–,19
This study helped clarify the controversy of adjuvanted versus adjuvant-free vaccines, because our results are similar to those of one-dose adjuvanted pH1N1 vaccine in 82 RA patients.13 In addition, when different preparations of influenza A H1N1 (2009) vaccine were tested in a large Chinese population, the adjuvant-free formulation was more effective with equivalent antigen content.20
Our data clearly settle the short-term safety of influenza A (pH1N1) 2009 vaccine in RA patients. Inflammatory tests remained unchanged and worsening of the disease, reported by nine out of 340 patients, cannot be distinguished from natural disease flares. The slight decrease in DAS28 and swollen joints after vaccination was well below the clinically significant value.9 Moreover, our patients, under regular treatment aimed at disease control, were not preselected, regarding medications or duration of treatment, as our study was developed during the national anti-pH1N1 vaccination campaign.5
The observed lower seroconversion rate in RA patients, particularly on methotrexate, has broader implications and may place significant numbers of immunosuppressed individuals at risk of infection despite vaccination. Inferring from the improved response to a second dose (booster) of adjuvanted pH1N1 vaccine13 and the response to seasonal influenza vaccine in RA patients, this study highlights the need for specific immunisation strategies for each new vaccine in this population.
Funding This study was supported by grants from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP #2010/10749-0 to EB), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPQ #303165/2008-1 to EFB, #300248/2008-3 to CAS, #300665/2009-1 to JFC and #301411/2009-3 to EB), Federico Foundation (to EFB, JFC, PDSB and EB) and Butantan Foundation. The authors would like to thank the subjects for their critical roles in this study, the staff of Hospital das Clinicas FMUSP, Adolfo Lutz Institute and Faculdade de Medicina da USP.
Competing interests None.
Ethics approval This study was conducted with the approval of the local Institutional Review Board, Faculdade Medicina, Universidade de São Paulo.
Provenance and peer review Not commissioned; externally peer reviewed.