Factors influencing the efficacy of two injections of a pandemic 2009 influenza A (H1N1) nonadjuvanted vaccine in systemic lupus erythematosus

Authors

  • A. Mathian,

    Corresponding author
    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Université Pierre et Marie Curie, Université Paris 06, and INSERM UMR-S 945, Paris, France
    • Service de Médecine Interne 2, Hôpital la Pitié-Salpêtrière, 47-83 Boulevard de l'Hôpital, 75651 Paris Cedex 13, France
    Search for more papers by this author
    • Drs. Mathian and Devilliers contributed equally to this work.

  • H. Devilliers,

    1. Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France and Centre Hospitalier Universitaire de Dijon, Dijon, France
    Search for more papers by this author
    • Drs. Mathian and Devilliers contributed equally to this work.

  • A. Krivine,

    1. Groupement Hospitalier Cochin, AP-HP, Paris, France
    Search for more papers by this author
  • N. Costedoat-Chalumeau,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, and Université Pierre et Marie Curie Université Paris 06, Paris, France
    Search for more papers by this author
  • J. Haroche,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France
    Search for more papers by this author
  • D. Boutin-Le Thi Huong,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France
    Search for more papers by this author
  • B. Wechsler,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France
    Search for more papers by this author
  • B. Hervier,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Université Pierre et Marie Curie, Université Paris 06, and INSERM UMR-S 945, Paris, France
    Search for more papers by this author
  • M. Miyara,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Université Pierre et Marie Curie, Université Paris 06, and INSERM UMR-S 945, Paris, France
    Search for more papers by this author
  • N. Morel,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France
    Search for more papers by this author
  • N. Le Corre,

    1. INSERM UMR-S 945, Groupement Hospitalier Pitié-Salpêtrière, Paris, France
    Search for more papers by this author
  • L. Arnaud,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Université Pierre et Marie Curie, Université Paris 06, and INSERM UMR-S 945, Paris, France
    Search for more papers by this author
  • J. C. Piette,

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France
    Search for more papers by this author
  • L. Musset,

    1. Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Paris, France
    Search for more papers by this author
  • B. Autran,

    1. Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Université Pierre et Marie Curie, Université Paris 06, and INSERM UMR-S 945, Paris, France
    Search for more papers by this author
  • F. Rozenberg,

    1. Groupement Hospitalier Cochin, AP-HP, and EA 1833, Université Paris Descartes, Sorbonne Paris Cité, Paris, France
    Search for more papers by this author
  • Z. Amoura

    1. Centre de Référence National pour les Lupus et le Syndrome des Antiphospholipides, Groupement Hospitalier Pitié-Salpêtrière, AP-HP, Université Pierre et Marie Curie, Université Paris 06, and INSERM UMR-S 945, Paris, France
    Search for more papers by this author

Abstract

Objective

To assess the factors influencing the efficacy of 2 injections of a pandemic 2009 influenza A (H1N1) vaccine in patients with systemic lupus erythematosus (SLE).

Methods

We conducted a single-center, observational prospective study of 111 patients who were vaccinated with a monovalent, inactivated, nonadjuvanted, split-virus vaccine during December 2009 and January 2010 and received a second dose of vaccine 3 weeks later. The antibody response was evaluated using the hemagglutination inhibition assay according to the guidelines recommended for the pandemic vaccine, consisting of 3 immunogenicity criteria (i.e., a seroprotection rate of 70%, a seroconversion rate of 40%, and a geometric mean ratio [GMR] of 2.5).

Results

The 3 immunogenicity criteria were met on day 42 (seroprotection rate 80.0% [95% confidence interval (95% CI) 72.5–87.5%], seroconversion rate 71.8% [95% CI 63.4–80.2%], and GMR 10.3 [95% CI 2.9–14.2]), while only 2 criteria were met on day 21 (seroprotection rate 66.7% [95% CI 57.9–75.4%], seroconversion rate 60.4% [95% CI 51.3–69.5%], and GMR 8.5 [95% CI 3.2–12.0]). The vaccine was well tolerated. Disease activity, assessed by the Safety of Estrogens in Lupus Erythematosus National Assessment version of the SLE Disease Activity Index, the British Isles Lupus Assessment Group score, and the Systemic Lupus Activity Questionnaire, did not increase. In the multivariate analysis, vaccination failure was significantly associated with immunosuppressive treatment or a lymphocyte count of ≤1.0 × 109/liter. The second injection significantly increased the immunogenicity in these subgroups, but not high enough to fulfill the seroprotection criterion in patients receiving immunosuppressive treatment.

Conclusion

Our findings indicate that the efficacy of the vaccine was impaired in patients who were receiving immunosuppressive drugs or who had lymphopenia. A second injection increased vaccine immunogenicity without reaching all efficacy criteria for a pandemic vaccine in patients receiving an immunosuppressive agent. These results open possibilities for improving anti-influenza vaccination in SLE.

Human infection with pandemic 2009 influenza A (H1N1) was first identified in April 2009 (1). On June 11, 2009, the World Health Organization declared the first influenza pandemic in 41 years. This novel strain is genetically and antigenically distinct from other H1N1 influenza strains that have been in circulation since 1977 (2), and most of the world's population, especially individuals younger than 60 years, are thought to have low levels of or no preexisting circulating antibody against the pandemic strain (3). As a consequence, this new strain was dominant in the Southern and Northern Hemispheres during the 2009–2010 influenza season.

Patients with systemic lupus erythematosus (SLE), who are often young and have an increased risk of infection related to the intrinsic disturbances of immune responses, the use of immunosuppressive drugs, and the associated organ complications of the disease (4–6), were considered to be at high risk of severe 2009 influenza A (H1N1) infection (7). Furthermore, viral illnesses are thought to be a potential cause of exacerbation of SLE. Since seasonal influenza vaccination has been shown in many studies to be well tolerated and to result in no change in disease activity in SLE (8–19), although it may trigger the short-term generation of autoantibodies in some patients (18, 20), it was recommended that SLE patients receive the pandemic influenza vaccine (21).

Data on vaccine immunogenicity in patients with autoimmune diseases are poor (22, 23). The immune dysfunction that increases the risk and complications of influenza infection in SLE may also compromise vaccine responses. Yet, factors influencing vaccine immunogenicity are not precisely known. As a consequence, there is still controversy regarding the efficacy of trivalent seasonal influenza vaccines in SLE, and while several studies have shown normal efficacy of seasonal influenza vaccination in SLE (11–14), others have suggested a reduced antibody response compared with that in healthy adults (8–10, 15–18, 24).

The present observational prospective study was conducted in a real-life setting to assess the factors influencing the immunogenicity of 2 injections of the 2009 influenza A (H1N1) nonadjuvanted vaccine in a large cohort of SLE patients, some of whom had active disease. Additionally, we studied the safety of the vaccine. These data should help to define new strategies to improve the efficacy of the influenza vaccine in SLE patients.

PATIENTS AND METHODS

Study design and patients.

Starting in November 2009 and following the recommendations of the French health authorities (21), SLE patients who were seen in our department (the French national reference center for SLE) were offered vaccination against pandemic 2009 influenza A (H1N1). Between November 30, 2009 and January 28, 2010, we conducted an observational prospective longitudinal cohort study. The primary objective was to determine the factors influencing vaccine immunogenicity. The secondary objective was to describe the safety of the vaccine. SLE patients were eligible for the study if they fulfilled at least 4 of the 1997 American College of Rheumatology criteria for SLE (25). Exclusion criteria were anaphylactic response to ovalbumin or other vaccine components, acute infection with a fever >38°C, and a history of Guillain-Barré syndrome. We did not exclude patients for particularly high disease activity or a particular visceral involvement.

We hypothesized that immunosuppressive drugs would induce a 2-fold decrease in the seroconversion rate (26). Prior data indicate that the seroconversion rate for healthy subjects is ∼0.75 (27). Using Fisher's exact test and assuming that a third of SLE patients are treated with immunosuppressive drugs, we would have to study at least 90 patients to be able to show a difference between the seroconversion rates of patients receiving immunosuppressive drugs and those not receiving immunosuppressive drugs, with a power of 0.9 and an alpha error of 0.05.

On day 0, SLE patients received an intramuscular injection into the deltoid (or a subcutaneous injection if the patient had thrombocytopenia or was receiving oral anticoagulants) of a dose of the influenza A (H1N1) nonadjuvanted vaccine (Panenza; Sanofi Pasteur). On day 21, patients received a second (booster) vaccination as recommended by French health authorities (21). Serum was obtained on days 0, 21, and 42. At each visit, the Safety of Estrogens in Lupus Erythematosus National Assessment (SELENA) version of the SLE Disease Activity Index (SLEDAI) (28–30), the British Isles Lupus Assessment Group (BILAG) score (31–32), the patient-reported disease activity Systemic Lupus Activity Questionnaire (SLAQ) results (33), and the physician's global assessment, a score on a visual analog scale of 0–3 measuring disease activity as judged by the physician, were recorded. Routine methods were used to determine anti–double-stranded DNA (anti-dsDNA) antibody titer (by Farr assay), complement C3 level, complete blood cell count, serum levels of creatinine, albumin, and total IgG and IgM, and the presence of proteinuria and hematuria. Information on seasonal influenza vaccination in the previous years and influenza-like illness preceding the vaccine was collected. On day 21, we collected reports of local and systemic adverse events using a standardized questionnaire, which included itching, pain, redness, induration, tenderness, or ecchymosis at the site of vaccination and chills, fever, headache, malaise, nausea, vomiting, and myalgia. Patients used a scale to categorize events as none, mild, moderate, or severe.

In order to monitor for postvaccine failure, patients were seen on days 21 and 42 and at months 6 and 12 and asked whether “acute respiratory symptoms with fever >38°C or systemic symptoms” had occurred. In addition, during the 1-year followup, we planned to take nasal swabs in patients who reported for visits within 72 hours of the onset of upper respiratory tract infection symptoms. The samples were to undergo real-time reverse transcriptase–polymerase chain reaction assay for 2009 H1N1 virus. However, since no patients reported for swab, this additional analysis was not performed. The study was approved by the local ethics committee, and informed consent was obtained from all participants.

Vaccine.

The vaccine (Panenza), a monovalent, inactivated, nonadjuvanted, split-virus influenza vaccine against the reassortant vaccine virus A/California/7/2009 (H1N1) v–like strain (NYMC X-179A), was prepared in embryonated chicken eggs using the same standard techniques that are used for the production of seasonal trivalent inactivated vaccine. The vaccine was packaged in 5-ml multidose vials with thimerosal added as a preservative (final concentration 0.01% weight per volume). There were 15 μg of hemagglutinin per 0.5-ml vaccine dose.

Antibody response to influenza.

Quantitative detection of antibodies against pandemic 2009 influenza A (H1N1) was performed by hemagglutination inhibition test using a modification of the method of Kendal et al (34). Briefly, after treatment with receptor destroying enzyme, 2-fold dilutions of sera, beginning with a dilution of 1:10, were tested against 4 hemagglutinin units of antigen (Panenza) on human type O Rh− red blood cells. The titer of hemagglutination-inhibiting antibodies was defined as the highest serum dilution that completely inhibited hemagglutination. Titers <1:10 were assigned a value of 1:5 for calculation purposes. The results of immunogenicity assays performed 21 days after the first vaccination (day 21) and 21 days after the second vaccination (day 42) were compared with the results obtained at baseline (day 0). Geometric mean titers (GMTs) of antibody were calculated. The antibody response was evaluated in 3 ways: the proportion of subjects with antibody titers ≥1:40 (seroprotection), the proportion of subjects with either a prevaccination hemagglutination-inhibiting titer <1:10 and a postvaccination titer ≥1:40 or a prevaccination titer ≥1:10 and an increase in the titer by a factor of 4 or more (seroconversion), and the factor increase in the GMT after vaccination (geometric mean ratio [GMR]). The 3 coprimary immunogenicity end points after vaccination were chosen according to international guidelines used to evaluate influenza vaccines for subjects ages 18–60 years (35, 36). Cutoff levels for vaccine immunogenicity were 70% for the seroprotection rate, 40% for the seroconversion rate, and 2.5 for the GMR (i.e., postvaccination GMT 2.5 fold higher than prevaccination GMT). In a pandemic context, all 3 criteria should be fulfilled (37).

Statistical analysis.

The 95% confidence intervals (95% CIs) for coprimary immunogenicity end points were calculated on the basis of the binomial distribution. For immunogenicity analyses, the proportions of subjects in whom seroconversion or seroprotection was achieved were compared after the first and the second injection using chi-square test. Significant variables were entered into a logistic regression model with stepwise selection of variables (P = 0.30 for entry and P = 0.10 for exit). Increases in seroconversion and seroprotection rates between injections were studied using generalized estimating equations (GEEs), with Bonferroni correction for multiple pairwise comparisons when appropriate. GMTs and 95% CIs were computed by taking the exponent (log10) of the mean and of the limits of the 95% CIs of the log10-transformed titers. GMTs were compared between each pair of vaccine groups by means of an analysis of variance on the log10-transformed titers. For safety analysis, GEEs and Friedman's nonparametric analysis of variance were used. We used 2-sided tests, and P values less than 0.05 were considered significant. Statistical analyses were performed using SAS software, version 9.1.3.

RESULTS

Patient characteristics.

According to the French guidelines (21), we invited the 322 SLE patients seen in our department between November 30, 2009 and January 28, 2010 to undergo vaccination against pandemic 2009 influenza A (H1N1) and to participate in an observational study on its efficacy and safety. Of the 322 patients, 135 (41.9%) agreed to be vaccinated, and 111 (34.5%) agreed to participate in the study. One patient was lost to followup between day 21 and day 42. The baseline characteristics of the patients are shown in Table 1. A total of 22 patients (19.8%) had a baseline SELENA–SLEDAI score of ≥6, and 28 patients (25.2%) had a least 1 BILAG grade of A or B, including 15 patients with 1 BILAG grade of A or 1 BILAG grade of B in the renal system. None of the patients had a BILAG grade of A or B in the nervous system.

Table 1. Baseline characteristics and disease parameters of the SLE patients*
 Patients (n = 111)
  • *

    Except where indicated otherwise, values are the number (%) of patients. SLE = systemic lupus erythematosus; BILAG = British Isles Lupus Assessment Group; SLAQ = Systemic Lupus Activity Questionnaire; GMT = geometric mean titer; 95% CI = 95% confidence interval.

  • Defined using the flare instrument of the Safety of Estrogens in Lupus Erythematosus National Assessment (SELENA) version of the SLE Disease Activity Index (SLEDAI).

  • Excluding antimalarials and prednisone.

Women102 (91.9)
Age, mean ± SD years35.2 ± 10.6
Disease duration, mean ± SD years10.3 ± 8.0
Flu-like symptoms since May 200925 (22.5)
Any seasonal influenza vaccine36 (32.4)
2008 seasonal influenza vaccine15 (13.5)
2009 seasonal influenza vaccine25 (22.5)
SELENA–SLEDAI score, median (range)2 (0–32)
SELENA–SLEDAI score ≥622 (19.8)
≥1 BILAG grade of A or B28 (25.2)
Physician's global assessment, median (range)0 (0–3)
SLAQ score, median (range)3 (0–38)
Mild/moderate flare3 (2.7)
Severe flare8 (7.2)
Hydroxychloroquine103 (92.8)
Daily prednisone87 (78.4)
Prednisone ≥0.15 mg/kg/day38 (34.2)
Immunosuppressive agents38 (34.2)
Positive Farr assay50 (45.1)
Low C313 (11.7)
Serum IgG ≤7 gm/liter10 (9.0)
Serum IgM ≤1 gm/liter52 (46.8)
Prevaccination hemagglutination-inhibiting titer ≥1:4017 (15.3)
Prevaccination hemagglutination-inhibiting GMT (95% CI)9.4 (8.0–11.2)

Of the patients who were receiving immunosuppressive agents, 21 were receiving mycophenolate mofetil, 8 were receiving azathioprine, 5 were receiving cyclophosphamide, and 4 were receiving methotrexate. In 7 patients, either the dose of prednisone had to be increased or an immunosuppressive drug was introduced less than 1 month before vaccination, due to SLE activity.

Immunogenicity of the pandemic 2009 influenza A (H1N1) vaccine in SLE patients.

At baseline, prevaccination hemagglutination-inhibiting titers of ≥1:40 were observed in 9 of 25 patients (36.0%) who had received the 2009 seasonal vaccine as compared with 8 of 78 patients (10.3%) who had not received the seasonal vaccine (P = 0.003 by Fisher's exact test). The proportion of patients with a baseline antibody titer of ≥1:40 was not significantly higher in patients older than 50 years, in those who had had flu-like symptoms since May 2009, or in those who had received the 2008 seasonal vaccine (data not shown). According to the international guidelines used to evaluate influenza vaccines (35–37), only 2 of the 3 immunogenicity criteria were met on day 21 in the whole study group. The seroconversion rate was 60.4% (95% CI 51.3–69.5%) (above the required level of 40%), and the GMR was 8.5 (95% CI 3.2–12.0) (above the required level of 2.5), while the seroprotection rate was 66.7% (95% CI 57.9–75.4%) (below the required level of 70%) (Table 2).

Table 2. Seroconversion rate, seroprotection rate, and GMR in SLE patients vaccinated against pandemic 2009 influenza A (H1N1)*
 Day 0 (n = 111)Day 21 (n = 111)Day 42 (n = 110)
  • *

    The cutoff levels for vaccine immunogenicity were 70% for the seroprotection rate, 40% for the seroconversion rate, and 2.5 for the geometric mean ratio (GMR). In a pandemic context, all 3 criteria (seroconversion rate, seroprotection rate, and significant increase in geometric mean titer) should be exceeded (37). SLE = systemic lupus erythematosus; 95% CI = 95% confidence interval; NA = not applicable.

  • P < 0.01 versus day 21.

  • P < 0.0001 versus day 0.

  • §

    P < 0.001 versus day 21.

  • P < 0.0001 versus day 21.

Seroconversion rate, % (95% CI)NA60.4 (51.3–69.5)71.8 (63.4–80.2)
Seroprotection rate, % (95% CI)15.3 (8.6–22)66.7 (57.9–75.4)80.0 (72.5–87.5)§
GMR (95% CI)NA8.5 (3.2–12)10.3 (2.9–14.2)

After the second vaccine administration, the 3 immunogenicity criteria were fully met on day 42. The seroconversion rate was 71.8% (95% CI 63.4–80.2%) (P = 0.003 versus day 21), the seroprotection rate was 80.0% (95% CI 72.5–87.5%) (P = 0.0002 versus day 21), and the GMR was 10.3 (95% CI 2.9–14.2) (P < 0.0001 versus day 21). A second (booster) injection allowed 12 additional patients (10.9%) and 14 additional patients (12.7%) to fulfill seroconversion and seroprotection criteria, respectively. During the 1-year followup, 1 patient developed symptoms suggestive of influenza infection; however, since the patient did not report for the followup visit, we were not able to confirm the presence of influenza virus.

Factors influencing the immunogenicity of the vaccine.

Seroconversion rate.

In bivariate analysis, failure to obtain seroconversion on day 21 was significantly associated with active disease (defined as a SELENA–SLEDAI score of ≥6 or at least 1 BILAG grade of A or B), a lymphocyte count of ≤1.0 × 109/liter, a serum level of total IgM of ≤1 gm/liter, a prednisone dosage of ≥0.15 mg/kg daily, and immunosuppressive treatment at baseline (Table 3). Failure to obtain seroconversion on day 42 was significantly associated with a lymphocyte count of ≤1.0 × 109/liter, a serum level of total IgG of ≤7 gm/liter, and immunosuppressive treatment on day 21. In the multivariate analysis, failure to obtain seroconversion on both days 21 and 42 was significantly associated with immunosuppressive treatment or a lymphocyte count of ≤1.0 × 109/liter at the previous visit (Table 3).

Table 3. Predictive factors of seroconversion in SLE patients after the first injection (day 21) and after the second (booster) injection (day 42) of pandemic 2009 influenza A (H1N1) vaccine*
Patient characteristicSeroconversion on day 21Seroconversion on day 42
No (n = 44)Yes (n = 67)PNo (n = 31)Yes (n = 79)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. Baseline data were used to determine factors associated with seroconversion on day 21, and data from day 21 were used to determine factors associated with seroconversion on day 42. See Table 1 for other definitions.

  • Bivariate analysis. P values were determined by Mann-Whitney U test for continuous variables and chi-square test for categorical variables. In multivariate analysis (stepwise logistic regression model with P = 0.3 for entry and P = 0.1 for exit), the factors associated with failure to obtain seroconversion on day 21 were use of immunosuppressive agents (P = 0.023) (odds ratio [OR] 2.8 [95% CI 1.1–6.7]), and lymphocyte count of ≤1,000/mm3 (P = 0.019) (OR 2.8 [95% CI 1.2–6.5]). In multivariate analysis, the factors associated with failure to obtain seroconversion on day 42 were use of immunosuppressive agents (P = 0.043) (OR 2.6 [95% CI 1.0–6.4]) and lymphocyte count of ≤1,000/mm3 (P = 0.046) (OR 2.6 [95% CI 1.0–6.4]).

  • Excluding antimalarials and prednisone.

Women40 (91)62 (93)0.7628 (90)73 (92)0.72
Age, median ± SD years36.0 ± 11.034.7 ± 10.40.4135.9 ± 11.234.8 ± 10.30.50
Disease duration, mean ± SD years9.8 ± 8.110.6 ± 8.00.569.5 ± 8.910.4 ± 7.60.32
Any seasonal influenza vaccine15 (34)21 (31)0.7614 (45)22 (28)0.08
SELENA–SLEDAI score ≥613 (30)9 (13)0.049 (29)11 (14)0.06
≥1 BILAG grade of A or B17 (39)11 (16)<0.0112 (39)18 (23)0.09
Hydroxychloroquine40 (91)63 (94)0.5328 (90)74 (94)0.54
Prednisone ≥0.15 mg/kg/day22 (50)16 (24)<0.0112 (39)25 (32)0.48
Immunosuppressive agents23 (52)15 (22)<0.0117 (55)21 (27)<0.01
Positive Farr assay22 (50)28 (42)0.4015 (48)35 (44)0.70
Low C38 (18)5 (7)0.095 (16)11 (14)0.77
Serum IgG ≤7 gm/liter5 (11)5 (7)0.487 (23)3 (4)<0.01
Serum IgM ≤1 gm/liter28 (64)24 (36)<0.0119 (61)38 (48)0.21
Lymphocyte count ≤1,000/mm325 (57)18 (27)<0.0116 (52)19 (24)<0.01

Seroprotection rate.

In bivariate analysis, failure to obtain seroprotection on day 21 was significantly associated with a low C3 level and with the same parameters associated with failure to obtain seroconversion on day 21 (Table 4). Failure to obtain seroprotection on day 42 was significantly associated with active disease (defined as a SELENA–SLEDAI score of ≥6 or at least 1 BILAG grade of A or B) and with the same parameters associated with failure to obtain seroconversion on day 42. In the multivariate analysis, failure to obtain seroprotection on days 21 and 42 was significantly associated with immunosuppressive treatment or a lymphocyte count of ≤1.0 × 109/liter at the previous visit (Table 4). Failure to obtain seroprotection on day 21 was also significantly associated with a serum level of total IgM of ≤1 gm/liter (Table 4).

Table 4. Predictive factors of seroprotection in SLE patients after the first injection (day 21) and after the second (booster) injection (day 42) of pandemic 2009 influenza A (H1N1) vaccine*
 Seroprotection on day 21Seroprotection on day 42
Patient characteristicNo (n = 37)Yes (n = 74)PNo (n = 22)Yes (n = 88)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. Baseline data were used to determine factors associated with seroprotection on day 21, and data from day 21 were used to determine factors associated with seroprotection on day 42. See Table 1 for other definitions.

  • Bivariate analysis. P values were determined by Mann-Whitney U test for continuous variables and chi-square test for categorical variables. In multivariate analysis (stepwise logistic regression model with P = 0.3 for entry and P = 0.1 for exit), the factors associated with failure to obtain seroprotection on day 21 were use of immunosuppressive agents (P = 0.039) (odds ratio [OR] 2.6 [95% CI 1.1–6.3]), serum IgM ≤1 gm/liter (P = 0.040) (OR 2.6 [95% CI 1.1–6.3]), and lymphocyte count of ≤1,000/mm3 (P = 0.020) (OR 2.8 [95% CI 1.2–6.9]). In multivariate analysis, the factors associated with failure to obtain seroconversion on day 42 were use of immunosuppressive agents (P = 0.005) (OR 4.6 [95% CI 1.6–13.2]) and lymphocyte count of ≤1,000/mm3 (P = 0.048) (OR 2.9 [95% CI 1.1–8.2]).

  • Excluding antimalarials and prednisone.

Women35 (95)67 (91)0.4620 (91)81 (92)0.86
Age, median ± SD years36.7 ± 10.734.5 ± 10.50.2236.5 ± 11.934.7 ± 10.20.46
Disease duration, mean ± SD years9.7 ± 7.910.6 ± 8.10.598.3 ± 8.910.6 ± 7.70.07
Any seasonal influenza vaccine11 (30)25 (34)0.678 (36)28 (32)0.68
SELENA–SLEDAI score ≥612 (32)10 (14)0.029 (41)11 (13)<0.01
≥1 BILAG grade of A or B15 (41)13 (18)<0.0111 (50)19 (22)<0.01
Hydroxychloroquine34 (92)69 (93)0.8020 (91)82 (93)0.71
Prednisone ≥0.15 mg/kg/day20 (54)18 (24)<0.0110 (45)27 (31)0.19
Immunosuppressive agents20 (54)18 (24)<0.0115 (68)23 (26)<0.001
Positive Farr assay21 (57)29 (39)0.0812 (55)38 (43)0.34
Low C38 (22)5 (7)0.024 (18)12 (14)0.59
Serum IgG ≤7 gm/liter4 (11)6 (8)0.645 (23)5 (6)0.01
Serum IgM ≤1 gm/liter25 (68)27 (36)<0.0115 (68)42 (48)0.09
Lymphocyte count ≤1,000/mm322 (59)21 (28)<0.0113 (59)22 (25)<0.01

Geometric mean ratios.

GMR values were significantly lower in SLE patients who were receiving immunosuppressive drugs than in those who were not. (Data are available from the author upon request.)

Effect of a second (booster) injection on vaccination efficacy.

To study the impact of the second injection on vaccination efficacy, we performed a subanalysis in which we studied the seroconversion rate, the seroprotection rate, and the GMR in different subgroups of patients (Table 5).

Table 5. Seroconversion rate, seroprotection rate, and GMR in different subgroups of SLE patients vaccinated against pandemic 2009 influenza A (H1N1)*
 Day 0Day 21Day 42
  • *

    The cutoff levels for vaccine immunogenicity were 70% for the seroprotection rate, 40% for the seroconversion rate, and 2.5 for the geometric mean ratio (GMR). In a pandemic context, all 3 criteria (seroconversion rate, seroprotection rate, and significant increase in geometric mean titer) should be exceeded (37). SLE = systemic lupus erythematosus; 95% CI = 95% confidence interval; NA = not applicable.

  • P < 0.05 versus day 21.

  • P < 0.001 versus day 0.

  • §

    P < 0.001 versus day 21.

  • P < 0.0001 versus day 21.

  • #

    P < 0.0001 versus day 0.

Patients taking immunosuppressive drugs (n = 38)   
 Seroconversion rate, % (95% CI)NA39.5 (23.9–55.0)55.3 (39.5–71.1)
 Seroprotection rate, % (95% CI)10.5 (0.8–20.3)47.4 (31.5–63.2)60.5 (45.0–76.1)
 GMR (95% CI)NA4.0 (2.2–7.0)5.5 (3.1–9.5)§
Patients with a lymphocyte count of ≤1.0 × 109/liter (n = 43)   
 Seroconversion rate, % (95% CI)NA41.9 (27.1–56.6)67.4 (53.4–81.4)§
 Seroprotection rate, % (95% CI)14.0 (3.6–24.3)48.8 (33.9–63.8)74.4 (61.4–87.5)§
 GMR (95% CI)NA5.3 (3.2–9.1)8.2 (5.0–13.5)
Patients who were not taking immunosuppressive drugs and  had a lymphocyte count of >1.0 × 109/liter (n = 50)   
 Seroconversion rate, % (95% CI)NA80.0 (68.9–91.1)83.7 (73.3–94.0)
 Seroprotection rate, % (95% CI)20.0 (8.9–31.1)84.0 (73.8–94.2)#91.8 (84.2–99.5)
 GMR (95% CI)NA15.7 (9.6–25.7)16.9 (10.8–26.5)

Patients receiving immunosuppressive drugs.

In the subgroup of SLE patients who were treated with an immunosuppressive drug (n = 38), only 1 immunogenicity criterion was met on day 21. The seroconversion rate was 39.5% (95% CI 23.9–55%) (below the required level of 40%), the seroprotection rate was 47.4% (95% CI 31.5–63.2%) (below the required level of 70%), and the GMR was 4 (95% CI 2.2–7.0) (above the required level of 2.5) (35–37). After the second vaccine injection, antibody levels increased, but only 2 of the immunogenicity criteria were met on day 42. On day 42, the seroconversion rate was 55.3% (95% CI 39.5–71.1%) (P = 0.014 versus day 21), the GMR was 5.5 (95% CI 3.1–9.5) (P < 0.001 versus day 21), and the seroprotection rate was 60.5% (95% CI 45.0–76.1%) (significantly increased compared with day 21 [P = 0.025] but below 70%).

Patients with a lymphocyte count of ≤1.0 × 109/liter.

In the subgroup of SLE patients with a lymphocyte count of ≤1.0 × 109/liter (n = 43), 2 immunogenicity criteria were met on day 21. The seroconversion rate was 41.9 (95% CI 27.1–56.6%), and the GMR was 5.3 (95% CI 3.2–9.1), while the seroprotection rate was 48.8% (95% CI 33.9–63.8%). After the second vaccine injection, antibody levels increased, and all 3 of the immunogenicity criteria were met on day 42. On day 42, the seroprotection rate was 74.4 (95% CI 61.4–87.5%) (P < 0.001 versus day 21), the seroconversion rate was 67.4% (95% CI 53.4–81.4%) (P < 0.001 versus day 21), and the GMR was 8.2 (95% CI 5.0–13.5) (P < 0.0001 versus day 21).

Patients who were not receiving immunosuppressive drugs and had a lymphocyte count of >1.0 × 109/liter.

In the subgroup of SLE patients who had a lymphocyte count of >1.0 × 109/liter and were not receiving immunosuppressive drugs (n = 50), all 3 immunogenicity criteria were met on day 21. The seroconversion rate was 80.0% (95% CI 68.9–91.1%), the seroprotection rate was 84.0% (95% CI 73.8–94.2%), and the GMR was 15.7 (95% CI 9.6–25.7). The increases in the seroconversion and seroprotection rates and in the GMR obtained on day 42, after the second injection, did not reach statistical significance. On day 42, the seroconversion rate was 83.7% (95% CI 73.3–94.0%) (P = 0.65 versus day 21), the seroprotection rate was 91.8% (95% CI 84.2–99.5%) (P = 0.08 versus day 21), and the GMR was 16.9 (95% CI 10.8–26.5) (P = 0.37 versus day 21). The immune response that was observed after the first dose of vaccine was sustained after the second dose. (Results are available from the author upon request.)

Adverse events.

None of the patients experienced serious side effects. At least 1 local adverse event was reported by 36.9% of the patients after the first vaccination and by 30.2% after the second vaccination. The most commonly reported local events were injection-site tenderness and pain (data not shown). All but 1 of the reported local adverse events were mild in intensity. At least 1 systemic adverse event was reported by 41.3% of the patients after the first vaccination and 42.3% of the patients after the second vaccination. The most commonly reported systemic adverse events were headache, myalgia, and chills. (Results are available from the author upon request.) The majority of systemic adverse events (93.3% of the events that occurred after the first injection and 93.2% of the events that occurred after the second injection) were mild in intensity. Generally, the pattern and frequency of adverse events after the second vaccination were similar to those observed after the first vaccination. Two adverse events of special interest were reported. One patient had a hysterical conversion reaction mimicking Guillain-Barré syndrome 3 days after the first injection, which resolved after 10 days. Another patient had a seizure 18 days after the second injection. This patient had epilepsy with recurrences of seizure several times a year prior to the vaccine injection.

Effects of the vaccine on disease activity and generation of autoantibodies.

The SELENA–SLEDAI, BILAG, and SLAQ scores on day 21 and day 42 did not differ significantly from the scores before vaccination (Table 6). On day 21, no patient had a new BILAG grade of A, 1 patient (0.9%) had 2 new BILAG grades of B (in the mucocutaneous and hematologic systems), 6 patients (5.4%) had 1 new BILAG grade of B (5 in the mucocutaneous system and 1 in the musculoskeletal system), and 8 patients (7.2%) had at least 1 new BILAG grade of C (7 in the musculoskeletal system and 1 in the mucocutaneous system). On day 42, no patient had a new BILAG grade of A, 1 patient (0.9%) had 1 new BILAG grade of B (in the mucocutaneous system), and 7 patients (6.3%) had at least 1 new BILAG grade of C (4 in the musculoskeletal system, 1 in the mucocutaneous system, 1 in the hematologic system, and 1 in both the mucocutaneous and hematologic systems). The symptoms observed were not responsible for a change in treatment in any of the patients. Anti-dsDNA antibody titers, as determined by Farr assay and enzyme-linked immunosorbent assay, and levels of anticardiolipin, anti–β2-glycoprotein I, anti-SSA, anti-SSB, antihistones, anti-Sm, anti-RNP, and C3 remained stable until the last visit (Table 6). (Additional data are available from the author upon request.)

Table 6. Summary of clinical and serologic disease activity in SLE patients vaccinated against pandemic 2009 influenza A (H1N1)*
 Day 0 (n = 111)Day 21 (n = 111)Day 42 (n = 110)P
  • *

    Except where indicated otherwise, values are the number (%) of patients. See Table 1 for definitions.

  • By Friedman's nonparametric repeated-measures analysis of variance for continuous values; by the generalized estimating equation method for percentages.

  • Defined using the SELENA–SLEDAI flare instrument.

SELENA–SLEDAI score, median (range)2 (0–32)2 (0–20)2 (0–14)0.09
SELENA–SLEDAI score ≥622 (19.8)21 (18.9)20 (18.2)0.93
≥1 BILAG grade of A or B28 (25.2)28 (25.2)27 (24.6)0.65
Physician's global assessment, median (range)0 (0–3)0 (0–2)0 (0–2)0.08
SLAQ score, median (range)3 (0–38)4 (0–24)3 (0–21)0.11
Positive Farr assay50 (45.1)51 (46.0)47 (42.7)0.66
Farr assay results, median (range) IU/ml7.1 (0–165)6.6 (0–168)6.8 (0–159)0.20
Low C313 (11.7)16 (14.4)14 (12.7)0.53
C3, mean ± SD gm/liter1.00 ± 0.261.00 ± 0.250.98 ± 0.230.13
1 new BILAG grade of A or 2 new BILAG grades of B1 (0.9)0 (0.0)0.92
Severe flare8 (7.2)1 (0.9)0 (0.0)<0.01
Mild/moderate flare3 (2.7)9 (8.1)3 (2.7)0.13

DISCUSSION

During the 2009 H1N1 influenza pandemic, SLE patients in France were advised to receive 2 injections of a monovalent, nonadjuvanted, pandemic 2009 influenza A (H1N1) vaccine (21). The two-dose regimen of hemagglutinin antigen was chosen because there was uncertainty as to whether a single dose would produce a satisfactory immune response (21). SLE patients enrolled in this prospective study were representative of SLE patients encountered in daily practice; the study included both inpatients and outpatients, patients displaying no disease activity and patients displaying a low to high level of disease activity, patients receiving low-dose steroids and those receiving intensive treatment such as high-dose steroids or immunosuppressive drugs, and patients with and without kidney disease. We noticed that a prevaccination hemagglutination-inhibiting titer ≥1:40 was significantly associated with a previous vaccination with the 2009 seasonal trivalent influenza vaccines but was not associated with a history of flu-like symptoms since May 2009. This suggests that the seasonal vaccine may have been involved in the presence of cross-reactive antibodies to 2009 H1N1, as previously reported (3).

The 2 vaccine injections were well tolerated. The pattern and the frequency of local and systemic adverse events were similar to those observed in healthy control subjects (27). The vaccine did not increase disease activity, even in patients with active disease at baseline. Patients with active SLE were included in the study because they were thought to be at high risk of influenza-associated complications (7). The fact that the vaccine did not increase disease activity in these patients is of particular interest because such patients are usually not included in vaccination protocols.

Our data indicate that a two-dose regimen of this vaccine can be used satisfactorily in SLE patients. Indeed, the 3 international guideline requirements recommended by the European Medicines Agency for a pandemic anti-influenza vaccine were met only after the injection of a second dose (35–37). Of note, 1 injection of this vaccine would have been judged acceptable in this SLE patient population for a seasonal, nonpandemic vaccine, for which only 1 of the 3 criteria is required to be met. Because our cohort was heterogeneous, we investigated which parameters influenced the immunogenicity of the vaccine. In multivariate analyses, immunosuppressive treatment was associated with a lower rate of seroconversion and seroprotection and a lower GMR on day 21 and day 42. The use of prednisone or azathioprine has previously been associated with a weaker vaccine immunogenicity (12, 16, 17) or a trend toward a weaker vaccine immunogenicity (9, 24) of seasonal influenza vaccine in SLE patients. Those studies used bivariate analyses and did not assess other important parameters, such as disease activity and routine lupus biomarkers. The other parameter independently associated with impaired vaccine immunogenicity in the present study was a lymphocyte count of ≤1.0 × 109/liter.

The strategy of using 2 injections gave us the opportunity to study the effect of the second dose on vaccine efficacy. In the subgroup of SLE patients who had a lymphocyte count of >1.0 × 109/liter and were not receiving immunosuppressive drugs, the vaccine was fully effective after a single injection, and a second dose of the vaccine did not significantly increase the values for any of the immunogenicity criteria. In this subgroup of SLE patients, the immunogenicity of the vaccine was similar to that described in nonimmunocompromised hosts (27, 38). By contrast, in the subgroups of SLE patients with a weaker vaccine response (i.e., patients who were receiving immunosuppressive treatment or had a lymphocyte count of ≤1.0 × 109/liter), a second dose of the vaccine administered 3 weeks after the first dose was beneficial in terms of immunogenicity.

A booster vaccine dose has been shown to be of interest for seasonal influenza vaccine in elderly patients (39) and in liver transplantation patients (40). Other studies have shown no effect or a marginal effect of a booster vaccine dose on the antibody titers in patients receiving hemodialysis or patients who were severely immunodepressed (41–45). In SLE patients, a booster dose given 1 month after the first vaccination has been shown to increase the GMT of antibody (10, 26), especially in patients treated with prednisone or azathioprine (26). However, the gain in immunogenicity was restricted to the subgroup of SLE patients who were not vaccinated against seasonal influenza during the previous year (26). In that study, a high proportion of patients was already seroprotected before the study started because of previous seasonal influenza vaccination or encounter with influenza virus. This could have partially masked a positive effect of a second (booster) vaccination. In our study, the low prevalence of prevaccination seroprotection allowed us to better analyze the effect of the second vaccination.

Even though the immunogenicity increased with the second injection, the extent of this improvement was too low to meet the seroprotection criterion in the subgroup of patients who were receiving immunosuppressive drugs. Thus, new influenza vaccination strategies should be explored to improve vaccine efficacy in this subgroup. MF59- or AS03-adjuvanted vaccines might elicit a better protective immune response, but the safety of these vaccines is unknown in SLE. Attenuated vaccines pose a risk to immunocompromised patients. More interestingly, other routes of influenza vaccine administration (e.g., transdermal inoculation by needle or patch) or more antigen per dose may also elicit a better immune response (42, 46–49) and should be well tolerated in SLE patients.

In the context of a pandemic, it would not have been ethical to have a group of SLE patients receive a placebo instead of the vaccine. Furthermore, healthy subjects in France were advised to receive 1 dose of adjuvanted vaccine (21). As a consequence, control groups of SLE patients who were not vaccinated, SLE patients who were vaccinated only once, and healthy subjects could not be included in this observational study, which might be a limitation. Yet, SLE patients functioned as their own controls for the purpose of comparing the effects of the first and booster vaccinations, and the study design permitted the analysis of the factors influencing the vaccine immunogenicity, which was our primary objective. Additionally, it allowed us to confirm that the vaccine does not increase disease activity, which has already been shown in other studies (8–19).

In summary, this study contributes new data on influenza vaccine efficacy in SLE patients in the setting of a pandemic, where selective and prioritized allocation of vaccine is necessary. In the subgroups of patients who had a lymphocyte count of >1.0 × 109/liter and were not receiving immunosuppressive drugs, the specific response was comparable to the response seen in healthy individuals, and a single injection was enough to meet all immunogenicity criteria. Patients who were receiving immunosuppressive drugs or had a lymphocyte count of ≤1.0 × 109/liter had a weaker immune response. The second injection significantly increased the efficacy of the vaccine in these subgroups, but not high enough to fulfill all immunogenicity criteria in patients who were receiving an immunosuppressive agent. Thus, new vaccination strategies should be developed for these patients. These data should help to define new strategies to improve influenza vaccine efficacy in SLE patients.

AUTHOR CONTRIBUTIONS

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Drs. Mathian and Devilliers had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Mathian, Devilliers, Krivine, Costedoat-Chalumeau, Hervier, Musset, Autran, Rozenberg, Amoura.

Acquisition of data. Mathian, Devilliers, Krivine, Costedoat-Chalumeau, Haroche, Boutin-Le Thi Huong, Wechsler, Miyara, Morel, Le Corre, Piette, Musset, Autran, Rozenberg, Amoura.

Analysis and interpretation of data. Mathian, Devilliers, Krivine, Arnaud, Autran, Rozenberg, Amoura.

Acknowledgements

The authors thank Ouidede Lajili and Laurent Dufat for technical assistance, Alain Mallet and Jean-Louis Golmards for assistance with the study design, and Romain Despalins for advice regarding statistical analysis.

Ancillary